WebKit-https.git
6 years agoUnreviewed. Some AppleWin VS2010 house cleaning.
roger_fong@apple.com [Tue, 4 Jun 2013 01:44:52 +0000 (01:44 +0000)]
Unreviewed. Some AppleWin VS2010 house cleaning.

* WebKit.vcxproj/Interfaces/InterfacesPreBuild.cmd:
* WebKit.vcxproj/WebKit/WebKitPreBuild.cmd:
* win/tools32/scripts/auto-version.sh: Replaced with win/tools32/scripts/auto-version2010.sh.
* win/tools32/scripts/auto-version2010.sh: Removed.
* win/tools32/scripts/feature-defines.sh: Replaced with win/tools32/scripts/feature-defines2010.sh.
* win/tools32/scripts/feature-defines2010.sh: Removed.
* WebCore.vcxproj/build-generated-files.sh:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151147 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoclearLayoutOverflow should never be called before calling layer()->updateScrollInfoAf...
roger_fong@apple.com [Tue, 4 Jun 2013 01:16:42 +0000 (01:16 +0000)]
clearLayoutOverflow should never be called before calling layer()->updateScrollInfoAfterLayout().
https://bugs.webkit.org/show_bug.cgi?id=116689.

Reviewed by Ojan Vafai.

Make sure that clearLayoutOverflow only gets called after layer()->updateScrollInfoAfterLayout().
Also move clearLayoutOverflow to RenderBlock so we can keep all the delayed scroll update logic together.

Test: fast/flexbox/clear-overflow-before-scroll-update.html

* rendering/RenderBlock.cpp:
(WebCore::RenderBlock::layout): Remove call to clearLayoutOverflow.
(WebCore::RenderBlock::finishDelayUpdateScrollInfo): Add call to clearLayoutOverflow.
(WebCore::RenderBlock::updateScrollInfoAfterLayout): Ditto.
* rendering/RenderBlock.h:
(RenderBlock):
* rendering/RenderBox.cpp:
* rendering/RenderBox.h:
(WebCore::RenderBox::layoutOverflowRect): Remove from layoutOverflowRect.
(RenderBox):
* rendering/RenderDeprecatedFlexibleBox.cpp:
(WebCore::RenderDeprecatedFlexibleBox::layoutBlock): This should call updateScrollInfoAfterLayout() intead of directly calling layer()->updateScrollInfoAfterLayout().
* rendering/RenderGrid.cpp:
(WebCore::RenderGrid::layoutBlock): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151146 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[EFL] Add ATK version 2.8.0 to efl jhbuild moduleset
commit-queue@webkit.org [Tue, 4 Jun 2013 01:09:37 +0000 (01:09 +0000)]
[EFL] Add ATK version 2.8.0 to efl jhbuild moduleset
https://bugs.webkit.org/show_bug.cgi?id=116726

Patch by Eduardo Lima Mitev <elima@igalia.com> on 2013-06-03
Reviewed by Gyuyoung Kim.

.:

Bump required version of ATK to 2.8.0 in EFL CMake's build.

* Source/cmake/FindATK.cmake: Adds macro to check required version
* Source/cmake/OptionsEfl.cmake: Specifies required version of ATK to be 2.8.0

Tools:

Current version of ATK in efl-ews environment is 2.4.0 and is taken from
the platform (it is not listed in jhbuild moduleset). That version is
old-ish and contain deprecated symbols and missing API, that are blocking
patches in other ports (gtk). This patch adds ATK version 2.8.0 to efl
moduleset, which is the same version the gtk port uses.

See bug https://bugs.webkit.org/show_bug.cgi?id=115647 for more details on
why this is needed.

* efl/jhbuild.modules: Adds ATK module as dependency

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151145 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[harfbuzz] WebKit fails to build with MinGW compiler because of invalid cast in HarfB...
commit-queue@webkit.org [Mon, 3 Jun 2013 22:47:43 +0000 (22:47 +0000)]
[harfbuzz] WebKit fails to build with MinGW compiler because of invalid cast in HarfBuzzShaper.cpp
https://bugs.webkit.org/show_bug.cgi?id=108373

Patch by Paweł Forysiuk <tuxator@o2.pl> on 2013-06-03
Reviewed by Darin Adler.

Fix building on platforms where UChar is defined as wchar_t.

* platform/graphics/harfbuzz/HarfBuzzShaper.cpp:
(WebCore::HarfBuzzShaper::shapeHarfBuzzRuns):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151141 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoREGRESSION: Undoing autocorrection doesn't prevent a future autocorrection
rniwa@webkit.org [Mon, 3 Jun 2013 22:45:11 +0000 (22:45 +0000)]
REGRESSION: Undoing autocorrection doesn't prevent a future autocorrection
https://bugs.webkit.org/show_bug.cgi?id=117165

Reviewed by Darin Adler.

Source/WebCore:

The bug was caused by SpellingCorrectionCommand returning true in callsAppliedEditingInDoApply, which prevented
the command from being registered to the undo manager. This behavior was introduced in r148306 because TypingCommand
registeres itself to the undo manager in appliedEditing.

Fixed the bug by reverting both r122176 and r148306 and implementing a better fix for the bug 90487.
Instead of using InsertTextCommand, simply don't specify ReplaceSelectionCommand::PreventNesting so that the command
doesn't try to break out of a blockquote.

Also fix a bug in ReplaceSelectionCommand::completeHTMLReplacement uncovered by this change. We were only merging
text nodes at the beginning of the replaced content but not at the end. Call mergeTextNodesAroundPosition for the
second time to fix whitespace at the end of the replaced content.

Test: platform/mac/editing/spelling/autocorrection-respets-undo.html

* editing/CompositeEditCommand.cpp: Explictily check isTypingCommand now that callsAppliedEditingInDoApply has been
removed from SpellingCorrectionCommand and TypingCommand and SpellingCorrectionCommand were only classes that
returned true in callsAppliedEditingInDoApply. See r148306 that introduced this function.
(WebCore::CompositeEditCommand::apply):
* editing/CompositeEditCommand.h: Removed callsAppliedEditingInDoApply.
* editing/ReplaceSelectionCommand.cpp:
(WebCore::ReplaceSelectionCommand::completeHTMLReplacement):
* editing/SpellingCorrectionCommand.cpp: Removed callsAppliedEditingInDoApply, which was the culprit of this bug,
so that CompositeEditCommand::apply correctly registers this command with the undo manager.
(WebCore::SpellingCorrectionCommand::doApply): Use ReplaceSelectionCommand instead of TypingCommand as noted above.
* editing/SpellingCorrectionCommand.h:
* editing/TypingCommand.cpp: Deleted callsAppliedEditingInDoApply.
* editing/TypingCommand.h:

LayoutTests:

Add a regression test and rebaseline other tests as ReplaceSelectionCommand now merges text nodes at
the end of replaced text.

* editing/deleting/merge-paragraph-from-address-expected.txt:
* editing/deleting/merge-paragraph-from-h6-expected.txt:
* editing/deleting/merge-paragraph-from-listing-expected.txt:
* editing/deleting/paste-with-transparent-background-color-expected.txt:
* editing/inserting/insert-3907422-fix-expected.txt:
* editing/pasteboard/copy-image-with-alt-text-expected.txt:
* editing/pasteboard/merge-end-4-expected.txt:
* editing/pasteboard/nested-blocks-with-text-area-expected.txt:
* editing/pasteboard/nested-blocks-with-text-field-expected.txt:
* editing/pasteboard/paste-text-001-expected.txt:
* editing/pasteboard/paste-unrendered-select-expected.txt:
* editing/pasteboard/smart-paste-003-trailing-whitespace-expected.txt:
* platform/mac/editing/deleting/delete-block-merge-contents-001-expected.txt:
* platform/mac/editing/deleting/delete-block-merge-contents-019-expected.txt:
* platform/mac/editing/deleting/delete-block-merge-contents-020-expected.txt:
* platform/mac/editing/pasteboard/4944770-2-expected.txt:
* platform/mac/editing/pasteboard/8145-2-expected.txt:
* platform/mac/editing/pasteboard/paste-line-endings-006-expected.txt:
* platform/mac/editing/pasteboard/paste-text-008-expected.txt:
* platform/mac/editing/pasteboard/paste-text-at-tabspan-003-expected.txt:
* platform/mac/editing/pasteboard/paste-xml-expected.txt:
* platform/mac/editing/spelling/autocorrection-blockquote-crash-expected.txt:
* platform/mac/editing/spelling/autocorrection-blockquote-crash.html: Added more asssertions to ensure that
the autocorrected text is in the blockquote. r122176 modified the test description saying that this test
already checks this condition but it never did so explicitly; it assumed that the autocorrected text was
inside a blockquote as long as it started at a certain offset in the text node.
* platform/mac/editing/spelling/autocorrection-respets-undo-expected.txt: Added.
* platform/mac/editing/spelling/autocorrection-respets-undo.html: Added.
* platform/mac/editing/spelling/autocorrection-with-multi-line-text-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151140 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoReduce indentation in Widget::paint in WidgetMac.mm
andersca@apple.com [Mon, 3 Jun 2013 22:29:18 +0000 (22:29 +0000)]
Reduce indentation in Widget::paint in WidgetMac.mm
https://bugs.webkit.org/show_bug.cgi?id=117171

Reviewed by Darin Adler.

Use an early return for the common case to reduce indentation in the complex case.

* platform/mac/WidgetMac.mm:
(WebCore::Widget::paint):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151138 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoUnreviewed. AppleWin VS2010 build fix.
roger_fong@apple.com [Mon, 3 Jun 2013 22:28:08 +0000 (22:28 +0000)]
Unreviewed. AppleWin VS2010 build fix.

* WebCore.vcxproj/migrate-scripts.sh:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151136 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[GTK] Respect PKG_CONFIG env variable when generating gtk-doc
commit-queue@webkit.org [Mon, 3 Jun 2013 22:25:19 +0000 (22:25 +0000)]
[GTK] Respect PKG_CONFIG env variable when generating gtk-doc
https://bugs.webkit.org/show_bug.cgi?id=117168

Patch by Erik van Pienbroek <epienbro@fedoraproject.org> on 2013-06-03
Reviewed by Martin Robinson.

The tool generate-gtkdoc internally calls pkg-config.
When cross-compiling it shouldn't use pkg-config, but
it should use the pkg-config which was detected by
the ./configure script (for example this can be
i686-w64-mingw32-pkg-config when cross-compiling)

* GNUmakefile.am:
* gtk/gtkdoc.py:
(PkgConfigGTKDoc.__init__):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151135 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[Windows] Unreviewed build fix.
bfulgham@apple.com [Mon, 3 Jun 2013 22:23:00 +0000 (22:23 +0000)]
[Windows] Unreviewed build fix.

* TestWebKitAPI/win/main.cpp: Embed proper manifest information to
allow VS2010 builds to use VS2005 runtime libraries as needed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151134 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoNuke VS2005 files from the tree.
roger_fong@apple.com [Mon, 3 Jun 2013 22:13:07 +0000 (22:13 +0000)]
Nuke VS2005 files from the tree.
<rdar://problem/14042021>.

Rubberstamped by Brent Fulgham.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151133 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[WK2] REGRESSION(r150798): scrolling coordinator is not being enabled when WebPage...
marcelo.lira@openbossa.org [Mon, 3 Jun 2013 22:06:39 +0000 (22:06 +0000)]
[WK2] REGRESSION(r150798): scrolling coordinator is not being enabled when WebPage is set to use fixed layout
https://bugs.webkit.org/show_bug.cgi?id=117163

Reviewed by Darin Adler.

Moved call to WebPage::settings()->setScrollingCoordinatorEnabled()
from CoordinatedLayerTreeHost constructor to WebPage::setUseFixedLayout().
The scrolling coordinator must be enabled or disabled following the
WebPage's use of fixed layout.

