Timeline



May 27, 2013:

11:57 PM Changeset in webkit [150788] by Patrick Gansterer
  • 3 edits in trunk/Source/WebCore

Build fix for WinCE after r150772.

  • platform/win/ClipboardWin.cpp:

(WebCore):

  • platform/win/PasteboardWin.cpp:

(WebCore::Pasteboard::readFilenames):
(WebCore):

11:48 PM Changeset in webkit [150787] by gyuyoung.kim@samsung.com
  • 3 edits in trunk/Source/WebKit2

[EFL][WK2] Remove unneeded WebCore namespace use in ewk
https://bugs.webkit.org/show_bug.cgi?id=116707

Reviewed by Christophe Dumez.

  • UIProcess/API/efl/ewk_security_origin.cpp: Remove "using namespace WebCore".
  • UIProcess/API/efl/ewk_text_checker.cpp: ditto.
11:31 PM Changeset in webkit [150786] by Patrick Gansterer
  • 2 edits in trunk/Source/WTF

Use correct stack size on Solaris and OpenBSD
https://bugs.webkit.org/show_bug.cgi?id=114978

Recommit after wrong rollout in r150621.

  • wtf/StackBounds.cpp:

(WTF):
(WTF::StackBounds::initialize):

5:02 PM Changeset in webkit [150785] by jonlee@apple.com
  • 14 edits
    1 add
    1 delete in trunk

[WK2] Notifications clobber each other with multiple processes
https://bugs.webkit.org/show_bug.cgi?id=116428
<rdar://problem/13935191>

Reviewed by Darin Adler.

.:

  • ManualTests/notification-in-multiple-windows.html: Added.

Source/WebKit2:

With multiple processes, the notification IDs, when passed up to the UI process, can clobber
each other. To fix this, we need to maintain a global map of notification IDs. This map is
keyed by its own unique notification ID, and maps to a pair containing the web page ID and that
web page's ID for the notification.

Now that we maintain groups of notifications based on the web page, we no longer send IPC messages
from WebNotificationManager to WebNotificationManagerProxy; instead we send messages to the
WebPageProxy. This removes the need for WebNotificationManagerProxy to be a message receiver.

When a page closes, all of the web notifications are cleared out. However, by the time the
WebPage::close() is called, the connection between WebPage and WebPageProxy is destroyed. Since
the WebPage is told to close from the UI process anyway, we clear out the notifications separately,
instead of waiting for a message from the WebPage.

  • UIProcess/Notifications/WebNotificationManagerProxy.h: Update to take into account the

notification's web page. Remove inheritance of CoreIPC::MessageReceiver. Expose the original message
handlers as public functions, since they will be called from WebPageProxy. Add a new map that
associates a global ID with a notification ID that came from a web page.

There are now two flavors of clearNotifications(). One clears out all notifications associated

with a web page. This is called when the page is closed. The other clears out a subset of
notifications associated with a web page. This is called when notifications associated with a sub-frame
is closed.

  • UIProcess/Notifications/WebNotificationManagerProxy.messages.in: Removed. All messages from

the web process go to WebPageProxy now.

  • UIProcess/Notifications/WebNotificationManagerProxy.cpp: Update to take into account the

notification's web page.

(WebKit::generateGlobalNotificationID): The manager proxy now maintains its own global notification
ID generator.
(WebKit::WebNotificationManagerProxy::WebNotificationManagerProxy): The proxy is no longer a
message receiver. Remove code that registers it as such.

(WebKit::WebNotificationManagerProxy::show): Refactor to differentiate between the notification ID
that came from the web process, and the global notification ID the proxy maintains. Add the mapping
from the global ID to the (web page ID, notification ID) pair.
(WebKit::WebNotificationManagerProxy::cancel): Refactor to take into consideration the web page.
(WebKit::WebNotificationManagerProxy::didDestroyNotification): Refactor to take into consideration
the web page. Fixes a leak where we did not remove the item from the maps. This function is called
from the web process, when the ScriptExecutionContext is destroyed, so we remove it from our maps
before we pass the message along to the provider.

Helper functions that evaluate when a given notification in the map matches the desired parameters.
(WebKit::pageIDsMatch): The notification is associated with the provided page.
(WebKit::pageAndNotificationIDsMatch): The notification is associated with the provided page and is
contained within the list of provided notifications.

(WebKit::WebNotificationManagerProxy::clearNotifications): Changed to only remove notifications
associated with the provided web page, and could include a specific list of notifications. This latter
situation occurs if notifications were associated with an iframe, and that iframe was removed.
There is an O(n) walk that could be make more efficient using another hash map, but that's overhead
for a map that should be small in size anyway.

(WebKit::WebNotificationManagerProxy::providerDidShowNotification): Refactor to take into
consideration the web page.
(WebKit::WebNotificationManagerProxy::providerDidClickNotification): Refactor to take into
consideration the web page.
(WebKit::WebNotificationManagerProxy::providerDidCloseNotifications): Now we need to comb through
the list of global IDs and put them in buckets based on the notification's web pages. After that
is done we can send the DidCloseNotifications() to those pages' processes. There is a possible
extra optimization here where we group based on the page's process instead, to reduce the number
of messages sent to processes.

  • UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::close): When a web page is closed, we clear the notifications associated
with the page.
(WebKit::WebPageProxy::cancelNotification): Forward call to WebNotificationManagerProxy.
(WebKit::WebPageProxy::clearNotifications): Ditto.
(WebKit::WebPageProxy::didDestroyNotification): Ditto.

  • UIProcess/WebPageProxy.h:
  • UIProcess/WebPageProxy.messages.in:
  • WebProcess/Notifications/NotificationPermissionRequestManager.cpp:
  • WebProcess/Notifications/WebNotificationManager.cpp:

(WebKit::WebNotificationManager::cancel):
(WebKit::WebNotificationManager::clearNotifications):
(WebKit::WebNotificationManager::didDestroyNotification):

  • WebProcess/Notifications/NotificationPermissionRequestManager.cpp: Remove extraneous include.
  • CMakeLists.txt: Remove WebNotificationManagerProxy.messages.in and related files.
  • DerivedSources.pri: Ditto.
  • DerivedSources.make: Ditto.
  • GNUmakefile.list.am: Ditto.
  • WebKit2.xcodeproj/project.pbxproj: Ditto.
4:55 PM Changeset in webkit [150784] by Darin Adler
  • 2 edits in trunk/Source/WebCore

Fix typo that broke the build.

4:53 PM Changeset in webkit [150783] by Darin Adler
  • 4 edits in trunk/Source/WebCore

Move hasNonEmptyBoundingBox from Node to HTMLAnchorElement
https://bugs.webkit.org/show_bug.cgi?id=116842

Reviewed by Antti Koivisto.

  • dom/Node.cpp: Removed hasNonEmptyBoundingBox.
  • dom/Node.h: Ditto.
  • html/HTMLAnchorElement.cpp:

(WebCore::HTMLAnchorElement::isMouseFocusable): Fixed #if to be more readable and
added more-explicit clearer comments. No behavior change.
(WebCore::hasNonEmptyBox): Added. Has the same code that was in
Node::hasNonEmptyBoundingBox with some additional comments.
(WebCore::HTMLAnchorElement::isKeyboardFocusable): Call hasNonEmptyBox.

4:40 PM Changeset in webkit [150782] by Darin Adler
  • 11 edits in trunk/Source/WebCore

Move isBlockFlowElement and related functions out of the Node class into editing code
https://bugs.webkit.org/show_bug.cgi?id=116846

Reviewed by Antti Koivisto.

  • dom/Node.cpp: Removed isBlockFlowElement, enclosingBlockFlowElement, and

inSameContainingBlockFlowElement. These are all editing functions that don't
belong in the Node class, and mostly-deprecated ones to boot.

  • dom/Node.h: Ditto.
  • dom/Position.cpp:

(WebCore::inSameEnclosingBlockFlowElement): Added. This editing-specific function
is used only here, so put it here.
(WebCore::Position::rendersInDifferentPosition): Updated to call new function above.
(WebCore::Position::leadingWhitespacePosition): Ditto.

  • editing/InsertLineBreakCommand.cpp:

(WebCore::InsertLineBreakCommand::insertNodeAfterPosition): Use
deprecatedEnclosingBlockFlowElement at its new location.
(WebCore::InsertLineBreakCommand::insertNodeBeforePosition): Ditto.

  • editing/InsertListCommand.cpp:

(WebCore::InsertListCommand::doApplyForSingleParagraph): use
isBlockFlowElement at its new location.

  • editing/ReplaceSelectionCommand.cpp:

(WebCore::enclosingInline): Ditto.

  • editing/VisiblePosition.cpp:

(WebCore::VisiblePosition::canonicalPosition): Use deprecatedEnclosingBlockFlowElement
at its new location.
(WebCore::enclosingBlockFlowElement): Ditto. Also added FIXME.

  • editing/VisibleSelection.cpp:

(WebCore::makeSearchRange): Ditto.

  • editing/htmlediting.cpp:

(WebCore::isBlockFlowElement): Added. Same as the Node::isBlockFlowElement
function, but here because this is an editing function, not suitable as a Node member.
(WebCore::deprecatedEnclosingBlockFlowElement): Ditto. Added "deprecated" to the name
since the Node member function had a comment saying it was deprecated. Also changed so
it tolerates a 0 for an argument, which made it easier for one caller.

  • editing/htmlediting.h: Added the two new functions.
3:31 PM Changeset in webkit [150781] by benjamin@webkit.org
  • 1 edit
    3 adds in trunk/PerformanceTests

Add a balanced benchmark for QuerySelector
https://bugs.webkit.org/show_bug.cgi?id=116811

Reviewed by Sam Weinig.

The goal of this benchmark is to have an overview of querySelector as typically used
by webpages.

It uses queries similar to what is used by popular websites and applies somewhat
similar weighting for each type of query.

The tree used for the queries is intentionally kept simple to ensure we measure QuerySelector
a not purely the overhead of traversal.

  • CSS/QuerySelector.html: Added.
  • CSS/resources/query-selector.html: Added.
3:09 PM Changeset in webkit [150780] by fpizlo@apple.com
  • 4 edits
    9 adds in branches/dfgFourthTier

fourthTier: DFG ArithMod should have the !nodeUsedAsNumber optimizations that ArithDiv has
https://bugs.webkit.org/show_bug.cgi?id=116841

Reviewed by Mark Hahnenberg.

Source/JavaScriptCore:

  • dfg/DFGSpeculativeJIT.cpp:

(JSC::DFG::SpeculativeJIT::compileArithMod):

LayoutTests:

  • fast/js/dfg-mod-by-neg1-and-then-or-zero-interesting-reg-alloc-expected.txt: Added.
  • fast/js/dfg-mod-by-neg1-and-then-or-zero-interesting-reg-alloc.html: Added.
  • fast/js/dfg-mod-by-zero-and-then-or-zero-interesting-reg-alloc-expected.txt: Added.
  • fast/js/dfg-mod-by-zero-and-then-or-zero-interesting-reg-alloc.html: Added.
  • fast/js/dfg-mod-neg2tothe31-by-one-and-then-or-zero-with-interesting-reg-alloc-expected.txt: Added.
  • fast/js/dfg-mod-neg2tothe31-by-one-and-then-or-zero-with-interesting-reg-alloc.html: Added.
  • fast/js/jsc-test-list:
  • fast/js/script-tests/dfg-mod-by-neg1-and-then-or-zero-interesting-reg-alloc.js: Added.

(foo):

  • fast/js/script-tests/dfg-mod-by-zero-and-then-or-zero-interesting-reg-alloc.js: Added.

(foo):

  • fast/js/script-tests/dfg-mod-neg2tothe31-by-one-and-then-or-zero-with-interesting-reg-alloc.js: Added.

(foo):

2:59 PM Changeset in webkit [150779] by commit-queue@webkit.org
  • 3 edits in trunk/Source/WebKit/blackberry

[BlackBerry] Clear supression flag when caret changes aren't propogated through
https://bugs.webkit.org/show_bug.cgi?id=116840

Patch by Nima Ghanavatian <nghanavatian@blackberry.com> on 2013-05-27
Reviewed by Rob Buis.

Internally reviewed by Mike Fenton.

PR339965
We set and clear the flag upon every use. This leaves us vulnerable if

  1. midway through processing the selection change event is dismissed and
  2. we don't receive a TouchRelease event

One example of this is while dragging the FCC handle we start typing. To
safeguard against this, if the process change guard is set during a
selection change event, we send out to clear the flag status before returning.

  • Api/WebPageClient.h:
  • WebKitSupport/InputHandler.cpp:

(BlackBerry::WebKit::InputHandler::selectionChanged):

2:11 PM Changeset in webkit [150778] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebCore

Add guard around paintForFixedRootBackground
https://bugs.webkit.org/show_bug.cgi?id=116754

Patch by Seokju Kwon <Seokju Kwon> on 2013-05-27
Reviewed by Simon Fraser.

paintForFixedRootBackground is only used when ACCELERATED_COMPOSITING is enabled.

No new tests, no behavior change.

  • rendering/RenderLayer.cpp:

(WebCore):

2:00 PM Changeset in webkit [150777] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebCore

Remove a leftover Pagepoup from PickerIndicatorElement.h
https://bugs.webkit.org/show_bug.cgi?id=116801

Patch by Seokju Kwon <Seokju Kwon> on 2013-05-27
Reviewed by Anders Carlsson.

No new tests, no behavior change.

  • html/shadow/PickerIndicatorElement.h:

(WebCore):

1:58 PM Changeset in webkit [150776] by Claudio Saavedra
  • 5 edits in trunk

[CSS] -webkit-var prefix is case-sensitive
https://bugs.webkit.org/show_bug.cgi?id=116829

Reviewed by Darin Adler.

Source/WebCore:

Tests: Added case to fast/css/variables/case-sensitive.html

  • css/CSSParser.cpp:

(WebCore::isEqualToCSSCaseSensitiveIdentifier): Add variant to
IsEqualToCSSIdentifier that respects case-sensitivity.
(WebCore::CSSParser::realLex): Use the above method for CSS
variable declarations.

LayoutTests:

  • fast/css/variables/case-sensitive.html: Add a

check for case-sensitiveness of the prefix of
variable declarations.

  • fast/css/variables/case-sensitive-expected.html: Update.
1:39 PM WebKitGTK/2.0.x edited by zandobersek@gmail.com
(diff)
1:07 PM Changeset in webkit [150775] by commit-queue@webkit.org
  • 3 edits in trunk/Source/WebCore

drop-shadow filter with overflow:hidden child misbehaves.
https://bugs.webkit.org/show_bug.cgi?id=109783

Patch by Zalan Bujtas <Alan Bujtas> on 2013-05-27
Reviewed by Simon Fraser.

Applying CSS filters only to the scrollbar rect leaves rendering
artefacts. Filters should skip such partial repaints.

Tests: Test framework uses mock theme for scrollbars. Such scrollbars do not
initiate partial repaints.

  • rendering/RenderLayer.cpp:

(WebCore::RenderLayer::setupFilters):
(WebCore::RenderLayer::paintLayerContents):

  • rendering/RenderLayer.h:

(RenderLayer):

1:04 PM Changeset in webkit [150774] by timothy_horton@apple.com
  • 2 edits in trunk/Source/WebKit2

Unreviewed build fix take 2.

  • WebProcess/Plugins/PDF/PDFPlugin.mm:
12:58 PM Changeset in webkit [150773] by timothy_horton@apple.com
  • 2 edits in trunk/Source/WebKit2

Unreviewed build fix.

  • WebProcess/Plugins/PDF/PDFPlugin.mm:
12:51 PM Changeset in webkit [150772] by commit-queue@webkit.org
  • 9 edits
    1 delete in trunk/Source

Move Windows port off legacy clipboard.
https://bugs.webkit.org/show_bug.cgi?id=116258

Patch by Xueqing Huang <huangxueqing@baidu.com> on 2013-05-27
Reviewed by Darin Adler.

Source/WebCore:

Clipboard refactor, no new tests.

  • dom/Clipboard.h:
  • page/win/EventHandlerWin.cpp:

(WebCore::EventHandler::createDraggingClipboard): Create Clipboard instead of ClipboardWin.

  • platform/Pasteboard.h:

(WebCore::Pasteboard::dataObject): Get IDataObject by Clipborad.
(WebCore::Pasteboard::writableDataObject): Get WCDataObject by Clipborad.

  • platform/win/ClipboardWin.cpp:

(WebCore::Clipboard::createDragImage): Change parameter name loc to dragLocation.
(WebCore::Clipboard::declareAndWriteDragImage): Implement declareAndWriteDragImage use WCDataObject exposed by Pasteboard.

  • platform/win/ClipboardWin.h: Removed.
  • platform/win/EditorWin.cpp:

(WebCore::Editor::newGeneralClipboard): Create Clipboard instead of ClipboardWin.

  • platform/win/PasteboardWin.cpp:

(WebCore::Pasteboard::createForCopyAndPaste): Create a Pasteboard.
(WebCore::Pasteboard::createPrivate): Ditto.
(WebCore::Pasteboard::createForDragAndDrop): Ditto.
(WebCore::Pasteboard::finishCreatingPasteboard): Create Pasteboard use Windows API.
(WebCore::Pasteboard::Pasteboard):
(WebCore::clipboardTypeFromMIMEType): Copied from ClipboardWin.cpp.
(WebCore::Pasteboard::clear): Ditto.
(WebCore::Pasteboard::hasData): Ditto.
(WebCore::addMimeTypesForFormat): Ditto.
(WebCore::Pasteboard::types): Ditto.
(WebCore::Pasteboard::readString): Ditto.
(WebCore::Pasteboard::readFilenames): Ditto.
(WebCore::writeURL): Ditto.
(WebCore::Pasteboard::writeString): Ditto.
(WebCore::Pasteboard::setDragImage): Ditto.
(WebCore::Pasteboard::writeRangeToDataObject): Write range to WCDataObject.
(WebCore::Pasteboard::writeSelection):
(WebCore::Pasteboard::writePlainTextToDataObject): Write PlainText to WCDataObject.
(WebCore::Pasteboard::writePlainText):
(WebCore::pathRemoveBadFSCharacters): Copied from ClipboardWin.cpp.
(WebCore::filesystemPathFromUrlOrTitle): Ditto.
(WebCore::writeFileToDataObject): Ditto.
(WebCore::Pasteboard::writeURLToDataObject): Write URL to WCDataObject.
(WebCore::Pasteboard::writeURL): Copied from ClipboardWin.cpp.
(WebCore::Pasteboard::writePasteboard): Copied from Pasteboard::writeClipboard.
(WebCore::Pasteboard::documentFragment): Copied from ClipboardWin.cpp.
(WebCore::Pasteboard::setExternalDataObject): Clipboard can set new IDataObject.
(WebCore::getCachedImage): Copied from ClipboardWin.cpp.
(WebCore::createGlobalImageFileDescriptor): Ditto.
(WebCore::createGlobalImageFileContent): Ditto.
(WebCore::createGlobalHDropContent): Ditto.
(WebCore::Pasteboard::writeImageToDataObject): Expose to make Clipboard implement declareAndWriteDragImage.
(WebCore::Pasteboard::writeURLToWritableDataObject): Ditto.

Source/WebKit/win:

  • WebCoreSupport/WebDragClient.cpp:

(WebDragClient::willPerformDragSourceAction): Get IDataObject from Pasteboard instead of Clipboard.
(WebDragClient::startDrag): Ditto.

12:50 PM Changeset in webkit [150771] by timothy_horton@apple.com
  • 13 edits in trunk/Source

PDFPlugin: Main-frame PDFPlugin Accessibility
https://bugs.webkit.org/show_bug.cgi?id=116563
<rdar://problem/13458534>

Reviewed by Sam Weinig.

  • WebProcess/Plugins/PDF/PDFLayerControllerDetails.h:

Forward-declare a variety of accessibility methods.

  • WebProcess/Plugins/PDF/PDFPlugin.h:

Add convertFromPDFViewToScreen, boundsOnScreen, accessibilityObject, and
storage for our accessibility wrapper.

  • WebProcess/Plugins/PDF/PDFPlugin.mm:

Add WKPDFPluginAccessibilityObject.

(-[WKPDFPluginAccessibilityObject accessibilityIsIgnored]):
We want to be included in the accessibility tree.

(-[WKPDFPluginAccessibilityObject accessibilityAttributeValue:]):
(-[WKPDFPluginAccessibilityObject accessibilityAttributeValue:forParameter:]):
Grab values for accessibility attributes from the relevant sources
(mostly from PDFLayerController).

(-[WKPDFPluginAccessibilityObject readingModel]): Ditto.
(-[WKPDFPluginAccessibilityObject accessibilityAttributeNames]):
(-[WKPDFPluginAccessibilityObject accessibilityParameterizedAttributeNames]):
(-[WKPDFPluginAccessibilityObject accessibilityActionNames]):
Return the set of accessibility attributes, actions, and parameterized
attributes that we support.

(-[WKPDFPluginAccessibilityObject accessibilityIsAttributeSettable:]):
(-[WKPDFPluginAccessibilityObject accessibilitySetValue:forAttribute:]):
Forward some other accessibility methods to PDFLayerController.

(-[WKPDFPluginAccessibilityObject accessibilityFocusedUIElement]):
(-[WKPDFPluginAccessibilityObject accessibilityHitTest:]):
We have no children, so if a hit test makes it to us, it's ours.

(WebKit::PDFPlugin::PDFPlugin):
Make a WKPDFPluginAccessibilityObject and hook it up.
[FIXME: this could use some de-crapifying]

(WebKit::PDFPlugin::convertFromPluginToPDFView):
Transform a point from plugin space to PDFLayerController space.

(WebKit::PDFPlugin::convertFromPDFViewToRootView):
(WebKit::PDFPlugin::convertFromPDFViewToScreen): Added.
(WebKit::PDFPlugin::boundsOnScreen):
Return the PDFPlugin's frame in screen coordinates.

(WebKit::PDFPlugin::accessibilityObject):
Return our accessibility wrapper.

  • WebProcess/Plugins/PDF/SimplePDFPlugin.h:
  • WebProcess/Plugins/Plugin.h:

(WebKit::Plugin::accessibilityObject):

  • WebProcess/Plugins/PluginView.cpp:

(WebKit::PluginView::accessibilityObject):

  • WebProcess/Plugins/PluginView.h:

Add accessibilityObject(), to acquire accessibility wrappers from plugins.

  • WebProcess/WebPage/WebPage.h:

(WebPage): Add accessibilityObjectForMainFramePlugin();

  • WebProcess/WebPage/mac/WKAccessibilityWebPageObject.mm:

(-[WKAccessibilityWebPageObject accessibilityRootObjectWrapper]):
If the main frame has a PluginDocument that has an accessibilityObject,
use that one instead of the WebPage's.

  • WebProcess/WebPage/mac/WebPageMac.mm:

(WebKit::WebPage::accessibilityObjectForMainFramePlugin):
Return the accessibilityObject from the main-frame PluginDocument, if
there is one.

  • WebCore.exp.in: Export AffineTransform::mapRect.
12:27 PM Changeset in webkit [150770] by commit-queue@webkit.org
  • 6 edits in trunk

[BlackBerry] Find-on-page should be able to convert the active match to selection when clearing all matches
https://bugs.webkit.org/show_bug.cgi?id=116837

Patch by Andy Chen <andchen@blackberry.com> on 2013-05-27
Reviewed by Rob Buis.

PR 291903
Internally reviewed by Mike Fenton.

  • Api/WebPage.cpp:

(BlackBerry::WebKit::WebPage::findNextString):

  • Api/WebPage.h:
  • WebKitSupport/InPageSearchManager.cpp:

(BlackBerry::WebKit::InPageSearchManager::findNextString):
(BlackBerry::WebKit::InPageSearchManager::clearTextMatches):

  • WebKitSupport/InPageSearchManager.h:

(InPageSearchManager):

11:39 AM Changeset in webkit [150769] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebCore

Checkbox UX update.
https://bugs.webkit.org/show_bug.cgi?id=116838.

Patch by Tiancheng Jiang <tijiang@rim.com> on 2013-05-27
Reviewed by Rob Buis.

BlackBerry PR 342302.
Internally Reviewed by Jeff Rogers.
Remove old background and change pressed checkmark image.

  • platform/blackberry/RenderThemeBlackBerry.cpp:

(WebCore::RenderThemeBlackBerry::paintCheckbox):

11:25 AM Changeset in webkit [150768] by weinig@apple.com
  • 2 edits in trunk/Source/WebCore

When the background is completely transparent we should be conservative and use the default scrollbar style, not the light one
https://bugs.webkit.org/show_bug.cgi?id=116836
<rdar://problem/13949905>

Reviewed by Simon Fraser.

  • page/FrameView.cpp:

(WebCore::FrameView::recalculateScrollbarOverlayStyle):
Don't use the light style when the background is completely transparent.

11:09 AM Changeset in webkit [150767] by Patrick Gansterer
  • 10 edits in trunk

Use ICU_INCLUDE_DIRS in BlackBerry CMake files
https://bugs.webkit.org/show_bug.cgi?id=116210

Reviewed by Rob Buis.

Set and use the ICU_INCLUDE_DIRS variable to avoid
duplicated adding of the ICU include directory.

.:

  • Source/cmake/OptionsBlackBerry.cmake:

Source/JavaScriptCore:

  • PlatformBlackBerry.cmake:

Source/WebCore:

  • PlatformBlackBerry.cmake:

Source/WebKit:

  • PlatformBlackBerry.cmake:

Source/WTF:

  • wtf/PlatformBlackBerry.cmake:
11:04 AM Changeset in webkit [150766] by efidler@rim.com
  • 2 edits in trunk/Source/WebKit/blackberry

[BlackBerry] Fix subframe target added to new requests
https://bugs.webkit.org/show_bug.cgi?id=116602

Reviewed by Rob Buis.

In some cases we might be adding TargetIsMainFrame to a subframe
request causing an isMainFrame assert.

  • WebCoreSupport/FrameLoaderClientBlackBerry.cpp:

(WebCore::FrameLoaderClientBlackBerry::dispatchWillSendRequest):
Check we are actually loading the main frame before setting
ResourceRequest::TargetIsMainFrame target.

10:39 AM Changeset in webkit [150765] by fpizlo@apple.com
  • 10 edits in branches/dfgFourthTier

fourthTier: clean up ArithDiv/ArithMod in the DFG
https://bugs.webkit.org/show_bug.cgi?id=116793

Source/JavaScriptCore:

Reviewed by Mark Hahnenberg.

This makes ArithDiv and ArithMod behave similarly, and moves both of their
implementations entirely into DFGSpeculativeJIT.cpp into methods named like
the ones for ArithSub/ArithMul.

Specifically, ArithMod now uses the wrap-in-conversion-nodes idiom that
ArithDiv used for platforms that don't support integer division. Previously
ArithMod had its own int-to-double and double-to-int conversions for this
purpose.

As well, this gets rid of confusing methods like compileSoftModulo() (which
did no such thing, there wasn't anything "soft" about it) and
compileIntegerArithDivForX86() (which is accurately named but we don't use
the platform-specific method convention anywhere else).

Finally, this takes the optimized power-of-two modulo operation that was
previously only for ARMv7s, and makes it available for all platforms. Well,
sort of: I actually rewrote it to do what latest LLVM appears to do, which
is a crazy straight-line power-of-2 modulo based on a combination of shifts,
ands, additions, and subtractions. I can kind of understand it well enough
to see that it complies with both C and JS power-of-2 modulo semantics. I've
also confirmed that it does by testing (hence the corresponding improvements
to one of the division tests). But, I don't claim to know exactly how this
code works other than to observe that it is super leet.

Overall, this patch has the effect of killing some code (no more hackish
int-to-double conversions in ArithMod), making some optimization work on
more platforms, and making the compiler less confusing by doing more things
with the same idiom.

  • dfg/DFGAbstractState.cpp:

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

  • dfg/DFGFixupPhase.cpp:

(JSC::DFG::FixupPhase::fixupNode):

  • dfg/DFGSpeculativeJIT.cpp:

(DFG):
(JSC::DFG::SpeculativeJIT::compileArithDiv):
(JSC::DFG::SpeculativeJIT::compileArithMod):

  • dfg/DFGSpeculativeJIT.h:

(SpeculativeJIT):

  • dfg/DFGSpeculativeJIT32_64.cpp:

(JSC::DFG::SpeculativeJIT::compile):

  • dfg/DFGSpeculativeJIT64.cpp:

(JSC::DFG::SpeculativeJIT::compile):

LayoutTests:

Reviewed by Mark Hahnenberg.

  • fast/js/script-tests/integer-division-neg2tothe32-by-neg1.js:

(myModBy2):
(myModBy1073741824):

9:52 AM Changeset in webkit [150764] by marcelo.lira@openbossa.org
  • 3 edits in trunk/Source/WebKit2

[WK2][CoordinatedGraphics] Misuse of DrawingAreaProxy::setVisibleContentsRect() in WebView::updateViewportSize()
https://bugs.webkit.org/show_bug.cgi?id=116688

Reviewed by Noam Rosenthal.

Visible contents area passed to DrawingAreaProxy::setVisibleContentsRect()
must be calculated including the page scale factor and avoid crossing the
boundaries of the page contents size. In other words the DrawingArea
must know the actual dimensions of what it being drawn. If the
DrawingArea thinks the visible rectangle is bigger than it is, the
contents will be drawn pixelated and blurry.

  • UIProcess/CoordinatedGraphics/WebView.cpp:

(WebKit::WebView::didChangeContentsSize):
(WebKit::WebView::updateViewportSize):
(WebKit::WebView::visibleContentsSize):
(WebKit):

  • UIProcess/CoordinatedGraphics/WebView.h:

(WebKit::WebView::contentsSize):
(WebView):

9:10 AM Changeset in webkit [150763] by Carlos Garcia Campos
  • 2 edits in trunk/Source/WebKit2

[GTK] WebKitWebView child widgets are not rendered with recent GTK+
https://bugs.webkit.org/show_bug.cgi?id=116794

Reviewed by Xan Lopez.

Since rev d22fd7223c75f4720ddb982c659efb0d8d7543c4 of GTK+ expose
events are not propagated to child GdkWindows anymore, and child
widgets are rendered via gtk_container_propagate_draw(). Since all
our children have its own GdkWindow (auth widget, inspector view
and windowed plugins) we don't propagate the draw signal making
them invisible with current GTK+.

  • UIProcess/API/gtk/WebKitWebViewBase.cpp:

(webkitWebViewBaseDraw): Chain up to the parent draw(), so it's
propagated to the children. This shouldn't have any effect in
previous versions of GTK+ because gtk_container_propagate_draw()
returned early when the child widget had its own GdkWindow.

9:01 AM Changeset in webkit [150762] by Antti Koivisto
  • 12 edits in trunk/Source/WebCore

Tighten FontGlyphs::glyphDataAndPageForCharacter to take FontDescription
https://bugs.webkit.org/show_bug.cgi?id=116823

Reviewed by Andreas Kling.

FontGlyphs does not need to know about Font anymore.

  • platform/graphics/Font.cpp:

(WebCore::Font::Font):
(WebCore::Font::operator=):

  • platform/graphics/Font.h:

(WebCore::Font::isPlatformFont):
(WebCore::Font::glyphDataAndPageForCharacter):
(Font):

  • platform/graphics/FontCache.h:
  • platform/graphics/FontFallbackList.cpp:

(WebCore::FontGlyphs::FontGlyphs):
(WebCore::FontGlyphs::glyphDataAndPageForCharacter):

  • platform/graphics/FontFallbackList.h:

(WebCore):
(WebCore::FontGlyphs::createForPlatformFont):

Tweaked the name.

(WebCore::FontGlyphs::isForPlatformFont):

Move knowledge about platform font to FontGlyphs as it is needed for resolving system fallback on Mac.

  • platform/graphics/blackberry/FontCacheBlackBerry.cpp:

(WebCore::FontCache::systemFallbackForCharacters):

  • platform/graphics/freetype/FontCacheFreeType.cpp:

(WebCore::FontCache::systemFallbackForCharacters):

  • platform/graphics/mac/FontCacheMac.mm:

(WebCore::FontCache::systemFallbackForCharacters):

  • platform/graphics/qt/FontCacheQt.cpp:

(WebCore::FontCache::systemFallbackForCharacters):

  • platform/graphics/win/FontCacheWin.cpp:

(WebCore::FontCache::systemFallbackForCharacters):

  • platform/graphics/wince/FontCacheWinCE.cpp:

(WebCore::FontCache::systemFallbackForCharacters):

More descriptive name and tighter interface.

8:52 AM Changeset in webkit [150761] by mihnea@adobe.com
  • 5 edits in trunk/Source/WebCore

[CSSRegions] Prevent unnecessary copy of LayoutRect objects
https://bugs.webkit.org/show_bug.cgi?id=116828

Reviewed by Andreas Kling.

Use const references for parameter passing.
No functional change, no new tests.

  • rendering/RenderFlowThread.cpp:

(WebCore::RenderFlowThread::paintFlowThreadPortionInRegion):
(WebCore::RenderFlowThread::hitTestFlowThreadPortionInRegion):

  • rendering/RenderFlowThread.h:
  • rendering/RenderRegion.cpp:

(WebCore::RenderRegion::overflowRectForFlowThreadPortion):

  • rendering/RenderRegion.h:
8:51 AM Changeset in webkit [150760] by commit-queue@webkit.org
  • 3 edits in trunk/Source/WebKit/blackberry

[BlackBerry] REGRESSION(r150071): Animation performance issues in some websites with CSS transforms
https://bugs.webkit.org/show_bug.cgi?id=116724

Patch by Carlos Garcia Campos <cgarcia@igalia.com> on 2013-05-27
Reviewed by Rob Buis.

Add another bool variable m_previousFrameDone to make sure we
don't start a new frame until the previous one has been done. Also
make sure we pass the animation start time to
serviceScriptedAnimations().

  • Api/WebPage.cpp:

(BlackBerry::WebKit::WebPagePrivate::WebPagePrivate): Initialize
m_previousFrameDone and m_monotonicAnimationStartTime.
(BlackBerry::WebKit::WebPagePrivate::animationFrameChanged):
Return early if previous frame is not done. Otherwise set
m_previousFrameDone to false and save the animation start time.
(BlackBerry::WebKit::WebPagePrivate::serviceAnimations): Call
serviceScriptedAnimations() passing the saved animation start time
and reset m_animationScheduled and m_previousFrameDone.
(BlackBerry::WebKit::WebPagePrivate::handleServiceScriptedAnimationsOnMainThread):
Call WebPagePrivate::serviceAnimations().

  • Api/WebPage_p.h:

(WebPagePrivate):

8:45 AM Changeset in webkit [150759] by leoyang@rim.com
  • 2 edits in trunk/Source/WebCore

[BlackBerry] Reveal HTTP error code for XHR even if there is no response body
https://bugs.webkit.org/show_bug.cgi?id=116543

Reviewed by Rob Buis.
Reviewed internally by Joe Mason.

It's a regression introduced by r132556.
BlackBerry is doing friendly HTTP error handling: if there is no response body
for an HTTP error it will display an error message made by the port. But for a
XHR request, which usually does not result in displaying error messages to the
end user, HTTP error code should be revealed to let the JavaScript know what
the actual server error code is.

Covered by existing tests:
ManualTests/blackberry/head-xhr-nonexistant-file.html
LayoutTests/http/tests/xmlhttprequest/xmlhttprequest-check-head-readystate-for-404.html
LayoutTests/http/tests/xmlhttprequest/xmlhttprequest-check-get-readystate-for-404-without-body.html

  • platform/network/blackberry/NetworkJob.cpp:

(WebCore::NetworkJob::shouldNotifyClientFailed):

8:43 AM Changeset in webkit [150758] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebKit2

Reset all clients on WebPage close
https://bugs.webkit.org/show_bug.cgi?id=112334

Patch by Xan Lopez <xlopez@igalia.com> on 2013-05-27
Reviewed by Anders Carlsson.

We do not reset the clients on WebPage::close(), so there's a
small chance that someone will try to access them after that. This
won't work and will generally cause runtime warnings, crashes or
other issues. Reset them all to null to prevent this situation.

  • WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::close):

8:37 AM Changeset in webkit [150757] by commit-queue@webkit.org
  • 3 edits in trunk/Source/WebKit/blackberry

[BlackBerry] Remove encoding parameter from FrameLoaderClientBlackBerry::receivedData()
https://bugs.webkit.org/show_bug.cgi?id=116598

Patch by Carlos Garcia Campos <cgarcia@igalia.com> on 2013-05-27
Reviewed by Rob Buis.

It's no longer needed since the encoder now checks the override encoding.

  • WebCoreSupport/FrameLoaderClientBlackBerry.cpp:

(WebCore::FrameLoaderClientBlackBerry::committedLoad): Call
receivedData() passing the DocumentLoader instead of the text
encoding.
(WebCore::FrameLoaderClientBlackBerry::receivedData): Call
DocumentLoader::commitData() directly.

  • WebCoreSupport/FrameLoaderClientBlackBerry.h:

(FrameLoaderClientBlackBerry):

8:35 AM Changeset in webkit [150756] by commit-queue@webkit.org
  • 6 edits
    10 moves in trunk/Source/WebKit

[BlackBerry] Move PagePopup implementation to WebKitSupport
https://bugs.webkit.org/show_bug.cgi?id=116824

Patch by Carlos Garcia Campos <cgarcia@igalia.com> on 2013-05-27
Reviewed by Rob Buis.

Source/WebKit:

  • PlatformBlackBerry.cmake: Update new file paths and names.

Source/WebKit/blackberry:

After r150434 PagePopup implementation in BlackBerry port is
independent from WebCore. The classes are now in the
BlackBerry::WebKit namespace so they can be moved from
WebCoresupport to WebKitSupport and renamed to remove the
BlackBerry suffix which is now redundant. Also, now that the
implementation is not shared, it can be simplified a bit more.

  • Api/WebPage.cpp:

(BlackBerry::WebKit::WebPagePrivate::WebPagePrivate): Update to
API changes.
(BlackBerry::WebKit::WebPage::initPopupWebView): Ditto.
(BlackBerry::WebKit::WebPagePrivate::openPagePopup): Ditto.
(BlackBerry::WebKit::WebPagePrivate::closePagePopup): Ditto.
(BlackBerry::WebKit::WebPagePrivate::hasOpenedPopup): Ditto.

  • Api/WebPage.h:

(WebKit): Ditto.

  • Api/WebPage_p.h:

(WebKit): Ditto.
(WebPagePrivate): Rename m_selectPopup as m_pagePopup, since it
can also point to a color picker or date picker page popup. Also
use RefPtr instead of raw pointer.

  • WebKitSupport/ColorPickerClient.cpp: Renamed from Source/WebKit/blackberry/WebCoreSupport/ColorPickerClient.cpp.

(BlackBerry::WebKit::ColorPickerClient::ColorPickerClient):
(BlackBerry::WebKit::ColorPickerClient::generateHTML):
(BlackBerry::WebKit::ColorPickerClient::setValueAndClosePopup):
Return early if popup has been closed already.
(BlackBerry::WebKit::ColorPickerClient::didClosePopup):

  • WebKitSupport/ColorPickerClient.h: Renamed from Source/WebKit/blackberry/WebCoreSupport/ColorPickerClient.h.
  • WebKitSupport/DatePickerClient.cpp: Renamed from Source/WebKit/blackberry/WebCoreSupport/DatePickerClient.cpp.

(BlackBerry::WebKit::DatePickerClient::DatePickerClient):
(BlackBerry::WebKit::DatePickerClient::~DatePickerClient):
(BlackBerry::WebKit::DatePickerClient::generateHTML):
(BlackBerry::WebKit::DatePickerClient::setValueAndClosePopup):
Return early if popup has been closed already.
(BlackBerry::WebKit::DatePickerClient::didClosePopup):
(BlackBerry::WebKit::DatePickerClient::generateDateLabels):

  • WebKitSupport/DatePickerClient.h: Renamed from Source/WebKit/blackberry/WebCoreSupport/DatePickerClient.h.
  • WebKitSupport/PagePopup.cpp: Renamed from Source/WebKit/blackberry/WebCoreSupport/PagePopupBlackBerry.cpp.

(BlackBerry::WebKit::PagePopup::PagePopup):
(BlackBerry::WebKit::PagePopup::~PagePopup):
(BlackBerry::WebKit::PagePopup::initialize): Renamed as initialize
and made void, since it always returned true and the callers
ignored the return value.
(BlackBerry::WebKit::PagePopup::writeDocument): Renamed from
generateHTML for consistency with PagePopupClient::writeDocument()
and to avoid confusion with generateHTML method in the clients
that create the HTML contents but they don't write the document.
(BlackBerry::WebKit::PagePopup::setValueAndClosePopupCallback):
Make it static member of PagePopup since the JS object now holds a
reference of PagePopup instance as private data.
(BlackBerry::WebKit::popUpExtensionInitialize): Take a reference
of the PagePopup instance saved as private data of the JS object.
(BlackBerry::WebKit::popUpExtensionFinalize): Unref the PagePopup
instance saved as private data of the JS object.
(BlackBerry::WebKit::PagePopup::installDOMFunction): Instead of
saving the client as private data of the JS object using a wrapper
refcounted object, save the PagePopup instance that is now
refcounted and access the client from the PagePopup instance in
the callback.
(BlackBerry::WebKit::PagePopup::close): Renamed from closePagePopup
since PagePopup::closePagePopup was a bit redundant.

  • WebKitSupport/PagePopup.h: Renamed from Source/WebKit/blackberry/WebCoreSupport/PagePopupBlackBerry.h.

(BlackBerry::WebKit::PagePopup::create): Static member to create
instances of PagePopup now that it's refcounted.

  • WebKitSupport/PagePopupClient.cpp: Renamed from Source/WebKit/blackberry/WebCoreSupport/PagePopupBlackBerryClient.cpp.

(BlackBerry::WebKit::PagePopupClient::PagePopupClient):
(BlackBerry::WebKit::PagePopupClient::closePopup):
(BlackBerry::WebKit::PagePopupClient::didClosePopup):
(BlackBerry::WebKit::PagePopupClient::contentSize):
(BlackBerry::WebKit::PagePopupClient::writeDocument):

  • WebKitSupport/PagePopupClient.h: Renamed from Source/WebKit/blackberry/WebCoreSupport/PagePopupBlackBerryClient.h.
  • WebKitSupport/SelectPopupClient.cpp: Renamed from Source/WebKit/blackberry/WebCoreSupport/SelectPopupClient.cpp.

(BlackBerry::WebKit::SelectPopupClient::SelectPopupClient):
(BlackBerry::WebKit::SelectPopupClient::~SelectPopupClient):
(BlackBerry::WebKit::SelectPopupClient::generateHTML):
(BlackBerry::WebKit::SelectPopupClient::setValueAndClosePopup):
Return early if popup has been closed already.
(BlackBerry::WebKit::SelectPopupClient::didClosePopup):
(BlackBerry::WebKit::SelectPopupClient::notifySelectionChange):

  • WebKitSupport/SelectPopupClient.h: Renamed from Source/WebKit/blackberry/WebCoreSupport/SelectPopupClient.h.
8:21 AM Changeset in webkit [150755] by commit-queue@webkit.org
  • 3 edits in trunk/LayoutTests

[EFL][GTK] fast/regions/offsetLeft-offsetTop-in-multiple-regions.html is failing
https://bugs.webkit.org/show_bug.cgi?id=116808

Patch by Radu Stavila <stavila@adobe.com> on 2013-05-27
Reviewed by Anders Carlsson.

The test was failing because of differences between different fonts.

  • fast/regions/offsetLeft-offsetTop-in-multiple-regions-expected.txt:
  • fast/regions/offsetLeft-offsetTop-in-multiple-regions.html:
8:03 AM Changeset in webkit [150754] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebCore

Adjust date input width.
https://bugs.webkit.org/show_bug.cgi?id=100760.

Patch by Tiancheng Jiang <tijiang@rim.com> on 2013-05-27
Reviewed by Rob Buis.

BlackBerry PR 343497.
Internally Reviewed by Jeff Rogers.

Set width to auto to fit the string length.

  • css/themeBlackBerry.css:

(input[type="date"], input[type="datetime"], input[type="datetime-local"], input[type="time"], input[type="month"]):

6:51 AM BadContent edited by zandobersek@gmail.com
Add a spamming email that was creating irrelevant attachments. (diff)
6:32 AM Changeset in webkit [150753] by commit-queue@webkit.org
  • 3 edits in trunk/Source/WebKit2

[GTK] Notify WebProcess in WebKitURISchemeRequest when we fail to read the user InputStream
https://bugs.webkit.org/show_bug.cgi?id=114347

Patch by Manuel Rego Casasnovas <Manuel Rego Casasnovas> on 2013-05-27
Reviewed by Carlos Garcia Campos.

While processing a WebKitURISchemeRequest if there is any error reading
the InputStream provided by the user (for example the stream is already
closed) we have to notify the WebProcess that the request has failed.

  • UIProcess/API/gtk/WebKitURISchemeRequest.cpp:

(webkitURISchemeRequestReadCallback): Get the error from
g_input_stream_read_finish() and use it to call
webkit_uri_scheme_request_finish_error() in order to finish the failing
WebKitURISchemeRequest properly.

  • UIProcess/API/gtk/tests/TestWebKitWebContext.cpp:

(testWebContextURIScheme): Modify test to check the new situation using
an already closed InputStream.

6:25 AM Changeset in webkit [150752] by akling@apple.com
  • 4 edits in trunk/Source/WebCore

Document::setHoveredNode() should be setHoveredElement().
<http://webkit.org/b/116819>

Reviewed by Antti Koivisto.

Only Elements can be hovered, teach Document about this.

  • dom/Document.h:
  • dom/Document.cpp:

(WebCore::Document::dispose):
(WebCore::Document::detach):

m_hoverNode => m_hoveredElement.
Removed hoverNode() and setHoverNode() since those were only used internally.

(WebCore::Document::updateHoverActiveState):

Updated for hovered things always being Elements.

(WebCore::Document::hoveredElementDidDetach):

  • dom/Element.cpp:

(WebCore::Element::detach):

Renamed Document::hoveredNodeDetached() to hoveredElementDidDetach() and fix it up
following the same pattern as elementInActiveChainDidDetach().

6:00 AM Changeset in webkit [150751] by kadam@inf.u-szeged.hu
  • 2 edits in trunk/Source/WTF

[Qt] Enable parallel GC. Probably it has been fixed in r131791.
https://bugs.webkit.org/show_bug.cgi?id=90957

Reviewed by Csaba Osztrogonác.

  • wtf/Platform.h:
5:48 AM Changeset in webkit [150750] by allan.jensen@digia.com
  • 2 edits in trunk/Source/WebCore

[Qt] REGRESSION(r150716): It made 15 inspector and 1 mozilla test to assert
https://bugs.webkit.org/show_bug.cgi?id=116822

Reviewed by Andreas Kling.

The input fontDescription may have more than one family defined.

  • platform/qt/RenderThemeQStyle.cpp:

(WebCore::RenderThemeQStyle::adjustButtonStyle):

5:33 AM Changeset in webkit [150749] by jocelyn.turcotte@digia.com
  • 6 edits in trunk/Source

[Qt][Win] Input events aren't mapped properly with windowless plugins.
https://bugs.webkit.org/show_bug.cgi?id=116094

Reviewed by Tor Arne Vestbø.

Source/WebCore:

The events are first sent properly but Flash then immediately repaints
and this causes flickering painting.
The issue is that Flash seems to be doing some input event tracking of
its own internally, using the HWND returned through NPN_GetValue(NPNVnetscapeWindow).

We are currently using two coordinate systems for windowless plugins on Windows with Qt:

  • FrameView coordinates: Used for input events and ajusted with the WM_WINDOWPOSCHANGED message
  • Drawable coordinates: Used by WM_PAINT and adjusted with NPP_SetWindow

This patch fixes the bug by mapping input events to the native window returned
as NPNVnetscapeWindow instead of the FrameView to ensure that those coordinates will match
the ones used by Flash internally.
With this we shouldn't be using FrameView coordinates anywhere for windowless plugins
on Windows with Qt.

  • platform/qt/QWebPageClient.h:

(QWebPageClient):

Added mapToOwnerWindow to the interface, mapping from the FrameView up to the wrapping nativeParentWidget.

  • plugins/win/PluginViewWin.cpp:

(WebCore::contentsToNativeWindow):
(WebCore::PluginView::paintIntoTransformedContext):
(WebCore::PluginView::handleMouseEvent):

Source/WebKit/qt:

  • WidgetSupport/PageClientQt.cpp:

(WebCore::PageClientQWidget::mapToOwnerWindow):
(WebCore::PageClientQGraphicsWidget::makeOpenGLContextCurrentIfAvailable):
(WebCore::PageClientQGraphicsWidget::screenNumber):
(WebCore::PageClientQGraphicsWidget::geometryRelativeToOwnerWidget):
(WebCore::PageClientQGraphicsWidget::mapToOwnerWindow):
(WebCore::PageClientQGraphicsWidget::graphicsItemVisibleRect):
(WebCore::PageClientQGraphicsWidget::firstGraphicsView):

Extracted this common logic from makeOpenGLContextCurrentIfAvailable,
screenNumber, geometryRelativeToOwnerWidget and graphicsItemVisibleRect
to be able to use it in mapToOwnerWindow.

  • WidgetSupport/PageClientQt.h:

(PageClientQWidget):
(PageClientQGraphicsWidget):

5:23 AM Changeset in webkit [150748] by rgabor@webkit.org
  • 2 edits in trunk/Source/JavaScriptCore

MacroAssemblerARM should use xor to swap registers instead of move
https://bugs.webkit.org/show_bug.cgi?id=116306

Reviewed by Zoltan Herczeg.

Change register swapping to xor from move and this way we don't need
temporary register anymore.

  • assembler/MacroAssemblerARM.h:

(JSC::MacroAssemblerARM::swap):

4:05 AM Changeset in webkit [150747] by Antti Koivisto
  • 13 edits in trunk/Source/WebCore

Tighten FontGlyphs interfaces to take FontDescription instead of Font
https://bugs.webkit.org/show_bug.cgi?id=116818

Reviewed by Andreas Kling.

This gives better understanding of what the actual inputs are.

  • WebCore.exp.in:
  • platform/graphics/Font.h:

(WebCore::Font::primaryFont):
(WebCore::Font::fontDataAt):
(WebCore::Font::isFixedPitch):

  • platform/graphics/FontCache.cpp:

(WebCore::FontCache::getFontData):

  • platform/graphics/FontCache.h:
  • platform/graphics/FontFallbackList.cpp:

(WebCore::FontGlyphs::determinePitch):
(WebCore::FontGlyphs::realizeFontDataAt):
(WebCore::FontGlyphs::glyphDataAndPageForCharacter):

  • platform/graphics/FontFallbackList.h:

(WebCore::FontGlyphs::primaryFontData):
(WebCore::FontGlyphs::isFixedPitch):
(WebCore::FontGlyphs::primarySimpleFontData):

  • platform/graphics/blackberry/FontCacheBlackBerry.cpp:
  • platform/graphics/freetype/FontCacheFreeType.cpp:
  • platform/graphics/mac/FontCacheMac.mm:

(WebCore::FontCache::similarFontPlatformData):

Drop get* naming and make this Mac only as no other platform has it.

  • platform/graphics/qt/FontCacheQt.cpp:
  • platform/graphics/win/FontCacheWin.cpp:
  • platform/graphics/wince/FontCacheWinCE.cpp:

(WebCore):

3:27 AM Changeset in webkit [150746] by commit-queue@webkit.org
  • 2 edits in trunk/LayoutTests

New test direct-pattern-compositing-contain-no-repeat.html fails on Mac
https://bugs.webkit.org/show_bug.cgi?id=116816

Modified the layer in the reference to still be composited, but have invisible painted content.

Patch by Noam Rosenthal <Noam Rosenthal> on 2013-05-27
Reviewed by Andreas Kling.

  • compositing/patterns/direct-pattern-compositing-contain-no-repeat-expected.html:
3:23 AM Changeset in webkit [150745] by Carlos Garcia Campos
  • 2 edits in trunk/Source/WebKit2

Unreviewed. Fix make distcheck.

  • GNUmakefile.list.am: Add missing header file.
2:45 AM Changeset in webkit [150744] by akling@apple.com
  • 5 edits in trunk/Source/WebCore

Strengthen typing of detaching an Element from Document's :active chain.
<http://webkit.org/b/116815>

Reviewed by Antti Koivisto.

Tiny cleanup now that the :active chain can only contain elements.
Rename Document::activeChainNodeDetached() to elementInActiveChainDidDetach().

  • dom/Element.cpp:

(WebCore::Element::detach):

  • dom/Document.h:
  • dom/Document.cpp:

(WebCore::Document::elementInActiveChainDidDetach):

Rename, take Element*, remove silly condition that was always true.

2:41 AM Changeset in webkit [150743] by abucur@adobe.com
  • 11 edits in trunk/Source/WebCore

[CSS Regions] Remove the offsetFromLogicalTopOfFirstPage parameter from layout functions
https://bugs.webkit.org/show_bug.cgi?id=116441

Reviewed by David Hyatt.

The patch changes the following things:

  • exposes offsetFromLogicalTopOfFirstPage() to all the RenderBoxes;
  • offsetFromLogicalTopOfFirstPage() is now safe to call from any box inside a flow thread, not only the

currently being laid out;

  • removes the offsetFromLogicalTopOfFirstPage parameter from all the layout functions

that didn't actually required it; when the value is necessary, the offsetFromLogicalTopOfFirstPage
function is called instead;

  • RenderFlowThread now handles its own logic for obtaining the offset from logical top of the first

region for a box, based on the layout states pushed on RenderView; it provides a better caching mechanism
that can be used for all the ancestor containing blocks of the current box being laid out.

The RenderBoxRegionInfo structures are cached only based on the box and the region. This means that
passing around a value for offsetFromLogicalTopOfFirstPage different in some cases than others was
risky. If there is a possibility of having a different box info result for a different offset
then the cache is not working correctly.

Tests: No function change, no tests.

  • rendering/RenderBlock.cpp:

(WebCore::RenderBlock::computeStartPositionDeltaForChildAvoidingFloats):
(WebCore::RenderBlock::logicalLeftOffsetForContent):
(WebCore::RenderBlock::logicalRightOffsetForContent):
(WebCore::RenderBlock::getClearDelta):
(WebCore::RenderBlock::lineWidthForPaginatedLineChanged):
(WebCore::RenderBlock::offsetFromLogicalTopOfFirstPage):
(WebCore::RenderBlock::logicalWidthChangedInRegions):

  • rendering/RenderBlock.h:

(WebCore::RenderBlock::availableLogicalWidthForLine):
(WebCore::RenderBlock::logicalRightOffsetForLine):
(WebCore::RenderBlock::logicalLeftOffsetForLine):
(WebCore::RenderBlock::startOffsetForLine):
(WebCore::RenderBlock::endOffsetForLine):
(WebCore::RenderBlock::availableLogicalWidthForContent):
(WebCore::RenderBlock::startOffsetForContent):
(WebCore::RenderBlock::endOffsetForContent):
(WebCore::RenderBlock::logicalLeftOffsetForContent):
(WebCore::RenderBlock::logicalRightOffsetForContent):

  • rendering/RenderBox.cpp:

(WebCore::RenderBox::borderBoxRectInRegion):
(WebCore::RenderBox::constrainLogicalWidthInRegionByMinMax):
(WebCore::RenderBox::shrinkLogicalWidthToAvoidFloats):
(WebCore::RenderBox::containingBlockLogicalWidthForContentInRegion):
(WebCore::RenderBox::containingBlockAvailableLineWidthInRegion):
(WebCore::RenderBox::computeLogicalWidthInRegion):
(WebCore::RenderBox::computeLogicalWidthInRegionUsing):
(WebCore::RenderBox::renderBoxRegionInfo):
(WebCore::RenderBox::containingBlockLogicalWidthForPositioned):
(WebCore::RenderBox::containingBlockLogicalHeightForPositioned):
(WebCore::computeInlineStaticDistance):
(WebCore::RenderBox::computePositionedLogicalWidth):
(WebCore::RenderBox::computePositionedLogicalWidthUsing):
(WebCore::RenderBox::computePositionedLogicalHeight):
(WebCore::RenderBox::computePositionedLogicalHeightUsing):
(WebCore::RenderBox::computePositionedLogicalWidthReplaced):
(WebCore::RenderBox::computePositionedLogicalHeightReplaced):
(WebCore::RenderBox::offsetFromLogicalTopOfFirstPage):

  • rendering/RenderBox.h:
  • rendering/RenderFlexibleBox.cpp:

(WebCore::RenderFlexibleBox::computeMainAxisExtentForChild):

  • rendering/RenderFlowThread.cpp:

(WebCore::RenderFlowThread::logicalWidthChangedInRegionsForBlock):
(WebCore::RenderFlowThread::hasCachedOffsetFromLogicalTopOfFirstRegion):
(WebCore::RenderFlowThread::cachedOffsetFromLogicalTopOfFirstRegion):
(WebCore::RenderFlowThread::setOffsetFromLogicalTopOfFirstRegion):
(WebCore::RenderFlowThread::clearOffsetFromLogicalTopOfFirstRegion):
(WebCore::RenderFlowThread::currentActiveRenderBox):
(WebCore::RenderFlowThread::pushFlowThreadLayoutState):
(WebCore::RenderFlowThread::popFlowThreadLayoutState):
(WebCore::RenderFlowThread::offsetFromLogicalTopOfFirstRegion):
(WebCore::RenderFlowThread::adjustedPositionRelativeToOffsetParent):

  • rendering/RenderFlowThread.h:
  • rendering/RenderTable.cpp:

(WebCore::RenderTable::updateLogicalWidth):

  • rendering/RenderView.cpp:

(WebCore::RenderView::pushLayoutState):
(WebCore::RenderView::pushLayoutStateForCurrentFlowThread):
(WebCore::RenderView::popLayoutStateForCurrentFlowThread):

  • rendering/RenderView.h:

(WebCore::RenderView::pushLayoutState):
(WebCore::RenderView::popLayoutState):

2:21 AM Changeset in webkit [150742] by zarvai@inf.u-szeged.hu
  • 3 edits in trunk/LayoutTests

[Qt] Unreviewed gardening. Greening bots.

  • platform/qt-5.0-wk1/TestExpectations:
  • platform/qt-5.0-wk2/TestExpectations:
2:13 AM Changeset in webkit [150741] by allan.jensen@digia.com
  • 2 edits in trunk/Source/WebCore

Active DOM objects stopped twice
https://bugs.webkit.org/show_bug.cgi?id=116524

Reviewed by Alexey Proskuryakov.

Only iterate over all active DOM object to stop them once.

To ensure later created active DOM objects are stopped,
suspendActiveDOMObjectIfNeeded now also stops active DOM objects
if needed. This also fixes an existing bug when active DOM
objects were created after stopActiveDOMObjects had been called.

Two existing tests creates new active DOM objects after stopActiveDOMObjects
has been called:

fast/dom/xmlhttprequest-constructor-in-detached-document.html
fast/dom/Window/timer-null-script-execution-context.html

  • dom/ScriptExecutionContext.cpp:

(WebCore::ScriptExecutionContext::stopActiveDOMObjects):
(WebCore::ScriptExecutionContext::suspendActiveDOMObjectIfNeeded):

1:34 AM Changeset in webkit [150740] by zandobersek@gmail.com
  • 1 edit
    1 add in trunk/LayoutTests

Unreviewed GTK gardening.

  • platform/gtk/plugins/mouse-events-expected.txt: Added. Rebaselining after r150558.
1:22 AM Changeset in webkit [150739] by zandobersek@gmail.com
  • 3 edits in trunk/LayoutTests

Unreviewed GTK gardening.
Adding failure expectations for the current test failures.

  • platform/gtk-wk1/TestExpectations:
  • platform/gtk/TestExpectations:
1:20 AM Changeset in webkit [150738] by charles.wei@torchmobile.com.cn
  • 3 edits in trunk/Source/WebCore

Remove redundant CachedRawResource::loader()
https://bugs.webkit.org/show_bug.cgi?id=116805

Reviewed by Andreas Kling.

Just remove redundant code, no test case needed.

  • loader/cache/CachedRawResource.cpp:
  • loader/cache/CachedRawResource.h:
1:02 AM WebKitGTK/2.0.x edited by alex
(diff)
12:50 AM Changeset in webkit [150737] by zandobersek@gmail.com
  • 2 edits
    1 add in trunk/LayoutTests

Unreviewed GTK gardening.

  • platform/gtk/fast/dynamic/002-expected.txt: Rebaselining after r150527.
  • platform/gtk/fast/js/global-constructors-attributes-worker-expected.txt: Added. Rebaselining after r150664.
12:20 AM Changeset in webkit [150736] by akling@apple.com
  • 2 edits in trunk/Source/WebCore

WinCE build TLC: Node::active() moved to Element.

  • rendering/RenderThemeWinCE.cpp:

(WebCore::RenderThemeWinCE::paintSearchFieldCancelButton):

May 26, 2013:

11:55 PM Changeset in webkit [150735] by commit-queue@webkit.org
  • 3 edits in trunk/Source/WebKit/blackberry

[BlackBerry] Make screen updates dependent on the existence of a drawing buffer.
https://bugs.webkit.org/show_bug.cgi?id=116539
Internal PR 330917

Patch by Jakob Petsovits <jpetsovits@blackberry.com> on 2013-05-26
Reviewed by Carlos Garcia Campos.
Internally reviewed by Xiaobo Wang and Arvid Nilsson.

This reduces complexity by relying on the current state,
rather than the suspend counter, which in turn makes it
possible to simplify the code in setCompositor().

  • Api/BackingStore.cpp:

(BlackBerry::WebKit::BackingStorePrivate::updateSuspendScreenUpdateState):
(BlackBerry::WebKit::BackingStorePrivate::createSurfaces):

  • Api/WebPage.cpp:

(BlackBerry::WebKit::WebPagePrivate::setCompositor):

7:33 PM Changeset in webkit [150734] by commit-queue@webkit.org
  • 3 edits in trunk/Source/WebCore

CSSParser::parseImageResolution doesn't need a parameter.
https://bugs.webkit.org/show_bug.cgi?id=116755

Patch by Jaehun Lim <ljaehun.lim@samsung.com> on 2013-05-26
Reviewed by Darin Adler.

parseImageResolution() can get CSSPraserValueList from m_valueList directly.

No new tests, no behavior change.

  • css/CSSParser.cpp:

(WebCore::CSSParser::parseImageResolution):

  • css/CSSParser.h:
7:07 PM Changeset in webkit [150733] by akling@apple.com
  • 7 edits in trunk/Source/WebCore

Turn TreeScope::focusedNode() into focusedElement().
<http://webkit.org/b/116802>

Reviewed by Darin Adler.

For the hundredth time, only Elements can be focused!

This change mostly reverts r121079 since that made things unnecessarily complicated
just to tweak the behavior of ShadowRoot.activeElement, an API that we don't expose.
Finding a TreeScope's focused element now does a simple walk up the parent chain instead
of running the full Shadow DOM re-targeting algorithm.

(WebCore::TreeScope::focusedElement):

Simplify and return Element* instead of Node*.

  • dom/TreeScope.h:
  • dom/TreeScope.cpp:

(WebCore::focusedFrameOwnerElement):

Make this return Element* as the name already suggested.

  • dom/Document.cpp:

(WebCore::Document::removeFocusedNodeOfSubtree):

  • dom/Element.cpp:

(WebCore::Element::blur):

  • dom/ShadowRoot.h:

(WebCore::ShadowRoot::activeElement):

  • html/HTMLDocument.cpp:

(WebCore::HTMLDocument::activeElement):

Simplified call sites that were paranoid about getting a non-Element focused node.

6:57 PM Changeset in webkit [150732] by commit-queue@webkit.org
  • 3 edits in trunk/Source/WebKit2

The public WK header file should not include WKAPICast.h
https://bugs.webkit.org/show_bug.cgi?id=115230

Patch by Changhun Kang <temoochin@company100.net> on 2013-05-26
Reviewed by Darin Adler.

Otherwise, WebKit internal headers are exposed.
No new tests. No change in behavior.

  • UIProcess/API/C/soup/WKSoupRequestManager.h:
  • UIProcess/soup/WebSoupRequestManagerClient.h:
6:44 PM Changeset in webkit [150731] by dino@apple.com
  • 3 edits in trunk/Source/WebCore

Plugins with no src but valid type should not snapshot
https://bugs.webkit.org/show_bug.cgi?id=116799
<rdar://problem/13695911>

Reviewed by Simon Fraser.

There are a few plug-ins around that don't take any source content
via the src attribute, but rather just load the binary based on the
mime-type. In those cases we shouldn't snapshot until we have a
mechanism to do so.

  • html/HTMLPlugInImageElement.cpp:

(WebCore::HTMLPlugInImageElement::subframeLoaderWillCreatePlugIn):

Add MIMEType logging, and detect the case of no src but valid type.

  • html/HTMLPlugInImageElement.h: Add another SnapshotReason enum for

this special case.

6:41 PM Changeset in webkit [150730] by Antti Koivisto
  • 8 edits in trunk/Source/WebCore

Encapsulate FontGlyphs
https://bugs.webkit.org/show_bug.cgi?id=116803

Reviewed by Andreas Kling.

FontGlyphs is unfriending Font. Give it a real class interface and some actual code.

  • platform/graphics/Font.cpp:

(WebCore::Font::Font):

  • platform/graphics/Font.h:

(WebCore::Font::glyphDataForCharacter):
(WebCore::Font::glyphDataAndPageForCharacter):

  • platform/graphics/FontFallbackList.cpp:

(WebCore::FontGlyphs::FontGlyphs):
(WebCore::isInRange):
(WebCore::shouldIgnoreRotation):
(WebCore::glyphDataAndPageForNonCJKCharacterWithGlyphOrientation):

Helpers from FontFastPath.cpp.

(WebCore::FontGlyphs::glyphDataAndPageForCharacter):

This is the primary interface. The code moves here from FontFastPath.cpp.

  • platform/graphics/FontFallbackList.h:

(WebCore::FontGlyphs::createForPlatformData):

Remove setPlatformData in favor of constructor.

(WebCore::FontGlyphs::primarySimpleFontData):

  • platform/graphics/FontFastPath.cpp:
  • platform/graphics/GraphicsContext.h:

Remove a stray function definition also called setPlatformFont().

  • rendering/svg/SVGTextRunRenderingContext.cpp:

(WebCore::SVGTextRunRenderingContext::glyphDataForCharacter):

6:25 PM Changeset in webkit [150729] by Patrick Gansterer
  • 2 edits in trunk/Source/WTF

[WINCE] Add wtf_bsearch()
https://bugs.webkit.org/show_bug.cgi?id=116528

Reviewed by Darin Adler.

r149833 introduced usage of ::bsearch(), which does not exist on Windwos CE.
Add our own implementation of this function as wtf_bsearch and define
bsearch as wtf_bsearch to fix compilation on Windwos CE.

  • wtf/StdLibExtras.h:

(wtf_bsearch):

5:00 PM Changeset in webkit [150728] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebCore

Remove a leftover CALENDAR_PICKER from LocaleNone.cpp
https://bugs.webkit.org/show_bug.cgi?id=116800

Patch by Seokju Kwon <Seokju Kwon> on 2013-05-26
Reviewed by Kent Tamura.

No new tests, no behavior change.

  • platform/text/LocaleNone.cpp:

(LocaleNone):

4:39 PM Changeset in webkit [150727] by Antti Koivisto
  • 9 edits in trunk/Source/WebCore

Rename FontFallbackList to FontGlyphs
https://bugs.webkit.org/show_bug.cgi?id=116797

Reviewed by Andreas Kling.

Maybe it was a fallback list once. Nowadays its purpose is to cache realized glyphs and related data.

File renaming will be done separately.

  • WebCore.exp.in:
  • platform/graphics/Font.cpp:

(WebCore::Font::Font):
(WebCore::Font::operator=):
(WebCore::Font::operator==):

Refactored to early return style.

(WebCore::Font::update):
(WebCore::Font::width):

  • platform/graphics/Font.h:

(Font):
(WebCore::Font::glyphs):
(WebCore::Font::loadingCustomFonts):
(WebCore::Font::primaryFont):
(WebCore::Font::fontDataAt):
(WebCore::Font::isFixedPitch):
(WebCore::Font::fontSelector):

  • platform/graphics/FontCache.h:

(FontCache):

  • platform/graphics/FontFallbackList.cpp:

(WebCore::FontGlyphs::FontGlyphs):
(WebCore::FontGlyphs::invalidate):
(WebCore::FontGlyphs::releaseFontData):
(WebCore::FontGlyphs::determinePitch):
(WebCore::FontGlyphs::realizeFontDataAt):

Renamed for clarity along with m_fontList -> m_realizedFontData.

(WebCore::FontGlyphs::setPlatformFont):

  • platform/graphics/FontFallbackList.h:

(FontGlyphs):
(WebCore::FontGlyphs::GlyphPagesStateSaver::GlyphPagesStateSaver):
(WebCore::FontGlyphs::GlyphPagesStateSaver::~GlyphPagesStateSaver):
(GlyphPagesStateSaver):
(WebCore::FontGlyphs::create):
(WebCore::FontGlyphs::~FontGlyphs):
(WebCore::FontGlyphs::primaryFontData):

  • platform/graphics/FontFastPath.cpp:

(WebCore::Font::glyphDataAndPageForCharacter):

  • rendering/svg/SVGTextRunRenderingContext.cpp:

(WebCore::SVGTextRunRenderingContext::glyphDataForCharacter):

4:23 PM Changeset in webkit [150726] by rniwa@webkit.org
  • 2 edits in trunk/Tools

Yet another WKR build fix. Sorry for the mess. It seems like I can't win today :(

  • Scripts/webkitpy/tool/commands/newcommitbot.py:

(NewCommitBot.next_work_item):

3:44 PM Changeset in webkit [150725] by rniwa@webkit.org
  • 3 edits in trunk/Tools

WKR and webkitbot build fixes.

  • Scripts/webkitpy/tool/bot/irc_command.py:

(RollChromiumDEPS.execute):

  • Scripts/webkitpy/tool/commands/newcommitbot.py:

(NewCommitBot.next_work_item):

3:35 PM Changeset in webkit [150724] by tkent@chromium.org
  • 13 edits
    2 deletes in trunk/Source

Remove ENABLE_CALENDAR_PICKER
https://bugs.webkit.org/show_bug.cgi?id=116795

Reviewed by Ryosuke Niwa.

Source/WebCore:

This feature was used only by Chromium port.

  • Resources/pagepopups/calendarPicker.css: Removed.
  • Resources/pagepopups/calendarPicker.js: Removed.
  • css/html.css:
  • platform/text/LocaleICU.cpp:

(WebCore::LocaleICU::LocaleICU):
(WebCore):

  • platform/text/LocaleICU.h:

(LocaleICU):

  • platform/text/PlatformLocale.h:

(Locale):

  • platform/text/mac/LocaleMac.h:

(LocaleMac):

  • platform/text/mac/LocaleMac.mm:

(WebCore::LocaleMac::monthLabels):

  • platform/text/win/LocaleWin.cpp:

(WebCore::LocaleWin::LocaleWin):
(WebCore::LocaleWin::monthLabels):

  • platform/text/win/LocaleWin.h:

(LocaleWin):

Source/WebKit/blackberry:

  • WebCoreSupport/AboutDataEnableFeatures.in:

The list was wrong. Blackberry port doesn't use CALENDAR_PICKER code.

Source/WTF:

  • wtf/FeatureDefines.h:
3:32 PM Changeset in webkit [150723] by rniwa@webkit.org
  • 2 edits in trunk/Tools

roll-chromium-deps doesn't repeat the full name.

Rubber-stamped by Andreas Kling.

  • Scripts/webkitpy/tool/bot/irc_command.py:

(RollChromiumDEPS.execute):

3:23 PM Changeset in webkit [150722] by akling@apple.com
  • 8 edits in trunk/Source/WebCore

Move :active chain participation state from Node to Element.
<http://webkit.org/b/116786>

Reviewed by Antti Koivisto.

Only Elements can be in the :active chain so move the logic there from Node.

  • dom/Document.cpp:

(WebCore::Document::updateHoverActiveState):

Add isElementNode() type checks when updating the :active chain.

  • dom/Node.h:
  • dom/Node.cpp:

(WebCore::Node::detach):

  • dom/Element.cpp:

(WebCore::Element::detach):

Move the remaining logic for detaching from the UserActionElementSet to Element.

(WebCore::Element::isUserActionElementInActiveChain):

  • dom/Element.h:

(WebCore::Element::inActiveChain):

Move all the :active chain stuff from Node to Element.

  • dom/UserActionElementSet.cpp:

(WebCore::UserActionElementSet::didDetach):

  • dom/UserActionElementSet.h:

(WebCore::UserActionElementSet::isInActiveChain):
(WebCore::UserActionElementSet::setInActiveChain):

UserActionElementSet now only takes Element* in its API.

11:44 AM Changeset in webkit [150721] by fpizlo@apple.com
  • 2 edits in branches/dfgFourthTier/Source/JavaScriptCore

fourthTier: cti_optimize shouldn't allow GCs to get in the way of it seeing the state of its CodeBlock
https://bugs.webkit.org/show_bug.cgi?id=116748

Reviewed by Geoffrey Garen.

This fixes the following race: an optimized version of our code block could be installed
by the GC just as we return from completeAllReadyPlansForVM(), leading us to believe
that the code block isn't ready yet even though it is. Currently this triggers a
RELEASE_ASSERT. We could remove that assertion, but then this case would lead to the
code in question entering into optimizeAfterWarmUp mode. That seems pretty wasteful.

Fix the bug, and hopefully close the door on these bugs for a while, by wrapping
cti_optimize in a DeferGC. There is little downside to doing so since the only
"allocations" in cti_optimize are the ones where we inform the GC about extra memory
usage.

I had a more comprehensive solution (see the bug, "work in progress" patch) but that
one involved adding *more* raciness to cti_optimize. I decided that was a less good
approach once I came to appreciate the simplicity of just using DeferGC.

  • jit/JITStubs.cpp:

(JSC::DEFINE_STUB_FUNCTION):

10:42 AM Changeset in webkit [150720] by fpizlo@apple.com
  • 4 edits
    9 adds in branches/dfgFourthTier

Merge trunk r150694.

Source/JavaScriptCore:

2013-05-25 Filip Pizlo <fpizlo@apple.com>


We broke (-231/-1)|0 in the DFG
https://bugs.webkit.org/show_bug.cgi?id=116767


Reviewed by Andreas Kling.


The bug is that we were assuming that in the -231 case, we already had -231
in the result register. This was a wrong assumption.


  • dfg/DFGSpeculativeJIT.cpp: (JSC::DFG::SpeculativeJIT::compileIntegerArithDivForX86):

LayoutTests:

2013-05-25 Filip Pizlo <fpizlo@apple.com>


We broke (-231/-1)|0 in the DFG
https://bugs.webkit.org/show_bug.cgi?id=116767


Reviewed by Andreas Kling.


  • fast/js/dfg-div-by-neg1-and-then-or-zero-interesting-reg-alloc-expected.txt: Added.
  • fast/js/dfg-div-by-neg1-and-then-or-zero-interesting-reg-alloc.html: Added.
  • fast/js/dfg-div-by-zero-and-then-or-zero-interesting-reg-alloc-expected.txt: Added.
  • fast/js/dfg-div-by-zero-and-then-or-zero-interesting-reg-alloc.html: Added.
  • fast/js/dfg-div-neg2tothe31-by-one-and-then-or-zero-with-interesting-reg-alloc-expected.txt: Added.
  • fast/js/dfg-div-neg2tothe31-by-one-and-then-or-zero-with-interesting-reg-alloc.html: Added.
  • fast/js/jsc-test-list:
  • fast/js/script-tests/dfg-div-by-neg1-and-then-or-zero-interesting-reg-alloc.js: Added. (foo):
  • fast/js/script-tests/dfg-div-by-zero-and-then-or-zero-interesting-reg-alloc.js: Added. (foo):
  • fast/js/script-tests/dfg-div-neg2tothe31-by-one-and-then-or-zero-with-interesting-reg-alloc.js: Added. (foo):
9:20 AM Changeset in webkit [150719] by Antti Koivisto
  • 8 edits
    2 deletes in trunk/Source/WebCore

Remove FontFamily.h/.cpp
https://bugs.webkit.org/show_bug.cgi?id=116788

Reviewed by Anders Carlsson.

Empty files.

  • CMakeLists.txt:
  • GNUmakefile.list.am:
  • Target.pri:
  • WebCore.vcproj/WebCore.vcproj:
  • WebCore.vcxproj/WebCore.vcxproj:
  • WebCore.xcodeproj/project.pbxproj:
  • platform/graphics/FontDescription.h:
  • platform/graphics/FontFamily.cpp: Removed.
  • platform/graphics/FontFamily.h: Removed.
8:15 AM Changeset in webkit [150718] by Antti Koivisto
  • 3 edits in trunk/Source/WebKit/win

Another build fix.

Not reviewed.

  • DOMCoreClasses.cpp:

(DOMElement::font):

  • FullscreenVideoController.cpp:

(FullscreenVideoController::draw):

8:11 AM Changeset in webkit [150717] by Antti Koivisto
  • 2 edits in trunk/Source/WebKit/win

Build fix

Not reviewed.

  • WebKitGraphics.cpp:

(makeFont):

7:50 AM Changeset in webkit [150716] by Antti Koivisto
  • 29 edits in trunk/Source/WebCore

Use Vector instead of custom linked list for font families
https://bugs.webkit.org/show_bug.cgi?id=116783

Reviewed by Andreas Kling.

Get rid of FontFamily and SharedFontFamily.

  • accessibility/AccessibilityRenderObject.cpp:

(WebCore::AccessibilityRenderObject::hasSameFont):

  • accessibility/atk/WebKitAccessibleInterfaceText.cpp:

(getAttributeSetForAccessibilityObject):

  • css/CSSComputedStyleDeclaration.cpp:

(WebCore::fontFamilyFromStyle):

  • css/DeprecatedStyleBuilder.cpp:

(WebCore::ApplyPropertyFontFamily::applyInheritValue):
(WebCore::ApplyPropertyFontFamily::applyInitialValue):
(WebCore::ApplyPropertyFontFamily::applyValue):

  • css/StyleResolver.cpp:

(WebCore::StyleResolver::styleForDocument):
(WebCore::StyleResolver::initializeFontStyle):

  • html/canvas/CanvasRenderingContext2D.cpp:

(WebCore::CanvasRenderingContext2D::font):
(WebCore::CanvasRenderingContext2D::setFont):

  • page/PrintContext.cpp:

(WebCore::PrintContext::pageProperty):

  • platform/efl/RenderThemeEfl.cpp:

(WebCore::RenderThemeEfl::systemFont):

  • platform/graphics/Font.h:

(WebCore::Font::firstFamily):
(WebCore::Font::familyCount):
(WebCore::Font::familyAt):

  • platform/graphics/FontCache.cpp:

(WebCore::FontCache::getFontData):

  • platform/graphics/FontDescription.cpp:

(SameSizeAsFontDescription):

  • platform/graphics/FontDescription.h:

(WebCore::FontDescription::FontDescription):
(WebCore::FontDescription::familyCount):
(WebCore::FontDescription::firstFamily):
(WebCore::FontDescription::familyAt):
(WebCore::FontDescription::families):
(WebCore::FontDescription::useFixedDefaultSize):
(WebCore::FontDescription::setOneFamily):
(WebCore::FontDescription::setFamilies):
(WebCore::FontDescription::adoptFamilies):
(WebCore::FontDescription::operator==):

  • platform/graphics/FontFamily.cpp:
  • platform/graphics/FontFamily.h:
  • platform/graphics/ca/win/PlatformCALayerWinInternal.cpp:

(PlatformCALayerWinInternal::displayCallback):

  • platform/graphics/mac/FontCacheMac.mm:

(WebCore::FontCache::getSimilarFontPlatformData):

  • platform/graphics/qt/FontCacheQt.cpp:

(WebCore::FontCache::getLastResortFallbackFont):

  • platform/gtk/RenderThemeGtk.cpp:

(WebCore::RenderThemeGtk::systemFont):

  • platform/mac/ThemeMac.mm:

(WebCore::ThemeMac::controlFont):

  • platform/qt/RenderThemeQStyle.cpp:

(WebCore::RenderThemeQStyle::adjustButtonStyle):

  • platform/text/transcoder/FontTranscoder.cpp:

(WebCore::FontTranscoder::converterType):

  • platform/win/DragImageWin.cpp:

(WebCore::dragLabelFont):

  • rendering/RenderText.cpp:

(WebCore::RenderText::styleDidChange):

  • rendering/RenderTextControl.cpp:

(WebCore::RenderTextControl::computeIntrinsicLogicalWidths):

  • rendering/RenderTextControlSingleLine.cpp:

(WebCore::RenderTextControlSingleLine::preferredContentLogicalWidth):

  • rendering/RenderThemeMac.mm:

(WebCore::RenderThemeMac::systemFont):
(WebCore::RenderThemeMac::setFontFromControlSize):

  • rendering/RenderThemeSafari.cpp:

(WebCore::RenderThemeSafari::systemFont):
(WebCore::RenderThemeSafari::setFontFromControlSize):

  • rendering/RenderThemeWin.cpp:

(WebCore::fillFontDescription):

7:29 AM Changeset in webkit [150715] by akling@apple.com
  • 17 edits in trunk/Source/WebCore

Move "active" state logic from Node to Element.
<http://webkit.org/b/116785>

Reviewed by Antti Koivisto.

Resolve another 7 year old FIXME by merging the "active" state logic from Node and ContainerNode
and moving it all to Element instead.

  • dom/UserActionElementSet.h:

(WebCore::UserActionElementSet::isActive):
(WebCore::UserActionElementSet::setActive):

  • dom/ContainerNode.cpp:
  • dom/ContainerNode.h:
  • dom/Node.cpp:
  • dom/Node.h:
  • dom/Element.h:

(WebCore::Element::active):

  • dom/Element.cpp:

(WebCore::Element::isUserActionElementActive):
(WebCore::Element::setActive):

Moved here from Node/ContainerNode.

  • accessibility/AccessibilityNodeObject.cpp:

(WebCore::AccessibilityNodeObject::isPressed):

  • dom/Document.cpp:

(WebCore::Document::setFocusedNode):
(WebCore::Document::updateHoverActiveState):

  • rendering/RenderTheme.cpp:

(WebCore::RenderTheme::isPressed):
(WebCore::RenderTheme::isSpinUpButtonPartPressed):

  • rendering/RenderThemeMac.mm:

(WebCore::getMediaUIPartStateFlags):
(WebCore::RenderThemeMac::updatePressedState):

  • platform/qt/RenderThemeQt.cpp:

(WebCore::RenderThemeQt::getMediaControlForegroundColor):

Check that the inspected Node is an Element before asking if it's active.

  • html/HTMLAnchorElement.cpp:

(WebCore::HTMLAnchorElement::setActive):

Call the right superclass.

  • html/HTMLAnchorElement.h:
  • html/HTMLLabelElement.h:
  • html/shadow/MediaControlElementTypes.h:

Sprinkle OVERRIDE>

6:32 AM Changeset in webkit [150714] by akling@apple.com
  • 7 edits in trunk/Source/WebCore

Move Node::dispatchSimulatedClick() to Element.
<http://webkit.org/b/116784>

Reviewed by Antti Koivisto.

Only Elements use the dispatchSimulatedClick() functionality, so move it there.

  • dom/Node.cpp:
  • dom/Node.h:
  • dom/Element.h:
  • dom/Element.cpp:

(WebCore::Element::dispatchSimulatedClick):

  • dom/EventDispatcher.h:
  • dom/EventDispatcher.cpp:

(WebCore::EventDispatcher::dispatchSimulatedClick):

5:10 AM Changeset in webkit [150713] by akling@apple.com
  • 8 edits in trunk/Source/WebCore

TreeScope::rootNode() should return a ContainerNode.
<http://webkit.org/b/116782>

Reviewed by Antti Koivisto.

The rootNode() of a TreeScope is always a ContainerNode, so update pointer types to reflect that.
This lets us take advantage of Antti's optimized traversal functions.

  • dom/ContainerNode.h:

(WebCore::Node::isTreeScope):

  • dom/Document.cpp:

(WebCore::Document::buildAccessKeyMap):

  • dom/Node.cpp:

(WebCore::Node::containingShadowRoot):

  • dom/Node.h:
  • dom/TreeScope.h:

(WebCore::TreeScope::rootNode):
(TreeScope):

  • page/FocusController.cpp:

(WebCore::FocusNavigationScope::rootNode):
(WebCore::FocusNavigationScope::owner):

  • page/FocusController.h:

(FocusNavigationScope):

4:39 AM Changeset in webkit [150712] by akling@apple.com
  • 18 edits in trunk/Source

FocusController::setFocusedNode() should be setFocusedElement().
<http://webkit.org/b/116780>

Reviewed by Antti Koivisto.

Source/WebCore:

Only Elements can have focus, so make the public API of FocusController take Element*.

  • WebCore.exp.in:
  • page/FocusController.h:
  • page/FocusController.cpp:

(WebCore::FocusController::setFocusedElement):

Out with the setFocusedNode(), in with the setFocusedElement().

  • dom/Element.cpp:

(WebCore::Element::focus):
(WebCore::Element::blur):

  • page/EventHandler.cpp:

(WebCore::EventHandler::dispatchMouseEvent):

  • plugins/PluginView.cpp:

(WebCore::PluginView::focusPluginElement):

  • editing/FrameSelection.cpp:

(WebCore::FrameSelection::setFocusedNodeIfNeeded):

Tweak for FocusController taking Element* instead of Node*.

(WebCore::FrameSelection::setFocusedElementIfNeeded):

Renamed from setFocusedNodeIfNeeded().

  • editing/htmlediting.h:
  • editing/htmlediting.cpp:

(WebCore::enclosingAnchorElement):

Cleaned up the loop for readability and made it return an Element.

Source/WebKit/blackberry:

  • Api/WebPage.cpp:

(BlackBerry::WebKit::WebPagePrivate::clearFocusNode):
(BlackBerry::WebKit::WebPage::setNodeFocus):

  • WebKitSupport/SelectionHandler.cpp:

(BlackBerry::WebKit::SelectionHandler::selectObject):

Source/WebKit/win:

  • WebView.cpp:

(WebView::clearFocusNode):

Source/WebKit2:

  • WebProcess/Plugins/PluginView.cpp:

(WebKit::PluginView::focusPluginElement):

2:43 AM Changeset in webkit [150711] by akling@apple.com
  • 7 edits in trunk/Source/WebCore

Move some form control things from Element to HTMLFormControlElement.
<http://webkit.org/b/116779>

Reviewed by Antti Koivisto.

The following methods sit better on HTMLFormControlElement than on Element:

  • formControlType()
  • wasChangedSinceLastFormControlChangeEvent()
  • setChangedSinceLastFormControlChangeEvent()
  • dispatchFormControlChangeEvent()

Moved them, and devirtualized as much as possible.

  • dom/Document.cpp:

(WebCore::Document::setFocusedNode):

Check if the old focus target is an Element before (possibly) dispatching a change event.

  • dom/Element.cpp:
  • dom/Element.h:

Removed some stub methods that were only used by HTMLFormControlElement subclasses.

  • html/HTMLFormControlElement.cpp:
  • html/HTMLFormControlElement.h:

(WebCore::isHTMLFormControlElement):
(WebCore::toHTMLFormControlElement):

Add toHTMLFormControlElement() helper with ASSERT_WITH_SECURITY_IMPLICATION.

  • html/HTMLTextFormControlElement.h:

Sprinkle OVERRIDE.

2:39 AM Changeset in webkit [150710] by akling@apple.com
  • 30 edits in trunk/Source/WebCore

Move Node::supportsFocus() to Element.
<http://webkit.org/b/116778>

Reviewed by Antti Koivisto.

Node never supports focus, so move supportsFocus() to Element instead.

  • dom/Element.h:
  • dom/Node.cpp:
  • dom/Node.h:

Removed Node::supportsFocus(), it always returned false anyway.

  • accessibility/AccessibilityNodeObject.cpp:

(WebCore::AccessibilityNodeObject::canSetFocusAttribute):

Check that the underlying Node is an Element before asking if it supports focus.

  • html/HTMLAnchorElement.h:
  • html/HTMLAreaElement.h:
  • html/HTMLBodyElement.h:
  • html/HTMLElement.h:
  • html/HTMLFieldSetElement.h:
  • html/HTMLFormControlElement.h:
  • html/HTMLFrameElementBase.h:
  • html/HTMLMediaElement.h:
  • html/HTMLOptGroupElement.h:
  • html/HTMLOptionElement.h:
  • html/HTMLOutputElement.h:
  • svg/SVGAElement.h:
  • svg/SVGCircleElement.h:
  • svg/SVGEllipseElement.h:
  • svg/SVGGElement.h:
  • svg/SVGImageElement.h:
  • svg/SVGLineElement.h:
  • svg/SVGPathElement.h:
  • svg/SVGPolyElement.h:
  • svg/SVGRectElement.h:
  • svg/SVGSVGElement.h:
  • svg/SVGSwitchElement.h:
  • svg/SVGSymbolElement.h:
  • svg/SVGTextElement.h:
  • svg/SVGUseElement.h:

OVERRIDE ALL THE THINGS!

May 25, 2013:

10:05 PM Changeset in webkit [150709] by akling@apple.com
  • 14 edits in trunk/Source/WebCore

Move Node::isFocusable() to Element.
<http://webkit.org/b/116777>

Reviewed by Anders Carlsson.

Nodes cannot be focusable, so move isFocusable() from Node to Element.

  • dom/Node.cpp:
  • dom/Node.h:
  • dom/Element.h:
  • dom/Element.cpp:

(WebCore::Element::isFocusable):

Moved here from Node.

  • dom/Document.cpp:

(WebCore::Document::setFocusedNode):

  • accessibility/AccessibilityNodeObject.cpp:

(WebCore::AccessibilityNodeObject::determineAccessibilityRole):

Check that the underlying node is an Element before asking if it's focusable.

  • page/FrameView.cpp:

(WebCore::FrameView::scrollToAnchor):

Renamed the 'anchorNode' variable to 'anchorElement' because reasons.

  • html/HTMLAreaElement.h:
  • html/HTMLFormControlElement.h:
  • html/HTMLLabelElement.h:
  • html/HTMLOptionElement.h:
  • svg/SVGAElement.h:

Sprinkle OVERRIDE.

  • html/ValidationMessage.cpp:

(WebCore::ValidationMessage::setMessage):

Update a comment to refer to Element::isFocusable() instead of Node.

9:42 PM Changeset in webkit [150708] by commit-queue@webkit.org
  • 2 edits in trunk/Websites/webkit.org

Broken link to committers list on contributing web page
https://bugs.webkit.org/show_bug.cgi?id=116722

Patch by Bem Jones-Bey <Bem Jones-Bey> on 2013-05-25
Reviewed by Andreas Kling.

  • coding/contributing.html:

Changed broken link to point to the team page instead.

9:33 PM Changeset in webkit [150707] by akling@apple.com
  • 9 edits in trunk/Source/WebCore

Move Node::tabIndex() to Element.
<http://webkit.org/b/116772>

Reviewed by Ryosuke Niwa.

Since only Elements are keyboard-focusable, it doesn't make sense for Node to have a tabIndex().

  • dom/Element.h:
  • dom/Node.cpp:
  • dom/Node.h:

A Node can't have a tab index, so move tabIndex() to Element.

  • page/FocusController.h:
  • page/FocusController.cpp:

(WebCore::adjustedTabIndex):
(WebCore::FocusController::findElementWithExactTabIndex):
(WebCore::nextElementWithGreaterTabIndex):
(WebCore::previousElementWithLowerTabIndex):
(WebCore::FocusController::nextFocusableNode):
(WebCore::FocusController::previousFocusableNode):

Make this code deal in Element* when doing tab index stuff. FocusController needs
more Node->Element cleanup, but let's do that separately.

  • html/HTMLAnchorElement.h:
  • html/HTMLElement.h:
  • html/HTMLFormControlElement.h:

Sprinkle OVERRIDE.

9:31 PM Changeset in webkit [150706] by rniwa@webkit.org
  • 2 edits in trunk/Tools

Revert another erroneous change.

  • Scripts/webkitpy/tool/commands/newcommitbot.py:

(NewCommitBot.next_work_item):

9:13 PM Changeset in webkit [150705] by rniwa@webkit.org
  • 2 edits in trunk/Tools

Revert an erroneous change.

  • Scripts/webkitpy/tool/commands/sheriffbot.py:

(SheriffBot):

8:39 PM Changeset in webkit [150704] by rniwa@webkit.org
  • 4 edits in trunk/Tools

Some build fixes for WKR and webkitbot.

  • Scripts/webkitpy/tool/bot/ircbot_unittest.py:

(IRCBotTest.test_help):

  • Scripts/webkitpy/tool/commands/newcommitbot.py:

(Agent):
(Agent.init):
(Agent.name):
(NewCommitBot):
(NewCommitBot.begin_work_queue):
(NewCommitBot.next_work_item):

  • Scripts/webkitpy/tool/commands/sheriffbot.py:

(SheriffBot):

8:35 PM Changeset in webkit [150703] by fpizlo@apple.com
  • 7 edits in branches/dfgFourthTier/Source/JavaScriptCore

fourthTier: FTL should support ArithDiv
https://bugs.webkit.org/show_bug.cgi?id=116771

Reviewed by Oliver Hunt.

  • ftl/FTLAbbreviations.h:

(JSC::FTL::buildDiv):
(JSC::FTL::buildRem):
(JSC::FTL::buildFDiv):

  • ftl/FTLCapabilities.cpp:

(JSC::FTL::canCompile):

  • ftl/FTLCommonValues.cpp:

(JSC::FTL::CommonValues::CommonValues):

  • ftl/FTLCommonValues.h:

(CommonValues):

  • ftl/FTLLowerDFGToLLVM.cpp:

(JSC::FTL::LowerDFGToLLVM::compileNode):
(JSC::FTL::LowerDFGToLLVM::compileArithMul):
(JSC::FTL::LowerDFGToLLVM::compileArithDiv):
(LowerDFGToLLVM):

  • ftl/FTLOutput.h:

(JSC::FTL::Output::div):
(JSC::FTL::Output::rem):
(JSC::FTL::Output::doubleDiv):

8:29 PM Changeset in webkit [150702] by akling@apple.com
  • 2 edits in trunk/Source/WebCore

Don't instantiate ChildNodeLists just to get a child count.
<http://webkit.org/b/116776>

Reviewed by Ryosuke Niwa.

Use ContainerNode::childNodeCount() instead of Node::childNodes()->length().

  • editing/ApplyStyleCommand.cpp:

(WebCore::ApplyStyleCommand::mergeEndWithNextIfIdentical):

7:57 PM Changeset in webkit [150701] by rniwa@webkit.org
  • 3 edits in trunk/Tools

Teach webkitbot and WKR how to respond to yt?
https://bugs.webkit.org/show_bug.cgi?id=116775

Reviewed by Andreas Kling.

  • Scripts/webkitpy/tool/bot/irc_command.py:

(Hi): Fixed the usage string.
(Hi.execute): Made it more generic so that it works in WKR.
(YouThere): Added.
(YouThere.execute): Added.

  • Scripts/webkitpy/tool/commands/newcommitbot.py:
7:48 PM Changeset in webkit [150700] by akling@apple.com
  • 7 edits in trunk/Source

PluginDocument::pluginNode() should be pluginElement() instead.
<http://webkit.org/b/116774>

Reviewed by Anders Carlsson.

Source/WebCore:

Make PluginDocument hand out a HTMLPlugInElement* instead of a Node*.

  • dom/Document.cpp:

(WebCore::eventTargetNodeForDocument):

  • html/PluginDocument.h:
  • html/PluginDocument.cpp:

(WebCore::PluginDocumentParser::createDocumentStructure):
(WebCore::PluginDocument::pluginWidget):
(WebCore::PluginDocument::setPluginElement):
(WebCore::PluginDocument::detach):

Source/WebKit2:

  • WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::focusedPluginViewForFrame):

7:27 PM Changeset in webkit [150699] by akling@apple.com
  • 7 edits in trunk/Source

Remove Document::getFocusableNodes().
<http://webkit.org/b/116773>

Reviewed by Joseph Pecoraro.

Source/WebCore:

This was added in r42191 and there are no remaining clients of the API.

  • dom/Document.cpp:
  • dom/Document.h:
  • WebCore.exp.in:

Source/WebKit/mac:

Remove the private _focusableNodes API on DOMDocument, it has no clients.

  • DOM/WebDOMOperations.mm:
  • DOM/WebDOMOperationsPrivate.h:
7:22 PM Changeset in webkit [150698] by Simon Fraser
  • 2 edits in trunk/Tools

Try to fix the EFL build.

  • DumpRenderTree/efl/TestRunnerEfl.cpp:

(TestRunner::setPageVisibility):

7:22 PM Changeset in webkit [150697] by akling@apple.com
  • 9 edits in trunk/Source/WebCore

Move Node::focusDelegate() to Element.
<http://webkit.org/b/116768>

Reviewed by Anders Carlsson.

Only Element can be a focus delegate (it's currently only ever HTMLInputElement actually)
so move Node::focusDelegate() to Element. Sprinkled isElementNode() checks as needed.

  • dom/Node.cpp:
  • dom/Node.h:
  • dom/Element.h:
  • dom/Element.cpp:
  • html/shadow/SliderThumbElement.cpp:

(WebCore::SliderThumbElement::focusDelegate):

  • html/shadow/SliderThumbElement.h:
  • rendering/RenderTheme.cpp:

(WebCore::RenderTheme::isFocused):

  • rendering/RenderThemeMac.mm:

(WebCore::RenderThemeMac::paintSliderThumb):

5:51 PM Changeset in webkit [150696] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebCore

Path: clean up addPathForRoundedRect()
https://bugs.webkit.org/show_bug.cgi?id=116765

Patch by Alberto Garcia <agarcia@igalia.com> on 2013-05-25
Reviewed by Andreas Kling.

Reorganize code that is a bit redundant.

No new tests, no behavior change.

  • platform/graphics/Path.cpp:

(WebCore::Path::addPathForRoundedRect):

4:56 PM Changeset in webkit [150695] by Simon Fraser
  • 35 edits in trunk

Unprefix Page Visibility API
https://bugs.webkit.org/show_bug.cgi?id=102340

Source/WebCore:

Reviewed by Tim Horton.

Remove prefixes from the document properties, and the visibilitychanged event.

Rename the "preview" state to "unloaded". The "unloaded" state is never
reached yet (see https://bugs.webkit.org/show_bug.cgi?id=116769).

Add some comments to places where we should update the visibility state
when going in and out of the page cache, but cannot safely do so yet.

Tested by existing Page Visibility tests.

  • dom/Document.cpp:

(WebCore::Document::pageVisibilityState):
(WebCore::Document::visibilityState):
(WebCore::Document::hidden):
(WebCore::Document::dispatchVisibilityStateChangeEvent):

  • dom/Document.h:
  • dom/Document.idl:
  • dom/EventNames.h:
  • history/CachedFrame.cpp:

(WebCore::CachedFrameBase::restore):

  • loader/FrameLoader.cpp:

(WebCore::FrameLoader::stopLoading):

  • page/Page.cpp:

(WebCore::Page::setVisibilityState):

  • page/PageVisibilityState.cpp:

(WebCore::pageVisibilityStateString):

  • page/PageVisibilityState.h:

Source/WebKit/efl:

Reviewed by Tim Horton.

Rename "preview" state to "unloaded" state.

  • WebCoreSupport/AssertMatchingEnums.cpp:
  • ewk/ewk_view.h:

Source/WebKit/mac:

Reviewed by Tim Horton.

Renaming of the "preview" state to "unloaded".

  • WebView/WebView.mm:
  • WebView/WebViewPrivate.h:

Source/WebKit2:

Reviewed by Tim Horton.

Renaming of the "preview" state to "unloaded".

  • Shared/API/c/WKPageVisibilityTypes.h:
  • Shared/API/c/WKSharedAPICast.h:

(WebKit::toPageVisibilityState):

Tools:

Reviewed by Tim Horton.

Rename "preview" state to "unloaded" state, and adjust the tests
to test for it.

  • DumpRenderTree/mac/TestRunnerMac.mm:

(TestRunner::setPageVisibility):

  • TestWebKitAPI/Tests/WebKit2/PageVisibilityState.cpp:

(TestWebKitAPI::setPageVisibilityStateWithEvalContinuation):
(TestWebKitAPI::didRunStep2StateChangeHiddenToPrerender):
(TestWebKitAPI::didRunStep3StateChangePrerenderToUnloaded):
(TestWebKitAPI::didRunStep4InStateUnloaded):

  • TestWebKitAPI/Tests/mac/PageVisibilityStateWithWindowChanges.html:
  • TestWebKitAPI/Tests/mac/PageVisibilityStateWithWindowChanges.mm:

(TestWebKitAPI::PageVisibilityStateWithWindowChanges::runTest):

  • WebKitTestRunner/InjectedBundle/TestRunner.cpp:

(WTR::TestRunner::setPageVisibility):

LayoutTests:

Reviewed by Tim Horton.

Remove prefixes, adjust to state name changes.

  • fast/events/page-visibility-iframe-delete-test.html:
  • fast/events/page-visibility-iframe-move-test-expected.txt:
  • fast/events/page-visibility-iframe-move-test.html:
  • fast/events/page-visibility-iframe-propagation-test-expected.txt:
  • fast/events/page-visibility-iframe-propagation-test.html:
  • fast/events/page-visibility-null-view-expected.txt:
  • fast/events/page-visibility-null-view.html:
  • fast/events/page-visibility-transition-test-expected.txt:
  • fast/events/page-visibility-transition-test.html:
4:31 PM Changeset in webkit [150694] by fpizlo@apple.com
  • 4 edits
    9 adds in trunk

We broke (-231/-1)|0 in the DFG
https://bugs.webkit.org/show_bug.cgi?id=116767

Source/JavaScriptCore:

Reviewed by Andreas Kling.

The bug is that we were assuming that in the -231 case, we already had -231
in the result register. This was a wrong assumption.

  • dfg/DFGSpeculativeJIT.cpp:

(JSC::DFG::SpeculativeJIT::compileIntegerArithDivForX86):

LayoutTests:

Reviewed by Andreas Kling.

  • fast/js/dfg-div-by-neg1-and-then-or-zero-interesting-reg-alloc-expected.txt: Added.
  • fast/js/dfg-div-by-neg1-and-then-or-zero-interesting-reg-alloc.html: Added.
  • fast/js/dfg-div-by-zero-and-then-or-zero-interesting-reg-alloc-expected.txt: Added.
  • fast/js/dfg-div-by-zero-and-then-or-zero-interesting-reg-alloc.html: Added.
  • fast/js/dfg-div-neg2tothe31-by-one-and-then-or-zero-with-interesting-reg-alloc-expected.txt: Added.
  • fast/js/dfg-div-neg2tothe31-by-one-and-then-or-zero-with-interesting-reg-alloc.html: Added.
  • fast/js/jsc-test-list:
  • fast/js/script-tests/dfg-div-by-neg1-and-then-or-zero-interesting-reg-alloc.js: Added.

(foo):

  • fast/js/script-tests/dfg-div-by-zero-and-then-or-zero-interesting-reg-alloc.js: Added.

(foo):

  • fast/js/script-tests/dfg-div-neg2tothe31-by-one-and-then-or-zero-with-interesting-reg-alloc.js: Added.

(foo):

4:13 PM Changeset in webkit [150693] by mitz@apple.com
  • 2 edits in trunk/LayoutTests

Updated expected results after r150688.

Reviewed by Andreas Kling.

  • http/tests/globalhistory/history-delegate-basic-refresh-redirect-expected.txt: Now shows

the URL that performed the client redirect, as expected.

4:09 PM Changeset in webkit [150692] by akling@apple.com
  • 21 edits in trunk/Source

Move Node::isMouseFocusable() to Element.
<http://webkit.org/b/116762>

Reviewed by Anders Carlsson.

Source/WebCore:

Node::isMouseFocusable() would just return isFocusable(), which is never true for a non-Element
since Node::supportsFocus() always returns false. So move it to Element!

  • dom/Node.h:
  • dom/Node.cpp:
  • dom/Element.h:
  • dom/Element.cpp:

(WebCore::Element::isMouseFocusable):

Moved here from Node.

  • editing/FrameSelection.cpp:

(WebCore::FrameSelection::setFocusedNodeIfNeeded):

  • page/EventHandler.cpp:

(WebCore::EventHandler::dispatchMouseEvent):

Walk up the parent chain with parentOrShadowHostElement() instead of parentOrShadowHostNode().
Removed a misleading no-op hunk about mouse-focusable ShadowRoots, since ShadowRoots are not
Elements and thus cannot be mouse-focusable.

  • page/TouchAdjustment.cpp:

(WebCore::TouchAdjustment::nodeRespondsToTapGesture):

Check that the Node is an Element before asking if it's mouse-focusable.

  • html/HTMLAnchorElement.h:
  • html/HTMLAreaElement.h:
  • html/HTMLFormControlElement.h:
  • html/HTMLInputElement.h:
  • html/HTMLMediaElement.h:
  • html/HTMLSelectElement.h:
  • html/HTMLTextAreaElement.h:
  • html/shadow/ClearButtonElement.h:
  • html/shadow/TextControlInnerElements.h:
  • svg/SVGAElement.h:
  • svg/SVGStyledElement.h:

Sprinkle OVERRIDE.

Source/WebKit2:

  • WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::highlightPotentialActivation):

Check that the potentially activated Node is an Element before asking if it's mouse-focusable.

2:34 PM Changeset in webkit [150691] by akling@apple.com
  • 2 edits in trunk/Source/WebKit2

REGRESSION: WebProcess is terminated when all Safari windows are closed.
<rdar://problem/13990901>
<http://webkit.org/b/116766>

Reviewed by Alexey Proskuryakov.

When closing the last page, only disconnect the web process if we're using a network process.
Otherwise there might be session state in the web process getting lost.

  • UIProcess/WebProcessProxy.cpp:

(WebKit::WebProcessProxy::removeWebPage):

1:17 PM Changeset in webkit [150690] by mark.lam@apple.com
  • 10 edits in branches/dfgFourthTier/Source

Remove Interpreter::retrieveLastCaller().
https://bugs.webkit.org/show_bug.cgi?id=116753.

Reviewed by Geoffrey Garen.

This is part of the refactoring effort to get rid of functions walking
the JS stack in their own way.

../JavaScriptCore:

  • API/JSContextRef.cpp:

(JSContextCreateBacktrace):

  • interpreter/CallFrame.cpp:
  • interpreter/Interpreter.cpp:

(JSC::Interpreter::Interpreter):
(JSC::Interpreter::getStackTrace):
(JSC::Interpreter::addStackTraceIfNecessary):

  • interpreter/Interpreter.h:

(StackFrame):
(JSC::StackFrame::StackFrame):
(Interpreter):

  • jsc.cpp:

(functionJSCStack):

  • profiler/ProfileGenerator.cpp:

(JSC::ProfileGenerator::addParentForConsoleStart):

../WebCore:

No new tests.

  • bindings/js/JSXMLHttpRequestCustom.cpp:

(WebCore::JSXMLHttpRequest::send):

  • bindings/js/ScriptCallStackFactory.cpp:

(WebCore::createScriptCallStack):

11:35 AM Changeset in webkit [150689] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebKit2

WebFrameProxy::didFailLoad clears frame title for no apparent reason
https://bugs.webkit.org/show_bug.cgi?id=116655

Patch by Santosh Mahto <santosh.ma@samsung.com> on 2013-05-25
Reviewed by Alexey Proskuryakov.

  • UIProcess/WebFrameProxy.cpp:

(WebKit::WebFrameProxy::didFailLoad):

10:57 AM Changeset in webkit [150688] by mitz@apple.com
  • 3 edits in trunk/Source/WebKit/mac

<rdar://problem/13924889> -webView:didPerformClientRedirectFromURL:toURL:inFrame: should be given the redirecting URL as the “from” URL
https://bugs.webkit.org/show_bug.cgi?id=116733

Reviewed by Adele Peterson.

By the time the history controller calls updateGlobalHistoryRedirectLinks(), the document
loader already reflects the destination URL. In order to be able to provide the immediate
source URL to the delegate, we make WebFrame track a URL and provisional URL, updating them
in response to FrameLoader callbacks in exactly the same way WebFrameProxy maintains its
m_url and m_provisionalURL member variables.

  • WebCoreSupport/WebFrameLoaderClient.mm:

(WebFrameLoaderClient::dispatchDidReceiveServerRedirectForProvisionalLoad): Update
provisionalURL.
(WebFrameLoaderClient::dispatchDidChangeLocationWithinPage): Update url.
(WebFrameLoaderClient::dispatchDidPushStateWithinPage): Ditto.
(WebFrameLoaderClient::dispatchDidReplaceStateWithinPage): Ditto.
(WebFrameLoaderClient::dispatchDidPopStateWithinPage): Ditto.
(WebFrameLoaderClient::dispatchDidStartProvisionalLoad): Update provisionalURL.
(WebFrameLoaderClient::dispatchDidCommitLoad): Update URL and provisionalURL.
(WebFrameLoaderClient::dispatchDidFailProvisionalLoad): Clear provisionalURL.
(WebFrameLoaderClient::dispatchDidFailLoad): Added an assertion.
(WebFrameLoaderClient::dispatchDidFinishLoad): Ditto.
(WebFrameLoaderClient::updateGlobalHistoryRedirectLinks): Pass the last committed URL as the
source, rather than clientRedirectSourceForHistory().

  • WebView/WebFrameInternal.h: Added url and provisionalURL ivars to WebFramePrivate.
5:02 AM Changeset in webkit [150687] by akling@apple.com
  • 16 edits in trunk/Source/WebCore

Move Node::isKeyboardFocusable() to Element.
<http://webkit.org/b/116761>

Reviewed by Antti Koivisto.

Only Elements can be keyboard-focusable. Move the base isKeyboardFocusable() implementation to Element
and sprinkle OVERRIDE on subclass overrides.
FocusController gets some temporary type checks that will go away as that code moves to dealing in
Elements instead of Node.

  • dom/Node.h:
  • dom/Node.cpp:
  • dom/Element.h:
  • dom/Element.cpp:

(WebCore::Element::isKeyboardFocusable):

Move isKeyboardFocusable() from Node to Element.

  • page/FocusController.cpp:

(WebCore::isNonFocusableShadowHost):
(WebCore::isFocusableShadowHost):
(WebCore::shouldVisit):
(WebCore::FocusController::advanceFocusInDocumentOrder):

Check that the inspected Node is an Element before querying isKeyboardFocusable().

  • html/HTMLAnchorElement.h:
  • html/HTMLAreaElement.h:
  • html/HTMLFormControlElement.h:
  • html/HTMLFrameOwnerElement.h:
  • html/HTMLInputElement.h:
  • html/HTMLPlugInElement.h:
  • html/HTMLSelectElement.h:
  • html/HTMLTextAreaElement.h:
  • svg/SVGAElement.h:
  • svg/SVGStyledElement.h:

Sprinkle OVERRIDE.

3:27 AM Changeset in webkit [150686] by akling@apple.com
  • 10 edits in trunk/Source/WebCore

Begin moving "focus" state logic from Node to Element.
<http://webkit.org/b/116760>

Reviewed by Antti Koivisto.

Only Elements can be focused, so merge the "focus" state logic from Node and ContainerNode
and move it into Element. There's a lot more iceberg under this tip, but we'll be starting
with setFocus(bool) and focused().

  • dom/Node.h:
  • dom/ContainerNode.cpp:
  • dom/ContainerNode.h:
  • dom/Element.cpp:

(WebCore::Element::isUserActionElementFocused):
(WebCore::Element::setFocus):

  • dom/Element.h:

(WebCore::Element::focused):

  • dom/Node.cpp:

Move focused() and setFocus(bool) from Node/ContainerNode to Element.

  • dom/Document.cpp:

(WebCore::Document::setFocusedNode):

  • page/EventHandler.cpp:

(WebCore::EventHandler::dispatchMouseEvent):

Check if the inspected Node is an Element before asking if it's focused.

2:51 AM Changeset in webkit [150685] by noam.rosenthal@nokia.com
  • 3 edits
    2 adds in trunk

Direct pattern compositing breaks when no-repeat is set on a large layer
https://bugs.webkit.org/show_bug.cgi?id=116140

Source/WebCore:

Reviewed by Antti Koivisto.

In some cases the directly composited background color logic overrides the layer's
contentsRect, which results in wrong rendering of tiled images.
This patch makes sure that the background image is applied after the solid color,
which ensures the correct contentsRect.

Test: compositing/patterns/direct-pattern-compositing-contain-no-repeat.html

  • rendering/RenderLayerBacking.cpp:

(WebCore::RenderLayerBacking::updateDirectlyCompositedContents):

LayoutTests:

Added a ref-test to make sure that direct pattern compositing works well with no-repeat
in a particular geometry.

Reviewed by Antti Koivisto.

  • compositing/patterns/direct-pattern-compositing-contain-no-repeat-expected.html: Added.
  • compositing/patterns/direct-pattern-compositing-contain-no-repeat.html: Added.
2:16 AM Changeset in webkit [150684] by akling@apple.com
  • 14 edits in trunk/Source/WebCore

Move "hover" state logic from Node to Element.
<http://webkit.org/b/116757>

Reviewed by Antti Koivisto.

Resolve a 7 year old FIXME by merging the "hover" state logic from Node and ContainerNode
and moving it all to Element instead.

  • dom/Node.cpp:

(WebCore::Node::detach):

  • dom/Element.cpp:

(WebCore::Element::detach):

Move Document::hoveredNodeDetached() call from Node::detach() to Element::detach().

(WebCore::Element::isUserActionElementHovered):
(WebCore::Element::setHovered):

  • dom/Element.h:

(WebCore::Element::hovered):

Moved here from Node/ContainerNode.

  • dom/UserActionElementSet.h:

(WebCore::UserActionElementSet::setHovered):
(WebCore::UserActionElementSet::setFlags):

Make setHovered() take an Element* instead of a Node*. I'd like to do this will all
of these methods eventually, so added a setFlags() helper with the tighter type.

  • html/HTMLLabelElement.h:
  • html/shadow/SpinButtonElement.h:

Sprinkle OVERRIDE on setHovered() overrides.

  • dom/Document.cpp:

(WebCore::Document::updateHoverActiveState):

  • accessibility/AccessibilityNodeObject.cpp:

(WebCore::AccessibilityNodeObject::isHovered):

  • rendering/RenderTheme.cpp:

(WebCore::RenderTheme::isHovered):

Check if the inspected Node is an element before asking if it's hovered.

2:08 AM Changeset in webkit [150683] by akling@apple.com
  • 6 edits in trunk/Source/WebCore

Document::formController() should return a reference.
<http://webkit.org/b/116758>

Reviewed by Antti Koivisto.

The formController() is created on demand, so return a reference instead.

  • dom/Document.h:
  • dom/Document.cpp:

(WebCore::Document::formController):
(WebCore::Document::setStateForNewFormElements):

  • html/HTMLFormControlElementWithState.cpp:

(WebCore::HTMLFormControlElementWithState::HTMLFormControlElementWithState):
(WebCore::HTMLFormControlElementWithState::~HTMLFormControlElementWithState):
(WebCore::HTMLFormControlElementWithState::didMoveToNewDocument):
(WebCore::HTMLFormControlElementWithState::finishParsingChildren):

  • html/HTMLFormElement.cpp:

(WebCore::HTMLFormElement::~HTMLFormElement):
(WebCore::HTMLFormElement::finishParsingChildren):

  • html/HTMLInputElement.cpp:

(WebCore::HTMLInputElement::~HTMLInputElement):
(WebCore::HTMLInputElement::didMoveToNewDocument):
(WebCore::HTMLInputElement::checkedRadioButtons):

1:29 AM Changeset in webkit [150682] by sergio@webkit.org
  • 2 edits in trunk/Source/WebCore

Reducing CSS code duplication in declaration list error recovery
https://bugs.webkit.org/show_bug.cgi?id=115157

Reviewed by Andreas Kling.

From Blink r148974 by <serya@chromium.org>

Refactor "invalid_block_list error | /* empty */" in a new
decl_list_recovery rule that allows to remove 4 redundant rules.

No new test required as we're just refactoring code, no new behaviour.

  • css/CSSGrammar.y.in:

May 24, 2013:

9:20 PM Changeset in webkit [150681] by commit-queue@webkit.org
  • 4 edits in trunk/Source/WebKit2

[Qt][GTK] Build fix after r150669
https://bugs.webkit.org/show_bug.cgi?id=116751

Patch by Seokju Kwon <Seokju Kwon> on 2013-05-24
Reviewed by Christophe Dumez.

  • GNUmakefile.list.am:
  • Target.pri:
  • WebProcess/WebCoreSupport/gtk/WebEditorClientGtk.cpp:

(WebKit::WebEditorClient::getEditorCommandsForKeyEvent):

9:01 PM Changeset in webkit [150680] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebKit/efl

[EFL] Remove obsolete Page forward declaration from InspectorClientEfl.h
https://bugs.webkit.org/show_bug.cgi?id=116752

Patch by Seokju Kwon <Seokju Kwon> on 2013-05-24
Reviewed by Christophe Dumez.

  • WebCoreSupport/InspectorClientEfl.h:
7:44 PM CommitterTips edited by Seokju Kwon
It is generated automatically from ​contributors.json after r150662. (diff)
7:19 PM Changeset in webkit [150679] by commit-queue@webkit.org
  • 6 edits in trunk/Source/WebKit2

[EFL] Build fix after r150669
https://bugs.webkit.org/show_bug.cgi?id=116750

Patch by Seokju Kwon <Seokju Kwon> on 2013-05-24
Reviewed by Anders Carlsson.

  • CMakeLists.txt:
  • WebProcess/Battery/WebBatteryManager.cpp:

(WebKit::WebBatteryManager::registerWebPage):
(WebKit::WebBatteryManager::unregisterWebPage):

  • WebProcess/NetworkInfo/WebNetworkInfoManager.cpp:

(WebKit::WebNetworkInfoManager::registerWebPage):
(WebKit::WebNetworkInfoManager::unregisterWebPage):
(WebKit::WebNetworkInfoManager::bandwidth):
(WebKit::WebNetworkInfoManager::metered):

  • WebProcess/WebCoreSupport/WebVibrationClient.cpp:

(WebKit::WebVibrationClient::vibrate):
(WebKit::WebVibrationClient::cancelVibration):

  • WebProcess/soup/WebSoupRequestManager.cpp:

(WebKit::WebSoupRequestManager::didReceiveURIRequestData):
(WebKit::WebSoupRequestManager::send):

6:28 PM Changeset in webkit [150678] by Brent Fulgham
  • 2 edits in trunk/LayoutTests

Unreviewed change to get Windows bots green. Disable a number
of crashes and failures. Crashes are tracked under
https://bugs.webkit.org/show_bug.cgi?id=116562, failures under
https://bugs.webkit.org/show_bug.cgi?id=116564.

  • platform/win/TestExpectations:
6:18 PM Changeset in webkit [150677] by Brent Fulgham
  • 2 edits in trunk/Tools

Unreviewed build fix. Don't use new NRWT logic needed
by Windows yet. Activate (and fix test) when we cut over.

  • Scripts/webkitpy/port/driver.py:

(Driver._setup_environ_for_driver):
(Driver._start):

6:12 PM Changeset in webkit [150676] by benjamin@webkit.org
  • 2 edits in trunk/Source/WTF

Inline TCMalloc_Central_FreeList's RemoveRange and FetchFromSpansSafe
https://bugs.webkit.org/show_bug.cgi?id=116700

Patch by Benjamin Poulain <bpoulain@apple.com> on 2013-05-24
Reviewed by Darin Adler.

The two functions only have one call site each.

  • wtf/FastMalloc.cpp:

(WTF::TCMalloc_Central_FreeList::RemoveRange):
(WTF::TCMalloc_Central_FreeList::FetchFromSpansSafe):

6:02 PM Changeset in webkit [150675] by rniwa@webkit.org
  • 2 edits in trunk/Source/WebCore

Build fix after r150664.

  • bindings/scripts/test/JS/JSTestObj.cpp:

(WebCore):

5:59 PM Changeset in webkit [150674] by Lucas Forschler
  • 4 edits in branches/safari-537.43-branch/Source

Versioning.

5:56 PM Changeset in webkit [150673] by Lucas Forschler
  • 1 copy in tags/Safari-537.43.2

New Tag.

5:49 PM Changeset in webkit [150672] by commit-queue@webkit.org
  • 3 edits
    2 adds in trunk

Fixing invalid block recovery in some declaration list.
https://bugs.webkit.org/show_bug.cgi?id=115159

Patch by Sergio Villar Senin <svillar@igalia.com> on 2013-05-24
Reviewed by Darin Adler.

From Blink r148889 by <serya@chromium.org>

Source/WebCore:

Improves the detection of some invalid block declarations inside a
declaration list. This allows the parser not to completely discard
a whole declaration list that contains an invalid block. In
particular grammar failed in tests 8, 12, 13, 15, 16 and 17 in the
new test added by this change.

Test: fast/css/parsing-invalid-block-recovery.html

  • css/CSSGrammar.y.in:

LayoutTests:

  • fast/css/parsing-invalid-block-recovery-expected.txt: Added.
  • fast/css/parsing-invalid-block-recovery.html: Added.
5:48 PM Changeset in webkit [150671] by Beth Dakin
  • 2 edits in trunk/Source/WebCore

DidFirstFlushForHeaderLayer can be called too soon on pages that redirect
https://bugs.webkit.org/show_bug.cgi?id=116749
-and corresponding-
<rdar://problem/13886753>

Reviewed by Simon Fraser.

Only start m_paintRelatedMilestonesTimer if we have actually painted.

  • rendering/RenderLayerCompositor.cpp:

(WebCore::RenderLayerCompositor::flushPendingLayerChanges):

5:19 PM Changeset in webkit [150670] by Brent Fulgham
  • 11 edits in trunk

[Windows] Expose database storage and cache locations via preferences.
https://bugs.webkit.org/show_bug.cgi?id=116729

Reviewed by Tim Horton.

Source/WebKit:

  • WebKit.vcxproj/WebKitExportGenerator/WebKitExports.def.in: Export

three symbols needed to implement the feature.

Source/WebKit/win:

  • WebDatabaseManager.cpp: Update to check preferences for the

desired location of the database store.
(databasesDirectory): Added.
(WebKitInitializeWebDatabasesIfNecessary): Use new databasesDirectory
method to determine what system path to use for file storage.

  • WebKit.vcproj/WebKitExports.def.in: Export three symbols

needed to implement the feature.

  • WebView.cpp: Update to check preferences for the desired location

of the various caches used by WebKit.
(WebView::setCacheModel): Update to check preferences for URL cache
storage.
(WebKitSetApplicationCachePathIfNecessary): Update to check
preferences for ccache storage.

Tools:

Update DumpRenderTree to use CFPreferences to control where WebKit
stores its local databases, URL caches, etc.

  • DumpRenderTree/win/DumpRenderTree.cpp:

(libraryPathForDumpRenderTree): Added.
(dllLauncherEntryPoint): Set up DRT-specific cache locations.

  • Scripts/webkitpy/port/base.py:

(Port._driver_tempdir): Added (to allow port-specific overload).
(Port._driver_tempdir_for_environment): Ditto

  • Scripts/webkitpy/port/driver.py:

(Driver._setup_environ_for_driver): Use new overload to set
environment variable.
(Driver._start): Use new overload for temp directory location.

  • Scripts/webkitpy/port/win.py:

(WinPort._driver_tempdir_for_environment): New overload to supply
Windows path to DumpRenderTree environment (while still using
cygwin paths for internal operations.

5:17 PM Changeset in webkit [150669] by andersca@apple.com
  • 40 edits
    1 copy in trunk/Source/WebKit2

MessageSender shouldn't be a class template
https://bugs.webkit.org/show_bug.cgi?id=116743

Reviewed by Andreas Kling.

Change MessageSender to be a proper abstract base class. Subclasses of MessageSender must
implement messageSenderConnection() as well as messageSenderDestinationID().

This reduces code size and will allow us to be more flexible with MessageSender in the future.

  • NetworkProcess/NetworkResourceLoader.cpp:

(WebKit::NetworkResourceLoader::sendAbortingOnFailure):
(WebKit::NetworkResourceLoader::messageSenderConnection):
(WebKit):
(WebKit::NetworkResourceLoader::messageSenderDestinationID):

  • NetworkProcess/NetworkResourceLoader.h:

(NetworkResourceLoader):

  • NetworkProcess/mac/DiskCacheMonitor.h:

(DiskCacheMonitor):

  • NetworkProcess/mac/DiskCacheMonitor.mm:

(WebKit::DiskCacheMonitor::messageSenderConnection):
(WebKit::DiskCacheMonitor::messageSenderDestinationID):
(WebKit):

  • Platform/CoreIPC/MessageSender.h:

(CoreIPC):
(MessageSender):
(CoreIPC::MessageSender::send):
(CoreIPC::MessageSender::sendSync):

  • Shared/ChildProcess.cpp:

(WebKit::ChildProcess::messageSenderConnection):
(WebKit):
(WebKit::ChildProcess::messageSenderDestinationID):

  • Shared/ChildProcess.h:

(ChildProcess):

  • Shared/Downloads/Download.cpp:

(WebKit::Download::messageSenderConnection):
(WebKit):
(WebKit::Download::messageSenderDestinationID):

  • Shared/Downloads/Download.h:

(Download):

  • Shared/Network/CustomProtocols/mac/CustomProtocolManagerMac.mm:

(WebKit::CustomProtocolManager::initialize):

  • Shared/WebConnection.h:

(WebConnection):

  • Shared/mac/SecItemShim.cpp:

(WebKit::sendSecItemRequest):

  • UIProcess/WebConnectionToWebProcess.cpp:

(WebKit::WebConnectionToWebProcess::hasValidConnection):
(WebKit::WebConnectionToWebProcess::messageSenderConnection):
(WebKit::WebConnectionToWebProcess::messageSenderDestinationID):

  • UIProcess/WebConnectionToWebProcess.h:

(WebConnectionToWebProcess):

  • WebKit2.xcodeproj/project.pbxproj:
  • WebProcess/Geolocation/WebGeolocationManager.cpp:

(WebKit::WebGeolocationManager::registerWebPage):
(WebKit::WebGeolocationManager::unregisterWebPage):

  • WebProcess/IconDatabase/WebIconDatabaseProxy.cpp:

(WebKit::WebIconDatabaseProxy::retainIconForPageURL):
(WebKit::WebIconDatabaseProxy::releaseIconForPageURL):
(WebKit::WebIconDatabaseProxy::synchronousIconForPageURL):
(WebKit::WebIconDatabaseProxy::loadDecisionForIconURL):
(WebKit::WebIconDatabaseProxy::setIconURLForPageURL):
(WebKit::WebIconDatabaseProxy::setIconDataForIconURL):

  • WebProcess/InjectedBundle/InjectedBundle.cpp:

(WebKit::InjectedBundle::postMessage):
(WebKit::InjectedBundle::postSynchronousMessage):

  • WebProcess/Network/WebResourceLoader.cpp:

(WebKit::WebResourceLoader::messageSenderConnection):
(WebKit::WebResourceLoader::messageSenderDestinationID):

  • WebProcess/Network/WebResourceLoader.h:

(WebResourceLoader):

  • WebProcess/Notifications/WebNotificationManager.cpp:

(WebKit::WebNotificationManager::show):
(WebKit::WebNotificationManager::cancel):
(WebKit::WebNotificationManager::clearNotifications):
(WebKit::WebNotificationManager::didDestroyNotification):

  • WebProcess/Plugins/PDF/PDFPlugin.mm:

(WebKit::PDFPlugin::writeItemsToPasteboard):

  • WebProcess/Plugins/PluginProcessConnectionManager.cpp:

(WebKit::PluginProcessConnectionManager::getPluginProcessConnection):

  • WebProcess/Storage/StorageAreaMap.cpp:

(WebKit::StorageAreaMap::StorageAreaMap):
(WebKit::StorageAreaMap::~StorageAreaMap):
(WebKit::StorageAreaMap::setItem):
(WebKit::StorageAreaMap::removeItem):
(WebKit::StorageAreaMap::clear):
(WebKit::StorageAreaMap::loadValuesIfNeeded):

  • WebProcess/WebConnectionToUIProcess.cpp:

(WebKit::WebConnectionToUIProcess::hasValidConnection):
(WebKit::WebConnectionToUIProcess::messageSenderConnection):
(WebKit::WebConnectionToUIProcess::messageSenderDestinationID):

  • WebProcess/WebConnectionToUIProcess.h:

(WebConnectionToUIProcess):

  • WebProcess/WebCoreSupport/WebChromeClient.cpp:

(WebKit::WebChromeClient::windowRect):
(WebKit::WebChromeClient::focusedFrameChanged):
(WebKit::WebChromeClient::createWindow):
(WebKit::WebChromeClient::toolbarsVisible):
(WebKit::WebChromeClient::statusbarVisible):
(WebKit::WebChromeClient::menubarVisible):
(WebKit::WebChromeClient::runBeforeUnloadConfirmPanel):
(WebKit::WebChromeClient::runJavaScriptAlert):
(WebKit::WebChromeClient::runJavaScriptConfirm):
(WebKit::WebChromeClient::runJavaScriptPrompt):
(WebKit::WebChromeClient::shouldInterruptJavaScript):
(WebKit::WebChromeClient::exceededDatabaseQuota):

  • WebProcess/WebCoreSupport/WebColorChooser.cpp:

(WebKit::WebColorChooser::WebColorChooser):
(WebKit::WebColorChooser::setSelectedColor):
(WebKit::WebColorChooser::endChooser):

  • WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:

(WebKit::WebFrameLoaderClient::dispatchDidReceiveIcon):
(WebKit::WebFrameLoaderClient::updateGlobalHistory):
(WebKit::WebFrameLoaderClient::updateGlobalHistoryRedirectLinks):
(WebKit::WebFrameLoaderClient::setTitle):

  • WebProcess/WebCoreSupport/WebPlatformStrategies.cpp:

(WebKit::WebPlatformStrategies::populatePluginCache):
(WebKit::WebPlatformStrategies::getTypes):
(WebKit::WebPlatformStrategies::bufferForType):
(WebKit::WebPlatformStrategies::getPathnamesForType):
(WebKit::WebPlatformStrategies::stringForType):
(WebKit::WebPlatformStrategies::copy):
(WebKit::WebPlatformStrategies::changeCount):
(WebKit::WebPlatformStrategies::uniqueName):
(WebKit::WebPlatformStrategies::color):
(WebKit::WebPlatformStrategies::url):
(WebKit::WebPlatformStrategies::addTypes):
(WebKit::WebPlatformStrategies::setTypes):
(WebKit::WebPlatformStrategies::setBufferForType):
(WebKit::WebPlatformStrategies::setPathnamesForType):
(WebKit::WebPlatformStrategies::setStringForType):

  • WebProcess/WebCoreSupport/WebPopupMenu.cpp:

(WebKit::WebPopupMenu::show):
(WebKit::WebPopupMenu::hide):

  • WebProcess/WebCoreSupport/WebSearchPopupMenu.cpp:

(WebKit::WebSearchPopupMenu::saveRecentSearches):
(WebKit::WebSearchPopupMenu::loadRecentSearches):

  • WebProcess/WebPage/EventDispatcher.cpp:

(WebKit::EventDispatcher::sendDidReceiveEvent):

  • WebProcess/WebPage/WebBackForwardListProxy.cpp:

(WebKit::updateBackForwardItem):
(WebKit::WebBackForwardListProxy::itemAtIndex):
(WebKit::WebBackForwardListProxy::backListCount):
(WebKit::WebBackForwardListProxy::forwardListCount):

  • WebProcess/WebPage/WebInspector.cpp:

(WebKit::WebInspector::createInspectorPage):
(WebKit::WebInspector::didClose):
(WebKit::WebInspector::bringToFront):
(WebKit::WebInspector::inspectedURLChanged):
(WebKit::WebInspector::save):
(WebKit::WebInspector::append):
(WebKit::WebInspector::attachBottom):
(WebKit::WebInspector::attachRight):
(WebKit::WebInspector::detach):
(WebKit::WebInspector::setAttachedWindowHeight):
(WebKit::WebInspector::setAttachedWindowWidth):
(WebKit::WebInspector::setToolbarHeight):
(WebKit::WebInspector::updateDockingAvailability):
(WebKit::WebInspector::sendMessageToRemoteFrontend):

  • WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::messageSenderConnection):
(WebKit):
(WebKit::WebPage::messageSenderDestinationID):

  • WebProcess/WebPage/WebPage.h:

(WebPage):

  • WebProcess/WebPage/mac/WebPageMac.mm:

(WebKit::WebPage::executeKeypressCommandsInternal):
(WebKit::WebPage::handleEditingKeyboardEvent):

  • WebProcess/WebProcess.cpp:

(WebKit::WebProcess::ensureNetworkProcessConnection):
(WebKit::WebProcess::shouldTerminate):

4:52 PM Changeset in webkit [150668] by andersca@apple.com
  • 15 edits in trunk/Source/WebKit/mac

Remove using namespace std from Mac WebKit1
https://bugs.webkit.org/show_bug.cgi?id=116747

Reviewed by Sam Weinig.

Use explicit std:: qualifiers instead.

  • History/BinaryPropertyList.cpp:

(IntegerArray::deletedValueSize):
(BinaryPropertyListPlan::invalidObjectReference):
(bytesNeeded):

  • History/WebHistory.mm:

(-[WebHistoryPrivate orderedLastVisitedDays]):
(WebHistoryWriter::WebHistoryWriter):

  • History/WebHistoryItem.mm:

(-[WebHistoryItem initFromDictionaryRepresentation:]):

  • Plugins/Hosted/NetscapePluginHostManager.mm:

(WebKit::NetscapePluginHostManager::instantiatePlugin):

  • Plugins/Hosted/NetscapePluginHostProxy.mm:
  • Plugins/Hosted/NetscapePluginInstanceProxy.mm:

(WebKit::NetscapePluginInstanceProxy::LocalObjectMap::idForObject):
(WebKit::NetscapePluginInstanceProxy::wheelEvent):
(WebKit::NetscapePluginInstanceProxy::print):
(WebKit::NetscapePluginInstanceProxy::snapshot):
(WebKit::NetscapePluginInstanceProxy::loadURL):
(WebKit::NetscapePluginInstanceProxy::createBindingsInstance):

  • Plugins/Hosted/ProxyInstance.mm:

(WebKit::ProxyInstance::invoke):
(WebKit::ProxyInstance::supportsInvokeDefaultMethod):
(WebKit::ProxyInstance::supportsConstruct):
(WebKit::ProxyInstance::getPropertyNames):
(WebKit::ProxyInstance::methodNamed):
(WebKit::ProxyInstance::fieldNamed):
(WebKit::ProxyInstance::fieldValue):
(WebKit::ProxyInstance::setFieldValue):

  • Plugins/WebBasePluginPackage.mm:

(-[WebBasePluginPackage supportsExtension:]):
(-[WebBasePluginPackage MIMETypeForExtension:]):

  • Plugins/WebNetscapePluginStream.mm:

(WebNetscapePluginStream::deliverData):

  • Plugins/WebNetscapePluginView.mm:

(PluginTimer::start):
(-[WebNetscapePluginView _postURL:target:len:buf:file:notifyData:sendNotification:allowHeaders:]):

  • WebCoreSupport/WebFrameLoaderClient.mm:
  • WebView/WebFrame.mm:
  • WebView/WebHTMLView.mm:

(-[WebHTMLView _adjustedBottomOfPageWithTop:bottom:limit:]):

  • WebView/WebTextCompletionController.mm:

(-[WebTextCompletionController _placePopupWindow:]):

4:46 PM Changeset in webkit [150667] by andersca@apple.com
  • 4 edits in trunk/Source/WebCore

Fix Windows build.

  • WebCore.vcproj/WebCore.vcproj:
  • WebCore.vcxproj/WebCore.vcxproj:
  • WebCore.vcxproj/WebCore.vcxproj.filters:
4:37 PM Changeset in webkit [150666] by timothy_horton@apple.com
  • 6 edits in trunk/Source

Animated GIFs don't repaint after TiledCoreAnimationDrawingArea un-suspends painting
https://bugs.webkit.org/show_bug.cgi?id=116744
<rdar://problem/13973514>

Reviewed by Simon Fraser.

  • WebCore.exp.in: Export FrameView::resumeAnimatingImages.
  • page/FrameView.cpp:

(WebCore::FrameView::setIsInWindow):
(WebCore::FrameView::resumeAnimatingImages):
Factor the code to restart animated images out of frameView::setIsInWindow.

  • page/FrameView.h:

(FrameView):

  • WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.mm:

(WebKit::TiledCoreAnimationDrawingArea::resumePainting):
Force WebCore to repaint animated images when painting is resumed.

4:30 PM Changeset in webkit [150665] by andersca@apple.com
  • 7 edits
    4 moves in trunk/Source

Move history property list writer to WebKit
https://bugs.webkit.org/show_bug.cgi?id=116746

Reviewed by Tim Horton.

Source/WebCore:

Remove files from the Xcode project and remove symbols that now live in WebKit.

  • WebCore.exp.in:
  • WebCore.xcodeproj/project.pbxproj:

Source/WebKit:

Add files to the Xcode project.

  • WebKit.xcodeproj/project.pbxproj:

Source/WebKit/mac:

Move WebCore history files to WebKit since they're only used by WebKit now.

  • History/BinaryPropertyList.cpp: Renamed from Source/WebCore/platform/cf/BinaryPropertyList.cpp.
  • History/BinaryPropertyList.h: Renamed from Source/WebCore/platform/cf/BinaryPropertyList.h.
  • History/HistoryPropertyList.cpp: Renamed from Source/WebCore/history/cf/HistoryPropertyList.cpp.
  • History/HistoryPropertyList.h: Renamed from Source/WebCore/history/cf/HistoryPropertyList.h.
  • History/WebHistory.mm:
4:15 PM Changeset in webkit [150664] by Christophe Dumez
  • 10 edits
    3 adds in trunk

Global constructors exposed in worker environment have wrong attributes
https://bugs.webkit.org/show_bug.cgi?id=116708

Reviewed by Geoffrey Garen.

Source/WebCore:

Fix JSC bindings generator so that Global constructors exposed in worker environment
have the following attributes:
{ Writable?: true, Enumerable?: false, Configurable?: true }

This is now consistent with global constructors on DOMWindow and with the Web IDL
specification:
http://dev.w3.org/2006/webapi/WebIDL/#es-interfaces

Test: fast/js/global-constructors-attributes-worker.html

  • bindings/scripts/CodeGeneratorJS.pm:

(GenerateAttributesHashTable):

LayoutTests:

Add fast/js/global-constructors-attributes-worker.html test case to check the
global constructors attributers in the worker environment.

  • fast/js/global-constructors-attributes-expected.txt:
  • fast/js/global-constructors-attributes-worker-expected.txt: Added.
  • fast/js/global-constructors-attributes-worker.html: Added.
  • fast/js/global-constructors-attributes.html:
  • fast/js/script-tests/global-constructors-attributes.js: Make sure script usable

in a Worker and stop reporting the "constructor" property as a global constructor.

  • platform/efl/fast/js/global-constructors-attributes-expected.txt:
  • platform/efl/fast/js/global-constructors-attributes-worker-expected.txt: Added.
  • platform/gtk/fast/js/global-constructors-attributes-expected.txt:
  • platform/mac-lion/fast/js/global-constructors-attributes-expected.txt:
  • platform/qt/fast/js/global-constructors-attributes-expected.txt:
3:57 PM Changeset in webkit [150663] by Christophe Dumez
  • 34 edits
    1 delete in trunk

Remove custom code for webkitAudioContext global constructor getter
https://bugs.webkit.org/show_bug.cgi?id=116530

Reviewed by Geoffrey Garen.

Source/WebCore:

Get rid of custom code for webkitAudioContext global constructor getter. The
global DOMWindow attribute is now automatically generated by leveraging support
for [EnabledAtRuntime] extended attribute and the existing methods in
RuntimeEnabledFeatures class.

No new tests, no behavior change.

  • GNUmakefile.list.am: Remove JSDOMWindowWebAudioCustom.*.
  • Modules/webaudio/AudioContext.idl:

Remove [NoInterfaceObject] extended attribute so that the global constructor is
automatically generated instead of automatically added. Add [EnabledAtRuntime]
extended attribute since this global constructor can be disabled at runtime and
so that we can getter rid of the [CustomGetter] for the corresponding global
constructor. Use [InterfaceName] extended attribute so that the interface name
matches the global constructor on DOMWindow (webkitAudioContext).

  • Target.pri: Remove JSDOMWindowWebAudioCustom.*.
  • UseJSC.cmake: Remove JSDOMWindowWebAudioCustom.*.
  • WebCore.exp.in: Export additional symbol.
  • WebCore.order: Export additional symbol.
  • WebCore.vcproj/WebCore.vcproj: Remove JSDOMWindowWebAudioCustom.*.
  • WebCore.vcxproj/WebCore.vcxproj: Remove JSDOMWindowWebAudioCustom.*.
  • WebCore.vcxproj/WebCore.vcxproj.filters: Remove JSDOMWindowWebAudioCustom.*.
  • WebCore.xcodeproj/project.pbxproj: Remove JSDOMWindowWebAudioCustom.*.
  • bindings/generic/RuntimeEnabledFeatures.h:

(WebCore::RuntimeEnabledFeatures::setWebAudioEnabled):

  • bindings/js/JSBindingsAllInOne.cpp: Remove JSDOMWindowWebAudioCustom.*.
  • bindings/js/JSDOMWindowWebAudioCustom.cpp: Removed. We no longer need any custom code for

the WebAudio global constructor getter because we use [EnabledAtRuntime] support in the
bindings generator now.

  • page/DOMWindow.idl: Remove global constructor for AudioContext as it is now automatically

generated.

  • page/Settings.in: Remove webAudioEnabled setting as we use the existing setting

in RuntimeEnabledFeatures now.

Source/WebKit/efl:

Use RuntimeEnabledFeatures instead of Settings to toggle Web Audio support.

  • WebCoreSupport/DumpRenderTreeSupportEfl.cpp:

(DumpRenderTreeSupportEfl::setWebAudioEnabled):

  • WebCoreSupport/DumpRenderTreeSupportEfl.h:
  • ewk/ewk_view.cpp:

(_Ewk_View_Private_Data):
(_ewk_view_priv_new):

  • ewk/ewk_view.h: Remove public API to toggle Web Audio support as it is no longer

a view specific feature.

Source/WebKit/gtk:

Use RuntimeEnabledFeatures instead of Settings to toggle Web Audio support.

  • webkit/webkitwebview.cpp:

(webkit_web_view_update_settings):
(webkit_web_view_settings_notify):

Source/WebKit/mac:

Use RuntimeEnabledFeatures instead of Settings to toggle Web Audio support.

  • WebView/WebView.mm:

(-[WebView _preferencesChanged:]):

Source/WebKit/qt:

Use RuntimeEnabledFeatures instead of Settings to toggle Web Audio support.

  • Api/qwebsettings.cpp:

(QWebSettingsPrivate::apply):

Source/WebKit/win:

Use RuntimeEnabledFeatures instead of Settings to toggle Web Audio support.

  • WebView.cpp:

(WebView::notifyPreferencesChanged):

Source/WebKit2:

Use RuntimeEnabledFeatures instead of WebCore Settings to toggle Web Audio support.

  • WebProcess/InjectedBundle/InjectedBundle.cpp:

(WebKit::InjectedBundle::overrideBoolPreferenceForTestRunner):

  • WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::updatePreferences):

Tools:

Update EFL's DumpRenderTree to use DumpRenderTreeSupport to toggle
WebAudio feature instead of public ewk_view API since that API was
removed.

  • DumpRenderTree/efl/DumpRenderTreeChrome.cpp:

(DumpRenderTreeChrome::resetDefaultsToConsistentValues):

  • DumpRenderTree/efl/TestRunnerEfl.cpp:

(TestRunner::overridePreference):

3:49 PM WebKit Team edited by Bem Jones-Bey
Update to point to new location. (diff)
3:34 PM Changeset in webkit [150662] by commit-queue@webkit.org
  • 4 edits in trunk

Make team.html have each contributor's area of expertise
https://bugs.webkit.org/show_bug.cgi?id=116737

Patch by Bem Jones-Bey <Bem Jones-Bey> on 2013-05-24
Reviewed by Benjamin Poulain.

Tools:

  • Scripts/webkitpy/common/config/contributors.json:

Add expertise to each contributor, taken from http://trac.webkit.org/wiki/WebKit%20Team

Changed formatting of the file because the lines are getting way to
long, and I generated this, so it's easier if the formatting of the
file is regular.

Websites/webkit.org:

  • team.html:

Update to use the new expertise field in contributors.json.

3:27 PM Changeset in webkit [150661] by fpizlo@apple.com
  • 3 edits in branches/dfgFourthTier/Source/JavaScriptCore

fourthTier: FTL boolify should support ObjectOrOtherUse
https://bugs.webkit.org/show_bug.cgi?id=116741

Reviewed by Geoffrey Garen.

Just reusing what was already there in equalNullOrUndefined(). Note that we will
sometimes generate some redundant IR - like having some spurious bitNot's in
places - but it's safe to assume that LLVM will simplify those, and that it won't
be the longest pole in the tent for compile times.

  • ftl/FTLCapabilities.cpp:

(JSC::FTL::canCompile):

  • ftl/FTLLowerDFGToLLVM.cpp:

(JSC::FTL::LowerDFGToLLVM::compileCompareEqConstant):
(JSC::FTL::LowerDFGToLLVM::compileCompareStrictEqConstant):
(JSC::FTL::LowerDFGToLLVM::boolify):
(JSC::FTL::LowerDFGToLLVM::equalNullOrUndefined):

3:05 PM Changeset in webkit [150660] by andersca@apple.com
  • 5 edits in trunk/Source/WebKit/win

Remove unused code to read/write history
https://bugs.webkit.org/show_bug.cgi?id=116738

Reviewed by Brent Fulgham.

  • Interfaces/IWebHistory.idl:

Replace loadFromURL and saveToURL with unused1/unused2.

  • Interfaces/IWebHistoryPrivate.idl:

Replace data with unused1.

  • WebHistory.cpp:

Remove history loading/saving code.

(WebHistory::unused1):
(WebHistory::unused2):
Add implementations.

  • WebHistory.h:

(WebHistory):

3:00 PM Changeset in webkit [150659] by fpizlo@apple.com
  • 7 edits
    3 adds in trunk

We broke !(0/0)
https://bugs.webkit.org/show_bug.cgi?id=116736

Reviewed by Gavin Barraclough.

Source/JavaScriptCore:

  • parser/ASTBuilder.h:

(JSC::ASTBuilder::createLogicalNot):

  • runtime/JSCJSValueInlines.h:

(JSC::JSValue::pureToBoolean):

Source/WTF:

  • wtf/MathExtras.h:

(isNotZeroAndOrdered):
(isZeroOrUnordered):

LayoutTests:

  • fast/js/constant-fold-not-nan.html: Added.
  • fast/js/constant-fold-not-nan-expected.txt: Added.
  • fast/js/jsc-test-list:
  • fast/js/script-tests/constant-fold-not-nan.js: Added.
2:45 PM Changeset in webkit [150658] by fpizlo@apple.com
  • 4 edits in branches/dfgFourthTier/Source/JavaScriptCore

fourthTier: FTL should support LogicalNot and Branch on Int32 and Number
https://bugs.webkit.org/show_bug.cgi?id=116739

Reviewed by Gavin Barraclough.

  • ftl/FTLCapabilities.cpp:

(JSC::FTL::canCompile):

  • ftl/FTLLowerDFGToLLVM.cpp:

(JSC::FTL::LowerDFGToLLVM::compileLogicalNot):
(JSC::FTL::LowerDFGToLLVM::compileBranch):
(JSC::FTL::LowerDFGToLLVM::boolify):
(LowerDFGToLLVM):

  • ftl/FTLOutput.h:

(JSC::FTL::Output::isZero32):
(JSC::FTL::Output::notZero32):

2:06 PM Changeset in webkit [150657] by fpizlo@apple.com
  • 11 edits in branches/dfgFourthTier/Source/JavaScriptCore

fourthTier: add heuristics to reduce the likelihood of a trivially inlineable function being independently compiled by the concurrent JIT
https://bugs.webkit.org/show_bug.cgi?id=116557

Reviewed by Geoffrey Garen.

This introduces a fairly comprehensive mechanism for preventing trivially inlineable
functions from being compiled independently of all of the things into which they end
up being inlined.

The trick is CodeBlock::m_shouldAlwaysBeInlined, or SABI for short (that's what the
debug logging calls it). A SABI function is one that we currently believe should
never be DFG optimized because it should always be inlined into the functions that
call it. SABI follows "innocent until proven guilty": all functions start out SABI
and have SABI set to false if we see proof that that function may be called in some
possibly non-inlineable way. So long as a function is SABI, it will not tier up to
the DFG: cti_optimize will perpetually postpone its optimization. Because SABI has
such a severe effect, we make the burden of proof of guilt quite low. SABI gets
cleared if any of the following happen:

  • You get called from native code (either through CallData or CachedCall).


  • You get called from an eval, since eval code takes a long time to get DFG optimized.


  • You get called from global code, since often global code doesn't tier-up since it's run-once.


  • You get called recursively, where recursion is detected by a stack walk of depth Options::maximumInliningDepth().


  • You get called through an unlinked virtual call.


  • You get called from DFG code, since if the caller was already DFG optimized and didn't inline you then obviously, you might not get inlined.


  • You've tiered up to the baseline JIT and you get called from the interpreter. The idea here is that this kind of ensures that you stay SABI only if you're called no more frequently than any of your callers.


  • You get called from a code block that isn't a DFG candidate.


  • You aren't an inlining candidate.


Most of the heuristics for SABI are in CodeBlock::noticeIncomingCall().

This is neutral on SunSpider and V8Spider, and appears to be a slight speed-up on
V8v7, which was previously adversely affected by concurrent compilation. I also
confirmed that for example on V8/richards, it dramatically reduces the number of
code blocks that get DFG compiled. It is a speed-up on those V8v7 benchmarks that
saw regressions from concurrent compilation.

  • bytecode/CodeBlock.cpp:

(JSC::CodeBlock::dumpAssumingJITType):
(JSC::CodeBlock::CodeBlock):
(JSC::CodeBlock::linkIncomingCall):
(JSC):
(JSC::CodeBlock::noticeIncomingCall):

  • bytecode/CodeBlock.h:

(CodeBlock):

  • dfg/DFGCapabilities.h:

(JSC::DFG::mightInlineFunction):
(DFG):

  • dfg/DFGPlan.cpp:

(JSC::DFG::Plan::compileInThread):

  • dfg/DFGRepatch.cpp:

(JSC::DFG::dfgLinkFor):

  • interpreter/Interpreter.cpp:

(JSC::Interpreter::executeCall):
(JSC::Interpreter::executeConstruct):
(JSC::Interpreter::prepareForRepeatCall):

  • jit/JIT.cpp:

(JSC::JIT::privateCompile):
(JSC::JIT::linkFor):

  • jit/JIT.h:

(JIT):

  • jit/JITStubs.cpp:

(JSC::DEFINE_STUB_FUNCTION):
(JSC::lazyLinkFor):

  • llint/LLIntSlowPaths.cpp:

(JSC::LLInt::setUpCall):

1:26 PM Changeset in webkit [150656] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebCore

[AC] Needs to be guarded by USE(ACCELERATED_COMPOSITING)
https://bugs.webkit.org/show_bug.cgi?id=116712

Patch by Seokju Kwon <Seokju Kwon> on 2013-05-24
Reviewed by Andreas Kling.

renderView is used when ACCELERATED_COMPOSITING is enabled.
So, this patch allows it to be guarded by USE(ACCELERATED_COMPOSITING).

No new tests, no behavior change.

  • page/FrameView.cpp:

(WebCore::FrameView::setFrameRect):

1:07 PM Changeset in webkit [150655] by sergio@webkit.org
  • 3 edits in trunk/LayoutTests

REGRESSION (r150215): broke a test case in inspector/console/console-css-warnings.html
https://bugs.webkit.org/show_bug.cgi?id=116731

Reviewed by Ryosuke Niwa.

The mentioned revision removed a layering violation added just to
support some IE7 hack. That was removed but the test case
supporting it was left untouched, remove it.

  • inspector/console/console-css-warnings-expected.txt:
  • inspector/console/console-css-warnings.html:
12:43 PM Changeset in webkit [150654] by g.czajkowski@samsung.com
  • 6 edits in trunk/LayoutTests

[WK2] Spelling tests gardening
https://bugs.webkit.org/show_bug.cgi?id=116718

Unreviewed gardening.

The most spelling tests are skipped for wk2 platforms and they have deprecated comments.
Move the tests from global wk2/TestExpectations to the ports' TestExpectations as
the core spellchecking implementation is delivered for wk2.
This change will allow to run spelling tests easier and using [ Pass ] keyword
in TestExpectations won't be needed.

  • platform/efl-wk2/TestExpectations:
  • platform/gtk-wk2/TestExpectations:
  • platform/mac-wk2/TestExpectations:
  • platform/qt-5.0-wk2/TestExpectations:
  • platform/wk2/TestExpectations:
12:41 PM Changeset in webkit [150653] by commit-queue@webkit.org
  • 4 edits in trunk/Source

Add more lines to the context during Dictionary lookups
https://bugs.webkit.org/show_bug.cgi?id=115256

Patch by Thomas Deniau <deniau@apple.com> on 2013-05-24
Reviewed by Ryosuke Niwa.

Source/WebCore:

No new tests - no new functionality.

  • WebCore.exp.in: Export a few methods in VisibleUnits needed to be able

to call previousLinePosition(), nextLinePosition(), startOfLine() and
endOfLine() in WebPageMac.mm when determining the context while looking
up a dictionary definition. These symbols used to be exported for iOS
only, export them on the desktop too.

Source/WebKit2:

  • WebProcess/WebPage/mac/WebPageMac.mm:

(WebKit::WebPage::performDictionaryLookupAtLocation):
Instead of using the current paragraph as the context (which stops at
line breaks), get a few lines before and after the point, so that the
context contains at least four lines in each direction.
This way the Dictionary will be able to look up things that extend past
the current line.

12:07 PM Changeset in webkit [150652] by andersca@apple.com
  • 18 edits
    17 deletes in trunk

Remove PagePopup code
https://bugs.webkit.org/show_bug.cgi?id=116732

Reviewed by Andreas Kling.

.:

  • Source/cmakeconfig.h.cmake:

Remove ENABLE_PAGE_POPUP.

Source/WebCore:

Remove all page popup code.

  • GNUmakefile.list.am:
  • WebCore.xcodeproj/project.pbxproj:
  • dom/ContextFeatures.cpp:
  • dom/ContextFeatures.h:
  • loader/EmptyClients.h:
  • page/ChromeClient.h:
  • page/DOMWindowPagePopup.cpp: Removed.
  • page/DOMWindowPagePopup.h: Removed.
  • page/DOMWindowPagePopup.idl: Removed.
  • page/PagePopup.h: Removed.
  • page/PagePopupClient.cpp: Removed.
  • page/PagePopupClient.h: Removed.
  • page/PagePopupController.cpp: Removed.
  • page/PagePopupController.h: Removed.
  • page/PagePopupController.idl: Removed.
  • page/PagePopupDriver.h: Removed.
  • testing/Internals.cpp:

(WebCore::Internals::resetToConsistentState):
(WebCore):

  • testing/Internals.h:

(WebCore):
(Internals):

  • testing/Internals.idl:
  • testing/MockPagePopupDriver.cpp: Removed.
  • testing/MockPagePopupDriver.h: Removed.

Source/WebKit/blackberry:

  • WebCoreSupport/AboutDataEnableFeatures.in:

Remove PAGE_POPUP.

Source/WTF:

Remove ENABLE_PAGE_POPUP.

  • wtf/FeatureDefines.h:

LayoutTests:

Remove tests that relied on the mock page popup driver.

  • fast/forms/date/no-page-popup-controller-expected.txt: Removed.
  • fast/forms/date/no-page-popup-controller.html: Removed.
  • fast/forms/resources/picker-common.js: Removed.
  • fast/forms/time-multiple-fields/time-multiple-fields-open-picker-key-bindings-expected.txt: Removed.
  • fast/forms/time-multiple-fields/time-multiple-fields-open-picker-key-bindings.html: Removed.
11:54 AM Changeset in webkit [150651] by jer.noble@apple.com
  • 13 edits
    3 copies
    1 add in trunk/Source

Mac: Set the default audio buffer size to a large value for <video> elements.
https://bugs.webkit.org/show_bug.cgi?id=116342

Source/WebCore:

Reviewed by Eric Carlson.
Roll-in rubber stamped by Eric Carlson.

To enable power savings by waking up the audio hardware less often, set the
requested buffer frame size to a large value, such as 4096. Since this results
in approximately 100ms worth of buffer, set the buffer size to a much lower
value when playing WebAudio, which is much more sensitive to latency than video
or audio elements.

Introduce a new class, AudioSessionManager, as well as a helper class,
AudioSessionManagerToken. Audio elements, video elements, and WebAudio destination
nodes will create and retain a token, and release the token in their destructor.
This allows the AudioSessionManager to track how many of what type of audio-
generating objects are in existence.

This requires implementing AudioSession for Mac platforms. Move the implementation
for retrieving the hardware sample rate and setting the buffer duration into
AudioSessionMac from AudioDestinationMac, to be shared with AudioSessionManagerMac.

Change the AudioSession method preferredBufferLength() into preferredBufferSize(),
as the callers really want to specify a buffer size, not a buffer duration. On iOS,
where the available API requires a duration, perform the conversion from duration to
size on behalf of the caller.

However, since the original version of this patch caused media test
failures on the Mac ML and Lion bots, only enable the buffer size
change for OS X > ML.

  • html/HTMLMediaElement.h: Add a AudioSessionManagerToken member.
  • html/HTMLMediaElement.cpp:

(WebCore::HTMLMediaElement::HTMLMediaElement): Initialize the token.

  • platform/audio/AudioSession.cpp:

(WebCore::AudioSession::preferredBufferSize): Renamed from preferredBufferLength.
(WebCore::AudioSession::setPreferredBufferSize): Renamed from setPreferredBufferLength.

  • platform/audio/AudioSession.h:
  • platform/audio/AudioSessionManager.cpp:

(AudioSessionManagerToken::create): Simple factory method.
(AudioSessionManagerToken::AudioSessionManagerToken): Call AudioSessionManager::incrementCount().
(AudioSessionManagerToken::~AudioSessionManagerToken): Call AudioSessionManager::decrementCount().
(AudioSessionManager::sharedManager): Simple singleton method.
(AudioSessionManager::AudioSessionManager): Simple constructor.
(AudioSessionManager::has): Return whether the type is present.
(AudioSessionManager::incrementCount): Increment, then call updateSessionState()
(AudioSessionManager::decrementCount): Decrement, then call updateSessionState()
(AudioSessionManager::updateSessionState): Stub, does nothing.

  • platform/audio/AudioSessionManager.h:
  • platform/audio/ios/AudioDestinationIOS.cpp:

(WebCore::AudioDestinationIOS::configure): Call setPreferredBufferSize() instead of setPreferredBufferLength()

  • platform/audio/ios/AudioSessionIOS.mm:

(WebCore::AudioSession::preferredBufferSize): Renamed from preferredBufferLength.
(WebCore::AudioSession::setPreferredBufferSize): Renamed from setPreferredBufferLength

  • platform/audio/mac/AudioDestinationMac.cpp:

(WebCore::AudioDestination::hardwareSampleRate): Call AudioSession::sampleRate().
(WebCore::AudioDestinationMac::AudioDestinationMac): Create the AudioSessionManagerToken.
(WebCore::AudioDestinationMac::configure): Do not set the buffer size (this is done in AudioSessionManagerMac).

  • platform/audio/mac/AudioDestinationMac.h:
  • platform/audio/mac/AudioSessionMac.cpp: Added.

(WebCore::defaultDevice): Added, returns the default audio device.
(WebCore::AudioSession::AudioSession): Simple constructor.
(WebCore::AudioSession::~AudioSession): Simple destructor.
(WebCore::AudioSession::category): Stub, unimplemented.
(WebCore::AudioSession::setCategory): Ditto.
(WebCore::AudioSession::categoryOverride): Ditto.
(WebCore::AudioSession::setCategoryOverride): Ditto.
(WebCore::AudioSession::numberOfOutputChannels): Ditto.
(WebCore::AudioSession::setActive): Ditto.
(WebCore::AudioSession::sampleRate): Use the HAL to return the default audio device sample rate.
(WebCore::AudioSession::preferredBufferSize): Return the current HAL setting.
(WebCore::AudioSession::setPreferredBufferSize): Set the buffer size.

  • platform/audio/mac/AudioSessionManagerMac.cpp:

(AudioSessionManager::updateSessionState): Set the buffer size depending on what audio outputs are present.

  • WebCore.xcodeproj/project.pbxproj: Add the new files to the project.

Source/WTF:

Reviewed by Eric Carlson.

  • wtf/Platform.h: Add a WTF_USE_AUDIO_SESSION setting.
11:46 AM Changeset in webkit [150650] by Lucas Forschler
  • 2 edits in branches/safari-537.43-branch/Source/WebCore

Merged r150631. <rdar://problem/13963712>

11:42 AM Changeset in webkit [150649] by Lucas Forschler
  • 7 edits in branches/safari-537.43-branch/Source/WebKit2

Merged r150620. <rdar://problem/13938510>

11:09 AM Changeset in webkit [150648] by andersca@apple.com
  • 2 edits in trunk/Source/WebKit2

Intermittent crash when typing in Address Bar
https://bugs.webkit.org/show_bug.cgi?id=116728
<rdar://problem/13976781>

Reviewed by Sam Weinig.

If a web process connection has been removed from the storage manager because we disconnected it
since there were no more active pages associated with the web process, we'll fail to find it here
so just return early instead of crashing.

  • UIProcess/Storage/StorageManager.cpp:

(WebKit::StorageManager::destroyStorageMap):

11:01 AM Changeset in webkit [150647] by Lucas Forschler
  • 4 edits in branches/safari-537.43-branch/Source

Versioning.

10:28 AM Changeset in webkit [150646] by commit-queue@webkit.org
  • 3 edits in trunk/Source/WebCore

Path: upstream the missing bits from the BlackBerry port
https://bugs.webkit.org/show_bug.cgi?id=116658

Patch by Alberto Garcia <agarcia@igalia.com> on 2013-05-24
Reviewed by Xan Lopez.

PathBlackBerry was upstreamed a while ago (r144612) but a few bits
were missing from the cross-platform code.

  • platform/graphics/Path.cpp:

(WebCore::Path::addPathForRoundedRect):

  • platform/graphics/Path.h:

(Graphics):
(Path):

10:16 AM Changeset in webkit [150645] by commit-queue@webkit.org
  • 3 edits in trunk/LayoutTests

Improve loader/go-back-cached-main-resource.html test
https://bugs.webkit.org/show_bug.cgi?id=116709

Patch by Manuel Rego Casasnovas <Manuel Rego Casasnovas> on 2013-05-24
Reviewed by Alexey Proskuryakov.

It is only needed to register the listeners on the initial load as they
are already present when restoring from page cache. Modified the test to
use event.persisted property to know if we are handling the first load
in order to register the listeners only once.

  • loader/resources/first-page.html: Rename registerListeners() function

to a more accurate name didShowPage() and use event.persisted to
register the listeners only in the first load.

  • loader/resources/other-page.html: Ditto.
9:50 AM WebKit Team edited by kbalazs@webkit.org
(diff)
9:46 AM Changeset in webkit [150644] by commit-queue@webkit.org
  • 4 edits in trunk/Source/JavaScriptCore

[sh4] Optimize LLINT generated code and fix few bugs in baseline JIT.
https://bugs.webkit.org/show_bug.cgi?id=116716

Patch by Julien Brianceau <jbrianceau@nds.com> on 2013-05-24
Reviewed by Geoffrey Garen.

  • assembler/MacroAssemblerSH4.h:

(JSC::MacroAssemblerSH4::mul32): Cosmetic changes.
(JSC::MacroAssemblerSH4::convertInt32ToDouble): Absolute address was not dereferenced.
(JSC::MacroAssemblerSH4::branch32): Absolute address was not dereferenced.
(JSC::MacroAssemblerSH4::revertJumpReplacementToBranchPtrWithPatch): Use all 32 bits of pointer for revertJump call.

  • assembler/SH4Assembler.h:

(JSC::SH4Assembler::revertJump): Use changePCrelativeAddress to patch the whole pointer.
(JSC::SH4Assembler::linkJump): Cosmetic change.

  • offlineasm/sh4.rb: Optimize LLINT generated code.
9:34 AM Changeset in webkit [150643] by commit-queue@webkit.org
  • 3 edits
    2 adds in trunk

WebProcess is crashing on http://achicu.github.io/css-presentation when direct pattern compositing is enabled
https://bugs.webkit.org/show_bug.cgi?id=115135

Patch by Noam Rosenthal <Noam Rosenthal> on 2013-05-24
Reviewed by Antti Koivisto.

Source/WebCore:

Issue is invoked when there is a visibility:hidden layer with a non-compositable
background image (e.g. a radial gradient).
The layer is treated as a no-contents layer because of its visibility, however
updateDirectlyCompositedBackgroundImage() later mistakenly treats it as a layer
with direct background-image compositing.

Fixed by returning early from updating directly composited contents for layers
with hidden contents.

Test: compositing/patterns/direct-pattern-compositing-hidden-radial-gradient.html

  • rendering/RenderLayerBacking.cpp:

(WebCore::RenderLayerBacking::updateDirectlyCompositedContents):
(WebCore::RenderLayerBacking::updateDirectlyCompositedBackgroundImage):
(WebCore::RenderLayerBacking::paintsChildren):
(WebCore::RenderLayerBacking::isDirectlyCompositedImage):

LayoutTests:

Added a ref-test that makes sure that a layer with a hidden radial gradient does not
cause the web process to crash.

  • compositing/patterns/direct-pattern-compositing-hidden-radial-gradient-expected.html: Added.
  • compositing/patterns/direct-pattern-compositing-hidden-radial-gradient.html: Added.
9:18 AM Changeset in webkit [150642] by robert@webkit.org
  • 2 edits in trunk/Source/WebCore

Refactor shouldAddBorderPaddingMargin()
https://bugs.webkit.org/show_bug.cgi?id=98803

Reviewed by Ryosuke Niwa.

Make this helper function less clever-stupid in its misguided sacrifice of intelligibility
for the sake of concision.

No new tests, refactoring.

  • rendering/RenderBlockLineLayout.cpp:

(WebCore::shouldAddBorderPaddingMargin):
(WebCore::inlineLogicalWidth):

8:51 AM Changeset in webkit [150641] by mifenton@rim.com
  • 2 edits in trunk/Source/WebKit/blackberry

[BlackBerry] Fix bad type warning in InputHandler log.
https://bugs.webkit.org/show_bug.cgi?id=116720

Reviewed by Carlos Garcia Campos.

Fix warning in inputLog string.

  • WebKitSupport/InputHandler.cpp:

(BlackBerry::WebKit::InputHandler::handleKeyboardInput):

7:48 AM Changeset in webkit [150640] by commit-queue@webkit.org
  • 2 edits in trunk/Tools

Update WTF::Vector pretty printers
https://bugs.webkit.org/show_bug.cgi?id=116719

Patch by Simon Pena <simon.pena@samsung.com> on 2013-05-24
Reviewed by Anders Carlsson.

After r148891, WTF::Vector pretty printers were left outdated, as
the WTF::Vector was made to extend a VectorBuffer, instead of keeping
one as a member. This patch updates the pretty printers so they work
again.

  • gdb/webkit.py:

(WTFVectorPrinter.children):
(WTFVectorPrinter.to_string):

7:20 AM Changeset in webkit [150639] by zarvai@inf.u-szeged.hu
  • 14 edits
    13 adds in trunk/LayoutTests

[Qt] Unreviewed gardening. Update qt-5.0-wk2 expected pictures after r150506.

Patch by Gabor Abraham <abrhm@inf.u-szeged.hu> on 2013-05-24

  • platform/qt-5.0-wk2/svg/dynamic-updates/SVGFEDisplacementMapElement-dom-in-attr-expected.png:
  • platform/qt-5.0-wk2/svg/dynamic-updates/SVGFEDisplacementMapElement-dom-in2-attr-expected.png:
  • platform/qt-5.0-wk2/svg/dynamic-updates/SVGFEDisplacementMapElement-dom-scale-attr-expected.png:
  • platform/qt-5.0-wk2/svg/dynamic-updates/SVGFEDisplacementMapElement-dom-xChannelSelector-attr-expected.png:
  • platform/qt-5.0-wk2/svg/dynamic-updates/SVGFEDisplacementMapElement-dom-yChannelSelector-attr-expected.png:
  • platform/qt-5.0-wk2/svg/dynamic-updates/SVGFEDisplacementMapElement-svgdom-in-prop-expected.png:
  • platform/qt-5.0-wk2/svg/dynamic-updates/SVGFEDisplacementMapElement-svgdom-in2-prop-expected.png:
  • platform/qt-5.0-wk2/svg/dynamic-updates/SVGFEDisplacementMapElement-svgdom-scale-prop-expected.png:
  • platform/qt-5.0-wk2/svg/dynamic-updates/SVGFEDisplacementMapElement-svgdom-xChannelSelector-prop-expected.png:
  • platform/qt-5.0-wk2/svg/dynamic-updates/SVGFEDisplacementMapElement-svgdom-yChannelSelector-prop-expected.png:
  • platform/qt-5.0-wk2/svg/dynamic-updates/SVGFEImageElement-dom-preserveAspectRatio-attr-expected.png:
  • platform/qt-5.0-wk2/svg/dynamic-updates/SVGFEImageElement-svgdom-preserveAspectRatio-prop-expected.png:
  • platform/qt-5.0-wk2/svg/filters/feImage-animated-transform-on-target-rect-expected.png:
  • platform/qt-5.0-wk2/svg/filters/feImage-late-indirect-update-expected.png: Added.
  • platform/qt-5.0-wk2/svg/filters/feImage-reference-invalidation-expected.png: Added.
  • platform/qt-5.0-wk2/svg/filters/feImage-target-add-to-document-expected.png: Added.
  • platform/qt-5.0-wk2/svg/filters/feImage-target-attribute-change-expected.png: Added.
  • platform/qt-5.0-wk2/svg/filters/feImage-target-attribute-change-with-use-indirection-2-expected.png: Added.
  • platform/qt-5.0-wk2/svg/filters/feImage-target-attribute-change-with-use-indirection-expected.png: Added.
  • platform/qt-5.0-wk2/svg/filters/feImage-target-changes-id-expected.png: Added.
  • platform/qt-5.0-wk2/svg/filters/feImage-target-id-change-expected.png: Added.
  • platform/qt-5.0-wk2/svg/filters/feImage-target-inline-style-change-expected.png: Added.
  • platform/qt-5.0-wk2/svg/filters/feImage-target-property-change-expected.png: Added.
  • platform/qt-5.0-wk2/svg/filters/feImage-target-reappend-to-document-expected.png: Added.
  • platform/qt-5.0-wk2/svg/filters/feImage-target-remove-from-document-expected.png: Added.
  • platform/qt-5.0-wk2/svg/filters/feImage-target-style-change-expected.png: Added.
6:25 AM Changeset in webkit [150638] by kadam@inf.u-szeged.hu
  • 3 edits in trunk/LayoutTests

[Qt] Unreviewed gardening. Skip failing tests.

  • platform/qt-5.0-wk2/TestExpectations: Move http/tests/security/cross-frame-access-put.html to qt/TestExpectations.
  • platform/qt/TestExpectations: Skip fast/frames/detach-frame-during-focus.html after r146523.
5:47 AM Changeset in webkit [150637] by mifenton@rim.com
  • 2 edits in trunk/Source/WebKit/blackberry

[BlackBerry] Respect tabindex when using form controls.
https://bugs.webkit.org/show_bug.cgi?id=116676

Reviewed by Xan Lopez.

PR 337419.

Update form control navigation tracking to take tabindex
order into account when calculating next and previous nodes.

The order to follow is tab index 1 - N, followed by all items
without a tab index or tab index 0 in rendering order to match
with standard desktop behavior.

Internally Reviewed by Genevieve Mak and Nima Ghanavatian.

  • WebKitSupport/InputHandler.cpp:

(BlackBerry::WebKit::InputHandler::updateFormState):

5:23 AM WebKitGTK/TrackingMemoryErrors edited by kbalazs@webkit.org
(diff)
1:41 AM Changeset in webkit [150636] by sfalken@apple.com
  • 2 edits in trunk/Tools

Update prepareConsistentTestingEnvironment to use a non-deprecated API.
https://bugs.webkit.org/show_bug.cgi?id=116710

Reviewed by Mark Rowe.

  • DumpRenderTree/mac/DumpRenderTree.mm:

(prepareConsistentTestingEnvironment): Use beginActivityWithOptions instead of beginSuspensionOfSystemBehaviors.

1:05 AM Changeset in webkit [150635] by commit-queue@webkit.org
  • 4 edits in trunk/Source/WebCore

[BlackBerry] DRT - crashed on GraphicsContext3D::makeContextCurrent
https://bugs.webkit.org/show_bug.cgi?id=116654

Patch by Xiaobo Wang <xiaobwang@blackberry.com> on 2013-05-24
Reviewed by Carlos Garcia Campos.

PR 331552.
Internally reviewed and amended by Arvid Nilsson.

No new tests, this patch is to fix DRT crashes in fast/canvas/webgl.
The instance of GraphicsContext3D has already been destroyed but
WebGLLayerWebKitThread is not aware of that. This is because compositor
layer of the WebGL graphics context was referenced outside the context.
So when destructing GraphicsContext3D the compositing layer was not
destructed.
Fix by setting m_webGLContext to 0 when destructing GrapthicsContext3D
after releasing the compositor layer, so that it won't be de-referenced
again later.

  • platform/graphics/blackberry/GraphicsContext3DBlackBerry.cpp:

(WebCore::GraphicsContext3D::~GraphicsContext3D):

  • platform/graphics/blackberry/WebGLLayerWebKitThread.cpp:

(WebCore::WebGLLayerWebKitThread::~WebGLLayerWebKitThread):
(WebCore::WebGLLayerWebKitThread::webGLContextDestroyed):
(WebCore):

  • platform/graphics/blackberry/WebGLLayerWebKitThread.h:

(WebGLLayerWebKitThread):

12:32 AM Changeset in webkit [150634] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebKit/blackberry

[BlackBerry] Remove SKIA leftovers from WebPage.cpp
https://bugs.webkit.org/show_bug.cgi?id=116656

Patch by Jeff Rogers <jrogers@rim.com> on 2013-05-24
Reviewed by Rob Buis.

  • Api/WebPage.cpp:

(BlackBerry::WebKit::WebPagePrivate::setLoadState):

12:12 AM Changeset in webkit [150633] by akling@apple.com
  • 6 edits in trunk/Source

Typing in Safari's unified field causes unnecessary web content repaints.
<http://webkit.org/b/116703>

Reviewed by Andy Estes.

Source/WebCore:

Break out an alternate version of Page::markAllMatchesForText() that only counts the number of times
a string occurs in the page, but doesn't try to mark the occurrences.
This allows Safari to count matches for the 'Find "foo" on This Page' functionality in the unified
location bar without causing tile repaints.

  • WebCore.exp.in:
  • page/Page.h:
  • page/Page.cpp:

(WebCore::Page::findMatchesForText):
(WebCore::Page::markAllMatchesForText):
(WebCore::Page::countFindMatches):

Renamed markAllMatchesForText() to findMatchesForText() and gave it a boolean parameter to control whether
matches should be marked or not. countFindMatches() is the new helper that doesn't mark.

Source/WebKit2:

  • WebProcess/WebPage/FindController.cpp:

(WebKit::FindController::countStringMatches):

Call WebCore::Page::countMatchesForText() instead of markAllMatchesForText(). If the search string
is found inside the page content, the renderer containing that text will no longer be repainted.

12:05 AM Changeset in webkit [150632] by rniwa@webkit.org
  • 3 edits in trunk/LayoutTests

Add back a failing test expectation on fast/canvas/webgl/object-deletion-behaviour.html
after r150558 on Lion Debug WK1.

  • platform/mac-wk2/TestExpectations:
  • platform/mac/TestExpectations:
12:02 AM Changeset in webkit [150631] by Beth Dakin
  • 2 edits in trunk/Source/WebCore

https://bugs.webkit.org/show_bug.cgi?id=116702
REGRESSION (r150399): Scrollbar missing for all overflow scroll
-and corresponding-
<rdar://problem/13963712>

Reviewed by Simon Fraser.

This was a silly copy-paste error.

  • rendering/RenderLayer.cpp:

(WebCore::RenderLayer::scrollbarAnimationsAreSuppressed):

May 23, 2013:

11:56 PM Changeset in webkit [150630] by commit-queue@webkit.org
  • 2 edits in trunk/Source/JavaScriptCore

CLoop llint backend should not use the d8 register as scratch register
https://bugs.webkit.org/show_bug.cgi?id=116019

Patch by Peter Wang <peter.wang@torchmobile.com.cn> on 2013-05-23
Reviewed by Csaba Osztrogonác.

  • offlineasm/cloop.rb:
11:26 PM Changeset in webkit [150629] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebKit/blackberry

[BlackBerry] Need to suspend/resume RootLayerCommit when the application becomes inactive/active
https://bugs.webkit.org/show_bug.cgi?id=115245

Patch by Xiaobo Wang <xiaobwang@blackberry.com> on 2013-05-23
Reviewed by Rob Buis.

PR 330917.
Internally reviewed by Arvid Nilsson.

  1. Suspend/resumeRootLayerCommit when notified app activation state

change.

  1. Schedule root layer commit in resumeRootLayerCommit() to explicitly

start root layer commit timer, so that there's a commit even if
BackingStore got disabled/removed.

  • Api/WebPage.cpp:

(BlackBerry::WebKit::WebPagePrivate::notifyAppActivationStateChange):
(BlackBerry::WebKit::WebPagePrivate::resumeRootLayerCommit):

11:23 PM Changeset in webkit [150628] by commit-queue@webkit.org
  • 4 edits in trunk/Source/WebKit/blackberry

[BlackBerry] Need to forward the opener frame url to client when creating a new window
https://bugs.webkit.org/show_bug.cgi?id=116566

Patch by Andy Chen <andchen@blackberry.com> on 2013-05-23
Reviewed by Rob Buis.

PR 337935
Internally reviewed by Arvid Nilsson.
When creating a new window, forward the opener frame url to client.

  • Api/WebPageClient.h:
  • WebCoreSupport/ChromeClientBlackBerry.cpp:

(WebCore::ChromeClientBlackBerry::createWindow):

  • WebCoreSupport/FrameLoaderClientBlackBerry.cpp:

(WebCore::FrameLoaderClientBlackBerry::dispatchCreatePage):

11:20 PM Changeset in webkit [150627] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebCore

[WinCairo] Build fails to link.
https://bugs.webkit.org/show_bug.cgi?id=116671

Patch by peavo@outlook.com <peavo@outlook.com> on 2013-05-23
Reviewed by Brent Fulgham.

  • page/FrameView.h: Added guard for ACCELERATED_COMPOSITING.
11:15 PM Changeset in webkit [150626] by lauro.neto@openbossa.org
  • 2 edits in trunk/Tools

Add missing dependencies to deps script.

[GTK] Missing packages in install-dependencies script.
https://bugs.webkit.org/show_bug.cgi?id=116632

Reviewed by Martin Robinson.

  • gtk/install-dependencies:
11:07 PM Changeset in webkit [150625] by commit-queue@webkit.org
  • 4 edits in trunk/Source/WebKit/blackberry

[BlackBerry] Compositor API change: Don't pass the unused "viewport" parameter.
https://bugs.webkit.org/show_bug.cgi?id=116545
Internal PR 189775

Patch by Jakob Petsovits <jpetsovits@blackberry.com> on 2013-05-23
Reviewed by Rob Buis.

Also change the name of the "documentContents" parameter
to the more descriptive "documentSrcRect".

  • Api/WebPageCompositor.cpp:

(BlackBerry::WebKit::WebPageCompositorPrivate::render):
(BlackBerry::WebKit::WebPageCompositor::render):

  • Api/WebPageCompositor.h:
  • Api/WebPageCompositor_p.h:

(WebPageCompositorPrivate):

10:14 PM Changeset in webkit [150624] by Csaba Osztrogonác
  • 2 edits in trunk/Source/WebKit2

[WK2][EFL][QT] Build break after r150610
https://bugs.webkit.org/show_bug.cgi?id=116699

Patch by Ryuan Choi <ryuan.choi@samsung.com> on 2013-05-23
Reviewed by Csaba Osztrogonác.

  • UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::findPlugin): Initialized pluginProcessSandboxPolicy.

7:59 PM Changeset in webkit [150623] by Lucas Forschler
  • 1 copy in tags/Safari-537.43.1

New Tag.

6:35 PM Changeset in webkit [150622] by fpizlo@apple.com
  • 11 edits in branches/dfgFourthTier/Source/JavaScriptCore

fourthTier: rationalize DFG::CapabilityLevel and DFGCapabilities.[h|cpp]
https://bugs.webkit.org/show_bug.cgi?id=116696

Reviewed by Sam Weinig.

Make it so that all capability calculation is funneled through one function, which tells
you everything you wanted to know: can it be inlined, and can it be compiled.

This work will help with https://bugs.webkit.org/show_bug.cgi?id=116557, since now the
JIT has a fairly authoritative answer to the "can it be inlined" question.

  • bytecode/CodeBlock.cpp:

(JSC::CodeBlock::CodeBlock):
(JSC::ProgramCodeBlock::capabilityLevelInternal):
(JSC::EvalCodeBlock::capabilityLevelInternal):
(JSC::FunctionCodeBlock::capabilityLevelInternal):

  • bytecode/CodeBlock.h:

(CodeBlock):
(JSC::CodeBlock::capabilityLevel):
(JSC::CodeBlock::capabilityLevelState):
(ProgramCodeBlock):
(EvalCodeBlock):
(FunctionCodeBlock):

  • dfg/DFGCapabilities.cpp:

(JSC::DFG::debugFail):
(DFG):
(JSC::DFG::canInlineResolveOperations):
(JSC::DFG::capabilityLevel):

  • dfg/DFGCapabilities.h:

(DFG):
(JSC::DFG::capabilityLevel):
(JSC::DFG::evalCapabilityLevel):
(JSC::DFG::programCapabilityLevel):
(JSC::DFG::functionForCallCapabilityLevel):
(JSC::DFG::functionForConstructCapabilityLevel):
(JSC::DFG::canInlineFunctionForCall):
(JSC::DFG::canInlineFunctionForClosureCall):
(JSC::DFG::canInlineFunctionForConstruct):

  • dfg/DFGCommon.h:

(JSC::DFG::canCompile):
(DFG):
(JSC::DFG::canInline):
(JSC::DFG::leastUpperBound):

  • dfg/DFGDriver.cpp:

(JSC::DFG::compile):

  • jit/JIT.cpp:

(JSC::JIT::privateCompile):

  • jit/JITPropertyAccess.cpp:

(JSC::JIT::privateCompilePutByIdTransition):

  • jit/JITPropertyAccess32_64.cpp:

(JSC::JIT::privateCompilePutByIdTransition):

  • tools/CodeProfile.cpp:

(JSC::CodeProfile::sample):

6:33 PM Changeset in webkit [150621] by Brent Fulgham
  • 2 edits in trunk/Source/WTF

[Windows] Rolling back r150600 as it breaks the VS2010 builds.

  • wtf/StackBounds.cpp: Rollback.
5:43 PM Changeset in webkit [150620] by andersca@apple.com
  • 7 edits in trunk/Source/WebKit2

Shim Flash plug-in to allow its updater to run
https://bugs.webkit.org/show_bug.cgi?id=116692
<rdar://problem/13938510>

Reviewed by Sam Weinig.

  • PluginProcess/PluginProcess.h:
  • PluginProcess/mac/PluginProcessMac.mm:

(WebKit::replacedNSConcreteTask_launch):
(WebKit::initializeCocoaOverrides):
Patch -[NSConcreteTask launch] to call PluginProcess::launchProcess.

(WebKit::PluginProcess::launchProcess):
Send a LaunchProcess message to the UI process.

  • UIProcess/Plugins/PluginProcessProxy.cpp:
  • UIProcess/Plugins/PluginProcessProxy.h:

Remove PluginProcessProxy::didReceiveMessage.

(PluginProcessProxy):

  • UIProcess/Plugins/PluginProcessProxy.messages.in:

Add LaunchProcess message and remove LegacyReceiver attribute.

  • UIProcess/Plugins/mac/PluginProcessProxyMac.mm:

(WebKit::isFlashUpdater):
Helper function to determine whether the process is the flash updater.

(WebKit::shouldLaunchProcess):
Return whether we should launch this process.

(WebKit::PluginProcessProxy::launchProcess):
Determine if we should launch this process and then do so.

5:41 PM Changeset in webkit [150619] by Brent Fulgham
  • 2 edits in trunk/Tools

[Windows] Unreviewed change for NRWT support.

  • Scripts/webkitpy/port/win.py:

(WinPort.default_child_processes): Added a default implementation
that limits NRWT to a single instance.

5:03 PM Changeset in webkit [150618] by roger_fong@apple.com
  • 3 edits in trunk/Source/WebCore

Unreviewed. Modify WebCore for AppleWin 2010 production build.

  • WebCore.vcxproj/WebCoreGenerated.make:

(WebCore::AXObjectCache::frameLoadingEventPlatformNotification):

  • platform/graphics/avfoundation/cf/MediaPlayerPrivateAVFoundationCF.cpp:
4:44 PM Changeset in webkit [150617] by commit-queue@webkit.org
  • 2 edits in trunk/LayoutTests

[EFL] Rebaseline after r150593
https://bugs.webkit.org/show_bug.cgi?id=116695

Unreviewed gardening.

Moved some editing tests from execCommand into pasteboard.

Patch by Seokju Kwon <Seokju Kwon> on 2013-05-23

  • platform/efl/TestExpectations:
4:22 PM Changeset in webkit [150616] by fpizlo@apple.com
  • 6 edits in trunk/LayoutTests

Make Math.imul tests run quicker, and fix some other bugs.

Rubber stampted by Oliver Hunt.

Also make tests not call print().

  • fast/js/regress/script-tests/cell-argument.js:
  • fast/js/regress/script-tests/imul-double-only.js:

(f):

  • fast/js/regress/script-tests/imul-int-only.js:

(f):

  • fast/js/regress/script-tests/imul-mixed.js:

(f):

  • fast/js/regress/script-tests/object-closure-call.js:
4:00 PM Changeset in webkit [150615] by rniwa@webkit.org
  • 2 edits in trunk/Tools

Remove meaningless "test_uses_apache" from Python test cases.
https://bugs.webkit.org/show_bug.cgi?id=38756

Patch by Brent Fulgham <Brent Fulgham> on 2013-05-23
Reviewed by Ryosuke Niwa.

  • Scripts/webkitpy/port/port_testcase.py:

(PortTestCase.test_driver_cmd_line): Remove the meaningless
"test_uses_apache" test case from the Python test infrastructure.
It was hard-coded to true, even though not all ports use
Apache.

2:44 PM Changeset in webkit [150614] by Lucas Forschler
  • 2 edits in branches/safari-537.43-branch/Source/WebKit2

Merged r150610. <rdar://problem/13975569>

2:17 PM Changeset in webkit [150613] by commit-queue@webkit.org
  • 3 edits in trunk/Source/WebCore

We need to clear main resource when detaching DocumentLoader from the frame.
https://bugs.webkit.org/show_bug.cgi?id=116680

Normally, when we detach the documentLoader in DocumentLoader::detachFromFrame, main resource is also cleared
in stopLoading(). There is possibility that main resource not being cleared, and this could cause crash later
since docLoader could still receive resource callbacks. To avoid that, we can remove docLoader from main resource's
client set in detachFromFrame.

Patch by Yongjun Zhang <yongjun_zhang@apple.com> on 2013-05-23
Reviewed by Brady Eidson.

No new tests, this doesn't happen on OS X.

  • loader/DocumentLoader.cpp:

(WebCore::DocumentLoader::detachFromFrame):
(WebCore::DocumentLoader::clearMainResource):

  • loader/cache/CachedResource.h: make hasClient accessible publicly.

(WebCore::CachedResource::hasClient):
(CachedResource):

1:59 PM Changeset in webkit [150612] by Brent Fulgham
  • 3 edits in trunk/Tools

[Windows] Update new-run-webkit-tests for Windows port.
https://bugs.webkit.org/show_bug.cgi?id=38756

Patch by Brent Fulgham <bfulgham@webkit.org> on 2013-05-23
Reviewed by Ryosuke Niwa.

  • Scripts/new-run-webkit-httpd:

(run): Switch from 'chromium' to 'win'

  • Scripts/webkitpy/port/win.py:

(WinPort): Add support methods for lighttpd server.
(WinPort._uses_apache):
(WinPort._path_to_lighttpd):
(WinPort._path_to_lighttpd_modules):
(WinPort._path_to_lighttpd_php):

1:56 PM Changeset in webkit [150611] by commit-queue@webkit.org
  • 3 edits
    2 adds in trunk
assertion failed: !node
node->isElementNode() in WebCore::RenderBlock::clone

https://bugs.webkit.org/show_bug.cgi?id=116614

Patch by Mihai Maerean <Mihai Maerean> on 2013-05-23
Reviewed by David Hyatt.

Source/WebCore:

Add a test for flow threads so they don't end up being cloned in region based multi columns with columns:1 where
the only child has column-span:all.

Test: fast/multicol/newmulticol/direct-child-column-span-all.html

  • rendering/RenderBlock.cpp:

(WebCore::RenderBlock::addChildIgnoringAnonymousColumnBlocks):

LayoutTests:

  • fast/multicol/newmulticol/direct-child-column-span-all-expected.txt: Added.
  • fast/multicol/newmulticol/direct-child-column-span-all.html: Added.
1:48 PM Changeset in webkit [150610] by andersca@apple.com
  • 2 edits in trunk/Source/WebKit2

Don't run all plug-ins unsandboxed
https://bugs.webkit.org/show_bug.cgi?id=116690
<rdar://problem/13975569>

Reviewed by Jessie Berlin.

Remove debugging code.

  • UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::findPlugin):

1:08 PM Changeset in webkit [150609] by andersca@apple.com
  • 7 edits in trunk/Source

Crash in convertMainResourceLoadToDownload when downloading file by option-return
https://bugs.webkit.org/show_bug.cgi?id=116451

Reviewed by Andreas Kling.

Source/WebCore:

Handle MainResourceLoader being null here which will happen when loading a cached main resource.

  • loader/DocumentLoader.cpp:

(WebCore::DocumentLoader::continueAfterContentPolicy):

Source/WebKit/mac:

If the main resource loader is null, create a new download instead since there's no existing loader to convert.

  • WebCoreSupport/WebFrameLoaderClient.mm:

(WebFrameLoaderClient::convertMainResourceLoadToDownload):

Source/WebKit2:

If the main resource loader is null, create a new download instead since there's no existing loader to convert.

  • NetworkProcess/NetworkConnectionToWebProcess.cpp:

(WebKit::NetworkConnectionToWebProcess::convertMainResourceLoadToDownload):

  • WebProcess/WebPage/WebFrame.cpp:

(WebKit::WebFrame::convertMainResourceLoadToDownload):

12:56 PM Changeset in webkit [150608] by rniwa@webkit.org
  • 2 edits in trunk/LayoutTests

Remove a stale test expectation.

  • platform/wk2/TestExpectations:
12:49 PM Changeset in webkit [150607] by rniwa@webkit.org
  • 2 edits in trunk/LayoutTests

Update Mac test expectations for the bug 85408 as bugs 116398 and 116477 have been merged into it.

  • platform/mac/TestExpectations:
12:46 PM Changeset in webkit [150606] by benjamin@webkit.org
  • 12 edits in trunk/Source/WebCore

KURL::createCFURL() should return a RetainPtr<CFURLRef>
https://bugs.webkit.org/show_bug.cgi?id=116644

Patch by Benjamin Poulain <bpoulain@apple.com> on 2013-05-23
Reviewed by Andreas Kling.

For consistency with the other createCFType() APIs in WebKit, KURL should
return a RetainPtr<> instead of raw +1 CFURLRef.

No leak were discovered while making this change.

  • platform/KURL.h:
  • platform/cf/KURLCFNet.cpp:

(WebCore):
(WebCore::createCFURLFromBuffer):
(WebCore::KURL::createCFURL):
(WebCore::KURL::fileSystemPath):

  • platform/graphics/avfoundation/cf/MediaPlayerPrivateAVFoundationCF.cpp:

(WebCore::AVFWrapper::createAssetForURL):

  • platform/graphics/cg/GraphicsContextCG.cpp:

(WebCore::GraphicsContext::setURLForRect):

  • platform/mac/KURLMac.mm:

(WebCore):
(WebCore::KURL::operator NSURL *):
(WebCore::KURL::createCFURL):

  • platform/network/cf/CookieJarCFNet.cpp:

(WebCore::setCookiesFromDOM):
(WebCore::cookiesForDOM):
(WebCore::cookieRequestHeaderFieldValue):
(WebCore::getRawCookies):
(WebCore::deleteCookie):

  • platform/network/cf/DNSCFNet.cpp:

(WebCore::DNSResolveQueue::platformProxyIsEnabledInSystemPreferences):

  • platform/network/cf/ProxyServerCFNet.cpp:

(WebCore::addProxyServersForURL):

  • platform/network/cf/ResourceRequestCFNet.cpp:

(WebCore::ResourceRequest::doUpdatePlatformRequest):
(WebCore::ResourceRequest::doUpdatePlatformHTTPBody):

  • platform/network/cf/ResourceResponseCFNet.cpp:

(WebCore::ResourceResponse::cfURLResponse):

  • platform/network/cf/SocketStreamHandleCFNet.cpp:

(WebCore::SocketStreamHandle::SocketStreamHandle):

12:42 PM Changeset in webkit [150605] by rniwa@webkit.org
  • 2 edits in trunk/LayoutTests

Add back failing test expectations for the bugs 88672 and 89528 after r150558 and r150559 on Lion.

  • platform/mac/TestExpectations:
12:37 PM Changeset in webkit [150604] by rniwa@webkit.org
  • 2 edits in trunk/LayoutTests

fast/workers/storage/use-same-database-in-page-and-workers.html times out on the lion bots
https://bugs.webkit.org/show_bug.cgi?id=81618

Add back the flaky test failure on Mac.

  • platform/mac/TestExpectations:
12:19 PM Changeset in webkit [150603] by Patrick Gansterer
  • 3 edits in trunk/Source/WTF

Use correct stack size on Solaris and OpenBSD
https://bugs.webkit.org/show_bug.cgi?id=114978

Reviewed by Oliver Hunt.

Original patch by David Hill <david@wmol.com>.

Use stack_t.ss_size for getting the size of the stack.

  • wtf/Platform.h:
  • wtf/StackBounds.cpp:

(WTF):
(WTF::StackBounds::initialize):

11:55 AM Changeset in webkit [150602] by robert@webkit.org
  • 5 edits
    2 adds in trunk

The ellipsis in a text overflow should not avoid floats
https://bugs.webkit.org/show_bug.cgi?id=115746

Reviewed by David Hyatt.

Source/WebCore:

When placing an ellipsis for text overflow we shouldn't avoid any floats on the line.
That is to say, if there is a float in the box we overflow we should let our no-wrap text,
including the ellipsis, render inside the float just as we would if overflow was hidden but
a text-overflow rendering had not been specified.

Test: fast/css/text-overflow-ellipsis-behind-floats.html

  • rendering/RenderBlockLineLayout.cpp:

(WebCore::RenderBlock::checkLinesForTextOverflow):

LayoutTests:

  • fast/css/text-overflow-ellipsis-behind-floats-expected.html: Added.
  • fast/css/text-overflow-ellipsis-behind-floats.html: Added.
  • fast/css/text-overflow-ellipsis-full-truncate-rtl-expected.html:
  • fast/css/text-overflow-ellipsis-full-truncate-rtl.html: Change this test to expect the text to cross into the floating input element and render the ellipsis at the left extreme of the element. The previous expected rendering relied on the incorrect premise that an overflow ellipsis should avoid floated elements.
11:51 AM Changeset in webkit [150601] by Lucas Forschler
  • 15 edits in branches/safari-537.43-branch/Source

Merged r150556. <rdar://problem/13918391>

11:51 AM Changeset in webkit [150600] by Patrick Gansterer
  • 2 edits in trunk/Source/WTF

[WIN] Implement correct detection of stack size
https://bugs.webkit.org/show_bug.cgi?id=116661

Reviewed by Oliver Hunt.

  • wtf/StackBounds.cpp:

(WTF):
(WTF::StackBounds::initialize):

11:49 AM Changeset in webkit [150599] by ap@apple.com
  • 2 edits in trunk/Source/WebKit2

<rdar://problem/13849260> Sandbox violations reading .CFUserTextEncoding

Reviewed by Dan Bernstein.

  • WebProcess/com.apple.WebProcess.sb.in: Allow this.
11:48 AM Changeset in webkit [150598] by Lucas Forschler
  • 2 edits in branches/safari-537.43-branch/Source/WebKit2

Merged r150543. <rdar://problem/13904054>

11:44 AM Changeset in webkit [150597] by Lucas Forschler
  • 6 edits in branches/safari-537.43-branch/Source/WebKit2

Merged r150537. <rdar://problem/13964515>

11:30 AM Changeset in webkit [150596] by rniwa@webkit.org
  • 3 edits in trunk/LayoutTests

[Lion] [WK2] compositing/overflow/overflow-compositing-descendant.html
https://bugs.webkit.org/show_bug.cgi?id=116684

Add a timeout/failure test expectation. Also update the expected result since the first frame
of the video must be shown in this test.

  • platform/mac-wk2/TestExpectations:
  • platform/mac/compositing/overflow/overflow-compositing-descendant-expected.png:
11:03 AM Changeset in webkit [150595] by barraclough@apple.com
  • 2 edits in trunk/Source/WebKit2

Move posix_spawn onto a zero delay timer
https://bugs.webkit.org/show_bug.cgi?id=116682

Reviewed by Anders Carlson.

<rdar://problem/13973468>

  • Shared/EntryPointUtilities/mac/XPCService/XPCServiceMain.Development.mm:

(ReexecInfo):

  • struct to hold info from the XPC message.

(WebKit::reexec):

  • moved call to posix_spawn out to this helper.

(WebKit::reexecCallBack):

  • calls reexec.

(WebKit::XPCServiceEventHandler):

  • removed call to posix_spawn; schedule a call to reexecCallBack.
10:52 AM Changeset in webkit [150594] by commit-queue@webkit.org
  • 4 edits in trunk/LayoutTests

Fix loader/go-back-cached-main-resource.html test
https://bugs.webkit.org/show_bug.cgi?id=116670

Patch by Manuel Rego Casasnovas <Manuel Rego Casasnovas> on 2013-05-23
Reviewed by Alexey Proskuryakov.

The page does not get an onload when being restored from page cache. The
test case case is modified to handle pageshow event instead.

  • loader/go-back-cached-main-resource.html: Wait for the third

'first-page' message to finish the test.

  • loader/resources/first-page.html: Use onpageshow instead of onload.
  • loader/resources/other-page.html: Ditto.
10:50 AM Changeset in webkit [150593] by rniwa@webkit.org
  • 1 edit
    13 moves in trunk/LayoutTests

Several editing tests that copy & paste slow down Mac WK2 EWS
https://bugs.webkit.org/show_bug.cgi?id=116681

Reviewed by Anders Carlsson.

Move editing tests that trigger copy and paste from execCommand into pasteboard.
Since NWRT uses a single instance of WebKitTestRunner to run tests in single directory,
this should mitigate the issue for now.

In the long term, we should really mock out the clipboard in WebKitTestRunner.

  • editing/execCommand/4128080-2-expected.txt: Removed.
  • editing/execCommand/4128080-2.html: Removed.
  • editing/execCommand/5939887-expected.txt: Removed.
  • editing/execCommand/5939887.html: Removed.
  • editing/execCommand/copy-without-selection-expected.txt: Removed.
  • editing/execCommand/copy-without-selection.html: Removed.
  • editing/execCommand/paste-2.html: Removed.
  • editing/pasteboard/copy-cut-paste-events-fired-when-disabled-expected.txt: Copied from LayoutTests/editing/execCommand/5939887-expected.txt.
  • editing/pasteboard/copy-cut-paste-events-fired-when-disabled.html: Copied from LayoutTests/editing/execCommand/5939887.html.
  • editing/pasteboard/copy-inside-h1-preserves-h1-expected.txt: Copied from LayoutTests/editing/execCommand/4128080-2-expected.txt.
  • editing/pasteboard/copy-inside-h1-preserves-h1.html: Copied from LayoutTests/editing/execCommand/4128080-2.html.
  • editing/pasteboard/copy-without-selection-expected.txt: Copied from LayoutTests/editing/execCommand/copy-without-selection-expected.txt.
  • editing/pasteboard/copy-without-selection.html: Copied from LayoutTests/editing/execCommand/copy-without-selection.html.
  • editing/pasteboard/paste-2.html: Copied from LayoutTests/editing/execCommand/paste-2.html.
  • platform/gtk/editing/execCommand/paste-2-expected.png: Removed.
  • platform/gtk/editing/execCommand/paste-2-expected.txt: Removed.
  • platform/gtk/editing/pasteboard/paste-2-expected.png: Copied from LayoutTests/platform/gtk/editing/execCommand/paste-2-expected.png.
  • platform/gtk/editing/pasteboard/paste-2-expected.txt: Copied from LayoutTests/platform/gtk/editing/execCommand/paste-2-expected.txt.
  • platform/mac/editing/execCommand/paste-2-expected.png: Removed.
  • platform/mac/editing/execCommand/paste-2-expected.txt: Removed.
  • platform/mac/editing/pasteboard/paste-2-expected.png: Copied from LayoutTests/platform/mac/editing/execCommand/paste-2-expected.png.
  • platform/mac/editing/pasteboard/paste-2-expected.txt: Copied from LayoutTests/platform/mac/editing/execCommand/paste-2-expected.txt.
  • platform/qt/editing/execCommand/paste-2-expected.png: Removed.
  • platform/qt/editing/execCommand/paste-2-expected.txt: Removed.
  • platform/qt/editing/pasteboard/paste-2-expected.png: Copied from LayoutTests/platform/qt/editing/execCommand/paste-2-expected.png.
  • platform/qt/editing/pasteboard/paste-2-expected.txt: Copied from LayoutTests/platform/qt/editing/execCommand/paste-2-expected.txt.
10:49 AM Changeset in webkit [150592] by Patrick Gansterer
  • 2 edits in trunk/Tools

Make default bug links in buildbot clickable
https://bugs.webkit.org/show_bug.cgi?id=116673

Reviewed by Ryosuke Niwa.

Use changecommentlink of html.WebStatus to make the links clickable.

  • BuildSlaveSupport/build.webkit.org-config/master.cfg:
10:39 AM WebKit Team edited by levin@chromium.org
(diff)
10:17 AM Changeset in webkit [150591] by Patrick Gansterer
  • 2 edits in trunk/Websites/bugs.webkit.org

EWS bubbles need more space
https://bugs.webkit.org/show_bug.cgi?id=116675

Reviewed by Andreas Kling.

  • template/en/custom/attachment/list.html.tmpl:
9:45 AM Changeset in webkit [150590] by Christophe Dumez
  • 13 edits in trunk

Unreviewed, rolling out r150586.
http://trac.webkit.org/changeset/150586
https://bugs.webkit.org/show_bug.cgi?id=116660

Those constructors should be exposed on the WorkerContext, not
the DOMWindow

Source/WebCore:

  • fileapi/FileReaderSync.idl:
  • workers/WorkerLocation.idl:

LayoutTests:

  • fast/js/global-constructors-attributes-expected.txt:
  • platform/efl/fast/js/constructor-length-expected.txt:
  • platform/efl/fast/js/global-constructors-attributes-expected.txt:
  • platform/gtk/fast/js/constructor-length-expected.txt:
  • platform/gtk/fast/js/global-constructors-attributes-expected.txt:
  • platform/mac-lion/fast/js/global-constructors-attributes-expected.txt:
  • platform/mac/fast/js/constructor-length-expected.txt:
  • platform/qt/fast/js/constructor-length-expected.txt:
  • platform/qt/fast/js/global-constructors-attributes-expected.txt:
9:13 AM Changeset in webkit [150589] by commit-queue@webkit.org
  • 2 edits in trunk/LayoutTests

[CSS Regions][Mac] [WK2] Layout Test fast/regions/fixed-pos-elem-in-region.html is flaky on Debug builders
https://bugs.webkit.org/show_bug.cgi?id=114571

Tests seems to be passing now, removed from TestExpectations.

Patch by Radu Stavila <stavila@adobe.com> on 2013-05-23
Reviewed by Anders Carlsson.

  • platform/mac-wk2/TestExpectations:
9:11 AM Changeset in webkit [150588] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebKit/blackberry

[BlackBerry] The web context does not contain any link properties if the context node is an image enclosed by a link node
https://bugs.webkit.org/show_bug.cgi?id=116627

Patch by Ed Baker <edbaker@blackberry.com> on 2013-05-23
Reviewed by Xan Lopez.

PR #341084

If the context node has an image tag and an enclosing link node was detected then add the link properties to the context.

Internally Reviewed by Gen Mak.

  • Api/WebPage.cpp:

(BlackBerry::WebKit::WebPagePrivate::webContext):

8:55 AM Changeset in webkit [150587] by mifenton@rim.com
  • 2 edits in trunk/Source/WebKit/blackberry

[BlackBerry] Validate form data before doing direct submission.
https://bugs.webkit.org/show_bug.cgi?id=116674

Reviewed by Xan Lopez.

PR 314202.

Check validation of the input form before triggering direct submission.

Internally Reviewed by Nima Ghanavatian.

  • WebKitSupport/InputHandler.cpp:

(BlackBerry::WebKit::InputHandler::submitForm):

8:20 AM Changeset in webkit [150586] by Christophe Dumez
  • 13 edits in trunk

Remove [NoInterfaceObject] from FileReaderSync and WorkerLocation
https://bugs.webkit.org/show_bug.cgi?id=116660

Reviewed by Kentaro Hara.

Source/WebCore:

Remove [NoInterfaceObject] from FileReaderSync and WorkerLocation to match their
respective specifications:

No new tests, covered by existing tests.

  • fileapi/FileReaderSync.idl:
  • workers/WorkerLocation.idl:

LayoutTests:

Rebaseline a couple of test cases now that FileReaderSync and WorkerLocation
attributes are exposed on the global Window object.

  • fast/js/global-constructors-attributes-expected.txt:
  • platform/efl/fast/js/constructor-length-expected.txt:
  • platform/efl/fast/js/global-constructors-attributes-expected.txt:
  • platform/gtk/fast/js/constructor-length-expected.txt:
  • platform/gtk/fast/js/global-constructors-attributes-expected.txt:
  • platform/mac-lion/fast/js/global-constructors-attributes-expected.txt:
  • platform/mac/fast/js/constructor-length-expected.txt:
  • platform/qt/fast/js/constructor-length-expected.txt:
  • platform/qt/fast/js/global-constructors-attributes-expected.txt:
7:14 AM Changeset in webkit [150585] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebKit/blackberry

[BlackBerry] Properly fill the ResourceError in FrameLoaderClientBlackBerry::cannotShowURLError
https://bugs.webkit.org/show_bug.cgi?id=116603

Patch by Rob Buis <rbuis@rim.com> on 2013-05-23
Reviewed by Xan Lopez.

PR 119789
Internally reviewed by Yong Li.

Provide a domain value for this ResourceError instance. Note that
this does not change behavior.

  • WebCoreSupport/FrameLoaderClientBlackBerry.cpp:

(WebCore::FrameLoaderClientBlackBerry::didRestoreFromPageCache):

7:00 AM Changeset in webkit [150584] by zarvai@inf.u-szeged.hu
  • 3 edits
    2 adds in trunk/LayoutTests

[Qt] Unreviewed gardening.

  • platform/qt-5.0-wk2/TestExpectations: Skip some failing test.
  • platform/qt-5.0-wk2/fast/forms/select-baseline-expected.png: Rebaselined after r148594.
  • platform/qt-5.0-wk2/fast/forms/select-baseline-expected.txt:
  • platform/qt-5.0-wk2/fast/images/repaint-subrect-grid-expected.txt: Rebaselined after r149292.
6:42 AM TestExpectations edited by simon.pena@samsung.com
Link to "Writing Reftests" to explain more on reference tests (diff)
6:17 AM Changeset in webkit [150583] by Csaba Osztrogonác
  • 2 edits in trunk/Source/WebKit2

[cmake] Fix generating-forwarding-headers rule after r150460
https://bugs.webkit.org/show_bug.cgi?id=116659

Reviewed by Christophe Dumez.

  • CMakeLists.txt: Typo fix.
5:45 AM Changeset in webkit [150582] by commit-queue@webkit.org
  • 3 edits in trunk/Source/WebKit2

[GTK] Remote Web Inspector sometimes won't serve any content
https://bugs.webkit.org/show_bug.cgi?id=116225

Patch by Anton Obzhirov <Anton Obzhirov> on 2013-05-23
Reviewed by Gustavo Noronha Silva.

Added an option to use environment variable to set up path to
inspector server files. Implementation is similar to InspectorClientGtk
implementation.

  • UIProcess/InspectorServer/WebInspectorServer.h:

(WebInspectorServer):

  • UIProcess/InspectorServer/gtk/WebInspectorServerGtk.cpp:

(WebKit::WebInspectorServer::platformResourceForPath):
(WebKit::WebInspectorServer::inspectorServerFilesPath):
(WebKit):

5:30 AM Changeset in webkit [150581] by gyuyoung.kim@samsung.com
  • 2 edits in trunk/Source/WebKit2

[EFL][WK2] Use WK2 C API in ewk_view_pagination_mode_set|get
https://bugs.webkit.org/show_bug.cgi?id=116657

Reviewed by Christophe Dumez.

WKPageGet|SetPaginationMode APIs is present in WK2 C APIs. So, we need to use them for
ewk_view_pagination_mode_set|get.

Besides we don't need to use WebCore namespace anymore.

  • UIProcess/API/efl/ewk_view.cpp:

(ewk_view_pagination_mode_set):
(ewk_view_pagination_mode_get):

5:22 AM Changeset in webkit [150580] by Christophe Dumez
  • 13 edits in trunk/Source/WebCore

Remove special case for MessagePortArray from bindings generator
https://bugs.webkit.org/show_bug.cgi?id=116653

Reviewed by Kentaro Hara.

Use MessagePort[] instead of MessagePortArray type in the IDL as MessagePortArray
is not defined on IDL side. This allows us to remove the special case for
MessagePortArray from the bindings generator.

No new tests, no behavior change.

  • bindings/js/JSMessageEventCustom.cpp:

(WebCore::JSMessageEvent::data):

  • bindings/scripts/CodeGeneratorJS.pm:

(NativeToJSValue): Remove special case for MessagePortArray.

  • bindings/scripts/test/CPP/WebDOMTestSerializedScriptValueInterface.cpp:
  • bindings/scripts/test/CPP/WebDOMTestSerializedScriptValueInterface.h:

(WebDOMTestSerializedScriptValueInterface):

  • bindings/scripts/test/GObject/WebKitDOMTestSerializedScriptValueInterface.cpp:

(webkit_dom_test_serialized_script_value_interface_get_property):
(webkit_dom_test_serialized_script_value_interface_class_init):

  • bindings/scripts/test/GObject/WebKitDOMTestSerializedScriptValueInterface.h:
  • bindings/scripts/test/JS/JSTestSerializedScriptValueInterface.cpp:

(WebCore::jsTestSerializedScriptValueInterfacePorts):

  • bindings/scripts/test/ObjC/DOMTestSerializedScriptValueInterface.h:
  • bindings/scripts/test/ObjC/DOMTestSerializedScriptValueInterface.mm:
  • bindings/scripts/test/TestSerializedScriptValueInterface.idl: Replace MessagePortArray

by MessagePort[].

  • dom/MessageEvent.h:

(WebCore::MessageEvent::ports): The method should return a Vector, not a Vector*
as the array on JavaScript side is not nullable.

  • dom/MessageEvent.idl: Replace MessagePortArray by MessagePort[].
3:32 AM Changeset in webkit [150579] by kbalazs@webkit.org
  • 3 edits in trunk/LayoutTests

media/W3C/video/networkState/networkState_during_progress.html is flaky
https://bugs.webkit.org/show_bug.cgi?id=76280

Unreviewed gardening.

Change expectation to the correct output and mark as flaky.

  • TestExpectations:
  • media/W3C/video/networkState/networkState_during_progress-expected.txt:
2:26 AM Changeset in webkit [150578] by kadam@inf.u-szeged.hu
  • 4 edits in trunk/Source/WebKit2

Support frame instrumentation for coordinated graphics
https://bugs.webkit.org/show_bug.cgi?id=115260

Patch by Roland Takacs <rtakacs@inf.u-szeged.hu> on 2013-05-23
Reviewed by Andreas Kling.

Enabled supportsFrameInstrumentation on ports that use coordinated graphics
by calling didBeginFrame().

  • WebProcess/WebCoreSupport/WebInspectorClient.cpp:

(WebKit::WebInspectorClient::supportsFrameInstrumentation):
(WebKit):

  • WebProcess/WebCoreSupport/WebInspectorClient.h:

(WebInspectorClient):

  • WebProcess/WebPage/CoordinatedGraphics/CoordinatedLayerTreeHost.cpp:

(WebKit::CoordinatedLayerTreeHost::syncDisplayState):

2:20 AM Changeset in webkit [150577] by zarvai@inf.u-szeged.hu
  • 3 edits in trunk/LayoutTests

[Qt] Unreviewed gardening.

  • platform/qt-5.0-wk1/TestExpectations: Remove a test failure that skipped in qt.
  • platform/qt/TestExpectations: Unskip now passing test.
1:39 AM Changeset in webkit [150576] by zarvai@inf.u-szeged.hu
  • 3 edits
    1 add in trunk/LayoutTests

[Qt] Unreviewed gardening.

  • platform/qt-5.0-wk2/TestExpectations: Unskip after rebaseline.
  • platform/qt/TestExpectations: Skip flaky test.
  • platform/qt/plugins/mouse-events-expected.txt: Rebaselining after r150558.
12:06 AM Changeset in webkit [150575] by commit-queue@webkit.org
  • 2 edits in trunk/Source/WebCore

[EFL] Back, Forward and Stop should be hidden in context menu
https://bugs.webkit.org/show_bug.cgi?id=116615

Patch by Michał Pakuła vel Rutka <Michał Pakuła vel Rutka> on 2013-05-23
Reviewed by Christophe Dumez.

Context menu items: Back, Forward, Stop, Reload, should be shown only when
they are able to perform action.

No new tests - no new functionality.

  • page/ContextMenuController.cpp:

(WebCore::ContextMenuController::populate):

Note: See TracTimeline for information about the timeline view.