Context Navigation

← Previous Period
Next Period →

Timeline

Dec 1, 2012:

11:55 PM Changeset in webkit [136317] by timothy_horton@apple.com

2 edits in trunk/Source/WebKit2

PDFPlugin: Support PDF form editing in <iframe>s
https://bugs.webkit.org/show_bug.cgi?id=103287
<rdar://problem/10326917>

Reviewed by Dan Bernstein.

PDFPlugin form editing depends on having a Document that has the same
size and position as the plugin; this is true not only in the full-main-frame
case, but also in the <iframe> case. Therefore, we can enable form editing
for <iframe> PDFPlugins.

WebProcess/Plugins/PDF/PDFPlugin.mm:

(WebKit::PDFPlugin::supportsForms): Support form editing whenever
our Document's widget is our PluginView (which will only be true in the
full-main-frame or <iframe> case).

11:51 PM Changeset in webkit [136316] by timothy_horton@apple.com

3 edits in trunk/Source/WebKit2

PDFPlugin: <embed> and <object> PDFs affect their parent frame's page scale
https://bugs.webkit.org/show_bug.cgi?id=103286
<rdar://problem/12752315>

Reviewed by Dan Bernstein.

Detecting whether or not a PDFPlugin is a full-page plugin or not is not
as simple as checking isMainFrame(), because <embed> and <object> will
cause PDFPlugin to live in the main frame, while still not being full-page.
We also have to check that the main frame's document is a PluginDocument,
and whether that PluginDocument's PluginWidget is our PDFPlugin's PluginView.

WebProcess/Plugins/PDF/PDFPlugin.h:

(PDFPlugin): Add isFullFramePlugin.

WebProcess/Plugins/PDF/PDFPlugin.mm:

(WebKit::PDFPlugin::isFullFramePlugin): Check whether the main frame is
backed by a PluginDocument, and that that PluginDocument corresponds
to this PDFPlugin.
(WebKit::PDFPlugin::handlesPageScaleFactor): Use isFullFramePlugin.
(WebKit::PDFPlugin::supportsForms): Use isFullFramePlugin.

11:44 PM Changeset in webkit [136315] by timothy_horton@apple.com

2 edits in trunk/Source/WebKit2

[wk2] Use spoolRect instead of snapshots for print preview
https://bugs.webkit.org/show_bug.cgi?id=103798
<rdar://problem/12686749>

Reviewed by Sam Weinig.

Revert to using spoolRect instead of snapshotting for printing; this was a bogus
change I made in http://trac.webkit.org/changeset/133935 which was irrelevant to
the core of that change, and incorrectly disregards the Frame parameter passed to
beginPrinting, instead using the WebPage's current Frame.

WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::drawRectToImage):

11:06 PM Changeset in webkit [136314] by zandobersek@gmail.com

4 edits in trunk/Tools

[XvfbDriver] XvfbDriverTest.test_next_free_display is flaky on builders using XvfbDriver
https://bugs.webkit.org/show_bug.cgi?id=103806

Reviewed by Dirk Pranke.

There are occasional mishaps on builders where XvfbDriver doesn't properly
clean up the Xvfb instance it was managing so the related guard lock files
are still existing when the webkitpy unit tests are run. This can cause failures
in XvfbDriverTest.test_next_free_display as the test tries to acquire displays
with a specific number but is unable to due to the stale locks.

The solution to this is to use MockFileLock when inside unit test. This is achievable
by creating a file lock through the SystemHost instance connected to the Port object,
the latter available to the Driver.

Scripts/webkitpy/common/system/file_lock_mock.py:

(MockFileLock.acquire_lock): Return True to signal lock acquiring succeeded.
(MockFileLock.release_lock): Return True to signal lock releasing succeeded.

Scripts/webkitpy/layout_tests/port/xvfbdriver.py:

(XvfbDriver._next_free_display): Create the file lock through a mockable way.

Scripts/webkitpy/layout_tests/port/xvfbdriver_unittest.py:

(XvfbDriverTest.test_next_free_display): Enable back the unit test.

10:35 PM WebInspector edited by mousewiki@yahoo.com

(diff)

4:31 PM Changeset in webkit [136313] by abarth@webkit.org

3 edits in trunk/Source/WebCore

Unreviewed, rolling out r136015.
http://trac.webkit.org/changeset/136015
https://bugs.webkit.org/show_bug.cgi?id=103821

Did not heal performance bot and caused other performance
regressions (Requested by abarth on #webkit).