* WebProcess/WebPage/CoordinatedGraphics/CoordinatedLayerTreeHost.cpp:
(WebKit::CoordinatedLayerTreeHost::CoordinatedLayerTreeHost):
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::setUseFixedLayout):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151132 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoAX: WebKit should join substrings on single space when concatenating the accessible...
cfleizach@apple.com [Mon, 3 Jun 2013 21:51:28 +0000 (21:51 +0000)]
AX: WebKit should join substrings on single space when concatenating the accessible name of an element
https://bugs.webkit.org/show_bug.cgi?id=117100

Reviewed by Darin Adler.

Source/WebCore:

When concatenating text from underneath an element append spaces between objects and trim whitespace.

Modified existing tests reflect the new behavior.

* accessibility/AccessibilityNodeObject.cpp:
(WebCore::AccessibilityNodeObject::textUnderElement):

LayoutTests:

* accessibility/text-role-with-aria-hidden-inside-expected.txt:
* accessibility/text-role-with-aria-hidden-inside.html:
* platform/mac/accessibility/internal-link-anchors2-expected.txt:
* platform/mac/accessibility/table-with-aria-role-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151131 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[BlackBerry] Expose show debug borders setting through BlackBerry::WebKit::WebPage
anlo@rim.com [Mon, 3 Jun 2013 21:47:12 +0000 (21:47 +0000)]
[BlackBerry] Expose show debug borders setting through BlackBerry::WebKit::WebPage
https://bugs.webkit.org/show_bug.cgi?id=117167

Reviewed by Rob Buis.
Internally reviewed by Arvid Nilsson.

Add WebKit::WebPage API for setting whether to show accelerated compositing
debug borders.

* Api/WebPage.cpp:
(BlackBerry::WebKit::WebPage::setShowDebugBorders):
(WebKit):
* Api/WebPage.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151129 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoRemove XMLMIMETypeRegExp
zandobersek@gmail.com [Mon, 3 Jun 2013 21:03:36 +0000 (21:03 +0000)]
Remove XMLMIMETypeRegExp
https://bugs.webkit.org/show_bug.cgi?id=116861

Reviewed by Alexey Proskuryakov.

Remove the XMLMIMETypeRegExp class. It contained a regular expression that was used
for checking for valid XML MIME types in DOMImplementation::isXMLMIMEType. The thread-global
instance was stored in ThreadGlobalData.

The class is removed, with DOMImplementation::isXMLMIMEType now testing that the given MIME type
is of proper form and contains only valid characters.

No new tests - no change in functionality.

* dom/DOMImplementation.cpp:
(WebCore::isValidXMLMIMETypeChar): A helper method that returns true only for characters that
are acceptable in a valid XML MIME type (apart from the forward slash).
(WebCore::DOMImplementation::isXMLMIMEType): Instead of using a regular expression, determine
whether the given string is a valid non-standard XML mime type by checking for the '+xml' ending,
proper forward slash position, and that no invalid characters are used in the string.
* dom/DOMImplementation.h:
* platform/ThreadGlobalData.cpp:
(WebCore::ThreadGlobalData::ThreadGlobalData): The OwnPtr<XMLMIMETypeRegExp> member variable is being removed.
(WebCore::ThreadGlobalData::destroy): Ditto.
* platform/ThreadGlobalData.h:
(WebCore): No need for the XMLMIMETypeRegExp forward declaration.
(ThreadGlobalData): Remove the OwnPtr<XMLMIMETypeRegExp> member variable.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151126 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[GTK] Use drawErrorUnderline() instead of Pango API for highlighting misspelled words
commit-queue@webkit.org [Mon, 3 Jun 2013 20:52:31 +0000 (20:52 +0000)]
[GTK] Use drawErrorUnderline() instead of Pango API for highlighting misspelled words
https://bugs.webkit.org/show_bug.cgi?id=115406

Patch by Eduardo Lima Mitev <elima@igalia.com> on 2013-06-03
Reviewed by Martin Robinson.

Source/WebCore:

Replace pango API for drawing underline error on misspelled words, in favor of
drawErrorUnderline() method, removing the platform #ifdef check. As this was the
only usage of Pango in GraphicsContextCairo, it no longer includes pango.h.

* platform/graphics/cairo/GraphicsContextCairo.cpp:
(WebCore::GraphicsContext::drawLineForDocumentMarker): Removes
pango_cairo_show_error_underline() in favor of drawErrorUnderline().
* platform/graphics/cairo/DrawErrorUnderline.h:
(drawErrorUnderline): Fills the cairo path at the end of the drawing.

LayoutTests:

Update PNG expectations for some spell checking tests to pass pixel hash comparison.
The images are not visually modified.

* platform/gtk/editing/spelling/inline-spelling-markers-hidpi-expected.png: Rebaseline
image to pass pixel hash
* platform/gtk/editing/spelling/inline_spelling_markers-expected.png: Rebasline
image to pass pixel hash
* platform/gtk/editing/spelling/spelling-expected.png: Rebasline
image to pass pixel hash

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151125 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoAX: MathML multiscripts not exposed in Accessibility
cfleizach@apple.com [Mon, 3 Jun 2013 19:41:22 +0000 (19:41 +0000)]
AX: MathML multiscripts not exposed in Accessibility
https://bugs.webkit.org/show_bug.cgi?id=115069

Reviewed by Tim Horton.

Add methods to help output components of math multiscripts for accessibility.

* DumpRenderTree/AccessibilityUIElement.cpp:
(mathPostscriptsDescriptionCallback):
(mathPrescriptsDescriptionCallback):
(AccessibilityUIElement::getJSClass):
* DumpRenderTree/AccessibilityUIElement.h:
(AccessibilityUIElement):
* DumpRenderTree/mac/AccessibilityUIElementMac.mm:
(convertMathMultiscriptPairsToString):
(AccessibilityUIElement::mathPostscriptsDescription):
(AccessibilityUIElement::mathPrescriptsDescription):
* WebKitTestRunner/InjectedBundle/AccessibilityUIElement.cpp:
(WTR::AccessibilityUIElement::mathPostscriptsDescription):
(WTR::AccessibilityUIElement::mathPrescriptsDescription):
* WebKitTestRunner/InjectedBundle/AccessibilityUIElement.h:
(AccessibilityUIElement):
* WebKitTestRunner/InjectedBundle/Bindings/AccessibilityUIElement.idl:
* WebKitTestRunner/InjectedBundle/mac/AccessibilityUIElementMac.mm:
(WTR::_convertMathMultiscriptPairsToString):
(WTR::AccessibilityUIElement::mathPostscriptsDescription):
(WTR::AccessibilityUIElement::mathPrescriptsDescription):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151124 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoAX: MathML multiscripts not exposed in Accessibility
cfleizach@apple.com [Mon, 3 Jun 2013 19:35:01 +0000 (19:35 +0000)]
AX: MathML multiscripts not exposed in Accessibility
https://bugs.webkit.org/show_bug.cgi?id=115069

Reviewed by Tim Horton.

Source/WebCore:

Math multiscript tags allow for an arbitrary number of prescripts and postscripts
around a base object.

This MathML tag currently is not supported in the render tree, but that doesn't need to
stop it from being exposed through accessibility, as a base object with an array of postscripts
and an array of prescripts.

Test: platform/mac/accessibility/mathml-multiscript.html

* accessibility/AccessibilityObject.h:
(WebCore::AccessibilityObject::isMathMultiscript):
(AccessibilityObject):
(WebCore::AccessibilityObject::mathPrescripts):
(WebCore::AccessibilityObject::mathPostscripts):
* accessibility/AccessibilityRenderObject.cpp:
(WebCore::AccessibilityRenderObject::isMathMultiscript):
(WebCore::AccessibilityRenderObject::isIgnoredElementWithinMathTree):
(WebCore::AccessibilityRenderObject::mathBaseObject):
(WebCore::AccessibilityRenderObject::mathPrescripts):
(WebCore::AccessibilityRenderObject::mathPostscripts):
* accessibility/AccessibilityRenderObject.h:
(AccessibilityRenderObject):
* accessibility/ios/WebAccessibilityObjectWrapperIOS.mm:
(-[WebAccessibilityObjectWrapper accessibilityPlatformMathSubscriptKey]):
(-[WebAccessibilityObjectWrapper accessibilityPlatformMathSuperscriptKey]):
(-[WebAccessibilityObjectWrapper accessibilityMathPostscripts]):
(-[WebAccessibilityObjectWrapper accessibilityMathPrescripts]):
(-[WebAccessibilityObjectWrapper accessibilityMathOverObject]):
(-[WebAccessibilityObjectWrapper accessibilityMathType]):
* accessibility/mac/WebAccessibilityObjectWrapperBase.h:
* accessibility/mac/WebAccessibilityObjectWrapperBase.mm:
(convertMathPairsToNSArray):
(-[WebAccessibilityObjectWrapperBase accessibilityPlatformMathSubscriptKey]):
(-[WebAccessibilityObjectWrapperBase accessibilityPlatformMathSuperscriptKey]):
(-[WebAccessibilityObjectWrapperBase accessibilityMathPostscriptPairs]):
(-[WebAccessibilityObjectWrapperBase accessibilityMathPrescriptPairs]):
* accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
(-[WebAccessibilityObjectWrapper additionalAccessibilityAttributeNames]):
(-[WebAccessibilityObjectWrapper subrole]):
(-[WebAccessibilityObjectWrapper accessibilityAttributeValue:]):
(-[WebAccessibilityObjectWrapper accessibilityPlatformMathSubscriptKey]):
(-[WebAccessibilityObjectWrapper accessibilityPlatformMathSuperscriptKey]):
* mathml/mathtags.in:

LayoutTests:

* platform/mac/accessibility/mathml-multiscript-expected.txt: Added.
* platform/mac/accessibility/mathml-multiscript.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151123 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[Layout Tests] Update testharness scripts, add WebIDLParser.js and idlharness.js
zandobersek@gmail.com [Mon, 3 Jun 2013 19:19:57 +0000 (19:19 +0000)]
[Layout Tests] Update testharness scripts, add WebIDLParser.js and idlharness.js
https://bugs.webkit.org/show_bug.cgi?id=116983

Reviewed by Sam Weinig.

Update the testharness scripts that are used by the W3C tests that are located under LayoutTests/http/tests/w3c. Also add the
WebIDLParser.js and idlharness.js scripts that will be used by tests imported in the future. The files are being updated or impoted
off the 06cab30bcee2 changeset in the http://w3c-test.org/resources repository.

No test is affected by this update, though some tests will require rebaselining due to the baselines containing stack traces of which
line numbers will change due to the updated scripts.

* http/tests/w3c/README: List the currently-imported directories, noting from where they originate, at which changeset were
they imported and what files can or should be modified. Also note what script should be used to import/update the files, if available.
* http/tests/w3c/resources/WebIDLParser.js: Added. Will be used by future tests imported from the W3C test suites.
* http/tests/w3c/resources/idlharness.js: Ditto.
* http/tests/w3c/resources/testharness.css: Updated.
* http/tests/w3c/resources/testharness.js: Updated.
* http/tests/w3c/resources/testharnessreport.js: Updated. Removed the default completion callback and readded
the completion callback that was in use until now.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151122 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[Qt][WK1][CSS Regions] New test fast/regions/hover-on-child-parent-different-regions...
commit-queue@webkit.org [Mon, 3 Jun 2013 18:33:17 +0000 (18:33 +0000)]
[Qt][WK1][CSS Regions]  New test fast/regions/hover-on-child-parent-different-regions.html fails
https://bugs.webkit.org/show_bug.cgi?id=117035

Forced relayout to ensure the hover event has propagated before checking the results.

Patch by Radu Stavila <stavila@adobe.com> on 2013-06-03
Reviewed by Alexandru Chiculita.

* fast/regions/hover-on-child-parent-different-regions.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151121 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoAdd radar number to ChangeLog.
andersca@apple.com [Mon, 3 Jun 2013 18:31:54 +0000 (18:31 +0000)]
Add radar number to ChangeLog.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151120 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoSend mouse events to the banner for as long as the mouse button is down
andersca@apple.com [Mon, 3 Jun 2013 18:30:07 +0000 (18:30 +0000)]
Send mouse events to the banner for as long as the mouse button is down
https://bugs.webkit.org/show_bug.cgi?id=117161

Reviewed by Beth Dakin.

