WebKit-https.git
8 years agoAddress one more review comment.
rniwa@webkit.org [Thu, 9 Aug 2012 21:59:19 +0000 (21:59 +0000)]
Address one more review comment.

* building/checkout.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125210 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoRecommend the use of https and mention git on building/checkout.html
rniwa@webkit.org [Thu, 9 Aug 2012 21:57:33 +0000 (21:57 +0000)]
Recommend the use of https and mention git on building/checkout.html
https://bugs.webkit.org/show_bug.cgi?id=93657

Reviewed by Mark Rowe.

This page was quite out-dated. Stop mentioning about how to install subversion on OS X 10.4
because the paragraph right above it refers to buliding/tools.html for that.

Recommend the use of HTTPS connection for the Subversion checkout and also mention how to
checkout using a Git client.

* building/checkout.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125208 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[BlackBerry] Export InRegionScroller to fix build
commit-queue@webkit.org [Thu, 9 Aug 2012 21:56:57 +0000 (21:56 +0000)]
[BlackBerry] Export InRegionScroller to fix build
https://bugs.webkit.org/show_bug.cgi?id=93635

Patch by Leo Yang <leoyang@rim.com> on 2012-08-09
Reviewed by Antonio Gomes.

Build fix.

* Api/InRegionScroller.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125207 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[Qt] Unreviewed gardening, paint the bots green.
ossy@webkit.org [Thu, 9 Aug 2012 21:51:56 +0000 (21:51 +0000)]
[Qt] Unreviewed gardening, paint the bots green.

* platform/qt-5.0-wk2/Skipped: Skip failing tests.
* platform/qt-5.0-wk2/fast/forms/mailto/advanced-get-expected.txt: Added because of whitespace differenced.
* platform/qt-5.0-wk2/fast/forms/mailto/advanced-put-expected.txt: Added because of whitespace differenced.
* platform/qt/TestExpectations: Mark an asserting test as crash.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125206 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[css3-text] Add suport for -webkit-text-decoration-line
commit-queue@webkit.org [Thu, 9 Aug 2012 21:49:36 +0000 (21:49 +0000)]
[css3-text] Add suport for -webkit-text-decoration-line
https://bugs.webkit.org/show_bug.cgi?id=90959

Patch by Bruno de Oliveira Abinader <bruno.abinader@basyskom.com> on 2012-08-09
Reviewed by Kenneth Rohde Christiansen.

This patch implements the "text-decoration-line" property specified in CSS3
working draft, with "-webkit-" prefix. The specification can be found below:
http://dev.w3.org/csswg/css3-text/#text-decoration-line

Additionally, Mozilla implementation details can be found here:
https://developer-dev.allizom.org/en-US/docs/CSS/text-decoration-line

It is important to notice that "-webkit-text-decoration-line" and
"text-decoration" properties shares the same property handlers, however, the
former resets the latter if value from latter doesn't have "important" value
set (same behavior as Firefox). In a rare case where both properties have
"important" value set, the text decoration gets value from the last parsed
property.

Source/WebCore:

Test: fast/css/text-decoration-line.html

* css/CSSComputedStyleDeclaration.cpp:
(WebCore::renderTextDecorationFlagsToCSSValue):
(WebCore::CSSComputedStyleDeclaration::getPropertyCSSValue):
* css/CSSParser.cpp:
(WebCore::CSSParser::parseValue):
(WebCore::CSSParser::addTextDecorationProperty):
(WebCore::CSSParser::parseTextDecoration):
* css/CSSParser.h:
* css/CSSProperty.cpp:
(WebCore::CSSProperty::isInheritedProperty):
* css/CSSPropertyNames.in:
* css/StyleBuilder.cpp:
(WebCore::StyleBuilder::StyleBuilder):
* css/StyleResolver.cpp:
(WebCore::StyleResolver::collectMatchingRulesForList):

LayoutTests:

* fast/css/getComputedStyle/computed-style-expected.txt:
* fast/css/getComputedStyle/computed-style-without-renderer-expected.txt:
* fast/css/getComputedStyle/resources/property-names.js:
* fast/css/text-decoration-line-expected.html: Added.
* fast/css/text-decoration-line.html: Added.
* svg/css/getComputedStyle-basic-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125205 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agonrwt: should log progress like ninja does
dpranke@chromium.org [Thu, 9 Aug 2012 21:35:55 +0000 (21:35 +0000)]
nrwt: should log progress like ninja does
https://bugs.webkit.org/show_bug.cgi?id=93563

Reviewed by Ojan Vafai.

This change modifies the output of nrwt in the
non--debug-rwt-logging modes.

In non-verbose mode, the output changes from:
  "Testing (50%): 96 ran as expected, 4 didn't, 100 left"
to:
  "[100/200] foo.html (+4)"
or:
  "[100/200] foo.html failed  unexpectedly (text diff)"

which indicates the earliest test currently running (optionally plus the
count of other currently-executing tests if we're running tests in parallel)
or the result of said test.

If the test produces unexpected results, that line is flushed
rather than erased (so we no longer need the %d didn't running
tally).

In --verbose mode, the output is similar, but every test gets a
line (and tests are logged in order of them being started, not
order of them being completed) (and in --details mode, the
first line of each test matches --verbose but the other lines
are unchanged).

The overall result should be easier to read and is simpler to
implement as well (apart from managing the number of tests that
have started but not completed, which is slightly annoying).

This also matches the output style for test-webkitpy, more-or-less.

* Scripts/webkitpy/layout_tests/controllers/layout_test_runner.py:
(LayoutTestRunner.run_tests):
(LayoutTestRunner._update_summary_with_result):
(LayoutTestRunner._handle_started_test):
* Scripts/webkitpy/layout_tests/controllers/layout_test_runner_unittest.py:
(FakePrinter):
(FakePrinter.print_started_test):
(FakePrinter.print_finished_test):
* Scripts/webkitpy/layout_tests/views/printing.py:
(Printer.__init__):
(Printer.print_started_test):
(Printer.print_finished_test):
(Printer._print_test_trace):
(Printer._print_progress):
* Scripts/webkitpy/layout_tests/views/printing_unittest.py:
(test_details):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125204 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoExtract the definition of NativeImagePtr from ImageSource.h into NativeImagePtr.h.
commit-queue@webkit.org [Thu, 9 Aug 2012 21:25:32 +0000 (21:25 +0000)]
Extract the definition of NativeImagePtr from ImageSource.h into NativeImagePtr.h.
https://bugs.webkit.org/show_bug.cgi?id=93466

Patch by Huang Dongsung <luxtella@company100.net> on 2012-08-09
Reviewed by Adam Barth.

Image.h includes ImageSource.h in order to know the definition of NativeImagePtr.
But Image does not need to know the definition of ImageSource. So, this patch
extracts the definition of NativeImagePtr from ImageSource.h into NativeImagePtr.h.

This is a preparation for asynchronous image decoding. Async decoding
will need a callback function, which sends a NativeImagePtr to clients, and the
callback must know the definition of NativeImagePtr, not ImageSource.

No new tests, behavior is unchanged.

* GNUmakefile.list.am:
* Target.pri:
* WebCore.gypi:
* WebCore.vcproj/WebCore.vcproj:
* WebCore.xcodeproj/project.pbxproj:
* platform/graphics/BitmapImage.h:
* platform/graphics/Image.h:
* platform/graphics/ImageSource.h:
(WebCore):
* platform/graphics/NativeImagePtr.h: Added.
(WebCore):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125203 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoCSS 2.1 failure: 'Text-indent' only affects a line if it is the first formatted line...
robert@webkit.org [Thu, 9 Aug 2012 21:12:44 +0000 (21:12 +0000)]
CSS 2.1 failure: 'Text-indent' only affects a line if it is the first formatted line of an element
https://bugs.webkit.org/show_bug.cgi?id=63185

Reviewed by Eric Seidel.

Source/WebCore:

Per CSS 2.1 (http://www.w3.org/TR/CSS21/text.html) : "'Text-indent' only affects a line if it is the
first formatted line of an element. For example, the first line of an anonymous block box is only
affected if it is the first child of its parent element."

Tests: fast/css/text-indent-first-line-001.html
       fast/css/text-indent-first-line-002.html
       fast/css/text-indent-first-line-003.html
       fast/css/text-indent-first-line-004.html
       fast/css/text-indent-first-line-005.html
       fast/css/text-indent-first-line-006.html

* rendering/RenderBlockLineLayout.cpp:
(WebCore::RenderBlock::computeInlineDirectionPositionsForLine): As well as being the first line under a parent block,
  check that this is also the first *formatted* line of an element by ensuring that we are at the start of a
  formatting context and not just a continuation of a previous anonymous block in an existing formatting context.

LayoutTests:

* css2.1/20110323/text-indent-014-expected.html: Added.
* css2.1/20110323/text-indent-014.htm: Added.
* fast/css/text-indent-first-line-001-expected.html: Added.
* fast/css/text-indent-first-line-001.html: Added.
* fast/css/text-indent-first-line-002-expected.html: Added.
* fast/css/text-indent-first-line-002.html: Added.
* fast/css/text-indent-first-line-003-expected.html: Added.
* fast/css/text-indent-first-line-003.html: Added.
* fast/css/text-indent-first-line-004-expected.html: Added.
* fast/css/text-indent-first-line-004.html: Added.
* fast/css/text-indent-first-line-005-expected.html: Added.
* fast/css/text-indent-first-line-005.html: Added.
  Text-indent in the span here is inherited from the div. CSS 2.1: "Since the 'text-indent'
  property inherits, when specified on a block element, it will affect descendant
  inline-block elements."
* fast/css/text-indent-first-line-006-expected.html: Added.
* fast/css/text-indent-first-line-006.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125202 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoWeb Inspector: replace the Web Inspector editor with CodeMirror
commit-queue@webkit.org [Thu, 9 Aug 2012 21:00:46 +0000 (21:00 +0000)]
Web Inspector: replace the Web Inspector editor with CodeMirror
https://bugs.webkit.org/show_bug.cgi?id=92769

Patch by Jan Keromnes <janx@linux.com> on 2012-08-09
Reviewed by Pavel Feldman.

Integrating CodeMirror as a devtools experiment. Currently supporting
basic editing, search/replacing, and saving. Missing proper breakpoints
wiring, debugging, message bubbles, popovers and context menus. WIP.

* WebCore.gypi:
* WebCore.vcproj/WebCore.vcproj:
* inspector/compile-front-end.py:
* inspector/front-end/CodeMirrorTextEditor.js: Added.
(WebInspector.CodeMirrorTextEditor.this._codeMirror.CodeMirror):
(WebInspector.CodeMirrorTextEditor):
(WebInspector.CodeMirrorTextEditor.prototype.set mimeType):
(WebInspector.CodeMirrorTextEditor.prototype.setReadOnly):
(WebInspector.CodeMirrorTextEditor.prototype.readOnly):
(WebInspector.CodeMirrorTextEditor.prototype.defaultFocusedElement):
(WebInspector.CodeMirrorTextEditor.prototype.focus):
(WebInspector.CodeMirrorTextEditor.prototype.revealLine):
(WebInspector.CodeMirrorTextEditor.prototype.addDecoration):
(WebInspector.CodeMirrorTextEditor.prototype.removeDecoration):
(WebInspector.CodeMirrorTextEditor.prototype.markAndRevealRange):
(WebInspector.CodeMirrorTextEditor.prototype.highlightLine):
(WebInspector.CodeMirrorTextEditor.prototype.clearLineHighlight):
(WebInspector.CodeMirrorTextEditor.prototype.freeCachedElements):
(WebInspector.CodeMirrorTextEditor.prototype.elementsToRestoreScrollPositionsFor):
(WebInspector.CodeMirrorTextEditor.prototype.inheritScrollPositions):
(WebInspector.CodeMirrorTextEditor.prototype.beginUpdates):
(WebInspector.CodeMirrorTextEditor.prototype.endUpdates):
(WebInspector.CodeMirrorTextEditor.prototype.onResize):
(WebInspector.CodeMirrorTextEditor.prototype.editRange):
(WebInspector.CodeMirrorTextEditor.prototype._onChange):
(WebInspector.CodeMirrorTextEditor.prototype._commitEditing):
(WebInspector.CodeMirrorTextEditor.prototype.scrollToLine):
(WebInspector.CodeMirrorTextEditor.prototype.selection):
(WebInspector.CodeMirrorTextEditor.prototype.lastSelection):
(WebInspector.CodeMirrorTextEditor.prototype.setSelection):
(WebInspector.CodeMirrorTextEditor.prototype.setText):
(WebInspector.CodeMirrorTextEditor.prototype.text):
(WebInspector.CodeMirrorTextEditor.prototype.range):
(WebInspector.CodeMirrorTextEditor.prototype.line):
(WebInspector.CodeMirrorTextEditor.prototype.get linesCount):
(WebInspector.CodeMirrorTextEditor.prototype.setAttribute):
(WebInspector.CodeMirrorTextEditor.prototype.getAttribute):
(WebInspector.CodeMirrorTextEditor.prototype.removeAttribute):
(WebInspector.CodeMirrorTextEditor.prototype.wasShown):
(WebInspector.CodeMirrorTextEditor.prototype.willHide):
(WebInspector.CodeMirrorTextEditor.prototype._textChanged):
(WebInspector.CodeMirrorTextEditor.prototype._toPos):
(WebInspector.CodeMirrorTextEditor.prototype._toRange):
* inspector/front-end/Settings.js:
(WebInspector.ExperimentsSettings):
* inspector/front-end/SourceFrame.js:
(WebInspector.SourceFrame):
* inspector/front-end/WebKit.qrc:
* inspector/front-end/cmdevtools.css: Added.
* inspector/front-end/inspector.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125201 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoGardening LayoutTests/platform/win/fast/canvas tests folder.
commit-queue@webkit.org [Thu, 9 Aug 2012 20:58:43 +0000 (20:58 +0000)]
Gardening LayoutTests/platform/win/fast/canvas tests folder.
https://bugs.webkit.org/show_bug.cgi?id=93633

Patch by Roger Fong <roger_fong@apple.com> on 2012-08-09
Reviewed by Tim Horton.

These tests pass now.
Adjusting expected results.
Removing other failing results.

* platform/win/fast/canvas/canvas-draw-canvas-on-canvas-shadow-expected.txt:
    This test is now passing. Modifying to passing results since Mac fallback results are failing.
* platform/win/fast/canvas/canvas-scale-drawImage-shadow-expected.txt: Removed.
* platform/win/fast/canvas/canvas-scale-fillPath-shadow-expected.txt: Removed.
* platform/win/fast/canvas/canvas-scale-fillRect-shadow-expected.txt: Removed.
* platform/win/fast/canvas/canvas-transforms-fillRect-shadow-expected.txt: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125200 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[QNX] Implement getCPUTime() for OS(QNX)
commit-queue@webkit.org [Thu, 9 Aug 2012 20:38:56 +0000 (20:38 +0000)]
[QNX] Implement getCPUTime() for OS(QNX)
https://bugs.webkit.org/show_bug.cgi?id=93516

Patch by Yong Li <yoli@rim.com> on 2012-08-09
Reviewed by George Staikos.

Implement getCPUTime() with CLOCK_THREAD_CPUTIME_ID so it will tell
exactly how long the current thread has spent without being impacted
by other things.

* runtime/TimeoutChecker.cpp:
(JSC::getCPUTime):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125199 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoVersioning.
lforschler@apple.com [Thu, 9 Aug 2012 20:35:13 +0000 (20:35 +0000)]
Versioning.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125198 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago [WK2] Add a process model constant for multiple processes
ap@apple.com [Thu, 9 Aug 2012 20:31:17 +0000 (20:31 +0000)]
    [WK2] Add a process model constant for multiple processes
        https://bugs.webkit.org/show_bug.cgi?id=93642

        Reviewed by Anders Carlsson.

        * UIProcess/ProcessModel.h: Added ProcessModelMultipleSecondaryProcesses.
        Deleted ProcessModelSecondaryProcess in favor of more descriptive ProcessModelSharedSecondaryProcess -
        these used to behave the same.

        * UIProcess/WebContext.cpp: (WebKit::WebContext::create): Changed to use ProcessModelSharedSecondaryProcess.

        * UIProcess/WebProcessProxy.h: Removed an unncessary include of ProcessModel.h.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125197 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[chromium] Do partial-swap scissoring on quads during draw instead of on layers
danakj@chromium.org [Thu, 9 Aug 2012 20:15:03 +0000 (20:15 +0000)]
[chromium] Do partial-swap scissoring on quads during draw instead of on layers
https://bugs.webkit.org/show_bug.cgi?id=91800

Reviewed by Adrienne Walker.

Source/WebCore:

Computing the scissor rect for each render pass is now done in
LayerRendererChromium::drawRenderPass() instead of being done for each
layer and then copied into the quads. This allows us to always push a
full frame of quads (and include damage rects) from a child compositor,
and do partial-swap scissoring in the host compositor.

This also make partial-swap functionality completely contained within
LayerRendererChromium!

The scissorRect is removed from all layers, and the concept of
scissorRect is now misplaced in CCOcclusionTracker, so the variables
and methods in the tracker should be renamed to more appropriate
things.

Covered by existing tests, as well as: CCLayerTreeHostImplTest.textureCachingWithScissor

* platform/graphics/chromium/LayerChromium.h:
(LayerChromium):
* platform/graphics/chromium/LayerRendererChromium.cpp:
(WebCore::LayerRendererChromium::clearFramebuffer):
(WebCore::LayerRendererChromium::renderPassTextureSize):
(WebCore::LayerRendererChromium::drawFrame):
(WebCore::LayerRendererChromium::drawRenderPass):
(WebCore::LayerRendererChromium::drawQuad):
(WebCore::LayerRendererChromium::drawDebugBorderQuad):
(WebCore::LayerRendererChromium::drawBackgroundFilters):
(WebCore::LayerRendererChromium::drawTileQuad):
(WebCore::LayerRendererChromium::swapBuffers):
(WebCore::LayerRendererChromium::useRenderPass):
(WebCore::LayerRendererChromium::setScissorToRect):
* platform/graphics/chromium/LayerRendererChromium.h:
(LayerRendererChromium):
(DrawingFrame):
* platform/graphics/chromium/RenderSurfaceChromium.cpp:
* platform/graphics/chromium/RenderSurfaceChromium.h:
(RenderSurfaceChromium):
* platform/graphics/chromium/cc/CCDrawQuad.h:
(WebCore::CCDrawQuad::clippedRectInTarget):
* platform/graphics/chromium/cc/CCLayerImpl.cpp:
(WebCore::CCLayerImpl::createSharedQuadState):
* platform/graphics/chromium/cc/CCLayerImpl.h:
(CCLayerImpl):
* platform/graphics/chromium/cc/CCLayerTreeHost.cpp:
(WebCore::CCLayerTreeHost::updateLayers):
(WebCore::CCLayerTreeHost::paintLayerContents):
* platform/graphics/chromium/cc/CCLayerTreeHostCommon.cpp:
(WebCore::calculateVisibleRectsInternal):
(WebCore::CCLayerTreeHostCommon::calculateVisibleRects):
* platform/graphics/chromium/cc/CCLayerTreeHostCommon.h:
(CCLayerTreeHostCommon):
* platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp:
(WebCore::CCLayerTreeHostImpl::calculateRenderSurfaceLayerList):
(WebCore::CCLayerTreeHostImpl::calculateRenderPasses):
(WebCore::CCLayerTreeHostImpl::drawLayers):
(WebCore::CCLayerTreeHostImpl::swapBuffers):
* platform/graphics/chromium/cc/CCLayerTreeHostImpl.h:
(CCLayerTreeHostImpl):
* platform/graphics/chromium/cc/CCOcclusionTracker.cpp:
(WebCore::::layerScissorRectInTargetSurface):
* platform/graphics/chromium/cc/CCRenderPass.cpp:
(WebCore::CCRenderPass::CCRenderPass):
(WebCore::CCRenderPass::appendQuadsToFillScreen):
* platform/graphics/chromium/cc/CCRenderPass.h:
(CCRenderPass):
(WebCore::CCRenderPass::transformToRootTarget):
(WebCore::CCRenderPass::outputRect):
(WebCore::CCRenderPass::damageRect):
(WebCore::CCRenderPass::setDamageRect):
* platform/graphics/chromium/cc/CCRenderSurface.cpp:
(WebCore::computeClippedRectInTarget):
(WebCore):
(WebCore::CCRenderSurface::createSharedQuadState):
(WebCore::CCRenderSurface::createReplicaSharedQuadState):
* platform/graphics/chromium/cc/CCRenderSurface.h:
(CCRenderSurface):
* platform/graphics/chromium/cc/CCRenderer.h:
(CCRenderer):
* platform/graphics/chromium/cc/CCSharedQuadState.cpp:
(WebCore::CCSharedQuadState::create):
(WebCore::CCSharedQuadState::CCSharedQuadState):
* platform/graphics/chromium/cc/CCSharedQuadState.h:
(CCSharedQuadState):
* platform/graphics/chromium/cc/CCTiledLayerImpl.cpp:
(WebCore::CCTiledLayerImpl::appendQuads):

Source/WebKit/chromium:

* tests/CCDamageTrackerTest.cpp:
(WebKitTests::executeCalculateDrawTransformsAndVisibility):
* tests/CCLayerIteratorTest.cpp:
* tests/CCLayerTreeHostCommonTest.cpp:
Removed tests of the scissor rect here, as they are no longer
computed within CCLayerTreeHostCommon. Verify that the scissor
does not affect texture caching in CCLayerTreeHostImplTest now.
Previous scissor would affect the texture cache decisions, but
this was incorrect, as damaged areas are always inside the
scissor rect, and anything outside the scissor rect should
be considered as valid and complete (assuming no external
occlusion, which is computed outside the scissor rect now also).

* tests/CCLayerTreeHostImplTest.cpp:
Removed didDrawNotCalledOnScissoredLayer test, as this no longer
applies. didDraw will be called on layers that are outside the
partial swap rect, regardless.

Removed partialSwapNoUpdate test, and merged its functionality into
the partialSwap test while updating its expectations.

Added textureCachingWithScissor, which is loosely based on the
CCLayerTreeHostCommonTest.scissorRectWithClip test.

* tests/CCOcclusionTrackerTest.cpp:
(WebKitTests::CCOcclusionTrackerTest::calcDrawEtc):
(WebKitTests::CCOcclusionTrackerTestChildInRotatedChild::runMyTest):
(WebKitTests::CCOcclusionTrackerTestVisitTargetTwoTimes::runMyTest):
* tests/CCQuadCullerTest.cpp:
* tests/CCRenderSurfaceTest.cpp:
* tests/CCSolidColorLayerImplTest.cpp:
(CCLayerTestCommon::TEST):
* tests/CCTiledLayerImplTest.cpp:
(CCLayerTestCommon::createLayer):
* tests/LayerRendererChromiumTest.cpp:
(LayerRendererChromiumTest::swapBuffers):
(TEST_F):
(TEST):
* tests/TiledLayerChromiumTest.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125196 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoSource/WebCore: Content Security Policy directives that begin with an invalid charact...
commit-queue@webkit.org [Thu, 9 Aug 2012 20:08:23 +0000 (20:08 +0000)]
Source/WebCore: Content Security Policy directives that begin with an invalid character should log a console warning.
https://bugs.webkit.org/show_bug.cgi?id=93622

Patch by Mike West <mkwst@chromium.org> on 2012-08-09
Reviewed by Adam Barth.

CSP directives containing unrecognized characters somewhere in the
middle were caught and logged by the current algorithm. We additionally
caught the case in which the directive begins with an invalid character,
but we silently ignored it. Now we're slightly more vocal.

This change also exits 'parseDirective' early in the case where the
directive text is empty, or filled only with whitespace. There doesn't
seem to be any value in flagging that case, as it has no impact on the
way the policy would have been interpreted (that is, 'img-src *;;...'
doesn't change in meaning by ignoring the empty directive between the
semicolons).

Test: http/tests/security/contentSecurityPolicy/directive-parsing-05.html

* page/ContentSecurityPolicy.cpp:
(WebCore::CSPDirectiveList::parseDirective):
    Two changes: first, we now exit early if the entire directive text
    is empty (e.g. ';;;' or ';      ;'); second, if the directive begins
    with a character that doesn't match 'isDirectiveNameCharacter', then
    we advance either to the next space, or the end of the value, and
    report an unrecognized directive to the policy.

LayoutTests: Content Security Policy directives with an invalid character should log a console warning.
https://bugs.webkit.org/show_bug.cgi?id=93622

Patch by Mike West <mkwst@chromium.org> on 2012-08-09
Reviewed by Adam Barth.

* http/tests/security/contentSecurityPolicy/directive-parsing-05-expected.txt: Added.
* http/tests/security/contentSecurityPolicy/directive-parsing-05.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125195 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoPerformanceTests: Another build fix after r125178.
rniwa@webkit.org [Thu, 9 Aug 2012 20:01:04 +0000 (20:01 +0000)]
PerformanceTests: Another build fix after r125178.

* resources/runner.js:

LayoutTests: Fix a test after r125178.

* fast/harness/perftests/runs-per-second-log-expected.txt:
* fast/harness/perftests/runs-per-second-log.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125194 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[chromium] Pass mask scale and offset to shaders for correct masking
shawnsingh@chromium.org [Thu, 9 Aug 2012 19:39:24 +0000 (19:39 +0000)]
[chromium] Pass mask scale and offset to shaders for correct masking
https://bugs.webkit.org/show_bug.cgi?id=93558

Reviewed by Adrienne Walker.

Source/WebCore:

In chromium compositor code, the mask was accidentally mapped 1:1
to the layer's renderSurface, rather than the layer itself. This
patch adds a layout test that reproduces the error and fixes the
problem. The solution we opted for in this patch is to pass a
scale and offset to the fragment shader, so that the correct
texture coordinate lookup is computed by the fragment shader.

Test: compositing/masks/mask-of-clipped-layer.html

* platform/graphics/chromium/LayerRendererChromium.cpp:
(WebCore::LayerRendererChromium::drawRenderPassQuad):
* platform/graphics/chromium/ShaderChromium.cpp:
(WebCore::FragmentShaderRGBATexAlphaMask::FragmentShaderRGBATexAlphaMask):
(WebCore::FragmentShaderRGBATexAlphaMask::init):
(WebCore::FragmentShaderRGBATexAlphaMask::getShaderString):
(WebCore::FragmentShaderRGBATexAlphaMaskAA::FragmentShaderRGBATexAlphaMaskAA):
(WebCore::FragmentShaderRGBATexAlphaMaskAA::init):
(WebCore::FragmentShaderRGBATexAlphaMaskAA::getShaderString):
* platform/graphics/chromium/ShaderChromium.h:
(WebCore::FragmentShaderRGBATexAlphaMask::maskTexCoordScaleLocation):
(WebCore::FragmentShaderRGBATexAlphaMask::maskTexCoordOffsetLocation):
(FragmentShaderRGBATexAlphaMask):
(WebCore::FragmentShaderRGBATexAlphaMaskAA::maskTexCoordScaleLocation):
(WebCore::FragmentShaderRGBATexAlphaMaskAA::maskTexCoordOffsetLocation):
(FragmentShaderRGBATexAlphaMaskAA):
* platform/graphics/chromium/cc/CCRenderPassDrawQuad.cpp:
(WebCore::CCRenderPassDrawQuad::create):
(WebCore::CCRenderPassDrawQuad::CCRenderPassDrawQuad):
* platform/graphics/chromium/cc/CCRenderPassDrawQuad.h:
(CCRenderPassDrawQuad):
(WebCore::CCRenderPassDrawQuad::maskTexCoordScaleX):
(WebCore::CCRenderPassDrawQuad::maskTexCoordScaleY):
(WebCore::CCRenderPassDrawQuad::maskTexCoordOffsetX):
(WebCore::CCRenderPassDrawQuad::maskTexCoordOffsetY):
* platform/graphics/chromium/cc/CCRenderSurface.cpp:
(WebCore::CCRenderSurface::appendQuads):

Source/WebKit/chromium:

Updated one unit test that needed additional args in a
constructor, but no change in behavior of the test.

* tests/CCLayerTreeHostImplTest.cpp:

LayoutTests:

Added a layout test that checks the mask does not accidentally get
resized when it is inside of an overflow div.

* compositing/masks/mask-of-clipped-layer-expected.png: Added.
* compositing/masks/mask-of-clipped-layer-expected.txt: Added.
* compositing/masks/mask-of-clipped-layer.html: Added.
* compositing/resources/alpha-blocks.png: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125193 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoCSSRegions: Crash when attaching a region to the removed named flow
mihnea@adobe.com [Thu, 9 Aug 2012 19:13:22 +0000 (19:13 +0000)]
CSSRegions: Crash when attaching a region to the removed named flow
https://bugs.webkit.org/show_bug.cgi?id=93204

Reviewed by Abhishek Arya.

Source/WebCore:

The crash happens because the flow thread reference is passed to the RenderRegion object
when the RenderRegion object is created. However, it is only in RenderRegion::attachRegion method
when the region is taken into account and stored in flow thread regions list. Under these circumstances,
it is possible that the flow thread passed to the region at the moment of region creation be destroyed
before the region gets the chance to be attached to the flow thread and a crash will occur.

This patch moves the moment when the RenderRegion object stores its 'weak' reference to the flow thread
from RenderRegion constructor to attachRegion method, preventing the situation in which the region is attached
to a wrong flow thread object.

Test: fast/regions/region-attached-wrong-flow-crash.html

* rendering/RenderObject.cpp:
(WebCore::RenderObject::createObject): Pass 0 instead of a reference to the flow thread.
* rendering/RenderRegion.cpp:
(WebCore::RenderRegion::checkRegionStyle): Added this function to check whether the region has region style rules attached.
(WebCore):
(WebCore::RenderRegion::styleDidChange): Use checkRegionStyle.
(WebCore::RenderRegion::attachRegion): Store the reference to the flow thread. If the flow thread does not exist, it will be created.
Also, check whether the region has region style rules attached, since at this moment the region has a flow thread. Checking whether
the region has region style rules is not enough in styleDidChange, because styleDidChange is sometimes triggered before attachRegion.
* rendering/RenderRegion.h:
(RenderRegion):

LayoutTests:

Add test reproducing the crash.

* fast/regions/region-attached-wrong-flow-crash-expected.txt: Added.
* fast/regions/region-attached-wrong-flow-crash.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125192 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago(r125185) http/tests/images/jpg-img-partial-load.html timeout on Lion
dino@apple.com [Thu, 9 Aug 2012 18:53:23 +0000 (18:53 +0000)]
(r125185) http/tests/images/jpg-img-partial-load.html timeout on Lion
https://bugs.webkit.org/show_bug.cgi?id=93636

Unreviewed test expectations update.