Patch by Sheriff Bot <webkit.review.bot@gmail.com> on 2012-12-01

bindings/scripts/CodeGeneratorV8.pm:

(GenerateHeader):

2:34 PM EFLWebKit edited by laszlo.gombos@gmail.com

fix the spelling for cmakeargs - see … (diff)

1:33 PM Changeset in webkit [136312] by mkwst@chromium.org

2 edits in trunk/LayoutTests

[chromium] Unreviewed gardening. Skipping crashing tests.

The DEPS roll containing r170422 is causing three crashes cross-platform
in the Chromium port. Skipping them here, filed crbug.com/163785 to fix.

platform/chromium/TestExpectations:

Marked media/encrypted-media/encrypted-media-events.html,
media/encrypted-media/encrypted-media-not-loaded.html, and
media/encrypted-media/encrypted-media-syntax.html as crashing.

12:00 PM Changeset in webkit [136311] by mkwst@chromium.org

2 edits in trunk/LayoutTests

[chromium] Unreviewed gardening. Skipping calendar-picker-datetime-with-step.html.

It looks like platform/chromium/fast/forms/calendar-picker/calendar-picker-datetime-with-step.html
started failing at midnight PST, December 1st. I don't think it handled
the month change correctly. Filed https://bugs.webkit.org/show_bug.cgi?id=103817
to cover the fix.

platform/chromium/TestExpectations:

10:43 AM Changeset in webkit [136310] by kov@webkit.org

4 edits in releases/WebKitGTK/webkit-1.10/Source

Merge 133059 - Enable webkit2 introspection
https://bugs.webkit.org/show_bug.cgi?id=94313

Patch by Jesse van den Kieboom <jessevdk@gnome.org> on 2012-10-31
Reviewed by Martin Robinson.

This patch enables generation of introspection information for webkit2
in the same was as it is done for webkit. The resulting gir and
typelib file are named WebKit2-3.0.{gir,typelib}.

GNUmakefile.am:
UIProcess/API/gtk/WebKitBackForwardList.cpp:

4:39 AM Changeset in webkit [136309] by commit-queue@webkit.org

2 edits in trunk/Source/WebCore

DRT - crashed in WebCore::SearchFieldCancelButtonElement::defaultEventHandler
https://bugs.webkit.org/show_bug.cgi?id=103592

Patch by Xiaobo Wang <xiaobo.wang@torchmobile.com.cn> on 2012-12-01
Reviewed by Kent Tamura.

The crash happened when the search field cancel button handles mouseout
event after the search input was detached. When it happens the input
element returned from shadowHost() is null, need to check the pointer
before dereferencing.
InputFieldSpeechButton and SearchFieldResultsButtonElement have the
similar issue.

No new tests. The patch fixes the following test on BlackBerry platform.
Test: fast/forms/search-delete-while-cancel-button-clicked.html

html/shadow/TextControlInnerElements.cpp:

(WebCore::SearchFieldResultsButtonElement::defaultEventHandler):
(WebCore::SearchFieldCancelButtonElement::defaultEventHandler):
(WebCore::SearchFieldCancelButtonElement::willRespondToMouseClickEvents):
(WebCore::InputFieldSpeechButtonElement::defaultEventHandler):
(WebCore::InputFieldSpeechButtonElement::willRespondToMouseClickEvents):
(WebCore::InputFieldSpeechButtonElement::setRecognitionResult):

3:09 AM Changeset in webkit [136308] by zandobersek@gmail.com

2 edits in trunk/Tools

[GTK] A11y tests in WK2 report badly-formatted AX roles
https://bugs.webkit.org/show_bug.cgi?id=102947

Reviewed by Xan Lopez.

As in DRT, convert the ATK accessibility roles to corresponding AXRoles.
These are expected in both tests and their baselines.

WebKitTestRunner/InjectedBundle/atk/AccessibilityUIElementAtk.cpp:

(WTR::roleToString):
(WTR):
(WTR::AccessibilityUIElement::role):

12:33 AM Changeset in webkit [136307] by mkwst@chromium.org

2 edits in trunk/LayoutTests

[gtk] Unreviewed gardening. Skipping CSP 1.1 tests.

The CSP 1.1 tests are failing under GTK after r136305 made the CSP_NEXT
flag runtime enabled. Will fix in webkit.org/b/103810, skipping for now.

platform/gtk/TestExpectations:

Skipping CSP 1.1 tests.

