⚠ Archived content — this site is no longer maintained.   Current WebKit documentation is at docs.webkit.org.

Timeline



Jul 29, 2013:

11:42 PM Changeset in webkit [153460] by Carlos Garcia Campos
  • 13 edits in trunk/Source

Unreviewed. Fix make distcheck.

Source/JavaScriptCore:

  • GNUmakefile.list.am: Add missing files to compilation.
  • bytecode/CodeBlock.cpp: Add a ENABLE(FTL_JIT) #if block to

include FTL header files not included in the compilation.

  • dfg/DFGDriver.cpp: Ditto.
  • dfg/DFGPlan.cpp: Ditto.

Source/ThirdParty/ANGLE:

  • GNUmakefile.am: Add missing header files to compilation.

Source/WebCore:

  • GNUmakefile.list.am: Add missing header file to compilation.

Source/WebKit2:

  • GNUmakefile.list.am: Add missing header file to compilation.

Source/WTF:

  • GNUmakefile.list.am: Add missing files to compilation.
9:51 PM Changeset in webkit [153459] by Lucas Forschler
  • 8 edits in branches/safari-537-branch/Source/WebKit2

Merged r153458. <rdar://problem/14568841>

9:45 PM Changeset in webkit [153458] by timothy_horton@apple.com
  • 8 edits in trunk/Source/WebKit2

[wk2] Flush the WebProcess’ implicit transaction when using endDeferringViewInWindowChangesSync
​https://bugs.webkit.org/show_bug.cgi?id=119225
<rdar://problem/14568841>

Reviewed by Simon Fraser.

Tell CoreAnimation to flush the implicit transaction before replying
when using endDeferringViewInWindowChangesSync, as that method's contract
is that the WebProcess is totally ready to be in-window when it returns.

  • UIProcess/API/mac/WKView.mm:

(-[WKView endDeferringViewInWindowChanges]):
Adopt viewInWindowStateDidChange.

(-[WKView endDeferringViewInWindowChangesSync]):
Adopt viewInWindowStateDidChange, asking it to send a reply only if we're going to wait for one.

  • UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::viewInWindowStateDidChange):
(WebKit::WebPageProxy::viewStateDidChange):
Pull viewInWindowStateDidChange out of viewStateDidChange.
Request a reply from SetIsInWindow if we're told to.

  • UIProcess/WebPageProxy.h: Add WantsReplyOrNot and viewInWindowStateDidChange.
  • WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::didUpdateInWindowStateTimerFired):
Don't build this version on Mac, we'll have a WebPageMac version.

(WebKit::WebPage::setIsInWindow):
Only start the timer to send the didUpdateInWindowState reply if we're asked to.

  • WebProcess/WebPage/WebPage.h:
  • WebProcess/WebPage/WebPage.messages.in:

Add an argument to the SetIsInWindow message for whether the WebProcess
should inform the UIProcess when SetIsInWindow completes or not.

  • WebProcess/WebPage/mac/WebPageMac.mm:

(WebKit::WebPage::didUpdateInWindowStateTimerFired):
Tell CA to flush the implicit transaction before telling the UIProcess that
we're finished moving in-window.

9:33 PM Changeset in webkit [153457] by commit-queue@webkit.org
  • 14 edits
    3 adds in trunk

Eager stack trace for error objects.
​https://bugs.webkit.org/show_bug.cgi?id=118918

Source/JavaScriptCore:

Patch by Chris Curtis <​chris_curtis@apple.com> on 2013-07-29
Reviewed by Geoffrey Garen.

Chrome and Firefox give error objects the stack property and we wanted to match
that functionality. This allows developers to see the stack without throwing an object.

  • runtime/ErrorInstance.cpp:

(JSC::ErrorInstance::finishCreation):

For error objects that are not thrown as an exception, we pass the stackTrace in
as a parameter. This allows the error object to have the stack property.

  • interpreter/Interpreter.cpp:

(JSC::stackTraceAsString):
Helper function used to eliminate duplicate code.

(JSC::Interpreter::addStackTraceIfNecessary):
When an error object is created by the user the vm->exceptionStack is not set.
If the user throws this error object later the stack that is in the error object
may not be the correct stack for the throw, so when we set the vm->exception stack,
the stack property on the error object is set as well.

  • runtime/ErrorConstructor.cpp:

(JSC::constructWithErrorConstructor):
(JSC::callErrorConstructor):

  • runtime/NativeErrorConstructor.cpp:

(JSC::constructWithNativeErrorConstructor):
(JSC::callNativeErrorConstructor):
These functions indicate that the user created an error object. For all error objects
that the user explicitly creates, the topCallFrame is at a new frame created to
handle the user's call. In this case though, the error object needs the caller's
frame to create the stack trace correctly.

  • interpreter/Interpreter.h:
  • runtime/ErrorInstance.h:

(JSC::ErrorInstance::create):

LayoutTests:

Patch by Chris Curtis <​chris_curtis@apple.com> on 2013-07-29
Reviewed by Geoffrey Garen.

Added tests to ensure that the stack property was present at creation for all
error Objects. This test will fail without this patch.

  • fast/js/script-tests/stack-at-creation-for-error-objects.js: Added.

(checkStack):

  • fast/js/stack-at-creation-for-error-objects-expected.txt: Added.
  • fast/js/stack-at-creation-for-error-objects.html: Added.
  • inspector/console/console-format-expected.txt:
  • inspector/console/console-format.html:

This test was modified by removing the error object from being evaluated. Prior to this patch
error objects did not have the stack property, so the stack information was not being
displayed. The stack trace includes a file path specific to the machine that is running
the test. The results would have differed from one computer to the next. There
is not an easy way to capture the error object to treat it differently. By removing
the error object there is no need to add extra code to treat it differently.
Also there are other tests inside inspector/console that test the stack trace,
so the testing suite does not lose error testing by removing it.

