11 years ago2011-05-23 Joe Wild <joseph.wild@nokia.com> qtwebkit-2.2-week20
commit-queue [Mon, 23 May 2011 08:14:10 +0000 (08:14 +0000)]
2011-05-23  Joe Wild  <joseph.wild@nokia.com>

        Reviewed by Andreas Kling.

        [Qt] The Qt WebKit Symbian .def file needs to be updated so Symbian will build

        * symbian/eabi/QtWebKitu.def:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@87060 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-22 Hui Huang <hui.2.huang@nokia.com>, Yi Shen <yi.4.shen@nokia.com>
Yi Shen [Sun, 22 May 2011 22:18:26 +0000 (22:18 +0000)]
2011-05-22  Hui Huang  <hui.2.huang@nokia.com>, Yi Shen  <yi.4.shen@nokia.com>

        Reviewed by Laszlo Gombos.

        [Qt] Upstream Symbian platform plugin

        Upstream Symbian platform plugin with html5 video player.

        * symbian/platformplugin/HTML5VideoPlugin.cpp: Added.
        * symbian/platformplugin/HTML5VideoPlugin.h: Added.
        * symbian/platformplugin/HTML5VideoWidget.cpp: Added.
        * symbian/platformplugin/HTML5VideoWidget.h: Added.
        * symbian/platformplugin/OverlayWidget.cpp: Added.
        * symbian/platformplugin/OverlayWidget.h: Added.
        * symbian/platformplugin/PlayerButton.cpp: Added.
        * symbian/platformplugin/PlayerButton.h: Added.
        * symbian/platformplugin/PlayerLabel.cpp: Added.
        * symbian/platformplugin/PlayerLabel.h: Added.
        * symbian/platformplugin/WebPlugin.cpp:
        * symbian/platformplugin/images: Added.
        * symbian/platformplugin/images/button_cannotplay.png: Added.
        * symbian/platformplugin/images/button_close.png: Added.
        * symbian/platformplugin/images/button_pause.png: Added.
        * symbian/platformplugin/images/button_play.png: Added.
        * symbian/platformplugin/images/button_sound_off.png: Added.
        * symbian/platformplugin/images/button_sound_on.png: Added.
        * symbian/platformplugin/images/loading_buffering_1.png: Added.
        * symbian/platformplugin/images/loading_buffering_2.png: Added.
        * symbian/platformplugin/images/loading_buffering_3.png: Added.
        * symbian/platformplugin/images/loading_buffering_4.png: Added.
        * symbian/platformplugin/platformplugin.pro:
        * symbian/platformplugin/platformplugin.qrc: Added.
        * symbian/platformplugin/qss: Added.
        * symbian/platformplugin/qss/OverlayWidget.qss: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@87044 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-21 Dirk Schulze <krit@webkit.org>
Dirk Schulze [Sun, 22 May 2011 06:58:30 +0000 (06:58 +0000)]
2011-05-21  Dirk Schulze  <krit@webkit.org>

        Reviewed by Darin Adler.

        REGRESSION(r66731): pointer-events are broken in some cases

        The SVGSVGElement shouldn't be the target of a mouse event, if its pointer-events attribute is set
        to 'none'. This matches the behavior on Firefox where an embedded SVG element is the target of an event,
        if none of its childs caught the event. This is the case for all pointer-events other than 'none'.

        Tests: svg/custom/pointer-events-on-svg-with-pointer.xhtml

        * rendering/svg/RenderSVGRoot.cpp:
2011-05-21  Dirk Schulze  <krit@webkit.org>

        Reviewed by Darin Adler.

        REGRESSION(r66731): pointer-events are broken in some cases

        Check that an embedded SVG element doesn't have a pointer, if pointer-events is set to 'none'.

        * svg/custom/pointer-events-on-svg-with-pointer-expected.txt: Added.
        * svg/custom/pointer-events-on-svg-with-pointer.xhtml: Added.
        * svg/custom/pointer-events-on-svg-without-pointer-expected.txt: Added.
        * svg/custom/pointer-events-on-svg-without-pointer.xhtml: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@87027 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago<rdar://problem/9479926> REGRESSION (r82144): Icon overlaps text in Twitter message...
Dan Bernstein [Sun, 22 May 2011 04:44:27 +0000 (04:44 +0000)]
<rdar://problem/9479926> REGRESSION (r82144): Icon overlaps text in Twitter message dialog

Reviewed by Darin Adler.


Test: fast/block/positioning/start-ignoring-before.html

* rendering/RenderBlockLineLayout.cpp:
(WebCore::TrailingObjects::updateMidpointsForTrailingBoxes): Added a boolean parameter saying
whether to merge the first trailing space with the line break.
(WebCore::RenderBlock::LineBreaker::nextLineBreak): When adding a midpoint behind the current
character, account for trailing positioned objects that occurred after the midpoint by calling
updateMidpointsForTrailingBoxes(), which adds midpoints for them.


* fast/block/positioning/start-ignoring-before-expected.png: Added.
* fast/block/positioning/start-ignoring-before-expected.txt: Added.
* fast/block/positioning/start-ignoring-before.html: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@87026 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-20 Andy Estes <aestes@apple.com>
Andy Estes [Sat, 21 May 2011 01:07:16 +0000 (01:07 +0000)]
2011-05-20  Andy Estes  <aestes@apple.com>

        Reviewed by Darin Adler.

        REGRESSION (r70748): WebKit cannot play videos created by Podcast Producer.

        Test that an object element with a non-empty classid, a valid MIME
        type and no fallback content is allowed to load.

        * fast/replaced/object-with-non-empty-classid-triggers-fallback-expected.txt:
        * fast/replaced/object-with-non-empty-classid-triggers-fallback.html:
2011-05-20  Andy Estes  <aestes@apple.com>

        Reviewed by Darin Adler.

        REGRESSION (r70748): WebKit cannot play videos created by Podcast Producer.

        Podcast Producer uses an object tag with a classid attribute to embed
        QuickTime Player into a page. In r70748, we changed our behavior to
        render the object's fallback content when a non-empty classid is
        encountered, per HTML5. Since Podcast Producer videos have no fallback
        content, this change in behavior causes the video to fail to load.

        Since the object tag has a valid type attribute, we would be able to
        load it if weren't for the non-empty classid. This patch changes our
        policy to allow objects with non-empty classids if there is no fallback
        content. We still continue to prefer fallback content if it exists,

        * html/HTMLObjectElement.cpp:
        (WebCore::HTMLObjectElement::hasValidClassId): Treat a non-empty
        classid as valid if the object has no fallback content.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@87007 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago<rdar://problem/9472883> and https://bugs.webkit.org/show_bug.cgi?id=61203
Brady Eidson [Fri, 20 May 2011 18:46:19 +0000 (18:46 +0000)]
<rdar://problem/9472883> and https://bugs.webkit.org/show_bug.cgi?id=61203
Horrendous bug in callOnMainThreadAndWait

Reviewed by Sam Weinig.

* wtf/MainThread.cpp:
(WTF::dispatchFunctionsFromMainThread): Before signaling the background thread with the
  syncFlag condition, reacquire the mutex first.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86972 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-20 Xan Lopez <xlopez@igalia.com>
Xan Lopez [Fri, 20 May 2011 15:55:16 +0000 (15:55 +0000)]
2011-05-20  Xan Lopez  <xlopez@igalia.com>

        Reviewed by Oliver Hunt.

        JIT requires VM overcommit (particularly on x86-64), Linux does not by default support this without swap?

        Use the MAP_NORESERVE flag for mmap on Linux to skip the kernel
        check of the available memory. This should give us an
        overcommit-like behavior in most systems, which is what we want.

        * wtf/OSAllocatorPosix.cpp:
        (WTF::OSAllocator::reserveAndCommit): pass MAP_NORSERVE to mmap.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86957 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-20 Dirk Schulze <krit@webkit.org>
Dirk Schulze [Fri, 20 May 2011 08:01:19 +0000 (08:01 +0000)]
2011-05-20  Dirk Schulze  <krit@webkit.org>

        Reviewed by Eric Seidel.

        SVG Large curve path segment OOM crash

        Limit the depth of repeatedly splitting a segment on length calculation to 20. The limitation
        is necessary for very big segments that would be splitter into millions of parts otherwise.
        The limitation just cause a less accurate approximation.
        At the moment the limit is fixed to 20. This is comparable with splitting the segment into
        ~1 million parts as a worst case. We might want to be more flexible later.

        Test: svg/custom/path-getTotalLength-on-big-segment-crash.svg

        * platform/graphics/PathTraversalState.cpp:
        * platform/graphics/PathTraversalState.h:
2011-05-20  Dirk Schulze  <krit@webkit.org>

        Reviewed by Eric Seidel.

        SVG Large curve path segment OOM crash

        Added a test to verify, that the browser does not crash on calculating the total length on big segments.
        It makes no sense to add the result of getTotalLength(), since they differ a lot bewteen platforms.
        This is caused by the platform graphic libraries. See comment #9 on the bug.

        * svg/custom/path-getTotalLength-on-big-segment-crash-expected.txt: Added.
        * svg/custom/path-getTotalLength-on-big-segment-crash.svg: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86928 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-19 Andrew Wilson <atwilson@chromium.org>
Drew Wilson [Thu, 19 May 2011 22:29:24 +0000 (22:29 +0000)]
2011-05-19  Andrew Wilson  <atwilson@chromium.org>

        Reviewed by Darin Adler.

        MessagePortArray cloning code needs to verify source before copying

        * fast/events/message-port-multi-expected.txt:
        * fast/events/resources/message-port-multi.js:
        Added test for "passing an array with an item at a really large index" to postMessage().
2011-05-19  Andrew Wilson  <atwilson@chromium.org>

        Reviewed by Darin Adler.

        MessagePortArray cloning code needs to verify source before copying.