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125191 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[Chromium] Updated expectations for script-change-externalResourcesRequired-while...
fmalita@chromium.org [Thu, 9 Aug 2012 18:46:19 +0000 (18:46 +0000)]
[Chromium] Updated expectations for script-change-externalResourcesRequired-while-loading.svg

Unreviewed gardening.

Patch by Florin Malita <fmalita@chromium.org> on 2012-08-09

* platform/chromium/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125190 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[chromium] Enable kerning on Android
commit-queue@webkit.org [Thu, 9 Aug 2012 18:41:05 +0000 (18:41 +0000)]
[chromium] Enable kerning on Android
https://bugs.webkit.org/show_bug.cgi?id=92737

Patch by Iain Merrick <husky@chromium.org> on 2012-08-09
Reviewed by Adam Barth.

This #ifdef was incorrectly copied from our forked WebKit repository
while unforking our Android-specific changes. Kerning works without
any known problems.

Covered by existing tests (layout and instrumentation tests).

* platform/graphics/harfbuzz/ComplexTextControllerHarfBuzz.cpp:
(WebCore::ComplexTextController::ComplexTextController):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125189 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoBuild fix after r125178.
rniwa@webkit.org [Thu, 9 Aug 2012 18:27:46 +0000 (18:27 +0000)]
Build fix after r125178.

* Scripts/webkitpy/performance_tests/perftest.py:
(PerfTest.parse_output): Don't include non-time results in the JSON output.
* Scripts/webkitpy/performance_tests/perftest_unittest.py: Updated the test outputs so to match
actual test outputs.
(MainTest.test_parse_output):
(MainTest.test_parse_output_with_failing_line):
* Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py: Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125188 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[EFL] Gardening after r125102
commit-queue@webkit.org [Thu, 9 Aug 2012 18:07:00 +0000 (18:07 +0000)]
[EFL] Gardening after r125102
https://bugs.webkit.org/show_bug.cgi?id=93593

Unreviewed gardening.

Patch by Mikhail Pozdnyakov <mikhail.pozdnyakov@intel.com> on 2012-08-09

* platform/efl/TestExpectations: Included newly added compositing/plugins/no-backing-store.html test.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125187 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoWeb Inspector: follow up to r125174 - fix subtype use.
pfeldman@chromium.org [Thu, 9 Aug 2012 17:56:04 +0000 (17:56 +0000)]
Web Inspector: follow up to r125174 - fix subtype use.

Source/WebCore:

Not reviewed, trivial fix.

* inspector/InjectedScriptSource.js:
(.):

LayoutTests:

Not reviewed.

* inspector/console/console-format-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125186 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoWeb Inspector: DocumentEventQueue memory instrumentation
yurys@chromium.org [Thu, 9 Aug 2012 17:37:06 +0000 (17:37 +0000)]
Web Inspector: DocumentEventQueue memory instrumentation
https://bugs.webkit.org/show_bug.cgi?id=93604

Reviewed by Pavel Feldman.

Provided DocumentEventQueue:: reportMemoryUsage implementation.

* dom/Document.cpp:
(WebCore::Document::reportMemoryUsage):
* dom/DocumentEventQueue.cpp:
(WebCore::DocumentEventQueue::reportMemoryUsage):
(WebCore):
* dom/DocumentEventQueue.h:
(WebCore):
* dom/Event.cpp:
(WebCore::Event::reportMemoryUsage):
(WebCore):
* dom/Event.h:
(WebCore):
(Event):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125185 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoWeb Inspector: add codemirror library as an experiment
pfeldman@chromium.org [Thu, 9 Aug 2012 17:26:40 +0000 (17:26 +0000)]
Web Inspector: add codemirror library as an experiment
https://bugs.webkit.org/show_bug.cgi?id=93620

Reviewed by Vsevolod Vlasov.

Source/WebCore:

This change adds codemirror, deploys it, but does not include it into the inspector.html. Clients need to fetch it
using XHR and eval into the inspector's context to use as a part of their experiments.

* WebCore.gypi:
* WebCore.vcproj/WebCore.vcproj:
* inspector/front-end/WebKit.qrc:
* inspector/front-end/cm/LICENSE: Added.
* inspector/front-end/cm/codemirror.css: Added.
(.CodeMirror):
* inspector/front-end/cm/xml.js: Added.
(CodeMirror.defineMode.):
(CodeMirror.defineMode):

Source/WebKit/chromium:

* WebKit.gyp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125184 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago1.9.6 drops symbols, breaking compatibility
commit-queue@webkit.org [Thu, 9 Aug 2012 17:19:42 +0000 (17:19 +0000)]
1.9.6 drops symbols, breaking compatibility
https://bugs.webkit.org/show_bug.cgi?id=93477

Patch by Xan Lopez <xlopez@igalia.com> on 2012-08-09
Reviewed by Carlos Garcia Campos.

Expose documentURI as writable again for the GObject DOM
bindings. DOM4 makes it read-only, but we need to revert to the
old state for compatibility reasons, just like Objective C.

* dom/Document.idl: ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125183 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoLayout Test fast/forms/autocomplete-off-with-default-value-does-not-clear.html is...
beidson@apple.com [Thu, 9 Aug 2012 17:15:42 +0000 (17:15 +0000)]
Layout Test fast/forms/autocomplete-off-with-default-value-does-not-clear.html is failing (on Chromium)
https://bugs.webkit.org/show_bug.cgi?id=93586

Not reviewed (fixing a test failure).

This test is only valid for platforms with the page cache enabled.

* platform/chromium/TestExpectations: Skip fast/forms/autocomplete-off-with-default-value-does-not-clear.html

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125182 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoREGRESSION (r124997): Flaky crashes in two tests
mario@webkit.org [Thu, 9 Aug 2012 17:11:54 +0000 (17:11 +0000)]
REGRESSION (r124997): Flaky crashes in two tests
https://bugs.webkit.org/show_bug.cgi?id=93501

Reviewed by Martin Robinson.

Add missing null check to make sure we early return if the
accessibility object does not have a valid parent exposed in the
accessibility hierarchy.

* accessibility/gtk/AXObjectCacheAtk.cpp:
(WebCore::emitTextChanged): Add missing null check.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125181 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[Chromium] Unreviewed gardening: updated expectations for console-format-collections...
fmalita@chromium.org [Thu, 9 Aug 2012 17:10:02 +0000 (17:10 +0000)]
[Chromium] Unreviewed gardening: updated expectations for console-format-collections.html.

Patch by Florin Malita <fmalita@chromium.org> on 2012-08-09

* platform/chromium/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125180 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoAdd memory measurement to balls.html and dromaeorunner.js
zoltan@webkit.org [Thu, 9 Aug 2012 16:22:05 +0000 (16:22 +0000)]
Add memory measurement to balls.html and dromaeorunner.js
https://bugs.webkit.org/show_bug.cgi?id=93617

Reviewed by Adam Barth.

Since we added memory measurement support to performance tests we need to modify these tests.

* Animation/balls.html:
* Dromaeo/resources/dromaeorunner.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125179 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoJSHeap and FastMallocStatistics based memory measurement for performance-tests
zoltan@webkit.org [Thu, 9 Aug 2012 16:21:09 +0000 (16:21 +0000)]
JSHeap and FastMallocStatistics based memory measurement for performance-tests
https://bugs.webkit.org/show_bug.cgi?id=90858

PerformanceTests:

Reviewed Ryosuke Niwa.

Measure the memory usage of the performancetests with the help of the windows.memory.usedJSHeapSize
object and the window.internals.fastMallocStatistics() function call.

* resources/runner.js:
(PerfTestRunner.logStatistics):
(PerfTestRunner.printStatistics):
(PerfTestRunner.storeHeapResults):
(PerfTestRunner.getUsedFastMallocHeap):
(PerfTestRunner.getUsedJSHeap):
(PerfTestRunner.getAndPrintMemoryStatistics):
(PerfTestRunner.ignoreWarmUpAndLog):
(PerfTestRunner.initAndStartLoop):

Tools:

Reviewed by Ryosuke Niwa.

Measure the memory usage of the performancetests with the help of the windows.memory.usedJSHeapSize
object and the window.internals.fastMallocStatistics() function call.

* Scripts/webkitpy/performance_tests/perftest.py:
(PerfTest):
(PerfTest.parse_output):
(PerfTest.output_statistics):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125178 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[Chromium] Unreviewed gardening: updated Win results for 125160.
fmalita@chromium.org [Thu, 9 Aug 2012 16:11:04 +0000 (16:11 +0000)]
[Chromium] Unreviewed gardening: updated Win results for 125160.

Patch by Florin Malita <fmalita@chromium.org> on 2012-08-09

* platform/chromium-win/svg/custom/use-detach-expected.png:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125177 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago<rdar://problem/12035637> 32-bit PluginProcess launches without appropriate DYLD...
mitz@apple.com [Thu, 9 Aug 2012 16:10:45 +0000 (16:10 +0000)]
<rdar://problem/12035637> 32-bit PluginProcess launches without appropriate DYLD environment variables if UI process is 64-bit only
https://bugs.webkit.org/show_bug.cgi?id=93596

Reviewed by Mark Rowe.

* UIProcess/Launcher/mac/ProcessLauncherMac.mm:
(WebKit::ProcessLauncher::launchProcess): Changed to pass the current architecture, rather than
the intended child process architecture, to the DynamicLinkerEnvironmentExtractor constructor.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125176 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[Chromium] Unreviewed gardening: updated results for 125167.
fmalita@chromium.org [Thu, 9 Aug 2012 15:49:58 +0000 (15:49 +0000)]
[Chromium] Unreviewed gardening: updated results for 125167.

Patch by Florin Malita <fmalita@chromium.org> on 2012-08-09

* platform/chromium-linux/svg/custom/pattern-userSpaceOnUse-userToBaseTransform-expected.txt:
* platform/chromium-mac/svg/custom/pattern-userSpaceOnUse-userToBaseTransform-expected.txt:
* platform/chromium-win/svg/custom/pattern-userSpaceOnUse-userToBaseTransform-expected.txt:
* platform/chromium/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125175 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago2012-08-09 Pavel Feldman <pfeldman@chromium.org>
pfeldman@chromium.org [Thu, 9 Aug 2012 15:13:56 +0000 (15:13 +0000)]
2012-08-09  Pavel Feldman  <pfeldman@chromium.org>

        Web Inspector: generate preview for the objects dumped into the console upon logging.
        https://bugs.webkit.org/show_bug.cgi?id=35801

        Reviewed by Vsevolod Vlasov.

        As of today, dumping an object (array) into console will result in objects' properties being
        read upon console object expansion (i.e. lazily). This means that dumping the same object while
        mutating it will be hard to debug using the console.

        This change starts generating abbreviated previews for objects / arrays at the moment of their
        logging and passes this information along into the front-end. This only happens when the front-end
        is already opened, it only works for console.log(), not live console interaction.

        Covered by the existing console tests + collections test has been expanded to capture both: pre and post-

        * inspector/ConsoleMessage.cpp:
        (WebCore::ConsoleMessage::addToFrontend):
        * inspector/ConsoleMessage.h:
        (ConsoleMessage):
        * inspector/InjectedScript.cpp:
        (WebCore::InjectedScript::wrapObject):
        * inspector/InjectedScript.h:
        (InjectedScript):
        * inspector/InjectedScriptSource.js:
        (.):
        * inspector/Inspector.json:
        * inspector/InspectorConsoleAgent.cpp:
        (WebCore::InspectorConsoleAgent::enable):
        (WebCore::InspectorConsoleAgent::addConsoleMessage):
        * inspector/front-end/ConsoleMessage.js:
        (WebInspector.ConsoleMessageImpl.prototype._format):
        (WebInspector.ConsoleMessageImpl.prototype._formatParameter):
        (WebInspector.ConsoleMessageImpl.prototype._formatParameterAsObject):
        (WebInspector.ConsoleMessageImpl.prototype._formatParameterAsArrayOrObject):
        (WebInspector.ConsoleMessageImpl.prototype._appendObjectPreview):
        (WebInspector.ConsoleMessageImpl.prototype._formatParameterAsNode):
        (WebInspector.ConsoleMessageImpl.prototype._formatParameterAsArray):
        (WebInspector.ConsoleMessageImpl.prototype._printArray):
        (WebInspector.ConsoleMessageImpl.prototype._formatAsArrayEntry):
        (WebInspector.ConsoleMessageImpl.prototype._formatWithSubstitutionString):
        * inspector/front-end/DOMExtension.js:
        (Element.prototype.createTextChild):
        * inspector/front-end/ObjectPropertiesSection.js:
        (WebInspector.ObjectPropertiesSection):
        * inspector/front-end/PropertiesSection.js:
        * inspector/front-end/RemoteObject.js:
        (WebInspector.RemoteObject):
        (WebInspector.RemoteObject.fromPayload):
        (WebInspector.RemoteObject.prototype.get preview):
        * inspector/front-end/Section.js:
        * inspector/front-end/inspector.css:
        (.console-group-messages .section .header .title):
        (.section .console-formatted-node):
        (.console-object-preview):
        (.section .properties .name, .event-properties .name, .console-formatted-object .name):

2012-08-09  Pavel Feldman  <pfeldman@chromium.org>

        Web Inspector: generate preview for the objects dumped into the console upon logging.
        https://bugs.webkit.org/show_bug.cgi?id=35801

        Reviewed by Vsevolod Vlasov.

        * inspector/console/command-line-api-inspect-expected.txt:
        * inspector/console/console-dir-expected.txt:
        * inspector/console/console-eval-global-expected.txt:
        * inspector/console/console-eval-global.html:
        * inspector/console/console-format-collections-expected.txt:
        * inspector/console/console-format-collections.html:
        * inspector/console/console-format-expected.txt:
        * inspector/console/console-format.html:
        * inspector/console/console-log-document-proto-expected.txt:
        * inspector/console/console-log-document-proto.html:
        * inspector/console/console-log-toString-object-expected.txt:
        * inspector/debugger/watch-expressions-panel-switch-expected.txt:
        * inspector/elements/event-listener-sidebar-expected.txt:
        * inspector/elements/event-listeners-about-blank-expected.txt:
        * platform/chromium/inspector/console/console-dir-expected.txt:
        * platform/chromium/inspector/console/console-eval-global-expected.txt: Removed.
        * platform/chromium/inspector/console/console-format-collections-expected.txt:
        * platform/chromium/inspector/console/console-format-expected.txt: Removed.
        * platform/chromium/inspector/console/console-log-document-proto-expected.txt:
        * platform/chromium/inspector/debugger/watch-expressions-panel-switch-expected.txt:
        * platform/chromium/inspector/elements/event-listener-sidebar-expected.txt: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125174 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoUnreviewed, rolling out r125171.
peter@chromium.org [Thu, 9 Aug 2012 15:05:18 +0000 (15:05 +0000)]
Unreviewed, rolling out r125171.
http://trac.webkit.org/changeset/125171

Broke the Chromium Android build.

* DEPS:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125173 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[Qt] Unreviewed gardening.
kkristof@inf.u-szeged.hu [Thu, 9 Aug 2012 14:17:17 +0000 (14:17 +0000)]
[Qt] Unreviewed gardening.

* platform/qt-5.0/fast/js/global-constructors-expected.txt:
* platform/qt/Skipped:
* platform/qt/fast/dom/constructed-objects-prototypes-expected.txt:
* platform/qt/fast/events/keyevent-iframe-removed-crash-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125172 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoUnreviewed. Rolled DEPS.
peter@chromium.org [Thu, 9 Aug 2012 14:05:45 +0000 (14:05 +0000)]
Unreviewed.  Rolled DEPS.

* DEPS:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125171 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoUnreviewed, update TestExpectations with one more test to rebase after r125167
allan.jensen@nokia.com [Thu, 9 Aug 2012 14:03:31 +0000 (14:03 +0000)]
Unreviewed, update TestExpectations with one more test to rebase after r125167

* platform/chromium/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125170 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoPage popups can show up at wrong locations
keishi@webkit.org [Thu, 9 Aug 2012 12:43:21 +0000 (12:43 +0000)]
Page popups can show up at wrong locations
https://bugs.webkit.org/show_bug.cgi?id=93556

Reviewed by Kent Tamura.

Source/WebCore:

No new tests. Cannot test popup position.

* html/shadow/CalendarPickerElement.cpp:
(WebCore::CalendarPickerElement::contentSize):

Source/WebKit/chromium:

We were showing the popup at wrong positions. When there isn't enough
room below and above the element it adjusts the position but we weren't
resetting the adjustment when we resize the popup. This patch will make
the popup adjust the popup position each time we resize.

* src/ColorChooserUIController.cpp:
(WebKit::ColorChooserUIController::contentSize):
* src/WebPagePopupImpl.cpp:
(WevKit::PagePopupChromeClient::setWindowRect):
(WebKit::WebPagePopupImpl::init): Use reposition().
(WebKit::WebPagePopupImpl::reposition): Repositions the page popup based on the popup size.
(WebKit):
(WebKit::WebPagePopupImpl::resize): Use reposition().
* src/WebPagePopupImpl.h:
(WebPagePopupImpl):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125169 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoWeb Inspector: display progress bar while loading timeline data
caseq@chromium.org [Thu, 9 Aug 2012 12:24:47 +0000 (12:24 +0000)]
Web Inspector: display progress bar while loading timeline data
https://bugs.webkit.org/show_bug.cgi?id=92762

Reviewed by Pavel Feldman.

Source/WebCore:

- display progress while loading timeline data;
- disable status bar buttons while operation is in progress;
- ignore save/load/record while operation is in progress;
- disable context menu items while operation is in progress;
- misc. compilation warning fixes;

* inspector/front-end/ProgressBar.js:
(WebInspector.CompositeProgress.prototype.createSubProgress):
* inspector/front-end/TimelineModel.js:
(WebInspector.TimelineModel.prototype._loadNextChunk):
(WebInspector.TimelineModel.prototype.loadFromFile.parseAndImportData):
(WebInspector.TimelineModel.prototype.loadFromFile.onLoad):
(WebInspector.TimelineModel.prototype.loadFromFile.onError):
(WebInspector.TimelineModel.prototype.loadFromFile.onProgress):
(WebInspector.TimelineModel.prototype.loadFromFile):
* inspector/front-end/TimelinePanel.js:
(WebInspector.TimelinePanel):
(WebInspector.TimelinePanel.prototype.get statusBarItems):
(WebInspector.TimelinePanel.prototype._createStatusBarItems):
(WebInspector.TimelinePanel.prototype._setOperationInProgress):
(WebInspector.TimelinePanel.prototype._contextMenu):
(WebInspector.TimelinePanel.prototype._saveToFile):
(WebInspector.TimelinePanel.prototype._loadFromFile):
(WebInspector.TimelinePanel.prototype.get _toggleTimelineButtonClicked):
(WebInspector.TimelinePanel.prototype.sidebarResized):
* inspector/front-end/utilities.js: added Array.prototype.select(field) to map array of objects to array containing particular field of each object.

LayoutTests:

- fix loading tests to pass mock progress indicator;

* inspector/timeline/timeline-load-incompatible.html:
* inspector/timeline/timeline-load.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125168 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoFloored and truncated rounded confused.
allan.jensen@nokia.com [Thu, 9 Aug 2012 12:20:30 +0000 (12:20 +0000)]
Floored and truncated rounded confused.
https://bugs.webkit.org/show_bug.cgi?id=93475

Reviewed by Levi Weintraub.

Source/WebCore:

Fix the common mistake of confusing truncating (round towards zero) and
flooring (round down). They are only identical for positive numbers,
not for negative numbers.

This patch fixes instances of misimplemented or misused floor in the
layout and geometric code. And also uses the new correct implementations
to clean up the code for enclosingRect.

* platform/FractionalLayoutUnit.h:
(WebCore::FractionalLayoutUnit::fromFloatFloor):
(FractionalLayoutUnit):
(WebCore::FractionalLayoutUnit::ceil):
(WebCore::FractionalLayoutUnit::floor):
* platform/graphics/FloatPoint.h:
(WebCore::FloatPoint::FloatPoint):
(WebCore::roundedIntPoint):
(WebCore::flooredIntPoint):
(WebCore::ceiledIntPoint):
(WebCore::flooredIntSize):
* platform/graphics/FloatRect.cpp:
(WebCore::enclosingIntRect):
(WebCore::enclosedIntRect):
* platform/graphics/FloatSize.h:
(WebCore::roundedIntSize):
(WebCore::flooredIntSize):
(WebCore::flooredIntPoint):
* platform/graphics/FractionalLayoutPoint.h:
(WebCore::flooredIntPoint):
(WebCore::flooredFractionalLayoutPoint):
(WebCore::ceiledFractionalLayoutPoint):
* platform/graphics/FractionalLayoutRect.cpp:
(WebCore::enclosingIntRect):
(WebCore::enclosingFractionalLayoutRect):
* platform/graphics/FractionalLayoutSize.h:
(WebCore::flooredIntSize):
* platform/graphics/IntRect.cpp:
(WebCore::IntRect::IntRect):
* rendering/LayoutTypes.h:
(WebCore::flooredLayoutPoint):
(WebCore::floorToInt):
(WebCore::isIntegerValue):
* rendering/RenderLayer.cpp:
(WebCore::RenderLayer::convertToLayerCoords):

LayoutTests:

* platform/chromium/TestExpectations:
* platform/mac/TestExpectations:
* platform/qt/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125167 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoWeb Inspector: use WebInspector.ProgressIndicator in AdvancedSearchController
caseq@chromium.org [Thu, 9 Aug 2012 12:18:18 +0000 (12:18 +0000)]
Web Inspector: use WebInspector.ProgressIndicator in AdvancedSearchController
https://bugs.webkit.org/show_bug.cgi?id=93362

Reviewed by Vsevolod Vlasov.

- drop custom status bar used in AdvancedSearchController, use WebInspector.ProgressIndicator;
- perform additional cancelation check after invoking searchResultCallback to account for stopSearch called from within callback;
- fix compilation warnings in ProgressBar;

* inspector/front-end/AdvancedSearchController.js:
(WebInspector.SearchView):
(WebInspector.SearchView.prototype.searchStarted):
(WebInspector.SearchView.prototype.addSearchResult):
(WebInspector.SearchView.prototype.searchFinished):
(WebInspector.SearchView.prototype._onCancel):
* inspector/front-end/ProgressBar.js:
(WebInspector.CompositeProgress.prototype.createSubProgress):
* inspector/front-end/ScriptsSearchScope.js:
(WebInspector.ScriptsSearchScope.prototype.performSearch.searchCallbackWrapper):
* inspector/front-end/inspector.css:
(.progress-bar-container progress):
(body.platform-mac .progress-bar-container progress):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125166 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoWeb Inspector: improve large array logging experience
pfeldman@chromium.org [Thu, 9 Aug 2012 12:08:49 +0000 (12:08 +0000)]
Web Inspector: improve large array logging experience
https://bugs.webkit.org/show_bug.cgi?id=93499

Reviewed by Vsevolod Vlasov.

Source/WebCore:

Group elements in chunks of 100, not smaller.

* inspector/front-end/ObjectPropertiesSection.js:
(WebInspector.FunctionScopeMainTreeElement):
(WebInspector.FunctionScopeMainTreeElement.prototype.onpopulate):
(WebInspector.ArrayGroupingTreeElement._populateRanges):

LayoutTests:

* inspector/console/console-big-array-expected.txt:
* inspector/console/console-big-array.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125165 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[Chromium] Pull in the android_tools directory for Android
peter@chromium.org [Thu, 9 Aug 2012 11:34:52 +0000 (11:34 +0000)]
[Chromium] Pull in the android_tools directory for Android
https://bugs.webkit.org/show_bug.cgi?id=84843

Reviewed by Jochen Eisinger.

Pull in the android_tools directory for Chromium Android checkouts. This
repository, living in Chromium's, contains checked in versions of the
Android NDK and SDK meant to make builds and updates easier.

* DEPS:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125164 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[Qt][Win] Build system bug revealed by r124835
ossy@webkit.org [Thu, 9 Aug 2012 11:00:56 +0000 (11:00 +0000)]
[Qt][Win] Build system bug revealed by r124835
https://bugs.webkit.org/show_bug.cgi?id=93339

Reviewed by Tor Arne Vestbø.

Don't add "LC_ALL=c" on Windows in svnRevisionForDirectory() and in pathRelativeToSVNRepositoryRootForPath().

* Scripts/VCSUtils.pm:
(isWindows): Copied from webkitdirs.pm.
(svnRevisionForDirectory):
(pathRelativeToSVNRepositoryRootForPath):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125163 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoDOMCharacterDataModified should not be fired inside shadows
morrita@google.com [Thu, 9 Aug 2012 08:16:54 +0000 (08:16 +0000)]
DOMCharacterDataModified should not be fired inside shadows
https://bugs.webkit.org/show_bug.cgi?id=93427

Reviewed by Ryosuke Niwa.

Source/WebCore:

CharacterData::dispatchModifiedEvent() fires DOMCharacterDataModified event even if
the node is in shadow. But it shouldn't. Check dispatchChildInsertionEvents() and
dispatchChildRemovalEvents() to see how other MutationEvents are suppressed behind shadows.
This change follows the same path to suppress DOMCharacterDataModified.

Tests: fast/dom/shadow/suppress-mutation-events-in-shadow-characterdata.html
       fast/forms/textarea-and-mutation-events-appending-text.html

* dom/CharacterData.cpp:
(WebCore::CharacterData::dispatchModifiedEvent):

LayoutTests:

Added two tests. This also fixes existing test, which cover the wrong behavior.

* fast/dom/shadow/suppress-mutation-events-in-shadow-characterdata-expected.txt: Added.
* fast/dom/shadow/suppress-mutation-events-in-shadow-characterdata.html: Added.
* fast/events/dom-character-data-modified-textarea-crash.html:
- DOMCharacterDataModified shouldn't be fired in this case.
  Fixed to allow test to terminate without the event.
  Note that The event was fired by <textarea>, which is fixed by this change.
* fast/forms/textarea-and-mutation-events-appending-text-expected.txt: Added.
* fast/forms/textarea-and-mutation-events-appending-text.html: Added.
* svg/custom/tref-nested-events-crash.svg:
- DOMCharacterDataModified shouldn't be fired in this case.
  Fixed to allow test to terminate without the event.
  Note that The event was fired by SVGShadowText, which is just an implentation detail and
  should not be exposed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125162 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoUnreviewed, update TestExpectations for chromium gardening.
toyoshim@chromium.org [Thu, 9 Aug 2012 08:09:23 +0000 (08:09 +0000)]
Unreviewed, update TestExpectations for chromium gardening.
https://bugs.webkit.org/show_bug.cgi?id=93589

Mark svg/dom/SVGScriptElement/script-change-externalResourcesRequired-while-loading.svg as slow for Mac10.6 (dbg).

Patch by Takashi Toyoshima <toyoshim@chromium.org> on 2012-08-09

* platform/chromium/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125161 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoUse m_everHadLayout in RenderObject::checkForRepaintDuringLayout()
eric@webkit.org [Thu, 9 Aug 2012 07:58:50 +0000 (07:58 +0000)]
Use m_everHadLayout in RenderObject::checkForRepaintDuringLayout()
https://bugs.webkit.org/show_bug.cgi?id=20885

Reviewed by Dan Bernstein.

Source/WebCore:

Fix FIXME in checkForRepaintDuringLayout, making RenderObjects
no longer paint on first layout, but instead require their
parent containers to paint them.
This appears to have fixed one minor SVG overpainting issue
in svg/use-detach-expected.png.

* rendering/RenderBlock.cpp:
(WebCore::RenderBlock::layoutBlock):
* rendering/RenderObject.cpp:
(WebCore::RenderObject::checkForRepaintDuringLayout):
* rendering/svg/SVGRenderSupport.cpp:
(WebCore::SVGRenderSupport::layoutChildren):

LayoutTests:

Updated result.  Appears to be a progression.  I suspect
we were over-drawing the <g> container before and now we're
just drawing the two objects which change.

* platform/chromium-mac/svg/custom/use-detach-expected.png:
* platform/chromium-linux/svg/custom/use-detach-expected.png:
* platform/mac/svg/custom/use-detach-expected.png:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125160 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoMicrodata: PropertyNodeList cache should be invalidated on id attribute change.
arko@motorola.com [Thu, 9 Aug 2012 07:47:55 +0000 (07:47 +0000)]
Microdata: PropertyNodeList cache should be invalidated on id attribute change.
https://bugs.webkit.org/show_bug.cgi?id=93485

Reviewed by Ryosuke Niwa.

Source/WebCore:

Invalidate PropertyNodeList cache when id attribute of an element changed/modified.

Test: fast/dom/MicroData/propertiescollection-on-changing-id-attr.html

* dom/DynamicNodeList.cpp: Moved ownerNodeHasItemRefAttribute() method to DynamicNodeList.h
(WebCore):
* dom/DynamicNodeList.h:
(WebCore):
(WebCore::DynamicNodeListCacheBase::isRootedAtDocument): Return true if owner has itemref
attribute specified with root type NodeListIsRootedAtDocumentIfOwnerHasItemrefAttr.
(WebCore::DynamicNodeListCacheBase::shouldInvalidateTypeOnAttributeChange): Return true if
attrName is id attribute in case of InvalidateOnItemAttrChange.
(WebCore::DynamicNodeListCacheBase::ownerNodeHasItemRefAttribute):

LayoutTests:

Added test to verify HTMLPropertiesCollection and PropertyNodeList interfaces
behavior on changing id attribute.

* fast/dom/MicroData/propertiescollection-on-changing-id-attr-expected.txt: Added.
* fast/dom/MicroData/propertiescollection-on-changing-id-attr.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125159 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoUnreviewed, TestExpectations update for a newly added test
toyoshim@chromium.org [Thu, 9 Aug 2012 07:45:07 +0000 (07:45 +0000)]
Unreviewed, TestExpectations update for a newly added test
https://bugs.webkit.org/show_bug.cgi?id=93586

Patch by Takashi Toyoshima <toyoshim@chromium.org> on 2012-08-09

* platform/chromium/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125158 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoMicrodata: HTMLPropertiesCollection does not contain all properties when item is...
arko@motorola.com [Thu, 9 Aug 2012 07:28:14 +0000 (07:28 +0000)]
Microdata: HTMLPropertiesCollection does not contain all properties when item is not attached to the DOM tree.
https://bugs.webkit.org/show_bug.cgi?id=93456

Reviewed by Ryosuke Niwa.

Source/WebCore:

HTMLPropertiesCollection does not contain the item properties added through
itemref attribute when item is not attached to the DOM tree. We need to
find the root node of the item if its not attached in the document.

It is also possible that itemref attribute of item refers to the root node.
Made changes in getItemRefElements() method so that the traversal should
start from the root node.

Tests: fast/dom/MicroData/add-remove-property-to-item-which-is-not-attached-to-domtree.html
       fast/dom/MicroData/item-not-attached-to-domtree.html
       fast/dom/MicroData/itemref-pointing-to-root-node.html

* dom/DynamicNodeList.cpp:
(WebCore::DynamicNodeListCacheBase::rootNode):
* html/HTMLElement.cpp:
(WebCore::HTMLElement::getItemRefElements):

LayoutTests:

Added tests to ensure that HTMLPropertiesCollection contains all item properties
when item is not attached to the DOM tree.

* fast/dom/MicroData/add-remove-property-to-item-which-is-not-attached-to-domtree-expected.txt: Added.
* fast/dom/MicroData/add-remove-property-to-item-which-is-not-attached-to-domtree.html: Added.
* fast/dom/MicroData/item-not-attached-to-domtree-expected.txt: Added.
* fast/dom/MicroData/item-not-attached-to-domtree.html: Added.
* fast/dom/MicroData/itemref-pointing-to-root-node-expected.txt: Added.
* fast/dom/MicroData/itemref-pointing-to-root-node.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125157 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoAdd optional debug logging for tiled scrolling
timothy_horton@apple.com [Thu, 9 Aug 2012 07:26:35 +0000 (07:26 +0000)]
Add optional debug logging for tiled scrolling
https://bugs.webkit.org/show_bug.cgi?id=93305

Reviewed by Simon Fraser.

Add scrolling performance logging to ScrollingTreeNodeMac and WebTileLayer, which will log
the amount of unpainted area exposed during scrolling and painting, respectively.

The logging output looks like this:

    SCROLLING: Exposed tileless area. Time: 1344464850.154839 Unfilled Pixels: 1122008
    SCROLLING: Filled visible fresh tile. Time: 1344464850.575303 Unfilled Pixels: 1078272
    SCROLLING: Filled visible fresh tile. Time: 1344464850.676569 Unfilled Pixels: 823808
    SCROLLING: Filled visible fresh tile. Time: 1344464850.795877 Unfilled Pixels: 569344
    SCROLLING: Filled visible fresh tile. Time: 1344464850.846940 Unfilled Pixels: 524288
    SCROLLING: Filled visible fresh tile. Time: 1344464850.984113 Unfilled Pixels: 262144
    SCROLLING: Filled visible fresh tile. Time: 1344464851.027985 Unfilled Pixels: 0

No new tests, as this is just debugging logging.

* WebCore.exp.in:
* page/FrameView.cpp:
(WebCore::FrameView::setScrollingPerformanceLoggingEnabled):
Added setScrollingPerformanceLoggingEnabled, passes flag into TiledBacking.

* page/FrameView.h:
* page/Settings.cpp:
(WebCore::Settings::Settings):
(WebCore::Settings::setScrollingPerformanceLoggingEnabled):
Added setScrollingPerformanceLoggingEnabled, passes flag into FrameView.

* page/Settings.h:
(WebCore::Settings::scrollingPerformanceLoggingEnabled):
* page/scrolling/ScrollingTree.cpp:
(WebCore::ScrollingTree::ScrollingTree):
(WebCore::ScrollingTree::setScrollingPerformanceLoggingEnabled):
(WebCore::ScrollingTree::scrollingPeformanceLoggingEnabled):
Added ScrollingPerformanceLoggingEnabled setter/getter; this is where it's accessed from ScrollingTreeNodeMac.

* page/scrolling/ScrollingTree.h:
* page/scrolling/mac/ScrollingTreeNodeMac.h:
(ScrollingTreeNodeMac):
* page/scrolling/mac/ScrollingTreeNodeMac.mm:
(WebCore::ScrollingTreeNodeMac::setScrollPosition):
When updating scroll position, conditionally log the amount of unpainted area that was just exposed.

(WebCore::ScrollingTreeNodeMac::logExposedTilelessArea):
* platform/graphics/TiledBacking.h:
(WebCore::TiledBacking::TiledBacking):
(WebCore::TiledBacking::setScrollingPerformanceLoggingEnabled):
(WebCore::TiledBacking::scrollingPerformanceLoggingEnabled):
Added ScrollingPerformanceLoggingEnabled setter/getter; this is where it's accessed from WebTileLayer.

* platform/graphics/ca/mac/TileCache.h:
(WebCore::TileCache::visibleRect):
* platform/graphics/ca/mac/TileCache.mm:
(WebCore::TileCache::blankPixelCount):
(WebCore::TileCache::blankPixelCountForTiles):
Compute the exposed, unpainted area of the tile cache by iterating over the layers,
checking if they've been painted, and subtracting painted layers from the total visible area.

* platform/graphics/ca/mac/WebTileLayer.h:
* platform/graphics/ca/mac/WebTileLayer.mm:
(-[WebTileLayer drawInContext:]):
(-[WebTileLayer repaintCount]):
(-[WebTileLayer logFilledFreshTile]):
When painting a WebTileLayer, conditionally log the amount of unpainted area remaining after painting said tile.

* rendering/RenderLayerBacking.cpp:
(WebCore::RenderLayerBacking::RenderLayerBacking):
Plumb ScrollingPerformanceLoggingEnabled through to TiledBacking.

Add a new boolean preference, ScrollingPerformanceLoggingEnabled, which will be used to toggle
the WebCore parts of this change.

* Shared/WebPreferencesStore.h: Add ScrollingPerformanceLoggingEnabled.
* UIProcess/API/C/WKPreferences.cpp:
(WKPreferencesSetScrollingPerformanceLoggingEnabled): Added.
(WKPreferencesGetScrollingPerformanceLoggingEnabled): Added.
* UIProcess/API/C/WKPreferencesPrivate.h: Add getters and setters for ScrollingPerformanceLoggingEnabled.
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::WebPage): Add ScrollingPerformanceLoggingEnabled.
(WebKit::WebPage::updatePreferences): Add ScrollingPerformanceLoggingEnabled.
(WebKit::WebPage::setScrollingPerformanceLoggingEnabled): Propagate settings changes through to FrameView.
* WebProcess/WebPage/WebPage.h:
(WebKit::WebPage::scrollingPerformanceLoggingEnabled): Add a getter for the setting so that TiledCoreAnimationDrawingArea can access it.
* WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.mm:
(WebKit::TiledCoreAnimationDrawingArea::updatePreferences): Pass ScrollingPerformanceLoggingEnabled down into the scrolling thread.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125156 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoSource/WebCore: Remove image decoding in some BitmapImage metadata functions
hclam@chromium.org [Thu, 9 Aug 2012 06:23:42 +0000 (06:23 +0000)]
Source/WebCore: Remove image decoding in some BitmapImage metadata functions
https://bugs.webkit.org/show_bug.cgi?id=93171

Reviewed by Simon Fraser.

These two metadata functions were decoding an entire frame:
- frameOrientationAtIndex
- frameHasAlphaAtIndex

This change removes image decoding from these two methods. This is for
preparation of having asynchronous image decoding, intending to reduce
code location that trigger image decoding.

frameOrientationAtIndex() doesn't require decoding a frame. This method
is only implemented in CG port in ImageSourceCG.cpp which doesn't do
image decoding.

frameHasAlphaAtIndex() is used to optimize certain drawing operations
and accelerated compositing. This change uses a heuristic for non-CG
port to determine if an image has alpha. If an image is not yet
decoded the function answers having alpha. Only if a frame is decoded
and cached that the alpha state of the frame is returned. This is an
admissible heuristic that postpone answering the question until a frame
is decoded.

Tested this change with a fully loaded image and partially loaded image
with background color.

Test: http/tests/images/jpg-img-partial-load.html

* platform/graphics/BitmapImage.cpp:
(WebCore::BitmapImage::frameHasAlphaAtIndex):
(WebCore::BitmapImage::frameOrientationAtIndex):
* platform/graphics/ImageSource.cpp:
(WebCore::ImageSource::frameHasAlphaAtIndex):
* platform/image-decoders/ImageDecoder.cpp:
(WebCore::ImageDecoder::frameHasAlphaAtIndex):
(WebCore):
* platform/image-decoders/ImageDecoder.h:
(ImageDecoder):

LayoutTests: Test rendering JPEG <img> with CSS background
https://bugs.webkit.org/show_bug.cgi?id=93171

Reviewed by Simon Fraser.

Tests defer answering the alpha state for BitmapImage not yet
decoded doesn't affect rendering results.

Test: http/tests/images/jpg-img-partial-load.html

Tests the rendering results of a partially loaded JPEG image.
The image should be green in the top region and blue in the
lower region. The blue is the background color revealed by
full alpha of the undecoded region.

* fast/images/resources/green-256x256.jpg: Added.
* http/tests/images/jpg-img-partial-load-expected.png: Added.
* http/tests/images/jpg-img-partial-load-expected.txt: Added.
* http/tests/images/jpg-img-partial-load.html: Added.
* http/tests/resources/load-and-stall.php: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125154 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoupdate bot parsing of new-run-webkit-tests' output
ossy@webkit.org [Thu, 9 Aug 2012 05:50:24 +0000 (05:50 +0000)]
update bot parsing of new-run-webkit-tests' output
https://bugs.webkit.org/show_bug.cgi?id=93537

Patch by Dirk Pranke <dpranke@chromium.org> on 2012-08-08
Reviewed by Ryosuke Niwa.

This change updates the way we parse the output from
new-run-webkit-tests on the bots so that we can show a summary
of the results properly after I re-land the changes that were
reverted in r124994 and r124870.

* BuildSlaveSupport/build.webkit.org-config/master.cfg:
(RunWebKitTests._parseNewRunWebKitTestsOutput):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125153 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoUnreviewed, rolling out r125146.
commit-queue@webkit.org [Thu, 9 Aug 2012 05:29:22 +0000 (05:29 +0000)]
Unreviewed, rolling out r125146.
http://trac.webkit.org/changeset/125146
https://bugs.webkit.org/show_bug.cgi?id=93578

Would like to study fast/dom/Window/timer-null-script-
execution-context.html behavior change (Requested by abarth on
#webkit).

Patch by Sheriff Bot <webkit.review.bot@gmail.com> on 2012-08-08

* bindings/js/JSDOMWindowBase.cpp:
(WebCore):
(WebCore::JSDOMWindowBase::crossDomainAccessErrorMessage):
(WebCore::JSDOMWindowBase::allowsAccessFrom):
* bindings/js/JSDOMWindowBase.h:
(JSDOMWindowBase):
* bindings/js/JSDOMWindowCustom.cpp:
(WebCore::namedItemGetter):
(WebCore::JSDOMWindow::getOwnPropertySlot):
(WebCore::JSDOMWindow::getOwnPropertyDescriptor):
(WebCore::JSDOMWindow::put):
(WebCore::JSDOMWindow::deleteProperty):
(WebCore::JSDOMWindow::getPropertyNames):
(WebCore::JSDOMWindow::getOwnPropertyNames):
(WebCore::JSDOMWindow::defineOwnProperty):
(WebCore::JSDOMWindow::setLocation):
* bindings/js/JSDOMWindowCustom.h:
(WebCore::JSDOMWindowBase::allowsAccessFrom):
(WebCore):
(WebCore::JSDOMWindowBase::allowsAccessFromNoErrorMessage):
(WebCore::JSDOMWindowBase::allowsAccessFromPrivate):
* bindings/js/JSInjectedScriptManager.cpp:
(WebCore::InjectedScriptManager::canAccessInspectedWindow):
* bindings/objc/WebScriptObject.mm:
(-[WebScriptObject _isSafeScript]):
* bindings/scripts/CodeGeneratorJS.pm:
(GenerateGetOwnPropertyDescriptorBody):
(GenerateImplementation):
* bindings/scripts/test/JS/JSTestActiveDOMObject.cpp:
(WebCore::jsTestActiveDOMObjectExcitingAttr):
(WebCore::jsTestActiveDOMObjectConstructor):
(WebCore::jsTestActiveDOMObjectPrototypeFunctionExcitingFunction):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125152 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoSkip CoreAnimation plugin model tests on android
commit-queue@webkit.org [Thu, 9 Aug 2012 05:27:30 +0000 (05:27 +0000)]
Skip CoreAnimation plugin model tests on android
https://bugs.webkit.org/show_bug.cgi?id=90586

Patch by Xingnan Wang <xingnan.wang@intel.com> on 2012-08-08
Reviewed by Adam Barth.

* platform/chromium/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125151 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoSkip platform/mac/editing/deleting/backward-delete.html on android
commit-queue@webkit.org [Thu, 9 Aug 2012 05:21:27 +0000 (05:21 +0000)]
Skip platform/mac/editing/deleting/backward-delete.html on android
https://bugs.webkit.org/show_bug.cgi?id=90585

Patch by Xingnan Wang <xingnan.wang@intel.com> on 2012-08-08
Reviewed by Adam Barth.

* platform/chromium/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125150 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoUnprefix window.webkitURL
kinuko@chromium.org [Thu, 9 Aug 2012 04:49:51 +0000 (04:49 +0000)]
Unprefix window.webkitURL
https://bugs.webkit.org/show_bug.cgi?id=93034

Reviewed by Jian Li.

Adding unprefixed window.URL as other browsers like FireFox and Opera 12 seem to have the API unprefixed.
(This patch doesn't remove webkitURL yet as Safari and Chrome has exposed the URL for a while now.)

Source/WebCore:

Test: fast/dom/DOMURL/check-instanceof-domurl-functions.html

* page/DOMWindow.cpp:
* page/DOMWindow.h:
(DOMWindow):
* page/DOMWindow.idl:
* workers/WorkerContext.idl:

LayoutTests:

Changed most of the window.URL test code to use unprefixed version.
To verify webkitURL is still available left webkitURL related tests in
fast/dom/DOMURL/check-instanceof-domurl-functions.html.

* editing/pasteboard/data-transfer-items-image-png.html:
* fast/dom/DOMURL/check-instanceof-domurl-functions-expected.txt:
* fast/dom/DOMURL/check-instanceof-domurl-functions.html:
* fast/dom/HTMLAnchorElement/anchor-download-unset.html:
* fast/dom/HTMLAnchorElement/anchor-download.html:
* fast/dom/HTMLAnchorElement/anchor-nodownload-set.html:
* fast/dom/HTMLAnchorElement/anchor-nodownload.html:
* fast/dom/constructed-objects-prototypes-expected.txt:
* fast/dom/window-domurl-crash.html:
* fast/files/apply-blob-url-to-img.html:
* fast/files/apply-blob-url-to-xhr.html:
* fast/files/create-blob-url-crash.html:
* fast/files/domurl-script-execution-context-crash.html:
* fast/files/revoke-blob-url.html:
* fast/files/url-null-expected.txt:
* fast/files/url-null.html:
* fast/files/url-required-arguments-expected.txt:
* fast/files/url-required-arguments.html:
* fast/files/workers/inline-worker-via-blob-url.html:
* fast/files/workers/resources/worker-apply-blob-url-to-xhr.js:
* fast/js/constructor-length.html:
* http/tests/fileapi/create-blob-url-from-data-url.html:
* http/tests/security/mixedContent/blob-url-in-iframe.html:
* platform/chromium-win/compositing/geometry/object-clip-rects-assertion-expected.txt:
* platform/chromium/fast/files/url-required-arguments-expected.txt:
* platform/chromium/media/video-capture-preview.html:
* platform/gtk/fast/dom/constructed-objects-prototypes-expected.txt:
* platform/gtk/fast/files/url-required-arguments-expected.txt:
* platform/gtk/fast/js/constructor-length-expected.txt:
* platform/gtk/fast/js/global-constructors-expected.txt:
* platform/mac/fast/js/constructor-length-expected.txt:
* platform/mac/fast/js/global-constructors-expected.txt:
* platform/qt-5.0/fast/js/global-constructors-expected.txt:
* platform/qt/fast/dom/constructed-objects-prototypes-expected.txt:
* platform/qt/fast/js/constructor-length-expected.txt:
* platform/qt/fast/js/global-constructors-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125149 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agonested fragment parser test slow on Android
commit-queue@webkit.org [Thu, 9 Aug 2012 04:26:44 +0000 (04:26 +0000)]
nested fragment parser test slow on Android
https://bugs.webkit.org/show_bug.cgi?id=90438

Patch by Wei James <james.wei@intel.com> on 2012-08-08
Reviewed by Ryosuke Niwa.