The .stack property was added to the error objects at creation time.

  • fast/js/exception-properties-expected.txt:
  • fast/js/script-tests/exception-properties.js:

The column numbers are modified in the following test. When error objects are explicitly
invoked, the column number points to the beginning "(" instead of end ")".
Functionality between browsers do not match either. Firefox does not output column
numbers. Chrome points columns numbers to the beginning of the "new" call.

  • fast/js/line-column-numbers-expected.txt:
  • fast/js/stack-trace-expected.txt:
9:14 PM Changeset in webkit [153456] by kseo@webkit.org
  • 2 edits in trunk/Source/WebCore

[Gtk] Improve string use in PasteboardGtk
​https://bugs.webkit.org/show_bug.cgi?id=119204

Reviewed by Martin Robinson.

Use emptyString() instead of "" and use ASCIILiteral where appropriate.

No behavior change, no new tests needed.

  • platform/gtk/PasteboardGtk.cpp:

(WebCore::Pasteboard::documentFragment):
Use emptyString() instead of "".
(WebCore::Pasteboard::types):
Use ASCIILiteral where appropriate.

8:33 PM Changeset in webkit [153455] by mrowe@apple.com
  • 2 edits in trunk/Source/WTF

<rdar://problem/14528244> Tons of FastMalloc leaks reported by leaks of objects that have already been deallocated

Reviewed by Sam Weinig.

  • wtf/FastMalloc.cpp:

(WTF::TCMalloc_Central_FreeList::enumerateFreeObjects): Mark objects in the transfer cache as being free.

6:15 PM Changeset in webkit [153454] by barraclough@apple.com
  • 8 edits in trunk/Source/JavaScriptCore

Some cleanup in PropertySlot
​https://bugs.webkit.org/show_bug.cgi?id=119189

Reviewed by Geoff Garen.

PropertySlot represents a property in one of four states - value, getter, custom, or custom-index.
The state is currently tracked redundantly by two mechanisms - the custom getter function (m_getValue)
is set to a special value to indicate the type (other than custom), and the type is also tracked by
an enum - but only if cacheable. Cacheability can typically be determined by the value of m_offset
(this is invalidOffset if not cacheable).

  • Internally, always track the type of the property using an enum value, PropertyType.
  • Use m_offset to indicate cacheable.
  • Keep the external interface (CachedPropertyType) unchanged.
  • Better pack data into the m_data union.

Performance neutral.

  • dfg/DFGRepatch.cpp:

(JSC::DFG::tryCacheGetByID):
(JSC::DFG::tryBuildGetByIDList):

  • cachedPropertyType() -> isCacheable*()
  • jit/JITPropertyAccess.cpp:

(JSC::JIT::privateCompileGetByIdProto):
(JSC::JIT::privateCompileGetByIdSelfList):
(JSC::JIT::privateCompileGetByIdProtoList):
(JSC::JIT::privateCompileGetByIdChainList):
(JSC::JIT::privateCompileGetByIdChain):

  • cachedPropertyType() -> isCacheable*()
  • jit/JITPropertyAccess32_64.cpp:

(JSC::JIT::privateCompileGetByIdProto):
(JSC::JIT::privateCompileGetByIdSelfList):
(JSC::JIT::privateCompileGetByIdProtoList):
(JSC::JIT::privateCompileGetByIdChainList):
(JSC::JIT::privateCompileGetByIdChain):

  • cachedPropertyType() -> isCacheable*()
  • jit/JITStubs.cpp:

(JSC::tryCacheGetByID):

  • cachedPropertyType() -> isCacheable*()
  • llint/LLIntSlowPaths.cpp:

(JSC::LLInt::LLINT_SLOW_PATH_DECL):

  • cachedPropertyType() -> isCacheable*()
  • runtime/PropertySlot.cpp:

(JSC::PropertySlot::functionGetter):

  • refactoring described above.
  • runtime/PropertySlot.h:

(JSC::PropertySlot::PropertySlot):
(JSC::PropertySlot::getValue):
(JSC::PropertySlot::isCacheable):
(JSC::PropertySlot::isCacheableValue):
(JSC::PropertySlot::isCacheableGetter):
(JSC::PropertySlot::isCacheableCustom):
(JSC::PropertySlot::cachedOffset):
(JSC::PropertySlot::customGetter):
(JSC::PropertySlot::setValue):
(JSC::PropertySlot::setCustom):
(JSC::PropertySlot::setCacheableCustom):
(JSC::PropertySlot::setCustomIndex):
(JSC::PropertySlot::setGetterSlot):
(JSC::PropertySlot::setCacheableGetterSlot):
(JSC::PropertySlot::setUndefined):
(JSC::PropertySlot::slotBase):
(JSC::PropertySlot::setBase):

  • refactoring described above.
6:10 PM Changeset in webkit [153453] by Lucas Forschler
  • 3 edits in branches/safari-537-branch/Source/WebKit2

Merged r153449. <rdar://problem/14440207>

6:07 PM Changeset in webkit [153452] by Lucas Forschler
  • 4 edits in branches/safari-537-branch/Source/WebCore

Merged r153447. <rdar://problem/14556632>

6:06 PM Changeset in webkit [153451] by Simon Fraser
  • 2 edits in trunk/Source/WebCore

Page count can be incorrect if there is a pending style recalc
​https://bugs.webkit.org/show_bug.cgi?id=119232

Reviewed by Beth Dakin.

If style is changed in a way that schedules a style recalc on the Document,
and then Page::pageCount() is called, we may give the wrong answer because
the existing code only checked FrameView::needsLayout(), not whether a style
recale was pending.

Fix by having Page::pageCount() call updateLayoutIgnorePendingStylesheets()
as we do for other properties that require layer, and are exposed to JS
or via API.