Nov 30, 2012:

11:09 PM Changeset in webkit [136306] by rniwa@webkit.org

2 edits
1 move in trunk/LayoutTests

Shorten a file name as the test renamed in r136122 is still hosing Chromium Win perf. bot.

http/tests/inspector/network/resources/script-as-text-with-a-very-very-very-very-very-very-very-very-very-very-very-very-very-very-long-url.php: Renamed.
http/tests/inspector/network/resources/script-as-text-with-a-very-very-very-very-very-very-very-very-very-very-very-very-very-very-very-very-very-very-very-long-url.php: Removed.
http/tests/inspector/network/script-as-text-loading-long-url.html:

10:53 PM Changeset in webkit [136305] by mkwst@chromium.org

11 edits in trunk

CSP 1.1: Make the CSP_NEXT flag runtime enabled.
https://bugs.webkit.org/show_bug.cgi?id=103652

Reviewed by Adam Barth.

Source/WebCore:

Content Security Policy 1.1 continues to live behind the CSP_NEXT flag,
this patch adds another layer on top of that in order to enable runtime
decisions about whether it should be active.

bindings/generic/RuntimeEnabledFeatures.cpp:

(WebCore):

bindings/generic/RuntimeEnabledFeatures.h:

(RuntimeEnabledFeatures):
(WebCore::RuntimeEnabledFeatures::experimentalContentSecurityPolicyFeaturesEnabled):
(WebCore::RuntimeEnabledFeatures::setExperimentalContentSecurityPolicyFeaturesEnabled):

Adds methods in order to correctly handle enabling and disabling
CSP 1.1 features.

dom/Document.idl:

Gate the 'document.securityPolicy' object on the runtime flag.

page/ContentSecurityPolicy.cpp:

(WebCore::CSPDirectiveList::addDirective):

Check that experimental features are runtime enabled before
processing 1.1 directives.

(WebCore::ContentSecurityPolicy::experimentalFeaturesEnabled):
(WebCore):

page/ContentSecurityPolicy.h:

Adds a new method which checks against the runtime flag to determine
whether CSP 1.1 features are enabled.

Source/WebKit/chromium:

The CSP_NEXT flag continues to be enabled on the Chromium port, but this
patch now locks the features away behind the securityPolicy runtime
flag.

public/WebRuntimeFeatures.h:

(WebRuntimeFeatures):

src/WebRuntimeFeatures.cpp:

(WebKit::WebRuntimeFeatures::enableExperimentalContentSecurityPolicyFeatures):
(WebKit):
(WebKit::WebRuntimeFeatures::isExperimentalContentSecurityPolicyFeaturesEnabled):

Adds the feature to WebRuntimeFeatures so it can be toggled from
inside Chromium.

Tools:

Ensures that the new SecurityPolicy runtime flag is enabled for Chromium's tests.

DumpRenderTree/chromium/TestShell.cpp:

(TestShell::TestShell):

10:18 PM Changeset in webkit [136304] by ryuan.choi@samsung.com

2 edits in trunk/LayoutTests

[EFL][WK1] Gardening test cases which are not crashed after r135662
https://bugs.webkit.org/show_bug.cgi?id=103803

Unreviewed, EFL gardening.

Patch by Sujin Park <sujjin.park@gmail.com> on 2012-11-30

platform/efl-wk1/TestExpectations:

10:12 PM Changeset in webkit [136303] by Simon Fraser

2 edits in trunk/Source/WebCore

Avoid calling into NSUserDefaults every time we start a scroll
https://bugs.webkit.org/show_bug.cgi?id=103804

Reviewed by Dan Bernstein.

Cache the NSScrollAnimationEnabled preference to avoid
hitting NSUserDefaults every time we start scrolling.

platform/mac/ScrollAnimatorMac.mm:

(WebCore::scrollAnimationEnabledForSystem):

6:24 PM Changeset in webkit [136302] by andersca@apple.com

2 edits in trunk/Source/WebKit2

Always listen for the DidFirstLayout layout milestone
https://bugs.webkit.org/show_bug.cgi?id=103795
<rdar://problem/12790268>

Reviewed by Dan Bernstein.

The DidFirstLayout layout milestone is used to determine when to unfreeze the layer tree, so we always want to listen for it.

WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::WebPage):

5:40 PM Changeset in webkit [136301] by roger_fong@apple.com

2 edits in trunk/Tools

Unreviewed. Removed temporary debugging change from DRT (r136300).

DumpRenderTree/win/DumpRenderTree.cpp:

(dump):

5:23 PM Changeset in webkit [136300] by roger_fong@apple.com

2 edits in trunk/Tools

Unreviewed. Temporary change to DRT for debugging purposes.

DumpRenderTree/win/DumpRenderTree.cpp:

(dump):

5:19 PM Changeset in webkit [136299] by pierre.rossi@gmail.com

4 edits in trunk/Source

[Qt] Unreviewed speculative Mac build fix after r136232

Source/JavaScriptCore:

Update the include path so that LLIntAssembly.h is picked up.
The bot didn't break until later when a clean build was triggered.

JavaScriptCore.pri:

Source/WebKit2:

cleanup an oversight from r136235

PluginProcess.pro: s/webkit/webkitwidgets/

4:53 PM Changeset in webkit [136298] by Beth Dakin

2 edits in trunk/Source/WebCore

https://bugs.webkit.org/show_bug.cgi?id=103790
The LayerPool m_pruneTimer should have a longer delay

Reviewed by Simon Fraser.

A 1 second delay is more appropriate here.

platform/graphics/ca/mac/LayerPool.mm:

(WebCore::LayerPool::schedulePrune):

4:50 PM Changeset in webkit [136297] by oliver@apple.com

2 edits in trunk/Source/JavaScriptCore

Optimise more cases of op_typeof
https://bugs.webkit.org/show_bug.cgi?id=103783

Reviewed by Mark Hahnenberg.

Increase our coverage of typeof based typechecks by
making sure that the codegenerators always uses
consistent operand ordering when feeding typeof operations
into equality operations.

bytecompiler/NodesCodegen.cpp:

(JSC::BinaryOpNode::emitBytecode):
(JSC::EqualNode::emitBytecode):
(JSC::StrictEqualNode::emitBytecode):

4:43 PM Changeset in webkit [136296] by jer.noble@apple.com

2 edits in trunk/Source/WebCore

Unreviewed Windows build fix.

Add a default: entry to an case statement to fix a build error when compiled in VS2005.

html/MediaController.cpp:

(MediaController::playbackState):

4:13 PM Changeset in webkit [136295] by jer.noble@apple.com

7 edits
1 copy
1 add in trunk

Add support for the 'unpause()' method on MediaController.
https://bugs.webkit.org/show_bug.cgi?id=103751

Reviewed by Eric Carlson.

Source/WebCore:

The MediaController spec has been updated to add an unpause() method
which is functionally identical to the old play() method, and changes
the semantics of play() to call play() on the underlying slaved media
elements. The spec also added a playbackState() method, returning a
string representing the last calculated playback state.

Test: media/media-controller-unpause.html

html/MediaController.cpp:

(MediaController::unpause): Renamed from play().
(MediaController::play): Added. Same as unpause(), but also call

play() on slaved media elements.

(playbackStateWaiting): Static string accessor.
(playbackStatePlaying): Ditto.
(playbackStateEnded): Ditto.
(MediaController::playbackState): Added. Returns string value of

m_playbackState.

html/MediaController.h:
html/MediaController.idl:

LayoutTests:

Add new test for unpause() and playbackState() methods.

media/media-controller-unpause.html:
media/media-controller-unpause-expected.txt:
media/media-controller-playback.html:
media/media-controller-playback-expected.txt:

3:59 PM Changeset in webkit [136294] by jchaffraix@webkit.org

5 edits in trunk/Source/WebCore

[CSS Grid Layout] Align the grid track code with the specification's production rules
https://bugs.webkit.org/show_bug.cgi?id=103703

Reviewed by Alexis Menard.

In order to support more of the syntax and make the code more readable, this change
splits following the grammar's production rules (see
http://dev.w3.org/csswg/css3-grid-layout/#grid-definition-columns-and-grid-definit).

Refactoring covered by existing tests.

css/CSSComputedStyleDeclaration.cpp:

(WebCore::valueForGridTrackMinMax):
(WebCore::valueForGridTrackGroup):
(WebCore::valueForGridTrackList):

css/CSSParser.cpp:

(WebCore::CSSParser::parseGridTrackList):
(WebCore::CSSParser::parseGridTrackGroup):
(WebCore::CSSParser::parseGridTrackMinMax):

css/CSSParser.h:
css/StyleResolver.cpp:

(WebCore::createGridTrackMinMax):
(WebCore::createGridTrackGroup):
(WebCore::createGridTrackList):
Mechanical split of the parsing, resolving and computed style code into
<track-list>, <track-group>, <track-minmax> and <track-breadth> parts.

css/CSSParser.cpp:

(WebCore::CSSParser::parseGridBreadth):
Changed this function to just parse a single value (without any extra logic).

css/StyleResolver.cpp:

(WebCore::createGridTrackBreadth):
Changed this function to take a Length as it is used to parse a <track-breadth>.
The goal is for it to be the low-level function createGridTrackMinmax will call on
each function parameter.