        * bindings/js/JSMessagePortCustom.cpp:
        Changed code to not pre-allocate the destination array.
        * bindings/v8/custom/V8MessagePortCustom.cpp:
        Changed code to not pre-allocate the destination array.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86899 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-06 Alexis Menard <alexis.menard@openbossa.org>
Alexis Menard [Fri, 6 May 2011 20:29:58 +0000 (20:29 +0000)]
2011-05-06  Alexis Menard  <alexis.menard@openbossa.org>

        Reviewed by Darin Adler.

        Fix two warnings of unused variables.

        Remove two unused local variable from the code.

        No new tests, the existing ones should cover.

        * rendering/RenderFlexibleBox.cpp:
        * svg/animation/SVGSMILElement.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85973 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-18 Alexis Menard <alexis.menard@openbossa.org>, Simon Hausmann <simon...
commit-queue [Wed, 18 May 2011 21:26:03 +0000 (21:26 +0000)]
2011-05-18  Alexis Menard  <alexis.menard@openbossa.org>, Simon Hausmann  <simon.hausmann@nokia.com>

        Reviewed by Eric Carlson.

        MediaElements fails to load the data in some cases.

        This test creates an hidden video element and make sure the loading works.

        * http/tests/media/media-can-load-when-hidden-expected.txt: Added.
        * http/tests/media/media-can-load-when-hidden.html: Added.
2011-05-18  Alexis Menard  <alexis.menard@openbossa.org>, Simon Hausmann  <simon.hausmann@nokia.com>

        Reviewed by Eric Carlson.

        MediaElements fails to load the data in some cases.

        WebKitWebSourceGStreamer is the interface between WebKit and GStreamer
        that uses the ResourceHandle API to request data and pass it down. For
        our builds it is absolutely essential that we have a NetworkingContext
        available there, in order to get access to the QNetworkAccessManager.
        No access means we basically cannot load the video. The WebSource gains
        access to the NetworkingContext through a WebCore::Frame pointer it has.

        MediaPlayerPrivateGStreamer is responsible for propagating a pointer of
        the WebCore::Frame to the WebKitWebSource in
        mediaPlayerPrivateSourceChangedCallback. In there we used the MediaPlayer's
        frameView() accessor to access the frame. However the frameView() member
        is only set through the render tree's RenderVideo, which is rather unreliable
        given that some sites create "fake" video tags initially that only become
        visible later (or never).

        A more reliable way is to simply use the document of the MediaPlayerClient,
        which is provided at constructor time.

        Test: http/tests/media/media-can-load-when-hidden.html

        * platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86790 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-11 Igor Oliveira <igor.oliveira@openbossa.org>
commit-queue [Thu, 12 May 2011 03:38:59 +0000 (03:38 +0000)]
2011-05-11  Igor Oliveira  <igor.oliveira@openbossa.org>

        Reviewed by Eric Seidel.

        WebKit does not build with GCCE

        Allow compile WebKit with GCCE

        * wtf/Alignment.h:
        * wtf/Platform.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86304 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-19 Andrew Wason <rectalogic@rectalogic.com>
commit-queue [Thu, 19 May 2011 18:39:14 +0000 (18:39 +0000)]
2011-05-19  Andrew Wason  <rectalogic@rectalogic.com>

        Reviewed by Darin Adler.

        Fix GraphicsContext3DQt.cpp compile error

        * platform/graphics/qt/GraphicsContext3DQt.cpp:
         clear() m_internal OwnPtr.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86861 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-19 Ryosuke Niwa <rniwa@webkit.org>
Ryosuke Niwa [Thu, 19 May 2011 17:22:24 +0000 (17:22 +0000)]
2011-05-19  Ryosuke Niwa  <rniwa@webkit.org>

        Reviewed by Darin Adler.

        REGRESSION (r83322): Many crashes in Mail.app in WebCore::Node::nodeIndex

        Added a test to ensure WebKit does not crash when inserting a content immediately after
        a styled element inside a Mail blockquote. Regrettably the expected result is incorrect,
        but it matches the behavior of WebKit before r83322.

        * editing/pasteboard/5065605-expected.txt: Reintroduced redundant style spans.
        * editing/pasteboard/paste-text-011-expected.txt: Ditto.
        * platform/chromium-win/editing/pasteboard/paste-text-011-expected.txt: Ditto.
        * editing/pasteboard/paste-after-inline-style-element-expected.txt: Added.
        * editing/pasteboard/paste-after-inline-style-element.html: Added.
2011-05-19  Ryosuke Niwa  <rniwa@webkit.org>

        Reviewed by Darin Adler.

        REGRESSION (r83322): Many crashes in Mail.app in WebCore::Node::nodeIndex

        The crash was caused by ReplaceSelectionCommand's inserting content into a middle of the paragraph
        being moved when the insertion position's container node is the node to split to. Fixed the crash
        by not changing the insertion position in such a case.

        Unfortunately, this fix caused markup to bloat in some tests but we'll take this regression since
        it's much better than crashing.

        Test: editing/pasteboard/paste-after-inline-style-element.html

        * editing/ReplaceSelectionCommand.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86852 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-19 Emil A Eklund <eae@chromium.org>
Emil A Eklund [Thu, 19 May 2011 09:21:58 +0000 (09:21 +0000)]
2011-05-19  Emil A Eklund  <eae@chromium.org>

        Reviewed by Alexey Proskuryakov.

        REGRESSION (r80808): Multiple <select> - Selection reset to first element from multiple selected ones

        * fast/dom/HTMLSelectElement/change-multiple-preserve-selection-expected.txt:
        * fast/dom/HTMLSelectElement/change-multiple-preserve-selection.html:
2011-05-19  Emil A Eklund  <eae@chromium.org>

        Reviewed by Alexey Proskuryakov.

        REGRESSION (r80808): Multiple <select> - Selection reset to first element from multiple selected ones

        * html/HTMLSelectElement.cpp:
        Don't restore selection if the multiple attribute hasn't changed.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86827 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-18 Yi Shen <yi.4.shen@nokia.com>
Ademar de Souza Reis Jr [Thu, 19 May 2011 20:16:21 +0000 (17:16 -0300)]
2011-05-18  Yi Shen  <yi.4.shen@nokia.com>

        Reviewed by Andreas Kling.

        [Qt] Enterkey to go to Newline does not work in the text area(in HTML form)

        Unskip the fast/events/onsearch-enter.html test.

        * platform/qt/Skipped:
2011-05-18  Yi Shen  <yi.4.shen@nokia.com>

        Reviewed by Andreas Kling.

        [Qt] Enterkey to go to Newline does not work in the text area(in HTML form)

        Fill the missing key text for the EnterKey event.

        Tests: fast/events/onsearch-enter.html

        * platform/qt/PlatformKeyboardEventQt.cpp:
2011-05-18  Yi Shen  <yi.4.shen@nokia.com>

        Reviewed by Andreas Kling.

        [Qt] Enterkey to go to Newline does not work in the text area(in HTML form)

        Remove the implementation of the handleInputMethodKeydown, which introduces
        a regression(r82243) on Linux. Also, add more Api tests for the EnterKey event.

        * WebCoreSupport/EditorClientQt.cpp:
        (WebCore::EditorClientQt::handleInputMethodKeydown): Remove implementation.
        * tests/qwebpage/tst_qwebpage.cpp:
        (tst_QWebPage::inputMethods): Add more tests.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86798 268f45cc-cd09-0410-ab3c-d52691b4dbfc



11 years ago2011-05-18 Oliver Hunt <oliver@apple.com>
Oliver Hunt [Wed, 18 May 2011 20:41:54 +0000 (20:41 +0000)]
2011-05-18  Oliver Hunt  <oliver@apple.com>

        Reviewed by Sam Weinig.

        JSGlobalObject and some others do GC allocation during initialization, which can cause heap corruption

        Remove the Structure-free JSGlobalObject constructor and instead always
        pass the structure into the JSGlobalObject constructor.
        Stop DebuggerActivation creating a new structure every time, and simply
        use a single shared structure held by the GlobalData.

        * API/JSContextRef.cpp:
        * debugger/DebuggerActivation.cpp:
        * jsc.cpp:
        * runtime/JSGlobalData.cpp:
        * runtime/JSGlobalData.h:
        * runtime/JSGlobalObject.h:
2011-05-18  Oliver Hunt  <oliver@apple.com>

        Reviewed by Sam Weinig.

        JSGlobalObject and some others do GC allocation during initialization, which can cause heap corruption

        Rather than having Constructor objects create their structure
        as part of initialisation, we now pass their expected structure
        in as an argument.  This required fixing the few custom Constructors
        and the code generator.

        * bindings/js/JSAudioConstructor.cpp:
        * bindings/js/JSAudioConstructor.h:
        * bindings/js/JSDOMGlobalObject.h:
          Pass the Constructor objects structure in as an argument
        * bindings/js/JSImageConstructor.cpp:
        * bindings/js/JSImageConstructor.h:
        * bindings/js/JSOptionConstructor.cpp:
        * bindings/js/JSOptionConstructor.h:
        * bindings/scripts/CodeGeneratorJS.pm:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86785 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-18 Caio Marcelo de Oliveira Filho <caio.oliveira@openbossa.org>
Caio Marcelo de Oliveira Filho [Wed, 18 May 2011 14:12:56 +0000 (14:12 +0000)]
2011-05-18  Caio Marcelo de Oliveira Filho  <caio.oliveira@openbossa.org>

        Reviewed by Andreas Kling.

        [Qt] Fix tst_QWebFrame::getSetStaticProperty() autotest

        The code for converting objects to QVariantMap was causing exception,
        that was "leaking" to the next evaluation. One situation was reading
        the property 'localStorage' when we do not have a proper security
        origin, which throws a SECURITY_ERR.

        Now, we will simply not include on the QVariantMap those properties,
        and make sure that we clean the exception if necessary.

        * bridge/qt/qt_runtime.cpp:
        Extracted function that performs conversion from JSObject to a QVariantMap. This
        functions makes sure that exception is clean after its execution.

        Use the previous function. Add a comment explaining the choice of distance value.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86758 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-17 Caio Marcelo de Oliveira Filho <caio.oliveira@openbossa.org>
Caio Marcelo de Oliveira Filho [Tue, 17 May 2011 19:35:37 +0000 (19:35 +0000)]
2011-05-17  Caio Marcelo de Oliveira Filho  <caio.oliveira@openbossa.org>

        Reviewed by Andreas Kling.

        [Qt] Simplify syntax in test code to make prepare-ChangeLog less confused

        Backslash to escape newlines was confusing both prepare-ChangeLog and the
        QtCreator highlight system.

        * tests/qwebframe/tst_qwebframe.cpp:
        Remove usage of backslash to escape newlines in string literal.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86697 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-18 Abhishek Arya <inferno@chromium.org>
Abhishek Arya [Wed, 18 May 2011 19:30:48 +0000 (19:30 +0000)]
2011-05-18  Abhishek Arya  <inferno@chromium.org>

        Reviewed by Beth Dakin.

        Tests that we do not crash when prematurely calling removeChild,
        followed by destroy call on table caption.

        * fast/table/table-captions-child-visible-crash-expected.txt: Added.
        * fast/table/table-captions-child-visible-crash.html: Added.
2011-05-18  Abhishek Arya  <inferno@chromium.org>

        Reviewed by Beth Dakin.

        Remove removeChild on table caption since destroy call
        already does that.

        Test: fast/table/table-captions-child-visible-crash.html

        * rendering/RenderTable.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86781 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago[Qt] Redirection of HTTP POST (3xx) incorrectly includes original POST data
Luiz Agostini [Tue, 17 May 2011 19:00:37 +0000 (19:00 +0000)]
[Qt] Redirection of HTTP POST (3xx) incorrectly includes original POST data

Reviewed by Kenneth Rohde Christiansen.


Makes sure that the HTTP headers Content-type and Content-length are not included in
the requests that do not have any content.

Tests: http/tests/navigation/post-301-response.html

* platform/network/qt/QNetworkReplyHandler.cpp:


These new tests check that no POST content is sent to the new URL after receiving http
status codes 301, 302 and 303, and checks that the POST content is sent to the new URL
after receiving a 307 http status code.

* http/tests/navigation/post-301-response-expected.txt: Added.
* http/tests/navigation/post-301-response.html: Added.
* http/tests/navigation/post-302-response-expected.txt: Added.
* http/tests/navigation/post-302-response.html: Added.
* http/tests/navigation/post-303-response-expected.txt: Added.
* http/tests/navigation/post-303-response.html: Added.
* http/tests/navigation/post-307-response-expected.txt: Added.
* http/tests/navigation/post-307-response.html: Added.
* http/tests/navigation/resources/redirected-post-request-contents.php: Added.
* http/tests/navigation/resources/redirection-response.php: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86693 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-17 Andreas Kling <kling@webkit.org>
Andreas Kling [Tue, 17 May 2011 18:37:32 +0000 (18:37 +0000)]
2011-05-17  Andreas Kling  <kling@webkit.org>

        Reviewed by Benjamin Poulain.

        [Qt] GraphicsLayerQtImpl: Remove an unused variable.

        * platform/graphics/qt/GraphicsLayerQt.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86691 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-17 Sam Magnuson <smagnuson@netflix.com>
Andreas Kling [Tue, 17 May 2011 17:16:26 +0000 (17:16 +0000)]
2011-05-17  Sam Magnuson  <smagnuson@netflix.com>

        Reviewed by Kenneth Rohde Christiansen.

        [Qt] Node that have both an opacity and a transform animation on them seem not to fire.

        Test: compositing/animation/busy-indicator.html

        * platform/graphics/qt/GraphicsLayerQt.cpp:
2011-05-17  Sam Magnuson  <smagnuson@netflix.com>

        Reviewed by Kenneth Rohde Christiansen.

        [Qt] Node that have both an opacity and a transform animation on them seem not to fire.

        * compositing/animation/busy-indicator-no.png: Added.
        * compositing/animation/busy-indicator.html: Added.
        * compositing/animation/busy-indicator.png: Added.
        * compositing/animation/busy-indicator-expected.txt: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86680 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-16 Adam Barth <abarth@webkit.org>
Adam Barth [Tue, 17 May 2011 01:42:14 +0000 (01:42 +0000)]
2011-05-16  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        Remove bogus ASSERT in Document::setCompatibilityMode

        This test case triggers the ASSERT.

        * fast/parser/append-child-followed-by-document-write-expected.txt: Added.
        * fast/parser/append-child-followed-by-document-write.html: Added.
2011-05-16  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        Remove bogus ASSERT in Document::setCompatibilityMode

        The ASSERT is invalid when the parser is in the initial state and the
        document is non-empty, which is strange but not impossible.

        Test: fast/parser/append-child-followed-by-document-write.html

        * dom/Document.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86649 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-15 Geoffrey Garen <ggaren@apple.com>
Geoffrey Garen [Mon, 16 May 2011 05:10:02 +0000 (05:10 +0000)]
2011-05-15  Geoffrey Garen  <ggaren@apple.com>

        Reviewed by Gavin Barraclough.

        Global object is recreated on teardown, for no good reason

        (Another partial fix for <rdar://problem/9417875> REGRESSION: SunSpider
        ~7% slower in browser than on command line (was 17%))

        I'm basically rolling out http://trac.webkit.org/changeset/49786 because

        (a) it created this performance problem


        (b) a more complete fix, which obsoletes http://trac.webkit.org/changeset/49786,
        was committed in http://trac.webkit.org/changeset/53439.

        Tested with the file attached to https://bugs.webkit.org/show_bug.cgi?id=29832.

        * page/Frame.cpp:
        (WebCore::Frame::~Frame): Don't create a new window every time we destroy
        a frame.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86523 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoPartial fix for <rdar://problem/9417875> REGRESSION: SunSpider ~17% slower
Geoffrey Garen [Sun, 15 May 2011 22:31:31 +0000 (22:31 +0000)]
Partial fix for <rdar://problem/9417875> REGRESSION: SunSpider ~17% slower
in browser than on command line

Reviewed by Maciej Stachowiak.

This patch fixes a few issues in generated code that could unreasonably
prolong object lifetimes.

* heap/Heap.cpp:
(JSC::Heap::collectAllGarbage): Throw away all function code before doing
a major collection. We want to clear polymorphic caches, since they can
keep alive large object graphs that have gone "stale". For the same reason,
but to a lesser extent, we also want to clear linked functions and other
one-off caches.

This has the side-benefit of reducing memory footprint from run-once
functions, and of allowing predictions and caches that have failed to

Eventually, if compilation costs rise far enough, we may want a more
limited strategy for de-specializing code without throwing it away
completely, but this works for now, and it's the simplest solution.

* jit/JITStubs.cpp:
* jit/JITStubs.h:
* runtime/JSFunction.cpp: Made the host function stub cache weak --
otherwise it's effectively a memory leak that can seriously fragment the
GC and JIT heaps.

(JSC::JSFunction::visitChildren): Cleared up some comments that confused
me when working with this code.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86510 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-13 Mikhail Naganov <mnaganov@chromium.org>
Mikhail Naganov [Fri, 13 May 2011 22:00:05 +0000 (22:00 +0000)]
2011-05-13  Mikhail Naganov  <mnaganov@chromium.org>

        Web Inspector: Unreviewed image glyph position fix in CSS after r85588.

        * inspector/front-end/heapProfiler.css:
        (.heapshot-help-status-bar-item .glyph):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86463 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago Reviewed by Joseph Pecoraro.
Alexey Proskuryakov [Fri, 13 May 2011 20:59:57 +0000 (20:59 +0000)]
    Reviewed by Joseph Pecoraro.

        Hide appcache status bar items

        We have a number of non-trivial bugs that make these more misleading than helpful to developers.

        * inspector/front-end/ApplicationCacheItemsView.js: (WebInspector.ApplicationCacheItemsView):
        Just hide the items for now. Removing all support code would be complicated, and not helpful.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86460 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-13 Abhishek Arya <inferno@chromium.org>
Abhishek Arya [Fri, 13 May 2011 18:17:12 +0000 (18:17 +0000)]
2011-05-13  Abhishek Arya  <inferno@chromium.org>

        Reviewed by Dave Hyatt.

        Regression(r75823): Revert initialization back to the first RenderBlock
        in the parent chain.

        Test: fast/block/float/float-not-removed-crash.html

        * rendering/RenderBox.cpp:
        In r75823, I forgot to initialize parentBlock, when the original code
        did and Hyatt did mention in c#5. Now, initializing parentBlock to the
        first RenderBlock in the parent chain.
2011-05-13  Abhishek Arya  <inferno@chromium.org>

        Reviewed by Dave Hyatt.

        Tests that we do not crash when trying to access a deleted float.

        * fast/block/float/float-not-removed-crash-expected.txt: Added.
        * fast/block/float/float-not-removed-crash.html: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86448 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoTell ScrollView's child Widgets that their frame rects have changed when its own...
Adam Roben [Fri, 13 May 2011 17:24:55 +0000 (17:24 +0000)]
Tell ScrollView's child Widgets that their frame rects have changed when its own frame rect changes

r79167 moved some code from setFrameRect to setBoundsSize, including a call to
frameRectsChanged. This was done because positionScrollbarLayers, which is called by
frameRectsChanged, only needs to be called when the bounds change, not when the frame rect
changes. But the recursive calls inside frameRectsChanged *do* need to be called when the
frame rect changes.

This patch moves the positionScrollbarLayers call out of frameRectsChanged, since it needs
to be called at different times from frameRectsChanged. Then it restores the
frameRectsChanged call to setFrameRect, which fixes the bug.

Test: platform/win/plugins/iframe-inside-overflow.html

Fixes <http://webkit.org/b/60194> <rdar://problem/9383760> REGRESSION (r79167): Windowed
plugins in Google Reader don't move when the article list is scrolled

Reviewed by Dan Bernstein.


* platform/ScrollView.cpp:
(WebCore::ScrollView::updateScrollbars): Added a positionScrollbarLayers call here now that
frameRectsChanged doesn't call it for us. Also added a FIXME because it seems strange to
call frameRectsChanged here when our frame rect hasn't changed.
(WebCore::ScrollView::setFrameRect): Added back the frameRectsChanged call that was removed
in r79167.
(WebCore::ScrollView::setBoundsSize): Replaced a frameRectsChanged call with a call to
positionScrollbarLayers. We were only calling frameRectsChanged here in order to get
positionScrollbarLayers to be called.
(WebCore::ScrollView::frameRectsChanged): Removed the call to positionScrollbarLayers. All
callers of frameRectsChanged have been updated to call positionScrollbarLayers if needed.


Add a plugin test that dumps the plugin window's rect

* DumpRenderTree/TestNetscapePlugIn/Tests/win/DumpWindowRect.cpp: Added.
(DumpWindowRect::DumpWindowRect): Just call up to the base class.
(DumpWindowRect::performWindowGeometryTest): Find our window rect relative to the test
harness window and log it.

* DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Added DumpWindowRect.


Add a test for windowed plugins inside iframes inside scrolled overflow areas

* platform/win/plugins/iframe-inside-overflow-expected.txt: Added.
* platform/win/plugins/iframe-inside-overflow.html: Added.
(loaded): Scrolls the div and tells the plugin to start its test (which will cause its
window rect to be logged).
* platform/win/plugins/resources/dump-window-rect-iframe.html: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86442 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-07 Dawit Alemayehu <adawit@kde.org>
Daniel Bates [Sat, 7 May 2011 22:59:09 +0000 (22:59 +0000)]
2011-05-07  Dawit Alemayehu  <adawit@kde.org>

        Reviewed by Daniel Bates.

        Fix compile with GCC 4.6.0

        Remove unused local variable from code.

        * runtime/StringPrototype.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86008 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-10 Chang Shu <cshu@webkit.org> qtwebkit-2.2-week19
commit-queue [Tue, 10 May 2011 17:36:41 +0000 (17:36 +0000)]
2011-05-10  Chang Shu  <cshu@webkit.org>

        Reviewed by Darin Adler.

        Convert editing/inserting/typing-tab-designmode-forms.html to dumpAsText

        * editing/inserting/typing-tab-designmode-forms-expected.txt: Added.
        * editing/inserting/typing-tab-designmode-forms.html:
        * platform/chromium-linux/editing/inserting/typing-tab-designmode-forms-expected.png: Removed.
        * platform/chromium-mac-leopard/editing/inserting/typing-tab-designmode-forms-expected.png: Removed.
        * platform/chromium-win/editing/inserting/typing-tab-designmode-forms-expected.png: Removed.
        * platform/chromium-win/editing/inserting/typing-tab-designmode-forms-expected.txt: Removed.
        * platform/gtk/editing/inserting/typing-tab-designmode-forms-expected.png: Removed.
        * platform/gtk/editing/inserting/typing-tab-designmode-forms-expected.txt: Removed.
        * platform/mac-leopard/editing/inserting/typing-tab-designmode-forms-expected.png: Removed.
        * platform/mac/editing/inserting/typing-tab-designmode-forms-expected.png: Removed.
        * platform/mac/editing/inserting/typing-tab-designmode-forms-expected.txt: Removed.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86166 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-09 Igor Oliveira <igor.oliveira@openbossa.org>
commit-queue [Mon, 9 May 2011 21:56:06 +0000 (21:56 +0000)]
2011-05-09  Igor Oliveira  <igor.oliveira@openbossa.org>

        Reviewed by Andreas Kling.

        [Qt] DRT eventSender.addTouchPoint method is not setting the start position

        When adding a new touch point, the start position of this point needs to be set.

        * DumpRenderTree/qt/EventSenderQt.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86090 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-11 Alexis Menard <alexis.menard@openbossa.org>
Alexis Menard [Thu, 12 May 2011 01:20:34 +0000 (01:20 +0000)]
2011-05-11  Alexis Menard  <alexis.menard@openbossa.org>

        Reviewed by Kenneth Rohde Christiansen.

        [Qt] Implements a disable appearance for Media Elements of Qt port.

        Implements a disable appearance for the media controls of the Qt port
        when the media is not yet available.

        * platform/qt/RenderThemeQt.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86299 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-16 Andreas Kling <kling@webkit.org>
Ademar de Souza Reis Jr [Mon, 16 May 2011 21:01:39 +0000 (18:01 -0300)]
2011-05-16  Andreas Kling  <kling@webkit.org>

        Reviewed by Kenneth Rohde Christiansen.

        REGRESSION(r83820): [Qt] Accelerated compositing no longer works in QGraphicsWebView.

        Don't set the ItemClipsChildrenToShape flag for the root platform layer,
        since that is now the overflow controls layer. The clip layer, which was
        previously the root platform layer, already gets the flag by way of
        the GraphicsLayer mask-to-bounds flag.

        * WebCoreSupport/PageClientQt.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86582 268f45cc-cd09-0410-ab3c-d52691b4dbfc



11 years ago2011-05-11 Chang Shu <cshu@webkit.org>
Ademar de Souza Reis Jr [Mon, 16 May 2011 20:50:19 +0000 (17:50 -0300)]
2011-05-11  Chang Shu  <cshu@webkit.org>

        Reviewed by Antonio Gomes.

        [Qt] Tab not working in editing/inserting/typing-tab-designmode-forms.html

        Unskip passed tests.

        * platform/qt/Skipped:
2011-05-11  Chang Shu  <cshu@webkit.org>

        Reviewed by Antonio Gomes.

        [Qt] Tab not working in editing/inserting/typing-tab-designmode-forms.html

        PlatformKeyboardEvent.m_text should be set to "\t" instead of leaving as null.
        The value is checked at EventHandler.cpp:2527 (r86166).

        Function keyTextForKeyEvent is partially implemented and can be enhanced in
        a need-base.

        * platform/qt/PlatformKeyboardEventQt.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86243 268f45cc-cd09-0410-ab3c-d52691b4dbfc



11 years agoYet another test expectation change after CONFIG(production)
Ademar de Souza Reis Jr [Mon, 16 May 2011 20:42:16 +0000 (17:42 -0300)]
Yet another test expectation change after CONFIG(production)

Hopefully the last one :-)

11 years agoAnother round of rebasing and skipping of tests after CONFIG(production)
Ademar de Souza Reis Jr [Mon, 16 May 2011 19:48:35 +0000 (16:48 -0300)]
Another round of rebasing and skipping of tests after CONFIG(production)

One more round may be necessary, let's see how the bots behave now.

11 years agoAdd a few more XSLT and Web Timing tests to the Skip list
Ademar de Souza Reis Jr [Mon, 16 May 2011 18:29:28 +0000 (15:29 -0300)]
Add a few more XSLT and Web Timing tests to the Skip list

Leftover from previous commit which added CONFIG(production), thus
disabling XSLT and Web Timing from our qtwebkit-2.2 branch.

There may be a few more tests failing because they depend on XSL
or Web Timing. They will be fixed/skipped in a next commit.

11 years agoTurn on CONFIG(production) on qtwebkit-2.2
Ademar de Souza Reis Jr [Mon, 16 May 2011 14:06:38 +0000 (11:06 -0300)]
Turn on CONFIG(production) on qtwebkit-2.2

This disables XSLT and Web Timing, which are buggy in QtWebKit
and therefore disabled in production branches.

See https://bugs.webkit.org/show_bug.cgi?id=60824 for details

Also adjust LayoutTests results and skip list (speculative, we
may have to skip or rebase a few more).

Patch partially based on changes from the 2.1 branch by Simon Hausmann.

11 years ago2011-05-16 Ademar de Souza Reis Jr. <ademar.reis@openbossa.org>
Ademar de Souza Reis Junior [Mon, 16 May 2011 14:49:17 +0000 (14:49 +0000)]
2011-05-16  Ademar de Souza Reis Jr.  <ademar.reis@openbossa.org>

        Reviewed by Csaba Osztrogonác.

        [Qt] Tools.pro misses include(common.pri)

        It includes features.pri, which depends on common.pri (the build is
        not broken today with the default options, but this is the right
        thing to do).

        * Tools.pro:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86572 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago[Qt] Fix build on MSVC.
Jocelyn Turcotte [Fri, 13 May 2011 16:55:26 +0000 (16:55 +0000)]
[Qt] Fix build on MSVC.

Reviewed by nobody, build fix.

qobject_cast<> requires the class to be exported on MSVC,
removing it since the code that needs it is commented out for
the same reason.

* tests/qdeclarativewebview/tst_qdeclarativewebview.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86438 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago[Qt] Fix build on MSVC by only enabling tst_MIMESniffing on linux.
Jocelyn Turcotte [Fri, 13 May 2011 16:54:20 +0000 (16:54 +0000)]
[Qt] Fix build on MSVC by only enabling tst_MIMESniffing on linux.

Reviewed by nobody, build fix.

* tests/tests.pro:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86437 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-12 Keith Kyzivat <keith.kyzivat@nokia.com>
commit-queue [Thu, 12 May 2011 07:31:43 +0000 (07:31 +0000)]
2011-05-12  Keith Kyzivat  <keith.kyzivat@nokia.com>

        Reviewed by Csaba Osztrogonác.

        [Qt] Arm debug build failing on ARMAssembler::debugOffset()

        Related to svn rev 85523

        * assembler/ARMAssembler.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86316 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-12 Rob Buis <rwlbuis@gmail.com>
Rob Buis [Thu, 12 May 2011 16:57:18 +0000 (16:57 +0000)]
2011-05-12  Rob Buis  <rwlbuis@gmail.com>

        Reviewed by Darin Adler.

        REGRESSION(79985): Changes in fill-opacity should trigger repaint but don't

        Improve test to not need a click. Offer a stroke-opacity variant as well.

        * platform/mac/svg/custom/fill-opacity-update-expected.png:
        * platform/mac/svg/custom/fill-opacity-update-expected.txt:
        * platform/mac/svg/custom/stroke-opacity-update-expected.png: Added.
        * platform/mac/svg/custom/stroke-opacity-update-expected.txt: Added.
        * svg/custom/fill-opacity-update.svg:
        * svg/custom/stroke-opacity-update.svg: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86352 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-12 Rob Buis <rwlbuis@gmail.com>
Rob Buis [Thu, 12 May 2011 12:32:38 +0000 (12:32 +0000)]
2011-05-12  Rob Buis  <rwlbuis@gmail.com>

        Reviewed by Eric Seidel.

        REGRESSION(79985): Changes in fill-opacity should trigger repaint but don't

        Also consider fill-opacity when determining whether StyleDifferenceRepaint is needed.

        Test: svg/custom/fill-opacity-update.svg

        * rendering/style/SVGRenderStyle.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86331 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-12 Adam Barth <abarth@webkit.org>
Adam Barth [Thu, 12 May 2011 18:29:01 +0000 (18:29 +0000)]
2011-05-12  Adam Barth  <abarth@webkit.org>

        Reviewed by Darin Adler.

        REGRESSION (r65868): createContextualFragment does not work with <style>

        These test results match Firefox.

        * fast/parser/style-in-create-contextual-fragment-expected.txt: Added.
        * fast/parser/style-in-create-contextual-fragment-head-expected.txt: Added.
        * fast/parser/style-in-create-contextual-fragment-head.html: Added.
        * fast/parser/style-in-create-contextual-fragment.html: Added.
2011-05-12  Adam Barth  <abarth@webkit.org>

        Reviewed by Darin Adler.

        REGRESSION (r65868): createContextualFragment does not work with <style>

        createContextualFragment doesn't have a spec, so it's hard to tell what
        we're supposed to do exactly.  This patch appears to make our behavior
        more like Firefox and should fix the compat issue with this site.

        Tests: fast/parser/style-in-create-contextual-fragment-head.html

        * dom/Element.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86364 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-12 Carol Szabo <carol@webkit.org>
Abhishek Arya [Thu, 12 May 2011 17:56:30 +0000 (17:56 +0000)]
2011-05-12  Carol Szabo  <carol@webkit.org>

        Reviewed by David Hyatt.

        Fix reparenting and destruction of counter nodes.

        Fixed several issues related to not met assertions.
        See below in the per file description.

        Test: fast/css/counters/element-removal-crash.xhtml

        * dom/ContainerNode.cpp:
        Fixed the fact that Node::detach() used to be called
        while the DOM tree was in an inconsistent state.
        * rendering/RenderCounter.cpp:
        Introduced this function to remove counters from
        descendents of renderers removed from the renderer
        tree not only from the removed renderers themselves.
        * rendering/RenderCounter.h:
        * rendering/RenderObjectChildList.cpp:
        Changed to call RenderCounter::rendererRemovedFromTree
        instead of RenderCounter::destroyCounters.
2011-05-08 Abhishek Arya <inferno@chromium.org> and Carol Szabo  <carol@webkit.org>

        Reviewed by David Hyatt.

        Fix reparenting and destruction of counter nodes.

        * fast/css/counters/element-removal-crash-expected.txt: Added.
        * fast/css/counters/element-removal-crash.xhtml: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86358 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago<rdar://problem/9331651> REGRESSION: RPRVT grows by 1MB / sec @ dvd2blu.com
Geoffrey Garen [Wed, 11 May 2011 20:55:30 +0000 (20:55 +0000)]
<rdar://problem/9331651> REGRESSION: RPRVT grows by 1MB / sec @ dvd2blu.com

Reviewed by Oliver Hunt.

SunSpider reports no change.

This bug was caused by changing Structure and Executable to being GC
objects, and by a long-standing bug that would thrash the global object
between dictionary and non-dictionary states.

* runtime/BatchedTransitionOptimizer.h:
(JSC::BatchedTransitionOptimizer::BatchedTransitionOptimizer): Don't
eagerly transition to dictionary -- this can cause pathological dictionary
churn, and it's not necessary, since objects know how to automatically
transition to dictionary when necessary.

* runtime/Executable.cpp:
(JSC::FunctionExecutable::compileForConstructInternal): Be sure to report
extra cost from compilation, because it can be quite high. This is especially
important for program code, since DOM timers can repeatedly allocate
program code without allocating any other objects.

* runtime/JSObject.cpp:
(JSC::JSObject::removeDirect): Don't transition to the uncacheable state
if the thing we're trying to remove doesn't exist. This can happen during
compilation, since the compiler needs to ensure that no pre-existing
conflicting definitions exist for certain declarations.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86263 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-11 Dimitri Glazkov <dglazkov@chromium.org>
Dimitri Glazkov [Wed, 11 May 2011 18:35:15 +0000 (18:35 +0000)]
2011-05-11  Dimitri Glazkov  <dglazkov@chromium.org>

        Unreviewed, rolling out r85650.

        Triggers m_numNodeListCaches > 0 assert in gc-heavy sites.

        * dom/ContainerNode.cpp:
        * dom/Node.cpp:
        * dom/Node.h:
        * html/HTMLFormControlElement.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86248 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago[MIPS] Fix compilation of the MIPS JIT
Holger Hans Peter Freyther [Wed, 11 May 2011 16:42:43 +0000 (16:42 +0000)]
[MIPS] Fix compilation of the MIPS JIT

Include the MIPSAssembler.h first to indirectly include
AssemblerBuffer.h before the AbstractMacroAssembler.h. This
order is used for the ARM and X86 MacroAssembler*.h.

* assembler/MacroAssemblerMIPS.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86239 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-09 Mahesh Kulkarni <mahesh.kulkarni@nokia.com>
Ademar de Souza Reis Jr [Tue, 10 May 2011 20:56:00 +0000 (17:56 -0300)]
2011-05-09  Mahesh Kulkarni  <mahesh.kulkarni@nokia.com>

        Reviewed by Andreas Kling.

        [Qt] fix fast/dom/Geolocation/window-close-crash.html

        Update all pages with Mock position/error/permission.

        * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2011-05-09  Mahesh Kulkarni  <mahesh.kulkarni@nokia.com>

        Reviewed by Andreas Kling.

        [Qt] fix fast/dom/Geolocation/window-close-crash.html

        Unskipping fast/dom/Geolocation/window-close-crash.html.

        * platform/qt/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86049 268f45cc-cd09-0410-ab3c-d52691b4dbfc



11 years ago2011-05-06 Eric Carlson <eric.carlson@apple.com>
Eric Carlson [Fri, 6 May 2011 15:23:01 +0000 (15:23 +0000)]
2011-05-06  Eric Carlson  <eric.carlson@apple.com>

        Reviewed by Darin Adler.

        REGRESSION (r77954): HTTP Live Streams have incorrect controller UI

        Don't assume that the only time the media controls UI may need to change is on a
        network state change, readyState changes can be significant as well.

        Tested manually because we don't have any live streams to test in DRT.

        * html/HTMLMediaElement.cpp:
        (WebCore::HTMLMediaElement::setNetworkState): changedNetworkState -> updateStatusDisplay.
        (WebCore::HTMLMediaElement::setReadyState): Call updateStatusDisplay.

        * html/shadow/MediaControlRootElement.cpp:
        (WebCore::MediaControlRootElement::reset): changedNetworkState -> updateStatusDisplay.
        (WebCore::MediaControlRootElement::updateStatusDisplay): Ditto.
        * html/shadow/MediaControlRootElement.h:

        * html/shadow/MediaControlRootElementChromium.cpp:
        (WebCore::MediaControlRootElementChromium::reset): Ditto.
        (WebCore::MediaControlRootElementChromium::updateStatusDisplay): Ditto.
        * html/shadow/MediaControlRootElementChromium.h:
        * html/shadow/MediaControls.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85949 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-09 Srinivasulu Chereddy <srinivasulu.chereddy@nokia.com>
commit-queue [Mon, 9 May 2011 11:17:23 +0000 (11:17 +0000)]
2011-05-09  Srinivasulu Chereddy  <srinivasulu.chereddy@nokia.com>

        Reviewed by Laszlo Gombos.

        [Qt] one skipped 2d-transforms layout test passes on Qt Linux - skip list to be updated

        * platform/qt/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86048 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agohttps://bugs.webkit.org/show_bug.cgi?id=60398 (<rdar://problem/9307696>)
David Hyatt [Fri, 6 May 2011 20:07:38 +0000 (20:07 +0000)]
https://bugs.webkit.org/show_bug.cgi?id=60398 (<rdar://problem/9307696>)
REGRESSION (r81684?): cox.com sign in button is missing

Reviewed by Dan Bernstein.

Fix a typo in r81684 where the ascent was used in a comparison instead of boxHeight.
This restores the original behavior before the refactoring messed it up.


Added fast/inline/nested-top-alignment.html

* rendering/InlineFlowBox.cpp:


* fast/inline/nested-top-alignment.html: Added.
* platform/mac/fast/inline/nested-top-alignment-expected.png: Added.
* platform/mac/fast/inline/nested-top-alignment-expected.txt: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85971 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agohttps://bugs.webkit.org/show_bug.cgi?id=60390 (<rdar://problem/9364449>)
David Hyatt [Fri, 6 May 2011 18:50:08 +0000 (18:50 +0000)]
https://bugs.webkit.org/show_bug.cgi?id=60390 (<rdar://problem/9364449>)
REGRESSION (r81992): portions of ticketmaster site render blank

Reviewed by Dan Bernstein.

Make simplified layout skip from a positioned object out to its containing block.  There is no
need to dirty the intermediate inlines, and in this case it's actively harmful, since the
block responsible for the layout of the positioned object doesn't get m_posChildNeedsLayout set.


Added fast/block/positioning/hiding-inside-relpositioned-inline.html

* rendering/RenderObject.h:


* fast/block/positioning/hiding-inside-relpositioned-inline.html: Added.
* platform/mac/fast/block/positioning/hiding-inside-relpositioned-inline-expected.png: Added.
* platform/mac/fast/block/positioning/hiding-inside-relpositioned-inline-expected.txt: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85964 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-05 Bharathwaaj Srinivasan <bharathwaaj.s@gmail.com>
commit-queue [Thu, 5 May 2011 19:21:39 +0000 (19:21 +0000)]
2011-05-05  Bharathwaaj Srinivasan  <bharathwaaj.s@gmail.com>

        Reviewed by Holger Freyther.

        Fix build-webkit --minimal.

        No new tests. This is just a build fix.

        * bindings/js/JSDOMBinding.cpp:
2011-05-05  Bharathwaaj Srinivasan  <bharathwaaj.s@gmail.com>

        Reviewed by Holger Freyther.

        Fix build-webkit --minimal.

        * webkit/webkitglobals.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85868 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-08 Jarkko Sakkinen <jarkko.j.sakkinen@gmail.com>
commit-queue [Sun, 8 May 2011 12:18:53 +0000 (12:18 +0000)]
2011-05-08  Jarkko Sakkinen  <jarkko.j.sakkinen@gmail.com>

        Reviewed by Kenneth Rohde Christiansen.

        [Qt] GraphicsContext3D::getImageData() does not retrieve image data correctly

        Tests: fast/canvas/webgl/gl-teximage.html

        * platform/graphics/qt/GraphicsContext3DQt.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86026 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-05 Prasanth Ullattil <prasanth.ullattil@nokia.com>
commit-queue [Thu, 5 May 2011 20:13:54 +0000 (20:13 +0000)]
2011-05-05  Prasanth Ullattil  <prasanth.ullattil@nokia.com>

        Reviewed by Simon Hausmann.

        Install correct header files for webkit.

        Since WebKit is no longer inside Qt, we can remove the detection for

        [Qt] Install targets are not working correctly for modularized Qt and QtWebkit

        * Source/sync.profile:
2011-05-05  Prasanth Ullattil  <prasanth.ullattil@nokia.com>

        Reviewed by Simon Hausmann.

        Install correct header files for webkit.

        Since WebKit is no longer inside Qt, we can remove the detection for

        [Qt] Install targets are not working correctly for modularized Qt and QtWebkit

        * QtWebKit.pro:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85870 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-09 Jeremy Noble <jer.noble@apple.com>
Jer Noble [Mon, 9 May 2011 23:41:07 +0000 (23:41 +0000)]
2011-05-09  Jeremy Noble  <jer.noble@apple.com>

        Reviewed by Simon Fraser.

        REGRESSION (r84706): Crazy rendering of vimeo.com after exiting fullscreen video

        Initialize m_containsFullScreenElement.

        * html/HTMLFrameElementBase.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86105 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-08 Jeremy Noble <jer.noble@apple.com>
Jer Noble [Tue, 10 May 2011 00:25:55 +0000 (00:25 +0000)]
2011-05-08  Jeremy Noble  <jer.noble@apple.com>

        Reviewed by Dan Bernstein.

        Full-screen video disappears behind black screen with scrollbar, followed by crash at jerryseinfeld.com

        * fullscreen/full-screen-remove-sibling-expected.txt: Added.
        * fullscreen/full-screen-remove-sibling.html: Added.
2011-05-08  Jeremy Noble  <jer.noble@apple.com>

        Reviewed by Dan Bernstein.

        Full-screen video disappears behind black screen with scrollbar, followed by crash at jerryseinfeld.com

        Make RenderFullScreen a non-anonymous block, so that it does not get deleted when a parent attempts to
        coalesce anonymous children.

        Test: fullscreen/full-screen-remove-sibling.html

        * rendering/RenderObject.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86109 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-10 MORITA Hajime <morrita@google.com>
MORITA Hajime [Tue, 10 May 2011 07:46:51 +0000 (07:46 +0000)]
2011-05-10  MORITA Hajime  <morrita@google.com>

        Reviewed by Ryosuke Niwa.

        REGRESSION(r73886): Frequent crashes in replaceSelectionWithFragment

        SpellChecker uses TextCheckerClient, which belongs Page object,
        which is possibly destroyed during SpellChecker's lifetime.
        This change added to a guard before using TextCheckerClient to
        ensure it being live.

        No new tests, this is a speculative fix for a real crash.

        * editing/Editor.cpp:
        * editing/SpellChecker.cpp:
        * editing/SpellChecker.h:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86132 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-10 Alexis Menard <alexis.menard@openbossa.org>
Alexis Menard [Tue, 10 May 2011 15:25:02 +0000 (15:25 +0000)]
2011-05-10  Alexis Menard  <alexis.menard@openbossa.org>

        Reviewed by Antonio Gomes.

        Warning fix on PluginPackage.cpp.

        Warning fix on conversion from time_t to unsigned on Linux with GCC 4.6.0.
        Let's use a struct rather than an array so we can pass everything to StringHasher.

        * plugins/PluginPackage.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86153 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-09 Darin Adler <darin@apple.com> qtwebkit-2.2-tp1
Darin Adler [Mon, 9 May 2011 23:45:06 +0000 (23:45 +0000)]
2011-05-09  Darin Adler  <darin@apple.com>

        Reviewed by Oliver Hunt.

        Wrong type used for return value from strlen

        * wtf/FastMalloc.cpp:
        (WTF::fastStrDup): Use size_t. Also don't bother checking for failure since
        fastMalloc won't return if it fails.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86106 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoUnreviewed.
Csaba Osztrogonác [Tue, 10 May 2011 12:42:32 +0000 (12:42 +0000)]

[Qt][Mac] Add platform specific expected result for plugins/plugin-javascript-access.html,
which is same as the platform independent one, but differs from the general Qt specific.

* platform/qt-mac/plugins: Added.
* platform/qt-mac/plugins/plugin-javascript-access-expected.txt: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86145 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoUnreviewed.
Csaba Osztrogonác [Tue, 10 May 2011 12:14:55 +0000 (12:14 +0000)]

Move plugins/plugin-javascript-access-expected.txt from qt-linux platform
to qt, because it should be used by qt-4.8, qt-arm Linux platforms too.

* platform/qt-linux/plugins: Removed.
* platform/qt-linux/plugins/plugin-javascript-access-expected.txt: Removed.
* platform/qt/plugins/plugin-javascript-access-expected.txt: Copied from LayoutTests/platform/qt-linux/plugins/plugin-javascript-access-expected.txt.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86142 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-10 Andreas Kling <andreas.kling@nokia.com>
andreas.kling@nokia.com [Tue, 10 May 2011 14:56:11 +0000 (14:56 +0000)]
2011-05-10  Andreas Kling  <andreas.kling@nokia.com>

        Rubber-stamped by Csaba Osztrogonác.

        [Qt] Disable HAVE_QRAWFONT until Qt API stabilizes.

        * WebCore.pro: Disable the line that enables HAVE_QRAWFONT
        if the feature is found. Added a comment about why.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86152 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-10 Keith Kyzivat <keith.kyzivat@nokia.com>
commit-queue [Tue, 10 May 2011 11:24:43 +0000 (11:24 +0000)]
2011-05-10  Keith Kyzivat  <keith.kyzivat@nokia.com>

        Reviewed by Kenneth Rohde Christiansen.

        [Qt] QtRawFont support should check based on feature, not version

        Build fix - no tests added.

        * WebCore.pro:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86140 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoUnreviewed Qt 4.8 fix.
Csaba Osztrogonác [Mon, 9 May 2011 14:13:40 +0000 (14:13 +0000)]
Unreviewed Qt 4.8 fix.

* WebCore.pro: 4.8 bots are updated with the full QRawFont APIs, so we can set HAVE_QRAWFONT now.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86054 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoOne more unreviewed typo fix after r86148. :(
Csaba Osztrogonác [Tue, 10 May 2011 13:56:11 +0000 (13:56 +0000)]
One more unreviewed typo fix after r86148. :(

* platform/qt-4.8/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86150 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years agoUnreviewed typo fix after r86148.
Csaba Osztrogonác [Tue, 10 May 2011 13:51:54 +0000 (13:51 +0000)]
Unreviewed typo fix after r86148.

* platform/qt-4.8/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86149 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago[Qt]http/tests/security/xss-DENIED-xsl-document.xml fails with Qt 4.8
Csaba Osztrogonác [Tue, 10 May 2011 13:48:20 +0000 (13:48 +0000)]
[Qt]http/tests/security/xss-DENIED-xsl-document.xml fails with Qt 4.8

[Qt]http/tests/xmlhttprequest/xmlhttprequest-missing-file-exception.html fails with Qt 4.8

* platform/qt-4.8/Skipped:
 - Add http/tests/security/xss-DENIED-xsl-document.xml
 - Add http/tests/xmlhttprequest/xmlhttprequest-missing-file-exception.html

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86148 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago[Qt] Add new qt-4.8 platform for layout testing
Csaba Osztrogonác [Tue, 10 May 2011 11:56:44 +0000 (11:56 +0000)]
[Qt] Add new qt-4.8 platform for layout testing

Reviewed by Kenneth Rohde Christiansen.