  • page/Page.cpp:

(WebCore::Page::pageCount):

5:25 PM Changeset in webkit [153450] by achristensen@apple.com
  • 5 edits
    2 moves in trunk/Source/WebCore

Moved GLContext.cpp and .h out of cairo directory to be used by non-cairo ports.
​https://bugs.webkit.org/show_bug.cgi?id=119223

Reviewed by Kenneth Rohde Christiansen.

  • GNUmakefile.list.am:
  • PlatformGTK.cmake:
  • WebCore.vcxproj/WebCore.vcxproj:
  • WebCore.vcxproj/WebCore.vcxproj.filters:

Changed location of GLContext.cpp and GLContext.h.

  • platform/graphics/GLContext.cpp: Copied from WebCore/platform/graphics/cairo/GLContext.cpp.
  • platform/graphics/GLContext.h: Copied from WebCore/platform/graphics/cairo/GLContext.h.
  • platform/graphics/cairo/GLContext.cpp: Removed.
  • platform/graphics/cairo/GLContext.h: Removed.
5:20 PM Changeset in webkit [153449] by timothy_horton@apple.com
  • 3 edits in trunk/Source/WebKit2

Null deref under PluginView::handlesPageScaleFactor()
​https://bugs.webkit.org/show_bug.cgi?id=119231
<rdar://problem/14440207>

Reviewed by Simon Fraser.

  • WebProcess/Plugins/PluginView.cpp:

(WebKit::PluginView::handlesPageScaleFactor):

  • WebProcess/Plugins/PluginView.h:

Null-check m_plugin and check m_isInitialized.
Make pageScaleFactor() and handlesPageScaleFactor const.

5:15 PM Changeset in webkit [153448] by jer.noble@apple.com
  • 3 edits in trunk/Source/WebCore

TextTrackRepresentation captions often disappear then appear again.
​https://bugs.webkit.org/show_bug.cgi?id=119228

Reviewed by Simon Fraser.

Calling updateTextTrackDisplay() will cause the entire DOM subtree containing the active cues
to be torn down and rebuilt. Only call into this method when text tracks have actually changed.

Add an enum parameter to configureTextTrackDisplay() which allows the caller to specify that
method should assume that the list of visible text tracks has changed, forcing a call to
updateTextTrackDisplay().

  • html/HTMLMediaElement.cpp:

(WebCore::HTMLMediaElement::updateActiveTextTrackCues): Remove the unnecessary call to updateTextTrackDisplay().
(WebCore::HTMLMediaElement::textTrackModeChanged): Pass AssumeVisibleTextTracksChanged to configureTextTrackDisplay().
(WebCore::HTMLMediaElement::configureTextTrackDisplay): Check the passed flags and exit early only if

AssumeVisibleTextTracksChanged is not set.

  • html/HTMLMediaElement.h:
5:12 PM Changeset in webkit [153447] by rniwa@webkit.org
  • 4 edits in trunk/Source/WebCore

Fix crash due to unexpected Node deletion during MutationObserver registration book-keeping
​https://bugs.webkit.org/show_bug.cgi?id=119124

Reviewed by Sam Weinig.

Merge ​https://chromium.googlesource.com/chromium/blink/+/b6afb927695b3acf2c75c25f05e99682660993e2

No new tests since I could not reproduce the crash with the test attached in the Blink change.

The bug was caused by Node::unregisterMutationObserver removing the MutationObserverRegistration
that holds the last ref to the node. Avoid that by explicitly allocating a local RefPtr to the node
in MutationObserverRegistration::unregister. Also rename it to unregisterAndDelete to clarify
the semantics and make it a static member function to be even safer.

  • dom/MutationObserver.cpp:

(WebCore::MutationObserver::disconnect):

  • dom/MutationObserverRegistration.cpp:

(WebCore::MutationObserverRegistration::unregisterAndDelete):

  • dom/MutationObserverRegistration.h:
4:57 PM Changeset in webkit [153446] by Lucas Forschler
  • 8 edits in branches/safari-537-branch/Source/JavaScriptCore

Merged r153071. <rdar://problem/14583789>

4:31 PM Changeset in webkit [153445] by Lucas Forschler
  • 2 edits in branches/safari-537-branch/Source/WebCore

Merged r153418. <rdar://problem/14572855>

4:00 PM Changeset in webkit [153444] by commit-queue@webkit.org
  • 2 edits in trunk/Tools

Unreviewed, rolling out r152860.
​http://trac.webkit.org/changeset/152860
​https://bugs.webkit.org/show_bug.cgi?id=119227

Disable WinEWS tests until tests stop crashing (Requested by
rfong on #webkit).

  • Scripts/webkitpy/common/config/ews.json:
3:40 PM Changeset in webkit [153443] by achristensen@apple.com
  • 2 edits
    2 moves in trunk/Source/WebCore

Moved GLTransportSurface.h and .cpp out of efl-specific directory.
​https://bugs.webkit.org/show_bug.cgi?id=119163

Reviewed by Alexis Menard.

  • PlatformEfl.cmake: Changed directory of GLTransportSurface.cpp.
  • platform/graphics/surfaces/GLTransportSurface.cpp: Copied from WebCore/platform/graphics/surfaces/efl/GLTransportSurface.cpp.
  • platform/graphics/surfaces/GLTransportSurface.h: Copied from WebCore/platform/graphics/surfaces/efl/GLTransportSurface.h.
  • platform/graphics/surfaces/efl/GLTransportSurface.cpp: Removed.
  • platform/graphics/surfaces/efl/GLTransportSurface.h: Removed.
2:55 PM Changeset in webkit [153442] by Lucas Forschler
  • 3 edits in branches/safari-537-branch/Source/JavaScriptCore

Merge patch from <rdar://problem/14579343>

2:25 PM Changeset in webkit [153441] by achristensen@apple.com
  • 3 edits
    1 add in trunk/Source/WebCore

Speculative implementation of missing GraphicsContext3D methods on Windows.
​https://bugs.webkit.org/show_bug.cgi?id=119167

Reviewed by Dean Jackson.

  • WebCore.vcxproj/WebCore.vcxproj:
  • WebCore.vcxproj/WebCore.vcxproj.filters:
  • platform/graphics/win/GraphicsContext3DWin.cpp: Added.

(WebCore::GraphicsContext3DPrivate::GraphicsContext3DPrivate):
(WebCore::GraphicsContext3DPrivate::~GraphicsContext3DPrivate):
(WebCore::GraphicsContext3D::create):
(WebCore::GraphicsContext3D::GraphicsContext3D):
(WebCore::GraphicsContext3D::~GraphicsContext3D):
(WebCore::GraphicsContext3D::setContextLostCallback):
(WebCore::GraphicsContext3D::setErrorMessageCallback):
(WebCore::GraphicsContext3D::makeContextCurrent):
(WebCore::GraphicsContext3D::platformGraphicsContext3D):
(WebCore::GraphicsContext3D::platformTexture):
(WebCore::GraphicsContext3D::isGLES2Compliant):
(WebCore::GraphicsContext3D::platformLayer):

1:10 PM Changeset in webkit [153440] by zandobersek@gmail.com
  • 2 edits in trunk/Source/WTF

Reintroduce convenience CheckedInt*, CheckedUint* types
​https://bugs.webkit.org/show_bug.cgi?id=119213

Reviewed by Oliver Hunt.

Reintroduce CheckedInt* and CheckedUint* types that were previously provided for convenience
through the CheckedInt header that was removed in r153095. The types are now based on the
Checked class, using the RecordOverflow class as the overflow handler.

  • wtf/CheckedArithmetic.h:
1:09 PM Changeset in webkit [153439] by zandobersek@gmail.com
  • 4 edits
    1 add in trunk/Tools

[GTK] Add WestonDriver, use it when appropriate
​https://bugs.webkit.org/show_bug.cgi?id=119100

Reviewed by Dirk Pranke.

Introduce the WestonDriver, a Driver-implementing class that sets up an instance of Weston, the Wayland compositor,
and sets up an environment that enables the test runner to run inside it.

The design very much mimics the XvfbDriver class. It is currently used by the GTK port when the WAYLAND_DISPLAY
environment variable is set, that is when NRWT is being run under a Wayland display.

  • Scripts/webkitpy/port/base.py:

(Port.to.setup_environ_for_server): Also copy the XDG_RUNTIME_DIR env from the original environment for Linux-using ports.

  • Scripts/webkitpy/port/gtk.py:

(GtkPort):
(GtkPort._driver_class): Make the method memoized - the display type is not going to change during the testing, so we
should only check once for the presence of the WAYLAND_DISPLAY env and determine the proper driver class to use based on that.
(GtkPort.check_sys_deps): Call the static check_driver method on the class as returned by the _driver_class method
instead of hard-coding the XvfbDriver class in the check.

  • Scripts/webkitpy/port/westondriver.py: Added.

(WestonDriver):
(WestonDriver.check_driver): Similarly to XvfbDriver, this checks for existance of the weston executable.
(WestonDriver.init): Again, like in XvfbDriver, we use a small delay between launching the Weston display and launching the driver.
(WestonDriver._start): Sets up the driver-specific temporary directory, launches the Weston display, sets up the environment for the
driver and then launches the driver in said environment, using the driver's Weston display.
(WestonDriver.stop): Cleans up the Weston instance and the temporary directory.
(WestonDriver._ensure_driver_tmpdir_subdirectory): A helper function that creates a subdirectory in the driver's temporary directory
and returns the absolute path to it.

  • Scripts/webkitpy/port/xvfbdriver.py:

(XvfbDriver.check_driver): Renamed from check_xvfb.
(XvfbDriver._start): Access the FileSystem object through the Port's Host object, instead of the Port's private member.

12:25 PM Changeset in webkit [153438] by dino@apple.com
  • 2 edits in trunk/Source/WebCore

Max IOSurfaceDimension should be an inclusive upper bound
​https://bugs.webkit.org/show_bug.cgi?id=119206
<rdar://problem/14555311>

Reviewed by Simon Fraser.

The maxIOSurfaceDimension value should be an inclusive upper bound.
Change the greater than or equal to into a greater than. This will
allow canvas elements of this dimension to be accelerated.

  • platform/graphics/cg/ImageBufferCG.cpp:

(WebCore::ImageBuffer::ImageBuffer): Change >= to >

12:20 PM Changeset in webkit [153437] by commit-queue@webkit.org
  • 4 edits
    6 adds
    4 deletes in trunk/LayoutTests

[CSS Shapes] New positioning model: update percentage rectangle and inset rectangle tests
​https://bugs.webkit.org/show_bug.cgi?id=118086

Patch by Bem Jones-Bey <Bem Jones-Bey> on 2013-07-29
Reviewed by Dirk Schulze.

Update percentage based tests to work with the new positioning model.
Also move the tests to be W3C spec tests, so they have been impored
back here from the CSSWG repo.