Keep track of whether the mouse was pressed in a banner and keep sending mouse events to it,
even if they are outside the banner's bounds for as long as the mouse is down.

* WebProcess/WebPage/PageBanner.h:
* WebProcess/WebPage/mac/PageBannerMac.mm:
(WebKit::PageBanner::PageBanner):
(WebKit::PageBanner::mouseEvent):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151119 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[BlackBerry] Selection handles are not cleared when displaying error pages.
mifenton@rim.com [Mon, 3 Jun 2013 17:57:23 +0000 (17:57 +0000)]
[BlackBerry] Selection handles are not cleared when displaying error pages.
https://bugs.webkit.org/show_bug.cgi?id=117158

Reviewed by Carlos Garcia Campos.

PR 342159.

Add explicit cancel of selection state when loading an error page
as it is not always cleared.

Internally Reviewed by Nima Ghanavatian.

* WebCoreSupport/FrameLoaderClientBlackBerry.cpp:
(WebCore::FrameLoaderClientBlackBerry::dispatchDidFailProvisionalLoad):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151118 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoASSERTION FAILED: node->parentNode(), Heap-use-after-free in WebCore::RenderBox:...
commit-queue@webkit.org [Mon, 3 Jun 2013 17:31:24 +0000 (17:31 +0000)]
ASSERTION FAILED: node->parentNode(), Heap-use-after-free in WebCore::RenderBox::exclusionShapeOutsideInfo
https://bugs.webkit.org/show_bug.cgi?id=117155

Patch by Bem Jones-Bey <bjonesbe@adobe.com> on 2013-06-03
Reviewed by David Hyatt.

Source/WebCore:

Ported from a Blink patch by kenrb@chromium.org:

Clear floats on RenderBlocks that haven't had layout

WebKit revision 150084 creates a condition where a RenderBlock
can have a populated floating object list even when it has not
previously had layout. Since clearing of floats is optimized
to not search through parts of the tree that have not had layout,
references to floating objects were not being cleared in some cases.

This patch causes markAllDescendantsWithFloatsForLayout() to proceed
in the above case, in order to clear floats for a RenderBlock that has
not had layout.

Test: fast/block/float/float-list-changed-before-layout-crash.html

* rendering/RenderBlock.cpp:
(WebCore::RenderBlock::markAllDescendantsWithFloatsForLayout):

LayoutTests:

Test for crash. Note that this will only crash when run under a memory
checking tool like ASAN.

* fast/block/float/float-list-changed-before-layout-crash-expected.txt: Added.
* fast/block/float/float-list-changed-before-layout-crash.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151117 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[CSS Shapes] inset-rectangle support for shape-outside
commit-queue@webkit.org [Mon, 3 Jun 2013 17:21:46 +0000 (17:21 +0000)]
[CSS Shapes] inset-rectangle support for shape-outside
https://bugs.webkit.org/show_bug.cgi?id=116641

Patch by Bem Jones-Bey <bjonesbe@adobe.com> on 2013-06-03
Reviewed by Alexandru Chiculita.

Source/WebCore:

Implement support for using inset-rectangles in a shape-outside.

Tests: fast/exclusions/shape-outside-floats/shape-outside-floats-simple-inset-rectangle.html
       fast/exclusions/shape-outside-floats/shape-outside-floats-simple-rounded-inset-rectangle.html

* rendering/ExclusionShape.cpp:
(WebCore::ExclusionShape::createExclusionShape): An inset-rectangle is
    simply a rectangle with the dimensions calculated differently. So
    just use ExclusionRectangle after converting the dimensions.
* rendering/ExclusionShapeOutsideInfo.cpp:
(WebCore::ExclusionShapeOutsideInfo::isEnabledFor): Enable inset
    rectangles for shape-outside.

LayoutTests:

Add tests for inset-rectangles.

* fast/exclusions/shape-outside-floats/shape-outside-floats-simple-inset-rectangle-expected.html: Added.
* fast/exclusions/shape-outside-floats/shape-outside-floats-simple-inset-rectangle.html: Added.
* fast/exclusions/shape-outside-floats/shape-outside-floats-simple-inset-rectangle-percentage-expected.html: Added.
* fast/exclusions/shape-outside-floats/shape-outside-floats-simple-inset-rectangle-percentage.html: Added.
* fast/exclusions/shape-outside-floats/shape-outside-floats-simple-rounded-inset-rectangle-expected.html: Added.
* fast/exclusions/shape-outside-floats/shape-outside-floats-simple-rounded-inset-rectangle.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151116 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[BlackBerry] ROI details are not always cleared when using back navigation.
mifenton@rim.com [Mon, 3 Jun 2013 17:17:38 +0000 (17:17 +0000)]
[BlackBerry] ROI details are not always cleared when using back navigation.
https://bugs.webkit.org/show_bug.cgi?id=117159

Reviewed by Carlos Garcia Campos.

PR 328557.

Secondary change to ensure when restoreViewState is triggered
ROI is reset.

Internally Reviewed by Nima Ghanavatian.

* WebKitSupport/InputHandler.cpp:
(BlackBerry::WebKit::InputHandler::restoreViewState):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151115 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[BlackBerry] Slider UX updates.
commit-queue@webkit.org [Mon, 3 Jun 2013 15:45:21 +0000 (15:45 +0000)]
[BlackBerry] Slider UX updates.
https://bugs.webkit.org/show_bug.cgi?id=117148

Patch by Tiancheng Jiang <tijiang@rim.com> on 2013-06-03
Reviewed by Rob Buis.

BlackBerry PR 342328
Internally Reviewed by Jeff Rogers.

Change image and reduce handle size when it is disabled.

* platform/blackberry/RenderThemeBlackBerry.cpp:
(WebCore::RenderThemeBlackBerry::paintSliderTrack):
(WebCore::RenderThemeBlackBerry::paintSliderThumb):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151114 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[Qt] [WK2] Unreviewed gardening. Unskipping passing tests and generate new test expec...
zarvai@inf.u-szeged.hu [Mon, 3 Jun 2013 14:45:23 +0000 (14:45 +0000)]
[Qt] [WK2] Unreviewed gardening. Unskipping passing tests and generate new test expectations where needed.

Patch by Gabor Abraham <abrhm@inf.u-szeged.hu> on 2013-06-03

* platform/qt-5.0-wk2/TestExpectations:
* platform/qt-5.0-wk2/compositing/geometry/video-opacity-overlay-expected.png: Added.
* platform/qt-5.0-wk2/compositing/geometry/video-opacity-overlay-expected.txt: Added.
* platform/qt-5.0-wk2/editing/spelling/grammar-markers-expected.png: Added.
* platform/qt-5.0-wk2/editing/spelling/grammar-markers-expected.txt: Added.
* platform/qt-5.0-wk2/editing/spelling/grammar-markers-hidpi-expected.png: Added.
* platform/qt-5.0-wk2/editing/spelling/grammar-markers-hidpi-expected.txt: Added.
* platform/qt-5.0-wk2/fast/ruby/rubyDOM-insert-rt-expected.png: Added.
* platform/qt-5.0-wk2/fast/ruby/rubyDOM-insert-rt-expected.txt: Added.
* platform/qt-5.0-wk2/fast/ruby/rubyDOM-insert-text1-expected.png: Added.
* platform/qt-5.0-wk2/fast/ruby/rubyDOM-insert-text1-expected.txt: Added.
* platform/qt-5.0-wk2/fast/ruby/rubyDOM-insert-text3-expected.png: Added.
* platform/qt-5.0-wk2/fast/ruby/rubyDOM-insert-text3-expected.txt: Added.
* platform/qt-5.0-wk2/fast/ruby/rubyDOM-remove-rt1-expected.png: Added.
* platform/qt-5.0-wk2/fast/ruby/rubyDOM-remove-rt1-expected.txt: Added.
* platform/qt-5.0-wk2/fast/ruby/rubyDOM-remove-text1-expected.png: Added.
* platform/qt-5.0-wk2/fast/ruby/rubyDOM-remove-text1-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151113 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[EFL] accessibility/placeholder.html is failing
commit-queue@webkit.org [Mon, 3 Jun 2013 14:26:26 +0000 (14:26 +0000)]
[EFL] accessibility/placeholder.html is failing
https://bugs.webkit.org/show_bug.cgi?id=112031

Unreviewed EFL gardening.

accessibility/placeholder.html passes after r151103.

Patch by Krzysztof Czech <k.czech@samsung.com> on 2013-06-03

* platform/efl-wk2/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151112 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoUnreviewed, rolling out r151109.
commit-queue@webkit.org [Mon, 3 Jun 2013 13:58:20 +0000 (13:58 +0000)]
Unreviewed, rolling out r151109.
http://trac.webkit.org/changeset/151109
https://bugs.webkit.org/show_bug.cgi?id=117145

"Plugins in WK2 aren't implemented yet." (Requested by
jturcotte on #webkit).

Source/WebCore:

* platform/FileSystem.h:
* platform/qt/FileSystemQt.cpp:
(WebCore::unloadModule):
* plugins/mac/PluginPackageMac.cpp:
(WebCore::PluginPackage::fetchInfo):
(WebCore::PluginPackage::load):
* plugins/mac/PluginViewMac.mm:

Source/WebKit/qt:

* WebCoreSupport/ChromeClientQt.cpp:
(WebCore::ChromeClientQt::windowResizerRect):

Tools:

* DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
* qmake/mkspecs/features/features.prf:
* qmake/mkspecs/features/unix/default_pre.prf:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151111 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoREGRESSION (r143070): Overflow:scroll content does not get clipped properly when...
commit-queue@webkit.org [Mon, 3 Jun 2013 13:37:38 +0000 (13:37 +0000)]
REGRESSION (r143070): Overflow:scroll content does not get clipped properly when the parent box has CSS3 filter on.
https://bugs.webkit.org/show_bug.cgi?id=116859

Patch by Zalan Bujtas <zalan@apple.com> on 2013-06-03
Reviewed by Simon Fraser.

Apply proper clipping for overflow:scroll, when parent dictates
full repaint to ensure correct CSS3 filter effect.

When a particular layer has 'pixel moving' CSS3 filter(drop shadow, blur), it sets clipToDirtyRect
to false to force no clipping on itself and on the child layers too.
No clipping makes overflow:scroll behave like overflow:display, because
overflow::scroll gets clipped at RenderLayer level (as opposed to overflow::hidden).
This patch makes sure that both full repaint and clipping are ensured.

Source/WebCore:

Tests: css3/filters/clipping-overflow-scroll-with-pixel-moving-effect-on-parent.html
       css3/filters/clipping-overflow-scroll-with-pixel-moving-effect-on.html

* rendering/RenderLayer.cpp:
(WebCore::RenderLayer::paintLayerContents):

LayoutTests:

* css3/filters/clipping-overflow-scroll-with-pixel-moving-effect-on-expected.html: Added.
* css3/filters/clipping-overflow-scroll-with-pixel-moving-effect-on-parent-expected.html: Added.
* css3/filters/clipping-overflow-scroll-with-pixel-moving-effect-on-parent.html: Added.
* css3/filters/clipping-overflow-scroll-with-pixel-moving-effect-on.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151110 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[Qt] Re-enable plugins on Mac.
jocelyn.turcotte@digia.com [Mon, 3 Jun 2013 13:13:26 +0000 (13:13 +0000)]
[Qt] Re-enable plugins on Mac.
https://bugs.webkit.org/show_bug.cgi?id=116622

Reviewed by Tor Arne Vestbø.

Source/WebCore:

* platform/FileSystem.h:
* platform/qt/FileSystemQt.cpp:
(WebCore::unloadModule):
  - Q_WS_* aren't defined since Qt5, use Q_OS_MACX instead.
* plugins/mac/PluginPackageMac.cpp:
(WebCore::PluginPackage::fetchInfo):
(WebCore::PluginPackage::load):
  - createCFString now does the adoption itself.
* plugins/mac/PluginViewMac.mm:
  - Add missing include.

Source/WebKit/qt:

* WebCoreSupport/ChromeClientQt.cpp:
(WebCore::ChromeClientQt::windowResizerRect):
  - Add missing parenthese.

