2010-04-01  Ojan Vafai  <ojan@chromium.org>

        Reviewed by Adam Barth.

        autoinstalled should be git ignored
        https://bugs.webkit.org/show_bug.cgi?id=36970

        * .gitignore:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56944 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore: Changing SQLiteStatement::prepare() to return an error when it's
given a string that has more than one statement in it. Currently,
everything past the first statement is silently ignored.

Reviewed by Dimitri Glazkov.

Test: storage/executesql-accepts-only-one-statement.html

* platform/sql/SQLiteStatement.cpp:
(WebCore::SQLiteStatement::prepare):

LayoutTests: Testing that executeSql() accepts only one-statement strings.

Reviewed by Dimitri Glazkov.

* storage/executesql-accepts-only-one-statement-expected.txt: Added.
* storage/executesql-accepts-only-one-statement.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56943 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Eric Seidel  <eric@webkit.org>

        Unreviewed, only ignoring chatty style errors.

        check-webkit-style: WebKit needs a python style checker
        https://bugs.webkit.org/show_bug.cgi?id=33639

        Remove the 79 char line limit by ignoring
        pep8/E501.  Because we have our own report_error
        implementation we have to ignore E501 by hand
        instead of passing --ignore=E501.

        Right now over 1400 lines of our existing python
        fail E501 so this rule just generates needless noise.
        The rest of WebKit has no wrapping rule so it makes
        little sense for our Python to differ here.

        * Scripts/webkitpy/style/processors/python.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56942 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  James Robinson  <jamesr@chromium.org>

        Reviewed by Simon Fraser.

        Keeps a transient optimistic offset to the root in RenderLayer::updateLayerPositions
        https://bugs.webkit.org/show_bug.cgi?id=33520

        RenderLayer::updateLayerPositions() makes a recursive walk through all RenderLayers and updates the repaint rectangles on each.
        These rectangles have to be calculated in the repaint container's coordinates using RenderObject::mapLocalToContainer to walk
        up to the repaint container.  This patch keeps track of the offset to the root and uses that offset instead of walking back up to
        the root every time.

        Tests: fast/layers/nested-layers-1.html
               fast/layers/nested-layers-2.html
               fast/layers/nested-layers-3.html
               fast/layers/nested-layers-4.html

        * page/FrameView.cpp:
        (WebCore::FrameView::layout):
        * rendering/RenderBox.cpp:
        (WebCore::RenderBox::outlineBoundsForRepaint):
        * rendering/RenderBox.h:
        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::updateLayerPositions):
        * rendering/RenderLayer.h:
        * rendering/RenderObject.h:
        (WebCore::RenderObject::outlineBoundsForRepaint):
        * rendering/RenderSVGModelObject.cpp:
        (WebCore::RenderSVGModelObject::outlineBoundsForRepaint):
        * rendering/RenderSVGModelObject.h:
2010-04-01  James Robinson  <jamesr@chromium.org>

        Reviewed by Simon Fraser.

        Mapping from local to container coord space in updateLayerPositions is quadratic
        https://bugs.webkit.org/show_bug.cgi?id=33520

        These tests move layers around in various ways to test that the offset caching in
        updateLayerPositions() is correct.

        * fast/layers/nested-layers-1-expected.txt: Added.
        * fast/layers/nested-layers-1.html: Added.
        * fast/layers/nested-layers-2-expected.txt: Added.
        * fast/layers/nested-layers-2.html: Added.
        * fast/layers/nested-layers-3-expected.txt: Added.
        * fast/layers/nested-layers-3.html: Added.
        * fast/layers/nested-layers-4-expected.txt: Added.
        * fast/layers/nested-layers-4.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56941 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Daniel Bates  <dbates@rim.com>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=36901

        Removed functions Range::operator == and Range::operator != as they
        were using C++ code that was not sound and hence may have undefined
        behavior.

        Test case: manual-tests/crash-on-find-with-no-selection.html

        * dom/Range.cpp:
        (WebCore::areRangesEqual): Added.
        * dom/Range.h:
        * editing/markup.cpp:
        (WebCore::createMarkup): Modified to call WebCore::areRangesEqual.
        * manual-tests/crash-on-find-with-no-selection.html: Added.
        * page/Frame.cpp:
        (WebCore::Frame::findString): Modified to call WebCore::areRangesEqual.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56940 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Touch WebKit.idl to force a rebuild of Interfaces

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56939 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fixed Windows (and hopefully Tiger) build

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56938 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Adam Barth  <abarth@webkit.org>

        Unreviewed.  Add cmarrin's IRC nickname.

        * Scripts/webkitpy/common/config/committers.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56937 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Adam Barth  <abarth@webkit.org>

        Unreviewed.

        EWS spins hot when unable to build
        https://bugs.webkit.org/show_bug.cgi?id=36981

        The problem is that the queue engine things we have more work to do,
        but the bot isn't actually able to do anything.  After this change, we
        back off the usual amount.

        * Scripts/webkitpy/tool/bot/queueengine.py:
        * Scripts/webkitpy/tool/bot/queueengine_unittest.py:
        * Scripts/webkitpy/tool/commands/queues.py:
        * Scripts/webkitpy/tool/commands/sheriffbot.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56936 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=36976
        <rdar://problem/7817498>
        REGRESSION(r54783): Silverlight plug-in causes Safari to crash if JavaScript is disabled

        * Plugins/Hosted/NetscapePluginInstanceProxy.mm:
        (WebKit::NetscapePluginInstanceProxy::LocalObjectMap::get): The HashTable assertions aren't
        there to catch potential future attempts to store empty/deleted values before these happen -
        it's actually wrong to try to look up these values. Added an early return.
        (WebKit::NetscapePluginInstanceProxy::LocalObjectMap::forget): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56935 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Added layerTreeAsText function to DRT (for Mac)
        https://bugs.webkit.org/show_bug.cgi?id=36782

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56934 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Synchronous rendering when setting form control values slows down JavaScript
https://bugs.webkit.org/show_bug.cgi?id=36967

Reviewed by Sam Weinig.

This patch basically reverts http://trac.webkit.org/changeset/19006.

Dan asked me to investigate why 19006 is no longer needed. I have two answers:

    (1) I was also able to remove the synchronous call to updateFromElement().
    That call was the proximate cause of the crash that 19006 fixed.

    (2) updateFromElement() no longer calls HTMLElement::setInnerText()
    in the way that it used to. (However, it doesn't seem prudent to
    rely on this happy coincidence.)

* html/HTMLInputElement.cpp:
(WebCore::HTMLInputElement::setValue): Simplified some logic here. Moved
setNeedsValidityCheck() outside of individual 'if' clauses, since they all
called it.

Removed call to updateStyleIfNeeded(), which does rendering synchronously,
since that was the performance problem. (setNeedsStyleRecalc() ensures
that rendering will happen asynchronously.) Also removed comment about
ordering dangers introduced by updateStyleIfNeeded().

Removed call to updateFromElement(), since it's dangerous and also a minor
performance problem. (setNeedsStyleRecalc() ensures that updateFromElement()
will happen asynchronously, too.)

* html/HTMLTextAreaElement.cpp:
(WebCore::HTMLTextAreaElement::setNonDirtyValue): Ditto. Here, I had to
add a call to setNeedsStyleRecalc(), since there wasn't one before.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56933 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Oliver Hunt.

        https://bugs.webkit.org/show_bug.cgi?id=36976
        <rdar://problem/7817498>
        REGRESSION(r54783): Silverlight plug-in causes Safari to crash if JavaScript is disabled

        * Plugins/Hosted/NetscapePluginInstanceProxy.mm:
        (WebKit::NetscapePluginInstanceProxy::LocalObjectMap::get): Use find() instead of get(),
        because the latter fails with an assertion when looking up 0 or -1.
        (WebKit::NetscapePluginInstanceProxy::LocalObjectMap::forget): Be prepared for unexpected
        object IDs coming from plug-in host.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56932 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Ilya Tikhonovsky  <loislo@chromium.org>

        Reviewed by Pavel Feldman.

        Web Inspector: Sometimes js code can detach page from it's frame and in that case
        Dispatch Events will stay in the TimelineAgent's events stack. Only immediate events will
        appear at frontend.
        https://bugs.webkit.org/show_bug.cgi?id=36890

        * bindings/v8/V8Proxy.cpp:
        (WebCore::V8Proxy::callFunction):
        * dom/Node.cpp:
        (WebCore::Node::dispatchGenericEvent):
        * page/DOMWindow.cpp:
        (WebCore::DOMWindow::dispatchEvent):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56931 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Bug 36977 - aria-label doesn't work on image map area
https://bugs.webkit.org/show_bug.cgi?id=36977

Reviewed by Beth Dakin.

WebCore:

Test: platform/mac/accessibility/area-with-aria-label.html

* accessibility/AccessibilityImageMapLink.cpp:
(WebCore::AccessibilityImageMapLink::accessibilityDescription):

LayoutTests:

* platform/mac/accessibility/area-with-aria-label-expected.txt: Added.
* platform/mac/accessibility/area-with-aria-label.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56930 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Jesus Sanchez-Palencia  <jesus.palencia@openbossa.org>

        Reviewed by Kenneth Rohde Christiansen.

        Add Single and Multiple Selection Popup for Maemo 5.

        [Qt] Maemo5 theme - popup dialogs
        https://bugs.webkit.org/show_bug.cgi?id=36789

        * WebCoreSupport/QtMaemoWebPopup.cpp:
        (WebCore::Maemo5Popup::populateList):
        (WebCore::Maemo5Popup::onItemSelected):
        (WebCore::QtMaemoWebPopup::createSingleSelectionPopup):
        (WebCore::QtMaemoWebPopup::createMultipleSelectionPopup):
        (WebCore::Maemo5SingleSelectionPopup::Maemo5SingleSelectionPopup):
        (WebCore::MultipleItemListDelegate::MultipleItemListDelegate):
        (WebCore::MultipleItemListDelegate::paint):
        (WebCore::Maemo5MultipleSelectionPopup::Maemo5MultipleSelectionPopup):
        * WebCoreSupport/QtMaemoWebPopup.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56929 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Julien Chaffraix  <jchaffraix@webkit.org>

        Reviewed by Eric Seidel.

        Part of:
        REGRESSION(56394):
        http/tests/xmlhttprequest/xmlhttprequest-50ms-download-dispatch.html
        fails on Tiger
        https://bugs.webkit.org/show_bug.cgi?id=36934

        * http/tests/xmlhttprequest/xmlhttprequest-50ms-download-dispatch.html:
        Add more explicit failing messages to help find out what is failing.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56928 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[wx] Build fix after addition of JavaScriptCore/wtf/text directory.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56927 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Eric Seidel  <eric@webkit.org>

        Reviewed by Maciej Stachowiak.

        REGRESSION(56489): security/block-test.html fails on Tiger Bot
        https://bugs.webkit.org/show_bug.cgi?id=36932

        It appears that r56489 just forgot to update the Tiger results
        to include the new parts of the test.  Tiger has had custom
        results for this test since r54193.

        I just pulled these off the bot.  I do not have the domain knowledge
        to know if these are right or not.  They look fine to me.

        * platform/mac-tiger/security/block-test-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56926 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Zhenyao Mo  <zmo@google.com>

        Reviewed by Eric Seidel.

        fast/canvas/webgl/index-validation.html failed on Leopard Commit Bot
        https://bugs.webkit.org/show_bug.cgi?id=36908

        * fast/canvas/webgl/script-tests/index-validation.js: Disable antialias for this test.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56925 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Dan Bernstein.

        Assertion failure: !repaintContainer || repaintContainer == this
        https://bugs.webkit.org/show_bug.cgi?id=36672

        RenderText's implementation of clippedOverflowRectForRepaint() uses containingBlock()
        to get the renderer to use for computing the repaint rect. However, the renderer returned
        by containingBlock() may be an ancestor of the repaintContainer, and containingBlock()
        doesn't have the 'repaintContainerSkipped' logic that container() has.

        So in this case, check to see whether repaintContainer is actually a descendant of the
        containing block, and in that case just repaint the entire repaintContainer.

        Test: compositing/repaint/inline-repaint-container.html

        * rendering/RenderText.cpp:
        (WebCore::RenderText::clippedOverflowRectForRepaint):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56924 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Laszlo Gombos  <laszlo.1.gombos@nokia.com>

        Reviewed by Kenneth Rohde Christiansen.

        [Qt]Add -Wextra warning level to the QtWebKit build
        https://bugs.webkit.org/show_bug.cgi?id=36971

        * WebKit.pri:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56923 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] REGRESSION(r56869): Windows build is broken
https://bugs.webkit.org/show_bug.cgi?id=36929

Reviewed by Jian Li.

* WebCore.pro: LIBS += -lOle32 added.
* platform/UUID.cpp: Define ARRAYSIZE macro if it isn't defined before.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56922 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rubber-stamped by David Levin.

Add myself to the reviewers list.

* Scripts/webkitpy/common/config/committers.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56921 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Bug 36968 - 1 crash in Safari at com.apple.WebCore: WebCore::Element::getAttribute const
https://bugs.webkit.org/show_bug.cgi?id=36968

Reviewed by Beth Dakin.

WebCore:

Test: accessibility/crash-with-noelement-selectbox.html

When a <select> element had no options, the selectedIndex == -1 and that was being
used to index into an empty array.

* accessibility/AccessibilityRenderObject.cpp:
(WebCore::AccessibilityRenderObject::stringValue):

LayoutTests:

* accessibility/crash-with-noelement-selectbox-expected.txt: Added.
* accessibility/crash-with-noelement-selectbox.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56920 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Adam Barth  <abarth@webkit.org>

        Unreviewed.

        Turns out the Chromium Windows bots don't have pdevenv installed.

        * Scripts/webkitdirs.pm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56919 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Adam Barth  <abarth@webkit.org>

        Reviewed by Dimitri Glazkov.

        Let Chromium Windows build with Visual Studio Express
        https://bugs.webkit.org/show_bug.cgi?id=36919

        This is horrible, horrible copy/paste code, but that seems to be the
        way of webkitdirs.pm.  :(

        Someone needs to go through an cleanup this code, but I don't have the
        heart to do it in this patch.

        * Scripts/webkitdirs.pm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56918 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Finnur Thorarinsson  <finnur.webkit@gmail.com>

        Reviewed by Dimitri Glazkov.

        [chromium] FindInPage should clear the focused node when a match has
        been found. This is because WebFrameImpl::setFocus will try to refocus
        editable elements if it thinks they have focus, causing the page to
        scroll.

        https://bugs.webkit.org/show_bug.cgi?id=36923

        * src/WebFrameImpl.cpp:
        (WebKit::WebFrameImpl::find):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56917 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Gustavo Noronha Silva  <gustavo.noronha@collabora.co.uk>

        Reviewed by Holger Freyther.

        [GTK] webkit_get_default_session() should make sure webkit_init() is called
        https://bugs.webkit.org/show_bug.cgi?id=36754

        Make sure global functions that do not require a WebKitWebView to
        be created call webkit_init() before doing their job. Also add an
        API test to check for that.

        * tests/testglobals.c: Added.
        (test_globals_default_session):
        (main):
        * webkit/webkitwebview.cpp:
        (webkit_get_default_session):
        (webkit_set_cache_model):
        (webkit_get_cache_model):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56916 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  MORITA Hajime  <morrita@google.com>

        Reviewed by Shinichiro Hamaji.

        https://bugs.webkit.org/show_bug.cgi?id=36948
        Refactoring: Position::primaryDirection() should be extracted.

        No new tests. No functinal change.

        * dom/Position.cpp:
        (WebCore::Position::getInlineBoxAndOffset):
        (WebCore::Position::primaryDirection): Added
        * dom/Position.h:
        * editing/VisiblePosition.cpp:
        (WebCore::VisiblePosition::leftVisuallyDistinctCandidate):
        (WebCore::VisiblePosition::rightVisuallyDistinctCandidate):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56915 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Leandro Pereira  <leandro@profusion.mobi>

        Reviewed by Gustavo Noronha Silva.

        Add EFL's pkg-config metadata file file to efl/.
        http://webkit.org/b/36766

        * efl/ewebkit.pc.in: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56914 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Ilya Tikhonovsky  <loislo@chromium.org>

        Reviewed by Yury Semikhatsky.

        Add myself to the committers list.
        https://bugs.webkit.org/show_bug.cgi?id=36953

        * Scripts/webkitpy/common/config/committers.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56913 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  yael aharon  <yael.aharon@nokia.com>

        Reviewed by Laszlo Gombos.

        set-progress-properties.html expected result should not be platform specific
        https://bugs.webkit.org/show_bug.cgi?id=36918

        * fast/dom/HTMLProgressElement/set-progress-properties-expected.txt: Added.
        * platform/qt/fast/dom/HTMLProgressElement/set-progress-properties-expected.txt: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56912 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Shu Chang  <chang.shu@nokia.com>

        Reviewed by Laszlo Gombos.

        [Qt] As Qt does not support dashboard compatibility mode, the red squares on
        the second row are expected. Rebaseline and unskip test case.
        https://bugs.webkit.org/show_bug.cgi?id=36869

        * platform/qt/Skipped:
        * platform/qt/fast/canvas/fill-stroke-clip-reset-path-expected.checksum: Added.
        * platform/qt/fast/canvas/fill-stroke-clip-reset-path-expected.png: Added.
        * platform/qt/fast/canvas/fill-stroke-clip-reset-path-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56911 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Yury Semikhatsky  <yurys@chromium.org>

        Reviewed by Pavel Feldman.

        Mark console-log-before-frame-navigation as pssing on Chromium Win.

        https://bugs.webkit.org/show_bug.cgi?id=36807

        * platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56910 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Alexander Pavlov  <apavlov@chromium.org>

        Reviewed by Pavel Feldman.

        Web Inspector: Audits: fix parsing of injected script-evaluated result for CssInHeadRule
        https://bugs.webkit.org/show_bug.cgi?id=36952

        * inspector/audits-panel-functional-expected.txt:
        * inspector/audits-panel-functional.html:
2010-04-01  Alexander Pavlov  <apavlov@chromium.org>

        Reviewed by Pavel Feldman.

        Web Inspector: Audits: fix parsing of injected script-evaluated result for CssInHeadRule
        https://bugs.webkit.org/show_bug.cgi?id=36952

        * inspector/front-end/AuditRules.js:
        (WebInspector.AuditRules.CssInHeadRule.prototype.doRun):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56909 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Marcus Bulach  <bulach@chromium.org>

        Reviewed by Jeremy Orlow.

        As a follow up on https://bugs.webkit.org/show_bug.cgi?id=36535, remove:
        GeolocationServiceBridgeChromium.h
        dettachBridge()
        getGeolocationService()
        https://bugs.webkit.org/show_bug.cgi?id=36895

        * WebKit.gyp:
        * public/GeolocationServiceBridgeChromium.h: Removed.
        * public/WebGeolocationService.h:
        (WebKit::WebGeolocationService::detachBridge):
        * public/WebViewClient.h:
        (WebKit::WebViewClient::geolocationService):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56908 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Jeremy Orlow  <jorlow@chromium.org>

        Reviewed by Nate Chapin.

        Misc IndexedDatabase cleanup
        https://bugs.webkit.org/show_bug.cgi?id=36889

        No functional changes.

        * bindings/v8/custom/V8CustomIDBCallbacks.h:
        (WebCore::V8CustomIDBCallbacks::onSuccess):
        (WebCore::V8CustomIDBCallbacks::onError):
        (WebCore::V8CustomIDBCallbacks::V8CustomIDBCallbacks):
            Get rid of 2 largely redundant bools

        * storage/IndexedDatabase.h:
        * storage/IndexedDatabaseImpl.cpp:
        (WebCore::IndexedDatabaseImpl::open):
        * storage/IndexedDatabaseImpl.h:
        * storage/IndexedDatabaseRequest.cpp:
        (WebCore::IndexedDatabaseRequest::open):
        * storage/IndexedDatabaseRequest.h:
            Plumb the Frame* and style cleanups.
2010-03-31  Jeremy Orlow  <jorlow@chromium.org>

        Reviewed by Nate Chapin.

        Misc IndexedDatabase cleanup
        https://bugs.webkit.org/show_bug.cgi?id=36889

        Plumb the Frame* so Chromium knows where the request originated from.

        * public/WebIndexedDatabase.h:
        * src/IndexedDatabaseProxy.cpp:
        (WebCore::IndexedDatabaseProxy::open):
        * src/IndexedDatabaseProxy.h:
        * src/WebIndexedDatabaseImpl.cpp:
        (WebKit::WebIndexedDatabaseImpl::open):
        * src/WebIndexedDatabaseImpl.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56907 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Nikolas Zimmermann  <nzimmermann@rim.com>

        Reviewed by Dirk Schulze.

        Fix filter references in svg/filters/filterRes.svg, to make two filters show up in Opera.
        Clearly shows we have a regression with filterRes + feMorphology.

        * svg/filters/filterRes.svg:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56906 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Nikolas Zimmermann  <nzimmermann@rim.com>

        Reviewed by Dirk Schulze.

        REGRESSION: document.documentElement.getScreenCTM() returns incorrect matrix.
        https://bugs.webkit.org/show_bug.cgi?id=27183

        Rewrite getCTM() / getScreenCTM() handling in an iterative way, fixing all known problems/limitations.
        The bug mentioned above is actually not a regression, getScreenCTM() only worked before, because we
        did not handle non-SVG CSS box parents properly. When support was added to handle those cases, the
        getScreenCTM() handling was completly off, causing a lot of trouble in real-life SVG applications (carto.net for instance)

        Tests: svg/custom/svgsvgelement-ctm.xhtml (fixed typo, missing unit identifier in CSS 'height' property, leading to incorrect results)
               svg/custom/svgsvgelement-ctm2.xhtml
               svg/custom/svgsvgelement-ctm3.xhtml
               svg/custom/svgsvgelement-ctm4.xhtml
               svg/custom/svgsvgelement-ctm5.xhtml

        * svg/SVGLocatable.cpp:
        (WebCore::SVGLocatable::getBBox):
        (WebCore::SVGLocatable::computeCTM):
        * svg/SVGLocatable.h:
        (WebCore::SVGLocatable::localCoordinateSpaceTransform):
        * svg/SVGSVGElement.cpp:
        (WebCore::SVGSVGElement::localCoordinateSpaceTransform):
        (WebCore::SVGSVGElement::createRenderer):
        * svg/SVGSVGElement.h:
        * svg/SVGStyledElement.cpp:
        (WebCore::SVGStyledElement::localCoordinateSpaceTransform):
        * svg/SVGStyledElement.h:
        * svg/SVGStyledLocatableElement.cpp:
        (WebCore::SVGStyledLocatableElement::getCTM):
        (WebCore::SVGStyledLocatableElement::getScreenCTM):
        * svg/SVGStyledLocatableElement.h:
        (WebCore::SVGStyledLocatableElement::localCoordinateSpaceTransform):
        * svg/SVGStyledTransformableElement.cpp:
        (WebCore::SVGStyledTransformableElement::getCTM):
        (WebCore::SVGStyledTransformableElement::getScreenCTM):
        * svg/SVGStyledTransformableElement.h:
        (WebCore::SVGStyledTransformableElement::localCoordinateSpaceTransform):
        * svg/SVGTextElement.cpp:
        (WebCore::SVGTextElement::getCTM):
        (WebCore::SVGTextElement::getScreenCTM):
        * svg/SVGTextElement.h:
        (WebCore::SVGTextElement::localCoordinateSpaceTransform):
        * svg/SVGTransformable.cpp:
        (WebCore::SVGTransformable::SVGTransformable):
        * svg/SVGTransformable.h:
        (WebCore::SVGTransformable::localCoordinateSpaceTransform):
2010-03-31  Nikolas Zimmermann  <nzimmermann@rim.com>

        Reviewed by Dirk Schulze.

        REGRESSION: document.documentElement.getScreenCTM() returns incorrect matrix.
        https://bugs.webkit.org/show_bug.cgi?id=27183

        Share code between the different svgsvgelement-ctm*.xhtm tests, placed in svg/custom/resources.
        Add several new tests covering getCTM()/getScreenCTM() on text, inner and outer svg elements and containers.

        * svg/custom/resources/svgsvgelement-ctm.js: Added.
        (log):
        (stringForMatrix):
        (printCTMs):
        * svg/custom/svgsvgelement-ctm-expected.txt:
        * svg/custom/svgsvgelement-ctm.xhtml:
        * svg/custom/svgsvgelement-ctm2-expected.txt: Added.
        * svg/custom/svgsvgelement-ctm2.xhtml: Added.
        * svg/custom/svgsvgelement-ctm3-expected.txt: Added.
        * svg/custom/svgsvgelement-ctm3.xhtml: Added.
        * svg/custom/svgsvgelement-ctm4-expected.txt: Added.
        * svg/custom/svgsvgelement-ctm4.xhtml: Added.
        * svg/custom/svgsvgelement-ctm5-expected.txt: Added.
        * svg/custom/svgsvgelement-ctm5.xhtml: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56905 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Not reviewed. Build fix.

Adding test expectations missing in r56873.

* platform/chromium-mac/media/audio-constructor-src-expected.txt: Added.
* platform/chromium-mac/media/video-source-add-src-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56904 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Vitaly Repeshko  <vitalyr@chromium.org>

        Reviewed by Darin Fisher.

        [V8] Quick fix for failure in SerializedScriptValue
        https://bugs.webkit.org/show_bug.cgi?id=36943

        This allows the web worker layout tests to pass again.

        * bindings/v8/SerializedScriptValue.h:
        (WebCore::SerializedScriptValue::deserializeAndSetProperty):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56903 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Eric Seidel  <eric@webkit.org>

        Rubber-stamped by Adam Barth.

        Add Snow Leopard Release bot to the list of "core builders" (builders which stop the commit-queue when they turn red)
        https://bugs.webkit.org/show_bug.cgi?id=33292

        Just adding the "Build" builder for now.
        We'll add the "Test" builders when the tests
        are less flaky.

        * Scripts/webkitpy/common/net/buildbot.py:
        * Scripts/webkitpy/common/net/buildbot_unittest.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56902 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Yury Semikhatsky  <yurys@chromium.org>

        Reviewed by David Levin.

        Add myself to the reviewers list.

        https://bugs.webkit.org/show_bug.cgi?id=36935

        * Scripts/webkitpy/common/config/committers.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56901 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Roland Steiner  <rolandsteiner@chromium.org>

        Unreviewed - more layout test result build fixes:
        .) Fix expected layout test filenames for Chromium.
        .) Add new layout test to gtk Skipped list.

        Original patch review by Darin Adler (no functional change):

        Bug 36803 - 'Inline-table' workaround for <ruby> broken with ruby support
        https://bugs.webkit.org/show_bug.cgi?id=36803

        * platform/chromium-linux/fast/ruby/ruby-inline-table-expected.checksum: Renamed from LayoutTests/platform/chromium-linux/fast/ruby/ruby-inline-table-actual.checksum.
        * platform/chromium-linux/fast/ruby/ruby-inline-table-expected.png: Renamed from LayoutTests/platform/chromium-linux/fast/ruby/ruby-inline-table-actual.png.
        * platform/chromium-linux/fast/ruby/ruby-inline-table-expected.txt: Renamed from LayoutTests/platform/chromium-linux/fast/ruby/ruby-inline-table-actual.txt.
        * platform/chromium-win/fast/ruby/ruby-inline-table-expected.checksum: Renamed from LayoutTests/platform/chromium-win/fast/ruby/ruby-inline-table-actual.checksum.
        * platform/chromium-win/fast/ruby/ruby-inline-table-expected.png: Renamed from LayoutTests/platform/chromium-win/fast/ruby/ruby-inline-table-actual.png.
        * platform/chromium-win/fast/ruby/ruby-inline-table-expected.txt: Renamed from LayoutTests/platform/chromium-win/fast/ruby/ruby-inline-table-actual.txt.
        * platform/gtk/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56900 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Added Python style checking to check-webkit-style using
the third-party pep8 module (via autoinstall).

Reviewed by Dave Levin.

https://bugs.webkit.org/show_bug.cgi?id=33639

* Scripts/webkitpy/style/checker.py:
  - Added PYTHON to FileType.
  - Updated ProcessorDispatcher to return a PythonProcessor
    for *.py files.

* Scripts/webkitpy/style/checker_unittest.py:
  - Updated the ProcessorDispatcher unit tests for *.py files.

* Scripts/webkitpy/style/processors/python.py: Added.
  - Added PythonProcessor class.

* Scripts/webkitpy/style/processors/python_unittest.py: Added.
  - Added PythonProcessor unit tests.

* Scripts/webkitpy/style/processors/python_unittest_input.py: Added.
  - Added a sample Python file to test the PythonProcessor.process()
    code path (since pep8 accepts a file path).

* Scripts/webkitpy/style/unittests.py:
  - Updated the style unit test file to import python_unittest.py.

* Scripts/webkitpy/style_references.py:
  - Adjusted style references to import pep8.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56899 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Eric Seidel  <eric@webkit.org>

        Reviewed by Adam Barth.

        fast/replaced/border-radius-clip.html failed on Snow Leopard Debug bot
        https://bugs.webkit.org/show_bug.cgi?id=36642

        This test is flaky (seemingly due to <video> load times)
        and is thus useless until fixed.  Skipping for now.

        * platform/mac/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56898 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rewrote autoinstall.py to support unzipping *.zip files after
download, unzipping and extracting *.tar.gz files after download,
and copying installed files to a given destination directory.

Reviewed by Dave Levin.

https://bugs.webkit.org/show_bug.cgi?id=35163

These changes will let us autoinstall pep8.py from the web and put
our third-party autoinstalled code in an explicit directory like
webkitpy/thirdparty/autoinstalled. These changes should also speed
up the execution of autoinstalled *.zip packages slightly since
*.pyc files cannot be generated when importing from zipped
packages using the current autoinstall.

* Scripts/test-webkitpy:
  - Addressed the FIXME to enable auto-install logging once
    autoinstall was rewritten not to log as verbosely.

* Scripts/webkitpy/common/net/bugzilla.py:
  - Updated mechanize import statement.

* Scripts/webkitpy/common/net/buildbot.py:
  - Updated mechanize import statement.

* Scripts/webkitpy/common/net/irc/ircbot.py:
  - Updated ircbot and irclib import statements.

* Scripts/webkitpy/common/net/networktransaction.py:
  - Updated mechanize import statement.

* Scripts/webkitpy/common/net/networktransaction_unittest.py:
  - Updated mechanize import statement.

* Scripts/webkitpy/common/net/statusserver.py:
  - Updated mechanize import statement.

* Scripts/webkitpy/common/system/autoinstall.py: Added.
  - Added AutoInstaller class.
  - Added sample/testing usage to the __main__ block.

* Scripts/webkitpy/thirdparty/__init__.py:
  - Updated the autoinstall lines to use the new autoinstall methods.
  - Added pep8.py to the list of auto-installed packages.
  - Added a README file to the target autoinstallation directory
    so users know that the directory is safe to delete.

* Scripts/webkitpy/thirdparty/autoinstall.py: Removed.
  - This is replaced by the rewritten autoinstall
    webkitpy/common/system/autoinstall.py.

* Scripts/webkitpy/thirdparty/autoinstalled/__init__.py: Removed.
  - The target autoinstallation directory is now auto-generated.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56897 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-04-01  Roland Steiner  <rolandsteiner@chromium.org>

        Unreviewed - add missing Chromium layout test results for:

        Bug 36803 - 'Inline-table' workaround for <ruby> broken with ruby support
        https://bugs.webkit.org/show_bug.cgi?id=36803

        (original patch review by Darin Adler).

        * platform/chromium-linux/fast/ruby/ruby-inline-table-expected.checksum: Added.
        * platform/chromium-linux/fast/ruby/ruby-inline-table-expected.png: Added.
        * platform/chromium-linux/fast/ruby/ruby-inline-table-expected.txt: Added.
        * platform/chromium-win/fast/ruby/ruby-inline-table-expected.checksum: Added.
        * platform/chromium-win/fast/ruby/ruby-inline-table-expected.png: Added.
        * platform/chromium-win/fast/ruby/ruby-inline-table-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56896 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  MORITA Hajime  <morrita@google.com>

        Reviewed by Darin Adler.

        Crash when writing into a detached TITLE element
        https://bugs.webkit.org/show_bug.cgi?id=25567

        * fast/dom/title-content-write-set-expected.txt: Added.
        * fast/dom/title-content-write-set.html: Added.
2010-03-31  MORITA Hajime  <morrita@google.com>

        Reviewed by Darin Adler.

        Crash when writing into a detached TITLE element
        https://bugs.webkit.org/show_bug.cgi?id=25567

        Document::setTitle() invoked HTMLTitleElement::setText(), which
        contains DOM tree modification, even when setTitle() is called
        from HTMLTitleElement::childrenChanged().  Fix to skip setText()
        when setTitle() is called childrenChanged() to avoid cascading
        DOM mutations between Document and HTMLTitleElement.

        Test: fast/dom/title-content-write-set.html

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

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56895 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Tweak webkitpy's logtesting.LogTesting class to get more mileage out
of our unit tests that test log messages.

Reviewed by Shinichiro Hamaji.

https://bugs.webkit.org/show_bug.cgi?id=36886

This patch adds to the LogTesting class's tearDown() method a line
asserting that the array of remaining log messages is empty.  This
ensures that no extra log messages are getting logged that the caller
might not be aware of or may have forgotten to check for.

* Scripts/webkitpy/common/system/logtesting.py:
  - Modified the tearDown() method as described above.
  - Also modified the assertMessages() method to clear the array
    of log messages after asserting.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56894 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Added support for a --verbose-logging flag to test-webkitpy.

Reviewed by Adam Barth.

https://bugs.webkit.org/show_bug.cgi?id=36429

Verbose logging is useful for debugging test-webkitpy code that
runs before the actual unit tests -- things like autoinstalling and
unit-test auto-detection logic.  This is different from verbose
logging of the unit tests themselves (which corresponds to the
unittest module's --verbose flag).

* Scripts/test-webkitpy:
  - In the configure_logging() method--
    - Added an is_verbose_logging parameter that sets the logging
      level to logging.DEBUG instead of logging.INFO.
    - Changed the method to throttle the logging level on the
      root logger's handler instead of directly on the root logger
      itself.
    - Enabled logging of the autoinstall module when the flag is set.

* Scripts/webkitpy/thirdparty/autoinstalled/__init__.py:
  - Added a work-around for a bug in Python 2.6's logging module
    that was discovered while working on this patch.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56893 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Enabled Python's logging module for webkit-patch, and replaced
deprecated_logging with Python logging in networktransaction.py.
This eliminates some spurious output when running test-webkitpy.

Reviewed by Adam Barth.

https://bugs.webkit.org/show_bug.cgi?id=36725

* Scripts/webkit-patch:
  - Replaced the configure_logging() function with a call to
    the new logutils.configure_logging() function.

* Scripts/webkitpy/common/net/networktransaction.py:
  - Replaced the use of deprecated_logging with Python logging.

* Scripts/webkitpy/common/system/logutils.py:
  - Added _default_handlers() which creates the default logging
    handler for webkitpy.
  - Added configure_logging() which configures default logging
    for webkitpy.

* Scripts/webkitpy/common/system/logutils_unittest.py:
  - Added unit tests for logutils.configure_logging().

* Scripts/webkitpy/style/checker.py:
  - Refactored check-webkit-style's configure_logging() method
    to call the new logutils.configure_logging().

* Scripts/webkitpy/style_references.py:
  - Updated references as necessary.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56892 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Eric Seidel  <eric@webkit.org>

        Reviewed by Adam Barth.

        webkit-patch land fails if not run from the root directory
        https://bugs.webkit.org/show_bug.cgi?id=35822

        The root of the problem was that ChangeLog.__init__ expects a path
        relative to the current working directory, and SCM expects to
        return paths relative to the SCM root.  Fix it by converting from
        SCM-relative to absolute paths in Checkout.modified_changelogs

        * Scripts/webkitpy/common/checkout/api.py:
        * Scripts/webkitpy/common/checkout/api_unittest.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56891 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-29  Roland Steiner  <rolandsteiner@chromium.org>

        Reviewed by Darin Adler.

        Bug 36803 - 'Inline-table' workaround for <ruby> broken with ruby support
        https://bugs.webkit.org/show_bug.cgi?id=36803

        Don't use ruby rendering if the underlying 'display' property has been
        changed into a value other than 'inline' or 'block' (such as 'inline-table').

        Test: fast/ruby/ruby-inline-table.html

        * rendering/RenderObject.cpp:
        (WebCore::RenderObject::createObject):

        Add layout test to confirm that 'inline-table' rendering is not affected
        by ruby implementation.

        * fast/ruby/ruby-inline-table.html: Added.
        * platform/mac/fast/ruby/ruby-inline-table-expected.checksum: Added.
        * platform/mac/fast/ruby/ruby-inline-table-expected.png: Added.
        * platform/mac/fast/ruby/ruby-inline-table-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56890 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        Add more tests for webkitpy.common.checkout
        https://bugs.webkit.org/show_bug.cgi?id=36926

        We don't have a great way of testing checkout, sadly.

        * Scripts/webkitpy/common/checkout/api_unittest.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56889 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        REGRESSION: EWS crashes on failure
        https://bugs.webkit.org/show_bug.cgi?id=36924

        Turns out we need to pass one more argument.  My test is kind of lame,
        but at least it's there.

        * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
        * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56888 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed.

Fixed typo in WebKitTools/ChangeLog: opsys -> ospath.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56887 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Zhenyao Mo  <zmo@google.com>

        Reviewed by Eric Seidel.

        Fix 3 failed tests on Snow Leopard bot from 33416
        https://bugs.webkit.org/show_bug.cgi?id=36910

        * fast/canvas/webgl/context-attributes-alpha-depth-stencil-antialias-expected.txt: Considering the case where EXT_packed_depth_stencil is unsupported.
        * fast/canvas/webgl/context-attributes-alpha-depth-stencil-antialias.html: Ditto.
        * fast/canvas/webgl/context-attributes-expected.txt: Ditto.
        * fast/canvas/webgl/context-attributes.html: Ditto.
        * fast/canvas/webgl/gl-get-calls-expected.txt: Ditto.
        * fast/canvas/webgl/gl-get-calls.html: Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56886 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Crash submitting display:none textarea in a form
https://bugs.webkit.org/show_bug.cgi?id=36905

Reviewed by Darin Adler.

WebCore:

Test: fast/forms/textarea-submit-crash.html

* html/HTMLTextAreaElement.cpp:
(WebCore::HTMLTextAreaElement::appendFormData): Do update layout before
asking our renderer for its text, since we can't rely on our renderer's
text if layout is needed.

* rendering/RenderTextControl.cpp:
(WebCore::RenderTextControl::textWithHardLineBreaks): Don't update layout
while being asked for our text, since doing so may delete us, causing a crash.

LayoutTests:

* fast/forms/textarea-submit-crash-expected.txt: Added.
* fast/forms/textarea-submit-crash.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56885 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        Make the EWS go faster by being optimistic
        https://bugs.webkit.org/show_bug.cgi?id=36916

        Have the EWS be optimistic that a patch will correctly build.  This
        should speed up the common case by not requiring two builds for every
        patch.

        * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
        * Scripts/webkitpy/tool/commands/queues.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56884 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        Add the ability to restart sheriffbot from IRC
        https://bugs.webkit.org/show_bug.cgi?id=36909

        * Scripts/webkitpy/tool/bot/irc_command.py:
        * Scripts/webkitpy/tool/bot/queueengine.py:
        * Scripts/webkitpy/tool/bot/sheriffircbot.py:
        * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56883 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore: Bug 36845 - AX: need a way to set the label of a AXWebArea through WebKit
https://bugs.webkit.org/show_bug.cgi?id=36845

Reviewed by Darin Adler.

Provide a way through WebKit to set an accessible label that describes the web area.

Test: platform/mac/accessibility/html-with-aria-label.html

* accessibility/AccessibilityObject.h:
(WebCore::AccessibilityObject::setAccessibleName):
* accessibility/AccessibilityRenderObject.cpp:
(WebCore::AccessibilityRenderObject::accessibilityDescription):
(WebCore::AccessibilityRenderObject::setAccessibleName):
* accessibility/AccessibilityRenderObject.h:

WebKit/mac: Bug 36845 - AX: need a way to set the label of a AXWebArea through WebKit
https://bugs.webkit.org/show_bug.cgi?id=36845

Reviewed by Darin Adler.

Provide a way through WebKit to set an accessible label that describes the web area.

* WebView/WebFrame.mm:
(-[WebFrame setAccessibleName:]):
* WebView/WebFramePrivate.h:

LayoutTests: Test for
Bug 36845 - AX: need a way to set the label of a AXWebArea through WebKit
https://bugs.webkit.org/show_bug.cgi?id=36845

Reviewed by Darin Adler.

* platform/mac/accessibility/html-with-aria-label-expected.txt: Added.
* platform/mac/accessibility/html-with-aria-label.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56882 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        Add win-ews to QueueStatusServer
        https://bugs.webkit.org/show_bug.cgi?id=36876

        The win-ews is still experimental, but it seems to be more or less
        running.  We should show its results to the people.

        * QueueStatusServer/model/queues.py:
        * QueueStatusServer/templates/dashboard.html:
        * QueueStatusServer/templates/statusbubble.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56881 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Not reviewed. Build fix.

Fixing a build break caused by 56872. One of the Chromium bots doesn't
like std::strstr(). Includes string.h and use strstr() instead.

* src/WebGraphicsContext3DDefaultImpl.cpp:
(WebKit::WebGraphicsContext3DDefaultImpl::validateAttributes):
Includes string.h and use strstr().

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56880 268f45cc-cd09-0410-ab3c-d52691b4dbfc

<webkit.org/b/36878> REGRESSION: Trailing colon on hostnames (with no port specified) causes "Not allowed to use restricted network port"

Reviewed by Darin Adler.

WebCore:

* platform/KURL.cpp:
(WebCore::KURL::port): Explicitly handle the case of a colon being present in the URL after the host name but with
no port number before the path.  This is handled in the same manner as the colon and port being omitted completely.

LayoutTests:

Add explicit tests for handling of allowed ports.  We currently only test the ports that map to port 80: 0, 80, and "".

* platform/mac/security/block-test-expected.txt:
* security/block-test.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56879 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Remove dependency on TextBreakIterator from StringImpl.cpp,
and from member methods of String.

Reviewed by Oliver Hunt.

Make 'numGraphemeClusters' & 'numCharactersInGraphemeClusters'
global function defined in PlatformString.h, rather than
member methods on String (these can be implemented purely
using the characters() and length() public interface),
and make 'makeCapitalized' a static function in RenderText.cpp.

* dom/InputElement.cpp:
(WebCore::InputElement::sanitizeUserInputValue):
(WebCore::InputElement::handleBeforeTextInsertedEvent):
* html/HTMLInputElement.cpp:
(WebCore::HTMLInputElement::tooLong):
* html/HTMLTextAreaElement.cpp:
(WebCore::HTMLTextAreaElement::handleBeforeTextInsertedEvent):
(WebCore::HTMLTextAreaElement::sanitizeUserInputValue):
(WebCore::HTMLTextAreaElement::tooLong):
* platform/text/PlatformString.h:
(WebCore::String::makeSecure):
* platform/text/String.cpp:
(WebCore::numGraphemeClusters):
(WebCore::numCharactersInGraphemeClusters):
* platform/text/StringImpl.cpp:
* platform/text/StringImpl.h:
* rendering/RenderText.cpp:
(WebCore::makeCapitalized):
(WebCore::RenderText::setTextInternal):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56878 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Vitaly Repeshko  <vitalyr@chromium.org>

        Reviewed by David Levin.

        [V8] SerializedScriptValue must be deserialized only once and in the right context
        https://bugs.webkit.org/show_bug.cgi?id=36892

        See also https://bugs.webkit.org/show_bug.cgi?id=34227 for the
        corresponding JSC change.

        General idea: SerializedScriptValue must be deserialized only once
        and in the context of the intended MessageEvent recepient. The
        approach we take for now is to eagerly deserialize when a
        JavaScript wrapper for MessageEvent is created.

        A better fix would be to keep a reference to the context in
        MessageEvent and use it when lazily deserializing. It's harder to
        do since the API doesn't have a clean method to have such a reference.

        Tested by fast/dom/Window/window-postmessage-clone-frames.html. This
        test still fails but only for the types which we can't serialize yet.

        * bindings/scripts/CodeGeneratorV8.pm:
        * bindings/v8/SerializedScriptValue.h:
        (WebCore::SerializedScriptValue::deserializeAndSetProperty):
        * bindings/v8/custom/V8MessageEventCustom.cpp:
        (WebCore::V8MessageEvent::initMessageEventCallback):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56877 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-30  Victor Wang  <victorw@chromium.org>

        Reviewed by Eric Seidel.

        Update rebaseline tool to check the release image diff binary and
        fallback to debug if the release version does not exist.

        https://bugs.webkit.org/show_bug.cgi?id=36245

        * Scripts/webkitpy/layout_tests/port/base.py:
        * Scripts/webkitpy/layout_tests/port/chromium.py:
        * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56876 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Adam Barth  <abarth@webkit.org>

        Reviewed by Darin Fisher.

        Cleanup RedirectScheduler
        https://bugs.webkit.org/show_bug.cgi?id=36874

        Removed the nutty ScheduledRedirection struct in favor of a hierarchy
        of classes to represent the various kinds of scheduled redirects.
        Doing this lets us get rid of the pseudo RTTI switch statements on
        "type" in favour of calling virtual functions.

        No new tests because this change should be API identical with the
        existing RedirectScheduler.

        * loader/RedirectScheduler.cpp:
        (WebCore::ScheduledNavigation::ScheduledNavigation):
        (WebCore::ScheduledNavigation::~ScheduledNavigation):
        (WebCore::ScheduledNavigation::isLocationChange):
        (WebCore::ScheduledNavigation::shouldStartTimer):
        (WebCore::ScheduledNavigation::didStartTimer):
        (WebCore::ScheduledNavigation::didStopTimer):
        (WebCore::ScheduledNavigation::delay):
        (WebCore::ScheduledNavigation::lockHistory):
        (WebCore::ScheduledNavigation::lockBackForwardList):
        (WebCore::ScheduledNavigation::wasDuringLoad):
        (WebCore::ScheduledURLNavigation::ScheduledURLNavigation):
        (WebCore::ScheduledURLNavigation::fire):
        (WebCore::ScheduledURLNavigation::didStartTimer):
        (WebCore::ScheduledURLNavigation::didStopTimer):
        (WebCore::ScheduledURLNavigation::url):
        (WebCore::ScheduledURLNavigation::referrer):
        (WebCore::ScheduledURLNavigation::wasUserGesture):
        (WebCore::ScheduledRedirect::ScheduledRedirect):
        (WebCore::ScheduledRedirect::isLocationChange):
        (WebCore::ScheduledRedirect::shouldStartTimer):
        (WebCore::ScheduledLocationChange::ScheduledLocationChange):
        (WebCore::ScheduledRefresh::ScheduledRefresh):
        (WebCore::ScheduledRefresh::fire):
        (WebCore::ScheduledHistoryNavigation::ScheduledHistoryNavigation):
        (WebCore::ScheduledHistoryNavigation::fire):
        (WebCore::ScheduledFormSubmission::ScheduledFormSubmission):
        (WebCore::ScheduledFormSubmission::fire):
        (WebCore::RedirectScheduler::redirectScheduledDuringLoad):
        (WebCore::RedirectScheduler::clear):
        (WebCore::RedirectScheduler::scheduleRedirect):
        (WebCore::RedirectScheduler::mustLockBackForwardList):
        (WebCore::RedirectScheduler::scheduleLocationChange):
        (WebCore::RedirectScheduler::scheduleFormSubmission):
        (WebCore::RedirectScheduler::scheduleRefresh):
        (WebCore::RedirectScheduler::locationChangePending):
        (WebCore::RedirectScheduler::scheduleHistoryNavigation):
        (WebCore::RedirectScheduler::timerFired):
        (WebCore::RedirectScheduler::schedule):
        (WebCore::RedirectScheduler::startTimer):
        (WebCore::RedirectScheduler::cancel):
        * loader/RedirectScheduler.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56875 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Eric Seidel  <eric@webkit.org>

        Reviewed by Adam Barth.

        Teach buildbot.py how to parse builder activity from /one_box_per_builder
        https://bugs.webkit.org/show_bug.cgi?id=36898

        I also removed some obsolete FIXMEs and
        refactored one_box_per_builder parsing into multiple
        methods for easier reading.

        * Scripts/webkitpy/common/net/buildbot.py:
        * Scripts/webkitpy/common/net/buildbot_unittest.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56874 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Andrew Scherkus  <scherkus@chromium.org>

        Reviewed by Dmitry Titov.

        Checking in Chromium media baselines for layout tests fixed in r56481.

        https://bugs.webkit.org/show_bug.cgi?id=36860

        * platform/chromium-linux/media/video-no-audio-expected.checksum: Added.
        * platform/chromium-linux/media/video-no-audio-expected.png: Added.
        * platform/chromium-mac/media/video-no-audio-expected.checksum: Added.
        * platform/chromium-mac/media/video-no-audio-expected.png: Added.
        * platform/chromium-mac/media/video-no-audio-expected.txt: Added.
        * platform/chromium-win/media/audio-constructor-src-expected.txt: Added.
        * platform/chromium-win/media/video-no-audio-expected.checksum: Added.
        * platform/chromium-win/media/video-no-audio-expected.png: Added.
        * platform/chromium-win/media/video-no-audio-expected.txt: Added.
        * platform/chromium-win/media/video-source-add-src-expected.txt: Added.
        * platform/chromium/test_expectations.txt: Tests are now passing.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56873 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Zhenyao Mo  <zmo@google.com>

        Reviewed by Darin Fisher.

        Hook up WebGLContextAttributes to OpenGL context creation code
        https://bugs.webkit.org/show_bug.cgi?id=33416

        * fast/canvas/webgl/context-attributes-alpha-depth-stencil-antialias-expected.txt: Added.
        * fast/canvas/webgl/context-attributes-alpha-depth-stencil-antialias.html: Added.
        * fast/canvas/webgl/context-attributes-expected.txt: WebGL context attributes behavior changed with this fix.
        * fast/canvas/webgl/context-attributes.html: Ditto.
        * fast/canvas/webgl/gl-get-calls-expected.txt: Stencil buffer is enabled, thus Stencil Bits is no longer 0.
        * fast/canvas/webgl/gl-get-calls.html: Ditto.
2010-03-31  Zhenyao Mo  <zmo@google.com>

        Reviewed by Darin Fisher.

        Hook up WebGLContextAttributes to OpenGL context creation code
        https://bugs.webkit.org/show_bug.cgi?id=33416

        Test: fast/canvas/webgl/context-attributes-alpha-depth-stencil-antialias.html

        * bindings/v8/custom/V8HTMLCanvasElementCustom.cpp: Fix an index bug.
        (WebCore::V8HTMLCanvasElement::getContextCallback):
        * platform/graphics/GraphicsContext3D.h: Add members/functions for multisampling/stencil buffer purpose.
        * platform/graphics/mac/Canvas3DLayer.h: Add GraphicsContext3D as a member of Canvas3DLayer.
        * platform/graphics/mac/Canvas3DLayer.mm: Add multisampling support.
        (-[Canvas3DLayer drawInCGLContext:pixelFormat:forLayerTime:displayTime:]):
        * platform/graphics/mac/GraphicsContext3DMac.cpp: Hook up WebGLContextAttributes to OpenGL context creation code for Mac.
        (WebCore::GraphicsContext3D::GraphicsContext3D):
        (WebCore::GraphicsContext3D::~GraphicsContext3D):
        (WebCore::GraphicsContext3D::validateAttributes):
        (WebCore::GraphicsContext3D::reshape):
        (WebCore::GraphicsContext3D::prepareTexture):
        (WebCore::GraphicsContext3D::bindFramebuffer):
        (WebCore::GraphicsContext3D::readPixels):
        * platform/graphics/mac/GraphicsLayerCA.mm: Adjust to modified Canvas3DLayer init call.
        (WebCore::GraphicsLayerCA::setContentsToGraphicsContext3D):
2010-03-31  Zhenyao Mo  <zmo@google.com>

        Reviewed by Darin Fisher.

        Hook up WebGLContextAttributes to OpenGL context creation code
        https://bugs.webkit.org/show_bug.cgi?id=33416

        * src/WebGraphicsContext3DDefaultImpl.cpp: Hook up WebGLContextAttributes to OpenGL context creation code for Chrome.
        (WebKit::WebGraphicsContext3DDefaultImpl::WebGraphicsContext3DDefaultImpl):
        (WebKit::WebGraphicsContext3DDefaultImpl::~WebGraphicsContext3DDefaultImpl):
        (WebKit::WebGraphicsContext3DDefaultImpl::initialize):
        (WebKit::WebGraphicsContext3DDefaultImpl::validateAttributes):
        (WebKit::WebGraphicsContext3DDefaultImpl::reshape):
        (WebKit::WebGraphicsContext3DDefaultImpl::readBackFramebuffer):
        (WebKit::WebGraphicsContext3DDefaultImpl::bindFramebuffer):
        (WebKit::WebGraphicsContext3DDefaultImpl::readPixels): Deal with wrong returned alpha values in Mac.
        * src/WebGraphicsContext3DDefaultImpl.h: Add a function.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56872 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Dirk Pranke  <dpranke@chromium.org>

        Reviewed by Eric Seidel.

        Fix logging in new-run-webkit-tests so that we don't report IMAGE
        expected failures as unexpected passes when we run with pixel tests
        disabled.

        This change splits some of the logic embedded into the TestExpectations
        classes out into separate pure functions (result_was_expected,
        remove_image_failures) to make them easier to test. This also adds
        a parameter to matches_an_expected_result() to indicate whether or
        not pixel test results should be included in the expectations.

        https://bugs.webkit.org/show_bug.cgi?id=36771

        * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
        * Scripts/webkitpy/layout_tests/layout_package/test_expectations_test.py:
        * Scripts/webkitpy/layout_tests/run_webkit_tests.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56871 268f45cc-cd09-0410-ab3c-d52691b4dbfc

In webkitpy, refactored two calls to os.path.relpath() replacements
to use a common method.

Reviewed by Eric Seidel.

https://bugs.webkit.org/show_bug.cgi?id=36891

* Scripts/webkitpy/common/net/bugzilla.py:
  - Replaced the relpath implementation with a call to
    webkitpy.common.system.opsys.relpath().

* Scripts/webkitpy/common/system/ospath.py: Added.
  - Moved the relpath() implementation from style/main.py.

* Scripts/webkitpy/common/system/ospath_unittest.py: Added.
  - Moved the relpath() unit tests from style/main_unittest.py.

* Scripts/webkitpy/style/main.py:
  - Replaced the relpath implementation with a call to
    webkitpy.common.system.opsys.relpath().

* Scripts/webkitpy/style/main_unittest.py:
  - Moved the relpath unit tests to opsys_unittest.py.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56870 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add support to create UUID string.
https://bugs.webkit.org/show_bug.cgi?id=36472

Reviewed by Dmitry Titov.

* GNUmakefile.am:
* WebCore.gypi:
* WebCore.pro:
* WebCore.vcproj/WebCore.vcproj:
* WebCore.xcodeproj/project.pbxproj:
* platform/UUID.cpp: Added.
(WebCore::createCanonicalUUIDString):
* platform/UUID.h: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56869 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:

Reviewed by NOBODY (windows build fix pt II).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56868 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:

Reviewed by NOBODY (windows build fix).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56867 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        Add a Builder.force_build method
        https://bugs.webkit.org/show_bug.cgi?id=36875

        We plan to eventually use this in SheriffBot to break deadlocks created
        by flaky tests.

        * Scripts/webkitpy/common/net/buildbot.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56866 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Qi Zhang  <qi.2.zhang@nokia.com>

        Reviewed by Laszlo Gombos.

        [Qt] Enable LayoutTests/fast/canvas/zero-size-fill-rect.html
        https://bugs.webkit.org/show_bug.cgi?id=36837

        Remove zero-size-fill-rect.html from skipped and create expected files.

        * platform/qt/Skipped:
        * platform/qt/fast/canvas/zero-size-fill-rect-expected.checksum: Added.
        * platform/qt/fast/canvas/zero-size-fill-rect-expected.png: Added.
        * platform/qt/fast/canvas/zero-size-fill-rect-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56865 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Bug 36871 - Remove JSC::CString
Use WTF::CString instead (which until recently was WebCore::CString).

Reviewed by Geoff Garen.

JavaScriptCore:

* JavaScriptCore.exp:
* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:
* runtime/UString.cpp:
* runtime/UString.h:

JavaScriptGlue:

* ForwardingHeaders/wtf/text: Added.
* ForwardingHeaders/wtf/text/CString.h: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56864 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        webkit-patch reads wrong bug url from unified diff context
        https://bugs.webkit.org/show_bug.cgi?id=36477

        Instead of trying to figure out the bug_id from the diff, we should
        just get the information from the Checkout object, which understands
        these concepts.

        * Scripts/webkitpy/common/checkout/api.py:
        * Scripts/webkitpy/common/checkout/commitinfo.py:
        * Scripts/webkitpy/tool/commands/download.py:
        * Scripts/webkitpy/tool/commands/upload.py:
        * Scripts/webkitpy/tool/mocktool.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56863 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* rendering/RenderThemeChromiumWin.h: Fix inconsistent line endings.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56862 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Alpha Lam  <hclam@chromium.org>

        Reviewed by Darin Fisher.

        Update test results due to changes in r56824.
        https://bugs.webkit.org/show_bug.cgi?id=36900

        * platform/chromium-win/fast/forms/input-file-re-render-expected.checksum:
        * platform/chromium-win/fast/forms/input-file-re-render-expected.png:
        * platform/chromium/test_expectations.txt:
        fast/forms/input-file-re-render.html generates incorrect results on
        Linux, mark it so.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56861 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Luiz Agostini  <luiz.agostini@openbossa.org>

        Reviewed by Kenneth Rohde Christiansen.

        [Qt] Maemo5 theme - form controls style
        https://bugs.webkit.org/show_bug.cgi?id=36370

        Adjusting Mameo5 form elements rendering.

        * WebCore.pri:
        * WebCore.pro:
        * css/themeQtMaemo5.css: Added.
        (select):
        (select:disabled):
        (select:active):
        (select:active:disabled):
        (textarea):
        (textarea:disabled):
        (textarea:active):
        * platform/qt/Maemo5Webstyle.cpp: Added.
        (Maemo5WebStyle::Maemo5WebStyle):
        (drawRectangularControlBackgorund):
        (Maemo5WebStyle::drawChecker):
        (Maemo5WebStyle::findChecker):
        (Maemo5WebStyle::drawRadio):
        (Maemo5WebStyle::findRadio):
        (Maemo5WebStyle::drawControl):
        (Maemo5WebStyle::drawComplexControl):
        * platform/qt/Maemo5Webstyle.h: Added.
        * platform/qt/RenderThemeQt.cpp:
        (WebCore::RenderThemeQt::RenderThemeQt):
        (WebCore::RenderThemeQt::isControlStyled):
        (WebCore::RenderThemeQt::popupInternalPaddingBottom):
        (WebCore::RenderThemeQt::extraDefaultStyleSheet):
        (WebCore::RenderThemeQt::adjustMenuListButtonStyle):
        * platform/qt/RenderThemeQt.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56860 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    https://bugs.webkit.org/show_bug.cgi?id=36897
        <rdar://problem/7804018> REGRESSION (r56429): Flash ads are clipped when main page is scrolled (boxofficemojo.com)

        Addressing additional review comments.

        * rendering/RenderWidget.cpp: Removed an unneeded include.
        * rendering/RenderWidget.h: Reworded the new comment.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56859 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Dan Bernstein.

        https://bugs.webkit.org/show_bug.cgi?id=36897
        <rdar://problem/7804018> REGRESSION (r56429): Flash ads are clipped when main page is scrolled (boxofficemojo.com)

        * manual-tests/plugin-in-iframe-scroll.html: Added.
        * manual-tests/resources/plugin-in-iframe-scroll-iframe.html: Added.

        * rendering/RenderWidget.cpp:
        (WebCore::RenderWidget::setWidgetGeometry): This method isn't called when an iframe containing
        the widget moves, so we should store a value that is not dependent on current scroll offset.
        (WebCore::RenderWidget::windowClipRect): Apply scroll offset and window clip now.

        * rendering/RenderWidget.h: Renamed m_windowClipRect to m_clipRect, because it is no longer
        in window coordinates.

        * WebCore.base.exp: Export RenderWidget::windowClipRect(), since it's no longer inline.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56858 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Adam Barth  <abarth@webkit.org>

        Reviewed by Eric Seidel.

        webkit-patch commit-queue should land patches optimistically
        https://bugs.webkit.org/show_bug.cgi?id=34187

        This patch adds an optimistic path to commit-queue and a "fail twice"
        requirement for rejecting patches.  That means we'll land good patches
        faster (via the optmistic first run) and we'll reject many fewer
        patches due to flaky tests.

        * Scripts/webkitpy/tool/commands/queues.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56857 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed crash fix.

Crash with frame flattening on after r56854
https://bugs.webkit.org/show_bug.cgi?id=36894

Solution tested by Greg Bolsinga, thanks.

* page/FrameView.cpp:
(WebCore::FrameView::avoidScrollbarCreation):
* platform/ScrollView.cpp:
(WebCore::ScrollView::setHasHorizontalScrollbar):
(WebCore::ScrollView::setHasVerticalScrollbar):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56856 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rubber-stamped by Kenneth Rohde Christiansen.

iframe flattening doesn't flatten
https://bugs.webkit.org/show_bug.cgi?id=36798

Platform specific expected files added, and updated after r56852.

* platform/mac/fast/frames/flattening/frameset-flattening-subframe-resize-expected.txt:
* platform/mac/fast/frames/flattening/iframe-flattening-fixed-height-expected.txt: Added.
* platform/mac/fast/frames/flattening/iframe-flattening-fixed-width-and-height-expected.txt: Added.
* platform/mac/fast/frames/flattening/iframe-flattening-fixed-width-and-height-no-scrolling-expected.txt: Added.
* platform/mac/fast/frames/flattening/iframe-flattening-fixed-width-expected.txt: Added.
* platform/mac/fast/frames/flattening/iframe-flattening-simple-expected.txt:
* platform/qt/fast/frames/flattening/frameset-flattening-subframe-resize-expected.txt:
* platform/qt/fast/frames/flattening/iframe-flattening-fixed-height-expected.txt: Added.
* platform/qt/fast/frames/flattening/iframe-flattening-fixed-width-and-height-expected.txt: Added.
* platform/qt/fast/frames/flattening/iframe-flattening-fixed-width-and-height-no-scrolling-expected.txt: Added.
* platform/qt/fast/frames/flattening/iframe-flattening-fixed-width-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56855 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Do not draw scrollbars for subframes when frame flattening is
enabled. Implemented using a virtual method in ScrollView as
suggested by Dave Hyatt.

Reviewed by Antti Koivisto.

Do not suppress scrollbars as that is wrong according to
Dave Hyatt.

Covered by current tests.

* page/FrameView.cpp:
(WebCore::FrameView::avoidScrollbarCreation):
* page/FrameView.h:
* platform/ScrollView.cpp:
(WebCore::ScrollView::setHasHorizontalScrollbar):
(WebCore::ScrollView::setHasVerticalScrollbar):
* platform/ScrollView.h:
(WebCore::ScrollView::avoidScrollbarCreation):
* rendering/RenderPart.cpp:
(WebCore::RenderPart::layoutWithFlattening):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56854 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Darin Fisher  <darin@chromium.org>

        Reviewed by David Levin.

        [chromium] including WebKit.gyp:webkit should automatically add Skia
        and NPAPI include paths

        https://bugs.webkit.org/show_bug.cgi?id=36887

        * WebKit.gyp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56853 268f45cc-cd09-0410-ab3c-d52691b4dbfc

iframe flattening doesn't flatten
https://bugs.webkit.org/show_bug.cgi?id=36798

WebCore: Fix wrongly reversing logic in frame flattening check.

Reviewed by Antti Koivisto.

Tests:
fast/frames/flattening/iframe-flattening-fixed-height.html
fast/frames/flattening/iframe-flattening-fixed-width.html
fast/frames/flattening/iframe-flattening-fixed-width-and-height.html
fast/frames/flattening/iframe-flattening-fixed-width-and-height-no-scrolling.html

* rendering/RenderPartObject.cpp:
(WebCore::RenderPartObject::flattenFrame):

LayoutTests: Add more tests for iframe flattening

Reviewed by Antti Koivisto.

* fast/frames/flattening/iframe-flattening-fixed-height.html
* fast/frames/flattening/iframe-flattening-fixed-width.html
* fast/frames/flattening/iframe-flattening-fixed-width-and-height.html
* fast/frames/flattening/iframe-flattening-fixed-width-and-height-no-scrolling.html

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56852 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Marcus Bulach  <bulach@chromium.org>

        Reviewed by Jeremy Orlow.

        Implements cancelGeolocationPermissionRequestForFrame.
        Similar to requestGeolocationPermissionForFrame(), passes Geolocation* to cancelGeolocationPermissionRequestForFrame(),
        so that the embedder can identify which Geolocation object is cancelling the pending permission request.
        Calls cancelGeolocationPermissionRequestForFrame() before stopUpdating() so it better matches the startUpdating() / requestPermission() flow.
        https://bugs.webkit.org/show_bug.cgi?id=35031

        * loader/EmptyClients.h:
        (WebCore::EmptyChromeClient::cancelGeolocationPermissionRequestForFrame):
        * page/Chrome.cpp:
        (WebCore::Chrome::cancelGeolocationPermissionRequestForFrame):
        * page/Chrome.h:
        * page/ChromeClient.h:
        * page/Geolocation.cpp:
        (WebCore::Geolocation::disconnectFrame):
2010-03-31  Marcus Bulach  <bulach@chromium.org>

        Reviewed by Jeremy Orlow.

        Implements cancelGeolocationPermissionRequestForFrame.
        https://bugs.webkit.org/show_bug.cgi?id=35031

        * public/WebGeolocationService.h:
        (WebKit::WebGeolocationService::cancelPermissionRequestForFrame):
        * src/ChromeClientImpl.cpp:
        (WebKit::ChromeClientImpl::cancelGeolocationPermissionRequestForFrame):
        * src/ChromeClientImpl.h:
2010-03-31  Marcus Bulach  <bulach@chromium.org>

        Reviewed by Jeremy Orlow.

        Adds Geolocation param for cancelGeolocationPermissionRequestForFrame.
        https://bugs.webkit.org/show_bug.cgi?id=35031

        * WebCoreSupport/ChromeClientGtk.cpp:
        (WebKit::ChromeClient::cancelGeolocationPermissionRequestForFrame):
        * WebCoreSupport/ChromeClientGtk.h:
2010-03-31  Marcus Bulach  <bulach@chromium.org>

        Reviewed by Jeremy Orlow.

        Adds Geolocation param for cancelGeolocationPermissionRequestForFrame.
        https://bugs.webkit.org/show_bug.cgi?id=35031

        * WebCoreSupport/ChromeClientHaiku.h:
        (WebCore::ChromeClientHaiku::cancelGeolocationPermissionRequestForFrame):
2010-03-31  Marcus Bulach  <bulach@chromium.org>

        Reviewed by Jeremy Orlow.

        Adds Geolocation param for cancelGeolocationPermissionRequestForFrame.
        https://bugs.webkit.org/show_bug.cgi?id=35031

        * WebCoreSupport/WebChromeClient.h:
        (WebChromeClient::cancelGeolocationPermissionRequestForFrame):
2010-03-31  Marcus Bulach  <bulach@chromium.org>

        Reviewed by Jeremy Orlow.

        Adds Geolocation param for cancelGeolocationPermissionRequestForFrame.
        https://bugs.webkit.org/show_bug.cgi?id=35031

        * WebCoreSupport/ChromeClientQt.h:
        (WebCore::ChromeClientQt::cancelGeolocationPermissionRequestForFrame):
2010-03-31  Marcus Bulach  <bulach@chromium.org>

        Reviewed by Jeremy Orlow.

        Adds Geolocation param for cancelGeolocationPermissionRequestForFrame.
        https://bugs.webkit.org/show_bug.cgi?id=35031

        * WebCoreSupport/WebChromeClient.h:
        (WebChromeClient::cancelGeolocationPermissionRequestForFrame):
2010-03-31  Marcus Bulach  <bulach@chromium.org>

        Reviewed by Jeremy Orlow.

        Adds Geolocation param for cancelGeolocationPermissionRequestForFrame.
        https://bugs.webkit.org/show_bug.cgi?id=35031

        * WebKitSupport/ChromeClientWx.h:
        (WebCore::ChromeClientWx::cancelGeolocationPermissionRequestForFrame):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56851 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add animation to progress element
https://bugs.webkit.org/show_bug.cgi?id=36664

Reviewed by Antti Koivisto.

Add a timer to control the animation. The timer is started after painting
or a state change in the progress bar, to prevent animation from running
when the progress bar is not visible.

* html/HTMLProgressElement.cpp:
* platform/qt/RenderThemeQt.cpp:
* platform/qt/RenderThemeQt.h:
* rendering/RenderProgress.cpp:
* rendering/RenderProgress.h:
* rendering/RenderTheme.cpp:
* rendering/RenderTheme.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56850 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Pavel Feldman  <pfeldman@chromium.org>

        Not reviewed. Rolling out r56829 since it broke chromium layout tests.

        [REGRESSION] Inspector tests started crashing since r56829
        https://bugs.webkit.org/show_bug.cgi?id=36888

        * bindings/scripts/CodeGeneratorV8.pm:
        * bindings/v8/V8Utilities.cpp:
        * bindings/v8/V8Utilities.h:
        * loader/FrameLoaderClient.h:
        WebKit/chromium:
        * public/WebFrameClient.h:
        * src/FrameLoaderClientImpl.cpp:
        * src/FrameLoaderClientImpl.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56849 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Sterling Swigart  <sswigart@google.com>

        Reviewed by David Levin.

        [Chromium] Enable fast/events/onbeforeunload-focused-iframe.html
        https://bugs.webkit.org/show_bug.cgi?id=36781

        * platform/chromium/test_expectations.txt: Enable test now that callShouldCloseOnWebView
        is implemented in test_shell.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56848 268f45cc-cd09-0410-ab3c-d52691b4dbfc

more tweaks to WebGL+CSS

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56847 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Pavel Feldman  <pfeldman@chromium.org>

        Reviewed by Timothy Hatcher.

        Web Inspector: creating new style bugfixing.

        https://bugs.webkit.org/show_bug.cgi?id=36884

        - InspectorDOMAgent should clear internal state upon reset (we are re-using
          single dom agent instance throughout entire inspector controller lifetime)
        - Brought back blank style 'refresh' processing logic - remove it by mistake earlier
        - Blocked couple of click handlers so that double-click to edit was not
          expanding / collapsing the pane
        - There is no need to reach out for matched rules upon 'refresh' update -
          getting computed style is sufficient.

        * css/CSSStyleSelector.cpp:
        (WebCore::CSSStyleSelector::matchRulesForList):
        * inspector/InspectorDOMAgent.cpp:
        (WebCore::InspectorDOMAgent::discardBindings):
        * inspector/front-end/StylesSidebarPane.js:
        (WebInspector.StylesSidebarPane.prototype.update.getStylesCallback):
        (WebInspector.StylesSidebarPane.prototype.update.getComputedStyleCallback):
        (WebInspector.StylesSidebarPane.prototype.update):
        (WebInspector.StylesSidebarPane.prototype._refreshUpdate):
        (WebInspector.StylesSidebarPane.prototype._rebuildUpdate):
        (WebInspector.StylesSidebarPane.prototype._refreshStyleRules):
        (WebInspector.StylesSidebarPane.prototype.addBlankSection):
        (WebInspector.StylePropertiesSection.prototype._dblclickEmptySpace):
        (WebInspector.StylePropertiesSection.prototype._clickSelector):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56846 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2010-03-31  Laszlo Gombos  <laszlo.1.gombos@nokia.com>

        Unreviewed, fix after r56842.

        Add UNUSED_PARAM a to silence warning.

        * jit/JITStubs.cpp:
        (JSC::DEFINE_STUB_FUNCTION):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@56845 268f45cc-cd09-0410-ab3c-d52691b4dbfc