  • TestExpectations: Remove skipping for replaced tests.
  • csswg/submitted/shapes/shape-outside/shape-outside-floats-horizontal-rectangle-003-expected.html: Added.
  • csswg/submitted/shapes/shape-outside/shape-outside-floats-horizontal-rectangle-003.html: Added.
  • csswg/submitted/shapes/shape-outside/shape-outside-floats-horizontal-rectangle-004-expected.html: Added.
  • csswg/submitted/shapes/shape-outside/shape-outside-floats-horizontal-rectangle-004.html: Added.
  • csswg/submitted/shapes/shape-outside/shape-outside-floats-inset-rectangle-002-expected.html: Added.
  • csswg/submitted/shapes/shape-outside/shape-outside-floats-inset-rectangle-002.html: Added.
  • csswg/submitted/shapes/shape-outside/w3c-import.log:
  • fast/shapes/shape-outside-floats/shape-outside-floats-simple-inset-rectangle-percentage-expected.html: Removed.
  • fast/shapes/shape-outside-floats/shape-outside-floats-simple-inset-rectangle-percentage.html: Removed.
  • fast/shapes/shape-outside-floats/shape-outside-floats-simple-rectangle-percentage-expected.html: Removed.
  • fast/shapes/shape-outside-floats/shape-outside-floats-simple-rectangle-percentage.html: Removed.
11:51 AM Changeset in webkit [153436] by achristensen@apple.com
  • 7 edits
    8 adds
    2 deletes in trunk/Source

Source/ThirdParty/ANGLE: Separated translator into preprocessor, translator_common,
translator_hlsl (for libGLESv2), and translator_glsl
(for ANGLEWebKitBridge).
​https://bugs.webkit.org/show_bug.cgi?id=119221

Reviewed by Brent Fulgham.

  • ANGLE.vcxproj/ANGLEGenerated.vcxproj:

Removed unused preprocessor definitions.

  • ANGLE.vcxproj/libEGL.vcxproj:
  • ANGLE.vcxproj/libGLESv2.vcxproj:

Removed dependency on removed translator project.

  • ANGLE.vcxproj/libGLESv2Common.props:

Link to translator_hlsl libraries.

  • ANGLE.vcxproj/preprocessor.vcxproj: Added.
  • ANGLE.vcxproj/preprocessor.vcxproj.filters: Added.
  • ANGLE.vcxproj/translator.vcxproj: Removed.
  • ANGLE.vcxproj/translator.vcxproj.filters: Removed.
  • ANGLE.vcxproj/translator_common.vcxproj: Added.
  • ANGLE.vcxproj/translator_common.vcxproj.filters: Added.
  • ANGLE.vcxproj/translator_glsl.vcxproj: Added.
  • ANGLE.vcxproj/translator_glsl.vcxproj.filters: Added.
  • ANGLE.vcxproj/translator_hlsl.vcxproj: Added.
  • ANGLE.vcxproj/translator_hlsl.vcxproj.filters: Added.

Source/WebKit: Separated translator into preprocessor, translator_common,
translator_hlsl (for libGLESv2), and translator_glsl (for ANGLEWebKitBridge).
These projects are still not built by default.
​https://bugs.webkit.org/show_bug.cgi?id=119221

Reviewed by Brent Fulgham.

  • WebKit.vcxproj/WebKit.sln:

Added separate ANGLE translator projects.

11:26 AM Changeset in webkit [153435] by mario@webkit.org
  • 6 edits
    3 adds in trunk

[ATK] Implement attributesOfChildren() for AccessibilityUIElement
​https://bugs.webkit.org/show_bug.cgi?id=118967

Reviewed by Gustavo Noronha Silva.

Tools:

Implemented AccessibilityUIElement::attributesOfChildren() both
in DumpRenderTree and the WebKitTestRunner.

  • DumpRenderTree/atk/AccessibilityUIElementAtk.cpp:

(AccessibilityUIElement::attributesOfChildren): Implemented.

  • WebKitTestRunner/InjectedBundle/atk/AccessibilityUIElementAtk.cpp:

(WTR::AccessibilityUIElement::attributesOfChildren): Implemented.

LayoutTests:

Removed fixed tests from TestExpectations and added/updated expected
results for tests that are now dumping more useful information.

  • platform/gtk/TestExpectations: Removed three fixed tests.
  • platform/gtk/accessibility/image-map2-expected.txt: Updated.
  • platform/gtk/accessibility/lists-expected.txt: Added.
  • platform/gtk/accessibility/plugin-expected.txt: Added.
  • platform/gtk/accessibility/table-with-aria-role-expected.txt: Added.
11:10 AM Changeset in webkit [153434] by andersca@apple.com
  • 5 edits in trunk/Source/WebKit2

Remove setUnderlayPage() and associated code
​https://bugs.webkit.org/show_bug.cgi?id=119220
<rdar://problem/14392426>

Reviewed by Jessie Berlin.

This is dead code, get rid of it.

  • WebProcess/InjectedBundle/API/c/WKBundlePage.cpp:
  • WebProcess/InjectedBundle/API/c/WKBundlePage.h:
  • WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::close):
(WebKit::WebPage::layoutIfNeeded):
(WebKit::WebPage::drawRect):

  • WebProcess/WebPage/WebPage.h:
10:45 AM Changeset in webkit [153433] by reni@webkit.org
  • 3 edits
    2 adds in trunk

ASSERT_NOT_REACHED was touched in WebCore::SVGAnimatedType::valueAsString
​https://bugs.webkit.org/show_bug.cgi?id=118744

Reviewed by Philip Rogers.

Source/WebCore:

If we have an <animateTransform> tag that sets the attributeName and attributeType
properties however they are not consistent and we shouldn't try to apply it.
It is already checked and catched in SVGAnimateElement::resetAnimatedType() but
DontApplyAnimation case isn't handled properly. The patch forces an early return
in this case.

Test: svg/animations/attributeNameAndAttributeTypeMissmatch.svg

  • svg/SVGAnimateElement.cpp:

(WebCore::SVGAnimateElement::resetAnimatedType):

LayoutTests:

Adding test to check whether attributeName and attributeType are consistent in <animateTransform>.

  • platform/qt/svg/animations/attributeNameAndAttributeTypeMissmatch-expected.txt: Added.
  • svg/animations/attributeNameAndAttributeTypeMissmatch.svg: Added.
10:16 AM Changeset in webkit [153432] by mario@webkit.org
  • 8 edits
    4 adds in trunk

[ATK] Implement allAttributes() for AccessibilityUIElement
​https://bugs.webkit.org/show_bug.cgi?id=118966

Reviewed by Gustavo Noronha Silva.