Tools:

Plugins on Mac were only enabled if !embedded but the later was set if
QPA is enabled, which is the default since Qt 5.0.

Remove references to 'embedded' and fix various build issues,
PluginViewMac.mm and PluginPackageMac.cpp haven't been compiled since a few
months.

* DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
* qmake/mkspecs/features/features.prf:
* qmake/mkspecs/features/unix/default_pre.prf:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151109 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[EFL] Remove obsolete test case from TestExpectations
kangil.han@samsung.com [Mon, 3 Jun 2013 13:12:49 +0000 (13:12 +0000)]
[EFL] Remove obsolete test case from TestExpectations

Unreviewed EFL gardening.

xmlviewer/extensions-api.html case has been removed since r151080.

* platform/efl/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151108 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[Qt][Mac] Allow drawing plugins with QWindow by always using an intermediate bitmap.
jocelyn.turcotte@digia.com [Mon, 3 Jun 2013 13:12:48 +0000 (13:12 +0000)]
[Qt][Mac] Allow drawing plugins with QWindow by always using an intermediate bitmap.
https://bugs.webkit.org/show_bug.cgi?id=116620

Reviewed by Tor Arne Vestbø.

The isolation of QWidget along with QPA in Qt5 is abstracting away the native CGContext
of a QWidget and we would need to create a non-public API to be able to achieve it.
Instead of adding complexity to this rarely touched code, always draw into an
intermediate bitmap that we then paint into our QPainter afterward.

- Use CGBitmapContextCreate to allocate a buffer directly instead of creating a QPixmap
  and extract an CGContextRef out of it.
- Get rid of the "if (platformPluginWidget())" code path since we don't paint directly
  into the QWidget's backing store anymore.
- Always use m_contextRef to draw the plugin instead of alternating between direct and
  indirect drawing.

* plugins/PluginView.cpp:
(WebCore::PluginView::PluginView):
* plugins/PluginView.h:
* plugins/mac/PluginViewMac.mm:
(WebCore::createBitmapContext):
(WebCore::PluginView::platformStart):
(WebCore::PluginView::platformDestroy):
(WebCore::PluginView::setFocus):
(WebCore::PluginView::setNPWindowIfNeeded):
(WebCore::PluginView::updatePluginWidget):
(WebCore::PluginView::paint):
(WebCore::PluginView::invalidateRect):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151107 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[Qt][Mac] Remove Carbon and QuickDraw support for plugins.
jocelyn.turcotte@digia.com [Mon, 3 Jun 2013 13:12:31 +0000 (13:12 +0000)]
[Qt][Mac] Remove Carbon and QuickDraw support for plugins.
https://bugs.webkit.org/show_bug.cgi?id=116619

Reviewed by Tor Arne Vestbø.

They are only enabled for 32bit builds of QtWebKit and are rarely tested.
Other browsers support them through a 32bit host process.
PluginViewMac.mm is only used by Qt since the Wx port was removed.

- Only support the Cocoa event model
- Only support the CoreGraphics drawing model
- Remove all NP_NO_CARBON and NP_NO_QUICKDRAW blocks in PluginViewMac.mm

* plugins/PluginView.cpp:
(WebCore::PluginView::setValue):
(WebCore::PluginView::PluginView):
* plugins/PluginView.h:
* plugins/mac/PluginViewMac.mm:
(WebCore::PluginView::platformStart):
(WebCore::PluginView::platformDestroy):
(WebCore::PluginView::platformGetValueStatic):
(WebCore::PluginView::setFocus):
(WebCore::PluginView::setNPWindowIfNeeded):
(WebCore::PluginView::paint):
(WebCore::PluginView::handleWheelEvent):
(WebCore::PluginView::handleMouseEvent):
(WebCore::PluginView::handleKeyboardEvent):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151106 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[Qt] Unreviewed gardening.
zarvai@inf.u-szeged.hu [Mon, 3 Jun 2013 13:07:39 +0000 (13:07 +0000)]
[Qt] Unreviewed gardening.

* platform/qt-5.0-wk1/TestExpectations: Unskip after r151104.
* platform/qt-5.0-wk2/TestExpectations:
* platform/qt/TestExpectations: Skip crashing test.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151105 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[Qt][WK1][CSS Regions] New test fast/regions/hover-on-child-parent-different-regions...
commit-queue@webkit.org [Mon, 3 Jun 2013 12:52:49 +0000 (12:52 +0000)]
[Qt][WK1][CSS Regions]  New test fast/regions/hover-on-child-parent-different-regions.html fails
https://bugs.webkit.org/show_bug.cgi?id=117035

Used waitUntilDone() to ensure correct results.

Patch by Radu Stavila <stavila@adobe.com> on 2013-06-03
Reviewed by Antti Koivisto.

* fast/regions/hover-on-child-parent-different-regions-expected.txt:
* fast/regions/hover-on-child-parent-different-regions.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151104 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[GTK][WK2] accessibility/placeholder.html is failing
commit-queue@webkit.org [Mon, 3 Jun 2013 12:40:37 +0000 (12:40 +0000)]
[GTK][WK2] accessibility/placeholder.html is failing
https://bugs.webkit.org/show_bug.cgi?id=106344

Patch by Krzysztof Czech <k.czech@samsung.com> on 2013-06-03
Reviewed by Gyuyoung Kim.

Tools:

Adds support for retrieving attribute value.

* WebKitTestRunner/InjectedBundle/atk/AccessibilityUIElementAtk.cpp:
(WTR::coreAttributeToAtkAttribute):
(WTR::AccessibilityUIElement::stringAttributeValue):

LayoutTests:

Unskipping failing test. It's been fixed.

* platform/gtk-wk2/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151103 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoUnreviewed WinCE build fix after r150833.
paroga@webkit.org [Mon, 3 Jun 2013 12:38:03 +0000 (12:38 +0000)]
Unreviewed WinCE build fix after r150833.

* wtf/DateMath.cpp:
(WTF::calculateDSTOffset):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151102 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago'onended' event is not received in AudioBufferSourceNode and OscillatorNode.
commit-queue@webkit.org [Mon, 3 Jun 2013 11:47:46 +0000 (11:47 +0000)]
'onended' event is not received in AudioBufferSourceNode and OscillatorNode.
https://bugs.webkit.org/show_bug.cgi?id=117132.

Patch by Praveen R Jadhav <praveen.j@samsung.com> on 2013-06-03
Reviewed by Christophe Dumez.

Source/WebCore:

The listener argument is a PassRefPtr which becomes NULL after the call to
setAttributeEventListener() in setOnended(). This causes the m_hasEndedListener
boolean to be incorrectly initialized. This patch reverses the two statements so
that the m_hasEndedListener is updated before the call to setAttributeEventListener().

No new tests, already covered by existing tests.

* Modules/webaudio/AudioScheduledSourceNode.cpp:
(WebCore::AudioScheduledSourceNode::setOnended):

LayoutTests:

Unskip webaudio test cases that were failing due to the 'onended' event not being fired.

* platform/efl/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151101 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[EFL]<select> element's text is clipped when a height is specified along with CSS...
commit-queue@webkit.org [Mon, 3 Jun 2013 10:30:07 +0000 (10:30 +0000)]
[EFL]<select> element's text is clipped when a height is specified along with CSS line-height.
https://bugs.webkit.org/show_bug.cgi?id=114862

Patch by Rashmi Kulakarni <rashmi.vijay@samsung.com> on 2013-06-03
Reviewed by Gyuyoung Kim.

Source/WebCore:

Fix <select> element getting clipped when element height becomes less than the line-height.
Set minimum height for <select> element.

No new tests, no behavior change.

* platform/efl/RenderThemeEfl.cpp:
(WebCore::RenderThemeEfl::adjustMenuListButtonStyle):

LayoutTests:

Rebaselined the test case fast/forms/001.html.

* platform/efl/fast/forms/001-expected.png:
* platform/efl/fast/forms/001-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151100 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoSource/WebCore: Webkit crashes while loading content from Application Cache.
charles.wei@torchmobile.com.cn [Mon, 3 Jun 2013 08:58:34 +0000 (08:58 +0000)]
Source/WebCore: Webkit crashes while loading content from Application Cache.
https://bugs.webkit.org/show_bug.cgi?id=116814

Reviewed by Darin Adler.

With the refacor for 112722, which removes MainResourceLoader and moves the functionality to DocumentLoader,
the processing of appcache for main resource error (because of network error) was moved DocumentLoader::
mainReceivedError(), which was trigged by ResourceHandle::didFail->SubResourceLoader::didFail, which already
in a Finishing state after notifying the DocumentLoader of the load failure.

Before patch for 112722, the appcache for main resource error was in MainResourceLoader::didFail(),
which will start processing the potential fallback before notifying DocumentLoader and moving the the state
of the resource loader to finishing state.

So we should move maybeLoadFallbackForMainError out of DocumentLoader::mainReceivedError, and move it to
ApplicationCacheHost::maybeLoadFallbackForError, if the resource that has error is the main resource.

Test: http/tests/appcache/main-resource-fallback-for-network-error-crash.html

* loader/DocumentLoader.cpp:
(WebCore::DocumentLoader::mainReceivedError):
* loader/appcache/ApplicationCacheHost.cpp:
(WebCore::ApplicationCacheHost::maybeLoadFallbackForError):
* loader/appcache/ApplicationCacheHost.h:

LayoutTests: Webkit crashes while loading  content from Application Cache.
https://bugs.webkit.org/show_bug.cgi?id=116814

Reviewed by Darin Adler.

* http/tests/appcache/main-resource-fallback-for-network-error-crash-expected.txt: Added.
* http/tests/appcache/main-resource-fallback-for-network-error-crash.html: Added.
* http/tests/appcache/resources/main-resource-fallback-for-network-error-crash.manifest: Added.
* http/tests/appcache/resources/success-and-notify-done.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151099 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[EFL][WK2] Process launcher uses system() for wrapping the WebProcess when using...
commit-queue@webkit.org [Mon, 3 Jun 2013 08:34:34 +0000 (08:34 +0000)]
[EFL][WK2] Process launcher uses system() for wrapping the WebProcess when using WEB_PROCESS_CMD_PREFIX
https://bugs.webkit.org/show_bug.cgi?id=105156

Patch by Sergio Correia <sergio.correia@openbossa.org> on 2013-06-03
Reviewed by Christophe Dumez.

When using WEB_PROCESS_CMD_PREFIX - which allows us for instance to analyze
WebProcess under tools like valgrind or gdb -, the ProcessLauncher would
spawn the new process using system(), which would, among other things, keep
an extra UIProcess waiting and executing the shell.

This patch handles the normal case and the case where we have something to
prefix WebProcess (i.e., by using WEB_PROCESS_CMD_PREFIX in a debug build)
the same way, through a call to execvp().

To achieve this a function was introduced to create an array with the given
arguments to the full command to be executed, to be used by execvp(). We use
a Vector<OwnArrayPtr<char>>, so that we can take advantage of the destructor
of OwnArrayPtr to handle the memory deallocation when it goes out of scope.

* UIProcess/Launcher/efl/ProcessLauncherEfl.cpp:
(WebKit::createArgsArray):  This function creates the array to be used by
execvp(), out of the Strings given as arguments.
(WebKit::ProcessLauncher::launchProcess): Rework the logic to accomodate
both the cases with and without WEB_PROCESS_CMD_PREFIX. The execl() call
was replaced with an execvp() call, since now we should deal with having
a variable number of arguments (WEB_PROCESS_CMD_PREFIX) as well.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151098 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoUnreviewed EFL gardening.
ch.dumez@sisa.samsung.com [Mon, 3 Jun 2013 08:02:00 +0000 (08:02 +0000)]
Unreviewed EFL gardening.

Rebaseline 2 printing test cases on EFL port.

* platform/efl/printing/compositing-layer-printing-expected.png:
* platform/efl/printing/compositing-layer-printing-expected.txt:
* platform/efl/printing/page-rule-in-media-query-expected.png:
* platform/efl/printing/page-rule-in-media-query-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151097 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoUnreviewed EFL gardening.
ch.dumez@sisa.samsung.com [Mon, 3 Jun 2013 07:48:55 +0000 (07:48 +0000)]
Unreviewed EFL gardening.

