WebKit-https.git
9 years ago2010-08-18 Andreas Kling <andreas.kling@nokia.com>
andreas.kling@nokia.com [Thu, 19 Aug 2010 05:56:30 +0000 (05:56 +0000)]
2010-08-18  Andreas Kling  <andreas.kling@nokia.com>

        Reviewed by Geoffrey Garen.

        REGRESSION(r58469): Math.pow() always returns double-backed JSValue which is extremely slow as array subscript
        https://bugs.webkit.org/show_bug.cgi?id=43742

        Add codegen for pow() to return Int32 values when possible.

        * jit/ThunkGenerators.cpp:
        (JSC::powThunkGenerator):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65653 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agofast/text/hyphenate-character.html is failing on the Leopard Commit Bot
mitz@apple.com [Thu, 19 Aug 2010 05:41:31 +0000 (05:41 +0000)]
fast/text/hyphenate-character.html is failing on the Leopard Commit Bot
https://bugs.webkit.org/show_bug.cgi?id=41093
-and some of-
fast/text/hyphen* tests have different results on Leopard
https://bugs.webkit.org/show_bug.cgi?id=43296

Reviewed by Mark Rowe.

* fast/text/hyphenate-character.html: Specify the hyphenation locale, which removes the
dependency on user preferences.
* fast/text/hyphens.html: Ditto.
* platform/mac-leopard/Skipped: Don’t skip the above tests.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65652 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Alexey Marinichev <amarinichev@chromium.org>
commit-queue@webkit.org [Thu, 19 Aug 2010 05:38:12 +0000 (05:38 +0000)]
2010-08-18  Alexey Marinichev  <amarinichev@chromium.org>

        Reviewed by Darin Fisher.

        [chromium] scrolling issues when accelerated compositor is enabled
        https://bugs.webkit.org/show_bug.cgi?id=43992

        Corrected scroll position not being updated, and an off-by-half error.
        Int cast and floorf are taken out because they produce incorrect
        transformation matrix.  The information lost there is not immediately
        obvious because of rounding that happens when texture is mapped with
        GL_NEAREST filtering mode.  These errors accumulate, and after some
        scrolling it becomes more apparent.

        To test the former, follow instructions in the bug.  To test the
        latter, change GL_NEAREST to GL_LINEAR in LayerRendererChromium.cpp.
        Scrolling should work properly for all window sizes, without blurring
        images and shifting them to the left.

        * platform/graphics/chromium/LayerRendererChromium.cpp:
        (WebCore::LayerRendererChromium::drawLayers):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65651 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Simon Hausmann <simon.hausmann@nokia.com>
commit-queue@webkit.org [Thu, 19 Aug 2010 05:20:22 +0000 (05:20 +0000)]
2010-08-18  Simon Hausmann  <simon.hausmann@nokia.com>

        Reviewed by Ariya Hidayat.

        [Qt] Implement GraphicsContext::clipOut more efficiently
        https://bugs.webkit.org/show_bug.cgi?id=43416

        The current implementation of clipOut uses QPainter::clipRegion().boundingRect(),
        which is a very slow function because it converts the entire clip region - which
        may potentially contain a complex path - into a set of QRegion rectangles, just
        to throw away all that information and use the bounding rect.

        QTBUG-12618 implements a faster function in QPainter: QPainter::clipBoundingRect().

        * platform/graphics/qt/GraphicsContextQt.cpp:
        (WebCore::GraphicsContext::clipOut): Use QPainter::clipBoundingRect() with Qt >= 4.8.
        (WebCore::GraphicsContext::clipOutEllipseInRect): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65650 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Martin Robinson <mrobinson@igalia.com>
mrobinson@webkit.org [Thu, 19 Aug 2010 04:35:14 +0000 (04:35 +0000)]
2010-08-18  Martin Robinson  <mrobinson@igalia.com>

        Reviewed by Gustavo Noronha Silva.

        [GTK] Since the GSEAL changes moz_gtk_scrollbar_button_paint fiddles the widget allocation but doesn't preserve it
        https://bugs.webkit.org/show_bug.cgi?id=44211

        No new tests as we do not currently use the scrollbar button painting
        code. Tests for this issue will be enabled when we have pixel tests
        for WebCore-drawn interior frame scrollbars.

        * platform/gtk/GtkVersioning.h: Add backward-compatible functions for getting
        and setting the widget allocation.
        * platform/gtk/gtk2drawing.c:
        (moz_gtk_scrollbar_button_paint): Use new backward-compatible functions for getting
        and setting the widget allocation.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65649 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 MORITA Hajime <morrita@google.com>
morrita@google.com [Thu, 19 Aug 2010 02:14:38 +0000 (02:14 +0000)]
2010-08-17  MORITA Hajime  <morrita@google.com>

        Reviewed by Tony Chang.

        [Chromium] editing/pasteboard/drop-text-events.html failed.
        https://bugs.webkit.org/show_bug.cgi?id=44051

        * editing/pasteboard/drop-text-events-expected.txt:
        * editing/pasteboard/script-tests/drop-text-events.js: Supressed side effect during a drag.
        * platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65648 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Ryosuke Niwa <rniwa@webkit.org>
rniwa@webkit.org [Thu, 19 Aug 2010 01:04:27 +0000 (01:04 +0000)]
2010-08-18  Ryosuke Niwa  <rniwa@webkit.org>

        Reviewed by Darin Adler.

        Group functions in markup.cpp into MarkupAccumulatorWrapper
        https://bugs.webkit.org/show_bug.cgi?id=43936

        Renamed MarkupAccumulatorWrapper to MarkupAccumulator and moved the MarkupAccumulatorWrapper
        to the top of markup.cpp. Made various functions private member functions of MarkupAccumulator.

        No new tests are added since this is a cleanup.

        * editing/markup.cpp:
        (WebCore::MarkupAccumulator::):
        (WebCore::MarkupAccumulator::MarkupAccumulator):
        (WebCore::MarkupAccumulator::appendString):
        (WebCore::MarkupAccumulator::appendStartTag):
        (WebCore::MarkupAccumulator::appendEndTag):
        (WebCore::MarkupAccumulator::wrapWithNode):
        (WebCore::MarkupAccumulator::wrapWithStyleNode):
        (WebCore::MarkupAccumulator::takeResults):
        (WebCore::MarkupAccumulator::appendAttributeValue):
        (WebCore::MarkupAccumulator::escapeContentText):
        (WebCore::MarkupAccumulator::appendQuotedURLAttributeValue):
        (WebCore::MarkupAccumulator::stringValueForRange):
        (WebCore::MarkupAccumulator::ucharRange):
        (WebCore::MarkupAccumulator::appendUCharRange):
        (WebCore::MarkupAccumulator::renderedText):
        (WebCore::MarkupAccumulator::shouldAddNamespaceElement):
        (WebCore::MarkupAccumulator::shouldAddNamespaceAttribute):
        (WebCore::MarkupAccumulator::appendNamespace):
        (WebCore::MarkupAccumulator::appendDocumentType):
        (WebCore::MarkupAccumulator::removeExteriorStyles):
        (WebCore::MarkupAccumulator::appendStartMarkup):
        (WebCore::MarkupAccumulator::shouldSelfClose):
        (WebCore::MarkupAccumulator::appendEndMarkup):
        (WebCore::serializeNodes):
        (WebCore::createMarkup):
        (WebCore::serializeNodesWithNamespaces):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65647 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoWebCore: Fix for https://bugs.webkit.org/show_bug.cgi?id=44207 Empty mfrac
bdakin@apple.com [Thu, 19 Aug 2010 00:26:20 +0000 (00:26 +0000)]
WebCore: Fix for https://bugs.webkit.org/show_bug.cgi?id=44207 Empty mfrac
and empty munderover cause crashes
       -and corresponding-
<rdar://problem/8325160>

Reviewed by Sam Weinig.

This is just a  matter of adding null checks in the right places.
* mathml/RenderMathMLFraction.cpp:
(WebCore::RenderMathMLFraction::baselinePosition):
* mathml/RenderMathMLUnderOver.cpp:
(WebCore::RenderMathMLUnderOver::baselinePosition):

LayoutTests: Tests for https://bugs.webkit.org/show_bug.cgi?id=44207 Empty mfrac
and empty munderover cause crashes
       -and corresponding-
<rdar://problem/8325160>
* mathml/EmptyMFracCrash-expected.txt: Added.
* mathml/EmptyMFracCrash.xhtml: Added.
* mathml/EmptyMunderOverCrash-expected.txt: Added.
* mathml/EmptyMunderOverCrash.xhtml: Added.

Reviewed by Sam Weinig.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65646 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoMove FileStream to platform.
jianli@chromium.org [Wed, 18 Aug 2010 23:52:03 +0000 (23:52 +0000)]
Move FileStream to platform.
https://bugs.webkit.org/show_bug.cgi?id=44213

Reviewed by David Levin.

In addition to moving files over, FileStream has been changed to remove
ExceptionCode dependency. Also update FileStreamProxy and FileRead to
account for this change.

* Android.mk:
* GNUmakefile.am:
* WebCore.gypi:
* WebCore.pro:
* WebCore.vcproj/WebCore.vcproj:
* WebCore.xcodeproj/project.pbxproj:
* html/FileReader.cpp:
(WebCore::FileReader::didOpen):
* html/FileReader.h:
* html/FileStream.cpp: Removed.
* html/FileStream.h: Removed.
* html/FileStreamClient.h:
(WebCore::FileStreamClient::didTruncate):
(WebCore::FileStreamClient::didOpen):
* html/FileStreamProxy.cpp:
(WebCore::didOpen):
(WebCore::FileStreamProxy::openForReadOnFileThread):
(WebCore::FileStreamProxy::openForWriteOnFileThread):
(WebCore::FileStreamProxy::write):
(WebCore::FileStreamProxy::writeOnFileThread):
(WebCore::didTruncate):
(WebCore::FileStreamProxy::truncateOnFileThread):
* html/FileStreamProxy.h:
* platform/FileStream.cpp: Renamed from WebCore/FileStream.cpp and updated.
* platform/FileStream.h: Renamed from WebCore/FileStream.h and updated.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65645 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoAdd ability to set custom userAgent for WebKit2
weinig@apple.com [Wed, 18 Aug 2010 23:46:21 +0000 (23:46 +0000)]
Add ability to set custom userAgent for WebKit2
https://bugs.webkit.org/show_bug.cgi?id=44215

Reviewed by Jon Honeycutt.

* Shared/CoreIPCSupport/WebPageMessageKinds.h:
(WebPageMessage::):
* UIProcess/API/C/WKPage.cpp:
(WKPageSetCustomUserAgent):
* UIProcess/API/C/WKPage.h:
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::setCustomUserAgent):
* UIProcess/WebPageProxy.h:
* WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:
(WebKit::WebFrameLoaderClient::userAgent):
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::setCustomUserAgent):
(WebKit::WebPage::userAgent):
(WebKit::WebPage::didReceiveMessage):
* WebProcess/WebPage/WebPage.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65644 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Lucas De Marchi <lucas.demarchi@profusion.mobi>
commit-queue@webkit.org [Wed, 18 Aug 2010 23:46:02 +0000 (23:46 +0000)]
2010-08-18  Lucas De Marchi  <lucas.demarchi@profusion.mobi>

        Reviewed by Antonio Gomes.

        [EFL] Bump library dependencies
        https://bugs.webkit.org/show_bug.cgi?id=44182

        Prepare for EFL release. All the EFL dependencies were bumped to
        1.0.0. Eina and Ecore_X needed a bump because of API has changed and
        the others were changed just to follow those two and because this is
        a stable release.

        * cmake/FindEFL.cmake: Bump dependencies.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65643 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoFix layout test crashes in chromium.
jianli@chromium.org [Wed, 18 Aug 2010 23:33:34 +0000 (23:33 +0000)]
Fix layout test crashes in chromium.

* src/BlobRegistryProxy.cpp:
(WebCore::BlobRegistryProxy::registerBlobURL):
(WebCore::BlobRegistryProxy::unregisterBlobURL):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65642 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoFix chromium build break.
jianli@chromium.org [Wed, 18 Aug 2010 23:26:57 +0000 (23:26 +0000)]
Fix chromium build break.

* src/WebBlobStorageData.cpp:
(WebKit::WebBlobStorageData::itemAt):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65641 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Fady Samuel <fsamuel@chromium.org>
commit-queue@webkit.org [Wed, 18 Aug 2010 23:23:43 +0000 (23:23 +0000)]
2010-08-18  Fady Samuel  <fsamuel@chromium.org>

        Reviewed by Dimitri Glazkov.

        Updated Test Expectations for Table Hit Testing Layout Tests to Produce Results from Canaries
        https://bugs.webkit.org/show_bug.cgi?id=44177

        * platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65640 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoMake it so the WebKit2 test slave runs the WebKit2 tests
mrowe@apple.com [Wed, 18 Aug 2010 23:10:51 +0000 (23:10 +0000)]
Make it so the WebKit2 test slave runs the WebKit2 tests

Reviewed by Sam Weinig.

* BuildSlaveSupport/build.webkit.org-config/master.cfg:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65638 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoWebCore: Chromium side implementation of blob data and blob registry.
jianli@chromium.org [Wed, 18 Aug 2010 22:48:25 +0000 (22:48 +0000)]
WebCore: Chromium side implementation of blob data and blob registry.
https://bugs.webkit.org/show_bug.cgi?id=43871

Reviewed by Darin Fisher.

Wrap !PLATFORM(CHROMIUM) around blobRegistry() so that chromium uses the
implementation in WebKit Chromium.

* platform/network/BlobRegistryImpl.cpp:

WebKit/chromium: [chromium] Chromium side implementation of blob data and blob registry.
https://bugs.webkit.org/show_bug.cgi?id=43871

Reviewed by Darin Fisher.

* WebKit.gyp:
* public/WebBlobData.h: Added.
* public/WebBlobRegistry.h: Added.
* public/WebBlobStorageData.h: Added.
* public/WebKitClient.h:
(WebKit::WebKitClient::blobRegistry):
* src/BlobRegistryProxy.cpp: Added.
* src/BlobRegistryProxy.h: Added.
* src/WebBlobData.cpp: Added.
* src/WebBlobRegistryImpl.cpp: Added.
* src/WebBlobRegistryImpl.h: Added.
* src/WebBlobStorageData.cpp: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65637 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoPropagate the m_downloadFilePath correctly.
dumi@chromium.org [Wed, 18 Aug 2010 22:46:38 +0000 (22:46 +0000)]
Propagate the m_downloadFilePath correctly.
https://bugs.webkit.org/show_bug.cgi?id=44198

Reviewed by Darin Fisher.

* src/WebURLRequest.cpp:
(WebKit::WebURLRequestPrivateImpl::WebURLRequestPrivateImpl):
* src/WebURLResponse.cpp:
(WebKit::WebURLResponsePrivateImpl::WebURLResponsePrivateImpl):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65636 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Martin Robinson <mrobinson@igalia.com>
mrobinson@webkit.org [Wed, 18 Aug 2010 22:45:19 +0000 (22:45 +0000)]
2010-08-18  Martin Robinson  <mrobinson@igalia.com>

        Reviewed by Gustavo Noronha Silva.

        [GTK] Bots are showing lots of GTK_IS_CONTAINER critical warnings
        https://bugs.webkit.org/show_bug.cgi?id=40990

        Instead of relying on the ScrollView's adjustment members to determine if a
        Scrollbar should be native or just a shell of the parent's scrollbar (main frame
        scrollbar), just check if this ScrollView has a parent. This will ensure the
        correct behavior when main frame scrollbar's are created after the ScrollView's
        containing adjustments go away.

        Lack of warnings during tests are the test for this fix.

        * platform/gtk/ScrollViewGtk.cpp:
        (WebCore::ScrollView::createScrollbar): Decide what type of scrollbar to make based
        on the result of the parent() method.
        * platform/gtk/ScrollbarGtk.cpp:
        (ScrollbarGtk::ScrollbarGtk): Allow for an m_adjustment which is null.
        (ScrollbarGtk::attachAdjustment): Ditto.
        (ScrollbarGtk::updateThumbPosition): Ditto.
        (ScrollbarGtk::updateThumbProportion): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65635 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Victor Wang <victorw@chromium.org>
victorw@chromium.org [Wed, 18 Aug 2010 22:34:31 +0000 (22:34 +0000)]
2010-08-18  Victor Wang  <victorw@chromium.org>

        Unreviewed. Fix chromium multi dll build.

        * public/WebIDBKeyPath.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65634 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Jay Civelli <jcivelli@chromium.org>
jcivelli@chromium.org [Wed, 18 Aug 2010 22:05:52 +0000 (22:05 +0000)]
2010-08-18  Jay Civelli  <jcivelli@chromium.org>

        Reviewed by Darin Fisher.

        Renaming the WebEvent to WebDOMEvent and adding a new event class
        for mouse events so that the event listener API can be used to listen
        to them.  The renaming is needed as there already is a WebMouseEvent class.
        https://bugs.webkit.org/show_bug.cgi?id=43453

        * WebKit.gyp:
        * public/WebDOMEvent.h: Added.
        * public/WebDOMEventListener.h: Added.
        * public/WebDOMMouseEvent.h: Added.
        * public/WebDOMMutationEvent.h: Added.
        * public/WebNode.h:
        * src/EventListenerWrapper.cpp:
        * src/EventListenerWrapper.h:
        * src/WebDOMEvent.cpp: Added.
        * src/WebDOMEventListener.cpp: Added.
        * src/WebDOMEventListenerPrivate.cpp: Added.
        * src/WebDOMEventListenerPrivate.h: Added.
        * src/WebDOMMouseEvent.cpp: Added.
        * src/WebDOMMutationEvent.cpp: Added.
        * src/WebNode.cpp:
        (WebKit::WebNode::isElementNode):
        (WebKit::WebNode::addDOMEventListener):
        (WebKit::WebNode::removeDOMEventListener):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65633 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Adam Barth <abarth@webkit.org>
abarth@webkit.org [Wed, 18 Aug 2010 21:45:53 +0000 (21:45 +0000)]
2010-08-18  Adam Barth  <abarth@webkit.org>

        Reviewed by David Levin.

        NOT_REACHED is reachable in SVGLength
        https://bugs.webkit.org/show_bug.cgi?id=44150

        The author of this code was confused.  This code is reachable.  We just
        haven't implemented it yet.

        * svg/SVGLength.cpp:
        (WebCore::SVGLength::setValue):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65632 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Balazs Kelemen <kb@inf.u-szeged.hu>
commit-queue@webkit.org [Wed, 18 Aug 2010 21:20:36 +0000 (21:20 +0000)]
2010-08-18  Balazs Kelemen  <kb@inf.u-szeged.hu>

        Reviewed by Antonio Gomes.

        [Qt] Use LAZY_NATIVE_CURSOR
        https://bugs.webkit.org/show_bug.cgi?id=44062

        No functional change so new tests.

        Change Cursor behaviour to match the LAZY_NATIVE_CURSOR policy.

        * platform/Cursor.h: Typedef PlatformCursor to be a QCursor* to be able create it dynamically.
        (WebCore::Cursor::Cursor): Remove the ifdef for Qt.
        * platform/qt/CursorQt.cpp: Remove the Cursors class since we have the static cursor instances
        for the common cursor types in Cursor.cpp. Move the logic that maps the cursor types to
        QCursor instances into ensurePlatformCursor.
        (WebCore::Cursor::Cursor):
        (WebCore::Cursor::~Cursor):
        (WebCore::Cursor::operator=):
        (WebCore::createCustomCursor):
        (WebCore::Cursor::ensurePlatformCursor):
        * platform/qt/WidgetQt.cpp:
        (WebCore::Widget::setCursor): Adjust to the PlatformCursor change.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65631 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoMake run-webkit-tests --webkit-test-runner build DumpRenderTree
weinig@apple.com [Wed, 18 Aug 2010 21:20:16 +0000 (21:20 +0000)]
Make run-webkit-tests --webkit-test-runner build DumpRenderTree
in addition to WebKitTestRunner to get the DumpRenderTreeSupport
module and the TestNetscapePlugin.

Reviewed by Mark Rowe.

* Scripts/old-run-webkit-tests:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65630 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Laszlo Gombos <laszlo.1.gombos@nokia.com>
laszlo.1.gombos@nokia.com [Wed, 18 Aug 2010 21:11:27 +0000 (21:11 +0000)]
2010-08-18  Laszlo Gombos  <laszlo.1.gombos@nokia.com>

        Unreviewed, trivial Symbian build fix.

        [Qt] Remove unnecessary QtGui include from NotificationPresenterClientQt.cpp

        * WebCoreSupport/NotificationPresenterClientQt.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65629 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Ojan Vafai <ojan@chromium.org>
ojan@chromium.org [Wed, 18 Aug 2010 20:51:44 +0000 (20:51 +0000)]
2010-08-18  Ojan Vafai  <ojan@chromium.org>

        Reviewed by Tony Chang.

        skip the new perf tests on non-chromium to gather data on which ones are flaky
        https://bugs.webkit.org/show_bug.cgi?id=44200

        * platform/chromium/test_expectations.txt:
        * platform/gtk/Skipped:
        * platform/mac/Skipped:
        * platform/qt/Skipped:
        * platform/win/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65628 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoFix my badly-merged ChangeLog entry.
senorblanco@chromium.org [Wed, 18 Aug 2010 20:43:36 +0000 (20:43 +0000)]
Fix my badly-merged ChangeLog entry.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65627 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Andreas Kling <andreas.kling@nokia.com>
andreas.kling@nokia.com [Wed, 18 Aug 2010 20:05:10 +0000 (20:05 +0000)]
2010-08-18  Andreas Kling  <andreas.kling@nokia.com>

        Reviewed by Darin Adler.

        CSS: Make rgb() and rgba() fast paths case-insensitive
        https://bugs.webkit.org/show_bug.cgi?id=44107

        Also inlined the string comparisons against "rgb(" and "rgba("
        which is faster and avoids creating a temporary String object.

        * css/CSSParser.cpp:
        (WebCore::mightBeRGBA):
        (WebCore::mightBeRGB):
        (WebCore::CSSParser::parseColor):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65626 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Sheriff Bot <webkit.review.bot@gmail.com>
commit-queue@webkit.org [Wed, 18 Aug 2010 19:22:51 +0000 (19:22 +0000)]
2010-08-18  Sheriff Bot  <webkit.review.bot@gmail.com>

        Unreviewed, rolling out r65619.
        http://trac.webkit.org/changeset/65619
        https://bugs.webkit.org/show_bug.cgi?id=44194

        Broke mac bots (Requested by jamesr on #webkit).

        * platform/chromium-linux/tables/hittesting/filltable-emptycells-expected.txt: Renamed from LayoutTests/tables/hittesting/filltable-emptycells-expected.txt.
        * platform/chromium-linux/tables/hittesting/filltable-levels-expected.txt: Renamed from LayoutTests/tables/hittesting/filltable-levels-expected.txt.
        * platform/chromium-linux/tables/hittesting/filltable-outline-expected.txt: Renamed from LayoutTests/tables/hittesting/filltable-outline-expected.txt.
        * platform/chromium-linux/tables/hittesting/filltable-rtl-expected.txt: Renamed from LayoutTests/tables/hittesting/filltable-rtl-expected.txt.
        * platform/chromium-linux/tables/hittesting/filltable-stress-expected.txt: Renamed from LayoutTests/tables/hittesting/filltable-stress-expected.txt.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65625 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoChange BlobStorageData to reuse BlobData.
jianli@chromium.org [Wed, 18 Aug 2010 19:03:45 +0000 (19:03 +0000)]
Change BlobStorageData to reuse BlobData.
https://bugs.webkit.org/show_bug.cgi?id=44188

Reviewed by Darin Fisher.

Also addressed Darin's feedbacks for bug 44116.

* html/ThreadableBlobRegistry.cpp:
(WebCore::registerBlobURLTask):
(WebCore::registerBlobURLFromTask):
(WebCore::unregisterBlobURLTask):
* platform/network/BlobData.cpp:
(WebCore::BlobData::appendData):
* platform/network/BlobData.h:
(WebCore::BlobDataItem::BlobDataItem):
* platform/network/BlobRegistry.h:
* platform/network/BlobRegistryImpl.cpp:
(WebCore::blobRegistry):
(WebCore::BlobRegistryImpl::appendStorageItems):
(WebCore::BlobRegistryImpl::registerBlobURL):
* platform/network/BlobRegistryImpl.h:
* platform/network/BlobStorageData.h:
(WebCore::BlobStorageData::create):
(WebCore::BlobStorageData::contentType):
(WebCore::BlobStorageData::contentDisposition):
(WebCore::BlobStorageData::items):
(WebCore::BlobStorageData::BlobStorageData):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65622 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agohttps://bugs.webkit.org/show_bug.cgi?id=44193
hyatt@apple.com [Wed, 18 Aug 2010 18:45:24 +0000 (18:45 +0000)]
https://bugs.webkit.org/show_bug.cgi?id=44193

Reviewed by Martin Robinson.

Fix a typo in my previous canvas checkin.  m_in->resultImage() was originally m_in2->resultImage(),
and I accidentally changed it when swapping the order of the arguments.

* platform/graphics/filters/FEComposite.cpp:
(WebCore::FEComposite::apply):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65621 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoMake WKBundleNodeHandleRef usable from WKRetainPtr.
weinig@apple.com [Wed, 18 Aug 2010 18:40:52 +0000 (18:40 +0000)]
Make WKBundleNodeHandleRef usable from WKRetainPtr.

Rubber-stamped by John Sullivan.

* WebKit2.xcodeproj/project.pbxproj: Make WKBundleNodeHandle.h public.
* WebProcess/InjectedBundle/API/c/WKBundleNodeHandle.h: Add WKRetain/WKRelease overloads.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65620 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 James Robinson <jamesr@chromium.org>
jamesr@google.com [Wed, 18 Aug 2010 18:21:47 +0000 (18:21 +0000)]
2010-08-18  James Robinson  <jamesr@chromium.org>

        Unreviewed.  Move some test expectations that are not platform specific to the right place.

        * platform/chromium-linux/tables/hittesting/filltable-emptycells-expected.txt: Removed.
        * platform/chromium-linux/tables/hittesting/filltable-levels-expected.txt: Removed.
        * platform/chromium-linux/tables/hittesting/filltable-outline-expected.txt: Removed.
        * platform/chromium-linux/tables/hittesting/filltable-rtl-expected.txt: Removed.
        * platform/chromium-linux/tables/hittesting/filltable-stress-expected.txt: Removed.
        * tables/hittesting/filltable-emptycells-expected.txt: Copied from LayoutTests/platform/chromium-linux/tables/hittesting/filltable-emptycells-expected.txt.
        * tables/hittesting/filltable-levels-expected.txt: Copied from LayoutTests/platform/chromium-linux/tables/hittesting/filltable-levels-expected.txt.
        * tables/hittesting/filltable-outline-expected.txt: Copied from LayoutTests/platform/chromium-linux/tables/hittesting/filltable-outline-expected.txt.
        * tables/hittesting/filltable-rtl-expected.txt: Copied from LayoutTests/platform/chromium-linux/tables/hittesting/filltable-rtl-expected.txt.
        * tables/hittesting/filltable-stress-expected.txt: Copied from LayoutTests/platform/chromium-linux/tables/hittesting/filltable-stress-expected.txt.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65619 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Jessie Berlin <jberlin@apple.com>
jberlin@webkit.org [Wed, 18 Aug 2010 18:16:57 +0000 (18:16 +0000)]
2010-08-18  Jessie Berlin  <jberlin@apple.com>

        Reviewed by Adam Roben.

        Bug 44180 - WebView::paint fails to paint a child WebView of a Layered Window.
        https://bugs.webkit.org/show_bug.cgi?id=44180

        Decide to end painting if the m_backingStoreBitmap is null after the call to
        ensureBackingStore() instead of when the rcPaint rect filled by BeginPaint is empty.
        The rcPaint rect filled by BeginPaint is always empty for a child WebView of a Layered
        Window, even if GetUpdateRect and GetUpdateRgn report a non-empty region that needs
        painting.

        * WebView.cpp:
        (WebView::paint):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65618 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Stephen White <senorblanco@chromium.org>
senorblanco@chromium.org [Wed, 18 Aug 2010 18:11:17 +0000 (18:11 +0000)]
2010-08-17  Stephen White  <senorblanco@chromium.org>

        Reviewed by Darin Fisher.

        [CHROMIUM] Eliminate a memcpy() from the canvas.drawImage(canvas, ...) path.
        https://bugs.webkit.org/show_bug.cgi?id=44115

        This is the Skia followup to Dave Hyatt's patch in r65449.
        Covered by canvas layout tests.

        * platform/graphics/skia/BitmapImageSingleFrameSkia.h:
        Add a "copyPixels" argument to the create() static to indicate if
        the caller wants the pixels to be copied or ref'ed.
        * platform/graphics/skia/ImageBufferSkia.cpp:
        (WebCore::ImageBuffer::drawsUsingCopy):
        We don't use the always-copy path anymore, so return false.
        (WebCore::ImageBuffer::copyImage):
        Force a copy here.
        (WebCore::ImageBuffer::draw):
        (WebCore::ImageBuffer::drawPattern):
        Create a temporary BitampImageSingleFrameSkia wrapped around the
        canvas's bitmap, and draw with it immediately.  If drawing to our own
        canvas, copy the pixels, otherwise just shallow copy.
        * platform/graphics/skia/ImageSkia.cpp:
        (WebCore::BitmapImageSingleFrameSkia::BitmapImageSingleFrameSkia):
        Make the constructor take an SkBitmap to be wrapped.
        (WebCore::BitmapImageSingleFrameSkia::create):
        Implement the "copyPixels" flag:  if true, do a deep copy, otherwise
        do a shallow copy.
        * platform/graphics/skia/NativeImageSkia.cpp:
        (WebCore::NativeImageSkia::NativeImageSkia):
        * platform/graphics/skia/NativeImageSkia.h:
        Implement a new constructor that shallow-copies the given SkBitmap.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65617 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago[Qt] Request for permission before starting Geolocation service
ossy@webkit.org [Wed, 18 Aug 2010 17:51:24 +0000 (17:51 +0000)]
[Qt] Request for permission before starting Geolocation service
https://bugs.webkit.org/show_bug.cgi?id=42027

Patch by Mahesh Kulkarni <mahesh.kulkarni@nokia.com> on 2010-08-18
Reviewed by Steve Block.

Handles starting location acquisition when request is granted for ports using
"!CLIENT_BASED_GEOLOCATION and PREEMPT_GEOLOCATION_PERMISSION" policy
Only Qt port as of today uses the above combination. Enable
PREEMPT_GEOLOCATION_PERMISSION for qt port.

Below tests cover these changes on platform using PREEMPT_GEOLOCATION_PERMISSION
- fast/dom/Geolocation/delayed-permission-allowed.html
- fast/dom/Geolocation/delayed-permission-denied.html
- fast/dom/Geolocation/delayed-permission-allowed-for-multiple-requests.html
- fast/dom/Geolocation/delayed-permission-denied-for-multiple-requests.html

* WebCore.pro:
* page/Geolocation.cpp:
(WebCore::Geolocation::handlePendingPermissionNotifiers):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65616 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Gustavo Noronha Silva <gustavo.noronha@collabora.co.uk>
kov@webkit.org [Wed, 18 Aug 2010 17:47:30 +0000 (17:47 +0000)]
2010-08-18  Gustavo Noronha Silva  <gustavo.noronha@collabora.co.uk>

        Reviewed by Martin Robinson.

        [GTK] WebBackForwardList's new_with_web_view should not be public
        https://bugs.webkit.org/show_bug.cgi?id=43802

        Deprecate the constructor. The introspection annotation has been
        fixed in a previous commit.

        * webkit/webkitwebbackforwardlist.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65615 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-13 Ojan Vafai <ojan@chromium.org>
ojan@chromium.org [Wed, 18 Aug 2010 17:34:51 +0000 (17:34 +0000)]
2010-08-13  Ojan Vafai  <ojan@chromium.org>

        Reviewed by Darin Adler.

        add ability to test order-of-magnitude in layout tests
        https://bugs.webkit.org/show_bug.cgi?id=43997

        Uses the R^2 value to determine likelihood the performance
        is linear and then logs the x and y axis appropriately to determine
        log n and n^2 cases. These are all estimates though, so we try
        up to three times if we don't get the correct expected result.

        In practice, with the tests below, we almost always get the correct
        result, but sometimes the O(1) tests get registered as O(n) or indeterminate.

        * perf/array-binary-search-expected.txt: Added.
        * perf/array-binary-search.html: Added.
        * perf/array-nested-loop-expected.txt: Added.
        * perf/array-nested-loop.html: Added.
        * perf/array-push-pop-expected.txt: Added.
        * perf/array-push-pop.html: Added.
        * perf/array-reverse-expected.txt: Added.
        * perf/array-reverse.html: Added.
        * perf/object-keys-expected.txt: Added.
        * perf/object-keys.html: Added.
        * perf/set-attribute-expected.txt: Added.
        * perf/set-attribute.html: Added.
        * perf/typing-at-end-of-line-expected.txt: Added.
        * perf/typing-at-end-of-line.html: Added.
        * resources/magnitude-perf.js: Added.
        (Magnitude.description):
        (Magnitude._log):
        (Magnitude._debug):
        (Magnitude.run):
        (Magnitude._run):
        (Magnitude._rSquared):
        (Magnitude._logIterationInfo):
        (Magnitude._bigOGuess):
        (Magnitude._runIteration):
        (Magnitude):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65614 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Andreas Kling <andreas.kling@nokia.com>
andreas.kling@nokia.com [Wed, 18 Aug 2010 17:28:11 +0000 (17:28 +0000)]
2010-08-18  Andreas Kling  <andreas.kling@nokia.com>

        Reviewed by Darin Adler.

        CanvasStyle method naming cleanup
        https://bugs.webkit.org/show_bug.cgi?id=44184

        Renamed the various create() overloads to reflect what they're
        constructing the CanvasStyle from.

        The unused CanvasStyle::create(grayLevel) and accompanying constructor
        were removed.

        Also renamed the isEquivalentColor overloads to isEquivalent{RGBA,CMYKA}

        * bindings/js/JSCanvasRenderingContext2DCustom.cpp:
        (WebCore::toHTMLCanvasStyle):
        * bindings/v8/custom/V8CanvasRenderingContext2DCustom.cpp:
        (WebCore::toCanvasStyle):
        * html/canvas/CanvasRenderingContext2D.cpp:
        (WebCore::CanvasRenderingContext2D::State::State):
        (WebCore::CanvasRenderingContext2D::setStrokeColor):
        (WebCore::CanvasRenderingContext2D::setFillColor):
        * html/canvas/CanvasStyle.cpp:
        (WebCore::CanvasStyle::createFromString):
        (WebCore::CanvasStyle::createFromStringWithOverrideAlpha):
        (WebCore::CanvasStyle::createFromGradient):
        (WebCore::CanvasStyle::createFromPattern):
        (WebCore::CanvasStyle::isEquivalentRGBA):
        (WebCore::CanvasStyle::isEquivalentCMYKA):
        * html/canvas/CanvasStyle.h:
        (WebCore::CanvasStyle::createFromRGBA):
        (WebCore::CanvasStyle::createFromGrayLevelWithAlpha):
        (WebCore::CanvasStyle::createFromRGBAChannels):
        (WebCore::CanvasStyle::createFromCMYKAChannels):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65613 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago[Qt] Implement Maemo5 local rendering NPAPI extension. See
girish@forwardbias.in [Wed, 18 Aug 2010 16:52:32 +0000 (16:52 +0000)]
[Qt] Implement Maemo5 local rendering NPAPI extension. See
https://wiki.mozilla.org/Plugins:NokiaMaemoImageSurface for details.

With the local rendering extension, Flash will paint into a 16-bit surface.
For wmode=transparent, Flash expects the surface to contain the contents
beneath it. As it is tricky to implement the content propagation across all
graphics systems, transparent Flash is not supported. We just fill the surface
with white and wmode=transparent behaves the same as wmode=opaque with a white
background.

Reviewed by Kenneth Rohde Christiansen.

https://bugs.webkit.org/show_bug.cgi?id=44043

* WebCore.pro:
* plugins/PluginView.cpp:
(WebCore::PluginView::setValue):
(WebCore::PluginView::PluginView):
* plugins/PluginView.h:
* plugins/qt/PluginViewQt.cpp:
(WebCore::PluginView::updatePluginWidget):
(WebCore::PluginView::paintUsingImageSurfaceExtension):
(WebCore::PluginView::paint):
(WebCore::PluginView::platformGetValueStatic):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65612 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Sheriff Bot <webkit.review.bot@gmail.com>
commit-queue@webkit.org [Wed, 18 Aug 2010 16:09:12 +0000 (16:09 +0000)]
2010-08-18  Sheriff Bot  <webkit.review.bot@gmail.com>

        Unreviewed, rolling out r65603.
        http://trac.webkit.org/changeset/65603
        https://bugs.webkit.org/show_bug.cgi?id=44179

        4 tests crash on Qt bot (Requested by Ossy on #webkit).

        * WebCore.pro:
        * page/Geolocation.cpp:
        (WebCore::Geolocation::handlePendingPermissionNotifiers):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65611 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Ryuan Choi <ryuan.choi@samsung.com>
commit-queue@webkit.org [Wed, 18 Aug 2010 16:02:33 +0000 (16:02 +0000)]
2010-08-18  Ryuan Choi  <ryuan.choi@samsung.com>

        Unreviewed.  Attempted build fix for EFL

        [EFL]REGRESSION(65530) need to fix build break
        https://bugs.webkit.org/show_bug.cgi?id=44154

        * CMakeListsEfl.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65610 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Ilya Tikhonovsky <loislo@chromium.org>
loislo@chromium.org [Wed, 18 Aug 2010 15:30:11 +0000 (15:30 +0000)]
2010-08-18  Ilya Tikhonovsky  <loislo@chromium.org>

        Unreviewed qt build fix.

        * WebCore.pri:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65609 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-16 Ilya Tikhonovsky <loislo@chromium.org>
loislo@chromium.org [Wed, 18 Aug 2010 14:36:58 +0000 (14:36 +0000)]
2010-08-16  Ilya Tikhonovsky  <loislo@chromium.org>

        Reviewed by Yury Semikhatsky.

        Web Inspector: replace hand written InspectorBackendStub.js with generated one.
        New command-line flag was added to combine-javascript-resources.
        It specifies the location of generated scripts.
        https://bugs.webkit.org/show_bug.cgi?id=43791

        * GNUmakefile.am:
        * WebCore.gyp/WebCore.gyp:
        * WebCore.gypi:
        * WebCore.pri:
        * WebCore.pro:
        * WebCore.vcproj/WebCore.vcproj:
        * WebCore.xcodeproj/project.pbxproj:
        * combine-javascript-resources:
        * inspector/CodeGeneratorInspector.pm:
        * inspector/front-end/InspectorBackendStub.js: Removed.
        * inspector/front-end/WebKit.qrc:

2010-08-16  Ilya Tikhonovsky  <loislo@chromium.org>

        Reviewed by Yury Semikhatsky.

        Web Inspector: replace hand written InspectorBackendStub.js with generated one.
        https://bugs.webkit.org/show_bug.cgi?id=43791

        * GNUmakefile.am:

2010-08-17  Ilya Tikhonovsky  <loislo@chromium.org>

        Reviewed by Yury Semikhatsky.

        WebInspector: inspector tests for GTK were added to Skip list.
        They should be enabled later.
        Bug for GTK https://bugs.webkit.org/show_bug.cgi?id=43977
        https://bugs.webkit.org/show_bug.cgi?id=43791

        * platform/gtk/Skipped:

2010-08-17  Ilya Tikhonovsky  <loislo@chromium.org>

        Reviewed by Yury Semikhatsky.

        Web Inspector: replace hand written InspectorBackendStub.js with generated one.
        https://bugs.webkit.org/show_bug.cgi?id=43791

        * WebKit.gyp:

2010-08-17  Ilya Tikhonovsky  <loislo@chromium.org>

        Reviewed by Yury Semikhatsky.

        Web Inspector: replace hand written InspectorBackendStub.js with generated one.
        https://bugs.webkit.org/show_bug.cgi?id=43791

        * WebCoreSupport/InspectorClientGtk.cpp:
        (WebKit::InspectorClient::openInspectorFrontend):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65608 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Andrey Kosyakov <caseq@chromium.org>
commit-queue@webkit.org [Wed, 18 Aug 2010 14:18:51 +0000 (14:18 +0000)]
2010-08-18  Andrey Kosyakov  <caseq@chromium.org>

        Reviewed by Pavel Feldman.

        Web Inspector: extension panels do not occupy entire panel client area.
        https://bugs.webkit.org/show_bug.cgi?id=44167

        * inspector/front-end/ExtensionServer.js:
        (WebInspector.ExtensionServer.prototype._onCreatePanel):
        (WebInspector.ExtensionServer.prototype._createClientIframe):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65607 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Pavel Podivilov <podivilov@chromium.org>
yurys@chromium.org [Wed, 18 Aug 2010 13:27:26 +0000 (13:27 +0000)]
2010-08-18  Pavel Podivilov  <podivilov@chromium.org>

        Reviewed by Yury Semikhatsky.

        Web Inspector: inspector tests should be frontend-driven

        * http/tests/inspector/inspector-test2.js: Added.
        (initializeInspectorTest):
        (initializeInspectorTest.):
        ():
        (completeTest):
        (dumpArray):
        (dumpObject):
        (output):
        (didEvaluateForTestInFrontend):
        * inspector/debugger-breakpoints-not-activated-on-reload-expected.txt:
        * inspector/debugger-breakpoints-not-activated-on-reload.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65606 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Dirk Schulze <krit@webkit.org>
krit@webkit.org [Wed, 18 Aug 2010 13:20:51 +0000 (13:20 +0000)]
2010-08-18  Dirk Schulze  <krit@webkit.org>

        Reviewed by Nikolas Zimmermann.

        Use SVGPathParser logic to traverse states of a Path
        https://bugs.webkit.org/show_bug.cgi?id=44009

        Move the getPathSegAtLength logic from SVGPathSegList into a new SVGPathConsumer
        class: SVGPathTraversalStateBuilder.
        This allows us to get a SVGPathSeg at a given length for SVGPathByteStreams and SVGPathSegLists.
        It is also possible to use unaltered SVG path data, instead of just normalized path data.
        This is the final preparation to provide normalized and unaltered SVGPathSegLists
        on SVGPathElement.

        * Android.mk:
        * CMakeLists.txt:
        * GNUmakefile.am:
        * WebCore.gypi:
        * WebCore.pro:
        * WebCore.vcproj/WebCore.vcproj:
        * WebCore.xcodeproj/project.pbxproj:
        * svg/SVGAllInOne.cpp:
        * svg/SVGPathBuilder.h:
        (WebCore::SVGPathBuilder::nextPathSegment):
        (WebCore::SVGPathBuilder::quitEarlier):
        * svg/SVGPathByteStreamBuilder.h:
        (WebCore::SVGPathByteStreamBuilder::nextPathSegment):
        (WebCore::SVGPathByteStreamBuilder::quitEarlier):
        * svg/SVGPathConsumer.h:
        * svg/SVGPathElement.cpp:
        (WebCore::SVGPathElement::getPathSegAtLength):
        * svg/SVGPathElement.h:
        * svg/SVGPathElement.idl:
        * svg/SVGPathParser.cpp:
        (WebCore::SVGPathParser::parsePathDataFromSource):
        * svg/SVGPathParserFactory.cpp:
        (WebCore::globalSVGPathTraversalStateBuilder):
        (WebCore::SVGPathParserFactory::getSVGPathSegAtLengthFromSVGPathSegList):
        * svg/SVGPathParserFactory.h:
        * svg/SVGPathSegList.cpp:
        * svg/SVGPathSegList.h:
        * svg/SVGPathSegListBuilder.h:
        (WebCore::SVGPathSegListBuilder::nextPathSegment):
        (WebCore::SVGPathSegListBuilder::quitEarlier):
        * svg/SVGPathStringBuilder.h:
        (WebCore::SVGPathStringBuilder::nextPathSegment):
        (WebCore::SVGPathStringBuilder::quitEarlier):
        * svg/SVGPathTraversalStateBuilder.cpp: Added. New Consumer for SVGPathParser.
        (WebCore::SVGPathTraversalStateBuilder::SVGPathTraversalStateBuilder):
        (WebCore::SVGPathTraversalStateBuilder::moveTo):
        (WebCore::SVGPathTraversalStateBuilder::lineTo):
        (WebCore::SVGPathTraversalStateBuilder::curveToCubic):
        (WebCore::SVGPathTraversalStateBuilder::closePath):
        (WebCore::SVGPathTraversalStateBuilder::setDesiredLength):
        (WebCore::SVGPathTraversalStateBuilder::quitEarlier):
        (WebCore::SVGPathTraversalStateBuilder::nextPathSegment):
        (WebCore::SVGPathTraversalStateBuilder::pathSegmentIndex):
        * svg/SVGPathTraversalStateBuilder.h: Added.
        (WebCore::SVGPathTraversalStateBuilder::setCurrentTraversalState):
        (WebCore::SVGPathTraversalStateBuilder::cleanup):
        (WebCore::SVGPathTraversalStateBuilder::lineToHorizontal):
        (WebCore::SVGPathTraversalStateBuilder::lineToVertical):
        (WebCore::SVGPathTraversalStateBuilder::curveToCubicSmooth):
        (WebCore::SVGPathTraversalStateBuilder::curveToQuadratic):
        (WebCore::SVGPathTraversalStateBuilder::curveToQuadraticSmooth):
        (WebCore::SVGPathTraversalStateBuilder::arcTo):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65605 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Sheriff Bot <webkit.review.bot@gmail.com>
hamaji@chromium.org [Wed, 18 Aug 2010 13:09:28 +0000 (13:09 +0000)]
2010-08-18  Sheriff Bot  <webkit.review.bot@gmail.com>

        Unreviewed, rolling out r65238.
        http://trac.webkit.org/changeset/65238
        https://bugs.webkit.org/show_bug.cgi?id=44168

        Check if clobbering chromium bot fixes the layout test
        failures (Requested by hamaji on #webkit).

        * public/WebFrameClient.h:
        (WebKit::WebFrameClient::frameDetached):
        * src/FrameLoaderClientImpl.cpp:
        (WebKit::FrameLoaderClientImpl::detachedFromParent3):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65604 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Mahesh Kulkarni <mahesh.kulkarni@nokia.com>
commit-queue@webkit.org [Wed, 18 Aug 2010 12:36:33 +0000 (12:36 +0000)]
2010-08-18  Mahesh Kulkarni  <mahesh.kulkarni@nokia.com>

        Reviewed by Steve Block.

        [Qt] Request for permission before starting Geolocation service
        https://bugs.webkit.org/show_bug.cgi?id=42027

        Handles starting location acquisition when request is granted for ports using
        "!CLIENT_BASED_GEOLOCATION and PREEMPT_GEOLOCATION_PERMISSION" policy
        Only Qt port as of today uses the above combination. Enable
        PREEMPT_GEOLOCATION_PERMISSION for qt port.

        Below tests cover these changes on platform using PREEMPT_GEOLOCATION_PERMISSION
        - fast/dom/Geolocation/delayed-permission-allowed.html
        - fast/dom/Geolocation/delayed-permission-denied.html
        - fast/dom/Geolocation/delayed-permission-allowed-for-multiple-requests.html
        - fast/dom/Geolocation/delayed-permission-denied-for-multiple-requests.html

        * WebCore.pro:
        * page/Geolocation.cpp:
        (WebCore::Geolocation::handlePendingPermissionNotifiers):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65603 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Yuta Kitamura <yutak@chromium.org>
morrita@google.com [Wed, 18 Aug 2010 12:25:33 +0000 (12:25 +0000)]
2010-08-18  Yuta Kitamura  <yutak@chromium.org>

        Unreviewed, test expectation update.

        Remove duplicate expectation.

        Update Chromium test expectations
        https://bugs.webkit.org/show_bug.cgi?id=44162

        * platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65602 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Zoltan Horvath <zoltan@webkit.org>
zoltan@webkit.org [Wed, 18 Aug 2010 12:15:23 +0000 (12:15 +0000)]
2010-08-18  Zoltan Horvath  <zoltan@webkit.org>

        Reviewed by Kenneth Rohde Christiansen.

        [Qt] Forwards QtWebProcess' stdout, stderr onto the UI process
        https://bugs.webkit.org/show_bug.cgi?id=44159

        For debugging purposes forwards stdout and stderr of QtWebProcess onto the UI process.

        * UIProcess/Launcher/qt/ProcessLauncherQt.cpp:
        (WebKit::ProcessLauncherHelper::launch):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65601 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoFix hang when saving crash logs on Windows
aroben@apple.com [Wed, 18 Aug 2010 11:42:45 +0000 (11:42 +0000)]
Fix hang when saving crash logs on Windows

* Scripts/old-run-webkit-tests:
(setUpWindowsCrashLogSaving):
(END):
Pass -s to regtool so it will write the Auto value as a string instead
of as a number. This was causing a "do you want to debug?" dialog to
appear.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65600 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Yuta Kitamura <yutak@chromium.org>
morrita@google.com [Wed, 18 Aug 2010 10:44:09 +0000 (10:44 +0000)]
2010-08-18  Yuta Kitamura  <yutak@chromium.org>

        Unreviewed. test expectation update.

        Update Chromium test expectations.
        https://bugs.webkit.org/show_bug.cgi?id=44162

        * platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65599 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Sheriff Bot <webkit.review.bot@gmail.com>
loislo@chromium.org [Wed, 18 Aug 2010 10:35:25 +0000 (10:35 +0000)]
2010-08-18  Sheriff Bot  <webkit.review.bot@gmail.com>

        Unreviewed, rolling out r65595.
        http://trac.webkit.org/changeset/65595
        https://bugs.webkit.org/show_bug.cgi?id=44161

        qt build failed (Requested by loislo on #webkit).

        * GNUmakefile.am:
2010-08-18  Sheriff Bot  <webkit.review.bot@gmail.com>

        Unreviewed, rolling out r65595.
        http://trac.webkit.org/changeset/65595
        https://bugs.webkit.org/show_bug.cgi?id=44161

        qt build failed (Requested by loislo on #webkit).

        * platform/gtk/Skipped:
2010-08-18  Sheriff Bot  <webkit.review.bot@gmail.com>

        Unreviewed, rolling out r65595.
        http://trac.webkit.org/changeset/65595
        https://bugs.webkit.org/show_bug.cgi?id=44161

        qt build failed (Requested by loislo on #webkit).

        * GNUmakefile.am:
        * WebCore.gyp/WebCore.gyp:
        * WebCore.gypi:
        * WebCore.pri:
        * WebCore.pro:
        * WebCore.vcproj/copyInspectorFiles.cmd:
        * WebCore.xcodeproj/project.pbxproj:
        * combine-javascript-resources:
        * inspector/CodeGeneratorInspector.pm:
        * inspector/front-end/InspectorBackendStub.js: Added.
        (WebInspector.InspectorBackendStub):
        (WebInspector.InspectorBackendStub.prototype._registerDelegate):
        (WebInspector.InspectorBackendStub.prototype.sendMessageToBackend):
        * inspector/front-end/InspectorBackendStub.qrc: Removed.
        * inspector/front-end/WebKit.qrc:
2010-08-18  Sheriff Bot  <webkit.review.bot@gmail.com>

        Unreviewed, rolling out r65595.
        http://trac.webkit.org/changeset/65595
        https://bugs.webkit.org/show_bug.cgi?id=44161

        qt build failed (Requested by loislo on #webkit).

        * WebKit.gyp:
2010-08-18  Sheriff Bot  <webkit.review.bot@gmail.com>

        Unreviewed, rolling out r65595.
        http://trac.webkit.org/changeset/65595
        https://bugs.webkit.org/show_bug.cgi?id=44161

        qt build failed (Requested by loislo on #webkit).

        * WebCoreSupport/InspectorClientGtk.cpp:
        (WebKit::InspectorClient::openInspectorFrontend):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65598 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Pavel Feldman <pfeldman@chromium.org>
pfeldman@chromium.org [Wed, 18 Aug 2010 10:09:12 +0000 (10:09 +0000)]
2010-08-18  Pavel Feldman  <pfeldman@chromium.org>

        Reviewed by Yury Semikhatsky.

        Web Inspector: make objectId of string type instead of object.
        https://bugs.webkit.org/show_bug.cgi?id=44160

        * WebCore.xcodeproj/project.pbxproj:
        * inspector/front-end/InjectedScript.js:
        (injectedScriptConstructor.):
        * inspector/front-end/InjectedScriptAccess.js:
        (get InjectedScriptAccess.getForObjectId):
        * inspector/front-end/RemoteObject.js:
        (WebInspector.RemoteObject.prototype.getProperties):
        (WebInspector.RemoteObject.prototype.setPropertyValue):
        (WebInspector.RemoteObject.prototype.pushNodeToFrontend):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65597 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Gabor Loki <loki@webkit.org>
loki@webkit.org [Wed, 18 Aug 2010 10:02:47 +0000 (10:02 +0000)]
2010-08-18  Gabor Loki  <loki@webkit.org>

        Reviewed by Gavin Barraclough.

        The JITStackFrame is wrong using Thumb-2 JIT with JSVALUE32_64
        https://bugs.webkit.org/show_bug.cgi?id=43897

        A 64 bits wide member in a structure is aligned to 8 bytes on ARM by
        default, but this is not taken into account in the offset defines of
        JITStackFrame.

        * jit/JITStubs.cpp:
        * jit/JITStubs.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65596 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-16 Ilya Tikhonovsky <loislo@chromium.org>
loislo@chromium.org [Wed, 18 Aug 2010 10:00:06 +0000 (10:00 +0000)]
2010-08-16  Ilya Tikhonovsky  <loislo@chromium.org>

        Reviewed by Yury Semikhatsky.

        Web Inspector: replace hand written InspectorBackendStub.js by generated one.
        New command-line flag was added to combine-javascript-resources.
        It specifies the location of generated scripts.
        https://bugs.webkit.org/show_bug.cgi?id=43791

        * GNUmakefile.am:
        * WebCore.gyp/WebCore.gyp:
        * WebCore.gypi:
        * WebCore.pri:
        * WebCore.pro:
        * WebCore.vcproj/WebCore.vcproj:
        * WebCore.xcodeproj/project.pbxproj:
        * combine-javascript-resources:
        * inspector/CodeGeneratorInspector.pm:
        * inspector/front-end/InspectorBackendStub.js: Removed.
        * inspector/front-end/WebKit.qrc:

2010-08-16  Ilya Tikhonovsky  <loislo@chromium.org>

        Reviewed by Yury Semikhatsky.

        Web Inspector: replace hand written InspectorBackendStub.js by generated one.
        https://bugs.webkit.org/show_bug.cgi?id=43791

        * GNUmakefile.am:

2010-08-17  Ilya Tikhonovsky  <loislo@chromium.org>

        Reviewed by Yury Semikhatsky.

        WebInspector: inspector tests for GTK were added to Skip list.
        They should be enabled later.
        Bug for GTK https://bugs.webkit.org/show_bug.cgi?id=43977
        https://bugs.webkit.org/show_bug.cgi?id=43791

        * platform/gtk/Skipped:

2010-08-17  Ilya Tikhonovsky  <loislo@chromium.org>

        Reviewed by Yury Semikhatsky.

        Web Inspector: replace hand written InspectorBackendStub.js by generated one.
        https://bugs.webkit.org/show_bug.cgi?id=43791

        * WebKit.gyp:

2010-08-17  Ilya Tikhonovsky  <loislo@chromium.org>

        Reviewed by Yury Semikhatsky.

        Web Inspector: replace hand written InspectorBackendStub.js by generated one.
        https://bugs.webkit.org/show_bug.cgi?id=43791

        * WebCoreSupport/InspectorClientGtk.cpp:
        (WebKit::InspectorClient::openInspectorFrontend):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65595 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Pavel Feldman <pfeldman@chromium.org>
pfeldman@chromium.org [Wed, 18 Aug 2010 09:37:58 +0000 (09:37 +0000)]
2010-08-17  Pavel Feldman  <pfeldman@chromium.org>

        Reviewed by Yury Semikhatsky.

        Chromium DevTools: Support runtime property name/value pairs
        instead of feature names as navigation state.
        https://bugs.webkit.org/show_bug.cgi?id=44054

        * public/WebDevToolsAgent.h:
        * public/WebDevToolsAgentClient.h:
        (WebKit::WebDevToolsAgentClient::runtimePropertyChanged):
        * src/WebDevToolsAgentImpl.cpp:
        (WebKit::WebDevToolsAgentImpl::setRuntimeProperty):
        (WebKit::WebDevToolsAgentImpl::setApuAgentEnabled):
        * src/WebDevToolsAgentImpl.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65594 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoRename UString::substr to substringSharingImpl, add to WTF::String.
barraclough@apple.com [Wed, 18 Aug 2010 08:56:49 +0000 (08:56 +0000)]
Rename UString::substr to substringSharingImpl, add to WTF::String.
Now WTF::String can do everything that JSC::UString can do!

Rubber stamped by Sam Weinig.

* JavaScriptCore.exp:
* bytecode/CodeBlock.cpp:
(JSC::escapeQuotes):
* bytecompiler/NodesCodegen.cpp:
(JSC::substitute):
* parser/SourceProvider.h:
(JSC::UStringSourceProvider::getRange):
* runtime/FunctionPrototype.cpp:
(JSC::insertSemicolonIfNeeded):
* runtime/JSGlobalObjectFunctions.cpp:
(JSC::parseInt):
* runtime/JSONObject.cpp:
(JSC::gap):
(JSC::Stringifier::indent):
(JSC::Stringifier::unindent):
* runtime/JSString.cpp:
(JSC::JSString::replaceCharacter):
* runtime/NumberPrototype.cpp:
(JSC::numberProtoFuncToFixed):
(JSC::numberProtoFuncToPrecision):
* runtime/StringPrototype.cpp:
(JSC::stringProtoFuncReplace):
(JSC::trimString):
* runtime/UString.cpp:
(JSC::UString::substringSharingImpl):
* runtime/UString.h:
* wtf/text/WTFString.cpp:
(WTF::String::substringSharingImpl):
* wtf/text/WTFString.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65593 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoWindows build fix.
barraclough@apple.com [Wed, 18 Aug 2010 08:44:12 +0000 (08:44 +0000)]
Windows build fix.

* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:
* JavaScriptCore.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65592 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Shinichiro Hamaji <hamaji@chromium.org>
hamaji@chromium.org [Wed, 18 Aug 2010 08:42:40 +0000 (08:42 +0000)]
2010-08-18  Shinichiro Hamaji  <hamaji@chromium.org>

        Reviewed by Dimitri Glazkov.

        Printer font is different from screen font
        https://bugs.webkit.org/show_bug.cgi?id=43340

        Not testable because PNG output in layoutTestController.setPrinting
        for chromium isn't implemented yet.

        * platform/graphics/FontDescription.h: Use screen font even for printing on Mac Chromium
        (WebCore::FontDescription::setUsePrinterFont):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65591 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoWebCore: Fix displaying problem of feTurbulence.
zherczeg@webkit.org [Wed, 18 Aug 2010 08:38:37 +0000 (08:38 +0000)]
WebCore: Fix displaying problem of feTurbulence.
https://bugs.webkit.org/show_bug.cgi?id=44111

Patch by Renata Hodovan <reni@inf.u-szeged.hu> on 2010-08-18
Reviewed by Dirk Schulze.

Change the order of calculating turbulenceFunctionResult
and clamping it. Furthermore remove the unnecessary rounding of m_seed.
Add a new test for checking the change of seed's value.

Test: svg/W3C-SVG-1.1/filters-turb-02-f.svg

* svg/graphics/filters/SVGFETurbulence.cpp:
(WebCore::FETurbulence::initPaint):
(WebCore::FETurbulence::calculateTurbulenceValueForPoint):
(WebCore::FETurbulence::apply):

LayoutTests: https://bugs.webkit.org/show_bug.cgi?id=44111

Patch by Renata Hodovan <reni@inf.u-szeged.hu> on 2010-08-18
Reviewed by Dirk Schulze.

Update expected result of test svg/W3C-SVG-1.1/filters-turb-01-f.svg and
add a new test for checking the change of the seed value.

* platform/mac/svg/W3C-SVG-1.1/filters-turb-01-f-expected.checksum:
* platform/mac/svg/W3C-SVG-1.1/filters-turb-01-f-expected.png:
* platform/mac/svg/W3C-SVG-1.1/filters-turb-02-f-expected.checksum: Added.
* platform/mac/svg/W3C-SVG-1.1/filters-turb-02-f-expected.png: Added.
* platform/mac/svg/W3C-SVG-1.1/filters-turb-02-f-expected.txt: Added.
* svg/W3C-SVG-1.1/filters-turb-02-f.svg: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65590 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoWindows build fix.
barraclough@apple.com [Wed, 18 Aug 2010 08:16:53 +0000 (08:16 +0000)]
Windows build fix.

* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:
* JavaScriptCore.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65589 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoBug 44146 - Remove toDouble/toUInt32 methods from UString.
barraclough@apple.com [Wed, 18 Aug 2010 07:41:22 +0000 (07:41 +0000)]
Bug 44146 - Remove toDouble/toUInt32 methods from UString.

Reviewed by Sam Weinig.

JavaScriptCore:

These methods all implement JavaScript language specific behaviour, and as such
are not suited to being on a generic string object.  They are also inefficient
and incorrectly used, refactor & cleanup.  Uses of these methods really divide
out into two cases.

ToNumber:
Uses of toDouble from JSString and from parseFloat are implementing ecma's
ToNumber conversion from strings (see ecma-262 9.3.1), so UString::toDouble
should largely just be moved out to a global jsToNumber function.  ToNumber is
capable of recognizing either decimal or hexadecimal numbers, but parseFloat
should only recognize decimal values.  This is currently handled by testing for
hexadecimal before calling toDouble, which should unnecessary - instead we can
just split out the two parts to the grammar into separate functions. Also,
strtod recognizes a set of literals (nan, inf, and infinity - all with any
capitalization) - which are not defined by any of the specs we are implementing.
To handle this we need to perform additional work in toDouble to convert the
unsupported cases of infinities back to NaNs.  Instead we should simply remove
support for this literals from strtod.  This should provide a more desirable
behaviour for all clients of strtod.

Indexed properties:
Uses of the toStrictUInt32 methods are were all converting property names to
indices, and all uses of toUInt32 were incorrect; in all cases we should have
been calling toUInt32.  This error results in some incorrect behaviour in the
DOM (accessing property "0 " of a NodeList should fail; it currently does not).
Move this method onto Identifier (our canonical property name), and make it
always perform a strict conversion. Add a layout test to check NodeList does
convert indexed property names correctly.

* JavaScriptCore.exp:
* runtime/Arguments.cpp:
(JSC::Arguments::getOwnPropertySlot):
(JSC::Arguments::getOwnPropertyDescriptor):
(JSC::Arguments::put):
(JSC::Arguments::deleteProperty):
* runtime/Identifier.cpp:
(JSC::Identifier::toUInt32):
* runtime/Identifier.h:
(JSC::Identifier::toUInt32):
* runtime/JSArray.cpp:
(JSC::JSArray::getOwnPropertySlot):
(JSC::JSArray::getOwnPropertyDescriptor):
(JSC::JSArray::put):
(JSC::JSArray::deleteProperty):
* runtime/JSArray.h:
(JSC::Identifier::toArrayIndex):
* runtime/JSByteArray.cpp:
(JSC::JSByteArray::getOwnPropertySlot):
(JSC::JSByteArray::getOwnPropertyDescriptor):
(JSC::JSByteArray::put):
* runtime/JSGlobalObjectFunctions.cpp:
(JSC::isInfinity):
(JSC::jsHexIntegerLiteral):
(JSC::jsStrDecimalLiteral):
(JSC::jsToNumber):
(JSC::parseFloat):
* runtime/JSGlobalObjectFunctions.h:
* runtime/JSString.cpp:
(JSC::JSString::getPrimitiveNumber):
(JSC::JSString::toNumber):
(JSC::JSString::getStringPropertyDescriptor):
* runtime/JSString.h:
(JSC::JSString::getStringPropertySlot):
* runtime/ObjectPrototype.cpp:
(JSC::ObjectPrototype::put):
* runtime/StringObject.cpp:
(JSC::StringObject::deleteProperty):
* runtime/UString.cpp:
* runtime/UString.h:
* wtf/dtoa.cpp:
(WTF::strtod):

WebCore:

These methods all implement JavaScript language specific behaviour, and as such
are not suited to being on a generic string object.  They are also inefficient
and incorrectly used, refactor & cleanup.  Uses of these methods really divide
out into two cases.

ToNumber:
Uses of toDouble from JSString and from parseFloat are implementing ecma's
ToNumber conversion from strings (see ecma-262 9.3.1), so UString::toDouble
should largely just be moved out to a global jsToNumber function.  ToNumber is
capable of recognizing either decimal or hexadecimal numbers, but parseFloat
should only recognize decimal values.  This is currently handled by testing for
hexadecimal before calling toDouble, which should unnecessary - instead we can
just split out the two parts to the grammar into separate functions. Also,
strtod recognizes a set of literals (nan, inf, and infinity - all with any
capitalization) - which are not defined by any of the specs we are implementing.
To handle this we need to perform additional work in toDouble to convert the
unsupported cases of infinities back to NaNs.  Instead we should simply remove
support for this literals from strtod.  This should provide a more desirable
behaviour for all clients of strtod.

Indexed properties:
Uses of the toStrictUInt32 methods are were all converting property names to
indices, and all uses of toUInt32 were incorrect; in all cases we should have
been calling toUInt32.  This error results in some incorrect behaviour in the
DOM (accessing property "0 " of a NodeList should fail; it currently does not).
Move this method onto Identifier (our canonical property name), and make it
always perform a strict conversion. Add a layout test to check NodeList does
convert indexed property names correctly.

Test: fast/dom/NodeList/nodelist-item-with-index.html

* WebCore.xcodeproj/project.pbxproj:
* bindings/js/JSDOMWindowCustom.cpp:
(WebCore::JSDOMWindow::getOwnPropertySlot):
(WebCore::JSDOMWindow::getOwnPropertyDescriptor):
* bindings/js/JSHTMLAllCollectionCustom.cpp:
(WebCore::callHTMLAllCollection):
(WebCore::JSHTMLAllCollection::item):
* bindings/js/JSHTMLCollectionCustom.cpp:
(WebCore::callHTMLCollection):
(WebCore::JSHTMLCollection::item):
* bindings/js/JSNodeListCustom.cpp:
(WebCore::callNodeList):
* bindings/scripts/CodeGeneratorJS.pm:
* bridge/runtime_array.cpp:
(JSC::RuntimeArray::getOwnPropertySlot):
(JSC::RuntimeArray::getOwnPropertyDescriptor):
(JSC::RuntimeArray::put):

LayoutTests:

Test that indexing into nodelists works correctly, particularly
wrt indices passed as strings that contain whitespace.

* fast/dom/NodeList/nodelist-item-with-index-expected.txt: Added.
* fast/dom/NodeList/nodelist-item-with-index.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65588 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-18 Adam Barth <abarth@webkit.org>
abarth@webkit.org [Wed, 18 Aug 2010 07:37:56 +0000 (07:37 +0000)]
2010-08-18  Adam Barth  <abarth@webkit.org>

        Reviewed by Adele Peterson.

        Null dereference in DOMSelection::deleteFromDocument
        https://bugs.webkit.org/show_bug.cgi?id=44153

        deleteFromDocument checks selection->isNone() before calling
        selection->selection().toNormalizedRange(), but toNormalizedRange()
        notes that it needs to updateLayout(), which can make the selection
        isNone() again.  In that case, we crash on a NULL pointer in
        deleteFromDocument.  I don't know how to trigger that situation in a
        test, but cross_fuzz was able to hit it, so we should fix it.

        * page/DOMSelection.cpp:
        (WebCore::DOMSelection::deleteFromDocument):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65587 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoReviewed by Antonio Gomes.
girish@forwardbias.in [Wed, 18 Aug 2010 03:53:08 +0000 (03:53 +0000)]
Reviewed by Antonio Gomes.

[Qt] On Maemo5, there is no libgdk-x11-2.0.so symlink. Look for
libgdk-x11-2.0.so.0 instead on X11 and Maemo5.

https://bugs.webkit.org/show_bug.cgi?id=44043

* plugins/qt/PluginViewQt.cpp:
(WebCore::getPluginDisplay):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65586 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Xan Lopez <xlopez@igalia.com>
xan@webkit.org [Wed, 18 Aug 2010 03:08:03 +0000 (03:08 +0000)]
2010-08-17  Xan Lopez  <xlopez@igalia.com>

        Reviewed by Gustavo Noronha.

        Fix the builds when HAVE_GSETTINGS is defined.

        * webkit/webkitprivate.cpp:
        (isSchemaAvailable):
        (inspectorGSettings):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65585 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Ademar de Souza Reis Jr <ademar.reis@openbossa.org>
commit-queue@webkit.org [Wed, 18 Aug 2010 01:55:42 +0000 (01:55 +0000)]
2010-08-17  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>

        Reviewed by Daniel Bates.

        REGRESSION(53060): http/tests/security/xssAuditor/malformed-HTML.html timed out on Qt Linux Release Bot
        https://bugs.webkit.org/show_bug.cgi?id=33460

        Enable http/tests/security/xssAuditor/malformed-HTML.html
        tested with qt-4.7 and qt-4.6.

        * platform/qt/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65584 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Ademar de Souza Reis Jr <ademar.reis@openbossa.org>
commit-queue@webkit.org [Wed, 18 Aug 2010 01:38:54 +0000 (01:38 +0000)]
2010-08-17  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>

        Reviewed by Ariya Hidayat.

        Enable http/tests/security/cross-frame-access-name-getter.html

        [Qt] http/tests/security/cross-frame-access-name-getter.html timed out
        https://bugs.webkit.org/show_bug.cgi?id=38540

        Test works with qt-4.7 and qt-4.6.

        * platform/qt/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65583 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Victor Wang <victorw@chromium.org>
victorw@chromium.org [Wed, 18 Aug 2010 01:37:11 +0000 (01:37 +0000)]
2010-08-17  Victor Wang  <victorw@chromium.org>

        Unreviewed. Fixed chromium incremental test json upload.
        Quote builder name and test results server in url.

        * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65582 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Laszlo Gombos <laszlo.1.gombos@nokia.com>
laszlo.1.gombos@nokia.com [Wed, 18 Aug 2010 01:22:05 +0000 (01:22 +0000)]
2010-08-17  Laszlo Gombos  <laszlo.1.gombos@nokia.com>

        Reviewed by Ariya Hidayat.

        [Qt] [Symbian] Consistently use Q_OS_SYMBIAN to guard all Symbian platform dependencies
        https://bugs.webkit.org/show_bug.cgi?id=44124

        Q_WS_S60 is not defined for Symbian^4 devices as Q_WS_S60 used to guard
        Avkon UI framework dependencies. Use Q_OS_SYMBIAN everywhere to mark
        Symbian dependencies.

        * WebCoreSupport/QtFallbackWebPopup.cpp:
        (WebCore::QtFallbackWebPopup::populate):
2010-08-17  Laszlo Gombos  <laszlo.1.gombos@nokia.com>

        Reviewed by Ariya Hidayat.

        [Qt] [Symbian] Consistently use Q_OS_SYMBIAN to guard all Symbian platform dependencies
        https://bugs.webkit.org/show_bug.cgi?id=44124

        Q_WS_S60 is not defined for Symbian^4 devices as Q_WS_S60 used to guard
        Avkon UI framework dependencies. Use Q_OS_SYMBIAN everywhere to mark
        Symbian dependencies.

        * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
        (WebCore::DumpRenderTree::open):
        * QtTestBrowser/launcherwindow.cpp:
        (LauncherWindow::init):
        (LauncherWindow::toggleFullScreenMode):
        (LauncherWindow::showFPS):
        (LauncherWindow::updateFPS):
        * QtTestBrowser/launcherwindow.h:
        (WindowOptions::WindowOptions):
        * QtTestBrowser/mainwindow.cpp:
        (MainWindow::buildUI):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65581 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago[Qt] Remove unused ContextShadowParameter class in GraphicsContextQt
ariya@webkit.org [Wed, 18 Aug 2010 01:20:58 +0000 (01:20 +0000)]
[Qt] Remove unused ContextShadowParameter class in GraphicsContextQt
https://bugs.webkit.org/show_bug.cgi?id=44142

Patch by Ariya Hidayat <ariya@sencha.com> on 2010-08-17
Reviewed by Antonio Gomes.

The class is not needed anymore after r65425.
We know have dedicated ContextShadow in its own file.

* platform/graphics/qt/GraphicsContextQt.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65580 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoAdd NetworkingContext to avoid layer violations
darin@apple.com [Wed, 18 Aug 2010 01:01:40 +0000 (01:01 +0000)]
Add NetworkingContext to avoid layer violations
https://bugs.webkit.org/show_bug.cgi?id=42292

Patch by Jesus Sanchez-Palencia <jesus.palencia@openbossa.org> on 2010-08-17
Reviewed by Darin Adler.

Preparation: Just add the files to the build system.

.:

* GNUmakefile.am: Added new files.

WebCore:

* GNUmakefile.am: Added new files.
* WebCore.gypi: Ditto.
* WebCore.pro: Ditto.
* WebCore.vcproj/WebCore.vcproj: Ditto.
* WebCore.xcodeproj/project.pbxproj: Ditto.

* loader/FrameNetworkingContext.h: Added.
* platform/network/NetworkingContext.h: Added.
Placeholders with tentative code that might be changed when landing
the rest of it.

WebKit:

* WebKit.xcodeproj/project.pbxproj: Added new files.

WebKit/chromium:

* WebKit.gyp: Added new files.
* src/FrameNetworkingContextImpl.h: Added.
Placeholder with tentative code that might be changed when landing
the rest of it.

WebKit/gtk:

* WebCoreSupport/FrameNetworkingContextGtk.h: Added.
Placeholder with tentative code that might be changed when landing
the rest of it.

WebKit/mac:

* WebCoreSupport/WebFrameNetworkingContext.h: Added.
Placeholder with tentative code that might be changed when landing
the rest of it.
* WebCoreSupport/WebFrameNetworkingContext.mm: Added.
Empty placeholder for now.

WebKit/qt:

* WebCoreSupport/FrameNetworkingContextQt.cpp: Added.
Empty placeholder for now.
* WebCoreSupport/FrameNetworkingContextQt.h: Added.
Placeholder with tentative code that might be changed when landing
the rest of it.

WebKit/win:

* WebCoreSupport/WebFrameNetworkingContext.cpp: Added.
Empty placeholder for now.
* WebCoreSupport/WebFrameNetworkingContext.h: Added.
Placeholder with tentative code that might be changed when landing
the rest of it.
* WebKit.vcproj/WebKit.vcproj: Added new files.

WebKit2:

* WebKit2.xcodeproj/project.pbxproj: Added new files.
* WebProcess/WebCoreSupport/mac/WebFrameNetworkingContext.h: Added.
Placeholder with tentative code that might be changed when landing
the rest of it.
* WebProcess/WebCoreSupport/mac/WebFrameNetworkingContext.mm: Added.
Empty placeholder for now.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65579 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Dirk Pranke <dpranke@chromium.org>
dpranke@chromium.org [Wed, 18 Aug 2010 00:59:47 +0000 (00:59 +0000)]
2010-08-17  Dirk Pranke  <dpranke@chromium.org>

        Reviewed by David Levin.

        new-run-webkit-tests: remove --show-sources option

        --show-sources is pretty much obsolete with --trace everything, so
        I'm removing it.

        Also rename a couple of methods in the TestTextDiff class to make their
        intended visibility (private) more obvious.

        https://bugs.webkit.org/show_bug.cgi?id=44143

        * Scripts/webkitpy/layout_tests/layout_package/printing.py:
        * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
        * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
        * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
        * Scripts/webkitpy/layout_tests/test_types/text_diff.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65578 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Dirk Pranke <dpranke@chromium.org>
dpranke@chromium.org [Wed, 18 Aug 2010 00:55:30 +0000 (00:55 +0000)]
2010-08-17  Dirk Pranke  <dpranke@chromium.org>

        Reviewed by David Levin.

        remove --fuzzy-image-diff in new-run-webkit-tests (it doesn't work)

        This code bit-rotted at some point more than a year ago, and nobody seems
        to miss it. old-run-webkit-tests has a --tolerance flag that new-run-webkit-tests
        should support instead, making this flag unnecessary anyway.

        https://bugs.webkit.org/show_bug.cgi?id=44141

        * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
        * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
        * Scripts/webkitpy/layout_tests/test_types/fuzzy_image_diff.py: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65577 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoAdding a MathML Demo and screenshot.
bdakin@apple.com [Wed, 18 Aug 2010 00:41:35 +0000 (00:41 +0000)]
Adding a MathML Demo and screenshot.

Rubber-stamped by Adele.

* blog-files/riemann-screenshot.png: Added.
* demos/mathml: Added.
* demos/mathml/MathMLDemo.xhtml: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65576 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Mihai Parparita <mihaip@chromium.org>
commit-queue@webkit.org [Tue, 17 Aug 2010 23:57:42 +0000 (23:57 +0000)]
2010-08-17  Mihai Parparita  <mihaip@chromium.org>

        Reviewed by Dimitri Glazkov.

        transitions/change-values-during-transition.html fails intermittently
        https://bugs.webkit.org/show_bug.cgi?id=28461

        Make test less flaky during high CPU usage by changing the expectation
        from a hardcoded value to an interpolated one based on when the timeout
        actually fires. Made the test much more reliable when running with
        --iterations=100 while maxing out the CPU on my machine with a
        background process.

        * transitions/change-values-during-transition.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65575 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Victor Wang <victorw@chromium.org>
victorw@chromium.org [Tue, 17 Aug 2010 23:56:24 +0000 (23:56 +0000)]
2010-08-17  Victor Wang  <victorw@chromium.org>

        Reviewed by Ojan Vafai.

        Update json results generator to have incremental json including
        results for tests that pass in current run but failed before.

        https://bugs.webkit.org/show_bug.cgi?id=44119

        * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
        * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
        * Scripts/webkitpy/layout_tests/run_webkit_tests.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65574 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Mihai Parparita <mihaip@chromium.org>
commit-queue@webkit.org [Tue, 17 Aug 2010 23:40:00 +0000 (23:40 +0000)]
2010-08-17  Mihai Parparita  <mihaip@chromium.org>

        Reviewed by Darin Fisher.

        Back button after Google search result click broken
        https://bugs.webkit.org/show_bug.cgi?id=44086

        Added a reduced test case of Google search result click behavior,
        renamed gesture-before-onload-target.html to redirect-target.html
        since it's useful for this test too.

        * fast/history/gesture-before-onload-form-submit-expected.txt:
        * fast/history/gesture-before-onload-form-submit.html:
        * fast/history/gesture-before-onload-location-href-expected.txt:
        * fast/history/gesture-before-onload-location-href.html:
        * fast/history/redirect-via-iframe-expected.txt: Added.
        * fast/history/redirect-via-iframe.html: Added.
        * fast/history/resources/iframe-redirect.html: Added.
        * fast/history/resources/redirect-target.html: Renamed from LayoutTests/fast/history/resources/gesture-before-onload-target.html.
2010-08-17  Mihai Parparita  <mihaip@chromium.org>

        Reviewed by Darin Fisher.

        Back button after Google search result click broken
        https://bugs.webkit.org/show_bug.cgi?id=44086

        Test: fast/history/redirect-via-iframe.html

        isLoadingInAPISense doesn't quite map to onload (it also returns true
        if frames are loading, even after the parent frame has fired its onload
        handler). Instead, add a flag for the onload event being fired, and
        use that.

        * loader/DocumentLoader.cpp:
        (WebCore::DocumentLoader::DocumentLoader):
        * loader/DocumentLoader.h:
        (WebCore::DocumentLoader::handledOnloadEvents):
        (WebCore::DocumentLoader::wasOnloadHandled):
        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::handledOnloadEvents):
        * loader/RedirectScheduler.cpp:
        (WebCore::RedirectScheduler::mustLockBackForwardList):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65573 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Dirk Pranke <dpranke@chromium.org>
dpranke@chromium.org [Tue, 17 Aug 2010 23:11:25 +0000 (23:11 +0000)]
2010-08-17  Dirk Pranke  <dpranke@chromium.org>

        Reviewed by Eric Seidel.

        fix test-webkitpy, add easy way to find a checkout root

        test-webkitpy currently doesn't work right if run from someplace other
        than the checkout root, and it spews a bunch of debug logging because
        the deduplicate_tests tests contaminates the test environment.

        This patch cleans up the deduplicate_tests unit tests, and creates
        two new methods in scm.py: find_checkout_root() and default_scm(),
        both of which use a single algorithm for guessing what checkout root
        to use if you aren't explicitly told one from a path.

        https://bugs.webkit.org/show_bug.cgi?id=44001

        * Scripts/deduplicate-tests:
        * Scripts/webkitpy/common/checkout/scm.py:
        * Scripts/webkitpy/common/checkout/scm_unittest.py:
        * Scripts/webkitpy/layout_tests/deduplicate_tests.py:
        * Scripts/webkitpy/layout_tests/deduplicate_tests_unittest.py:
        * Scripts/webkitpy/layout_tests/port/test.py:
        * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
        * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
        * Scripts/webkitpy/tool/main.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65572 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoBug 44099 - REGRESSION(r65468): Crashes in StringImpl::find
barraclough@apple.com [Tue, 17 Aug 2010 23:05:50 +0000 (23:05 +0000)]
Bug 44099 - REGRESSION(r65468): Crashes in StringImpl::find

Reviewed by Sam Weinig.

Bug 44080 introuduced a couple of cases in which array bounds could be overrun.
One of these was fixed in r65493, this patch fixes the other and address the
concerns voiced in comment #6 by restructuring the loops to remove the code
dupliction without introducing an additional if check.

* wtf/text/StringImpl.cpp:
(WTF::StringImpl::find):
(WTF::StringImpl::findIgnoringCase):
(WTF::StringImpl::reverseFind):
(WTF::StringImpl::reverseFindIgnoringCase):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65571 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Steve Block <steveblock@google.com>
steveblock@google.com [Tue, 17 Aug 2010 23:05:42 +0000 (23:05 +0000)]
2010-08-17  Steve Block  <steveblock@google.com>

        Reviewed by Jeremy Orlow.

        Geolocation clearWatch() needs to protect against invalid IDs
        https://bugs.webkit.org/show_bug.cgi?id=44096

        * fast/dom/Geolocation/clear-watch-invalid-id-crash-expected.txt: Added.
        * fast/dom/Geolocation/clear-watch-invalid-id-crash.html: Added.
        * fast/dom/Geolocation/script-tests/clear-watch-invalid-id-crash.js: Added.
        * fast/dom/Geolocation/script-tests/notimer-after-unload.js:
        (document.body.onload):
2010-08-17  Steve Block  <steveblock@google.com>

        Reviewed by Jeremy Orlow.

        Geolocation clearWatch() needs to protect against invalid IDs
        https://bugs.webkit.org/show_bug.cgi?id=44096

        If the ID passed to clearWatch() is invalid, we early-out.

        Test: fast/dom/Geolocation/clear-watch-invalid-id-crash.html

        * page/Geolocation.cpp:
        (WebCore::Geolocation::Watchers::set):
        (WebCore::Geolocation::Watchers::remove):
        (WebCore::Geolocation::watchPosition):
        (WebCore::Geolocation::clearWatch):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65570 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Stephen White <senorblanco@chromium.org>
senorblanco@chromium.org [Tue, 17 Aug 2010 23:00:15 +0000 (23:00 +0000)]
2010-08-17  Stephen White  <senorblanco@chromium.org>

        Reviewed by David Levin.

        Allow non-Skia ports to compile with ENABLE(ACCELERATED_2D_CANVAS).
        https://bugs.webkit.org/show_bug.cgi?id=44130

        No tests; if it compiles, it's good.

        * html/canvas/CanvasRenderingContext2D.cpp:
        (WebCore::CanvasRenderingContext2D::CanvasRenderingContext2D):
        Fix a warning (treated as error).
        * platform/graphics/GraphicsContext.cpp:
        (WebCore::GraphicsContext::setGraphicsContext3D):
        (WebCore::GraphicsContext::syncSoftwareCanvas):
        * platform/graphics/GraphicsContext.h:
        Make setGraphicsContext3D() and syncSoftwareCanvas() stubbed out,
        but still present in non-Skia ports.  This allows Chrome Mac to
        compile with ENABLE(ACCELERATED_2D_CANVAS) (although it isn't hooked
        up yet, it should at least compile).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65569 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Victor Wang <victorw@chromium.org>
victorw@chromium.org [Tue, 17 Aug 2010 22:46:32 +0000 (22:46 +0000)]
2010-08-17  Victor Wang  <victorw@chromium.org>

        Reviewed by Ojan Vafai.

        Add support to the test results server for downloading json that
        contains test list only.

        This is for json results generator to generate incremental json
        results so that it includes results not only for tests failed in
        current run, but also tests failed before.

        Also set the results type to "N" (no data) instead of "P" (pass)
        if test results cannot be found in incremental json file.

        https://bugs.webkit.org/show_bug.cgi?id=44117

        * TestResultServer/handlers/testfilehandler.py:
        * TestResultServer/model/jsonresults.py:
        * TestResultServer/model/jsonresults_unittest.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65567 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoWebCore: Navigating back/forward during a modal dialog causes a crash when the modal...
beidson@apple.com [Tue, 17 Aug 2010 22:44:33 +0000 (22:44 +0000)]
WebCore: Navigating back/forward during a modal dialog causes a crash when the modal dialog is dismissed.
<rdar://problem/8313579> and https://bugs.webkit.org/show_bug.cgi?id=44131

Reviewed by Sam Weinig.

No new tests. (Especially since DRT doesn't spin inner run loops at all, it would be a significant
effort to give it the ability to perform the back/forward navigation in the background to reproduce
this bug)

Manual tests in place of layout tests:
* manual-tests/back-forward-during-alert-1.html: Added.
* manual-tests/back-forward-during-alert-2.html: Added.

* page/Page.cpp:
(WebCore::Page::goToItem):

WebKit: Remove developmentRegion from the project file as it shouldn't have been checked in.

Reviewed by Sam Weinig.

* WebKit.xcodeproj/project.pbxproj:

WebKit/mac: Navigating back/forward during a modal dialog causes a crash when the modal dialog is dismissed.
<rdar://problem/8313579> and https://bugs.webkit.org/show_bug.cgi?id=44131

Reviewed by Sam Weinig.

* WebView/WebView.mm:
(-[WebView canGoBack]): Return false if loads are deferred.
(-[WebView canGoForward]): Ditto.

WebKit/win: Navigating back/forward during a modal dialog causes a crash when the modal dialog is dismissed.
<rdar://problem/8313579> and https://bugs.webkit.org/show_bug.cgi?id=44131

Reviewed by Sam Weinig.

* WebView.cpp:
(WebView::canGoBack): Return false if loads are deferred.
(WebView::canGoForward): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65566 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Joseph Pecoraro <joepeck@webkit.org>
joepeck@webkit.org [Tue, 17 Aug 2010 22:36:15 +0000 (22:36 +0000)]
2010-08-17  Joseph Pecoraro  <joepeck@webkit.org>

        Reviewed by Pavel Feldman.

        Web Inspector: Add XMLHttpRequest Logging Tests
        https://bugs.webkit.org/show_bug.cgi?id=44092

        * http/tests/inspector/console-xhr-logging-expected.txt: Added.
        * http/tests/inspector/console-xhr-logging.html: Added.
        * http/tests/inspector/resources/xhr-exists.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65565 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Ojan Vafai <ojan@chromium.org>
ojan@chromium.org [Tue, 17 Aug 2010 22:33:57 +0000 (22:33 +0000)]
2010-08-17  Ojan Vafai  <ojan@chromium.org>

        Reviewed by Tony Chang.

        selectionStart clobbered if textarea isn't focused
        https://bugs.webkit.org/show_bug.cgi?id=43814

        * fast/forms/text-control-select-blurred-expected.txt: Added.
        * fast/forms/text-control-select-blurred.html: Added.
2010-08-17  Ojan Vafai  <ojan@chromium.org>

        Reviewed by Tony Chang.

        selectionStart clobbered if textarea isn't focused
        https://bugs.webkit.org/show_bug.cgi?id=43814

        If a text control is blurred, setting selectionStart/end should use
        the cached selection value, which is currently stored on the form
        control element.

        Test: fast/forms/text-control-select-blurred.html

        * rendering/RenderTextControl.cpp:
        (WebCore::RenderTextControl::setSelectionStart):
        (WebCore::RenderTextControl::setSelectionEnd):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65564 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoMessagePort.close() crashes if the owning context was destroyed.
dimich@chromium.org [Tue, 17 Aug 2010 22:29:42 +0000 (22:29 +0000)]
MessagePort.close() crashes if the owning context was destroyed.
https://bugs.webkit.org/show_bug.cgi?id=43140

Reviewed by David Levin.

WebCore:

Test: fast/events/message-port-context-destroyed.html

* dom/MessagePort.cpp:
(WebCore::MessagePort::postMessage):
(WebCore::MessagePort::start):
(WebCore::MessagePort::close):
(WebCore::MessagePort::contextDestroyed):
Use isEntangled() method to gate various operations on MessagePort. This method also takes into account m_closed bit.

* dom/ScriptExecutionContext.cpp:
(WebCore::ScriptExecutionContext::stopActiveDOMObjects):
(WebCore::ScriptExecutionContext::closeMessagePorts):
* dom/ScriptExecutionContext.h:
Add closeMessagePorts() method that closes message ports at the same moments when other ActiveDOMObjects are stopped.

LayoutTests:

* fast/events/message-port-context-destroyed-expected.txt: Added.
* fast/events/message-port-context-destroyed.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65563 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoUnreviewed. Build fix.
hclam@chromium.org [Tue, 17 Aug 2010 22:22:49 +0000 (22:22 +0000)]
Unreviewed. Build fix.

Submitted for Victoria Kirst <vrk@google.com>.

Rebaseline straggler Mac UI video regressions.
Fixing test expectations to match output from r65215 bug fix.

https://bugs.webkit.org/show_bug.cgi?id=44134

* platform/chromium-mac/fast/layers/video-layer-expected.checksum:
* platform/chromium-mac/fast/layers/video-layer-expected.png:
* platform/chromium-mac/media/video-empty-source-expected.checksum:
* platform/chromium-mac/media/video-empty-source-expected.png:
* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65562 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 John Gregg <johnnyg@google.com>
johnnyg@google.com [Tue, 17 Aug 2010 22:13:43 +0000 (22:13 +0000)]
2010-08-17  John Gregg  <johnnyg@google.com>

        Unreviewed, test expectations.

        Skip the filltable-*.html tests on mac and windows.  They have no expectations, so expecting
        them to FAIL is insufficient to make the bots happy.

        * platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65561 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoUse the right path style
aroben@apple.com [Tue, 17 Aug 2010 22:10:36 +0000 (22:10 +0000)]
Use the right path style

* Scripts/old-run-webkit-tests:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65560 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoFix typo
aroben@apple.com [Tue, 17 Aug 2010 22:08:51 +0000 (22:08 +0000)]
Fix typo

* Scripts/old-run-webkit-tests:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65559 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoDon't hang when running run-webkit-tests as a non-Administrator on Vista/7
aroben@apple.com [Tue, 17 Aug 2010 22:02:58 +0000 (22:02 +0000)]
Don't hang when running run-webkit-tests as a non-Administrator on Vista/7

* Scripts/old-run-webkit-tests:
(setUpWindowsCrashLogSaving): Use regtool to set NTSD as the
post-mortem debugger, rather than using NTSD itself. The latter waits
for user input when it fails to set the registry values; the former
does not.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65557 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago[Qt] Skip media/video-poster-delayed.html
vestbo@webkit.org [Tue, 17 Aug 2010 21:36:26 +0000 (21:36 +0000)]
[Qt] Skip media/video-poster-delayed.html

Reviewed by NOBODY.

The QtMultimedia backend has issues with video size hints, skip the test
until this is sorted out.

* platform/qt/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65547 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoInstruct developers to (optionally) install Debugging Tools for Windows
aroben@apple.com [Tue, 17 Aug 2010 20:58:23 +0000 (20:58 +0000)]
Instruct developers to (optionally) install Debugging Tools for Windows

These tools are used by some of our scripts, so are useful for
developers to have installed.

Fixes <http://webkit.org/b/44128> webkit.org building instructions
should recommend installing Debugging Tools for Windows

Reviewed by Sam Weinig.

* building/tools.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65540 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Fady Samuel <fsamuel@chromium.org>
jamesr@google.com [Tue, 17 Aug 2010 20:57:15 +0000 (20:57 +0000)]
2010-08-17  Fady Samuel  <fsamuel@chromium.org>

        Reviewed by David Hyatt.

        Updated table cell hit testing and painting to use binary search instead of linear scan.

        Improved Table Hit Testing and Painting Performance
        https://bugs.webkit.org/show_bug.cgi?id=43933

        Tests: fast/table/simple_paint.html
               tables/hittesting/filltable-emptycells.html
               tables/hittesting/filltable-levels.html
               tables/hittesting/filltable-outline.html
               tables/hittesting/filltable-rtl.html
               tables/hittesting/filltable-stress.html

        * rendering/RenderTableSection.cpp:
        (WebCore::RenderTableSection::paintObject): Dirty rect determined using binary search.
        (WebCore::RenderTableSection::nodeAtPoint): Cell hit determined using binary search
2010-08-17  Fady Samuel  <fsamuel@chromium.org>

        Reviewed by David Hyatt.

        Updated table cell hit testing and painting to use binary search instead of a linear scan.

        Layout tests check that hit testing of empty cells, overlapping cells,
        cells with outlines, right-to-left tables, and large
        tables all work as expected.

        Improved Table Hit Testing and Painting Performance
        https://bugs.webkit.org/show_bug.cgi?id=43933

        * fast/table/simple_paint.html: Added.
        * platform/chromium-linux/fast/table/simple_paint-expected.checksum: Added.
        * platform/chromium-linux/fast/table/simple_paint-expected.png: Added.
        * platform/chromium-linux/fast/table/simple_paint-expected.txt: Added.
        * platform/chromium-linux/tables/hittesting/filltable-emptycells-expected.txt: Added.
        * platform/chromium-linux/tables/hittesting/filltable-levels-expected.txt: Added.
        * platform/chromium-linux/tables/hittesting/filltable-outline-expected.txt: Added.
        * platform/chromium-linux/tables/hittesting/filltable-rtl-expected.txt: Added.
        * platform/chromium-linux/tables/hittesting/filltable-stress-expected.txt: Added.
        * platform/chromium/test_expectations.txt:
        * platform/gtk/tables/hittesting/filltable-emptycells-expected.txt: Added.
        * platform/gtk/tables/hittesting/filltable-levels-expected.txt: Added.
        * platform/gtk/tables/hittesting/filltable-outline-expected.txt: Added.
        * platform/gtk/tables/hittesting/filltable-rtl-expected.txt: Added.
        * platform/gtk/tables/hittesting/filltable-stress-expected.txt: Added.
        * tables/hittesting/filltable-emptycells.html: Added.
        * tables/hittesting/filltable-levels.html: Added.
        * tables/hittesting/filltable-outline.html: Added.
        * tables/hittesting/filltable-rtl.html: Added.
        * tables/hittesting/filltable-stress.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65539 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Abhishek Arya <inferno@chromium.org>
inferno@chromium.org [Tue, 17 Aug 2010 20:49:50 +0000 (20:49 +0000)]
2010-08-17  Abhishek Arya  <inferno@chromium.org>

        Reviewed by Dave Hyatt.

        Only destroy empty anonymous block when it is columns or column span block
        in RenderBlock::removeChild.
        https://bugs.webkit.org/show_bug.cgi?id=44035

        Test: fast/block/basic/empty-anonymous-block-remove-crash.html

        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::removeChild):
2010-08-17  Abhishek Arya  <inferno@chromium.org>

        Reviewed by Dave Hyatt.

        Tests that removing an empty anonymous block does not result in crash.
        https://bugs.webkit.org/show_bug.cgi?id=44035

        * fast/block/basic/empty-anonymous-block-remove-crash-expected.txt: Added.
        * fast/block/basic/empty-anonymous-block-remove-crash.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65538 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years agoTeach run-webkit-tests, DumpRenderTree, and WebKitTestRunner how to save crash logs...
aroben@apple.com [Tue, 17 Aug 2010 20:45:10 +0000 (20:45 +0000)]
Teach run-webkit-tests, DumpRenderTree, and WebKitTestRunner how to save crash logs on Windows

The crash logs are saved to /tmp/layout-test-results and have names
like CrashLog_02bc_2010-08-17_14-36-20-108.txt.

Unfortunately, crashes in the WebKit2 web process are recorded as
time-outs by run-webkit-tests. Fixing this is covered by
<http://webkit.org/b/44121>.

Fixes <http://webkit.org/b/37859> <rdar://problem/7883560>
DumpRenderTree should save a stack trace and/or dump file when it
crashes

Reviewed by Sam Weinig.

* DumpRenderTree/win/DumpRenderTree.cpp:
(exceptionFilter):
(main):
* WebKitTestRunner/win/TestControllerWin.cpp:
(WTR::exceptionFilter):
(WTR::TestController::platformInitialize):
These changes set up an exception filter that prints #CRASHED to
stderr, then lets the exception propagate normally. This allows
run-webkit-tests to detect when we've crashed even when a post-mortem
debugger attaches to the process.

* Scripts/old-run-webkit-tests:
(top level): Declared some variables used by the crash-log-saving
code.
(openDumpTool): Copy _NT_SYMBOL_PATH into the clean environment so
that the post-mortem debugger has access to it.
(toCygwinPath): Added. This is the opposite of toWindowsPath.
(readFromDumpToolWithTimer): If the dump tool prints #CRASHED to
stderr, consider it a crash.
(setUpWindowsCrashLogSaving): Added. Saves the current post-mortem
debugger, then sets ntsd as the post-mortem debugger. ntsd is
configured to save a crash log and then quit automatically.
(END): Added. Restores the previous post-mortem debugger when the
script exits.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65537 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Sheriff Bot <webkit.review.bot@gmail.com>
commit-queue@webkit.org [Tue, 17 Aug 2010 20:17:09 +0000 (20:17 +0000)]
2010-08-17  Sheriff Bot  <webkit.review.bot@gmail.com>

        Unreviewed, rolling out r65516.
        http://trac.webkit.org/changeset/65516
        https://bugs.webkit.org/show_bug.cgi?id=44126

        broke chromium unit tests on mac and linux (Requested by
        johnny_g on #webkit).

        * WebKit.gyp:
        * src/WebFrameImpl.cpp:
        (WebKit::frameContentAsPlainText):
        * tests/RunAllTests.cpp:
        (main):
        * tests/WebFrameTest.cpp: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65536 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 John Gregg <johnnyg@google.com>
johnnyg@google.com [Tue, 17 Aug 2010 20:14:46 +0000 (20:14 +0000)]
2010-08-17  John Gregg  <johnnyg@google.com>

        Unreviewed, test expectation.

        Excluding the new geolocation multiple-requests layout tests from chromium.
        See https://bugs.webkit.org/show_bug.cgi?id=44118

        * platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65535 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9 years ago2010-08-17 Darin Fisher <darin@chromium.org>
darin@chromium.org [Tue, 17 Aug 2010 20:07:45 +0000 (20:07 +0000)]
2010-08-17  Darin Fisher  <darin@chromium.org>

        Reviewed by Darin Adler.

        [Chromium] REGRESSION: Incorrect repaint on scrolling with position:fixed elements
        https://bugs.webkit.org/show_bug.cgi?id=42949

        This is a partial revert of http://trac.webkit.org/changeset/63907

        * page/FrameView.cpp:
        (WebCore::FrameView::scrollPositionChangedViaPlatformWidget):
        (WebCore::FrameView::scrollPositionChanged):
        * platform/ScrollView.cpp:
        (WebCore::ScrollView::valueChanged):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@65534 268f45cc-cd09-0410-ab3c-d52691b4dbfc