Tools:

Implemented AccessibilityUIElement::allAttributes() both in
DumpRenderTree and the WebKitTestRunner.

  • DumpRenderTree/atk/AccessibilityUIElementAtk.cpp:

(getAtkAttributeSetAsString): Moved up in the file, and refactored
to use StringBuilder and return a String instead of a gchar*.
(roleToString): Changed not to return the "AXRole: " prefix, so
we can use it as well from attributesOfElement().
(attributesOfElement): New function to dump attributes for an object.
(AccessibilityUIElement::allAttributes): Call attributesOfElement().
(AccessibilityUIElement::role): Updated to prefix "AXRole: ".

  • WebKitTestRunner/InjectedBundle/atk/AccessibilityUIElementAtk.cpp:

(getAtkAttributeSetAsString): Refactored to use StringBuilder and return
a String instead of a gchar*.

(WTF::roleToString): Changed not to return the "AXRole: " prefix,
so we can use it as well from attributesOfElement().
(WTF::attributesOfElement): New function to dump attributes for an object.
(WTR::AccessibilityUIElement::allAttributes): Call attributesOfElement().
(WTR::AccessibilityUIElement::role): Updated to prefix "AXRole: ".

LayoutTests:

Removed fixed tests from TestExpectations and added/updated expected
results for tests that are now dumping more useful information.

  • platform/gtk/TestExpectations: Removed four fixed tests.
  • platform/gtk/accessibility/image-link-expected.txt: Updated.
  • platform/gtk/accessibility/table-cell-spans-expected.txt: Updated
  • platform/gtk/accessibility/table-cells-expected.txt: Updated
  • platform/gtk/accessibility/table-detection-expected.txt: Added.
  • platform/gtk/accessibility/table-one-cell-expected.txt: Added.
  • platform/gtk/accessibility/table-with-rules-expected.txt: Added.
  • platform/gtk/accessibility/transformed-element-expected.txt: Added.
10:10 AM Changeset in webkit [153431] by allan.jensen@digia.com
  • 2 edits in trunk/Source/WebCore

REGRESSION(148300) GIFs not reanimated after resumeActiveDOMObjectAndAnimations
​https://bugs.webkit.org/show_bug.cgi?id=119217

Reviewed by Tim Horton.

Commit 148300 introduced a new better was of restarting GIF animations after
they have been suspended. Unfortunately the method was not activated when a page
is resumed but only when it is returning from background.

This fixes GIF animations that are stopped during user interaction on Qt and other
platforms using suspendActiveDOMObjectAndAnimations.

  • page/Frame.cpp:

(WebCore::Frame::resumeActiveDOMObjectsAndAnimations):

9:35 AM Changeset in webkit [153430] by dbates@webkit.org
  • 2 edits in trunk/Tools

Add support infrastructure for iOS WebKit port
​https://bugs.webkit.org/show_bug.cgi?id=119168

Reviewed by Ryosuke Niwa.

Implement support infrastructure towards building iOS WebKit.

  • Scripts/webkitdirs.pm:

(xcodeSDKPlatformName): Added.
(determineConfigurationProductDir):
(builtDylibPathForName):
(willUseIOSDeviceSDKWhenBuilding): Added.
(willUseIOSSimulatorSDKWhenBuilding): Added.
(isIOSWebKit): Added.
(copyInspectorFrontendFiles):

9:34 AM Changeset in webkit [153429] by achristensen@apple.com
  • 4 edits in trunk/Source/WebCore

Made EGL useable by non-cairo ports.
​https://bugs.webkit.org/show_bug.cgi?id=119177

Reviewed by Brent Fulgham.

  • platform/graphics/cairo/GLContext.h:
  • platform/graphics/egl/GLContextEGL.cpp:

(WebCore::GLContextEGL::GLContextEGL):

  • platform/graphics/egl/GLContextEGL.h:

Protected Cairo-specific code with #if USE(CAIRO)

9:25 AM Changeset in webkit [153428] by commit-queue@webkit.org
  • 9 edits
    2 adds in trunk

CoordinatedGraphics: Add API to get and set the active state of a WebView
​https://bugs.webkit.org/show_bug.cgi?id=119067

Patch by Sergio Correia <Sergio Correia> on 2013-07-29
Reviewed by Noam Rosenthal.

Source/WebCore:

Added a method to return whether a scene is active or not.

No new tests, no behavior change.

  • platform/graphics/texmap/coordinated/CoordinatedGraphicsScene.h:

(WebCore::CoordinatedGraphicsScene::isActive): Added method to return the
active state of the scene.

Source/WebKit2:

A view being in the "active" state generally means that it is not in the
"background", although this definition can vary depending on the platform
under consideration.

This patch adds the following APIs to WKView:

  • bool WKViewIsActive(WKViewRef)

This API accepts a WKViewRef as parameter and returns a boolean indicating
whether the given view is active.

  • void WKViewSetIsActive(WKViewRef, bool)

This API accepts a WKViewRef and a boolean as parameteres, and it sets the
active state of the given view to the given boolean argument.

  • UIProcess/API/C/CoordinatedGraphics/WKView.cpp:

(WKViewIsActive): API added.
(WKViewSetIsActive): Ditto.

  • UIProcess/API/C/CoordinatedGraphics/WKView.h: Export aforementioned API.
  • UIProcess/CoordinatedGraphics/WebView.cpp:

(WebKit::WebView::initialize): Use new setActive method.
(WebKit::WebView::setActive): Added method to set the active state of the
view.
(WebKit::WebView::isActive): Added method to return the active state of
view.
(WebKit::WebView::enterAcceleratedCompositingMode): Use new setActive
method.
(WebKit::WebView::exitAcceleratedCompositingMode): Ditto.

  • UIProcess/CoordinatedGraphics/WebView.h: Added declarations of isActive

and setActive.

Tools:

Added API tests for WKViewIsActive and WKViewSetIsActive.

  • TestWebKitAPI/PlatformEfl.cmake:
  • TestWebKitAPI/Tests/WebKit2/CoordinatedGraphics/WKViewIsActiveSetIsActive.cpp: Added.

(TestWebKitAPI::didFinishLoadForFrame):
(TestWebKitAPI::webProcessCrashed):
(TestWebKitAPI::webProcessRelaunched):
(TestWebKitAPI::TEST):

  • TestWebKitAPI/Tests/WebKit2/CoordinatedGraphics/WKViewIsActiveSetIsActive_Bundle.cpp:

Bundle added to allow crashing the WebProcess and testing the active state of a view
under such condition.
(TestWebKitAPI::WKViewIsActiveSetIsActiveTest::WKViewIsActiveSetIsActiveTest):
(TestWebKitAPI::WKViewIsActiveSetIsActiveTest::didReceiveMessage):

8:46 AM Changeset in webkit [153427] by mario@webkit.org
  • 4 edits in trunk/Source

[ATK] Issues with edge cases when getting offsets for a text range in AtkText
​https://bugs.webkit.org/show_bug.cgi?id=118908

Reviewed by Martin Robinson.

Source/WebCore:

Reimplement getSelectionOffsetsForObject() just in term of
Positions instead of using ranges, which makes it simpler and
works better. Also, make sure we use the right Node as reference
for the accessibility object by getting the proper one both for
text control objects (e.g. input, text area) and normal ones.

  • accessibility/atk/WebKitAccessibleInterfaceText.cpp:

(getNodeForAccessibilityObject): New helper function to ensure we
always get the relevant object from the DOM tree for a given
accessibility object, in the context of the implementation of
AtkText, so it works both with Text Controls and other elements.
(getSelectionOffsetsForObject): Rewritten this function in terms
of VisiblePositions and avoiding weird operations with ranges, so
we have more control to fine tune the results and give more
accurate ones. Besides, now it works better with edge cases.

Source/WebKit/gtk:

  • tests/testatk.c:

(runGetTextTests): Updated unit tests to check more cases of
calling the atk_text_get_text_*_offset() functions for WORD.

8:36 AM Changeset in webkit [153426] by commit-queue@webkit.org
  • 3 edits in trunk/Source/WebKit2

Remove duplicated LayerTreeHost::supportsAcceleratedCompositing
​https://bugs.webkit.org/show_bug.cgi?id=118949

Patch by Jae Hyun Park <​jae.park@company100.net> on 2013-07-29
Reviewed by Noam Rosenthal.

  • WebProcess/WebPage/CoordinatedGraphics/CoordinatedLayerTreeHost.cpp:
  • WebProcess/WebPage/LayerTreeHost.h:

(WebKit::LayerTreeHost::supportsAcceleratedCompositing):

8:30 AM Changeset in webkit [153425] by Antoine Quint
  • 2 edits in trunk/LayoutTests

fast/spatial-navigation test fail on mac-wk2
​https://bugs.webkit.org/show_bug.cgi?id=96438

This test now passes on Mac so we don't need to skip it any longer.

Reviewed by Antonio Gomes.

  • platform/mac/TestExpectations:
7:37 AM Changeset in webkit [153424] by zandobersek@gmail.com
  • 5 edits in trunk/Source/WebCore

Clean up DragImage
​https://bugs.webkit.org/show_bug.cgi?id=118981

Reviewed by Anders Carlsson.

  • dom/Clipboard.h: Add forward declaration for the Range class. This was not necessary until now as the forward

declaration was introduced in the DragImage header, from where it's being removed due to not being required in that class.

  • page/Frame.h: Ditto.
  • platform/DragImage.cpp: Remove the unnecessary DragController and FontRenderingMode header inclusions.

The latter is already included in the header.

  • platform/DragImage.h: Remove the forward declaration for the Range class - it's not used by the DragImage class at all.
6:50 AM Changeset in webkit [153423] by kov@webkit.org
  • 2 edits in trunk/Source/WebCore/platform/gtk/po

Updated Telugu [te] Translations for WebKitGTK+ HEAD
​https://bugs.webkit.org/show_bug.cgi?id=117501

Patch by Krishnababu Krothapalli <​k.meetme@gmail.com> on 2013-07-29
Reviewed by Gustavo Noronha.

  • te.po: Updated.
6:34 AM Changeset in webkit [153422] by kov@webkit.org
  • 2 edits in trunk/Source/WebCore/platform/gtk/po

[l10n] [mr] Updated WebKitGTK+ Translation(s) in Marathi [mr] language
​https://bugs.webkit.org/show_bug.cgi?id=119001

Patch by Sandeep Shedmake <​sandeep.shedmake@gmail.com> on 2013-07-29
Reviewed by Gustavo Noronha.

  • mr.po: Updated.
6:23 AM Changeset in webkit [153421] by zandobersek@gmail.com
  • 6 edits in trunk

[Automake] Remove references to non-existent *_CFLAGS and *_LIBS variables
​https://bugs.webkit.org/show_bug.cgi?id=119212

Reviewed by Carlos Garcia Campos.

Source/WebKit2:

  • GNUmakefile.am: Remove references to GLOBALDEPS_(CFLAGS|LIBS) and CLUTTER_(CFLAGS|LIBS)

variables that are not defined during the configuration process.

Tools:

  • GNUmakefile.am: Remove references to GLOBALDEPS_(CFLAGS|LIBS) and CLUTTER_(CFLAGS|LIBS)

variables that are not defined during the configuration process.

  • TestWebKitAPI/GNUmakefile.am: Ditto.
  • WebKitTestRunner/GNUmakefile.am: Ditto.
12:45 AM Changeset in webkit [153420] by kseo@webkit.org
  • 2 edits in trunk/Source/WebCore

[EFL] Improve string use in ErrorsEfl
​https://bugs.webkit.org/show_bug.cgi?id=119207

Reviewed by Christophe Dumez.

Use ASCIILiteral for constant strings becoming WTFStrings.

No behavior change.

  • platform/efl/ErrorsEfl.cpp:

(WebCore::cancelledError):
(WebCore::blockedError):
(WebCore::cannotShowURLError):
(WebCore::interruptedForPolicyChangeError):
(WebCore::cannotShowMIMETypeError):
(WebCore::fileDoesNotExistError):
(WebCore::pluginWillHandleLoadError):
(WebCore::downloadCancelledByUserError):
(WebCore::printerNotFoundError):
(WebCore::invalidPageRangeToPrint):

12:17 AM Changeset in webkit [153419] by g.czajkowski@samsung.com
  • 2 edits in trunk/Source/WebKit2

[EFL][WK2] Simplify context_menu_spelling_items_availability unit test
​https://bugs.webkit.org/show_bug.cgi?id=119085

Reviewed by Gyuyoung Kim.

Use findContextMenuItem method to check whether the item
appears in context menu.

  • UIProcess/API/efl/tests/test_ewk2_text_checker.cpp:

(EWK2TextCheckerTest::findContextMenuItem):
Do not report failure inside the method.
It doesn't allow to test the negative test cases.

(EWK2TextCheckerTest::checkCorrectnessOfSpellingItems):
Use findContextMenuItem method.

(EWK2TextCheckerTest::toogleCheckSpellingWhileTyping):
(EWK2TextCheckerTest::selectLearnSpelling):
(EWK2TextCheckerTest::selectIgnoreSpelling):
Verify the return value of findContextMenuItem.

(TEST_F):
Reset wasContextMenuShown variable, it's needed after r152153.
Use waitUntilTrue method.

Jul 28, 2013:

11:23 PM Changeset in webkit [153418] by akling@apple.com
  • 2 edits in trunk/Source/WebCore

Don't update the text track override CSS every time a media element is created.
<​http://webkit.org/b/119199>
<rdar://problem/14572855>

Reviewed by Darin Adler.

We should only have to update the CSS if the accessibility caption preferences are changed.
This was forcing a full style recalc in every Document any time a media element is instantiated.

  • page/CaptionUserPreferencesMediaAF.cpp:

(WebCore::CaptionUserPreferencesMediaAF::setInterestedInCaptionPreferenceChanges):

9:37 PM Changeset in webkit [153417] by Lucas Forschler
  • 4 edits
    2 copies in branches/safari-537-branch

Merged r153399. <rdar://problem/14549536>

9:27 PM Changeset in webkit [153416] by Lucas Forschler
  • 8 edits in branches/safari-537-branch/Source/WebKit2

Merged r153404. <rdar://problem/14444331>

9:22 PM Changeset in webkit [153415] by Lucas Forschler
  • 2 edits in branches/safari-537-branch/Source/WebKit2

Merged r153402. <rdar://problem/14517068>

8:08 PM Changeset in webkit [153414] by aestes@apple.com
  • 2 edits in trunk/Tools

Include <stdio.h> to declare perror()
​https://bugs.webkit.org/show_bug.cgi?id=119198

Reviewed by Andreas Kling.

Silences an implicit declaration warning.

  • DumpRenderTree/mac/PerlSupport/DumpRenderTreeSupport.c:
6:16 PM Changeset in webkit [153413] by aestes@apple.com
  • 2 edits in trunk/Source/WebCore

Stop exporting Widget::frameRectsChanged()
​https://bugs.webkit.org/show_bug.cgi?id=119196

Reviewed by Darin Adler.

The definition is inlined in a private header, so there's no need to
export a symbol to link against. On the Mac, since we compile with
-fvisibility-inlines-hidden, removing the symbol from the export file
silences an ld warning about exporting a hidden symbol.

  • WebCore.exp.in: Removed ZN7WebCore6Widget17frameRectsChangedEv.
4:36 PM Changeset in webkit [153412] by weinig@apple.com
  • 2 edits in trunk/Source/WebCore

Clean up CSSPrimitiveValue::equals a bit
​https://bugs.webkit.org/show_bug.cgi?id=119195

Reviewed by Andreas Kling.

  • css/CSSPrimitiveValue.cpp:

(WebCore::CSSPrimitiveValue::equals):
This should not change behavior, but is quite a bit clearer.

3:08 PM Changeset in webkit [153411] by Antoine Quint
  • 2 edits in trunk/LayoutTests

fast/forms/input-in-table-cell-no-value.html fails on platform without ENABLE(INPUT_TYPE_DATE)
​https://bugs.webkit.org/show_bug.cgi?id=90987

This test now passes on Mac so we don't need to skip it any longer.

Reviewed by Anders Carlsson.

  • platform/mac/TestExpectations:
11:04 AM Changeset in webkit [153410] by oliver@apple.com
  • 3 edits in trunk/Source/JavaScriptCore

REGRESSION: Crash when opening Facebook.com
​https://bugs.webkit.org/show_bug.cgi?id=119155

Reviewed by Andreas Kling.

Scope nodes are always objects, so we should be using SpecObjectOther
rather than SpecCellOther. Marking Scopes as CellOther leads to a
contradiction in the CFA, resulting in bogus codegen.

  • dfg/DFGAbstractInterpreterInlines.h:

(JSC::DFG::::executeEffects):

  • dfg/DFGPredictionPropagationPhase.cpp:

(JSC::DFG::PredictionPropagationPhase::propagate):

Note: See TracTimeline for information about the timeline view.