Skip 2 webaudio tests that are failing due to the onended event not getting
fired.

* platform/efl/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151096 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoUnreviewed EFL gardening.
ch.dumez@sisa.samsung.com [Mon, 3 Jun 2013 07:44:38 +0000 (07:44 +0000)]
Unreviewed EFL gardening.

Rebaseline several test cases on EFL.

* platform/efl-wk2/compositing/overflow/content-gains-scrollbars-expected.txt: Added.
* platform/efl-wk2/compositing/overflow/content-loses-scrollbars-expected.txt: Added.
* platform/efl-wk2/fast/dynamic/002-expected.txt:
* platform/efl/fast/block/float/float-not-removed-from-next-sibling2-expected.png:
* platform/efl/fast/block/float/float-not-removed-from-next-sibling2-expected.txt:
* platform/efl/fast/dynamic/002-expected.png:
* platform/efl/fast/table/007-expected.png:
* platform/efl/fast/table/007-expected.txt:
* platform/efl/fast/table/colspanMinWidth-expected.png:
* platform/efl/fast/table/colspanMinWidth-expected.txt:
* platform/efl/fast/table/spanOverlapRepaint-expected.png:
* platform/efl/fast/table/spanOverlapRepaint-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151095 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoUnreviewed EFL gardening.
ch.dumez@sisa.samsung.com [Mon, 3 Jun 2013 07:14:37 +0000 (07:14 +0000)]
Unreviewed EFL gardening.

Rebaseline several mozilla tables layout tests for EFL port.

* platform/efl/tables/mozilla/bugs/bug13169-expected.png:
* platform/efl/tables/mozilla/bugs/bug13169-expected.txt:
* platform/efl/tables/mozilla/bugs/bug133756-1-expected.png:
* platform/efl/tables/mozilla/bugs/bug133756-1-expected.txt:
* platform/efl/tables/mozilla/bugs/bug133756-2-expected.png:
* platform/efl/tables/mozilla/bugs/bug133756-2-expected.txt:
* platform/efl/tables/mozilla/bugs/bug17138-expected.png:
* platform/efl/tables/mozilla/bugs/bug17138-expected.txt:
* platform/efl/tables/mozilla/bugs/bug17548-expected.png:
* platform/efl/tables/mozilla/bugs/bug17548-expected.txt:
* platform/efl/tables/mozilla/bugs/bug220536-expected.png:
* platform/efl/tables/mozilla/bugs/bug220536-expected.txt:
* platform/efl/tables/mozilla/bugs/bug6304-expected.png:
* platform/efl/tables/mozilla/bugs/bug6304-expected.txt:
* platform/efl/tables/mozilla/bugs/bug73321-expected.png:
* platform/efl/tables/mozilla/bugs/bug73321-expected.txt:
* platform/efl/tables/mozilla/bugs/bug7714-expected.txt:
* platform/efl/tables/mozilla/bugs/bug8858-expected.png:
* platform/efl/tables/mozilla/bugs/bug8858-expected.txt:
* platform/efl/tables/mozilla/core/bloomberg-expected.png:
* platform/efl/tables/mozilla/core/bloomberg-expected.txt:
* platform/efl/tables/mozilla/core/cell_heights-expected.png:
* platform/efl/tables/mozilla/core/cell_heights-expected.txt:
* platform/efl/tables/mozilla/other/test6-expected.png:
* platform/efl/tables/mozilla/other/test6-expected.txt:
* platform/efl/tables/mozilla_expected_failures/bugs/bug1010-expected.png:
* platform/efl/tables/mozilla_expected_failures/bugs/bug1010-expected.txt:
* platform/efl/tables/mozilla_expected_failures/bugs/bug131020-3-expected.png:
* platform/efl/tables/mozilla_expected_failures/bugs/bug131020-3-expected.txt:
* platform/efl/tables/mozilla_expected_failures/bugs/bug23847-expected.png:
* platform/efl/tables/mozilla_expected_failures/bugs/bug23847-expected.txt:
* platform/efl/tables/mozilla_expected_failures/bugs/bug58402-2-expected.png:
* platform/efl/tables/mozilla_expected_failures/bugs/bug58402-2-expected.txt:
* platform/efl/tables/mozilla_expected_failures/bugs/bug65372-expected.png:
* platform/efl/tables/mozilla_expected_failures/bugs/bug65372-expected.txt:
* platform/efl/tables/mozilla_expected_failures/bugs/bug6933-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151094 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[WTF] Add OwnArrayPtr vectortraits template
gyuyoung.kim@samsung.com [Mon, 3 Jun 2013 07:09:56 +0000 (07:09 +0000)]
[WTF] Add OwnArrayPtr vectortraits template
https://bugs.webkit.org/show_bug.cgi?id=117131

Reviewed by Darin Adler.

To support Vector<OwnArrayPtr<P>>, which will be used by EFL port.

* wtf/VectorTraits.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151093 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoUnreviewed EFL Gardening.
gyuyoung.kim@samsung.com [Mon, 3 Jun 2013 05:32:22 +0000 (05:32 +0000)]
Unreviewed EFL Gardening.

* platform/efl-wk2/TestExpectations: Unskipping some of webgl passing tests.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151092 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[Qt] Move Qt port off legacy clipboard
commit-queue@webkit.org [Mon, 3 Jun 2013 04:53:33 +0000 (04:53 +0000)]
[Qt] Move Qt port off legacy clipboard
https://bugs.webkit.org/show_bug.cgi?id=116416

Patch by Arunprasad Rajkumar <arurajku@cisco.com> on 2013-06-02
Reviewed by Darin Adler.

Source/WebCore:

No change in behavior, no new tests added.

* Target.pri: Removed ClipboardQt.h and EditorQt.cpp from compilation.
* dom/Clipboard.h:
* editing/qt/EditorQt.cpp: Removed.
* page/qt/EventHandlerQt.cpp: Use Clipboard::createForDragAndDrop().
(WebCore::EventHandler::createDraggingClipboard):
* platform/Pasteboard.h:
(WebCore::Pasteboard::clipboardData): Expose writable QMimeData* to outside, so that it is accessible
from DragClientQt.cpp and WebDragClientQt.cpp.
(WebCore::Pasteboard::invalidateWritableData):
(WebCore::Pasteboard::isForDragAndDrop):
(WebCore::Pasteboard::isForCopyAndPaste):
* platform/qt/ClipboardQt.cpp:
(WebCore::Clipboard::createDragImage):
(WebCore::Clipboard::declareAndWriteDragImage):
* platform/qt/ClipboardQt.h: Removed.
* platform/qt/PasteboardQt.cpp: Moved required methods from ClipboardQt.cpp.
(WebCore::isTextMimeType):
(WebCore::isHtmlMimeType):
(WebCore::Pasteboard::create):
(WebCore::Pasteboard::createForCopyAndPaste):
(WebCore::Pasteboard::createPrivate):
(WebCore::Pasteboard::createForDragAndDrop):
(WebCore::Pasteboard::Pasteboard):
(WebCore::Pasteboard::~Pasteboard):
(WebCore::Pasteboard::generalPasteboard):
(WebCore::Pasteboard::isSelectionMode):
(WebCore::Pasteboard::setSelectionMode):
(WebCore::Pasteboard::readData):
(WebCore::Pasteboard::hasData):
(WebCore::Pasteboard::clear):
(WebCore::Pasteboard::readString):
(WebCore::Pasteboard::writeString):
(WebCore::Pasteboard::types):
(WebCore::Pasteboard::readFilenames):
(WebCore::Pasteboard::setDragImage):
(WebCore::Pasteboard::writePasteboard):

Source/WebKit/qt:

* WebCoreSupport/DragClientQt.cpp:
(WebCore::DragClientQt::startDrag):

Source/WebKit2:

* WebProcess/WebCoreSupport/qt/WebDragClientQt.cpp:
(WebKit::WebDragClient::startDrag):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151091 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoGoing "back" to a cached page from a page with a main resource error breaks scrolling...
beidson@apple.com [Mon, 3 Jun 2013 01:43:24 +0000 (01:43 +0000)]
Going "back" to a cached page from a page with a main resource error breaks scrolling, amongst other issues.
<rdar://problem/13751844> and https://bugs.webkit.org/show_bug.cgi?id=117112

Reviewed by Alexey Proskuryakov.

Source/WebCore:

If a main resource load ends in error, the Document's parser is never cleared out.

If you then return to a CachedPage, we run the Document->clearParser() step for the
old page and that incorrectly dispatches didFinishLoad for the previous page load
in the middle of the load for the cached page.

The parser should never be needed after a load completes (even if it fails) and
holding on to the parser after the page load failed but before a new navigation is
actually using unnecessary resources.

So we should just clear the parser right when the main resource fails.

Test: http/tests/loading/unfinished-load-back-to-cached-page-callbacks.html

* loader/DocumentLoader.cpp:
(WebCore::DocumentLoader::stopLoading): After canceling the main resource load, also clear the parser.

LayoutTests:

* http/tests/loading/resources/resource-that-goes-back-while-still-loading.php: Added.
* http/tests/loading/unfinished-load-back-to-cached-page-callbacks-expected.txt: Added.
* http/tests/loading/unfinished-load-back-to-cached-page-callbacks.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151088 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago The empty directory did not get deleted by commit queue, deleting manually.
ap@apple.com [Sun, 2 Jun 2013 20:37:05 +0000 (20:37 +0000)]
    The empty directory did not get deleted by commit queue, deleting manually.

        * xmlviewer: Removed.
        * xmlviewer/resources: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151084 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[CMake] Unreviewed buildfix after r148896 and r150940.
ossy@webkit.org [Sun, 2 Jun 2013 19:08:20 +0000 (19:08 +0000)]
[CMake] Unreviewed buildfix after r148896 and r150940.

* CMakeLists.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151083 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[WK2][CoordinatedGraphics]: Use a properly initialized WebPage when creating a PageClient
commit-queue@webkit.org [Sun, 2 Jun 2013 08:33:09 +0000 (08:33 +0000)]
[WK2][CoordinatedGraphics]: Use a properly initialized WebPage when creating a PageClient
https://bugs.webkit.org/show_bug.cgi?id=117116

Patch by Sergio Correia <sergio.correia@openbossa.org> on 2013-06-02
Reviewed by Tim Horton.

Valgrind reported a "conditional jump or move depends on uninitialised value"
at WebPageProxy constructor, when checking the m_isVisible variable inside
the ENABLE(PAGE_VISIBILITY_API) guard.

The issue was tracked down to the WebView constructor, which was creating a
WebKit::PageClient* making use of `this' before having initialized all of its
attributes.

The solution was to move the initialization of the `page' member to inside the
constructor itself - out of the initialization list -, where we can guarantee
all the relevant stuff has already been properly initialized.

Here's the relevant valgrind trace for reference:

Conditional jump or move depends on uninitialised value(s)
    at 0x65A35A4: WebKit::WebPageProxy::WebPageProxy(WebKit::PageClient*,
                  WTF::PassRefPtr<WebKit::WebProcessProxy>,
                  WebKit::WebPageGroup*, unsigned long) (WebPageProxy.cpp:322)
    by 0x65A2BA2: WebKit::WebPageProxy::create(WebKit::PageClient*,
                  WTF::PassRefPtr<WebKit::WebProcessProxy>,
                  WebKit::WebPageGroup*, unsigned long) (WebPageProxy.cpp:233)
    by 0x65E94BB: WebKit::WebProcessProxy::createWebPage(WebKit::PageClient*,
                  WebKit::WebContext*, WebKit::WebPageGroup*)
                  (WebProcessProxy.cpp:172)
    by 0x6570957: WebKit::WebContext::createWebPage(WebKit::PageClient*,
                  WebKit::WebPageGroup*, WebKit::WebPageProxy*)
                  (WebContext.cpp:735)
    by 0x67673E3: WebKit::WebView::WebView(WebKit::WebContext*,
                  WebKit::WebPageGroup*) (WebView.cpp:52)
    by 0x6775F18: WebKit::WebViewEfl::WebViewEfl(WebKit::WebContext*,
                  WebKit::WebPageGroup*) (WebViewEfl.cpp:54)
    by 0x6775EB4: WebKit::WebView::create(WebKit::WebContext*,
                  WebKit::WebPageGroup*) (WebViewEfl.cpp:49)
    by 0x673E13D: WKViewCreate (WKView.cpp:33)
    by 0x6763ECE: EWKViewCreate (ewk_view.cpp:92)

* UIProcess/CoordinatedGraphics/WebView.cpp:
(WebKit::WebView::WebView):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151081 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoUnstyled XML viewer crashes when XML contains an element with id="tree"
ap@apple.com [Sun, 2 Jun 2013 08:14:48 +0000 (08:14 +0000)]
Unstyled XML viewer crashes when XML contains an element with id="tree"
https://bugs.webkit.org/show_bug.cgi?id=117117
<rdar://problem/13727267>

Reviewed by Darin Adler.

Source/WebCore:

Test: http/tests/xmlviewer/dumpAsText/existing-tree-element-crash.xml

* xml/XMLViewer.css: Removed the rule to hide source document, as we no longer
keep it in the tree.

* xml/XMLViewer.js:
(prepareWebKitXMLViewer): Make sourceXML a detached tree, as otherwise it can
have ID conflicts with elements we add.
(sourceXMLLoaded): Removed some commented out code. Removed code that was checking
whether an extension took over XML display, as it makes no sense to duplicate this
functionality in an extension.

LayoutTests:

* http/tests/xmlviewer/dumpAsText/existing-tree-element-crash-expected.txt: Added.
* http/tests/xmlviewer/dumpAsText/existing-tree-element-crash.xml: Added.

* xmlviewer: Removed.
* xmlviewer/extensions-api-expected.txt: Removed.
* xmlviewer/extensions-api.html: Removed.
* xmlviewer/resources: Removed.
* xmlviewer/resources/sample.xml: Removed.
The "API" no longer exists.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151080 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoWebKit ignores aria-hidden when traversing nodes for name calculation
cfleizach@apple.com [Sun, 2 Jun 2013 04:42:57 +0000 (04:42 +0000)]
WebKit ignores aria-hidden when traversing nodes for name calculation
https://bugs.webkit.org/show_bug.cgi?id=117101

Reviewed by Tim Horton.

Source/WebCore:

When calculating the inner text, aria-hidden should be considered.

Test: accessibility/text-role-with-aria-hidden-inside.html

* accessibility/AccessibilityNodeObject.cpp:
(WebCore::shouldUseAccessiblityObjectInnerText):

LayoutTests:

* accessibility/text-role-with-aria-hidden-inside-expected.txt: Added.
* accessibility/text-role-with-aria-hidden-inside.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151079 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agowindow.speechSynthesis needs to be cheap
darin@apple.com [Sat, 1 Jun 2013 23:26:02 +0000 (23:26 +0000)]
window.speechSynthesis needs to be cheap
https://bugs.webkit.org/show_bug.cgi?id=117111
rdar://problem/14042030

Reviewed by Dean Jackson.

Add the traditional laziness to all of the speech synthesis code, wherever
it was omitted.

* Modules/speech/SpeechSynthesis.cpp:
(WebCore::SpeechSynthesis::create): Tweaked style (took out unneeded parentheses).
(WebCore::SpeechSynthesis::SpeechSynthesis): Don't create m_platformSpeechSynthesizer.
(WebCore::SpeechSynthesis::setPlatformSynthesizer): Clear state when changing the
platform speech synthesizer. Since this code is only used to set up a mock in the
test runner, the fact that it was wrong before was harmless, but still not good.
(WebCore::SpeechSynthesis::getVoices): Create the platform speech synthesizer here
so we can get the voice list from it.
(WebCore::SpeechSynthesis::startSpeakingImmediately): Create the platform speech
synthesizer here.
(WebCore::SpeechSynthesis::cancel): Check the platform speech synthesizer for
null and do nothing if it's not present.
(WebCore::SpeechSynthesis::pause): Ditto.
(WebCore::SpeechSynthesis::resume): Ditto.

* platform/PlatformSpeechSynthesizer.cpp:
(WebCore::PlatformSpeechSynthesizer::create): Don't call initializeVoiceList just
to create a synthesizer.
(WebCore::PlatformSpeechSynthesizer::voiceList): Do call initializeVoiceList once
when asked for a voice list.

* platform/PlatformSpeechSynthesizer.h: The voiceList function is no longer inlined.
The unused setVoiceList function has been removed. The initializeVoiceList is now
private rather than protected. Added a new m_voiceListIsInitialized boolean.

* platform/mac/PlatformSpeechSynthesizerMac.mm:
(WebCore::PlatformSpeechSynthesizer::PlatformSpeechSynthesizer): Initialize
m_voiceListIsInitialized to false.

* platform/mock/PlatformSpeechSynthesizerMock.cpp:
(WebCore::PlatformSpeechSynthesizerMock::create): Don't call initializeVoiceList just
to create a synthesizer.
(WebCore::PlatformSpeechSynthesizerMock::~PlatformSpeechSynthesizerMock):
Removed unneeded call to m_speakingFinishedTimer.stop() since timers automatically
stop when you destroy them.
(WebCore::PlatformSpeechSynthesizerMock::initializeVoiceList): Removed unneeded
call to m_voiceList.clear(), since the caller only calls this once when the
voice list is already clear.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151076 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoAdd a preferred email address and "expertise" for me.
timothy_horton@apple.com [Sat, 1 Jun 2013 23:09:44 +0000 (23:09 +0000)]
Add a preferred email address and "expertise" for me.

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

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151075 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoFix typo in r151071.
akling@apple.com [Sat, 1 Jun 2013 18:02:36 +0000 (18:02 +0000)]
Fix typo in r151071.

* dom/Element.cpp:
(WebCore::Element::attributeChanged):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151072 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoMove Node::hasName() to Element.
akling@apple.com [Sat, 1 Jun 2013 10:20:25 +0000 (10:20 +0000)]
Move Node::hasName() to Element.
<http://webkit.org/b/117107>

Reviewed by Antti Koivisto.

A Node can't have attributes, and thus can't have a name.
Use a bit on ElementData instead of (half) a Node flag to track whether we have a name.

* dom/Element.cpp:
(WebCore::Element::attributeChanged):
(WebCore::ElementData::ElementData):
* dom/Element.h:
(WebCore::ElementData::hasName):
(WebCore::Element::hasName):
* dom/Node.h:
(WebCore::Node::isEditingText):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151071 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoRemove unnused WebDOMNodeOperationsInternal category.
akling@apple.com [Sat, 1 Jun 2013 09:48:33 +0000 (09:48 +0000)]
Remove unnused WebDOMNodeOperationsInternal category.
<http://webkit.org/b/117105>

Reviewed by Ryosuke Niwa.

_subresourceURLs had no clients.

* DOM/WebDOMOperations.mm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151070 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoMove Node::hasID() and hasClass() to Element.
akling@apple.com [Sat, 1 Jun 2013 09:39:59 +0000 (09:39 +0000)]
Move Node::hasID() and hasClass() to Element.
<http://webkit.org/b/117104>

Reviewed by Ryosuke Niwa.

Source/WebCore:

A Node can't have attributes, and thus can't have an ID or a class.
Move these to Element instead, and fix up a few call sites that called through on a Node*.

* dom/Element.h:
* dom/LiveNodeList.cpp:
(WebCore::LiveNodeList::namedItem):
* dom/Node.h:
* rendering/RenderLayer.cpp:
(WebCore::RenderLayer::name):
* rendering/RenderTreeAsText.cpp:
(WebCore::RenderTreeAsText::writeRenderObject):
(WebCore::writeRenderRegionList):

Source/WebKit/mac:

* WebView/WebRenderLayer.mm:
(+[WebRenderLayer nameForLayer:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151069 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoUnreviewed, EFL gardening. Remove passed accessibility test from EFL WK1.
gyuyoung.kim@samsung.com [Sat, 1 Jun 2013 06:00:19 +0000 (06:00 +0000)]
Unreviewed, EFL gardening. Remove passed accessibility test from EFL WK1.

* platform/efl-wk1/TestExpectations:
* platform/efl-wk2/TestExpectations:
* platform/efl/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151068 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[Curl] Unable to download files.
commit-queue@webkit.org [Sat, 1 Jun 2013 05:41:42 +0000 (05:41 +0000)]
[Curl] Unable to download files.
https://bugs.webkit.org/show_bug.cgi?id=116150

Patch by peavo@outlook.com <peavo@outlook.com> on 2013-05-31
Reviewed by Brent Fulgham.

Implemented basic download functionality for Curl.

Source/WebCore:

* WebCore.vcproj/WebCore.vcproj: Added file.
* WebCore.vcxproj/WebCore.vcxproj: Added file.
* WebCore.vcxproj/WebCore.vcxproj.filters: Added file.
* platform/network/curl/CurlDownload.h: Added.
* platform/network/curl/CurlDownload.cpp: Added.

Source/WebKit/win:

* WebDownload.h:
* WebDownloadCurl.cpp:
(WebDownload::init):
(WebDownload::start):
(WebDownload::cancel):
(WebDownload::deletesFileUponFailure):
(WebDownload::setDeletesFileUponFailure):
(WebDownload::setDestination):
(WebDownload::useCredential):
(WebDownload::didReceiveResponse):
(WebDownload::didReceiveDataOfLength):
(WebDownload::didFinish):
(WebDownload::didFail):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151067 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoUnreviewed. Make windows test bots search in the correct directory for build output.
roger_fong@apple.com [Sat, 1 Jun 2013 04:21:09 +0000 (04:21 +0000)]
Unreviewed. Make windows test bots search in the correct directory for build output.

* BuildSlaveSupport/build.webkit.org-config/master.cfg:
(RunWebKitTests.start):
* Scripts/webkitdirs.pm:
(determineConfigurationProductDir):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151066 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoUnreviewed. Windows build fix...
roger_fong@apple.com [Sat, 1 Jun 2013 03:54:25 +0000 (03:54 +0000)]
Unreviewed. Windows build fix...

* BuildSlaveSupport/built-product-archive:
(createZip):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151065 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoUnreviewed. Windows build fix. One more time!
roger_fong@apple.com [Sat, 1 Jun 2013 03:34:35 +0000 (03:34 +0000)]
Unreviewed. Windows build fix. One more time!

* BuildSlaveSupport/built-product-archive:
(archiveBuiltProduct):
(extractBuiltProduct):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151064 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoUnreviewed. Windows build fix.
roger_fong@apple.com [Sat, 1 Jun 2013 03:04:24 +0000 (03:04 +0000)]
Unreviewed. Windows build fix.

* Scripts/build-webkit:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151063 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[EFL][WK2] Add a null checking condition to in hide() of EwkContextMenu
gyuyoung.kim@samsung.com [Sat, 1 Jun 2013 02:59:08 +0000 (02:59 +0000)]
[EFL][WK2] Add a null checking condition to in hide() of EwkContextMenu
https://bugs.webkit.org/show_bug.cgi?id=117014

Reviewed by Christophe Dumez.

m_viewImpl is able to be null. So, we need to protect this function against such case.

* UIProcess/API/efl/ewk_context_menu.cpp:
(EwkContextMenu::hide):
(ewk_context_menu_hide):
* UIProcess/API/efl/ewk_context_menu_private.h:
(EwkContextMenu):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151062 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoFix bug in active page tracking that lead to plugins hanging during destruction
oliver@apple.com [Sat, 1 Jun 2013 01:39:17 +0000 (01:39 +0000)]
Fix bug in active page tracking that lead to plugins hanging during destruction
https://bugs.webkit.org/show_bug.cgi?id=117099

Reviewed by Anders Carlsson.

NPRuntimeObjectMap no longer keeps its own reference to the
throttler, and no longer reports activity during destruction.