3:56 PM Changeset in webkit [136293] by Simon Fraser

2 edits in trunk/Source/WebCore

Fix potential crash when canceling animations on renderers with no node
https://bugs.webkit.org/show_bug.cgi?id=103784

Reviewed by Dan Bernstein.

We could crash when tearing down RenderScrollbars because we hit
AnimationController::cancelAnimations() for a renderer with no node.
The assertion already check for a null node, but the line after
does not, so fix that.

I tried to make a test with animations/transitions on a custom scrollbar, but was not able
to get it to crash.

page/animation/AnimationController.cpp:

(WebCore::AnimationController::cancelAnimations):

3:36 PM Changeset in webkit [136292] by bfulgham@webkit.org

6 edits
1 move
10 adds
1 delete in trunk/Source

Source/WebKit/win: [Windows, WinCairo] Generate library export definition file.
https://bugs.webkit.org/show_bug.cgi?id=103687

Reviewed by Tim Horton.

Update project solutions to use the new generated *.def files.

WebKit.vcproj/WebKit.sln: Update for new project and file

locations.

Source/WebKit2: [Windows, WinCairo] Generate library export symbol file.
https://bugs.webkit.org/show_bug.cgi?id=103687

Reviewed by Tim Horton.

Generate the export definition file based on the compile options
and port settings. Get rid of the old hand-written file, and use
a revised master input file that allows using conditional
settings to generate the output, much like the Mac and other
ports do.

win/WebKit2.def: Removed.
win/WebKit2.def.in: Added.
win/WebKit2Apple.vsprops:
win/WebKit2CFLite.def: Removed.
win/WebKit2CFLite.vsprops:
win/WebKit2Common.vsprops:
win/WebKit2ExportGenerator.vcproj: Added.
win/WebKit2ExportGeneratorCommon.vsprops: Added.
win/WebKit2ExportGeneratorDebug.vsprops: Added.
win/WebKit2ExportGeneratorDebugAll.vsprops: Added.
win/WebKit2ExportGeneratorDebugCairoCFLite.vsprops: Added.
win/WebKit2ExportGeneratorPostBuild.cmd: Added.
win/WebKit2ExportGeneratorPreBuild.cmd: Added.
win/WebKit2ExportGeneratorProduction.vsprops: Added.
win/WebKit2ExportGeneratorRelease.vsprops: Added.
win/WebKit2ExportGeneratorReleaseCairoCFLite.vsprops: Added.

3:33 PM Changeset in webkit [136291] by cevans@google.com

1 edit in branches/chromium/1312/Source/WebCore/platform/graphics/chromium/FontPlatformDataChromiumWin.h

Merge 134902
BUG=157875
Review URL: https://codereview.chromium.org/11419266

3:33 PM Changeset in webkit [136290] by cevans@google.com

9 edits in branches/chromium/1312/Source/WebCore/platform/graphics

Merge 134871
BUG=157875
Review URL: https://codereview.chromium.org/11348334

3:30 PM Changeset in webkit [136289] by timothy_horton@apple.com

2 edits in trunk/Source/WebCore

PluginDocument fires didFinishDocumentLoadForFrame upon receiving initial bytes instead of when load completes
https://bugs.webkit.org/show_bug.cgi?id=103702
<rdar://problem/12762534>

Reviewed by Alexey Proskuryakov.

Remove the call to finish() from PluginDocumentParser::appendBytes().

This was introduced in http://trac.webkit.org/changeset/14838, when finish() wouldn't get
called for PluginDocuments any other way. In the time since, DocumentWriter::end() was made
to call finish() anyway, at the correct time (the call from appendBytes makes a PluginDocument
appear to be finished loading after the first bytes are received).