* Scripts/old-run-webkit-tests:
* Scripts/webkitdirs.pm:


* platform/qt-4.8/Skipped: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86141 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago[Qt] Use qt-arm platform instead of qt-linux on ARM platform.
Gabor Rapcsanyi [Fri, 6 May 2011 13:24:51 +0000 (13:24 +0000)]
[Qt] Use qt-arm platform instead of qt-linux on ARM platform.

Reviewed by Csaba Osztrogonác.

* Scripts/old-run-webkit-tests:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85942 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-06 Andreas Kling <andreas.kling@nokia.com>
Andreas Kling [Fri, 6 May 2011 14:35:27 +0000 (14:35 +0000)]
2011-05-06  Andreas Kling  <andreas.kling@nokia.com>

        Unreviewed Qt 4.8 build fix.

        * WebCore.pro: Don't set HAVE_QRAWFONT yet. Left a FIXME so we can
        enable it after the 4.8 bots are updated with the full QRawFont APIs.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85944 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-09 Andreas Kling <kling@webkit.org>
Andreas Kling [Mon, 9 May 2011 19:40:17 +0000 (19:40 +0000)]
2011-05-09  Andreas Kling  <kling@webkit.org>

        Reviewed by Kenneth Rohde Christiansen.

        [Qt] Simple small-caps text is displayed in uniform size.

        Covered by existing tests.

        * platform/graphics/qt/FontPlatformDataQt.cpp:
        (WebCore::FontPlatformData::FontPlatformData): In the FontPlatformData constructor
        used by SimpleFontData::scaledFontData(), pass the pixel size to both QFont and QRawFont
        to make small caps and emphasis marks have the correct size when rendered through
        the fast font path.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86081 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-09 Alexis Menard <alexis.menard@openbossa.org>
Alexis Menard [Mon, 9 May 2011 19:57:46 +0000 (19:57 +0000)]
2011-05-09  Alexis Menard  <alexis.menard@openbossa.org>

        Reviewed by Andreas Kling.

        [Qt] Fix the build on Embedded Linux.

        qpa is the name for lighthouse but the old name for Qt for Embedded Linux should
        still be used in order to build properly.

        * Source/WebKit.pri:
2011-05-09  Alexis Menard  <alexis.menard@openbossa.org>

        Reviewed by Andreas Kling.

        [Qt] Fix the build on Embedded Linux.

        qpa is the name for lighthouse but the old name for Qt for Embedded Linux should
        still be used in order to build properly.

        * WebCore.pri:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@86083 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-06 Alexis Menard <alexis.menard@openbossa.org>
Alexis Menard [Fri, 6 May 2011 07:35:46 +0000 (07:35 +0000)]
2011-05-06  Alexis Menard  <alexis.menard@openbossa.org>

        Reviewed by Alexey Proskuryakov.

        Build fix with gcc 4.6 and c++0x support.

        When enabling support of c++0x the compilation fails because of an
        ambiguous overload error. In this code when constructing the pair
        we use the new c++0x rvalue reference feature (&&). It means we are
        calling String(const WTF::AtomicString&) which becomes ambiguous because
        it has an overload WTF::String::String(const WTF::String&) and usually one
        with the native port string type (e.g. QString). In this code we want the
        String version because the pair store Strings.

        No new tests, build fix.

        * loader/FormSubmission.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85928 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-05 Tony Gentilcore <tonyg@chromium.org>
Tony Gentilcore [Thu, 5 May 2011 22:38:09 +0000 (22:38 +0000)]
2011-05-05  Tony Gentilcore  <tonyg@chromium.org>

        Reviewed by Adam Barth.

        ASSERT(m_state == ParsingState) fires @ www.canalplus.fr

        * fast/parser/close-while-stopping-expected.txt: Added.
        * fast/parser/close-while-stopping.html: Added.
2011-05-05  Tony Gentilcore  <tonyg@chromium.org>

        Reviewed by Adam Barth.

        ASSERT(m_state == ParsingState) fires @ www.canalplus.fr

        Test: fast/parser/close-while-stopping.html

        * dom/Document.cpp:
        (WebCore::Document::close): According to http://www.whatwg.org/specs/web-apps/current-work/#dom-document-close,
        the close() steps should be aborted if there is no script-created parser
        associated with the document. Our parser lives throughout
        http://www.whatwg.org/specs/web-apps/current-work/#the-end, but it seems
        the spec doesn't consider the parser active any more. So to properly
        respect this, we need to check that the parser is still parsing.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85894 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-05 Alexis Menard <alexis.menard@openbossa.org> qtwebkit-2.2-base-from-trunk
Alexis Menard [Thu, 5 May 2011 15:44:42 +0000 (15:44 +0000)]
2011-05-05  Alexis Menard  <alexis.menard@openbossa.org>

        Reviewed by Benjamin Poulain.

        [Qt] Make QtWebKit build when using gcc 4.6.0

        If QtWebKit is compiled with gcc 4.6.0 and the build is not done
        with c++0x support we need to deactivate warnings about compatibility.
        Some WebKit types have names which conflicts with upcoming c++0x
        types. If QtWebKit is built with c++0x support, WebKit has code to
        fallback to the standard type rather than the WebKit implementation.
        For example nullptr will be the one of the std if build options turns on
        c++0x support.

        * Source/WebKit.pri:
2011-05-05  Alexis Menard  <alexis.menard@openbossa.org>

        Reviewed by Benjamin Poulain.

        [Qt] Make QtWebKit build when using gcc 4.6.0

        If QtWebKit is compiled with gcc 4.6.0 or later we don't want to deactivate
        the c++0x support because it works.

        * JavaScriptCore.pro:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85855 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-05 Carlos Garcia Campos <cgarcia@igalia.com>
Carlos Garcia Campos [Thu, 5 May 2011 15:40:18 +0000 (15:40 +0000)]
2011-05-05  Carlos Garcia Campos  <cgarcia@igalia.com>

        Reviewed by Martin Robinson.

        [GTK] Module::load() should return true when g_module_open() returns a valid pointer

        * Platform/gtk/ModuleGtk.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85854 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-05 Andreas Kling <andreas.kling@nokia.com>
Andreas Kling [Thu, 5 May 2011 15:20:07 +0000 (15:20 +0000)]
2011-05-05  Andreas Kling  <andreas.kling@nokia.com>

        Reviewed by Simon Hausmann.

        [Qt] Implement the fast font path for Qt.

        Use the new QRawFont and QGlyphs APIs in Qt 4.8 to implement the fast paths for
        rendering and measurement of simple text.

        Since this is still unreleased API, it's guarded by HAVE(QRAWFONT) until the new
        classes are fully integrated into the Qt 4.8 release branch.

        * WebCore.pro: Add HAVE_QRAWFONT define (for Qt >= 4.8) and new files to build.

        * platform/graphics/Font.cpp:
        (WebCore::Font::drawText): Disable fast font path for stroked text, and when
        painting text with a shadow. (Qt-only)
        (WebCore::Font::codePath): Try to use fast font path in more cases.

        * platform/graphics/qt/FontCacheQt.cpp:
        (WebCore::rawFontForCharacters): Helper function, returns a suitable QRawFont
        to use for rendering a given string. Goes through QTextLayout to find the best
        font based on the original QFont query.
        (WebCore::FontCache::getFontDataForCharacters): Implemented using helper above.

        * platform/graphics/qt/FontPlatformData.h:
        * platform/graphics/qt/FontPlatformDataQt.cpp:
        (WebCore::FontPlatformData::FontPlatformData): Add a QRawFont member to FontPlatformData.

        * platform/graphics/qt/FontQt.cpp:
        (WebCore::drawTextCommon): Factored QPen creation out of drawTextCommon
        to share code between complex and simple font rendering paths.
        (WebCore::Font::drawGlyphs): Implemented using QPainter::drawGlyphs().

        * platform/graphics/qt/GlyphPageTreeNodeQt.cpp:
        (WebCore::GlyphPage::fill): Implemented using QRawFont::glyphIndexesForString().

        * platform/graphics/qt/SimpleFontDataQt.cpp:
        (WebCore::SimpleFontData::determinePitch): Return false when using QRawFont since we
        have no way of knowing the pitch.
        (WebCore::SimpleFontData::containsCharacters): Implemented using QRawFont::supportsCharacter().
        (WebCore::SimpleFontData::platformWidthForGlyph): Implemented using QRawFont::advancesForGlyphIndexes().
        (WebCore::SimpleFontData::scaledFontData): Added, based on other ports.
        (WebCore::SimpleFontData::smallCapsFontData): Ditto.
        (WebCore::SimpleFontData::emphasisMarkFontData): Ditto.
        (WebCore::SimpleFontData::platformBoundsForGlyph): Stub.
        (WebCore::SimpleFontData::platformInit): Use QRawFont APIs to retrieve font metrics.
        (WebCore::SimpleFontData::platformCharWidthInit): Ditto.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85853 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-05 Andrey Kosyakov <caseq@chromium.org>
Andrey Kosyakov [Thu, 5 May 2011 14:15:38 +0000 (14:15 +0000)]
2011-05-05  Andrey Kosyakov  <caseq@chromium.org>

        Unreviewed. Skip inspector/console/console-shadow-dom-access.html (crashes in debug)

        * platform/mac/Skipped:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85852 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-05 Alexis Menard <alexis.menard@openbossa.org>
Alexis Menard [Thu, 5 May 2011 14:14:09 +0000 (14:14 +0000)]
2011-05-05  Alexis Menard  <alexis.menard@openbossa.org>

        Reviewed by Andreas Kling.

        [Qt] RenderThemeQt and DumpRenderTreeSupportQt should use nullptr rather than 0.

        We should use nullptr rather than 0. nullptr will be added in the new C++ standard
        but WebKit already has a nullptr class if there is no c++0x support.

        * WebCoreSupport/DumpRenderTreeSupportQt.cpp:
2011-05-05  Alexis Menard  <alexis.menard@openbossa.org>

        Reviewed by Andreas Kling.

        [Qt] RenderThemeQt and DumpRenderTreeSupportQt should use nullptr rather than 0.

        We should use nullptr rather than 0. nullptr will be added in the new C++ standard
        but WebKit already has a nullptr class if there is no c++0x support.

        * platform/qt/RenderThemeQt.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85851 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-05 Ilya Tikhonovsky <loislo@chromium.org>
Ilya Tikhonovsky [Thu, 5 May 2011 13:20:24 +0000 (13:20 +0000)]
2011-05-05  Ilya Tikhonovsky  <loislo@chromium.org>

        Reviewed by Yury Semikhatsky.

        Web Inspector: rename BrowserDebugger to DOMDebugger.

        InspectorBrowserDebuggerAgent => InspectorDOMDebuggerAgent
        browserDebugger => domDebugger

        * CMakeLists.txt:
        * GNUmakefile.list.am:
        * WebCore.gypi:
        * WebCore.pro:
        * WebCore.vcproj/WebCore.vcproj:
        * WebCore.xcodeproj/project.pbxproj:
        * inspector/CodeGeneratorInspector.pm:
        * inspector/InspectorAgent.cpp:
        * inspector/InspectorAgent.h:
        * inspector/InspectorController.cpp:
        * inspector/InspectorDOMDebuggerAgent.cpp: Renamed from Source/WebCore/inspector/InspectorBrowserDebuggerAgent.cpp.
        * inspector/InspectorDOMDebuggerAgent.h: Renamed from Source/WebCore/inspector/InspectorBrowserDebuggerAgent.h.
        * inspector/InspectorInstrumentation.cpp:
        * inspector/InstrumentingAgents.h:
        * inspector/WorkerInspectorController.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85850 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-04 Andrey Kosyakov <caseq@chromium.org>
Andrey Kosyakov [Thu, 5 May 2011 12:14:17 +0000 (12:14 +0000)]
2011-05-04  Andrey Kosyakov  <caseq@chromium.org>

        Reviewed by Yury Semikhatsky.

        Web Inspector: add test for shadow DOM access in the console API

        * http/tests/inspector/elements-test.js:
        * inspector/console/console-shadow-dom-access-expected.txt: Added.
        * inspector/console/console-shadow-dom-access.html: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85849 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-05 Philippe Normand <pnormand@igalia.com>
Philippe Normand [Thu, 5 May 2011 10:58:38 +0000 (10:58 +0000)]
2011-05-05  Philippe Normand  <pnormand@igalia.com>

        Unreviewed, GTK baselines for new tests.

        * platform/gtk/editing/inserting/typing-at-end-of-line-expected.png: Added.
        * platform/gtk/editing/inserting/typing-at-end-of-line-expected.txt: Added.
        * platform/gtk/editing/selection/select-across-readonly-input-1-expected.png: Added.
        * platform/gtk/editing/selection/select-across-readonly-input-1-expected.txt: Added.
        * platform/gtk/editing/selection/select-across-readonly-input-2-expected.png: Added.
        * platform/gtk/editing/selection/select-across-readonly-input-2-expected.txt: Added.
        * platform/gtk/editing/selection/select-across-readonly-input-3-expected.png: Added.
        * platform/gtk/editing/selection/select-across-readonly-input-3-expected.txt: Added.
        * platform/gtk/editing/selection/select-across-readonly-input-4-expected.png: Added.
        * platform/gtk/editing/selection/select-across-readonly-input-4-expected.txt: Added.
        * platform/gtk/editing/selection/select-across-readonly-input-5-expected.png: Added.
        * platform/gtk/editing/selection/select-across-readonly-input-5-expected.txt: Added.
        * platform/gtk/fast/forms/input-placeholder-text-indent-expected.png: Added.
        * platform/gtk/fast/forms/input-placeholder-text-indent-expected.txt: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85848 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-04 Yury Semikhatsky <yurys@chromium.org>
Yury Semikhatsky [Thu, 5 May 2011 10:41:36 +0000 (10:41 +0000)]
2011-05-04  Yury Semikhatsky  <yurys@chromium.org>

        Reviewed by Pavel Feldman.

        Web Inspector: show only Console and Scripts panels in worker inspector front-end

        * inspector/front-end/WorkerManager.js:
        * inspector/front-end/inspector.js: show only Scripts and Console panels in the worker inspector front-end.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85847 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-05 MORITA Hajime <morrita@google.com>
MORITA Hajime [Thu, 5 May 2011 09:17:40 +0000 (09:17 +0000)]
2011-05-05  MORITA Hajime  <morrita@google.com>

        Unreviewed, marking flaky tests.

        * platform/chromium/test_expectations.txt:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85846 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-05 MORITA Hajime <morrita@google.com>
MORITA Hajime [Thu, 5 May 2011 08:44:12 +0000 (08:44 +0000)]
2011-05-05  MORITA Hajime  <morrita@google.com>

        Unreviewed, rolling out r85817.

        It looks these lines are mis-skipped.

        * platform/chromium/test_expectations.txt:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85845 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-03 Hans Wennborg <hans@chromium.org>
Hans Wennborg [Thu, 5 May 2011 08:36:27 +0000 (08:36 +0000)]
2011-05-03  Hans Wennborg  <hans@chromium.org>

        Reviewed by Steve Block.

        IndexedDB: Unit tests for LevelDB key coding functions

        Fix some embarrassing bugs uncovered by unit tests.

        No new functionality, but this is now covered by unit tests in the Chromium WebKit port.

        * storage/IDBLevelDBCoding.cpp:
2011-05-03  Hans Wennborg  <hans@chromium.org>

        Reviewed by Steve Block.

        IndexedDB: Unit tests for LevelDB key coding functions

        Add unit tests for the basic key coding functions used by the LevelDB back-end.

        * WebKit.gypi:
        * tests/IDBLevelDBCodingTest.cpp: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85844 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-05 Philippe Normand <pnormand@igalia.com>
Philippe Normand [Thu, 5 May 2011 08:25:33 +0000 (08:25 +0000)]
2011-05-05  Philippe Normand  <pnormand@igalia.com>

        Unreviewed, GTK baselines for new fullscreen test.

        * platform/gtk/fullscreen/full-screen-stacking-context-expected.png: Added.
        * platform/gtk/fullscreen/full-screen-stacking-context-expected.txt: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85843 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-05 MORITA Hajime <morrita@google.com>
MORITA Hajime [Thu, 5 May 2011 08:12:29 +0000 (08:12 +0000)]
2011-05-05  MORITA Hajime <morrita@google.com>

        Unreviewed, fixed a syntax error on expectations file.

        * platform/chromium/test_expectations.txt:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85842 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-05 Carlos Garcia Campos <cgarcia@igalia.com>
Carlos Garcia Campos [Thu, 5 May 2011 07:57:33 +0000 (07:57 +0000)]
2011-05-05  Carlos Garcia Campos  <cgarcia@igalia.com>

        Unreviewed. Fix WebKit2 GTK build after r85808.

        * Shared/NativeWebWheelEvent.h:
        * UIProcess/API/gtk/WebKitWebViewBase.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85841 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-05 MORITA Hajime <morrita@google.com>
MORITA Hajime [Thu, 5 May 2011 07:25:26 +0000 (07:25 +0000)]
2011-05-05  MORITA Hajime  <morrita@google.com>

        Unreviewed expectations update for MAC GPU flakiness.
        - video-zoom-controls.html: Added IMAGE
        - media/controls-without-preload.html, media/controls-styling.html, media/controls-strict.html, media/controls-after-reload.html
          Added TEXT
        * platform/chromium/test_expectations.txt:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85840 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-05 Young Han Lee <joybro@company100.net>
commit-queue [Thu, 5 May 2011 07:00:38 +0000 (07:00 +0000)]
2011-05-05  Young Han Lee  <joybro@company100.net>

        Reviewed by Csaba Osztrogonác.

        [Texmap][Qt] Enable strict OwnPtr for Qt with texmap enabled.

        No new tests. Build fix.

        * platform/graphics/texmap/GraphicsLayerTextureMapper.cpp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85839 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-04 MORITA Hajime <morrita@google.com>
MORITA Hajime [Thu, 5 May 2011 06:48:08 +0000 (06:48 +0000)]
2011-05-04  MORITA Hajime  <morrita@google.com>

        Unreviewed expectations update.

        * platform/chromium/test_expectations.txt:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85838 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-04 Jay Civelli <jcivelli@chromium.org>
Jay Civelli [Thu, 5 May 2011 06:41:40 +0000 (06:41 +0000)]
2011-05-04  Jay Civelli  <jcivelli@chromium.org>

        Reviewed by Adam Barth.

        Adding a utility class to read a SharedBuffer line by line.

        * CMakeLists.txt:
        * GNUmakefile.list.am:
        * WebCore.gypi:
        * WebCore.pro:
        * WebCore.vcproj/WebCore.vcproj:
        * WebCore.xcodeproj/project.pbxproj:
        * platform/SharedBufferCRLFLineReader.cpp: Added.
        * platform/SharedBufferCRLFLineReader.h: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85837 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-04 MORITA Hajime <morrita@google.com>
MORITA Hajime [Thu, 5 May 2011 06:32:46 +0000 (06:32 +0000)]
2011-05-04  MORITA Hajime <morrita@google.com>

        r85835 skipped too much. narrowed skipped target.

        * platform/chromium/test_expectations.txt:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85836 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11 years ago2011-05-04 MORITA Hajime <morrita@google.com>
MORITA Hajime [Thu, 5 May 2011 06:29:16 +0000 (06:29 +0000)]
2011-05-04  MORITA Hajime  <morrita@google.com>

        Unreviewed marking failing gpu tests:
        - controls-without-preload.html, controls-after-reload.html, controls-strict.html, controls-styling.html

        * platform/chromium/test_expectations.txt:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@85835 268f45cc-cd09-0410-ab3c-d52691b4dbfc