* platform/chromium/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125148 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[SVG] load events shouldn't be fired during Node::insrtedInto()
morrita@google.com [Thu, 9 Aug 2012 04:16:38 +0000 (04:16 +0000)]
[SVG] load events shouldn't be fired during Node::insrtedInto()
https://bugs.webkit.org/show_bug.cgi?id=92969

Reviewed by Ryosuke Niwa.

Source/WebCore:

Event dispatches during insertedInto() allow event handlers to
break DOM tree cosistency. This chagne makes them async for load
events which are dispatched during insertedInto() call. This
prevents event handlers from breaking tree consistency while the
notification traversal.

Test: svg/custom/loadevents-async.html

* svg/SVGElement.cpp:
(WebCore::SVGElement::sendSVGLoadEventIfPossibleAsynchronously): Added.
(WebCore):
(WebCore::SVGElement::svgLoadEventTimerFired): Added.
(WebCore::SVGElement::svgLoadEventTimer):
- Added a stub. Implemented in SVGScriptElement, SVGStopElement, SVGUseElement
  where the load event happens.
* svg/SVGElement.h:
(SVGElement):
* svg/SVGExternalResourcesRequired.cpp:
(WebCore::SVGExternalResourcesRequired::insertedIntoDocument):
- Replaces event dispatch call with async version.
* svg/SVGScriptElement.h:
* svg/SVGStyleElement.h:
* svg/SVGUseElement.h:

LayoutTests:

* svg/custom/loadevents-async-expected.txt: Added.
* svg/custom/loadevents-async.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125147 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoImplement JSDOMWindow*::allowsAccessFrom* in terms of BindingSecurity
abarth@webkit.org [Thu, 9 Aug 2012 03:52:41 +0000 (03:52 +0000)]
Implement JSDOMWindow*::allowsAccessFrom* in terms of BindingSecurity
https://bugs.webkit.org/show_bug.cgi?id=93407

Reviewed by Eric Seidel.

This patch removes allowsAccessFrom and implements the security checks
in terms of shouldAllowAccessToFrame directly. There shouldn't be any
change in behavior.

* bindings/js/JSDOMWindowBase.cpp:
(WebCore):
(WebCore::shouldAllowAccessFrom):
* bindings/js/JSDOMWindowBase.h:
(JSDOMWindowBase):
* bindings/js/JSDOMWindowCustom.cpp:
(WebCore::namedItemGetter):
(WebCore::JSDOMWindow::getOwnPropertySlot):
(WebCore::JSDOMWindow::getOwnPropertyDescriptor):
(WebCore::JSDOMWindow::put):
(WebCore::JSDOMWindow::deleteProperty):
(WebCore::JSDOMWindow::getPropertyNames):
(WebCore::JSDOMWindow::getOwnPropertyNames):
(WebCore::JSDOMWindow::defineOwnProperty):
(WebCore::JSDOMWindow::setLocation):
* bindings/js/JSDOMWindowCustom.h:
* bindings/js/JSInjectedScriptManager.cpp:
(WebCore::InjectedScriptManager::canAccessInspectedWindow):
* bindings/objc/WebScriptObject.mm:
(-[WebScriptObject _isSafeScript]):
* bindings/scripts/CodeGeneratorJS.pm:
(GenerateGetOwnPropertyDescriptorBody):
(GenerateImplementation):
* bindings/scripts/test/JS/JSTestActiveDOMObject.cpp:
(WebCore::jsTestActiveDOMObjectExcitingAttr):
(WebCore::jsTestActiveDOMObjectConstructor):
(WebCore::jsTestActiveDOMObjectPrototypeFunctionExcitingFunction):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125146 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoUnreviewed, rolling out r125143.
tony@chromium.org [Thu, 9 Aug 2012 03:51:20 +0000 (03:51 +0000)]
Unreviewed, rolling out r125143.
http://trac.webkit.org/changeset/125143
https://bugs.webkit.org/show_bug.cgi?id=93573

Doesn't compile on Mac (Requested by tony^work on #webkit).

Patch by Sheriff Bot <webkit.review.bot@gmail.com> on 2012-08-08

* GNUmakefile.list.am:
* Target.pri:
* WebCore.gypi:
* WebCore.vcproj/WebCore.vcproj:
* WebCore.xcodeproj/project.pbxproj:
* platform/graphics/BitmapImage.h:
* platform/graphics/Image.h:
* platform/graphics/ImageSource.h:
(WebCore):
* platform/graphics/NativeImagePtr.h: Removed.
* platform/graphics/cairo/GraphicsContext3DCairo.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125145 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoExtract the definition of NativeImagePtr from ImageSource.h into NativeImagePtr.h.
kseo@webkit.org [Thu, 9 Aug 2012 02:35:43 +0000 (02:35 +0000)]
Extract the definition of NativeImagePtr from ImageSource.h into NativeImagePtr.h.
https://bugs.webkit.org/show_bug.cgi?id=93466

Patch by Huang Dongsung <luxtella@company100.net> on 2012-08-08
Reviewed by Adam Barth.

Image.h includes ImageSource.h in order to know the definition of NativeImagePtr.
But Image does not need to know the definition of ImageSource. So, this patch
extracts the definition of NativeImagePtr from ImageSource.h into NativeImagePtr.h.

This is a preparation for asynchronous image decoding. Async decoding
will need a callback function, which sends a NativeImagePtr to clients, and the
callback must know the definition of NativeImagePtr, not ImageSource.

No new tests, behavior is unchanged.

* GNUmakefile.list.am:
* Target.pri:
* WebCore.gypi:
* WebCore.vcproj/WebCore.vcproj:
* WebCore.xcodeproj/project.pbxproj:
* platform/graphics/BitmapImage.h:
* platform/graphics/Image.h:
* platform/graphics/ImageSource.h:
(WebCore):
* platform/graphics/NativeImagePtr.h: Added.
(WebCore):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125143 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoCompile flag for CSS Hierarchies
commit-queue@webkit.org [Thu, 9 Aug 2012 02:29:35 +0000 (02:29 +0000)]
Compile flag for CSS Hierarchies
https://bugs.webkit.org/show_bug.cgi?id=92433

Patch by Shane Stephens <shanestephens@google.com> on 2012-08-08
Reviewed by Tony Chang.

.:

* Source/cmake/WebKitFeatures.cmake:

Source/JavaScriptCore:

* Configurations/FeatureDefines.xcconfig:

Source/WebCore:

Note that this flag is disabled on all platforms for now.

No new tests since there's no functional change.

* Configurations/FeatureDefines.xcconfig:

Source/WebKit/chromium:

* features.gypi:

Source/WebKit/mac:

* Configurations/FeatureDefines.xcconfig:

Source/WebKit2:

* Configurations/FeatureDefines.xcconfig:

Tools:

* Scripts/webkitperl/FeatureList.pm:
* qmake/mkspecs/features/features.pri:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125142 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[Chromium-mac] Switch to LocalizedNumberMac from LocalizedNumberICU
tkent@chromium.org [Thu, 9 Aug 2012 02:26:01 +0000 (02:26 +0000)]
[Chromium-mac] Switch to LocalizedNumberMac from LocalizedNumberICU
https://bugs.webkit.org/show_bug.cgi?id=93432

Reviewed by Hajime Morita.

Source/WebCore:

No new tests. Covered by fast/forms/number/number-lossless-localization.html,
and WebKit/chromium/tests/LocaleMacTest.cpp.

* WebCore.gyp/WebCore.gyp:
Exclude locale-relatef files for ICU, and include files for Mac.

Source/WebKit/chromium:

* tests/LocaleMacTest.cpp:
Add some tests for number localization. They are similar to tests in
LocaleWinTest and LocalizedNumberICUTest.cpp
(LocaleMacTest):
(LocaleMacTest::decimalSeparator):
(TEST_F):
(testNumberIsReversible):
(testNumbers):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125141 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agocss3/flexbox/content-height-with-scrollbars.html failing on non-fractional pixel...
tony@chromium.org [Thu, 9 Aug 2012 02:22:52 +0000 (02:22 +0000)]
css3/flexbox/content-height-with-scrollbars.html failing on non-fractional pixel layout machines
https://bugs.webkit.org/show_bug.cgi?id=93530

Unreviewed, fixing the test to not depend on rounding behavior.

* css3/flexbox/content-height-with-scrollbars-expected.html: Use a fixed width.
* css3/flexbox/content-height-with-scrollbars.html: Use a fixed width.
* platform/efl/TestExpectations: Remove failing expectation.
* platform/gtk/TestExpectations: Remove failing expectation.
* platform/mac/TestExpectations: Remove failing expectation.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125140 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoUnreviewed TestExpectations cleanup and a new baseline.
dpranke@chromium.org [Thu, 9 Aug 2012 02:09:30 +0000 (02:09 +0000)]
Unreviewed TestExpectations cleanup and a new baseline.

* platform/chromium/TestExpectations:
* platform/chromium/http/tests/security/sandboxed-iframe-modify-self-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125139 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoFix invalid test
arv@chromium.org [Thu, 9 Aug 2012 02:05:36 +0000 (02:05 +0000)]
Fix invalid test
https://bugs.webkit.org/show_bug.cgi?id=93543

Reviewed by Kentaro Hara.

The HTMLMediaElement.prototype.NETWORK_NO_SOURCE is non writable so assigning to the element instance should
be ignored.

* media/network-no-source-const-shadow-expected.txt:
* media/network-no-source-const-shadow.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125138 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoUse char* instead of LChar* for the public interface of String construction from...
benjamin@webkit.org [Thu, 9 Aug 2012 01:29:44 +0000 (01:29 +0000)]
Use char* instead of LChar* for the public interface of String construction from literals
https://bugs.webkit.org/show_bug.cgi?id=93402

Patch by Benjamin Poulain <bpoulain@apple.com> on 2012-08-08
Reviewed by Michael Saboff.

Source/JavaScriptCore:

Update JSC' Identifier to use StringImpl::createFromLiteral with a char*.

* runtime/Identifier.cpp:
(JSC::IdentifierASCIIStringTranslator::translate):

Source/WTF:

When the initialization from literal was added, some constructor/initialization function were
using LChar for convenience.

Since those function should only take ASCII characters, using LChar* could cause confusion. This
patch intents to clarify this by using char* for all those APIs.

* wtf/text/AtomicString.cpp:
(WTF::CharBufferFromLiteralDataTranslator::hash):
(WTF::CharBufferFromLiteralDataTranslator::equal):
(WTF::CharBufferFromLiteralDataTranslator::translate):
(WTF::AtomicString::addFromLiteralData):
* wtf/text/AtomicString.h:
(WTF::AtomicString::AtomicString):
(AtomicString):
* wtf/text/StringImpl.cpp:
(WTF::StringImpl::createFromLiteral):
* wtf/text/StringImpl.h:
(StringImpl):
(WTF::StringImpl::StringImpl):
(WTF::StringImpl::createFromLiteral):

Tools:

* TestWebKitAPI/Tests/WTF/StringImpl.cpp:
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125135 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoEventDispatcher::dispatchSimulatedClick should not reuse the same EventDispatcher...
hayato@chromium.org [Thu, 9 Aug 2012 01:11:35 +0000 (01:11 +0000)]
EventDispatcher::dispatchSimulatedClick should not reuse the same EventDispatcher instance.
https://bugs.webkit.org/show_bug.cgi?id=93452

Reviewed by Dimitri Glazkov.

This is a follow up patch after r124975.  There is yet another
place where the same EventDispatcher instance is reused across
event dispatching.

In addition to that, in order to simulate a real click, we must
different dispatchers because we must recalculate event ancestors
each time.  Current implementation wrongly freezes event ancestors
at the first event dispatching.

No new tests, no change in functionality.

* dom/EventDispatcher.cpp:
(WebCore::EventDispatcher::dispatchSimulatedClick):
(WebCore::EventDispatcher::EventDispatcher):
(WebCore::EventDispatcher::dispatchEvent):
* dom/EventDispatcher.h:
(EventDispatcher):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125133 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agowriteHelperGetString in V8HTMLDocument uses += to concatenate strings
commit-queue@webkit.org [Thu, 9 Aug 2012 01:10:01 +0000 (01:10 +0000)]
writeHelperGetString in V8HTMLDocument uses += to concatenate strings
https://bugs.webkit.org/show_bug.cgi?id=93035

Patch by Jae Hyun Park <jae.park@company100.net> on 2012-08-08
Reviewed by Adam Barth.

Use a StringBuilder instead of String concatenation because StringBuilder is generally faster.

No new functionality, so no new tests.

* bindings/v8/custom/V8HTMLDocumentCustom.cpp:
(WebCore::writeHelperGetString):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125132 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agonrwt: rename finder.py to layout_test_finder.py
dpranke@chromium.org [Thu, 9 Aug 2012 01:07:24 +0000 (01:07 +0000)]
nrwt: rename finder.py to layout_test_finder.py
https://bugs.webkit.org/show_bug.cgi?id=93557

Reviewed by Ryosuke Niwa.

One last cleanup per old review feedback from rniwa.

* Scripts/webkitpy/layout_tests/controllers/layout_test_finder.py: Renamed from Tools/Scripts/webkitpy/layout_tests/controllers/finder.py.
(LayoutTestFinder):
(LayoutTestFinder.__init__):
(LayoutTestFinder.find_tests):
(LayoutTestFinder._strip_test_dir_prefixes):
(LayoutTestFinder._strip_test_dir_prefix):
(LayoutTestFinder._read_test_names_from_file):
(LayoutTestFinder._strip_comments):
(LayoutTestFinder.skip_tests):
(LayoutTestFinder.split_into_chunks):
* Scripts/webkitpy/layout_tests/controllers/manager.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125131 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[chromium] Move scrollbar pointer into WebScrollbarThemePainter
enne@google.com [Thu, 9 Aug 2012 01:04:31 +0000 (01:04 +0000)]
[chromium] Move scrollbar pointer into WebScrollbarThemePainter
https://bugs.webkit.org/show_bug.cgi?id=93541

Reviewed by James Robinson.

Source/Platform:

Update WebScrollbarThemePainter to encapsulate a pointer to the
scrollbar as well. Lion scrollbars and RenderScrollbars are both
incompatible with wrapping a scrollbar in a WebScrollbar.

ScrollbarTheme(Chromium)Mac attaches extra data to a scrollbar based
on the pointer value (via a static map keyed on the pointer) so
passing an object that returns all the same values for the
ScrollbarThemeClient interface but has a different pointer will fail
to paint.

RenderScrollbar does static casts on the ScrollbarThemeClient pointer
that it is passed, assuming that it is the same. Therefore, it also
cannot use a WebScrollbar.

To fix this, push the real scrollbar pointer into the painter.