No new tests, because there doesn't seem to be any state in a PluginDocument that can be accessed
from JavaScript to determine whether or not it has finished loading.

html/PluginDocument.cpp:

(WebCore::PluginDocumentParser::appendBytes):

3:21 PM Changeset in webkit [136288] by loislo@chromium.org

3 edits in trunk/Source/WebCore

RenderBlock minor clean-up: replace raw pointers with OwnPtrs.
https://bugs.webkit.org/show_bug.cgi?id=103690

Reviewed by Eric Seidel.

raw pointers were replaced with OwnPtrs.
Probably we also gain a small performance improvements because
I replaced get && remove pair with take.

rendering/RenderBlock.cpp:

(WebCore):
(WebCore::removeBlockFromDescendantAndContainerMaps):
(WebCore::RenderBlock::~RenderBlock):
(WebCore::RenderBlock::addOverflowFromFloats):
(WebCore::RenderBlock::addContinuationWithOutline):
(WebCore::RenderBlock::paintContinuationOutlines):
(WebCore::RenderBlock::insertIntoTrackedRendererMaps):
(WebCore::RenderBlock::removeFromTrackedRendererMaps):
(WebCore::RenderBlock::setDesiredColumnCountAndWidth):

rendering/RenderBlock.h:

(WebCore):

3:21 PM Changeset in webkit [136287] by zmo@google.com

2 edits in trunk/LayoutTests

Unreviewed, webkit gardening, fix webkit lint.

platform/chromium/TestExpectations:

3:17 PM Changeset in webkit [136286] by cevans@google.com

3 edits
2 copies in branches/chromium/1312

Merge 136060
BUG=158693
Review URL: https://codereview.chromium.org/11416300

3:12 PM Changeset in webkit [136285] by cevans@google.com

1 edit
2 copies in branches/chromium/1312

Merge 135914
BUG=158695
Review URL: https://codereview.chromium.org/11415245

3:12 PM Changeset in webkit [136284] by Simon Fraser

2 edits in trunk/LayoutTests

Use the correct path to a test in TestExpectations.

platform/chromium/TestExpectations:

3:07 PM Changeset in webkit [136283] by cevans@google.com

1 edit in branches/chromium/1312/Source/WebCore/page/EventHandler.cpp

Merge 136062
BUG=159219
Review URL: https://codereview.chromium.org/11415242

3:04 PM Changeset in webkit [136282] by commit-queue@webkit.org

11 edits in trunk/Source/WebCore

WebGL: Add a class to abstract the status of the Image in texImage2D() and texSubImage2D()
https://bugs.webkit.org/show_bug.cgi?id=103606

Patch by Jun Jiang <jun.a.jiang@intel.com> on 2012-11-30
Reviewed by Kenneth Russell.

In texImage2D() and texSubImage2D() for WebGL, the status of the Image is extracted and kept in the function GraphicsContext3D::getImageData() but provides no interface or guaranteed way to use the status of the Image outside GraphicsContext3D::getImageData() safely. For example, you can not get the address of the raw Image data and operate it outside the scope of the GraphicsContext3D::getImageData() and there is at least one memory copy existed from the Image to the intermediate vector.

This patch refactors the code by adding a ImageExtractor class to abstract and keep the status of the Image. The lifetime and validity of Image status are determined by the lifetime of the object instead of the scope of getImageData(). It provides flexibility on future optimizations.

No new tests. It is a code refactoring patch.

html/canvas/WebGLRenderingContext.cpp:

(WebCore):
(WebCore::WebGLRenderingContext::texImage2DImpl):
(WebCore::WebGLRenderingContext::texSubImage2DImpl):

platform/graphics/GraphicsContext3D.cpp:

(WebCore::GraphicsContext3D::ImageExtractor::ImageExtractor):
(WebCore):
(WebCore::GraphicsContext3D::packImageData):

platform/graphics/GraphicsContext3D.h:

(WebCore):
(GraphicsContext3D):
(ImageExtractor):
(WebCore::GraphicsContext3D::ImageExtractor::extractSucceeded):
(WebCore::GraphicsContext3D::ImageExtractor::imagePixelData):
(WebCore::GraphicsContext3D::ImageExtractor::imageWidth):
(WebCore::GraphicsContext3D::ImageExtractor::imageHeight):
(WebCore::GraphicsContext3D::ImageExtractor::imageSourceFormat):
(WebCore::GraphicsContext3D::ImageExtractor::imageAlphaOp):
(WebCore::GraphicsContext3D::ImageExtractor::imageSourceUnpackAlignment):

platform/graphics/NativeImagePtr.h:
platform/graphics/cairo/GraphicsContext3DCairo.cpp: