Timeline
Jun 28, 2016:
- 11:33 PM Changeset in webkit [202615] by
-
- 4 edits in trunk/Source/WebCore
[GStreamer] Adaptive streaming issues
https://bugs.webkit.org/show_bug.cgi?id=144040
Reviewed by Philippe Normand.
There are multiple deadlocks in the web process when HLS content is loaded by GStreamer. It happens because gst
is using several threads to download manifest, fragments, monitor the downloads, etc. To download the fragments
and manifest it always creates the source element in a separate thread, something that is not actually expected
to happen in WebKit source element. Our source element is always scheduling tasks (start, stop, need-data,
enough-data and seek) to the main thread, and those downloads that use the ResourceHandleStreamingClient
(there's no player associated) also happen in the main thread, because libsoup calls all its async callbacks in
the main thread. So, the result is that it can happen that we end up blocking the main thread in a lock until
the download finishes, but the download never finishes because tasks are scheduled in the main thread that is
blocked in a lock. This can be prevented by always using a secondary thread for downloads made by
ResourceHandleStreamingClient, using its own run loop with a different GMainContext so that libsoup sends
callbacks to the right thread. We also had to refactor the tasks a bit, leaving the thread safe parts to be run
in the calling thread always, and only scheduling to the main thread in case of not using
ResourceHandleStreamingClient and only for the non thread safe parts.
This patch also includes r200455 that was rolled out, but it was a perfectly valid workaround for GST bug.
- platform/graphics/gstreamer/GRefPtrGStreamer.cpp:
(WTF::ensureGRef): Consume the floating ref if needed.
- platform/graphics/gstreamer/GRefPtrGStreamer.h:
- platform/graphics/gstreamer/WebKitWebSourceGStreamer.cpp:
(webkit_web_src_init): Check if object is being created in the main thread.
(webKitWebSrcStop): Stop the media resource loader in the main thread and the resource handle streaming in the
current thread.
(webKitWebSrcStart): Start the media resource loader in the main thread and the resource handle streaming in
the current thread.
(webKitWebSrcChangeState): Call webKitWebSrcStart and webKitWebSrcStop in the current thread.
(webKitWebSrcNeedData): Update status in the current thread and notify the media resource loader in the main thread.
(webKitWebSrcEnoughData): Ditto.
(webKitWebSrcSeek): Ditto.
(webKitWebSrcSetMediaPlayer): Add an assert to ensure that source elements used by WebKit are always created in
the main thread.
(ResourceHandleStreamingClient::ResourceHandleStreamingClient): Use a secondary thread to do the download.
(ResourceHandleStreamingClient::~ResourceHandleStreamingClient): Stop the secondary thread.
(ResourceHandleStreamingClient::setDefersLoading): Notify the secondary thread.
- 11:25 PM Changeset in webkit [202614] by
-
- 9 edits in trunk/Source/WebCore
Remove ThreadableLoaderOptions origin
https://bugs.webkit.org/show_bug.cgi?id=159221
Patch by Youenn Fablet <youennf@gmail.com> on 2016-06-28
Reviewed by Sam Weinig.
No change of behavior.
- loader/DocumentThreadableLoader.cpp:
(WebCore::DocumentThreadableLoader::loadResourceSynchronously): Adding origing parameter.
(WebCore::DocumentThreadableLoader::create): Ditto.
(WebCore::DocumentThreadableLoader::DocumentThreadableLoader): Ditto.
(WebCore::DocumentThreadableLoader::redirectReceived): Setting m_origin.
(WebCore::DocumentThreadableLoader::securityOrigin): Checking m_origin.
- loader/DocumentThreadableLoader.h: Adding m_origin member.
- loader/ThreadableLoader.cpp:
(WebCore::ThreadableLoaderOptions::ThreadableLoaderOptions): Removing origin.
(WebCore::ThreadableLoaderOption::isolatedCopy): Deleted.
- loader/ThreadableLoader.h: Removing origin parameter and isolatedCopy function.
- loader/WorkerThreadableLoader.cpp:
(WebCore::LoaderTaskOptions::LoaderTaskOptions): Structure to pass loader task options from one thread to another.
(WebCore::WorkerThreadableLoader::MainThreadBridge::MainThreadBridge):
- page/EventSource.cpp:
(WebCore::EventSource::connect): Removing setting of the origin.
- workers/WorkerScriptLoader.cpp:
(WebCore::WorkerScriptLoader::loadSynchronously): Ditto.
(WebCore::WorkerScriptLoader::loadAsynchronously): Ditto.
- xml/XMLHttpRequest.cpp:
(WebCore::XMLHttpRequest::createRequest): Ditto.
- 10:55 PM Changeset in webkit [202613] by
-
- 2 edits8 adds in trunk/PerformanceTests
ES6SampleBench should have a harness
https://bugs.webkit.org/show_bug.cgi?id=159246
Reviewed by Saam Barati.
This adds a simple web harness for ES6SampleBench. It runs Air and Basic 10 times for 200
iterations each and reports three metrics:
First iteration: the time it takes for the first iteration to run. This is the first
iteration after the benchmark is loaded into the iframe, so it's representative of what
would happen if one of these workloads only ran for a short time.
Worst 2%: of the last 199 iterations, the average of the worst 2% iterations. If code like
any of these workloads was used in an important event handler, you'd want that code to run
well in the worst case in addition to having great throughput.
Steady state: the total of the last 199 iterations. This is representative of what would
happen if you ran code like this for a long time.
The total score is the geomean of the firstIteration/worstCase/steadyState numbers of the
two benchmarks.
The harness does statistics using Student's T-distribution confidence intervals.
- ES6SampleBench/Basic/benchmark.js:
(Benchmark):
- ES6SampleBench/air_benchmark.js: Added.
- ES6SampleBench/basic_benchmark.js: Added.
- ES6SampleBench/driver.js: Added.
(Driver):
(Driver.prototype.addBenchmark):
(Driver.prototype.start):
(Driver.prototype.reportResult):
(Driver.prototype.reportError):
(Driver.prototype._recomputeSummary.Geomean):
(Driver.prototype._recomputeSummary.Geomean.prototype.add):
(Driver.prototype._recomputeSummary.Geomean.prototype.get result):
(Driver.prototype._recomputeSummary):
(Driver.prototype._iterate):
(Driver.prototype._updateIterations):
- ES6SampleBench/glue.js: Added.
- ES6SampleBench/index.html: Added.
- ES6SampleBench/results.js: Added.
(Results):
(Results.prototype.get benchmark):
(Results.prototype.reset):
(Results.prototype.reportRunning):
(Results.prototype.reportDone):
(Results.prototype.reportResult.averageAbovePercentile):
(Results.prototype.reportResult):
(Results.prototype.reportError):
- ES6SampleBench/stats.js: Added.
(Stats):
(Stats.prototype.reset):
(Stats.prototype.add):
(Stats.prototype.get numIterations):
(Stats.prototype.valueForIteration):
(Stats.get result.tDist):
(Stats.prototype.get result):
(Stats.prototype.toString):
(Stats.prototype._update):
- ES6SampleBench/style.css: Added.
(body):
(body, th, tr):
(h1, h2, h3, h4):
(h1):
(h2):
(h3):
(hr):
(address):
(img):
(.underline):
(ol.loweralpha):
(ol.upperalpha):
(ol.lowerroman):
(ol.upperroman):
(ol.arabic):
(.banner-link:link, .banner-link:visited):
(:link, :visited):
(h4 :link, h4 :visited, h5 :link, h5 :visited, h6 :link, h6 :visited):
(.anchor:link, .anchor:visited):
(* > .anchor:link, * > .anchor:visited):
(span:hover .anchor):
(a.forbidden, span.forbidden):
(a.missing:hover):
(a.closed:link, a.closed:visited, span.closed):
(pre):
(div.code):
(div.code pre):
(dt):
(dd):
(dd:last-child):
(.site-logo):
(.site-logo .tagline):
(table):
(#contents):
(p):
(p:last-child):
(th):
(td):
- 10:04 PM Changeset in webkit [202612] by
-
- 4 edits in trunk/Source/WebKit2
Unreviewed, rolling out r202586 and r202595.
https://bugs.webkit.org/show_bug.cgi?id=159247
These changes broke the iOS build (Requested by ryanhaddad on
#webkit).
Reverted changesets:
"Keep track of when a WKWebView is blank before the initial
non-empty layout"
https://bugs.webkit.org/show_bug.cgi?id=159217
http://trac.webkit.org/changeset/202586
"Try to fix the build."
http://trac.webkit.org/changeset/202595
Patch by Commit Queue <commit-queue@webkit.org> on 2016-06-28
- 9:49 PM Changeset in webkit [202611] by
-
- 89 edits in trunk
Unreviewed, rolling out r202580.
https://bugs.webkit.org/show_bug.cgi?id=159245
Caused all WKTR tests to fail on GuardMalloc and Production
only for unknown reasons, investigating offline. (Requested by
brrian on #webkit).
Reverted changeset:
"RunLoop::Timer should use constructor templates instead of
class templates"
https://bugs.webkit.org/show_bug.cgi?id=159153
http://trac.webkit.org/changeset/202580
- 9:31 PM Changeset in webkit [202610] by
-
- 1 edit3 adds in trunk/LayoutTests
AX: Add accessibility tests for MathML operators
https://bugs.webkit.org/show_bug.cgi?id=124836
Patch by Frederic Wang <fwang@igalia.com> on 2016-06-28
Reviewed by Joanmarie Diggs.
We add one test to verify the (sub)roles of MathML operators with various fence and separator
properties. These properties can be set via an explicit attribute or read from the operator
dictionary.
- accessibility/math-mo-fences-and-separators.html: Added.
- platform/gtk/accessibility/math-mo-fences-and-separators-expected.txt: Added.
- platform/mac/accessibility/math-mo-fences-and-separators-expected.txt: Added.
- 9:17 PM Changeset in webkit [202609] by
-
- 23 edits in trunk/LayoutTests
Enhance shouldNotThrow()/shouldThrow() to accept functions and a descriptive message
<https://webkit.org/b/159232>
Reviewed by Brent Fulgham.
Based on a Blink change (patch by <hongchan@chromium.org>):
<https://src.chromium.org/viewvc/blink?view=revision&revision=192204>
Currently, shouldNotThrow() and shouldThrow() take the following
arguments:
shouldNotThrow(evalString)
shouldThrow(evalString, expectedExceptionString)
The challenges with this design are:
1) The 'evalString' must capture every variable that it
needs, which means the code can be long, and concatenated
into a single line. It would be really nice to be able
to use an (anonymous) function to capture variables
instead.
2) The 'evalString' is literally printed out in the test
results, which isn't always the most descriptive. A
descriptive message would make it clearer what failed.
3) When changing a shouldThrow() into a shouldNotThrow()
or copying/pasting code, it's possible to forget to
remove 'expectedExceptionString' from the function call.
This patch changes the methods to take the following arguments:
shouldNotThrow(evalString|function [, message])
shouldThrow(evalString|function, expectedExceptionString [, message])
If 'function' is passed in, then it is invoked instead of
evaluated, and 'message' replaces the literal code in the
pass/fail output.
This patch also adds the global 'didFailSomeTests' variable to
js-test.js, which already exists in js-test-pre.js. This was
added to js-test-pre.js in r153203 by Oliver Hunt to
LayoutTests/fast/js/resources/js-test-pre.js.
- fast/canvas/webgl/canvas-supports-context-expected.txt:
- fast/canvas/webgl/gl-bind-attrib-location-before-compile-test-expected.txt:
- fast/css-grid-layout/grid-element-auto-repeat-get-set-expected.txt:
- fast/dom/getElementsByClassName/ASCII-case-insensitive-expected.txt:
- storage/indexeddb/cursor-basics-expected.txt:
- storage/indexeddb/cursor-basics-private-expected.txt:
- Update expected results to include "Some tests fail." since some subtests actually do fail during these tests.
- fast/css/parsing-css-lang.html:
- fast/css/parsing-css-matches-1.html:
- fast/css/parsing-css-matches-2.html:
- fast/css/parsing-css-matches-3.html:
- fast/css/parsing-css-matches-4.html:
- fast/css/parsing-css-not-1.html:
- fast/css/parsing-css-not-2.html:
- fast/css/parsing-css-not-3.html:
- fast/css/parsing-css-not-4.html:
- fast/css/parsing-css-nth-child-of-1.html:
- fast/css/parsing-css-nth-child-of-2.html:
- fast/css/parsing-css-nth-last-child-of-1.html:
- fast/css/parsing-css-nth-last-child-of-2.html:
- js/script-tests/arrowfunction-supercall.js:
- Remove expectedExceptionString from shouldNotThrow() calls after they were changed from shouldThrow() calls.
- resources/js-test-pre.js:
(shouldNotThrow): Change to invoke first argument if it is a
function, else use eval() as before. Use second argurment in
place of first argument (if set) when printing results. NOTE:
Care was taken not to add any lines of code to prevent changes
to test results.
(shouldThrow): Ditto. Reformat code.
- resources/js-test.js: Declare 'didFailSomeTests'.
(testFailed): Set 'didFailSomeTests' to true when a test fails.
(shouldNotThrow): Same changes as js-test-pre.js.
(shouldThrow): Ditto.
(isSuccessfullyParsed): Output a message if 'didFailSomeTests'
is true.
- 9:06 PM Changeset in webkit [202608] by
-
- 3 edits1 add in trunk/Source/JavaScriptCore
We should not crash there is a finally inside a for-in loop
https://bugs.webkit.org/show_bug.cgi?id=159243
<rdar://problem/27018910>
Reviewed by Benjamin Poulain.
Previously we would swap the m_forInContext with an empty vector
then attempt to shrink the size of m_forInContext by the amount
we expected. This meant that if there was more than one ForInContext
on the stack and we wanted to pop exactly one off we would crash.
This patch makes ForInContexts RefCounted so they can be duplicated
into other vectors. It also has ForInContexts copy the entire stack
rather than do the swap that we did before. This makes ForInContexts
work the same as the other contexts.
- bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::emitComplexPopScopes):
(JSC::BytecodeGenerator::pushIndexedForInScope):
(JSC::BytecodeGenerator::pushStructureForInScope):
- bytecompiler/BytecodeGenerator.h:
- tests/stress/finally-for-in.js: Added.
(repeat):
(createSimple):
- 9:03 PM Changeset in webkit [202607] by
-
- 5 edits in trunk/Source/WebCore
Rename ChildrenAffectedByActive to StyleAffectedByActive
https://bugs.webkit.org/show_bug.cgi?id=159187
Reviewed by Antti Koivisto.
Flags named "ChildrenAffectedBy" are used when the invalidation
of children is so crazy that we invalidate the whole parent subtree instead.
That's not the case for :active. It is a straightforward element invalidation.
Consequently, the property is renamed to StyleAffectedByActive.
- dom/Element.cpp:
(WebCore::Element::setActive):
(WebCore::Element::setStyleAffectedByActive):
(WebCore::Element::hasFlagsSetDuringStylingOfChildren):
(WebCore::Element::rareDataStyleAffectedByActive):
(WebCore::Element::setChildrenAffectedByActive): Deleted.
(WebCore::Element::rareDataChildrenAffectedByActive): Deleted.
- dom/Element.h:
(WebCore::Element::styleAffectedByActive):
(WebCore::Element::childrenAffectedByActive): Deleted.
- dom/ElementRareData.h:
(WebCore::ElementRareData::styleAffectedByActive):
(WebCore::ElementRareData::setStyleAffectedByActive):
(WebCore::ElementRareData::ElementRareData):
(WebCore::ElementRareData::resetDynamicRestyleObservations):
(WebCore::ElementRareData::childrenAffectedByActive): Deleted.
(WebCore::ElementRareData::setChildrenAffectedByActive): Deleted.
- style/StyleRelations.cpp:
(WebCore::Style::commitRelations):
- 8:56 PM Changeset in webkit [202606] by
-
- 4 edits in trunk/Tools
Use a regex to check if a test step is for JavaScriptCore
https://bugs.webkit.org/show_bug.cgi?id=159224
Patch by Srinivasan Vijayaraghavan <svijayaraghavan@apple.com> on 2016-06-28
Reviewed by Geoffrey Garen.
- BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/BuildbotCombinedQueueView.js:
(BuildbotCombinedQueueView.prototype.update): Replace array membership test with regex test.
- BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/BuildbotIteration.js:
(BuildbotIteration.prototype._parseData): Replace array membership test with regex test.
- BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/BuildbotTesterQueueView.js:
(BuildbotTesterQueueView.prototype.appendBuilderQueueStatus): Replace array membership test with regex test.
- 8:53 PM Changeset in webkit [202605] by
-
- 3 edits in trunk/Source/WebKit2
Unreviewed, rolling out r202598.
https://bugs.webkit.org/show_bug.cgi?id=159244
broke the build. (Requested by keith_miller on #webkit).
Reverted changeset:
"Try to fix the build."
http://trac.webkit.org/changeset/202598
- 7:26 PM Changeset in webkit [202604] by
-
- 2 edits in trunk/Tools
btjs no longer accepts optional frame count argument
https://bugs.webkit.org/show_bug.cgi?id=159235
Reviewed by Saam Barati.
Fix the detection of optional backtrace-depth parameter to use the length
of the command.
- lldb/lldb_webkit.py:
(btjs):
- 7:07 PM Changeset in webkit [202603] by
-
- 2 edits in trunk/Tools
Update animometer.plan
Rubber-stamped by Said Abou-Hallawa.
- Scripts/webkitpy/benchmark_runner/data/plans/animometer.plan: Update to r202601.
- 7:06 PM Changeset in webkit [202602] by
-
- 2 edits1 add in trunk/Source/JavaScriptCore
Assertion failure or crash when accessing let-variable in TDZ with eval with a function in it that returns let variable
https://bugs.webkit.org/show_bug.cgi?id=158796
<rdar://problem/26984659>
Reviewed by Michael Saboff.
There was a bug where some functions inside of an eval were
omitting a necessary TDZ check. This obviously leads to bad
things because a variable under TDZ is the null pointer.
The eval's bytecode was generated with the correct TDZ set, but
it created all its functions before pushing that TDZ set onto
the stack. That's a mistake. Those functions need to be created with
that TDZ set. The solution is simple, the TDZ set that the eval
is created with needs to be pushed onto the TDZ stack before
the eval creates any functions.
- bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::BytecodeGenerator):
- tests/stress/variable-under-tdz-eval-tricky.js: Added.
(assert):
(throw.new.Error):
(assert.try.underTDZ):
- 6:52 PM Changeset in webkit [202601] by
-
- 4 edits2 copies in trunk/PerformanceTests
Update focus test
https://bugs.webkit.org/show_bug.cgi?id=159242
rdar://problem/27070007
Reviewed by Dean Jackson.
Provisionally reviewed by Said Abou-Hallawa.
Move previous test to dom suite, and update the test for better reporting of frame rate, although
it uses a different rendering path.
- Animometer/resources/debug-runner/tests.js: Add to dom suite.
- Animometer/tests/dom/focus.html: Copied from PerformanceTests/Animometer/tests/master/focus.html.
- Animometer/tests/dom/resources/focus.js: Copied from PerformanceTests/Animometer/tests/master/resources/focus.js.
- Animometer/tests/master/focus.html: Remove center element.
- Animometer/tests/master/resources/focus.js: Use narrower size range with smaller particles. Remove the
container elements. Inline getBlurValue and getOpacityValue since they are only called once.
- 6:37 PM Changeset in webkit [202600] by
-
- 1 edit1 add in trunk/Websites/webkit.org
Add a more interactive version of the demo.
- blog-files/color-gamut/comparison.html: Added.
- 6:04 PM Changeset in webkit [202599] by
-
- 14 edits4 adds in trunk
Implement "replacement" codec
https://bugs.webkit.org/show_bug.cgi?id=159180
<rdar://problem/26015178>
Reviewed by Brent Fulgham.
LayoutTests/imported/w3c:
- web-platform-tests/dom/nodes/Document-characterSet-normalization-expected.txt:
Source/WebCore:
Test: fast/encoding/charset-replacement.html
Add support for "replacement" codec according to the spec:
https://encoding.spec.whatwg.org/#replacement
According to the spec, encoding labels {"csiso2022kr", "hz-gb-2312", "iso-2022-cn",
"iso-2022-cn-ext", "iso-2022-kr"} are used to conduct certain attacks that abuse
a mismatch between encodings supported on the server and the client. Therefore,
they are grouped under the "replacement" codec, which does the following things
to prevent those attacks.
1) Decode: terminates with a single U+FFFD.
2) Encode: treated as UTF-8.
Furthermore, the "replacement" codec is a specification convenience to group those
vulnerable encoding labels. Therefore, it should not be able to use directly.
This change is based on the following Blink changes:
https://codereview.chromium.org/265973003, and
https://codereview.chromium.org/261013007.
- CMakeLists.txt:
- WebCore.xcodeproj/project.pbxproj:
- platform/text/TextAllInOne.cpp:
- platform/text/TextCodecReplacement.cpp: Added.
(WebCore::TextCodecReplacement::create):
(WebCore::TextCodecReplacement::TextCodecReplacement):
(WebCore::TextCodecReplacement::registerEncodingNames):
(WebCore::TextCodecReplacement::registerCodecs):
(WebCore::TextCodecReplacement::decode):
- platform/text/TextCodecReplacement.h: Added.
- platform/text/TextEncoding.cpp:
(WebCore::TextEncoding::TextEncoding):
- platform/text/TextEncodingRegistry.cpp:
(WebCore::isReplacementEncoding):
(WebCore::extendTextCodecMaps):
- platform/text/TextEncodingRegistry.h:
LayoutTests:
- fast/encoding/char-decoding-expected.txt:
- fast/encoding/char-decoding.html:
- fast/encoding/char-encoding-expected.txt:
- fast/encoding/char-encoding.html:
- fast/encoding/charset-replacement-expected.txt: Added.
- fast/encoding/charset-replacement.html: Added.
- 5:47 PM Changeset in webkit [202598] by
-
- 3 edits in trunk/Source/WebKit2
- UIProcess/API/Cocoa/WKWebView.mm:
(convertSystemLayoutDirection):
(-[WKWebView _setIsBlankBeforeFirstNonEmptyLayout:]):
(-[WKWebView _initializeWithConfiguration:]):
- UIProcess/API/Cocoa/WKWebViewInternal.h:
- 5:38 PM Changeset in webkit [202597] by
-
- 5 edits in trunk
REGRESSION (r200946): Improper backtracking from last alternative in sticky patterns
https://bugs.webkit.org/show_bug.cgi?id=159233
Reviewed by Mark Lam.
Source/JavaScriptCore:
Jump to fail exit code when the last alternative of a sticky pattern fails.
- yarr/YarrJIT.cpp:
(JSC::Yarr::YarrGenerator::backtrack):
LayoutTests:
Updated tests.
- js/regexp-sticky-expected.txt:
- js/script-tests/regexp-sticky.js:
- 5:18 PM Changeset in webkit [202596] by
-
- 2 edits in trunk/LayoutTests
Marking fast/images/composited-animated-gif-outside-viewport.html as flaky on ios-simulator
https://bugs.webkit.org/show_bug.cgi?id=157592
Unreviewed test gardening.
- platform/ios-simulator/TestExpectations:
- 5:12 PM BadContent edited by
- (diff)
- 4:07 PM Changeset in webkit [202595] by
-
- 3 edits in trunk/Source/WebKit2
Try to fix the build.
- UIProcess/API/Cocoa/WKWebView.mm:
- UIProcess/API/Cocoa/WKWebViewInternal.h:
- 3:16 PM Changeset in webkit [202594] by
-
- 2 edits in trunk/Tools
Unreviewed, fix JSC tests. Air.js moved, and the harness needs to know.
- Scripts/run-javascriptcore-tests:
(runJSCStressTests):
- 3:12 PM Changeset in webkit [202593] by
-
- 2 edits in trunk/Websites/webkit.org
Fixed tile height for mobile phone-sized viewports.
https://bugs.webkit.org/show_bug.cgi?id=159228
Reviewed by Timothy Hatcher.
Prevents cutoff text for featured tiles and prevents overlapping text and controls on the Twitter tile for phone-sized viewports.
- wp-content/themes/webkit/style.css:
(.twitter-tile.text-only .tile-content):
(.tile):
- 3:11 PM Changeset in webkit [202592] by
-
- 2 edits in trunk/Source/WebCore
Remove incorrect comments in HTMLCanvasElement
https://bugs.webkit.org/show_bug.cgi?id=159229
Reviewed by Sam Weinig.
These comments are wrong.
- html/HTMLCanvasElement.cpp:
(WebCore::HTMLCanvasElement::probablySupportsContext):
(WebCore::HTMLCanvasElement::getContext): Deleted.
- 3:09 PM Changeset in webkit [202591] by
-
- 2 edits2 moves1 add in trunk/PerformanceTests
Move Air.js and Basic into ES6SampleBench
Rubber stamped by Geoffrey Garen.
We want to group our ES6 benchmarks together, and eventually, we want to give them a
common harness.
- Air.js: Removed.
- Air.js/README.md: Removed.
- Air.js/airjs-tests.yaml: Removed.
- Air.js/all.js: Removed.
- Air.js/allocate_stack.js: Removed.
- Air.js/arg.js: Removed.
- Air.js/basic_block.js: Removed.
- Air.js/benchmark.js: Removed.
- Air.js/code.js: Removed.
- Air.js/custom.js: Removed.
- Air.js/frequented_block.js: Removed.
- Air.js/insertion_set.js: Removed.
- Air.js/inst.js: Removed.
- Air.js/liveness.js: Removed.
- Air.js/make_dist.sh: Removed.
- Air.js/opcode.js: Removed.
- Air.js/payload-airjs-ACLj8C.js: Removed.
- Air.js/payload-gbemu-executeIteration.js: Removed.
- Air.js/payload-imaging-gaussian-blur-gaussianBlur.js: Removed.
- Air.js/payload-typescript-scanIdentifier.js: Removed.
- Air.js/reg.js: Removed.
- Air.js/stack_slot.js: Removed.
- Air.js/stress-test.js: Removed.
- Air.js/strip-hash.rb: Removed.
- Air.js/symbols.js: Removed.
- Air.js/test.html: Removed.
- Air.js/test.js: Removed.
- Air.js/tmp.js: Removed.
- Air.js/tmp_base.js: Removed.
- Air.js/util.js: Removed.
- Basic: Removed.
- Basic/ast.js: Removed.
- Basic/basic.js: Removed.
- Basic/benchmark.js: Removed.
- Basic/caseless_map.js: Removed.
- Basic/lexer.js: Removed.
- Basic/number.js: Removed.
- Basic/parser.js: Removed.
- Basic/random.js: Removed.
- Basic/state.js: Removed.
- Basic/test.html: Removed.
- Basic/test.js: Removed.
- Basic/util.js: Removed.
- ES6SampleBench: Added.
- ES6SampleBench/Air: Copied from PerformanceTests/Air.js.
- ES6SampleBench/Basic: Copied from PerformanceTests/Basic.
- Skipped:
- 2:35 PM Changeset in webkit [202590] by
-
- 4 edits in trunk/Source/WebCore
CrashTracer beneath JSC::MarkedBlock::specializedSweep
https://bugs.webkit.org/show_bug.cgi?id=159223
Reviewed by Saam Barati.
This crash is caused by a media element re-entering JS during the GC
sweep phase.
In theory, other CachedResourceClients in the DOM might also trigger
similar bugs, but our data only implicates the media elements, so this
fix targets them.
- html/HTMLDocument.h: Document has no reason to inherit from
CachedResourceClient. I found this becuase I had to search for all
CachedResourceClients in researching this patch.
- platform/graphics/avfoundation/cf/WebCoreAVCFResourceLoader.cpp:
(WebCore::WebCoreAVCFResourceLoader::invalidate): Delay our call to
stopLoading because it might re-enter JS, and we might have been called
by the GC sweep phase destroying a media element.
- platform/graphics/avfoundation/objc/WebCoreAVFResourceLoader.mm:
(WebCore::WebCoreAVFResourceLoader::invalidate): Ditto.
- 2:33 PM Changeset in webkit [202589] by
-
- 3 edits in trunk/Source/WebInspectorUI
REGRESSION (r188730): Web Inspector: Warning icons incorrectly positioned in CSS Rules sidebar
https://bugs.webkit.org/show_bug.cgi?id=157869
<rdar://problem/26356520>
Reviewed by Timothy Hatcher.
Before r188730, CSS text always had a line break as a prefix. r188730 started trimming text.
This patch keeps trimming unnecessary whitespace but brings back required line break prefix.
- UserInterface/Models/CSSStyleDeclaration.js:
(WebInspector.CSSStyleDeclaration.prototype.set text):
- UserInterface/Views/CSSStyleDeclarationTextEditor.js:
Make PrefixWhitespace and SuffixWhitespace public.
(WebInspector.CSSStyleDeclarationTextEditor.prototype._formattedContent):
(WebInspector.CSSStyleDeclarationTextEditor.prototype._updateTextMarkers.update):
(WebInspector.CSSStyleDeclarationTextEditor.prototype._updateTextMarkers):
Remove an if conditional as it's always true.
- 2:30 PM Changeset in webkit [202588] by
-
- 29 edits in trunk/Source
some Watchpoints' ::fireInternal method will call operations that might GC where the GC will cause the watchpoint itself to destruct
https://bugs.webkit.org/show_bug.cgi?id=159198
<rdar://problem/26302360>
Reviewed by Filip Pizlo.
Source/JavaScriptCore:
Firing a watchpoint may cause a GC to happen. This GC could destroy various
Watchpoints themselves while they're in the process of firing. It's not safe
for most Watchpoints to be destructed while they're in the middle of firing.
This GC could also destroy the WatchpointSet itself, and it's not in a safe
state to be destroyed. WatchpointSet::fireAllWatchpoints now defers gc for a
while. This prevents a GC from destructing any Watchpoints while they're
in the process of firing. This bug was being hit by the stress GC bots
because we would destruct a particular Watchpoint while it was firing,
and then we would access its field after it had already been destroyed.
This was causing all kinds of weird symptoms. Also, this was easier to
catch when running with guard malloc because the first access after
destruction would lead to a crash.
- bytecode/AdaptiveInferredPropertyValueWatchpointBase.cpp:
(JSC::AdaptiveInferredPropertyValueWatchpointBase::fire):
- bytecode/CodeBlock.cpp:
(JSC::CodeBlock::finishCreation):
- bytecode/VariableWriteFireDetail.cpp:
(JSC::VariableWriteFireDetail::dump):
(JSC::VariableWriteFireDetail::touch):
- bytecode/VariableWriteFireDetail.h:
- bytecode/Watchpoint.cpp:
(JSC::WatchpointSet::add):
(JSC::WatchpointSet::fireAllSlow):
(JSC::WatchpointSet::fireAllWatchpoints):
(JSC::InlineWatchpointSet::add):
(JSC::InlineWatchpointSet::fireAll):
(JSC::InlineWatchpointSet::inflateSlow):
- bytecode/Watchpoint.h:
(JSC::WatchpointSet::startWatching):
(JSC::WatchpointSet::fireAll):
(JSC::WatchpointSet::touch):
(JSC::WatchpointSet::invalidate):
(JSC::WatchpointSet::isBeingWatched):
(JSC::WatchpointSet::offsetOfState):
(JSC::WatchpointSet::addressOfSetIsNotEmpty):
(JSC::InlineWatchpointSet::startWatching):
(JSC::InlineWatchpointSet::fireAll):
(JSC::InlineWatchpointSet::invalidate):
(JSC::InlineWatchpointSet::touch):
- bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::BytecodeGenerator):
- dfg/DFGOperations.cpp:
- interpreter/Interpreter.cpp:
(JSC::Interpreter::execute):
- jit/JITOperations.cpp:
- jsc.cpp:
(WTF::Masquerader::create):
- llint/LLIntSlowPaths.cpp:
(JSC::LLInt::LLINT_SLOW_PATH_DECL):
- runtime/ArrayBufferNeuteringWatchpoint.cpp:
(JSC::ArrayBufferNeuteringWatchpoint::fireAll):
- runtime/FunctionRareData.cpp:
(JSC::FunctionRareData::clear):
- runtime/InferredType.cpp:
(JSC::InferredType::willStoreValueSlow):
(JSC::InferredType::makeTopSlow):
(JSC::InferredType::set):
(JSC::InferredType::removeStructure):
(JSC::InferredType::InferredStructureWatchpoint::fireInternal):
- runtime/InferredValue.cpp:
(JSC::InferredValue::notifyWriteSlow):
(JSC::InferredValue::ValueCleanup::finalizeUnconditionally):
- runtime/InferredValue.h:
(JSC::InferredValue::notifyWrite):
(JSC::InferredValue::invalidate):
- runtime/JSGlobalObject.cpp:
(JSC::JSGlobalObject::haveABadTime):
- runtime/JSSymbolTableObject.h:
(JSC::symbolTablePutTouchWatchpointSet):
(JSC::symbolTablePutInvalidateWatchpointSet):
- runtime/Structure.cpp:
(JSC::Structure::didCachePropertyReplacement):
(JSC::Structure::startWatchingInternalProperties):
(JSC::DeferredStructureTransitionWatchpointFire::~DeferredStructureTransitionWatchpointFire):
(JSC::DeferredStructureTransitionWatchpointFire::add):
(JSC::Structure::didTransitionFromThisStructure):
(JSC::Structure::prototypeForLookup):
- runtime/StructureInlines.h:
(JSC::Structure::didReplaceProperty):
(JSC::Structure::propertyReplacementWatchpointSet):
- runtime/SymbolTable.h:
(JSC::SymbolTableEntry::isDontEnum):
(JSC::SymbolTableEntry::disableWatching):
- runtime/VM.cpp:
(JSC::VM::addImpureProperty):
(JSC::enableProfilerWithRespectToCount):
Source/WebCore:
- bindings/js/JSDOMWindowBase.cpp:
(WebCore::JSDOMWindowBase::fireFrameClearedWatchpointsForWindow):
- bindings/scripts/CodeGeneratorJS.pm:
(GenerateHeader):
- bindings/scripts/test/JS/JSTestEventTarget.h:
(WebCore::JSTestEventTarget::create):
- 2:27 PM Changeset in webkit [202587] by
-
- 2 edits in trunk/LayoutTests
Rebaseline http/tests/media/video-auth.html for WK2
https://bugs.webkit.org/show_bug.cgi?id=159195
Unreviewed test gardening.
- platform/wk2/http/tests/media/video-auth-expected.txt:
- 2:10 PM Changeset in webkit [202586] by
-
- 4 edits in trunk/Source/WebKit2
Keep track of when a WKWebView is blank before the initial non-empty layout
https://bugs.webkit.org/show_bug.cgi?id=159217
<rdar://problem/26071766>
Reviewed by Beth Dakin.
- UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _initializeWithConfiguration:]):
(-[WKWebView _setIsBlankBeforeFirstNonEmptyLayout:]):
(-[WKWebView _didFirstVisuallyNonEmptyLayoutForMainFrame]):
- UIProcess/API/Cocoa/WKWebViewInternal.h:
- UIProcess/ios/PageClientImplIOS.mm:
(WebKit::PageClientImpl::didFirstVisuallyNonEmptyLayoutForMainFrame):
- 2:03 PM Changeset in webkit [202585] by
-
- 2 edits in trunk/Source/JavaScriptCore
JSRopeString should use release asserts, not debug asserts, about substring bounds
https://bugs.webkit.org/show_bug.cgi?id=159227
Reviewed by Saam Barati.
According to my experiments this change costs nothing. That's not surprising since the
most common way to construct a rope these days is inlined into the JIT, which does its own
safety checks. This makes us crash sooner rather than corrupting memory.
- runtime/JSString.h:
- 2:02 PM Changeset in webkit [202584] by
-
- 2 edits in trunk/Source/WebCore
Move the user gesture requirement to the ApplePaySession constructor
https://bugs.webkit.org/show_bug.cgi?id=159225
rdar://problem/26507267
Reviewed by Tim Horton.
By doing this, clients can do pre-validation before showing the sheet, while we still maintain the user gesture requirement.
- Modules/applepay/ApplePaySession.cpp:
(WebCore::ApplePaySession::create):
(WebCore::ApplePaySession::begin): Deleted.
- 1:19 PM Changeset in webkit [202583] by
-
- 7 edits in trunk
Iterable interfaces should have their related prototype @@iterator property writable
https://bugs.webkit.org/show_bug.cgi?id=159211
Source/WebCore:
<rdar://problem/26950766>
Patch by Youenn Fablet <youenn@apple.com> on 2016-06-28
Reviewed by Chris Dumez.
Updating @@iterator property according http://heycam.github.io/webidl/#es-iterator.
Covered by updated test.
- bindings/scripts/CodeGeneratorJS.pm:
(GenerateImplementation): Removing ReadOnly flag from @@iterator property of iterable interfaces.
- bindings/scripts/test/JS/JSTestNode.cpp:
(WebCore::JSTestNodePrototype::finishCreation): Rebasing expectation.
- bindings/scripts/test/JS/JSTestObj.cpp:
(WebCore::JSTestObjPrototype::finishCreation): Ditto.
LayoutTests:
Patch by Youenn Fablet <youenn@apple.com> on 2016-06-28
Reviewed by Chris Dumez.
- fast/dom/nodeListIterator-expected.txt:
- fast/dom/nodeListIterator.html: Overriding NodeList @@iterator by Array one and checking everything is fine.
- 12:50 PM Changeset in webkit [202582] by
-
- 2 edits in trunk/Source/WebCore
"Total amount is too big" error message is displaying on clicking Pay button
https://bugs.webkit.org/show_bug.cgi?id=159219
rdar://problem/26722110
Reviewed by Tim Horton.
Match the PassKit max amount.
- Modules/applepay/PaymentRequestValidator.cpp:
(WebCore::PaymentRequestValidator::validateTotal):
- 12:31 PM Changeset in webkit [202581] by
-
- 12 edits1 copy in trunk/Source
PaymentMerchantSession should wrap a PKPaymentMerchantSession
https://bugs.webkit.org/show_bug.cgi?id=159218
rdar://problem/26872118
Reviewed by Tim Horton.
Source/WebCore:
- Modules/applepay/ApplePaySession.cpp:
(WebCore::ApplePaySession::completeMerchantValidation):
Use PaymentMerchantSession::fromJS.
(WebCore::createMerchantSession): Deleted.
- Modules/applepay/PaymentCoordinator.h:
PaymentMerchantSession is now a class.
- Modules/applepay/PaymentCoordinatorClient.h:
PaymentMerchantSession is now a class.
- Modules/applepay/PaymentMerchantSession.h:
(WebCore::PaymentMerchantSession::PaymentMerchantSession):
(WebCore::PaymentMerchantSession::~PaymentMerchantSession):
(WebCore::PaymentMerchantSession::pkPaymentMerchantSession):
Store a PKPaymentMerchantSession in a RetainPtr inside the PaymentMerchantSession object.
- Modules/applepay/cocoa/PaymentMerchantSessionCocoa.mm:
(WebCore::PaymentMerchantSession::fromJS):
Convert the JS object to a PKPaymentMerchantSession and return a PaymentMerchantSession that wraps it.
- WebCore.xcodeproj/project.pbxproj:
Add new files.
- bindings/js/Dictionary.h:
(WebCore::Dictionary::initializerObject):
Add new getter.
Source/WebKit2:
- Shared/Cocoa/WebCoreArgumentCodersCocoa.mm:
(IPC::ArgumentCoder<WebCore::PaymentMerchantSession>::encode):
(IPC::ArgumentCoder<WebCore::PaymentMerchantSession>::decode):
Use NSCoder.
- UIProcess/ApplePay/WebPaymentCoordinatorProxy.h:
PaymentMerchantSession is now a class.
- UIProcess/ApplePay/WebPaymentCoordinatorProxy.messages.in:
PaymentMerchantSession is now a class.
- UIProcess/ApplePay/cocoa/WebPaymentCoordinatorProxyCocoa.mm:
(WebKit::WebPaymentCoordinatorProxy::platformCompleteMerchantValidation):
Just get the underlying PKPaymentMerchantSession from the PaymentMerchantSession object.
(WebKit::toPKPaymentMerchantSession): Deleted.
- 12:10 PM Changeset in webkit [202580] by
-
- 89 edits in trunk
RunLoop::Timer should use constructor templates instead of class templates
https://bugs.webkit.org/show_bug.cgi?id=159153
Reviewed by Alex Christensen.
Source/JavaScriptCore:
Remove the RunLoop::Timer class template argument, and pass its constructor
a reference tothisinstead of a pointer tothis.
- inspector/agents/InspectorHeapAgent.cpp:
(Inspector::SendGarbageCollectionEventsTask::SendGarbageCollectionEventsTask):
Source/WebCore:
Remove the RunLoop::Timer class template argument, and pass its constructor
a reference tothisinstead of a pointer tothis.
- Modules/mediasession/WebMediaSessionManager.cpp:
(WebCore::WebMediaSessionManager::WebMediaSessionManager):
- Modules/mediasession/WebMediaSessionManager.h:
- page/WheelEventTestTrigger.cpp:
(WebCore::WheelEventTestTrigger::WheelEventTestTrigger):
- page/WheelEventTestTrigger.h:
- page/mac/TextIndicatorWindow.h:
- page/mac/TextIndicatorWindow.mm:
(WebCore::TextIndicatorWindow::TextIndicatorWindow):
- platform/MainThreadSharedTimer.h:
- platform/cocoa/ScrollController.h:
- platform/cocoa/ScrollController.mm:
(WebCore::ScrollController::ScrollController):
- platform/glib/MainThreadSharedTimerGLib.cpp:
(WebCore::MainThreadSharedTimer::MainThreadSharedTimer):
- platform/graphics/MediaPlaybackTargetPicker.cpp:
(WebCore::MediaPlaybackTargetPicker::MediaPlaybackTargetPicker):
- platform/graphics/MediaPlaybackTargetPicker.h:
- platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:
(WebCore::MediaPlayerPrivateGStreamer::MediaPlayerPrivateGStreamer):
- platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.h:
- platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.cpp:
(WebCore::MediaPlayerPrivateGStreamerBase::MediaPlayerPrivateGStreamerBase):
- platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.h:
- platform/graphics/gstreamer/VideoSinkGStreamer.cpp:
(VideoRenderRequestScheduler::VideoRenderRequestScheduler):
- platform/graphics/texmap/TextureMapperPlatformLayerProxy.cpp:
(WebCore::TextureMapperPlatformLayerProxy::TextureMapperPlatformLayerProxy):
(WebCore::TextureMapperPlatformLayerProxy::activateOnCompositingThread):
- platform/graphics/texmap/TextureMapperPlatformLayerProxy.h:
- platform/mock/MediaPlaybackTargetPickerMock.cpp:
(WebCore::MediaPlaybackTargetPickerMock::MediaPlaybackTargetPickerMock):
- platform/mock/MediaPlaybackTargetPickerMock.h:
- platform/mock/MockRealtimeVideoSource.cpp:
(WebCore::MockRealtimeVideoSource::MockRealtimeVideoSource):
- platform/mock/MockRealtimeVideoSource.h:
- platform/network/ResourceHandleInternal.h:
(WebCore::ResourceHandleInternal::ResourceHandleInternal):
Source/WebKit2:
Remove the RunLoop::Timer class template argument, and pass its constructor
a reference tothisinstead of a pointer tothis.
- NetworkProcess/Downloads/soup/DownloadSoup.cpp:
- Platform/IPC/mac/ConnectionMac.mm:
(IPC::ConnectionTerminationWatchdog::ConnectionTerminationWatchdog):
- PluginProcess/PluginControllerProxy.cpp:
(WebKit::PluginControllerProxy::PluginControllerProxy):
- PluginProcess/PluginControllerProxy.h:
- PluginProcess/PluginProcess.cpp:
(WebKit::PluginProcess::PluginProcess):
- PluginProcess/PluginProcess.h:
- Shared/ChildProcess.cpp:
(WebKit::ChildProcess::ChildProcess):
- Shared/ChildProcess.h:
- Shared/CoordinatedGraphics/threadedcompositor/CompositingRunLoop.h:
- UIProcess/API/Cocoa/APISerializedScriptValueCocoa.mm:
(API::SharedJSContext::SharedJSContext):
- UIProcess/API/gtk/WebKitWebViewBase.cpp:
(_WebKitWebViewBasePrivate::_WebKitWebViewBasePrivate):
- UIProcess/DrawingAreaProxy.cpp:
(WebKit::DrawingAreaProxy::DrawingAreaProxy):
- UIProcess/DrawingAreaProxy.h:
- UIProcess/DrawingAreaProxyImpl.cpp:
(WebKit::DrawingAreaProxyImpl::DrawingAreaProxyImpl):
- UIProcess/DrawingAreaProxyImpl.h:
- UIProcess/Plugins/gtk/PluginInfoCache.cpp:
(WebKit::PluginInfoCache::PluginInfoCache):
- UIProcess/Plugins/gtk/PluginInfoCache.h:
- UIProcess/ProcessThrottler.cpp:
(WebKit::ProcessThrottler::ProcessThrottler):
- UIProcess/ProcessThrottler.h:
- UIProcess/ResponsivenessTimer.cpp:
(WebKit::ResponsivenessTimer::ResponsivenessTimer):
- UIProcess/ResponsivenessTimer.h:
- UIProcess/ViewGestureController.cpp:
(WebKit::ViewGestureController::ViewGestureController):
(WebKit::ViewGestureController::SnapshotRemovalTracker::SnapshotRemovalTracker):
- UIProcess/VisitedLinkStore.cpp:
(WebKit::VisitedLinkStore::VisitedLinkStore):
- UIProcess/VisitedLinkStore.h:
- UIProcess/WebInspectorProxy.cpp:
(WebKit::WebInspectorProxy::WebInspectorProxy):
- UIProcess/WebInspectorProxy.h:
- UIProcess/WebProcessPool.cpp:
(WebKit::m_hiddenPageThrottlingTimer):
- UIProcess/WebProcessPool.h:
- UIProcess/gtk/GestureController.cpp:
(WebKit::GestureController::DragGesture::DragGesture):
(WebKit::GestureController::ZoomGesture::ZoomGesture):
- UIProcess/gtk/GestureController.h:
- UIProcess/gtk/WebPopupMenuProxyGtk.cpp:
(WebKit::WebPopupMenuProxyGtk::WebPopupMenuProxyGtk):
- UIProcess/gtk/WebPopupMenuProxyGtk.h:
- UIProcess/mac/ViewGestureController.h:
- WebProcess/Network/WebLoaderStrategy.cpp:
(WebKit::WebLoaderStrategy::WebLoaderStrategy):
- WebProcess/Network/WebLoaderStrategy.h:
- WebProcess/Plugins/Netscape/NPRuntimeObjectMap.cpp:
(WebKit::NPRuntimeObjectMap::NPRuntimeObjectMap):
- WebProcess/Plugins/Netscape/NPRuntimeObjectMap.h:
- WebProcess/Plugins/Netscape/NetscapePlugin.cpp:
(WebKit::NetscapePlugin::NetscapePlugin):
(WebKit::NetscapePlugin::Timer::Timer):
- WebProcess/Plugins/Netscape/NetscapePlugin.h:
- WebProcess/Plugins/Netscape/NetscapePluginStream.cpp:
(WebKit::NetscapePluginStream::NetscapePluginStream):
- WebProcess/Plugins/Netscape/NetscapePluginStream.h:
- WebProcess/Plugins/PluginView.cpp:
(WebKit::PluginView::PluginView):
- WebProcess/Plugins/PluginView.h:
- WebProcess/WebPage/CoordinatedGraphics/CoordinatedDrawingArea.cpp:
(WebKit::CoordinatedDrawingArea::CoordinatedDrawingArea):
- WebProcess/WebPage/CoordinatedGraphics/CoordinatedDrawingArea.h:
- WebProcess/WebPage/CoordinatedGraphics/ThreadedCoordinatedLayerTreeHost.cpp:
(WebKit::ThreadedCoordinatedLayerTreeHost::ThreadedCoordinatedLayerTreeHost):
- WebProcess/WebPage/CoordinatedGraphics/ThreadedCoordinatedLayerTreeHost.h:
- WebProcess/WebPage/DrawingAreaImpl.cpp:
(WebKit::DrawingAreaImpl::DrawingAreaImpl):
- WebProcess/WebPage/DrawingAreaImpl.h:
- WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::WebPage):
- WebProcess/WebPage/WebPage.h:
- WebProcess/WebPage/gtk/LayerTreeHostGtk.cpp:
(WebKit::LayerTreeHostGtk::RenderFrameScheduler::RenderFrameScheduler):
- WebProcess/WebPage/gtk/LayerTreeHostGtk.h:
- WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.h:
- WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.mm:
(WebKit::TiledCoreAnimationDrawingArea::TiledCoreAnimationDrawingArea):
Source/WTF:
Refactor RunLoop::Timer to align with WebCore::Timer. Use a constructor
template instead of a class template. Store a bound std::function rather than
a templated member and function. Add a constructor that takes a std::function.
- wtf/RunLoop.h:
(WTF::RunLoop::Timer::Timer):
(WTF::RunLoop::Timer::fired):
Tools:
Remove the RunLoop::Timer class template argument, and pass its constructor
a reference tothisinstead of a pointer tothis.
- WebKitTestRunner/InjectedBundle/TestRunner.cpp:
(WTR::TestRunner::TestRunner):
- WebKitTestRunner/InjectedBundle/TestRunner.h:
- TestWebKitAPI/Tests/WTF/RunLoop.cpp:
(TestWebKitAPI::TEST):
- 11:25 AM Changeset in webkit [202579] by
-
- 5 edits in trunk
Cross-domain video loads do not prompt for authorization.
https://bugs.webkit.org/show_bug.cgi?id=159195
<rdar://problem/26234612>
Reviewed by Brent Fulgham.
Source/WebCore:
Test: http/tests/media/video-auth.html (modified)
We should prompt for authorization when a cross-origin <video> is embedded
in a web page.
- loader/MediaResourceLoader.cpp:
(WebCore::MediaResourceLoader::requestResource):
LayoutTests:
Add a cross-origin authorization sub-test.
- http/tests/media/video-auth.html:
- http/tests/media/video-auth-expected.txt:
- 11:20 AM Changeset in webkit [202578] by
-
- 6 edits in trunk
REGRESSION(r201471): FormClient.textFieldDidEndEditing is no longer called when a text field is removed
https://bugs.webkit.org/show_bug.cgi?id=159199
Source/WebCore:
<rdar://problem/26748189>
Reviewed by Alexey Proskuryakov.
The bug was caused by HTMLInputElement's endEditing no longer getting called due to the behavior change.
Preserve the WebKit2 API semantics by manually calling HTMLInputElement::endEditing in setFocusedElement.
Tests: WebKit2TextFieldDidBeginAndEndEditing
- dom/Document.cpp:
(WebCore::Document::setFocusedElement):
Tools:
Reviewed by Alexey Proskuryakov.
Added a test case for removing a text field. Also fixed the flakiness and re-enabled it on Mac.
- TestWebKitAPI/Tests/WebKit2/TextFieldDidBeginAndEndEditing.cpp:
- TestWebKitAPI/Tests/WebKit2/TextFieldDidBeginAndEndEditing_Bundle.cpp:
- TestWebKitAPI/Tests/WebKit2/input-focus-blur.html: Focus a div upfront to avoid the flakiness from
an input element getting automatically focused on Mac.
- 11:12 AM Changeset in webkit [202577] by
-
- 2 edits1 add in trunk/Websites/webkit.org
Fixed Open Graph meta data and image entries.
https://bugs.webkit.org/show_bug.cgi?id=159167.
Reviewed by Timothy Hatcher.
This patch addresses Open Graph meta data issues:
- Added a high-resolution PNG of the WebKit logo because Facebook doesn't support SVG images
- Uses the first image of a page or post if not featured image is set
- Article date/time information should not be given for the homepage
- Added Facebook admins as per the Facebook Open Graph debugger
- wp-content/plugins/social-meta.php:
- wp-content/themes/webkit/images/ogimage.png: Added.
- 11:08 AM Changeset in webkit [202576] by
-
- 2 edits in trunk/Source/WebKit2
[WK2] Add some logging related to NetworkProcess suspension
https://bugs.webkit.org/show_bug.cgi?id=159178
Reviewed by Andreas Kling.
Add some logging related to NetworkProcess suspension to help debug
suspension issues.
- NetworkProcess/NetworkProcess.cpp:
(WebKit::NetworkProcess::prepareToSuspend):
(WebKit::NetworkProcess::cancelPrepareToSuspend):
(WebKit::NetworkProcess::processDidResume):
- 11:08 AM Changeset in webkit [202575] by
-
- 5 edits in trunk/Source/WebKit2
Add some more UIProcess-side logging related to process assertions
https://bugs.webkit.org/show_bug.cgi?id=159188
Reviewed by Andreas Kling.
Add some more UIProcess-side logging related to process assertions
to help debug process suspension issues.
- UIProcess/Network/NetworkProcessProxy.cpp:
(WebKit::NetworkProcessProxy::fetchWebsiteData):
(WebKit::NetworkProcessProxy::deleteWebsiteData):
(WebKit::NetworkProcessProxy::deleteWebsiteDataForOrigins):
- UIProcess/ProcessThrottler.cpp:
(WebKit::ProcessThrottler::updateAssertionNow):
(WebKit::ProcessThrottler::updateAssertion):
- UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::updateActivityToken):
- UIProcess/WebProcessProxy.cpp:
(WebKit::WebProcessProxy::fetchWebsiteData):
(WebKit::WebProcessProxy::deleteWebsiteData):
(WebKit::WebProcessProxy::deleteWebsiteDataForOrigins):
(WebKit::WebProcessProxy::didSetAssertionState):
- 11:07 AM Changeset in webkit [202574] by
-
- 2 edits in trunk/Source/WebKit2
[WK2] Do not suppress navigation snapshotting on session restore if we are not navigating right away
https://bugs.webkit.org/show_bug.cgi?id=159216
<rdar://problem/27058360>
Reviewed by Andreas Kling.
Do not suppress navigation snapshotting on session restore if we are not navigating
right away. The snapshot we already have may be outdated by the time we actually
navigate (e.g. Because the user scrolled) so we don't want to suppress snapshotting
in this case.
- UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::restoreFromSessionState):
- 10:23 AM Changeset in webkit [202573] by
-
- 3 edits in trunk/LayoutTests
Marking fast/multicol/fixed-stack.html as failing on ios-simulator
Unreviewed test gardening.
- platform/ios-simulator-wk1/TestExpectations:
- platform/ios-simulator/TestExpectations:
- 10:01 AM Changeset in webkit [202572] by
-
- 5 edits in trunk
Phrasing content should be accepted in <mo> elements
https://bugs.webkit.org/show_bug.cgi?id=130245
Patch by Frederic Wang <fwang@igalia.com> on 2016-06-28
Reviewed by Brent Fulgham.