* chromium/public/WebScrollbarThemePainter.h:
(WebCore):
(WebKit::WebScrollbarThemePainter::WebScrollbarThemePainter):
(WebScrollbarThemePainter):

Source/WebCore:

Update clients of WebScrollbarThemePainter.

* page/scrolling/chromium/ScrollingCoordinatorChromium.cpp:
(WebCore::createScrollbarLayer):
* platform/graphics/chromium/ScrollbarLayerChromium.cpp:

Source/WebKit/chromium:

Fix WebScrollbarThemePainter implementation to make direct calls on
the scrollbar rather than through the WebScrollbarThemeClientImpl
adapter.

* src/WebScrollbarThemePainter.cpp:
(WebKit::WebScrollbarThemePainter::assign):
(WebKit::WebScrollbarThemePainter::paintScrollbarBackground):
(WebKit::WebScrollbarThemePainter::paintTrackBackground):
(WebKit::WebScrollbarThemePainter::paintBackTrackPart):
(WebKit::WebScrollbarThemePainter::paintForwardTrackPart):
(WebKit::WebScrollbarThemePainter::paintBackButtonStart):
(WebKit::WebScrollbarThemePainter::paintBackButtonEnd):
(WebKit::WebScrollbarThemePainter::paintForwardButtonStart):
(WebKit::WebScrollbarThemePainter::paintForwardButtonEnd):
(WebKit::WebScrollbarThemePainter::paintTickmarks):
(WebKit::WebScrollbarThemePainter::paintThumb):
(WebKit::WebScrollbarThemePainter::WebScrollbarThemePainter):
* tests/ScrollbarLayerChromiumTest.cpp:
(WebCore::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125130 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[CSS Shaders] Invalid shaders should act as pass-through filters
achicu@adobe.com [Thu, 9 Aug 2012 00:56:13 +0000 (00:56 +0000)]
[CSS Shaders] Invalid shaders should act as pass-through filters
https://bugs.webkit.org/show_bug.cgi?id=93405

Reviewed by Dean Jackson.

Source/WebCore:

If the shader fails to apply then clearShaderResult will just copy the result of the previous filter
to the output of the current filter.

Test: css3/filters/custom/invalid-custom-filter-shader.html

* platform/graphics/filters/FECustomFilter.cpp:
(WebCore::FECustomFilter::platformApplySoftware):
(WebCore):
(WebCore::FECustomFilter::clearShaderResult):
(WebCore::FECustomFilter::applyShader):
* platform/graphics/filters/FECustomFilter.h:
(FECustomFilter):
* rendering/style/StyleCustomFilterProgram.h: The test was exposing the fact that if the shaders were referencing the same
file then StyleCustomFilterProgram will never complete the load. Having the same file for both the vertex and the fragment shader
cannot really work, because the shaders would not compile anyway, thus triggering an invalid shader.
I'm fixing it part of this change because the current test actually exposes that on Safari Mac builds. See the note in the LayoutTest/ChangeLog.
(WebCore::StyleCustomFilterProgram::notifyFinished):

LayoutTests:

Note that invalid-custom-filter-shader-expected.html is actually using another builtin filter,
just to make sure we trigger the whole FilterEffectRenderer on the images, otherwise the result will be
slightly different because of the conversion errors. Having another filter like grayscale(0) will trigger
the same precission errors on the result of the expected html.

* css3/filters/custom/invalid-custom-filter-shader-expected.html: Added.
* css3/filters/custom/invalid-custom-filter-shader.html: Added.
* css3/filters/resources/invalid-shader.vs: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125128 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoRewire the same-origin checks for the JavaScriptCore bindings through BindingSecurity
abarth@webkit.org [Thu, 9 Aug 2012 00:41:02 +0000 (00:41 +0000)]
Rewire the same-origin checks for the JavaScriptCore bindings through BindingSecurity
https://bugs.webkit.org/show_bug.cgi?id=93382

Reviewed by Eric Seidel.

This patch rewires the same-origin policy checks in the JavaScriptCore
bindings to use the implementation in BindingSecurity.cpp, which is now
shared by JavaScriptCore and V8. There are still a few places were we
use the JSDOMWindowCustom-based code path, but I plan to change those
in a follow up patch in the interest of keeping this patch as small as
possible.

This patch as two main benefits:

1) We no longer need to maintain duplicate code in the JSC and the V8
   bindings for as delicate an area as the same-origin check.
   Previously, the two implementations accomplished the same task using
   a slightly different mechansim. After this patch, they use the same
   mechanism, which means we only need to convince ourselves that one
   implementation is correct.

2) This patch will make it easier to remove DOMWindow::m_securityOrigin
   because there will be only one piece of code that needs to change.
   Prior to this patch, we would have had to change both
   implementations of the same-origin policy not to rely upon
   DOMWindow::m_securityOrigin.

* bindings/js/BindingState.cpp:
(WebCore::immediatelyReportUnsafeAccessTo):
(WebCore):
* bindings/js/BindingState.h:
(WebCore):
* bindings/js/JSDOMBinding.cpp:
(WebCore::shouldAllowAccessToNode):
(WebCore::shouldAllowAccessToFrame):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125126 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[chromium/mac] Map NSEventPhaseMayBegin appropriately on 10.8
commit-queue@webkit.org [Thu, 9 Aug 2012 00:40:37 +0000 (00:40 +0000)]
[chromium/mac] Map NSEventPhaseMayBegin appropriately on 10.8
https://bugs.webkit.org/show_bug.cgi?id=93535

Patch by Alexei Svitkine <asvitkine@chromium.org> on 2012-08-08
Reviewed by Dimitri Glazkov.

This enables putting two fingers on the track pad to show overlay scroll bars on 10.8+.

* src/mac/WebInputEventFactory.mm:
(WebKit::phaseForNSEventPhase):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125125 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoGoogle search query text reverts to original search query after multiple searches
beidson@apple.com [Thu, 9 Aug 2012 00:34:11 +0000 (00:34 +0000)]
Google search query text reverts to original search query after multiple searches
<rdar://problem/10800686> and https://bugs.webkit.org/show_bug.cgi?id=93544

Reviewed by Darin Adler.

Source/WebCore:

For security sensitive fields we normally clear "autocomplete=off" form elements when
restoring a page from the page cache.

If the element is textual and has a defaultValue then "clearing" it actually restores
the default value.

There's no scenario we can imagine where that makes sense so we should not reset the
value in such fields.

Test: fast/forms/autocomplete-off-with-default-value-does-not-clear.html

* html/HTMLInputElement.cpp:
(WebCore::HTMLInputElement::parseAttribute): Update suspension callback registration as needed.
(WebCore::HTMLInputElement::needsSuspensionCallback): Don't reset text fields with a non-empty default value.

LayoutTests:

* fast/forms/autocomplete-off-with-default-value-does-not-clear-expected.txt: Added.
* fast/forms/autocomplete-off-with-default-value-does-not-clear.html: Added.
* fast/forms/resources/go-back.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125124 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoAdd Perf EWS IRC bot
rniwa@webkit.org [Thu, 9 Aug 2012 00:19:23 +0000 (00:19 +0000)]
Add Perf EWS IRC bot
https://bugs.webkit.org/show_bug.cgi?id=92913

Reviewed by Dirk Pranke.

Adds the preliminary implementation of perfalizer, a IRC bot that runs performance tests
based on IRC requests.

* Scripts/webkitpy/common/checkout/scm/scm_mock.py:
(MockSCM.head_svn_revision): The actual implementation returns string, so match that.
(MockSCM.svn_revision): Ditto.
* Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
(Bugzilla.add_attachment_to_bug): Added mimetype.
* Scripts/webkitpy/common/net/bugzilla/bugzilla_mock.py:
(MockBugzilla.add_attachment_to_bug): Ditto.
* Scripts/webkitpy/common/system/filesystem.py:
(FileSystem.copytree): Added.
* Scripts/webkitpy/common/system/filesystem_mock.py:
(MockFileSystem.copytree): Added.
* Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py: Rebaselined tests
that had assumed revision numbers were integers.
(test_run_with_json_output):
(test_run_with_description):
(test_run_generates_json_by_default):
(test_run_generates_and_show_results_page):
(test_run_with_json_source):
(test_run_with_multiple_repositories):
* Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py:
* Scripts/webkitpy/tool/commands/__init__.py:
* Scripts/webkitpy/tool/commands/perfalizer.py: Added.
(PerfalizerTask):
(PerfalizerTask.__init__):
(PerfalizerTask._copy_build_product_without_patch):
(PerfalizerTask.run):
(PerfalizerTask.parent_command):
(PerfalizerTask.run_webkit_patch):
(PerfalizerTask._json_path):
(PerfalizerTask._results_page_path):
(PerfalizerTask._run_perf_test):
(PerfalizerTask.run_command):
(PerfalizerTask.command_passed):
(PerfalizerTask.command_failed):
(PerfalizerTask.refetch_patch):
(PerfalizerTask.expected_failures):
(PerfalizerTask.build_style):
(PerfTest): IRC command.
(PerfTest.execute):
(Perfalizer): IRC bot.
(Perfalizer.begin_work_queue):
(Perfalizer.work_item_log_path):
(Perfalizer._is_old_failure):
(Perfalizer.next_work_item):
(Perfalizer.process_work_item):
(Perfalizer.handle_unexpected_error):
(Perfalizer.handle_script_error):
* Scripts/webkitpy/tool/commands/perfalizer_unittest.py: Added.
(PerfalizerTaskTest):
(PerfalizerTaskTest._create_and_run_perfalizer):
(PerfalizerTaskTest._create_and_run_perfalizer.logger):
(PerfalizerTaskTest._create_and_run_perfalizer.run_webkit_patch):
(PerfalizerTaskTest._create_and_run_perfalizer.run_perf_test):
(PerfalizerTaskTest.test_run):
(PerfalizerTaskTest.test_run_with_clean_fails):
(PerfalizerTaskTest.test_run_with_update_fails):
(PerfalizerTaskTest.test_run_with_build_fails):
(PerfalizerTaskTest.test_run_with_perf_test_fails):
(PerfalizerTaskTest.test_run_without_results_page):
* Scripts/webkitpy/tool/commands/queues_unittest.py:
(AbstractPatchQueueTest.test_upload_results_archive_for_patch): Rebaselined.
* Scripts/webkitpy/tool/commands/upload_unittest.py:
(test_attach_to_bug): Ditto.
(test_attach_to_bug_no_description_or_comment): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125123 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoCreating "basic waveform" Oscillator nodes is not efficient
commit-queue@webkit.org [Thu, 9 Aug 2012 00:18:23 +0000 (00:18 +0000)]
Creating "basic waveform" Oscillator nodes is not efficient
https://bugs.webkit.org/show_bug.cgi?id=93194

Patch by Raymond Toy <rtoy@google.com> on 2012-08-08
Reviewed by Chris Rogers.

Source/WebCore:

Cache the wavetables for the basic types so they don't have to be
recomputed every time.

Also fix a bug where oscillator type was always set to CUSTOM
instead of the specified oscillator type.  Test added for this.

Test: webaudio/oscillator-basic.html

* Modules/webaudio/Oscillator.cpp:
(WebCore):
(WebCore::Oscillator::setType): Use cached wavetables; fix bug in
setting the oscillator type.
* Modules/webaudio/Oscillator.h:
(Oscillator): Define static variables to hold cached wavetables.

LayoutTests:

Add test to verify that the returned oscillator type is the same
as what was set.

* webaudio/oscillator-basic-expected.txt: Added.
* webaudio/oscillator-basic.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125122 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[BlackBerry] Fix "-fPIC" define in the BlackBerry build
commit-queue@webkit.org [Thu, 9 Aug 2012 00:18:01 +0000 (00:18 +0000)]
[BlackBerry] Fix "-fPIC" define in the BlackBerry build
https://bugs.webkit.org/show_bug.cgi?id=93548

Patch by Ming Xie <mxie@rim.com> on 2012-08-08
Reviewed by Rob Buis.

In the current CMake release (version 2.8.5), ${CMAKE_SHARED
_LIBRARY_C_FLAGS} and ${CMAKE_SHARED_LIBRARY_CXX_FLAGS} is
set to empty (See cmake/Modules/Platform/QNX.cmake)

This breaks the assumption which WebKit CMake build system
makes in Source/cmake/WebKitHelper.cmake

* Source/cmake/OptionsBlackBerry.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125121 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoAvoid ASSERT(m_workerContext->isSharedWorkerContext()) in WorkerScriptController...
tsepez@chromium.org [Thu, 9 Aug 2012 00:16:38 +0000 (00:16 +0000)]
Avoid ASSERT(m_workerContext->isSharedWorkerContext()) in WorkerScriptController::initScript()
https://bugs.webkit.org/show_bug.cgi?id=93521

Reviewed by Adam Barth.

Source/WebCore:

Changes WorkerContext and its subclasses so that we do not call functions that rely upon subclass
virtual methods from the superclass constructor.

Test: http/tests/security/contentSecurityPolicy/worker-set-timeout-blocked.html

* workers/DedicatedWorkerContext.cpp:
(WebCore):
(WebCore::DedicatedWorkerContext::create):
(WebCore::DedicatedWorkerContext::DedicatedWorkerContext):
* workers/DedicatedWorkerContext.h:
(DedicatedWorkerContext):
* workers/SharedWorkerContext.cpp:
(WebCore):
(WebCore::SharedWorkerContext::create):
(WebCore::SharedWorkerContext::SharedWorkerContext):
* workers/SharedWorkerContext.h:
(SharedWorkerContext):
* workers/WorkerContext.cpp:
(WebCore::WorkerContext::WorkerContext):
(WebCore::WorkerContext::applyContentSecurityPolicyFromString):
(WebCore):
* workers/WorkerContext.h:
(WorkerContext):

LayoutTests:

* http/tests/security/contentSecurityPolicy/resources/worker-set-timeout.js: Added.
* http/tests/security/contentSecurityPolicy/worker-set-timeout-blocked-expected.txt: Added.
* http/tests/security/contentSecurityPolicy/worker-set-timeout-blocked.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125120 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago(r124484) inspector/device-orientation-success.html failing on Mac ports
dino@apple.com [Thu, 9 Aug 2012 00:01:30 +0000 (00:01 +0000)]
(r124484) inspector/device-orientation-success.html failing on Mac ports
https://bugs.webkit.org/show_bug.cgi?id=93552

Unreviewed test expectations update. inspector/device-orientation-success.html missing console output.

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125119 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoCanvasRenderContext2D::setFont() should ignore inherited properties and default keywo...
commit-queue@webkit.org [Wed, 8 Aug 2012 23:59:34 +0000 (23:59 +0000)]
CanvasRenderContext2D::setFont() should ignore inherited properties and default keyword value
https://bugs.webkit.org/show_bug.cgi?id=93491

Patch by Thiago Marcos P. Santos <thiago.santos@intel.com> on 2012-08-08
Reviewed by Kenneth Rohde Christiansen.

Source/WebCore:

Treat "inherit" as an invalid value. Also ignore the "default"
keyword that was previously treated as a font family. This second
change was done in the property validation and will also affect the
CSS font parsing, but it is probably harmless since font doesn't
recognize the default keyword anyway.

* css/CSSParser.cpp:
(WebCore::CSSParser::parseFontFamily):
* html/canvas/CanvasRenderingContext2D.cpp:
(WebCore::CanvasRenderingContext2D::setFont):

LayoutTests:

Unskipped tests that are now passing after this fix and removed wrong
expectations.

* platform/chromium/TestExpectations:
* platform/chromium/canvas/philip/tests/2d.text.font.parse.invalid-expected.txt: Removed.
* platform/efl/Skipped:
* platform/gtk/TestExpectations:
* platform/mac/canvas/philip/tests/2d.text.font.parse.invalid-expected.txt: Removed.
* platform/qt/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125118 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[chromium] Fix up includes and types in chromium compositor unit tests
jamesr@google.com [Wed, 8 Aug 2012 23:59:09 +0000 (23:59 +0000)]
[chromium] Fix up includes and types in chromium compositor unit tests
https://bugs.webkit.org/show_bug.cgi?id=93553

Reviewed by Adrienne Walker.

This removes some unused GraphicsContext3DPrivate.h includes and adds some includes that were being implicitely
depending upon - mostly for wtf/PassOwnPtr.h.

* tests/CCLayerTreeHostTest.cpp:
(::CCLayerTreeHostTestCommit::beginTest):
(::CCLayerTreeHostTestCommit::commitCompleteOnCCThread):
* tests/CCRenderSurfaceFiltersTest.cpp:
* tests/CompositorFakeWebGraphicsContext3D.h:
* tests/FakeGraphicsContext3DTest.cpp:
* tests/FakeWebScrollbarThemeGeometry.h:
* tests/TextureCopierTest.cpp:
* tests/ThrottledTextureUploaderTest.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125117 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[chromium] Add missing OVERRIDE and virtual annotations in compositor
jamesr@google.com [Wed, 8 Aug 2012 23:56:50 +0000 (23:56 +0000)]
[chromium] Add missing OVERRIDE and virtual annotations in compositor
https://bugs.webkit.org/show_bug.cgi?id=93438

Reviewed by Adrienne Walker.

As the title suggests.

* platform/graphics/chromium/BitmapCanvasLayerTextureUpdater.h:
(BitmapCanvasLayerTextureUpdater):
* platform/graphics/chromium/ScrollbarLayerChromium.h:
(ScrollbarLayerChromium):
* platform/graphics/chromium/SolidColorLayerChromium.h:
(SolidColorLayerChromium):
* platform/graphics/chromium/cc/CCDelayBasedTimeSource.h:
* platform/graphics/chromium/cc/CCFrameRateController.h:
(CCFrameRateController):
* platform/graphics/chromium/cc/CCScrollbarLayerImpl.h:
(CCScrollbarLayerImpl):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125116 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agowebkit-patch rebaseline-expectations needs to support multiple expectations files
dpranke@chromium.org [Wed, 8 Aug 2012 23:55:46 +0000 (23:55 +0000)]
webkit-patch rebaseline-expectations needs to support multiple expectations files
https://bugs.webkit.org/show_bug.cgi?id=89051

Reviewed by Adam Barth.

fixed the bug described above, added unit test in
test_expectations_unittest, and updated the tool tests
to not contain all of the "missing skia expectations" warnings.

* Scripts/webkitpy/layout_tests/models/test_expectations.py:
(TestExpectations.remove_rebaselined_tests):
(TestExpectations.remove_rebaselined_tests.without_rebaseline_modifier):
* Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
(RebaseliningTest.assertRemove):
(RebaseliningTest.test_remove):
* Scripts/webkitpy/tool/commands/rebaseline.py:
(RebaselineExpectations._update_expectations_files):
(RebaselineExpectations.execute):
* Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
(TestRebaseline.test_rebaseline_updates_expectations_file_noop):
(test_rebaseline_expectations):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125115 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[chromium] Refactor tile flags.
commit-queue@webkit.org [Wed, 8 Aug 2012 23:50:35 +0000 (23:50 +0000)]
[chromium] Refactor tile flags.
https://bugs.webkit.org/show_bug.cgi?id=93059

Patch by Eric Penner <epenner@google.com> on 2012-08-08
Reviewed by Adrienne Walker.

Source/WebCore:

This has become a general refactoring patch to make updateTiles easier
to work with. The 'updated' flag had become convoluted and incorrect
and the function had lots of special cases piled into it which
makes for tangled depedancies that are hard to reason with.

This refactors the updated flag into two flags ('validForFrame',
and 'occluded') for the two purposes it is actually used for,
simplifies the use of those flags.

'validForFrame' is improved to work for all tiles that are not dirty
before painting begins, rather than just the tiles that get touched via
updateTiles().

This refactors updateTiles() into several functions that more clearly
illustrate all the depedancies that have accumulated in there. After
doing that I managed to separate the occlusion pass from the painting
passes, ecapsulating the difficult dependancies in one documented
function 'markOcclusionsAndRequestTextures'.

Now update tiles doesn't need the 'idle' flag, and returns didPaint
as an out-param. The last remaining wierd output is m_failedPaint
which is needed in place of m_skipsIdlePaint to prevent infinite
paints from being requested.

As a last step I was going to merge LayerChromium::update() and
LayerChromium::needsMoreUpdates() by having update() just return
a boolean, but this proved to be a big change so I'm holding off
on that. That would let us remove the m_failedPaint and get rid
of needsIdlePaint() altogether.

Tested by many existing tests.

* platform/graphics/chromium/TiledLayerChromium.cpp:
(UpdatableTile):
(WebCore::UpdatableTile::resetUpdateState):
(WebCore::UpdatableTile::markForUpdate):
(WebCore::UpdatableTile::UpdatableTile):
(WebCore::TiledLayerChromium::TiledLayerChromium):
(WebCore::TiledLayerChromium::pushPropertiesTo):
(WebCore):
(WebCore::TiledLayerChromium::updateTiles):
(WebCore::TiledLayerChromium::markOcclusionsAndRequestTextures):
(WebCore::TiledLayerChromium::haveTexturesForTiles):
(WebCore::TiledLayerChromium::markTilesForUpdate):
(WebCore::TiledLayerChromium::updateTileTextures):
(WebCore::TiledLayerChromium::resetUpdateState):
(WebCore::TiledLayerChromium::updateContentRect):
(WebCore::TiledLayerChromium::needsIdlePaint):
* platform/graphics/chromium/TiledLayerChromium.h:
(TiledLayerChromium):

Source/WebKit/chromium:

Minor fixes to existing tests.

* tests/TiledLayerChromiumTest.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125114 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago [WK2] MiniBrowser crashes on quit if any windows were previously closed
ap@apple.com [Wed, 8 Aug 2012 23:47:42 +0000 (23:47 +0000)]
    [WK2] MiniBrowser crashes on quit if any windows were previously closed
        https://bugs.webkit.org/show_bug.cgi?id=93529

        Reviewed by John Sullivan.

        * MiniBrowser/mac/AppDelegate.h:
        * MiniBrowser/mac/AppDelegate.m:
        (-[BrowserAppDelegate init]):
        (-[BrowserAppDelegate newWindow:]):
        (-[BrowserAppDelegate browserWindowWillClose:]):
        (-[BrowserAppDelegate applicationWillTerminate:]):
        (-[BrowserAppDelegate frontmostBrowserWindowController]):
        Track browser windows explicitly, not relying on [NSApp windows]. Closed windows
        are not automatically removed from the list until deallocated, so a refcounting
        error can result in working on a closed window.

        * MiniBrowser/mac/BrowserWindowController.m:
        (-[BrowserWindowController dealloc]): Moved code from -windowWillClose: here.
        Releasing data members is -dealloc's job.
        (-[BrowserWindowController windowWillClose:]): Notify BrowserAppDelegate that
        the window is no longer open. Autorelease self, making sure that WKView and
        then WebPageProxy get released.
        (-[BrowserWindowController applicationTerminating]): Don't release _webView.pageRef.
        WKView is responsible for its lifetime.
        (closePage): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125113 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agocss3/calc/transitions-dependent.html was causing css3/calc/transitions.html to fail
commit-queue@webkit.org [Wed, 8 Aug 2012 23:39:50 +0000 (23:39 +0000)]
css3/calc/transitions-dependent.html was causing css3/calc/transitions.html to fail
https://bugs.webkit.org/show_bug.cgi?id=93515

Patch by Roger Fong <roger_fong@apple.com> on 2012-08-08
Reviewed by Tim Horton.

The transitions-dependent.html test was causing the transitions.html test to fail.
This was a result of a timeout callback set in the first test that was only being fired
in the second test because the first test did not use waitUntilDone/notifyDone.
The second test printed out an extra TypeError: 'undefined' error message because the callback
was defined in the first test and was thus 'undefined' by the time it was called while the second test
was running.

In addition, the test contained the following line:
window.addEventListener("load", function() { waitForAnimationStart(runTest(expectedValues)); }, false);
which does not work as intended because the parameter to waitForAnimationStart should have been a function
that calls runTest, not the call to the runTest function itself, which calls the method immediately.

* css3/calc/transitions-dependent.html:
Added waitUntilDone/notifyDone.
Fixed function parameter.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125111 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[Chromium] Unreviewed gardening: Chromium results for compositing/repaint/newly-compo...
fmalita@chromium.org [Wed, 8 Aug 2012 23:36:00 +0000 (23:36 +0000)]
[Chromium] Unreviewed gardening: Chromium results for compositing/repaint/newly-composited-on-scroll.html.

* platform/chromium-mac/compositing/repaint/newly-composited-on-scroll-expected.png: Added.
* platform/chromium-win/compositing/repaint/newly-composited-on-scroll-expected.png: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125109 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[chromium] Clean up WebKit.h and WebKitPlatformSupport.h includes in chromium/src
jamesr@google.com [Wed, 8 Aug 2012 23:27:47 +0000 (23:27 +0000)]
[chromium] Clean up WebKit.h and WebKitPlatformSupport.h includes in chromium/src
https://bugs.webkit.org/show_bug.cgi?id=93539

Reviewed by Adam Barth.

Most of these files don't need either WebKit.h or WebKitPlatformSupport.h. Some need <public/Platform.h>, and
some are relying on the forwarding headers. This updates includes to what is actually needed and points platform
includes at <public/Foo.h> instead of "platform/Foo.h" where possible.

* src/ApplicationCacheHostInternal.h:
* src/AssociatedURLLoader.cpp:
* src/AsyncFileSystemChromium.cpp:
(WebCore::AsyncFileSystemChromium::AsyncFileSystemChromium):
* src/BlobRegistryProxy.cpp:
* src/ChromiumCurrentTime.cpp:
* src/ChromiumOSRandomSource.cpp:
* src/ChromiumThreading.cpp:
* src/EditorClientImpl.cpp:
* src/FrameLoaderClientImpl.cpp:
* src/IDBFactoryBackendProxy.cpp:
* src/LocalizedStrings.cpp:
* src/PlatformMessagePortChannel.cpp:
* src/WebCompositorInputHandlerImpl.cpp:
* src/WebMediaPlayerClientImpl.cpp:
* src/WebPluginContainerImpl.cpp:
* src/WebViewImpl.cpp:
* src/WebWorkerClientImpl.cpp:
* src/WorkerAsyncFileSystemChromium.cpp:
* src/WorkerFileWriterCallbacksBridge.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125108 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[Chromium] Unreviewed gardening: compositing/plugins/no-backing-store.html yields...
fmalita@chromium.org [Wed, 8 Aug 2012 23:19:42 +0000 (23:19 +0000)]
[Chromium] Unreviewed gardening: compositing/plugins/no-backing-store.html yields incorrect results.

* platform/chromium/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125106 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agohttps://bugs.webkit.org/show_bug.cgi?id=93393
bdakin@apple.com [Wed, 8 Aug 2012 23:09:37 +0000 (23:09 +0000)]
https://bugs.webkit.org/show_bug.cgi?id=93393
Overflow regions sometimes repaint incorrectly after going into or
coming out of compositing mode
-and corresponding-
<rdar://problem/12006463>

Reviewed by Simon Fraser.

My first patch to fix this bug removed an if (parent()) check that is
needed to prevent a table crash seen in
fast/table/table-row-compositing-repaint-crash.html
The parent() check was actually added originally to prevent this same
crash. See http://trac.webkit.org/changeset/110456
This patch adds that check back, but really we should delay the
computation of repaint rects if layout has not happened yet.
* rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::updateBacking):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125104 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoREGRESSION (r124416): fast/forms/basic-selects.html failing on Lion Debug Tests
dino@apple.com [Wed, 8 Aug 2012 22:40:50 +0000 (22:40 +0000)]
REGRESSION (r124416): fast/forms/basic-selects.html failing on Lion Debug Tests
https://bugs.webkit.org/show_bug.cgi?id=93538

Unreviewed test expectations update after regression from r124416 on Mac ports.

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125103 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoAvoid backing store on layers created for CoreAnimation plugins
scherkus@chromium.org [Wed, 8 Aug 2012 22:28:16 +0000 (22:28 +0000)]
Avoid backing store on layers created for CoreAnimation plugins
https://bugs.webkit.org/show_bug.cgi?id=93526
<rdar://problem/12052828>

Patch by Simon Fraser <simon.fraser@apple.com> on 2012-08-08
Reviewed by Dean Jackson.

Test for backing store on CoreAnimation-based plugins in various
configurations.

* compositing/plugins/no-backing-store-expected.txt: Added.
* compositing/plugins/no-backing-store.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125102 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoAvoid backing store on layers created for CoreAnimation plugins
simon.fraser@apple.com [Wed, 8 Aug 2012 22:17:34 +0000 (22:17 +0000)]
Avoid backing store on layers created for CoreAnimation plugins
https://bugs.webkit.org/show_bug.cgi?id=93526
<rdar://problem/12052828>

Reviewed by Dean Jackson.

Source/WebCore:

We create a compositing layer for <embed> objects whose plug-in supports
the CoreAnimation rendering model. When the embed's renderer is not styled,
we can avoid allocating backing store for it.

Test: compositing/plugins/no-backing-store.html

* rendering/RenderLayerBacking.cpp:
(WebCore::isCompositedPlugin):
(WebCore::RenderLayerBacking::isSimpleContainerCompositingLayer):

LayoutTests:

Test for backing store on CoreAnimation-based plugins in various
configurations.

* compositing/plugins/no-backing-store-expected.txt: Added.
* compositing/plugins/no-backing-store.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125101 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years agoREGRESSION: Crash when plug-in initialization fails
andersca@apple.com [Wed, 8 Aug 2012 22:16:35 +0000 (22:16 +0000)]
REGRESSION: Crash when plug-in initialization fails
https://bugs.webkit.org/show_bug.cgi?id=93534
<rdar://problem/12059786>

Reviewed by Simon Fraser.

Protect the PluginProxy while calling out to functions that could cause it to be destroyed.

* WebProcess/Plugins/PluginProxy.cpp:
(WebKit::PluginProxy::didFailToCreatePluginInternal):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125100 268f45cc-cd09-0410-ab3c-d52691b4dbfc

8 years ago[Chromium] Updating fast/canvas/webgl/gl-teximage.html to prevent valid variations...
commit-queue@webkit.org [Wed, 8 Aug 2012 22:13:19 +0000 (22:13 +0000)]
[Chromium] Updating fast/canvas/webgl/gl-teximage.html to prevent valid variations in results from causing test failures.
https://bugs.webkit.org/show_bug.cgi?id=89494

Patch by Jeff Timanus <twiz@chromium.org> on 2012-08-08
Reviewed by Adrienne Walker.

* fast/canvas/webgl/gl-teximage-expected.txt:
* fast/canvas/webgl/gl-teximage.html: Only display the actual percentage of passing pixels when outside of DRT.
* platform/chromium/TestExpectations:
* platform/efl/TestExpectations:
* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@125099 268f45cc-cd09-0410-ab3c-d52691b4dbfc