* WebProcess/Plugins/Netscape/NPRuntimeObjectMap.cpp:
(WebKit::NPRuntimeObjectMap::NPRuntimeObjectMap):
(WebKit::NPRuntimeObjectMap::evaluate):
* WebProcess/Plugins/Netscape/NPRuntimeObjectMap.h:
* WebProcess/Plugins/PluginView.cpp:
(WebKit::PluginView::PluginView):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151060 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoUnreviewed. Windows build fix.
roger_fong@apple.com [Sat, 1 Jun 2013 01:07:15 +0000 (01:07 +0000)]
Unreviewed. Windows build fix.

* Scripts/build-webkit:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151059 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[wk2] All intrinsic content size changes should go through the 0-delay timer
timothy_horton@apple.com [Sat, 1 Jun 2013 00:01:54 +0000 (00:01 +0000)]
[wk2] All intrinsic content size changes should go through the 0-delay timer
https://bugs.webkit.org/show_bug.cgi?id=117058
<rdar://problem/13935042>

Reviewed by Anders Carlsson.

In r146913, I added a 0-delay timer so that we'd coalesce intrinsic
content size changes. However, didUpdateGeometry was still immediately
invalidating the intrinsic content size without coalescing.

Move that update to the WebProcess to share the timer.

* UIProcess/DrawingAreaProxy.h:
(WebKit::DrawingAreaProxy::didUpdateGeometry):
* UIProcess/DrawingAreaProxy.messages.in:
* UIProcess/mac/TiledCoreAnimationDrawingAreaProxy.h:
* UIProcess/mac/TiledCoreAnimationDrawingAreaProxy.mm:
(WebKit::TiledCoreAnimationDrawingAreaProxy::didUpdateGeometry):
Remove the newIntrinsicContentSize argument, and don't invalidate intrinsic
content size inside didUpdateGeometry.

* WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.mm:
(WebKit::TiledCoreAnimationDrawingArea::updateGeometry):
Start the intrinsic content size update timer when we finish updateGeometry.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151057 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoMake build-webkit output VCExpress 2010 build logs properly.
roger_fong@apple.com [Fri, 31 May 2013 23:47:23 +0000 (23:47 +0000)]
Make build-webkit output VCExpress 2010 build logs properly.
https://bugs.webkit.org/show_bug.cgi?id=117096

Reviewed by Brent Fulgham.

Make it so that output from VCExpress is always printed to console.
Get build log results from AssembleBuildLogs project output.
We don't need those old VCExpress scripts anymore either.

* Scripts/build-webkit:
* Scripts/print-msvc-project-dependencies: Removed.
* Scripts/print-vse-failure-logs: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151056 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago Sandbox violations when trying to access fonts in non-standard locations
ap@apple.com [Fri, 31 May 2013 23:40:40 +0000 (23:40 +0000)]
    Sandbox violations when trying to access fonts in non-standard locations
        https://bugs.webkit.org/show_bug.cgi?id=117097
        <rdar://problem/9264746>

        Reviewed by Anders Carlsson.

        The issue was that lack of a main xib was resulting in font machinery being
        initialized too early, before we entered sandbox.

        * Resources/WebContentProcess.xib: Added.
        * WebKit2.xcodeproj/project.pbxproj:
        * WebProcess/EntryPoint/mac/LegacyProcess/Info.plist:
        * WebProcess/EntryPoint/mac/XPCService/WebContentService.Development/Info.plist:
        * WebProcess/EntryPoint/mac/XPCService/WebContentService/Info.plist:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151055 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoNeed the ability to get the information for a plug-in with a particular process id...
jberlin@webkit.org [Fri, 31 May 2013 23:32:53 +0000 (23:32 +0000)]
Need the ability to get the information for a plug-in with a particular process id that may
be running on a page
https://bugs.webkit.org/show_bug.cgi?id=117090

Rubber-stamped by Anders Carlsson.

Follow-up to r151043 incorprating some feedback I got from Darin after I had already
committed the patch.

* UIProcess/API/C/mac/WKPagePrivateMac.cpp:
(callGetPlugInInformationBlockAndDispose):
Renamed to be less wordy.
(WKPageGetPlugInInformation):
Ditto.
* UIProcess/API/C/mac/WKPagePrivateMac.h:
Ditto.

* UIProcess/Plugins/PluginProcessManager.h:
* UIProcess/Plugins/mac/PluginProcessManagerMac.mm:
(WebKit::PluginProcessManager::findPlugInProcessByID):
Renamed to not use "Cocoa method style" in non-Cocoa functions.
(WebKit::PluginProcessManager::findPlugInProcessByToken):
Ditto.

* UIProcess/WebPageProxy.h:
* UIProcess/mac/WebPageProxyMac.mm:
(WebKit::WebPageProxy::getPlugInInformation):
Reduce ref count churn and fix a variable name.
Updated for the change to findPlugInProcessByID.
(WebKit::WebPageProxy::containsPlugInCallback):
Updated for the change to findPlugInProcessByToken.

* WebProcess/Plugins/PDF/SimplePDFPlugin.h:
Move the override of a virtual function to the private section.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151054 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago <rdar://problem/14036464> Sandbox violation when using Web SpeechSynthesis
ap@apple.com [Fri, 31 May 2013 22:32:27 +0000 (22:32 +0000)]
    <rdar://problem/14036464> Sandbox violation when using Web SpeechSynthesis

        Reviewed by Anders Carlsson

        * WebProcess/com.apple.WebProcess.sb.in: Allow com.apple.speech.synthesis.console.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151049 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[Windows] Use WinLauncher to display LayoutTest results
bfulgham@apple.com [Fri, 31 May 2013 22:14:28 +0000 (22:14 +0000)]
[Windows] Use WinLauncher to display LayoutTest results
https://bugs.webkit.org/show_bug.cgi?id=117085

Reviewed by Oliver Hunt.

* Scripts/webkitdirs.pm:
(runSafari): Use WinLauncher to display content
* WinLauncher/WinLauncher.cpp:
(dllLauncherEntryPoint): Check for URL argument.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151047 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoIncorrect assertion in DFG::Graph::uncheckedActivationRegisterFor()
fpizlo@apple.com [Fri, 31 May 2013 22:08:16 +0000 (22:08 +0000)]
Incorrect assertion in DFG::Graph::uncheckedActivationRegisterFor()
<rdar://problem/13989324>

Rubber stamped by Mark Hahnenberg.

This has a bogus assertion that checks that the passed CodeOrigin doesn't have
an inline call frame. This was well intentioned in the sense that it is true
that inlined call frames wouldn't have an activation register. But that doesn't
mean that people won't ask. Removing the assertion fixes a debug-only crash and
has no impact on production code. This change adds a comment to that effect.

* dfg/DFGGraph.h:
(JSC::DFG::Graph::uncheckedActivationRegisterFor):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151045 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoUnreviewed. Modify AssembleBuildLogs project to WebKit solution in preparation for...
roger_fong@apple.com [Fri, 31 May 2013 22:01:56 +0000 (22:01 +0000)]
Unreviewed. Modify AssembleBuildLogs project to WebKit solution in preparation for moving buildbots to VCExpress 2010.
Add it to the WebKit solution as well.

* win/AssembleBuildLogs/AssembleLogs.cmd:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151044 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoNeed the ability to get the information for a plug-in with a particular process id...
jberlin@webkit.org [Fri, 31 May 2013 21:59:58 +0000 (21:59 +0000)]
Need the ability to get the information for a plug-in with a particular process id that may
be running on a page
https://bugs.webkit.org/show_bug.cgi?id=117090

Reviewed by Anders Carlsson.

* UIProcess/API/C/mac/WKPagePrivateMac.cpp:
(callGetInformationForPlugInWithProcessIDBlockAndDispose):
Get the block from the context, invoke it with the result, and dispose of it.
(WKPageGetInformationForPlugInWithProcessID):
Ask the WebPageProxy and provide it with a dictionary callback that uses the
above function to invoke the passed in block.
* UIProcess/API/C/mac/WKPagePrivateMac.h:
Add WKPageGetInformationForPlugInWithProcessID and a typedef for the callback block.

* UIProcess/Plugins/PluginProcessManager.h:
Add ways to get the PluginProcessProxy associated with a process id or process token.
* UIProcess/Plugins/mac/PluginProcessManagerMac.mm:
(WebKit::PluginProcessManager::plugInProcessWithProcessID):
Return the first PluginProcessProxy with the passed in process ID.
(WebKit::PluginProcessManager::plugInProcessWithToken):
Return the first PluginProcessProxy with the passed in process token.

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::close):
Invalidate any outstanding requests for plug-in information.
* UIProcess/WebPageProxy.h:
Keep track of the callbacks for the requests for plug-in information.
* UIProcess/WebPageProxy.messages.in:
* UIProcess/mac/WebPageProxyMac.mm:
(WebKit::WebPageProxy::getInformationForPlugInWithProcessID):
If there are no known plug-ins with the provided process ID, invoke the callback with a null
dictionary of plug-in information.
Otherwise, keep track of the callback and send a message to the web process asking for the
information.
(WebKit::WebPageProxy::containsPlugInCallback):
If the page did not contain the plug-in, invoke the callback with a null dictionary of
plug-in information.
Invoke the callback with a dictionary containing the PlugInModuleInfo.

* WebProcess/Plugins/Netscape/NetscapePlugin.h:
Return 0 for the plugInProcessToken.
* WebProcess/Plugins/PDF/SimplePDFPlugin.h:
Ditto.
* WebProcess/Plugins/Plugin.h:
Add a pure virtual method for plugInProcessToken so it can be accessed from from the Plugin
on the PluginView.
* WebProcess/Plugins/PluginProxy.h:
Return the actual plug-in process token.

* WebProcess/Plugins/PluginView.h:
(WebKit::PluginView::plugIn):
Make it possible to get at the plug-in process token through the Plugin.

* WebProcess/WebPage/WebPageMac.mm:
(WebKit::WebPage::containsPluginViewsWithPluginProcessToken):
If there is an instance of the plug-in with that token, send a confirmation to the ui
process.
* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/WebPage.messages.in:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151043 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoTake the normal NSView code path when drawing into an inclusive layer
andersca@apple.com [Fri, 31 May 2013 21:55:24 +0000 (21:55 +0000)]
Take the normal NSView code path when drawing into an inclusive layer
https://bugs.webkit.org/show_bug.cgi?id=117094
<rdar://problem/13812355>

Reviewed by Sam Weinig.

For now, take the normal code path when drawing into an inclusive layer.

* platform/mac/WidgetMac.mm:
(WebCore::Widget::paint):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151042 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoSource/WebCore: Update low memory handler to use new memory pressure notifications...
slewis@apple.com [Fri, 31 May 2013 21:44:26 +0000 (21:44 +0000)]
Source/WebCore: Update low memory handler to use new memory pressure notifications on new OS versions.
<rdar://problem/14027095>

Reviewed by Mark Rowe.

No change in functionality.

* WebCore.exp.in:
* platform/mac/MemoryPressureHandlerMac.mm:
(WebCore::MemoryPressureHandler::install): Call new API
* platform/mac/WebCoreSystemInterface.h:
* platform/mac/WebCoreSystemInterface.mm:

Source/WebKit/mac: Update low memory handler to use new memory pressure notifications on new OS versions.
<rdar://problem/14027095>

Reviewed by Mark Rowe.

* WebCoreSupport/WebSystemInterface.mm:
(InitWebCoreSystemInterface):

Source/WebKit2: Update low memory handler to use new memory pressure notifications on new OS versions.
<rdar://problem/14027095>

Reviewed by Mark Rowe.

* WebProcess/WebCoreSupport/mac/WebSystemInterface.mm:
(InitWebCoreSystemInterface):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151040 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[BlackBerry] Update Button UX.
commit-queue@webkit.org [Fri, 31 May 2013 20:44:53 +0000 (20:44 +0000)]
[BlackBerry] Update Button UX.
https://bugs.webkit.org/show_bug.cgi?id=116992

Patch by Tiancheng Jiang <tijiang@rim.com> on 2013-05-31
Reviewed by Rob Buis.

BlackBerry PR 342298.
Internally Review by Jeff Rogers.

* css/themeBlackBerry.css:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151037 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoDerivedSources.make: any rules that use $(FEATURE_DEFINES) should be rebuilt when...
ddkilzer@apple.com [Fri, 31 May 2013 20:27:41 +0000 (20:27 +0000)]
DerivedSources.make: any rules that use $(FEATURE_DEFINES) should be rebuilt when FeatureDefines.xcconfig changes
<http://webkit.org/b/117081>

Reviewed by Mark Rowe.

* DerivedSources.make: Add $(PLATFORM_FEATURE_DEFINES)
dependency for all rules that use $(FEATURE_DEFINES) to generate
files.
(PLATFORM_FEATURE_DEFINES): Set this variable to
Configurations/FeatureDefines.xcconfig if not already set on
Mac OS X.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151036 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago REGRESSION (r141445): MPEG-4 videos do not play on certain hardware configura...
ap@apple.com [Fri, 31 May 2013 20:16:26 +0000 (20:16 +0000)]
    REGRESSION (r141445): MPEG-4 videos do not play on certain hardware configurations
        https://bugs.webkit.org/show_bug.cgi?id=117089
        <rdar://problem/13839717>

        Reviewed by Mark Rowe.

        * Configurations/WebContentProcess.xcconfig: Statically link to a framework that
        uses an XPC service to work around a libxpc limitation on Lion.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151035 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoText Field UX updates.
commit-queue@webkit.org [Fri, 31 May 2013 19:08:58 +0000 (19:08 +0000)]
Text Field UX updates.
https://bugs.webkit.org/show_bug.cgi?id=116996

Patch by Tiancheng Jiang <tijiang@rim.com> on 2013-05-31
Reviewed by Rob Buis.

BlackBerry PR 342333.
Internally Reviewed by Jeff Rogers.

Update disabled text field font color.

* css/themeBlackBerry.css:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151034 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[sh4] Fix Overflow case of branchMul32 in baseline JIT.
commit-queue@webkit.org [Fri, 31 May 2013 18:22:57 +0000 (18:22 +0000)]
[sh4] Fix Overflow case of branchMul32 in baseline JIT.
https://bugs.webkit.org/show_bug.cgi?id=117057

Patch by Julien Brianceau <jbrianceau@nds.com> on 2013-05-31
Reviewed by Oliver Hunt.

Current implementation of Overflow case in branchMul32 performs an
unsigned multiplication whereas a signed multiplication is expected.

* assembler/MacroAssemblerSH4.h:
(JSC::MacroAssemblerSH4::branchMul32):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151033 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[sh4] Fix floating point comparisons in baseline JIT.
commit-queue@webkit.org [Fri, 31 May 2013 18:22:05 +0000 (18:22 +0000)]
[sh4] Fix floating point comparisons in baseline JIT.
https://bugs.webkit.org/show_bug.cgi?id=117066.

Patch by Julien Brianceau <jbrianceau@nds.com> on 2013-05-31
Reviewed by Oliver Hunt.

Current implementation of branchDouble function in baseline JIT is wrong
for some conditions and overkill for others. For instance:
- With DoubleGreaterThanOrEqual condition, branch will be taken if either
  operand is NaN with current implementation whereras it should not.
- With DoubleNotEqualOrUnordered condition, performed NaN checks are
  useless (because comparison result is false if either operand is NaN).

* assembler/MacroAssemblerSH4.h:
(JSC::MacroAssemblerSH4::branchDouble):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151032 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[sh4] Fix double floating point transfer in baseline JIT.
commit-queue@webkit.org [Fri, 31 May 2013 18:21:23 +0000 (18:21 +0000)]
[sh4] Fix double floating point transfer in baseline JIT.
https://bugs.webkit.org/show_bug.cgi?id=117054

Patch by Julien Brianceau <jbrianceau@nds.com> on 2013-05-31
Reviewed by Oliver Hunt.

In current implementation, dmovRegReg function transfers only one single
FPRegister as PR=1 and SZ=0 in floating point status/control register.
Double transfers must be performed with two fmov.s opcodes.

* assembler/MacroAssemblerSH4.h:
(JSC::MacroAssemblerSH4::moveDouble):
(JSC::MacroAssemblerSH4::addDouble): Handle (op2==dest) case properly.
(JSC::MacroAssemblerSH4::sqrtDouble):
* assembler/SH4Assembler.h:
(JSC::SH4Assembler::fmovsRegReg):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151031 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[sh4] Handle branchType properly in branchTruncateDoubleToInt32.
commit-queue@webkit.org [Fri, 31 May 2013 18:19:15 +0000 (18:19 +0000)]
[sh4] Handle branchType properly in branchTruncateDoubleToInt32.
https://bugs.webkit.org/show_bug.cgi?id=117062

Patch by Julien Brianceau <jbrianceau@nds.com> on 2013-05-31
Reviewed by Oliver Hunt.

Current implementation of branchTruncateDoubleToInt32 is incorrect
when branchType == BranchIfTruncateSuccessful in sh4 baseline JIT.

* assembler/MacroAssemblerSH4.h:
(JSC::MacroAssemblerSH4::branchTruncateDoubleToInt32):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151030 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoUpdate CheckedInt.h to latest version from Mozilla
dino@apple.com [Fri, 31 May 2013 17:45:51 +0000 (17:45 +0000)]
Update CheckedInt.h to latest version from Mozilla
https://bugs.webkit.org/show_bug.cgi?id=117009

Reviewed by Darin Adler.

Mozilla have updated their version of CheckedInt and suggested
we sync. Here are the improvements:

- CheckedInt now consistently rejects unsupported types. Before, you
could sometimes construct a CheckedInt from an unsupported type (like,
float) or multiply a CheckedInt by a float, and have it not caught! See
https://bugzilla.mozilla.org/show_bug.cgi?id=869194
 - Fixed a possibility of undefined behavior in unary operator-. See bug
https://bugzilla.mozilla.org/show_bug.cgi?id=849666
 - CheckedInt now supports the "long long" and "unsigned long long"
types. We don't like "long long" (and neither does the Google C++ style
guide!) but we had to do that because libogg exposes typedefs to those
types. See Bug https://bugzilla.mozilla.org/show_bug.cgi?id=869685
 - CheckedInt also now supports "signed char" in addition to "char" ---
they are not the same type! See Bug https://bugzilla.mozilla.org/show_bug.cgi?id=847521
 - Fixed and simplified usage of CheckedInt outside of Mozilla. You'll
like this one! See Bug https://bugzilla.mozilla.org/show_bug.cgi?id=849667
 - Fixed compilation with the Intel C++ compiler. See bug https://bugzilla.mozilla.org/show_bug.cgi?id=784309

* html/canvas/CheckedInt.h: Updated.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151029 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoDon't create an accelerated compositing layer unless we have to for snapshotting...
dino@apple.com [Fri, 31 May 2013 17:04:13 +0000 (17:04 +0000)]
Don't create an accelerated compositing layer unless we have to for snapshotting plug-ins
https://bugs.webkit.org/show_bug.cgi?id=117076
<rdar://problem/14030364>

Reviewed by Beth Dakin and Anders Carlsson.

If a plug-in can support software snapshotting, then don't create an
accelerated compositing layer for it while in snapshot mode. This is a
better user experience -- otherwise the page/plug-in looks like it broke.

* WebProcess/Plugins/PluginView.cpp:
(WebKit::PluginView::isAcceleratedCompositingEnabled): If we are in snapshotting mode
and we are a Flash plugin (the only one that supports software snapshots at the moment)
then do not create an accelerated compositing layer.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151026 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[Windows] Unreviewed build fix for VS2005 builders.
bfulgham@apple.com [Fri, 31 May 2013 17:02:14 +0000 (17:02 +0000)]
[Windows] Unreviewed build fix for VS2005 builders.

* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreExports.def: Add
missing export for WTF::SHA1::computeHexDigest

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151025 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoUnreviewed, rolling out r151022.
commit-queue@webkit.org [Fri, 31 May 2013 16:57:56 +0000 (16:57 +0000)]
Unreviewed, rolling out r151022.
http://trac.webkit.org/changeset/151022
https://bugs.webkit.org/show_bug.cgi?id=117084

Also breaks build by introducing duplicate symbols (Requested
by bfulgham on #webkit).

* TestWebKitAPI/Configurations/TestWebKitAPICommon.vsprops:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151024 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[Windows] Unreviewed build fix to NRWT
bfulgham@apple.com [Fri, 31 May 2013 16:45:31 +0000 (16:45 +0000)]
[Windows] Unreviewed build fix to NRWT

* Scripts/webkitpy/port/win.py:
(WinPort.create_debugger_command_file): Use proper argument for _filesystem.write_text_file.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151023 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[Windows] Unreviewed build fix after r151012
bfulgham@apple.com [Fri, 31 May 2013 16:31:29 +0000 (16:31 +0000)]
[Windows] Unreviewed build fix after r151012

* TestWebKitAPI/Configurations/TestWebKitAPICommon.vsprops: Add dependency on
WTF library to support SHA tests

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151022 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoMake sure gstreamer source element is thread-safe
commit-queue@webkit.org [Fri, 31 May 2013 16:17:47 +0000 (16:17 +0000)]
Make sure gstreamer source element is thread-safe
https://bugs.webkit.org/show_bug.cgi?id=115352

Patch by Andre Moreira Magalhaes <andre.magalhaes@collabora.co.uk> on 2013-05-31
Reviewed by Philippe Normand.

GStreamer source element may be created by any gstreamer element on any thread by calling
gst_element_make_from_uri with the URIs handled by the source element.
This patch makes sure the gstreamer source element is thread-safe to avoid issues with it
being created outside the main thread.

* platform/graphics/gstreamer/WebKitWebSourceGStreamer.cpp:
(_WebKitWebSrcPrivate):
(webkit_web_src_init):
(webKitWebSrcFinalize):
(webKitWebSrcSetProperty):
(webKitWebSrcGetProperty):
(webKitWebSrcStop):
(webKitWebSrcStart):
(webKitWebSrcChangeState):
(webKitWebSrcQueryWithParent):
(webKitWebSrcGetUri):
(webKitWebSrcSetUri):
(webKitWebSrcNeedDataMainCb):
(webKitWebSrcEnoughDataMainCb):
(webKitWebSrcSeekMainCb):
(webKitWebSrcSeekDataCb):
(webKitWebSrcSetMediaPlayer):
(StreamingClient::StreamingClient):
(StreamingClient::~StreamingClient):
(StreamingClient::didReceiveResponse):
(StreamingClient::didReceiveData):
(StreamingClient::didFinishLoading):
(StreamingClient::wasBlocked):
(StreamingClient::cannotShowURL):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151021 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoFix double hash lookup in PageMapMemoryUsageRecorder::visit
rafael.lobo@openbossa.org [Fri, 31 May 2013 16:16:11 +0000 (16:16 +0000)]
Fix double hash lookup in PageMapMemoryUsageRecorder::visit
https://bugs.webkit.org/show_bug.cgi?id=117071

Reviewed by Andreas Kling.

* wtf/FastMalloc.cpp:
(WTF::TCMalloc_ThreadCache::CreateCacheIfNecessary):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151020 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[Windows] Unreviewed build fix after r151012
bfulgham@apple.com [Fri, 31 May 2013 16:14:22 +0000 (16:14 +0000)]
[Windows] Unreviewed build fix after r151012

* TestWebKitAPI/Tests/WTF/MD5.cpp: Include <wtf/StringExtras.h> to provide
snprintf definition for Windows.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151019 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years ago[Windows] Speculative build fix for pre-VS2010 compilers
bfulgham@apple.com [Fri, 31 May 2013 16:01:35 +0000 (16:01 +0000)]
[Windows] Speculative build fix for pre-VS2010 compilers

* wtf/Platform.h: Only treat WTF as DLL for VS2010 builds

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151018 268f45cc-cd09-0410-ab3c-d52691b4dbfc

6 years agoUnreviewed gardening. Update test-expectations after r151000.
jer.noble@apple.com [Fri, 31 May 2013 15:53:54 +0000 (15:53 +0000)]
Unreviewed gardening. Update test-expectations after r151000.

* fast/js/date-DST-time-cusps-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@151017 268f45cc-cd09-0410-ab3c-d52691b4dbfc