Source/WebCore:
After r202420, the RenderMathMLOperator element no longer messes with anonymous block and
text nodes. Hence it is now safe to allow foreign content inside <mo>.
We extend foreign-element-in-token.html to cover the mo case.
- mathml/MathMLTextElement.cpp:
(WebCore::MathMLTextElement::childShouldCreateRenderer): Remove the early return for <mo> so
that it accepts phrasing content children.
LayoutTests:
- mathml/presentation/foreign-element-in-token.html: Copy test cases to
check <mo> elements too.
- mathml/presentation/foreign-element-in-token-expected.txt: Add the
expectations for <mo>.
- 9:53 AM Changeset in webkit [202571] by
-
- 2 edits2 moves1 add in trunk/LayoutTests
Sierra test expectation gardening.
- platform/mac-elcapitan/fast/canvas: Added.
- platform/mac-elcapitan/fast/canvas/canvas-strokePath-gradient-shadow-expected.txt: Copied from platform/mac/fast/canvas/canvas-strokePath-gradient-shadow-expected.txt.
- platform/mac-elcapitan/fast/canvas/canvas-strokeRect-gradient-shadow-expected.txt: Copied from platform/mac/fast/canvas/canvas-strokeRect-gradient-shadow-expected.txt.
- platform/mac/fast/canvas/canvas-strokePath-gradient-shadow-expected.txt: Removed.
- platform/mac/fast/canvas/canvas-strokeRect-gradient-shadow-expected.txt: Removed.
Sierra matches cross-platform expectations.
- platform/mac/TestExpectations: Removed expectations for fast/canvas/canvas-strokeRect-alpha-shadow.html
- 9:45 AM Changeset in webkit [202570] by
-
- 10 edits in trunk/Source
WebKit::WebPaymentCoordinator leak
https://bugs.webkit.org/show_bug.cgi?id=159168
rdar://problem/26929772
Reviewed by Beth Dakin.
Source/WebCore:
- Modules/applepay/PaymentCoordinator.cpp:
(WebCore::PaymentCoordinator::~PaymentCoordinator):
Call paymentCoordinatorDestroyed().
- Modules/applepay/PaymentCoordinatorClient.h:
Rename mainFrameDestroyed to paymentCoordinatorDestroyed().
- loader/EmptyClients.cpp:
Source/WebKit/mac:
Update for WebCore changes.
- WebCoreSupport/WebPaymentCoordinatorClient.h:
- WebCoreSupport/WebPaymentCoordinatorClient.mm:
(WebPaymentCoordinatorClient::paymentCoordinatorDestroyed):
(WebPaymentCoordinatorClient::mainFrameDestroyed): Deleted.
Source/WebKit2:
Update for WebCore changes.
- WebProcess/ApplePay/WebPaymentCoordinator.cpp:
(WebKit::WebPaymentCoordinator::paymentCoordinatorDestroyed):
(WebKit::WebPaymentCoordinator::mainFrameDestroyed): Deleted.
- WebProcess/ApplePay/WebPaymentCoordinator.h:
- 8:36 AM Changeset in webkit [202569] by
-
- 4 edits in trunk/Source/WebCore
Remove anonymous in renderName for all MathML renderers but RenderMathMLOperator
https://bugs.webkit.org/show_bug.cgi?id=159114
Patch by Frederic Wang <fwang@igalia.com> on 2016-06-28
Reviewed by Martin Robinson.
After r202420, the only anonymous MathML renderers are the RenderMathMLOperators created by
the mfenced element. Hence we remove the special case for anonymous in the renderName
implementation of most MathML renderers.
No new tests, behavior unchanged.
- rendering/mathml/RenderMathMLRow.h:
- rendering/mathml/RenderMathMLSpace.h:
- rendering/mathml/RenderMathMLToken.h:
- 8:33 AM Changeset in webkit [202568] by
-
- 6 edits in trunk
Web Inspector: selectElement.options shows unexpected entries in console (named indexes beyond collection length)
https://bugs.webkit.org/show_bug.cgi?id=159192
Patch by Joseph Pecoraro <Joseph Pecoraro> on 2016-06-28
Reviewed by Timothy Hatcher.
Source/JavaScriptCore:
- inspector/InjectedScriptSource.js:
(InjectedScript.prototype.arrayIndexPropertyNames):
Start with an empty array because we just push valid indexes.
(InjectedScript.prototype._propertyDescriptors):
Avoid the >100 length requirement, and always treat the
array-like objects the same. The frontend currently
doesn't show named indexes for arrays anyways, so they
would have been unused.
LayoutTests:
- inspector/model/remote-object-get-properties-expected.txt:
- inspector/model/remote-object-get-properties.html:
- inspector/runtime/getProperties-expected.txt:
- 8:32 AM Changeset in webkit [202567] by
-
- 2 edits in trunk/Source/JavaScriptCore
[Win] Skip failing INTL test.
https://bugs.webkit.org/show_bug.cgi?id=159141
Reviewed by Brent Fulgham.
INTL is not enabled on Windows.
- tests/stress/intl-constructors-with-proxy.js:
(shouldBe):
- 8:12 AM Changeset in webkit [202566] by
-
- 7 edits in trunk
Web Inspector: QuickConsole should update its selection when RuntimeManager.defaultExecutionContextIdentifier changes
https://bugs.webkit.org/show_bug.cgi?id=159183
Reviewed by Timothy Hatcher.
Source/WebInspectorUI:
Currently, the UI listens for state changes in the Frame tree to decide when to reset
the selected execution context back to its default value. This is no good, because this
should happen even if we have no UI (i.e., testing models only). The UI should simply
display changes to the model rather than driving them based on other model changes.
Move the logic to reset the execution context into RuntimeManager. When the selected
context changes, an event is fired which causes the QuickConsole to rebuild its path components.
- UserInterface/Controllers/RuntimeManager.js:
(WebInspector.RuntimeManager):
(WebInspector.RuntimeManager.prototype.set defaultExecutionContextIdentifier):
Fire an event when the execution context actually changes.
(WebInspector.RuntimeManager.prototype._frameExecutionContextsCleared):
Added. If the selected context was cleared, reset back to default.
- UserInterface/Models/Frame.js:
(WebInspector.Frame.prototype.clearExecutionContexts):
Include the contexts that were cleared so clients can match against them.
- UserInterface/Views/QuickConsole.js:
(WebInspector.QuickConsole):
Use the proper constant name. Both the old and new names evaluated toundefined.
No need to keep track of the selected path component, it will always match the
defaultExecutionContextIdentifier in RuntimeManager.
(WebInspector.QuickConsole.prototype.get selectedExecutionContextIdentifier):
(WebInspector.QuickConsole.prototype.set selectedExecutionContextIdentifier):
Forward to RuntimeManager. This name is less awkward for the UI code that manages selections.
(WebInspector.QuickConsole.prototype._executionContextPathComponentsToDisplay):
Special-case for the main frame execution context.
(WebInspector.QuickConsole.prototype._framePageExecutionContextsChanged):
Remove indirection.
(WebInspector.QuickConsole.prototype._frameExecutionContextsCleared):
Fix the guard to handle undefined execution contexts in the case where it represents the main frame (undefined).
(WebInspector.QuickConsole.prototype._defaultExecutionContextChanged):
Rebuild when the model changes.
(WebInspector.QuickConsole.prototype._pathComponentSelected): Simplify.
(WebInspector.QuickConsole.prototype.get executionContextIdentifier):
(WebInspector.QuickConsole.prototype._removeExecutionContextPathComponentForFrame):
Move the fallback selection behavior into RuntimeManager.
(WebInspector.QuickConsole.prototype._updateExecutionContextPathComponentForFrame): Deleted.
This has been dead code ever since we removed iOS 6 legacy support.
LayoutTests:
Add a new test case for reverting to the top-level execution context when the selected context is destroyed.
- inspector/runtime/change-execution-context-identifier-expected.txt:
- inspector/runtime/change-execution-context-identifier.html:
- 8:10 AM Changeset in webkit [202565] by
-
- 6 edits in trunk
WebRTC: Robustify 'this' type check in RTCPeerConnection JS built-ins
https://bugs.webkit.org/show_bug.cgi?id=158831
Reviewed by Youenn Fablet.
Source/WebCore:
Use @operations slot in RTCPeerConnection type check.
Updated results of existing test.
- Modules/mediastream/RTCPeerConnection.js:
(initializeRTCPeerConnection):
Initialize @operations slot in constructor.
- Modules/mediastream/RTCPeerConnectionInternals.js:
(isRTCPeerConnection):
Use @operations slot in type check.
LayoutTests:
- fast/mediastream/RTCPeerConnection-js-built-ins-check-this-expected.txt:
Flip some result lines from FAIL to PASS. Clean out unnecessary tests (third set).
- 8:05 AM Changeset in webkit [202564] by
-
- 4 edits in trunk/LayoutTests
Remove flakiness expectations for mathml/wbr-in-mroot-crash.html
https://bugs.webkit.org/show_bug.cgi?id=130353
Unreviewed gardening.
The flakiness seemed to be due to the timeout of mathml/very-large-stretchy-operators.html.
The latter was fixed in r202489, so we update the test expectations.
Patch by Frederic Wang <fwang@igalia.com> on 2016-06-28
- platform/efl/TestExpectations:
- platform/gtk/TestExpectations:
- platform/ios-simulator-wk2/TestExpectations:
- 7:43 AM Changeset in webkit [202563] by
-
- 4 edits in trunk/LayoutTests
Increase opacity of stretched operators in mo-stacked-glyphs.html
https://bugs.webkit.org/show_bug.cgi?id=159201
Patch by Frederic Wang <fwang@igalia.com> on 2016-06-28
Reviewed by Martin Robinson.
The reftest mo-stacked-glyphs.html fails with small pixel differences invisible at the naked
eyes. We increase opacity of stretched operators in that test to reduce these differences and
make the test pass.
- mathml/presentation/mo-stacked-glyphs-expected.html: Increase opacity of stretched operators.
- mathml/presentation/mo-stacked-glyphs.html: Ditto.
- platform/gtk/TestExpectations: Remove failure expectation.
- 7:16 AM Changeset in webkit [202562] by
-
- 2 edits in trunk/Source/WebCore
AX: Remove dead code in AccessibilityRenderObject::textUnderElement
https://bugs.webkit.org/show_bug.cgi?id=159205
Patch by Frederic Wang <fwang@igalia.com> on 2016-06-28
Reviewed by Joanmarie Diggs.
RenderMathMLOperator used to destroy its descendants and to replace them with an anonymous
text node wrapped inside anonymous blocks. After r202420, it just behaves as any other token
elements. Hence we remove the code in AccessibilityRenderObject::textUnderElement that was
used to handle this specific render tree structure.
No new tests, already covered by accessibility/math-text.html.
- accessibility/AccessibilityRenderObject.cpp:
(WebCore::AccessibilityRenderObject::textUnderElement): Remove dead code for RenderText and RenderMathMLOperator.
- 7:10 AM Changeset in webkit [202561] by
-
- 2 edits in trunk/LayoutTests
[GTK] Update expectations for subtle crypto layout tests
Unreviewed.
- platform/gtk/TestExpectations:
- 6:32 AM WebKitGTK/2.12.x edited by
- (diff)
- 6:31 AM WebKitGTK/2.12.x edited by
- (diff)
- 6:25 AM Changeset in webkit [202560] by
-
- 2 edits in trunk/Source/WTF
THUMB2 support not correctly detected on Fedora with GCC 6.1.
https://bugs.webkit.org/show_bug.cgi?id=159083
Patch by Tomas Popela <tpopela@redhat.com> on 2016-06-28
Reviewed by Carlos Garcia Campos.
On Fedora 24 with GCC 6.1. the thumb2 and thumb are not
defined so the detection of THUMB2 support will fail. Look also
whether the ARM_ARCH_ISA_THUMB is defined to fix the THUMB2
detection.
- wtf/Platform.h:
- 4:49 AM Changeset in webkit [202559] by
-
- 22 edits in trunk
[Win] Custom elements tests are failing.
https://bugs.webkit.org/show_bug.cgi?id=159139
Reviewed by Alex Christensen.
.:
Enable custom element API on Windows.
- Source/cmake/OptionsWin.cmake:
Source/WebCore:
Fix compile errors after enabling custom element API.
- bindings/js/JSHTMLElementCustom.cpp:
(WebCore::constructJSHTMLElement):
- dom/CustomElementDefinitions.cpp:
(WebCore::CustomElementDefinitions::addElementDefinition):
- dom/Document.cpp:
(WebCore::createHTMLElementWithNameValidation):
(WebCore::createFallbackHTMLElement):
- dom/Element.cpp:
(WebCore::Element::attributeChanged):
- dom/LifecycleCallbackQueue.cpp:
(WebCore::LifecycleQueueItem::LifecycleQueueItem):
(WebCore::LifecycleCallbackQueue::enqueueElementUpgrade):
(WebCore::LifecycleCallbackQueue::enqueueAttributeChangedCallback):
- html/parser/HTMLConstructionSite.cpp:
(WebCore::HTMLConstructionSite::insertHTMLElementOrFindCustomElementInterface):
(WebCore::HTMLConstructionSite::createHTMLElementOrFindCustomElementInterface):
- html/parser/HTMLDocumentParser.cpp:
(WebCore::HTMLDocumentParser::runScriptsForPausedTreeBuilder):
- html/parser/HTMLTreeBuilder.cpp:
(WebCore::CustomElementConstructionData::CustomElementConstructionData):
(WebCore::HTMLTreeBuilder::insertGenericHTMLElement):
- html/parser/HTMLTreeBuilder.h:
Source/WebKit/win:
Add preference for enabling custom element API.
- Interfaces/IWebPreferencesPrivate.idl:
- WebPreferenceKeysPrivate.h:
- WebPreferences.cpp:
(WebPreferences::initializeDefaultSettings):
(WebPreferences::customElementsEnabled):
(WebPreferences::setCustomElementsEnabled):
- WebPreferences.h:
- WebView.cpp:
(WebView::notifyPreferencesChanged):
Tools:
Enable custom element API when running tests.
- DumpRenderTree/win/DumpRenderTree.cpp:
(resetWebPreferencesToConsistentValues):
LayoutTests:
Update test expectations for passing custom elements tests.
- platform/win/TestExpectations:
- 4:26 AM Changeset in webkit [202558] by
-
- 4 edits in trunk
[GStreamer] usec rounding is wrong during accurate seeking
https://bugs.webkit.org/show_bug.cgi?id=90734
Reviewed by Carlos Garcia Campos.
Use floor() to round the microseconds value, this is more robust
than roundf.
- platform/graphics/gstreamer/GStreamerUtilities.cpp:
(WebCore::toGstClockTime):
- platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:
(WebCore::MediaPlayerPrivateGStreamer::playbackPosition):
- 4:24 AM Changeset in webkit [202557] by
-
- 6 edits in trunk/LayoutTests
[GTK] Web audio tests failing since GStreamer 1.6 upgrade
https://bugs.webkit.org/show_bug.cgi?id=158927
Unreviewed GTK rebaseline of webaudio tests.
- platform/gtk/TestExpectations:
- platform/gtk/webaudio/codec-tests/aac/vbr-128kbps-44khz-expected.wav:
- platform/gtk/webaudio/codec-tests/vorbis/vbr-128kbps-44khz-expected.wav:
- platform/gtk/webaudio/codec-tests/vorbis/vbr-70kbps-44khz-expected.wav:
- platform/gtk/webaudio/codec-tests/vorbis/vbr-96kbps-44khz-expected.wav:
- 3:20 AM Changeset in webkit [202556] by
-
- 2 edits in trunk/Source/WebCore
[GStreamer] improved duration query support in the HTTP source element
https://bugs.webkit.org/show_bug.cgi?id=159204
Reviewed by Carlos Garcia Campos.
When we have the Content-Length value it is possible to infer the TIME
duration in most cases by performing a convert query in the downstream
elements. This is especially useful when the duration query wasn't
managed by the sinks and thus reached the source element.
- platform/graphics/gstreamer/WebKitWebSourceGStreamer.cpp:
(webKitWebSrcQueryWithParent):
- 3:15 AM Changeset in webkit [202555] by
-
- 2 edits in trunk/Source/JavaScriptCore
[JSC] Fix build break since r202502 - 2
https://bugs.webkit.org/show_bug.cgi?id=159194
Reviewed by Gyuyoung Kim.
Fix about the error message below.
error: control reaches end of non-void function [-Werror=return-type]
- b3/B3TypeMap.h: add #pragma GCC diagnostic ignored "-Wreturn-type".
- 12:17 AM Changeset in webkit [202554] by
-
- 3 edits in trunk/LayoutTests
http/tests/media/hls/video-controller-getStartDate.html is failing on non-PDT time-based computers
https://bugs.webkit.org/show_bug.cgi?id=159134
Patch by Youenn Fablet <youenn@apple.com> on 2016-06-28
Reviewed by Alex Christensen.
- http/tests/media/hls/video-controller-getStartDate-expected.txt: Converting date to UTC string before making the comparison.
- http/tests/media/hls/video-controller-getStartDate.html: Rebasing.
- 12:10 AM Changeset in webkit [202553] by
-
- 2 edits in trunk/Source/JavaScriptCore
[JSC] Fix build break since r202502
https://bugs.webkit.org/show_bug.cgi?id=159194
Reviewed by Alex Christensen.
Fix about the error message below.
error: control reaches end of non-void function [-Werror=return-type]
- b3/B3TypeMap.h:
(JSC::B3::TypeMap::at): add missing ASSERT_NOT_REACHED().
- 12:09 AM Changeset in webkit [202552] by
-
- 13 edits10 adds in trunk/LayoutTests
Unreviewed EFL Gardening on 28th Jun.
- platform/efl/accessibility/aria-hidden-negates-no-visibility-expected.txt: Added.
- platform/efl/accessibility/aria-labelledby-overrides-aria-label-expected.txt: Added.
- platform/efl/accessibility/aria-labelledby-overrides-label-expected.txt: Added.
- platform/efl/accessibility/aria-option-role-expected.txt: Added.
- platform/efl/accessibility/aria-role-on-label-expected.txt: Added.
- platform/efl/accessibility/aria-switch-text-expected.txt:
- platform/efl/accessibility/auto-fill-crash-expected.txt: Added.
- platform/efl/accessibility/empty-image-with-title-expected.txt: Added.
- platform/efl/accessibility/form-control-value-settable-expected.txt: Added.
- platform/efl/accessibility/image-link-expected.txt:
- platform/efl/accessibility/image-map1-expected.txt:
- platform/efl/accessibility/image-map2-expected.txt:
- platform/efl/accessibility/image-with-alt-and-map-expected.txt:
- platform/efl/accessibility/img-fallsback-to-title-expected.txt:
- platform/efl/accessibility/math-fenced-expected.txt: Added.
- platform/efl/accessibility/math-foreign-content-expected.txt: Added.
- platform/efl/accessibility/math-multiscript-attributes-expected.txt:
- platform/efl/accessibility/math-text-expected.txt:
- platform/efl/accessibility/media-element-expected.txt:
- platform/efl/accessibility/meter-element-expected.txt:
- platform/efl/accessibility/radio-button-title-label-expected.txt:
- platform/efl/accessibility/table-detection-expected.txt:
- 12:04 AM Changeset in webkit [202551] by
-
- 15 edits in trunk/Source/WebCore
Binding generator should generate accessors for constructors safely accessed from JS builtin
https://bugs.webkit.org/show_bug.cgi?id=159087
Patch by Youenn Fablet <youenn@apple.com> on 2016-06-28
Reviewed by Alex Christensen.
Removed constructor private slots direct additions in JSDOMGlobalObject.
Added support for generating the code that will do that.
Advantage of the implementation:
- Private slots will expose constructors that are also publically visible (previously workers had some private slots filled with WebRTC constructors).
- Private slots no longer require the creation of the constructors at window creation time.
Although PublicIdentifier and PrivateIdentifier are both added where needed, the binding generator does not
support the case of a constructor accessible only privately.
Covered by existing test set and adding binding test.
- Modules/mediastream/MediaStream.idl: Marked as PublicIdentifier/PrivateIdentifier.
- Modules/mediastream/MediaStreamTrack.idl: Ditto.
- Modules/mediastream/RTCIceCandidate.idl: Ditto.
- Modules/mediastream/RTCSessionDescription.idl: Ditto.
- Modules/streams/ReadableStream.idl: Ditto.
- bindings/js/JSDOMGlobalObject.cpp:
(WebCore::JSDOMGlobalObject::addBuiltinGlobals): Removed unneeded additions.
- bindings/scripts/CodeGeneratorJS.pm:
(GenerateImplementation): Added support for private slots for interface constructors marked as
PrivateIdentifier.
- bindings/scripts/preprocess-idls.pl:
(GenerateConstructorAttribute): Make PublicIdentifier/PrivateIdentifier copied interface attributes.
- bindings/scripts/test/GObject/WebKitDOMTestGlobalObject.cpp:
(webkit_dom_test_global_object_set_property):
(webkit_dom_test_global_object_get_property):
(webkit_dom_test_global_object_class_init):
(webkit_dom_test_global_object_get_public_and_private_attribute):
(webkit_dom_test_global_object_set_public_and_private_attribute):
- bindings/scripts/test/GObject/WebKitDOMTestGlobalObject.h:
- bindings/scripts/test/JS/JSTestGlobalObject.cpp:
(WebCore::JSTestGlobalObject::finishCreation):
(WebCore::jsTestGlobalObjectPublicAndPrivateAttribute):
(WebCore::setJSTestGlobalObjectPublicAndPrivateAttribute):
- bindings/scripts/test/ObjC/DOMTestGlobalObject.h:
- bindings/scripts/test/ObjC/DOMTestGlobalObject.mm:
(-[DOMTestGlobalObject publicAndPrivateAttribute]):
(-[DOMTestGlobalObject setPublicAndPrivateAttribute:]):
- bindings/scripts/test/TestGlobalObject.idl:
- 12:03 AM Changeset in webkit [202550] by
-
- 5 edits in tags/Safari-601.7.5.1/Source
Versioning.
- 12:01 AM Changeset in webkit [202549] by
-
- 1 copy in tags/Safari-601.7.5.1
New tag.
Jun 27, 2016:
- 11:53 PM Changeset in webkit [202548] by
-
- 5 edits in tags/Safari-602.1.38.2.1/Source
Versioning.
- 11:52 PM Changeset in webkit [202547] by
-
- 1 copy in tags/Safari-602.1.38.2.1
New tag.
- 11:44 PM Changeset in webkit [202546] by
-
- 4 edits in trunk/Source/WebCore
REGRESSION?(r202466): http/tests/security/canvas-remote-read-remote-video-redirect.html failing on Sierra
https://bugs.webkit.org/show_bug.cgi?id=159172
<rdar://problem/27030025>
Reviewed by Brent Fulgham.
Add a hasSingleSecurityOrigin property to WebCoreNSURLSession that gets updated each time one of that
sessions' tasks receieves a response or a redirect request. Check that property from the MediaPlayerPrivate.
- platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:
(WebCore::MediaPlayerPrivateAVFoundationObjC::hasSingleSecurityOrigin):
- platform/network/cocoa/WebCoreNSURLSession.h:
- platform/network/cocoa/WebCoreNSURLSession.mm:
(-[WebCoreNSURLSession updateHasSingleSecurityOrigin:]):
(-[WebCoreNSURLSession dataTaskWithRequest:]):
(-[WebCoreNSURLSession dataTaskWithURL:]):
(-[WebCoreNSURLSessionDataTask resource:receivedResponse:]):
(-[WebCoreNSURLSessionDataTask resource:receivedRedirect:request:]):
- 11:38 PM Changeset in webkit [202545] by
-
- 2 edits in trunk/Source/WebCore
CMake build fix.
- PlatformMac.cmake:
- 11:22 PM Changeset in webkit [202544] by
-
- 2 edits in trunk/LayoutTests
Unreviewed, rebaseline test on iOS after r202539.
- platform/ios-simulator/fast/dom/event-handler-attributes-expected.txt:
- 11:11 PM Changeset in webkit [202543] by
-
- 1 edit in trunk/Tools/ChangeLog
Test commit. Please ignore.
- 11:09 PM Changeset in webkit [202542] by
-
- 14 edits in trunk
Remove didFailAccessControlCheck ThreadableLoaderClient callback
https://bugs.webkit.org/show_bug.cgi?id=159149
Patch by Youenn Fablet <youenn@apple.com> on 2016-06-27
Reviewed by Daniel Bates.
LayoutTests/imported/w3c:
- web-platform-tests/XMLHttpRequest/send-authentication-cors-setrequestheader-no-cred-expected.txt:
Source/WebCore:
Adding an AccessControl ResourceError type.
Replacing didFailAccessControlCheck callback by a direct call to didFail with an error of type AccessControl.
Making CrossOriginPreflightChecker always return an AccessControl error. Previously some errors created below
were passed directly to threadable loader clients.
When doing preflight on unauthorized web sites, WTR/DRT will trigger a cancellation error which was translating into an abort event in XMLHttpRequest.
This patch is changing the error type to AccessControl, which translates into an error event in XMLHttpReauest.
This change of behavior is seen in imported/w3c/web-platform-tests/XMLHttpRequest/send-authentication-cors-setrequestheader-no-cred.htm.
No other observable change of behavior should be expected.
- inspector/InspectorNetworkAgent.cpp: Computing error message in didFail according the error type.
- loader/CrossOriginPreflightChecker.cpp:
(WebCore::CrossOriginPreflightChecker::validatePreflightResponse): Setting preflightFailure error type to AccessControl.
(WebCore::CrossOriginPreflightChecker::notifyFinished): Ditto.
(WebCore::CrossOriginPreflightChecker::doPreflight): Ditto.
- loader/DocumentThreadableLoader.cpp:
(WebCore::DocumentThreadableLoader::makeSimpleCrossOriginAccessRequest): Replacing didFailAccessControlCheck
callback by a direct call to didFail with an error of type AccessControl.
(WebCore::reportContentSecurityPolicyError): Ditto.
(WebCore::reportCrossOriginResourceSharingError): Ditto.
(WebCore::DocumentThreadableLoader::didReceiveResponse): Ditto.
(WebCore::DocumentThreadableLoader::preflightFailure): Calling didFail directly.
- loader/ThreadableLoaderClient.h: Removing didFailAccessControlCheck.
- loader/ThreadableLoaderClientWrapper.h: Ditto.
- loader/WorkerThreadableLoader.cpp: Ditto.
- loader/WorkerThreadableLoader.h: Ditto.
- page/EventSource.cpp:
(WebCore::EventSource::didFail): Removing didFailAccessControlCheck and putting handling code in didFail.
- page/EventSource.h:
- platform/network/ResourceErrorBase.cpp:
(WebCore::ResourceErrorBase::setType): Softening the assertion to cover the case of migration to AccessControl.
- platform/network/ResourceErrorBase.h: Adding AccessControl error type.
(WebCore::ResourceErrorBase::isAccessControl):
- 11:07 PM Changeset in webkit [202541] by
-
- 2 edits in trunk/LayoutTests
Mark some tests as image failures.
These tests started to fail with the render server snapshotting enabled in r202532.
- platform/ios-simulator-wk2/TestExpectations:
- 10:56 PM Changeset in webkit [202540] by
-
- 2 edits in trunk/Source/WebKit2
Fix TestWebKitAPI crash introduced in r202532.
WebPageProxy::forceRepaint() could trigger a call to the didForceRepaintCallback()
after the WebPageProxy had been invalidated, causing a null de-ref of m_drawingArea.
Fix by checking for error and isValid().
- UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::forceRepaint):
- 10:10 PM Changeset in webkit [202539] by
-
- 9 edits in trunk
HTMLElement / SVGElement should implement GlobalEventHandlers, not Element
https://bugs.webkit.org/show_bug.cgi?id=159191
<rdar://problem/27019299>
Reviewed by Ryosuke Niwa.
LayoutTests/imported/w3c:
Rebaseline test now that more checks are passing.
- web-platform-tests/html/dom/interfaces-expected.txt:
Source/WebCore:
HTMLElement / SVGElement should implement GlobalEventHandlers, not Element:
- https://html.spec.whatwg.org/multipage/dom.html#htmlelement
- https://www.w3.org/TR/SVG2/types.html#InterfaceSVGElement
Firefox and Chrome behave as per the specification.
Fixing this also fixes rendering on http://survey123.arcgis.com/.
No new tests, covered by existing tests that were rebaselined.
- dom/Element.idl:
- html/HTMLElement.idl:
- svg/SVGElement.idl:
LayoutTests:
- fast/dom/event-handler-attributes-expected.txt:
Rebaseline tests now that more checks are passing.
- js/dom/dom-static-property-for-in-iteration-expected.txt:
Rebaseline test now that enumeration order is different.
- 9:04 PM Changeset in webkit [202538] by
-
- 2 edits in trunk/Source/WebCore
[macOS] Test gardening: Generic font families should not map to fonts which aren't installed
https://bugs.webkit.org/show_bug.cgi?id=159111
<rdar://problem/25807529>
Unreviewed.
Osaka-Mono does not come preinstalled on macOS Sierra. However, many Japanese users
will have the font installed. Before setting the generic font family, we should check
to see if the font is present.
- page/cocoa/SettingsCocoa.mm:
(WebCore::osakaMonoIsInstalled):
(WebCore::Settings::initializeDefaultFontFamilies):
- 8:25 PM Changeset in webkit [202537] by
-
- 2 edits in trunk/Source/WebCore
Don't keep all newly created potential custom elements alive when the feature is disabled
https://bugs.webkit.org/show_bug.cgi?id=159113
Reviewed by Daniel Bates.
Don't keep all HTML unknown elements which has a valid custom element alive when the feature is turned off.
Ideally we want to conform to the behavior in the Custom Elements specification and only upgrade an element
that is inserted into the document. We'll implement that later.
- dom/Document.cpp:
(WebCore::createHTMLElementWithNameValidation):
- 7:58 PM Changeset in webkit [202536] by
-
- 7 edits4 adds in trunk
[iOS] -webkit-overflow-scrolling: touch prevents repaint with RTL
https://bugs.webkit.org/show_bug.cgi?id=159186
rdar://problem/26659341
Reviewed by Zalan Bujtas.
Source/WebCore:
There were two issues with repaints in -webkit-overflow-scrolling:touch scrolling
layers.
First, if the scrolled contents were inline (e.g. a <span>), then repaints were
broken because RenderInline didn't call shouldApplyClipAndScrollPositionForRepaint().
Fix by making shouldApplyClipAndScrollPositionForRepaint() a member function of RenderBox
and calling it from RenderBox::computeRectForRepaint() and RenderInline::clippedOverflowRectForRepaint().
Second, repaints were broken in RTL because RenderLayerBacking::setContentsNeedDisplayInRect()
confused scroll offset and scroll position; it needs to subtract scrollPosition.
Finally renamed to applyCachedClipAndScrollOffsetForRepaint() to applyCachedClipAndScrollPositionForRepaint()
to make it clear that it uses scrollPosition, not scrollOffset.
Tests: compositing/scrolling/touch-scrolling-repaint-spans.html
compositing/scrolling/touch-scrolling-repaint.html
- rendering/RenderBox.cpp:
(WebCore::RenderBox::applyCachedClipAndScrollPositionForRepaint):
(WebCore::RenderBox::shouldApplyClipAndScrollPositionForRepaint):
(WebCore::RenderBox::computeRectForRepaint):
(WebCore::RenderBox::applyCachedClipAndScrollOffsetForRepaint): Deleted.
(WebCore::shouldApplyContainersClipAndOffset): Deleted.
- rendering/RenderBox.h:
- rendering/RenderInline.cpp:
(WebCore::RenderInline::clippedOverflowRectForRepaint):
(WebCore::RenderInline::computeRectForRepaint):
- rendering/RenderLayerBacking.cpp:
(WebCore::RenderLayerBacking::setContentsNeedDisplayInRect):
- rendering/RenderObject.cpp:
(WebCore::RenderObject::computeRectForRepaint):
LayoutTests:
- compositing/scrolling/touch-scrolling-repaint-expected.html: Added.
- compositing/scrolling/touch-scrolling-repaint-spans-expected.html: Added.
- compositing/scrolling/touch-scrolling-repaint-spans.html: Added.
- compositing/scrolling/touch-scrolling-repaint.html: Added.
- 7:47 PM Changeset in webkit [202535] by
-
- 2 edits in trunk/Tools
[GTK][EFL] Disable subtle-crypto in FeatureList.pm
https://bugs.webkit.org/show_bug.cgi?id=155073
Reviewed by Daniel Bates.
- Scripts/webkitperl/FeatureList.pm:
- 7:38 PM Changeset in webkit [202534] by
-
- 12 edits in trunk/Websites/webkit.org
Updated content for the color blog post.
- blog-files/color-gamut/Iceland-P3.jpg:
- blog-files/color-gamut/Iceland-oog.jpg:
- blog-files/color-gamut/Iceland-sRGB.jpg:
- blog-files/color-gamut/Italy-P3.jpg:
- blog-files/color-gamut/Italy-sRGB.jpg:
- blog-files/color-gamut/Sunset-P3.jpg:
- blog-files/color-gamut/Sunset-sRGB.jpg:
- blog-files/color-gamut/YellowFlower-P3.jpg:
- blog-files/color-gamut/YellowFlower-oog.jpg:
- blog-files/color-gamut/YellowFlower-sRGB.jpg:
- blog-files/color-gamut/index.html:
- 7:00 PM Changeset in webkit [202533] by
-
- 2 edits in trunk/Source/WebCore
Unreviewed, rolling out r202436.
https://bugs.webkit.org/show_bug.cgi?id=159190
We don't need to make this change. (Requested by thorton on
#webkit).
Reverted changeset:
"Do not use iOS specific telephone detection on macOS."
https://bugs.webkit.org/show_bug.cgi?id=159096
http://trac.webkit.org/changeset/202436
- 6:44 PM Changeset in webkit [202532] by
-
- 9 edits in trunk
[iOS] Make DumpRenderTree and WebKitTestRunner in the simulator use render server snapshotting
https://bugs.webkit.org/show_bug.cgi?id=159077
Reviewed by Tim Horton.
Source/WebKit2:
Fix WebPageProxy::forceRepaint() to correctly wait for the next commit from the web process,
which is necessary for UI-side compositing.
- UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::forceRepaint):
Tools:
Re-enable render server snapshotting for iOS WTR via the change in InjectedBundlePage.cpp.
Fix WebPageProxy::forceRepaint() to correctly wait for the next commit from the web process,
which is necessary for UI-side compositing.
Add some null checks to fix issues when the WKWebView gets resized to be empty, which
seems to happen for some tests that call window.resizeTo().
- WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::dump): Deleted.
- WebKitTestRunner/TestInvocation.cpp:
(WTR::TestInvocation::dumpResults):
- WebKitTestRunner/cg/TestInvocationCG.cpp:
(WTR::createCGContextFromImage):
(WTR::computeMD5HashStringForContext):
(WTR::TestInvocation::dumpPixelsAndCompareWithExpected):
- WebKitTestRunner/ios/PlatformWebViewIOS.mm:
(WTR::PlatformWebView::windowSnapshotImage):
LayoutTests:
fast/harness/snapshot-captures-compositing.html passes now.
- platform/ios-simulator-wk2/TestExpectations:
- 6:13 PM Changeset in webkit [202531] by
-
- 6 edits12 adds in trunk
Adopt the iOS TouchEventHandler API for cases that must have synchronous dispatch
https://bugs.webkit.org/show_bug.cgi?id=159179
rdar://problem/27006387
Reviewed by Simon Fraser.
Source/WebCore:
Tests: fast/events/touch/ios/block-without-overflow-scroll-and-passive-observer-on-block-scrolling-state.html
fast/events/touch/ios/block-without-overflow-scroll-and-passive-observer-on-document-scrolling-state.html
fast/events/touch/ios/block-without-overflow-scroll-scrolling-state.html
fast/events/touch/ios/drag-block-without-overflow-scroll-and-passive-observer-on-block.html
fast/events/touch/ios/drag-block-without-overflow-scroll-and-passive-observer-on-document.html
fast/events/touch/ios/drag-block-without-overflow-scroll.html
- dom/Document.cpp:
(WebCore::Document::prepareForDestruction):
(WebCore::Document::removeAllEventListeners):
- dom/Node.cpp:
(WebCore::Node::willBeDeletedFrom):
(WebCore::tryAddEventListener):
(WebCore::tryRemoveEventListener):
- html/shadow/SliderThumbElement.cpp:
(WebCore::SliderThumbElement::registerForTouchEvents):
(WebCore::SliderThumbElement::unregisterForTouchEvents):
- rendering/RenderLayer.cpp:
(WebCore::RenderLayer::registerAsTouchEventListenerForScrolling):
(WebCore::RenderLayer::unregisterAsTouchEventListenerForScrolling):
LayoutTests:
Several of those tests expose existing bugs with overflow scrolling.
They are not using PASS/FAIL at the moment, just dump the called listeners.
- fast/events/touch/ios/block-without-overflow-scroll-and-passive-observer-on-block-scrolling-state-expected.txt: Added.
- fast/events/touch/ios/block-without-overflow-scroll-and-passive-observer-on-block-scrolling-state.html: Added.
- fast/events/touch/ios/block-without-overflow-scroll-and-passive-observer-on-document-scrolling-state-expected.txt: Added.
- fast/events/touch/ios/block-without-overflow-scroll-and-passive-observer-on-document-scrolling-state.html: Added.
- fast/events/touch/ios/block-without-overflow-scroll-scrolling-state-expected.txt: Added.
- fast/events/touch/ios/block-without-overflow-scroll-scrolling-state.html: Added.
- fast/events/touch/ios/drag-block-without-overflow-scroll-and-passive-observer-on-block-expected.txt: Added.
- fast/events/touch/ios/drag-block-without-overflow-scroll-and-passive-observer-on-block.html: Added.
- fast/events/touch/ios/drag-block-without-overflow-scroll-and-passive-observer-on-document-expected.txt: Added.
- fast/events/touch/ios/drag-block-without-overflow-scroll-and-passive-observer-on-document.html: Added.
- fast/events/touch/ios/drag-block-without-overflow-scroll-expected.txt: Added.
- fast/events/touch/ios/drag-block-without-overflow-scroll.html: Added.
- 6:00 PM Changeset in webkit [202530] by
-
- 2 edits in trunk/Source/WebCore
Fix Windows build.
- bindings/js/SerializedScriptValue.h:
WTF
- 5:55 PM Changeset in webkit [202529] by
-
- 5 edits3 adds in trunk
Web Inspector: When modifying sessionStorage, localStorage gets updated
https://bugs.webkit.org/show_bug.cgi?id=159181
<rdar://problem/27043447>
Patch by Joseph Pecoraro <Joseph Pecoraro> on 2016-06-27
Reviewed by Timothy Hatcher.
Source/WebInspectorUI:
- UserInterface/Test/Test.js:
(WebInspector.loaded):
Add registration for StorageManager and StorageObserver.
Source/WebKit2:
- WebProcess/Storage/StorageAreaMap.cpp:
(WebKit::StorageAreaMap::dispatchSessionStorageEvent):
This should be dispatching storage events.
LayoutTests:
- inspector/storage/domStorage-events-expected.txt: Added.
- inspector/storage/domStorage-events.html: Added.
Add a new test for DOMStorage domain events. Ensures that sessionStorage
and localStorage events are dispatched for the appropriate DOMStorageObject.
- 5:42 PM Changeset in webkit [202528] by
-
- 2 edits1 add in trunk/Source/JavaScriptCore
Fix bad assert in StructureRareData::setObjectToStringValue
https://bugs.webkit.org/show_bug.cgi?id=159171
<rdar://problem/26987355>
Reviewed by Mark Lam.
We should not have expected the generateConditionsForPrototypePropertyHit would succeed.
There are many reasons it might fail including that there is a proxy somewhere on the
prototype chain of the object.
- runtime/StructureRareData.cpp:
(JSC::StructureRareData::setObjectToStringValue):
- tests/stress/object-toString-with-proxy.js: Added.
(get target):
- 5:36 PM Changeset in webkit [202527] by
-
- 4 edits in trunk/Source/WebCore
Unreviewed, rolling out r202520.
https://bugs.webkit.org/show_bug.cgi?id=159185
This change broke the 32-bit El Capitan build (Requested by
ryanhaddad on #webkit).
Reverted changeset:
"REGRESSION?(r202466): http/tests/security/canvas-remote-read-
remote-video-redirect.html failing on Sierra"
https://bugs.webkit.org/show_bug.cgi?id=159172
http://trac.webkit.org/changeset/202520
Patch by Commit Queue <commit-queue@webkit.org> on 2016-06-27
- 5:36 PM Changeset in webkit [202526] by
-
- 2 edits in trunk/LayoutTests
[Cocoa] Test gardening for the system font in macOS Sierra and iOS 10
https://bugs.webkit.org/show_bug.cgi?id=159175
<rdar://problem/26478081>
Unreviewed.
- fast/text/trak-optimizeLegibility.html:
- 5:34 PM Changeset in webkit [202525] by
-
- 5 edits3 deletes in trunk
Fix flakiness on Sierra after r202511
https://bugs.webkit.org/show_bug.cgi?id=159071
Source/WebKit2:
- NetworkProcess/cocoa/NetworkSessionCocoa.mm:
(WebKit::NetworkSession::clearCredentials):
Replacing the NSURLSessions immediately caused assertions because the tasks are not always all stopped immediately when tearing down a test.
Stopping an NSURLSession needs to be done asynchronously. I'll try a different testing technique for r202511.
LayoutTests:
- http/tests/xmlhttprequest/sync-delegate-callbacks-expected.txt: Removed.
- http/tests/xmlhttprequest/sync-delegate-callbacks.html: Removed.
- platform/ios-simulator-wk2/TestExpectations:
- platform/mac/TestExpectations:
- platform/wk2/http/tests/xmlhttprequest/sync-delegate-callbacks-expected.txt: Removed.
- 5:33 PM Changeset in webkit [202524] by
-
- 4 edits in trunk/Source/WebKit/mac
REGRESSION (r189052): Clipping occurs when using context menu to Look Up words within the Dictionary app
https://bugs.webkit.org/show_bug.cgi?id=159184
<rdar://problem/26370206>
Reviewed by Beth Dakin.
- WebView/WebHTMLView.mm:
(-[WebHTMLView _lookUpInDictionaryFromMenu:]):
- WebView/WebImmediateActionController.h:
- WebView/WebImmediateActionController.mm:
(+[WebImmediateActionController _dictionaryPopupInfoForRange:inFrame:withLookupOptions:indicatorOptions:transition:]):
(-[WebImmediateActionController _animationControllerForText]):
(dictionaryPopupInfoForRange): Deleted.
WebImmediateActionController's code to make a DictionaryPopupInfo and TextIndicator
from a Range in WebKit1 is much better than WebHTMLView's; use it in both cases.
- 5:26 PM Changeset in webkit [202523] by
-
- 2 edits in trunk/Tools
[EFL] Support to build on Debian linux
https://bugs.webkit.org/show_bug.cgi?id=159123
Reviewed by Antonio Gomes.
- efl/install-dependencies: Install libgnutls28-dev.
- 5:23 PM Changeset in webkit [202522] by
-
- 6 edits4 adds in trunk
Web Inspector: RuntimeManager should not use view object WebInspector.quickConsole
https://bugs.webkit.org/show_bug.cgi?id=128092
<rdar://problem/15966526>
Reviewed by Timothy Hatcher.
Source/WebInspectorUI:
This is a layering violation which makes it harder to use RuntimeManager.evaluateInInspectedWindow
from a testing context where the QuickConsole view does not exist.
Store the selected execution context identifier on RuntimeManager and use it
when doing subsequent evaluations that act on the currently selected frame.
- UserInterface/Controllers/RuntimeManager.js:
(WebInspector.RuntimeManager.prototype.evaluateInInspectedWindow):
(WebInspector.RuntimeManager.prototype.saveResult):
Use local state.
(WebInspector.RuntimeManager.prototype.get defaultExecutionContextIdentifier):
(WebInspector.RuntimeManager.prototype.set defaultExecutionContextIdentifier):
Added.
(WebInspector.RuntimeManager):
- UserInterface/Models/ExecutionContext.js:
(WebInspector.ExecutionContext):
Move the symbolic name for the top level execution context to RuntimeManager.
- UserInterface/Test/Test.js:
(WebInspector.loaded): No need to stub out WebInspector.QuickConsole any more.
- UserInterface/Views/QuickConsole.js:
(WebInspector.QuickConsole.prototype._framePageExecutionContextsChanged):
(WebInspector.QuickConsole.prototype._removeExecutionContextPathComponentForFrame):
(WebInspector.QuickConsole.prototype._updateExecutionContextPathComponentForFrame):
(WebInspector.QuickConsole.prototype._pathComponentSelected):
For now, set RuntimeManager's selected execution context whenever we set the
selected path component. In a future patch, we should invert the dependency and have
the selected component change whenever RuntimeManager.defaultExecutionContext changes.
LayoutTests:
Add some really basic coverage for RuntimeManager.defaultExecutionContextIdentifier
and using it in RuntimeManager.evaluateInInspectedWindow.
- inspector/runtime/change-execution-context-identifier-expected.txt: Added.
- inspector/runtime/change-execution-context-identifier.html: Added.
- inspector/runtime/resources/change-execution-context-identifier-subframe.html: Added.
- 4:59 PM Changeset in webkit [202521] by
-
- 2 edits in trunk/Websites/webkit.org
Set a default image for sharing webkit.org links via social.
https://bugs.webkit.org/show_bug.cgi?id=159167
Reviewed by Timothy Hatcher.
- wp-content/plugins/social-meta.php:
- 4:56 PM Changeset in webkit [202520] by
-
- 4 edits in trunk/Source/WebCore
REGRESSION?(r202466): http/tests/security/canvas-remote-read-remote-video-redirect.html failing on Sierra
https://bugs.webkit.org/show_bug.cgi?id=159172
<rdar://problem/27030025>
Reviewed by Brent Fulgham.
Add a hasSingleSecurityOrigin property to WebCoreNSURLSession that gets updated each time one of that
sessions' tasks receieves a response or a redirect request. Check that property from the MediaPlayerPrivate.
- platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:
(WebCore::MediaPlayerPrivateAVFoundationObjC::hasSingleSecurityOrigin):
- platform/network/cocoa/WebCoreNSURLSession.h:
- platform/network/cocoa/WebCoreNSURLSession.mm:
(-[WebCoreNSURLSession updateHasSingleSecurityOrigin:]):
(-[WebCoreNSURLSession dataTaskWithRequest:]):
(-[WebCoreNSURLSession dataTaskWithURL:]):
(-[WebCoreNSURLSessionDataTask resource:receivedResponse:]):
(-[WebCoreNSURLSessionDataTask resource:receivedRedirect:request:]):
- 4:26 PM Changeset in webkit [202519] by
-
- 3 edits in trunk/Source/JavaScriptCore
Crashing at an unreachable code trap in FTL should give more information
https://bugs.webkit.org/show_bug.cgi?id=159177
Reviewed by Saam Barati.
This stuffs information into registers so that we have some chance of seeing what happened
by looking at the register dumps.
- assembler/AbortReason.h:
- ftl/FTLLowerDFGToB3.cpp:
(JSC::FTL::DFG::ftlUnreachable):
(JSC::FTL::DFG::LowerDFGToB3::compileBlock):
(JSC::FTL::DFG::LowerDFGToB3::crash):
- 3:26 PM Changeset in webkit [202518] by
-
- 4 edits in trunk/Source/JavaScriptCore
Clean up resetting reachability in B3/Air
https://bugs.webkit.org/show_bug.cgi?id=159170
Reviewed by Geoffrey Garen.
When I fixed bug 159165, I took the brute force approach. I still used the
B3::resetReachability() method, and changed the callback to record the set of deleted values
instead of deleting them eagerly. But this means tracking the set of deleted values, even
though resetReachability() already internally tracks the set of deleted blocks. You can find
out if a value is deleted by asking if its owning block was deleted.
So, this change refactors B3::resetReachability() into a new helper called
B3::recomputePredecessors(). This new helper skips the block deletion step, and lets the
client delete blocks. This lets Air delete blocks the same way that it did before, and it
lets B3 use the isBlockDead() method (which is a glorified proxy for
block->predecessors().isEmpty()) to track which values are deleted. This allows B3 to turn
Upsilons that point to dead Phis into Nops before deleting the blocks.
This shouldn't affect performance or anything real. It just makes the code cleaner.
- b3/B3BasicBlockUtils.h:
(JSC::B3::updatePredecessorsAfter):
(JSC::B3::recomputePredecessors):
(JSC::B3::isBlockDead):
(JSC::B3::resetReachability): Deleted.
- b3/B3Procedure.cpp:
(JSC::B3::Procedure::resetReachability):
(JSC::B3::Procedure::invalidateCFG):
- b3/air/AirCode.cpp:
(JSC::B3::Air::Code::resetReachability):
(JSC::B3::Air::Code::dump):
- 3:20 PM Changeset in webkit [202517] by
-
- 5 edits14 adds in trunk
Fix style invalidation for :active when the activated node has no renderer
https://bugs.webkit.org/show_bug.cgi?id=159125
Reviewed by Antti Koivisto.
Source/WebCore:
Same old bug: a style invalidation path was depending
on the style.
Here we really need both flags. An element can have
childrenAffectedByActive() false and renderStyle->affectedByActive() true
if it was subject to style sharing.
The element state "childrenAffectedByActive" should be renamed
"styleAffectedByActive" since it is not a parent invalidation flag.
That will be done separately.
Tests: fast/css/pseudo-active-on-labeled-control-without-renderer.html
fast/css/pseudo-active-style-sharing-1.html
fast/css/pseudo-active-style-sharing-2.html
fast/css/pseudo-active-style-sharing-3.html
fast/css/pseudo-active-style-sharing-4.html
fast/css/pseudo-active-style-sharing-5.html
fast/css/pseudo-active-style-sharing-6.html
- dom/Element.cpp:
(WebCore::Element::setActive):
- style/StyleRelations.cpp:
(WebCore::Style::commitRelationsToRenderStyle):
LayoutTests:
There was no bug with style sharing but I wanted that covered anyway.
Style sharing depends on 2 flags which is uncommon.
There was no test coverage whatsoever, breaking it did not fail any test.
- fast/css/pseudo-active-on-labeled-control-without-renderer-expected.txt: Added.
- fast/css/pseudo-active-on-labeled-control-without-renderer.html: Added.
- fast/css/pseudo-active-style-sharing-1-expected.txt: Added.
- fast/css/pseudo-active-style-sharing-1.html: Added.
- fast/css/pseudo-active-style-sharing-2-expected.txt: Added.
- fast/css/pseudo-active-style-sharing-2.html: Added.
- fast/css/pseudo-active-style-sharing-3-expected.txt: Added.
- fast/css/pseudo-active-style-sharing-3.html: Added.
- fast/css/pseudo-active-style-sharing-4-expected.txt: Added.
- fast/css/pseudo-active-style-sharing-4.html: Added.
- fast/css/pseudo-active-style-sharing-5-expected.txt: Added.
- fast/css/pseudo-active-style-sharing-5.html: Added.
- fast/css/pseudo-active-style-sharing-6-expected.txt: Added.
- fast/css/pseudo-active-style-sharing-6.html: Added.
- 3:13 PM Changeset in webkit [202516] by
-
- 3 edits3 adds in trunk
AX: REGRESSION (r202063): ARIA role attribute is being ignored for label element
https://bugs.webkit.org/show_bug.cgi?id=159162
Reviewed by Chris Fleizach.
Source/WebCore:
createFromRenderer() was creating an AccessibilityLabel for any HTMLLabelElement which
lacked an explicitly-handled ARIA role. We should instead create an AccessibilityLabel
when there is no ARIA role.
Test: accessibility/aria-role-on-label.html
- accessibility/AXObjectCache.cpp:
(WebCore::createFromRenderer):
LayoutTests:
This regression was caught by a GTK-specific layout test. Adding a cross-platform
test for additional coverage.
- accessibility/aria-role-on-label.html: Added.
- platform/gtk/accessibility/aria-role-on-label-expected.txt: Added.
- platform/mac/accessibility/aria-role-on-label-expected.txt: Added.
- 2:50 PM Changeset in webkit [202515] by
-
- 2 edits in trunk/Source/JavaScriptCore
Web Inspector: CRASH in backend at Inspector::HeapFrontendDispatcher::garbageCollected + 552 when closing frontend/inspected page
https://bugs.webkit.org/show_bug.cgi?id=159075
<rdar://problem/26094341>
Reviewed by Filip Pizlo.
This change caused JSC stress tests to all hit an assertion in RunLoop.
We should use RunLoop::current() to create the RunLoop::Timer since JSC-only
clients like testapi and jsc don't ever call initializeMainRunLoop().
- inspector/agents/InspectorHeapAgent.cpp:
(Inspector::SendGarbageCollectionEventsTask::SendGarbageCollectionEventsTask):
- 2:45 PM Changeset in webkit [202514] by
-
- 8 edits2 deletes in trunk
Unreviewed, rolling out r202505.
https://bugs.webkit.org/show_bug.cgi?id=159169
The test added with this change is flaky and it caused an
existing test to time out on El Capitan. (Requested by
ryanhaddad on #webkit).
Reverted changeset:
"[iOS] Media controls are too cramped with small video"
https://bugs.webkit.org/show_bug.cgi?id=158815
http://trac.webkit.org/changeset/202505
Patch by Commit Queue <commit-queue@webkit.org> on 2016-06-27
- 2:44 PM Changeset in webkit [202513] by
-
- 2 edits in trunk/Source/WebCore
Add :focus-within to the status page
Patch by Benjamin Poulain <bpoulain@apple.com> on 2016-06-27
- features.json:
I forgot to update the json file when landing the feature.
- 2:36 PM Changeset in webkit [202512] by
-
- 5 edits in trunk/Source/JavaScriptCore
B3::Procedure::resetReachability() can create dangling references from Upsilons to Phis
https://bugs.webkit.org/show_bug.cgi?id=159165
Reviewed by Mark Lam.
You can delete an unreachable block that has a Phi but some prior block may still have an
Upsilon pointing to that Phi. This can happen if the Upsilon precedes a Check that always
exits or it can happen if we remove some successor of a block and this block had an Upsilon
for one of the removed successors. These things are valid IR even if they are not canonical.
Our policy for not-canonical-but-valid IR is that the compiler should still emit valid code
in the end.
The solution is to have Procedure::resetReachability() turn those Upsilons into Nops.
- b3/B3Procedure.cpp:
(JSC::B3::Procedure::resetReachability): Fix the bug.
- b3/B3Validate.h:
- b3/testb3.cpp:
(JSC::B3::testResetReachabilityDanglingReference): Add a test. This always crashes prior to this change.
- dfg/DFGGraph.cpp:
(JSC::DFG::Graph::killUnreachableBlocks): Add a FIXME about a possible similar bug.
- 1:59 PM Changeset in webkit [202511] by
-
- 32 edits3 adds in trunk
Send canAuthenticateAgainstProtectionSpace calls from NetworkProcess directly to UIProcess
https://bugs.webkit.org/show_bug.cgi?id=159071
Patch by Alex Christensen <achristensen@webkit.org> on 2016-06-27
Reviewed by Brady Eidson.
Source/WebKit2:
When CFNetwork asks the NetworkProcess if we can authenticate against this protection space, we used to ask
the WebProcess, which would call ResourceLoader::canAuthenticateAgainstProtectionSpace which would ask the FrameLoaderClient,
which would synchronously ask the UIProcess. Going directly from the NetworkProcess to the UIProcess gives us many advantages:
- It reduces IPC because we ask fewer loads.
- It reduces synchronous IPC because we are not using the same code path as WebKit1 any more.
- It allows us to check if the NetworkProcess is still valid in sendToNetworkingProcess, possibly fixing rdar://problem/26825408
- It allows us to ask the UIProcess when we get authentication challenges for synchronous xhr.
I added a new test http/tests/xmlhttprequest/sync-delegate-callbacks.html to verify the change in behavior of synchronous xhr.
In order to make this test not flaky on platforms using NSURLSession, we clear the NSURLSessions used by WebKitTestRunner
to make sure that the TLS session cache is cleared so we get NSURLAuthenticationChallenges corresponding to the NSURLConnection
canAuthenticateAgainstProtectionSpace callback when we initiate the first HTTPS connection to 127.0.0.1 with this session.
- NetworkProcess/NetworkLoad.cpp:
(WebKit::NetworkLoad::didReceiveChallenge):
(WebKit::NetworkLoad::didReceiveResponseNetworkSession):
(WebKit::NetworkLoad::canAuthenticateAgainstProtectionSpaceAsync):
- NetworkProcess/NetworkProcess.cpp:
(WebKit::NetworkProcess::clearCachedCredentials):
(WebKit::NetworkProcess::ensurePrivateBrowsingSession):
(WebKit::NetworkProcess::cancelDownload):
(WebKit::NetworkProcess::canAuthenticateAgainstProtectionSpace):
(WebKit::NetworkProcess::continueCanAuthenticateAgainstProtectionSpace):
(WebKit::NetworkProcess::continueCanAuthenticateAgainstProtectionSpaceDownload):
- NetworkProcess/NetworkProcess.h:
- NetworkProcess/NetworkProcess.messages.in:
- NetworkProcess/NetworkResourceLoader.cpp:
(WebKit::NetworkResourceLoader::canAuthenticateAgainstProtectionSpaceAsync):
- NetworkProcess/NetworkResourceLoader.h:
- NetworkProcess/NetworkResourceLoader.messages.in:
- NetworkProcess/NetworkSession.h:
- NetworkProcess/cocoa/NetworkSessionCocoa.mm:
(WebKit::NetworkSession::networkStorageSession):
(WebKit::NetworkSession::clearCredentials):
(WebKit::NetworkSession::dataTaskForIdentifier):
- UIProcess/Downloads/DownloadProxy.cpp:
(WebKit::DownloadProxy::canAuthenticateAgainstProtectionSpace):
(WebKit::DownloadProxy::willSendRequest):
- UIProcess/Network/NetworkProcessProxy.cpp:
(WebKit::NetworkProcessProxy::logSampledDiagnosticMessageWithValue):
(WebKit::NetworkProcessProxy::canAuthenticateAgainstProtectionSpace):
(WebKit::NetworkProcessProxy::sendProcessWillSuspendImminently):
- UIProcess/Network/NetworkProcessProxy.h:
- UIProcess/Network/NetworkProcessProxy.messages.in:
- UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::backForwardClear):
(WebKit::WebPageProxy::canAuthenticateAgainstProtectionSpace):
(WebKit::WebPageProxy::didReceiveAuthenticationChallenge):
(WebKit::WebPageProxy::canAuthenticateAgainstProtectionSpaceInFrame): Deleted.
- UIProcess/WebPageProxy.h:
- UIProcess/WebPageProxy.messages.in:
- WebProcess/Network/WebResourceLoader.cpp:
(WebKit::WebResourceLoader::didReceiveResource):
(WebKit::WebResourceLoader::isAlwaysOnLoggingAllowed):
(WebKit::WebResourceLoader::canAuthenticateAgainstProtectionSpace): Deleted.
- WebProcess/Network/WebResourceLoader.h:
- WebProcess/Network/WebResourceLoader.messages.in:
- WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:
(WebKit::WebFrameLoaderClient::canAuthenticateAgainstProtectionSpace):
- WebProcess/WebProcess.cpp:
(WebKit::WebProcess::clearCachedCredentials):
(WebKit::WebProcess::focusedWebPage):
Tools:
- WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
- WebKitTestRunner/InjectedBundle/TestRunner.cpp:
(WTR::TestRunner::setHandlesAuthenticationChallenges):
(WTR::TestRunner::setShouldLogCanAuthenticateAgainstProtectionSpace):
(WTR::TestRunner::setAuthenticationUsername):
- WebKitTestRunner/InjectedBundle/TestRunner.h:
- WebKitTestRunner/TestController.cpp:
(WTR::TestController::TestController):
(WTR::TestController::resetStateToConsistentValues):
(WTR::TestController::didFinishNavigation):
(WTR::TestController::canAuthenticateAgainstProtectionSpace):
(WTR::TestController::didReceiveAuthenticationChallenge):
(WTR::TestController::didCommitNavigation):
- WebKitTestRunner/TestController.h:
(WTR::TestController::setBlockAllPlugins):
(WTR::TestController::setShouldLogHistoryClientCallbacks):
(WTR::TestController::setShouldLogCanAuthenticateAgainstProtectionSpace):
(WTR::TestController::isCurrentInvocation):
- WebKitTestRunner/TestInvocation.cpp:
(WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
Added output so we can see when canAuthenticateAgainstProtectionSpace is called from the test expectations.
LayoutTests:
- http/tests/xmlhttprequest/sync-delegate-callbacks-expected.txt: Added.
- http/tests/xmlhttprequest/sync-delegate-callbacks.html: Added.
- platform/ios-simulator-wk2/TestExpectations:
- platform/mac/TestExpectations:
- platform/wk2/http/tests/xmlhttprequest/sync-delegate-callbacks-expected.txt: Added.
- 1:52 PM Changeset in webkit [202510] by
-
- 1 edit1 delete in trunk/LayoutTests
Remove OS X Mavericks test results
Rubber-stamped by Lucas Forschler.
- platform/mac-mavericks: Removed.
- 1:37 PM Changeset in webkit [202509] by
-
- 5 edits in trunk
[Mac] PiP placeholder should remain visible when 'controls' attribute is removed
https://bugs.webkit.org/show_bug.cgi?id=159158
<rdar://problem/26727435>
Reviewed by Jer Noble.
Source/WebCore:
No new tests, existing test updated.
- Modules/mediacontrols/mediaControlsApple.js:
(Controller.prototype.shouldHaveControls): Always return true when in PiP or AirPlay mode.
LayoutTests:
- media/controls/picture-in-picture-expected.txt: Updated.
- media/controls/picture-in-picture.html: Ditto.
- 12:52 PM Changeset in webkit [202508] by
-
- 2 edits in trunk/Source/JavaScriptCore
Add comment to Module feature in features.json
https://bugs.webkit.org/show_bug.cgi?id=159159
Reviewed by Saam Barati.
- features.json:
- 12:17 PM Changeset in webkit [202507] by
-
- 3 edits in trunk/Source/WebCore
Update ATS WebContent exception for more robust framework information
https://bugs.webkit.org/show_bug.cgi?id=159151
Reviewed by Alex Christensen.
We found some unexpected poor interaction with AVFoundation in the existing
CFNetwork SPI. This new SPI is more solid and let's us provide more useful
information while also being more future proof against new frameworks and
ATS modes.
- platform/network/mac/ResourceHandleMac.mm:
(WebCore::ResourceHandle::createNSURLConnection):
- 12:11 PM Changeset in webkit [202506] by
-
- 4 edits in trunk/Source/WTF
Remove now unused WTF::findNextLineStart
https://bugs.webkit.org/show_bug.cgi?id=159157
Patch by Joseph Pecoraro <Joseph Pecoraro> on 2016-06-27
Reviewed by Mark Lam.
Unused after r202498.
- wtf/text/StringImpl.cpp:
(WTF::StringImpl::findNextLineStart): Deleted.
- wtf/text/StringImpl.h:
(WTF::findNextLineStart): Deleted.
- wtf/text/WTFString.h:
(WTF::String::findNextLineStart): Deleted.
- 12:06 PM Changeset in webkit [202505] by
-
- 8 edits2 adds in trunk
[iOS] Media controls are too cramped with small video
https://bugs.webkit.org/show_bug.cgi?id=158815
<rdar://problem/26824238>
Patch by Antoine Quint <Antoine Quint> on 2016-06-27
Reviewed by Dean Jackson.
Source/WebCore:
In updateLayoutForDisplayedWidth(), we try to ensure a minimum width is guaranteed
for the progress indicator. However, we were not accounting for the width used by
the current and remaining time labels on either side of it, so we would incorrectly
conclude that we were guaranteeing the minimum time and yield incorrect layouts since
we were trying to fit more buttons than we had room for.
In order to correctly compute the available width for the progress indicator, we now
have clones of the current and remaining time labels, hidden from video and VoiceOver,
that we update along with the originals. The same styles apply to both clones and
originals, so we may measure the clones to determine the space used by the time labels.
The reason we need to use clones is that if the time labels had previously been hidden
from view, precisely because there was not enough space to display them along with the
progress indicator, then trying to obtain metrics from them would yield 0 since they had
"display: none" styles applied. In order to avoid extra layouts and possible flashing, we
use the clones so that we never have to toggle the "display" property of the originals
just to obtain their measurements.
As a result of this change, we adjust the constant used to set the minimum required
width available to display the progress indicator after all other essential controls
and labels have been measured. That constant used to account for the width of the
time labels, and this is no longer correct.
Test: media/video-controls-drop-and-restore-timeline.html
- Modules/mediacontrols/mediaControlsApple.css:
(::-webkit-media-controls-time-remaining-display.clone):
- Modules/mediacontrols/mediaControlsApple.js:
(Controller):
(Controller.prototype.createTimeClones):
(Controller.prototype.removeTimeClass):
(Controller.prototype.addTimeClass):
(Controller.prototype.updateDuration):
(Controller.prototype.updateLayoutForDisplayedWidth):
(Controller.prototype.updateTime):
(Controller.prototype.updateControlsWhileScrubbing):
- Modules/mediacontrols/mediaControlsiOS.css:
(::-webkit-media-controls-time-remaining-display.clone):
- Modules/mediacontrols/mediaControlsiOS.js:
LayoutTests:
Adjust the output of this test to account for the time label clones and add a new test.
- media/video-controls-drop-and-restore-timeline-expected.txt: Added.
- media/video-controls-drop-and-restore-timeline.html: Added.
- platform/mac-yosemite/http/tests/media/hls/video-controls-live-stream-expected.txt:
- platform/mac/http/tests/media/hls/video-controls-live-stream-expected.txt:
- 12:06 PM Changeset in webkit [202504] by
-
- 2 edits in trunk/Source/WebKit2
Discounted items are displayed with positive values
https://bugs.webkit.org/show_bug.cgi?id=159160
rdar://problem/26980772
Reviewed by Dean Jackson.
- UIProcess/ApplePay/cocoa/WebPaymentCoordinatorProxyCocoa.mm:
(WebKit::toDecimalNumber):
Add new helper function. Correctly compute the mantissa.
(WebKit::toPKPaymentSummaryItem):
(WebKit::toPKShippingMethod):
Use toDecimalNumber.
- 11:52 AM Changeset in webkit [202503] by
-
- 2 edits in trunk/Source/JavaScriptCore
Update features.json for ES6 completed features.
https://bugs.webkit.org/show_bug.cgi?id=159152
Reviewed by Mark Lam.
- features.json:
- 11:35 AM Changeset in webkit [202502] by
-
- 12 edits2 adds in trunk/Source/JavaScriptCore
B3 should not use Nops when deleting unreachable code
https://bugs.webkit.org/show_bug.cgi?id=159120
rdar://problem/26500743
Reviewed by Michael Saboff.
Prior to this change, transformations that obviated the need for some value could choose
from these ways to kill it:
- replaceWithIdentity() if we're replacing with another value.
- replaceWithNop() if the type is Void or if we know that we'll fix any users of this value.
- deleteValue() if the code is unreachable.
The bug here is that reduceStrength() was being clever about how to get rid of a value.
reduceStrength() may find a Check that must always exit. The goal is to remove any code
dominated by the Check. But it would be awkward to eagerly delete all of the blocks
dominated by this one. So this code took a much simpler approach: it would
replaceWithNop() for all of the values in this block after the Check and it would replace
the terminal with Oops.
But this corrupts the IR in a subtle way: some of those values may have been non-Void but
now they are Nops so they are Void. reduceStrength() will not yet realize that the blocks
dominated by the one with the Check are unreachable, so it will run all sorts of
optimizations on those blocks. This could have probably manifested as many different kinds
of badness, but the way I found out about this issue was through a crash in
IntRange::top(Type) when inlined into ReduceStrength::rangeFor(). We'd die in a switch
statement over a child's type.
We could fix this by making rangeFor() tolerate Void. But I think that this would be
dangerous. There could easily be other places in reduceStrength() that assume that value's
children are non-Void. So, this change fixes the Check optimization and adds mechanisms to
prevent other optimizations from breaking the children-are-not-Void rule.
This introduces two high-level changes:
- It's no longer legal to replaceWithNop() if the value is not Void. This change alone would cause reduceStrength() to instacrash in its Check optimization. Almost all other uses of replaceWithNop() were already following this rule, so they were fine. One other place was using replaceWithNop() on non-Void values after arranging for them to no longer have any parents. That was changed to call replaceWithNopIgnoringType(), which doesn't have any type assertions.
- For reduceStrength() there is a new Value::replaceWithBottom() method that works with Void or non-Void and behaves like you would want replaceWithNop() to behave: if you know that the code is unreachable then it produces something that is guaranteed to be deleted by later optimizations, and if it's not unreachable, then it's guaranteed to be compiled to something harmless and cheap. This means replacing the value with an identity that points to a bottom constant (the 0 for whatever type we have), or just replacing it with Nop if it's Void.
This also adds a test case for the reason why we do this: we may have two blocks, where
the first block unconditionally exits while dominating the second block. The second block
references values in the part of the first block that is unreachable. In trunk, this test
would assert in ReduceStrength::rangeFor() because the CheckAdd in the second block would
reference a Nop in the first block.
This fixes a high volume crash in ReduceStrength::rangeFor(). This crash was very
confusing. Even though we were crashing at a RELEASE_ASSERT_NOT_REACHED() in a switch
statement in IntRange::top(Type), clang was merging that trap with the trap it used for
Vector OOB. The top of the stack in crash dumps looked like:
JSC::B3::(anonymous namespace)::ReduceStrength::rangeFor(JSC::B3::Value*, unsigned int) + 4477 (Vector.h:655)
Where Vector.h:655 is:
OverflowHandler::overflowed();
But this crash was not at Vector.h:655. It was at B3ReduceStrength.cpp:121. The two lines
are both traps, so they got merged despite differences in debug info. This bug would have
been so much easier to fix if I had the right line number.
- b3/B3BottomProvider.h: Added. This is a utility for creating bottom values.
(JSC::B3::BottomProvider::BottomProvider):
(JSC::B3::BottomProvider::operator()):
- b3/B3InsertionSet.cpp: Optimized adding bottom values a bit. We will no longer create pointless duplicates.
(JSC::B3::InsertionSet::insertBottom):
(JSC::B3::InsertionSet::execute):
(JSC::B3::InsertionSet::bottomForType):
- b3/B3InsertionSet.h:
- b3/B3MoveConstants.cpp: Use replaceWithNopIgnoringType() because we *know* that we can replaceWithNop even for non-Void.
- b3/B3Procedure.h:
- b3/B3ReduceStrength.cpp: Use replaceWithBottom().
- b3/B3ReduceStrength.h:
- b3/B3TypeMap.h: I figured if I wrote type-casing code like this once then I'd never want to write it again.
- b3/B3Value.cpp:
(JSC::B3::Value::replaceWithIdentity):
(JSC::B3::Value::replaceWithNop):
(JSC::B3::Value::replaceWithNopIgnoringType):
- b3/B3Value.h:
- b3/B3ValueInlines.h:
(JSC::B3::Value::replaceWithBottom): This is the new method of killing unreachable code.
(JSC::B3::Value::as):
- b3/testb3.cpp: Add new tests!
(JSC::B3::testLateRegister):
(JSC::B3::testReduceStrengthCheckBottomUseInAnotherBlock):
(JSC::B3::zero):
(JSC::B3::run):
- 11:30 AM Changeset in webkit [202501] by
-
- 5 edits in branches/safari-601.1.46-branch/Source
Versioning.
- 11:29 AM Changeset in webkit [202500] by
-
- 5 edits in branches/safari-601-branch/Source
Versioning.
- 11:28 AM Changeset in webkit [202499] by
-
- 2 edits in trunk/Source/WebCore
No error message when passing an invalid API version to ApplePaySession constructor
https://bugs.webkit.org/show_bug.cgi?id=159154
Reviewed by Tim Horton.
Log an error message if the version is not supported. Also, check for version 0 since that is also not supported.
- Modules/applepay/ApplePaySession.cpp:
(WebCore::ApplePaySession::create):
- 11:18 AM Changeset in webkit [202498] by
-
- 5 edits in trunk
REGRESSION: Web Inspector: Text search broken in resources with <CR>
https://bugs.webkit.org/show_bug.cgi?id=159110
<rdar://problem/27008485>
Reviewed by Brian Burg.
Source/JavaScriptCore:
- inspector/ContentSearchUtilities.cpp:
(Inspector::ContentSearchUtilities::lineEndings):
The frontend moved to only treated newlines as line endings in
the TextEditor. The backend however was looking for many
different types of line endings (\r\n, \r, \n). This caused
the line endings to ultimately differ between the frontend
and the backend, so the frontend couldn't find the lines that
the backend was claiming search results were on. Change the
backend to only look for \n line endings.
LayoutTests:
- inspector/debugger/searchInContent-linebreaks-expected.txt:
- inspector/debugger/searchInContent-linebreaks.html:
Now that the backend responds with lines that end in \n, this test changes
the number of line results. The frontend interprets this correctly.
- 10:57 AM Changeset in webkit [202497] by
-
- 11 edits6 adds in trunk
AX: Anonymous RenderMathMLOperators are not exposed to the accessibility tree
https://bugs.webkit.org/show_bug.cgi?id=139582
<rdar://problem/26938849>
Reviewed by Chris Fleizach.
This is based on a patch by Frederic Wang <fwang@igalia.com>.
Source/WebCore:
WebCore assigns the generic MathElementRole AccessibilityRole to elements
which are expected to be included in the accessibility tree. This assignment
is based on the AccessibilityRenderObject's node being a MathMLElement. The
anonymous RenderMathMLOperators fail that test.
From the perspective of accessibility support, these operators function
like MathMLElements. Furthermore, both WebCore and the platforms rely
upon MathElementRole to identify accessible MathML objects. The simplest
fix is to have AccessibilityRenderObject::isMathElement() treat anonymous
MathML operators as if they were MathMLElements.
Now that these operators are being exposed, we need to handle them in
AccessibilityRenderObject::textUnderElement() which assumes that anonymous
objects either have nodes or have children with nodes. And crashes when
that fails to be the case. Making RenderMathMLOperator::textContent()
public and then using it to get the text under anonymous operators solves
this problem. We also assign StaticTextRole to these operators on the Mac
because the default platform mapping of MathElementRole is GroupRole, which
made sense when we had a child RenderText object holding the operator.
Lastly, AccessibilityRenderObject::isIgnoredElementWithinMathTree() no
longer needs to special-case anonymous operators because they now have
MathElementRole.
Tests: accessibility/math-fenced.html
accessibility/math-foreign-content.html
- accessibility/AccessibilityObject.h:
(WebCore::AccessibilityObject::isAnonymousMathOperator):
- accessibility/AccessibilityRenderObject.cpp:
(WebCore::AccessibilityRenderObject::textUnderElement):
(WebCore::AccessibilityRenderObject::stringValue):
(WebCore::AccessibilityRenderObject::isMathElement):
(WebCore::AccessibilityRenderObject::isAnonymousMathOperator):
(WebCore::AccessibilityRenderObject::isIgnoredElementWithinMathTree):
- accessibility/AccessibilityRenderObject.h:
- accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
(-[WebAccessibilityObjectWrapper role]):
- rendering/mathml/RenderMathMLMath.h:
- rendering/mathml/RenderMathMLOperator.h:
(WebCore::RenderMathMLOperator::textContent):
LayoutTests:
One of the things making it difficult to tell that the operators were not
exposed on the Mac is the fact that the operator text is present. But the
operator text is simply a StaticTextRole object corresponding to the
RenderText descendant of the operator. Furthermore, on the Mac, accessible
math operators have subroles which are missing from the StaticTextRole
object which is exposed. In order to make issues like this more obvious,
add an option to include the subrole to dumpAccessibilityTree() and use it
in the new math-fenced.html.
This change also fixes the mfenced in mac/mathml-elements.html that was
broken after r202420. We enable it again and update the expectations of
that test because the operators are now in the accessibility tree.
We also add a new test to verify the render tree of foreign content in
MathML formulas.
- accessibility/mac/mathml-elements-expected.txt: Updated.
- accessibility/mac/mathml-elements.html: Re-enable the mfenced test.
- accessibility/math-fenced.html: Added.
- accessibility/math-foreign-content.html: Added.
- platform/gtk/accessibility/math-fenced-expected.txt: Added.
- platform/gtk/accessibility/math-foreign-content-expected.txt: Added.
- platform/mac/accessibility/math-fenced-expected.txt: Added.
- platform/mac/accessibility/math-foreign-content-expected.txt: Added.
- resources/accessibility-helper.js: Add option to include subrole in tree.
(dumpAccessibilityTree):
- 10:57 AM Changeset in webkit [202496] by
-
- 2 edits in trunk/LayoutTests
Remove flaky expectation for imported/w3c/web-platform-tests/html/semantics/embedded-content/media-elements/event_loadeddata.html
https://bugs.webkit.org/show_bug.cgi?id=150956
Unreviewed test gardening.
- platform/mac/TestExpectations:
- 10:53 AM Changeset in webkit [202495] by
-
- 4 edits in trunk/LayoutTests
WebRTC: Unskip RTCPeerConnection-AddRemoveStream test
https://bugs.webkit.org/show_bug.cgi?id=159133
Reviewed by Eric Carlson.
Unskip RTCPeerConnection-AddRemoveStream test (on GTK+ port). RTCPeerConnection implements
the legacy MediaStream-based API as JS built-ins so we can run this test again.
- fast/mediastream/RTCPeerConnection-AddRemoveStream-expected.txt:
- fast/mediastream/RTCPeerConnection-AddRemoveStream.html:
- platform/gtk/TestExpectations:
- 10:52 AM Changeset in webkit [202494] by
-
- 1 edit1 delete in trunk/Source/WebCore
WebRTC: Remove unused RTCOfferAnswerOptionsPrivate.h
https://bugs.webkit.org/show_bug.cgi?id=159130
Reviewed by Eric Carlson.
Remove unused RTCOfferAnswerOptionsPrivate.h file.
- platform/mediastream/RTCOfferAnswerOptionsPrivate.h: Removed.
- 10:50 AM Changeset in webkit [202493] by
-
- 2 edits in trunk/Source/WebCore
Crash in layout test /media/video-buffered-range-contains-currentTime.html
https://bugs.webkit.org/show_bug.cgi?id=159109
<rdar://problem/26535750>
Reviewed by Alex Christensen.
Guard against a dealloc race condition by holding a retain on the session
until the task's _resource:loadFinishedWithError: completes, including
main thread callbacks.
- platform/network/cocoa/WebCoreNSURLSession.mm:
(-[WebCoreNSURLSessionDataTask _resource:loadFinishedWithError:]):
- 10:50 AM Changeset in webkit [202492] by
-
- 3 edits in trunk/Source/JavaScriptCore
Web Inspector: CRASH in backend at Inspector::HeapFrontendDispatcher::garbageCollected + 552 when closing frontend/inspected page
https://bugs.webkit.org/show_bug.cgi?id=159075
<rdar://problem/26094341>
Reviewed by Timothy Hatcher.
Move the asynchronous work to a task class that can be cancelled when the
heap agent is reset, disabled or destroyed.
- inspector/agents/InspectorHeapAgent.cpp:
(Inspector::SendGarbageCollectionEventsTask::SendGarbageCollectionEventsTask):
(Inspector::SendGarbageCollectionEventsTask::addGarbageCollection):
(Inspector::SendGarbageCollectionEventsTask::reset):
(Inspector::SendGarbageCollectionEventsTask::timerFired):
Added. This holds onto GarbageCollectionData that needs to be sent asynchronously.
It uses the RunLoop variant of Timer and can queue multiple collections to be sent.
The data vector is guarded with a lock so that garbageCollected() can safely add
collection data from a non-main thread while the main thread sends out events.
(Inspector::InspectorHeapAgent::InspectorHeapAgent):
(Inspector::InspectorHeapAgent::~InspectorHeapAgent):
(Inspector::InspectorHeapAgent::disable):
Reset the task when disabling or tearing down the agent so the timer doesn't fire after destruction.
(Inspector::InspectorHeapAgent::didGarbageCollect):
Add the collection data to the task, which will dispatch an event for it asynchronously.
- inspector/agents/InspectorHeapAgent.h:
- 10:44 AM Changeset in webkit [202491] by
-
- 5 edits in trunk/LayoutTests
Update TestExpectations for some dynamic MathML tests
https://bugs.webkit.org/show_bug.cgi?id=159142
Patch by Frederic Wang <fwang@igalia.com> on 2016-06-27
Reviewed by Brent Fulgham.
- TestExpectations: Enable this test again after r199980.
- platform/efl/TestExpectations: This is already claimed to pass in the platform independent TestExpectations.
- platform/gtk/TestExpectations: Ditto.
- platform/ios-simulator/TestExpectations: This test passes after the refactoring of RenderMathMLOperator.
- 10:38 AM Changeset in webkit [202490] by
-
- 7 edits in trunk
ES6 Change: Unify handling of RegExp CharacterClassEscapes \w and \W and Word Asserts \b and \B
https://bugs.webkit.org/show_bug.cgi?id=158505
Reviewed by Geoffrey Garen.
Source/JavaScriptCore:
This change makes it so that the CharacterClassEscape \w matches the inverse of
\W and vice versa for unicode, ignore case RegExp's.
Before this change, both /\w/ui and /\W/ui RegExp's would match the characters
k, K, s, S, \u017f (Latin Small Letter Long S) and \u212a (Kelvin Sign).
This was due to how the ES6 standard defined matching of character classes
specifically that the abstract operation "Canonicalize()" is called for the
character to be matched AND for the characters in the character class we are
matching against. This change is to make \W always be the inverse of \w.
It is still the case that the characters that match against \w changes
depending on a regular expression's flags.
The only real changes occur for regular expressions with both the unicode and
ignore case flags set. Updated the character class generator to make
nonwordUnicodeIgnoreCaseChar not include k, K, s, S, \u017f and \u212a.
Changed BytecodePattern.wordcharCharacterClass to use the correct
word character class for the flags. Simplfied character class set up in
in the pattern to use m_pattern.wordUnicodeIgnoreCaseCharCharacterClass and
invert as appropriate when unicode and ignore case are both set.
- create_regex_tables:
- yarr/YarrInterpreter.h:
(JSC::Yarr::BytecodePattern::BytecodePattern):
- yarr/YarrPattern.cpp:
(JSC::Yarr::YarrPatternConstructor::atomBuiltInCharacterClass):
LayoutTests:
Updated and added test cases.
- js/regexp-unicode-expected.txt:
- js/script-tests/regexp-unicode.js:
- 10:36 AM Changeset in webkit [202489] by
-
- 8 edits in trunk
Set an upper limit for the size or number of pieces of stretchy operators
https://bugs.webkit.org/show_bug.cgi?id=155434
Source/WebCore:
Patch by Frederic Wang <fwang@igalia.com> on 2016-06-27
Reviewed by Brent Fulgham.
Stretchy MathML operators can currently use an arbitrary number of extension glyphs to cover
a target size. This may result in hangs if large stretch sizes are requested. This change
only allow at most the 128 first extensions to be painted by the MathOperator class, which
should really be enough for mathematical formulas used in practice.
No new tests, already tested by very-large-stretchy-operators.
- rendering/mathml/MathOperator.cpp: Add a new kMaximumExtensionCount constant.
(WebCore::MathOperator::fillWithVerticalExtensionGlyph): Limit the number of step in this loop to kMaximumExtensionCount.
(WebCore::MathOperator::fillWithHorizontalExtensionGlyph): Ditto.
LayoutTests:
Update test expectations for very-large-stretchy-operators.
Patch by Frederic Wang <fwang@igalia.com> on 2016-06-27
Reviewed by Brent Fulgham.
- platform/efl/TestExpectations:
- platform/gtk/TestExpectations:
- platform/ios-simulator/TestExpectations:
- platform/mac/TestExpectations:
- platform/win/TestExpectations:
- 10:33 AM Changeset in webkit [202488] by
-
- 2 edits in trunk/Source/WebCore
Small refactoring MathMLInlineContainerElement::createElementRenderer
https://bugs.webkit.org/show_bug.cgi?id=159131
Reviewed by Brent Fulgham.
Many of the MathML renderer classes have been merged during the MathML refactoring. We
simplify how instances are created in MathMLInlineContainerElement::createElementRenderer
by removing duplicate createRenderer calls.
No new tests, behavior unchanged.
- mathml/MathMLInlineContainerElement.cpp:
(WebCore::MathMLInlineContainerElement::createElementRenderer):
- 9:48 AM Changeset in webkit [202487] by
-
- 2 edits1 add in trunk/Source/JavaScriptCore
DFGByteCodeParsing does not handle calling the Object constructor with no arguments correctly
https://bugs.webkit.org/show_bug.cgi?id=159117
<rdar://problem/26996781>
Reviewed by Saam Barati.
DFGByteCodeParsing always assumed there would be an argument to the Object constructor.
This is clearly not always the case and we should be able to handle it.
- dfg/DFGByteCodeParser.cpp:
(JSC::DFG::ByteCodeParser::handleConstantInternalFunction):
- tests/stress/indirect-call-object-constructor-with-no-arguments.js: Added.
(let.foo.Object.test):
- 9:01 AM Changeset in webkit [202486] by
-
- 2 edits in trunk/LayoutTests
Skip two LayoutTests that rely on mouse events on ios-simulator
Unreviewed test gardening.
- platform/ios-simulator/TestExpectations:
- 8:10 AM WebKitGTK/Gardening/Calendar edited by
- (diff)
- 7:47 AM Changeset in webkit [202485] by
-
- 2 edits in trunk/LayoutTests
[GTK] Unreviewed test gardening. Added Failure Test Expectations because of r194502 and r194513.
https://bugs.webkit.org/show_bug.cgi?id=159147
- platform/gtk/TestExpectations:
- 7:23 AM WebKitGTK/Gardening/Calendar edited by
- (diff)
- 3:57 AM Changeset in webkit [202484] by
-
- 5 edits in trunk/Source
[GTK][EFL] Build with threaded compositor enabled is broken
https://bugs.webkit.org/show_bug.cgi?id=159138
Patch by Miguel Gomez <magomez@igalia.com> on 2016-06-27
Reviewed by Carlos Garcia Campos.
Source/WebCore:
No need to set the device scale. The compositor buffer is only used for the accelerated
canvas scenario, and the device scale is always 1 there.
This change was introduced in r202421.
Covered by existing tests.
- platform/graphics/cairo/ImageBufferCairo.cpp:
(WebCore::ImageBufferData::createCompositorBuffer):
Source/WebKit2:
Replace the usage of NoncopyableFunction with Function, to follow the changes in r202439.
- Shared/CoordinatedGraphics/threadedcompositor/CompositingRunLoop.cpp:
(WebKit::CompositingRunLoop::performTask):
(WebKit::CompositingRunLoop::performTaskSync):
- Shared/CoordinatedGraphics/threadedcompositor/CompositingRunLoop.h:
- 2:22 AM Changeset in webkit [202483] by
-
- 1 edit2 adds in trunk/LayoutTests
LayoutTests/imported/w3c/web-platform-tests/html/semantics/forms/form-control-infrastructure/ is failing on ios-simulator
https://bugs.webkit.org/show_bug.cgi?id=159136
Unreviewed.
Patch by Youenn Fablet <youennf@gmail.com> on 2016-06-27
- platform/ios-simulator/imported/w3c/web-platform-tests/html/semantics/forms/form-control-infrastructure/form-expected.txt: Added.
- 2:21 AM Changeset in webkit [202482] by
-
- 2 edits in trunk/Source/WebCore
[GStreamer] top/bottom black bars added needlessly in fullscreen
https://bugs.webkit.org/show_bug.cgi?id=158980
Patch by Philippe Normand <philn@igalia.com> on 2016-06-27
Reviewed by Carlos Garcia Campos.
The natural video size calculation depends on the validity of the
current sample, so whenever the first sample reached the sink it's a
good idea to reflect this on the player which will update its natural
size accordingly.
Fixes an issue where black borders were added on top and bottom of
fullscreen video.
- platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.cpp:
(WebCore::MediaPlayerPrivateGStreamerBase::triggerRepaint):
- 2:01 AM Changeset in webkit [202481] by
-
- 2 edits in trunk/Tools
[Win] The test fast/events/wheel-event-outside-body.html is timing out.
https://bugs.webkit.org/show_bug.cgi?id=159129
Reviewed by Alex Christensen.
Implement the function mouseScrollBy in the Windows event sender.
- DumpRenderTree/win/EventSender.cpp:
(mouseScrollBy):
(mouseScrollByWithWheelAndMomentumPhasesCallback):
- 1:28 AM Changeset in webkit [202480] by
-
- 21 edits in trunk
Remove didFailRedirectCheck ThreadableLoaderClient callback
https://bugs.webkit.org/show_bug.cgi?id=159085
Patch by Youenn Fablet <youenn@apple.com> on 2016-06-27
Reviewed by Daniel Bates.
Source/WebCore:
Removing didFailRedirectCheck and using didFailAccessControlCheck instead.
The change in behavior is that additional error messages are outputted in the console.
These messages give additional debugging information.
Covered by rebased tests.
- Modules/fetch/FetchLoader.cpp: Removing didFailRedirectCheck.
- Modules/fetch/FetchLoader.h: Ditto.
- inspector/InspectorNetworkAgent.cpp: Ditto.
- loader/DocumentThreadableLoader.cpp:
(WebCore::DocumentThreadableLoader::redirectReceived): Calling didFailAccessControlCheck with information on failing
URL.
(WebCore::DocumentThreadableLoader::loadRequest): Ditto.
- loader/ThreadableLoaderClient.h: Removing didFailRedirectCheck.
- loader/ThreadableLoaderClientWrapper.h: Ditto.
- loader/WorkerThreadableLoader.cpp: Ditto.
- loader/WorkerThreadableLoader.h: Ditto.
- page/EventSource.cpp: Ditto.
- page/EventSource.h: Ditto.
- workers/WorkerScriptLoader.cpp: Ditto.
- workers/WorkerScriptLoader.h: Ditto.
- xml/XMLHttpRequest.cpp: Ditto.
- xml/XMLHttpRequest.h: Ditto.
LayoutTests:
- http/tests/security/contentSecurityPolicy/connect-src-eventsource-redirect-to-blocked-expected.txt:
- http/tests/security/contentSecurityPolicy/connect-src-xmlhttprequest-redirect-to-blocked-expected.txt:
- http/tests/security/contentSecurityPolicy/worker-csp-blocks-xhr-redirect-cross-origin-expected.txt:
- http/tests/xmlhttprequest/access-control-and-redirects-async-expected.txt:
- http/tests/xmlhttprequest/access-control-and-redirects-expected.txt:
