Unreviewed, update my list of email addresses in contributors.json.
[WebKit-https.git] / Tools / ChangeLog
index 4bb64f7..ebdf722 100644 (file)
@@ -1,3 +1,907 @@
+2014-06-10  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
+
+        Unreviewed, update my list of email addresses in contributors.json.
+
+        * Scripts/webkitpy/common/config/contributors.json:
+
+2014-06-10  Simon Fraser  <simon.fraser@apple.com>
+
+        Always switch to an sRGB screen color profile when running tests
+        https://bugs.webkit.org/show_bug.cgi?id=133712
+
+        Reviewed by Tim Horton.
+        
+        Switch the screen color profile to sRGB when running tests, because even ref
+        tests need the screen to use the correct color profile.
+
+        * DumpRenderTree/mac/LayoutTestHelper.m:
+        (installLayoutTestColorProfile):
+        * Scripts/webkitpy/port/mac.py:
+        (MacPort.start_helper):
+
+2014-06-10  Martin Hodovan  <mhodovan@partner.samsung.com>
+
+        [EFL] Bump EFL libraries to 1.9
+        https://bugs.webkit.org/show_bug.cgi?id=125479
+
+        Reviewed by Gyuyoung Kim.
+
+        Based on the work of Ryuan Choi <ryuan.choi@samsung.com>
+        and László Langó <llango.u-szeged@partner.samsung.com>.
+
+        * ImageDiff/CMakeLists.txt:
+        * MiniBrowser/efl/CMakeLists.txt:
+        * efl/install-dependencies:
+        * efl/jhbuild.modules:
+        * efl/jhbuildrc:
+
+2014-06-10  David Farler  <dfarler@apple.com>
+
+        old-run-webkit-tests: Create CoreSimulator device on demand and find it by name
+        https://bugs.webkit.org/show_bug.cgi?id=133663
+
+        Reviewed by David Kilzer.
+
+        * Scripts/old-run-webkit-tests:
+        (installAndLaunchDumpToolAppUsingNotification):
+        - Remove obsolete notification keys
+        - Key off of mandatory device UDID (a UUID)
+        - Create a device on demand via CoreSimulator if necessary
+        -- Grab the UDID from the created device plist.
+        * Scripts/webkitdirs.pm:
+        (iOSSimulatorDevicesPath): Added
+        (iOSSimulatorDevices): Added
+        (createiOSSimulatorDevice): Added
+        (deleteiOSSimulatorDevice): Added
+
+2014-06-10  Víctor Manuel Jáquez Leal  <vjaquez@igalia.com>
+
+        [GTK][gtkdoc] remove -Wcast-align
+        https://bugs.webkit.org/show_bug.cgi?id=133640
+
+        Apparently gcc warns that GParamSpec is not castable to
+        GParamSpecInt64/GParamSpecUInt64/GParamSpecDouble due they are 64bit, even
+        though ARM hackers claim that those only need 4byte alignment. As long as gcc
+        behaves that way, this warning is not very useful, also they break the Debian
+        packaging.
+
+        This patch appends the compiler flag -Wno-align-cast for the gtkdoc
+        scanner compilation, thus the compiler ignores the previous
+        -Waling-cast flag.
+
+        Reviewed by Martin Robinson.
+
+        * gtk/generate-gtkdoc:
+        (get_generator_for_config):
+
+2014-06-09  Alexey Proskuryakov  <ap@apple.com>
+
+        Disable button animations in regression tests
+        https://bugs.webkit.org/show_bug.cgi?id=133647
+        <rdar://problem/16677530>
+
+        Reviewed by Tim Horton.
+
+        * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm:
+        (WTR::InjectedBundle::platformInitialize):
+
+2014-06-09  Simon Fraser  <simon.fraser@apple.com>
+
+        Move scrolling-related tests to tiled-drawing/scrolling subdirectory, and minor cleanup
+        https://bugs.webkit.org/show_bug.cgi?id=133650
+
+        Reviewed by Tim Horton.
+        
+        No point checking for a Windows-style path in Cocoa code.
+
+        * WebKitTestRunner/TestInvocation.cpp:
+        (WTR::shouldUseThreadedScrolling):
+
+2014-06-09  Simon Fraser  <simon.fraser@apple.com>
+
+        Fix more latched scrolling test flakiness and slowness
+        https://bugs.webkit.org/show_bug.cgi?id=133601
+
+        Reviewed by Brent Fulgham.
+        
+        Latched scrolling tests were flakey for two reasons. First, the EventSender
+        wheel events were async from the web to the UI process, and the resulting
+        synthetic events also async from the UI to the web process. Make them sync
+        for testing.
+        
+        Secondly, the timer-based scrolling tree commit could result in the scrolling
+        tree (specifically the frame node's notion of the scrollable area and content
+        size) not being updated before the scrolling thread starts getting wheel events.
+        Fix by having WebPage::wheelEventSyncForTesting() commit if necessary.
+        
+        Finally remove all the now-unnecessary timeouts from the tests.
+
+        * WebKitTestRunner/mac/EventSenderProxy.mm:
+        (WTR::EventSenderProxy::mouseMoveTo):
+        (WTR::EventSenderProxy::mouseScrollBy):
+        (WTR::EventSenderProxy::mouseScrollByWithWheelAndMomentumPhases):
+
+2014-06-09  Andrzej Badowski  <a.badowski@samsung.com>
+
+        [EFL] Add parameter to MiniBrowser to disable web security
+        https://bugs.webkit.org/show_bug.cgi?id=121095
+
+        Reviewed by Gyuyoung Kim.
+
+        -w (--web-security) commandline allows to disable/enable web security (enabled by default).
+        The reason for this is to test and debug (using, for example, MiniBrowser) scripts accessing
+        resources from other domains (cross-domain requests) such as access map location on another server.
+        Another possibility is to launch in a webkit browser local file tests that will refer to the links
+        in the http website loaded in an iframe. It will not be possible without deactivation of web security.
+
+        * MiniBrowser/efl/main.c:
+        (window_create):
+        (elm_main):
+
+2014-06-06  Simon Fraser  <simon.fraser@apple.com>
+
+        32-bit build fix.
+
+        * WebKitTestRunner/mac/EventSenderProxy.mm:
+        (WTR::EventSenderProxy::mouseMoveTo):
+        (WTR::EventSenderProxy::mouseScrollByWithWheelAndMomentumPhases):
+
+2014-06-06  Simon Fraser  <simon.fraser@apple.com>
+
+        Fix builds on 10.8; -firstObject isn't declared on NSArray there.
+
+        * DumpRenderTree/mac/DumpRenderTree.mm:
+        (createWebViewAndOffscreenWindow):
+        * DumpRenderTree/mac/EventSendingController.mm:
+        (-[EventSendingController mouseScrollByX:andY:continuously:]):
+        (-[EventSendingController mouseScrollByX:andY:withWheel:andMomentumPhases:]):
+        * WebKitTestRunner/mac/EventSenderProxy.mm:
+        (WTR::EventSenderProxy::mouseScrollBy):
+        (WTR::EventSenderProxy::mouseScrollByWithWheelAndMomentumPhases):
+
+2014-06-06  Simon Fraser  <simon.fraser@apple.com>
+
+        Latched scrolling tests are flakey on Mavericks
+        https://bugs.webkit.org/show_bug.cgi?id=133578
+        <rdar://problem/17180591>
+
+        Reviewed by Brent Fulgham.
+        
+        The latched scrolling tests rely on synthetic mousewheel events whose coordinates can
+        be wrong on machines with more than one screen. This is because we use
+        [NSScreen mainScreen] when computing the global coordinates of the events, but
+        [NSScreen mainScreen] returns the screen with the key window, not the first screen.
+        
+        So replace calls to [NSScreen mainScreen] with [[NSScreen screens] firstObject]. Also
+        add comments clarifying why the coordinate math works (which is not obvious given
+        the events have no target window).
+        
+        Some scrolling tests also dispatched events outside the 800x600 test window, so
+        fix those tests, and add logging that will appear in test output when this happens.
+
+        * DumpRenderTree/mac/DumpRenderTree.mm:
+        (createWebViewAndOffscreenWindow):
+        * DumpRenderTree/mac/EventSendingController.mm:
+        (-[EventSendingController mouseScrollByX:andY:continuously:]):
+        (-[EventSendingController mouseScrollByX:andY:withWheel:andMomentumPhases:]):
+        * WebKitTestRunner/mac/EventSenderProxy.mm:
+        (WTR::EventSenderProxy::mouseMoveTo):
+        (WTR::EventSenderProxy::mouseScrollBy):
+        (WTR::EventSenderProxy::continuousMouseScrollBy):
+        (WTR::EventSenderProxy::mouseScrollByWithWheelAndMomentumPhases):
+
+2014-06-05  Carlos Alberto Lopez Perez  <clopez@igalia.com>
+
+        [GTK] run-perf-tests should run the tests inside the jhbuild shell.
+        https://bugs.webkit.org/show_bug.cgi?id=133533
+
+        Reviewed by Carlos Garcia Campos.
+
+        * Scripts/webkitpy/common/multiprocessing_bootstrap.py:
+        (run): Remove the jhbuild-wrapper wrapper from here. It should be
+        on Scripts/webkitpy/port/gtk.py.
+        * Scripts/webkitpy/port/base.py:
+        (Port.set_option): Add a set_option(name, value) function to set
+        the optparse parameter "name" to the value "value", even if the
+        parameter "name" is already initialized.
+        * Scripts/webkitpy/port/gtk.py:
+        (GtkPort.__init__): Set the default wrapper to jhbuild-wrapper.
+        If another wrapper is specified via the option "-l --wrapper=x",
+        then prepend the jhbuild-wrapper to the one specified.
+
+2014-06-04  David Farler  <dfarler@apple.com>
+
+        Build ASan WebKit for iOS Simulator
+        https://bugs.webkit.org/show_bug.cgi?id=133489
+
+        Reviewed by Darin Adler.
+
+        * DumpRenderTree/mac/Configurations/DumpRenderTree.xcconfig:
+        * DumpRenderTree/mac/Configurations/DumpRenderTreeApp.xcconfig:
+        - Add missing ASan LDFLAGS.
+        * asan/asan.xcconfig:
+        - Split compiler-rt linker flags for OS X and iOS Simulator
+
+2014-06-03  Yoav Weiss  <yoav@yoav.ws>
+
+        Add myself (yoav@yoav.ws) to contributors.json
+        https://bugs.webkit.org/show_bug.cgi?id=133492
+
+        * Scripts/webkitpy/common/config/contributors.json:
+
+2014-06-03  Eva Balazsfalvi  <evab.u-szeged@partner.samsung.com>
+
+        REGRESSION(r169092 and r169102): Skip failing JSC tests poperly on non-x86 Darwin platforms
+        https://bugs.webkit.org/show_bug.cgi?id=133149
+
+        Reviewed by Csaba Osztrogonác.
+
+        * Scripts/run-jsc-stress-tests: Added determineOS to determine the host operating system the script run on.
+
+2014-06-02  Tibor Meszaros  <tmeszaros.u-szeged@partner.samsung.com>
+
+        Add a verbose flag to binding tests
+        https://bugs.webkit.org/show_bug.cgi?id=122303
+
+        Reviewed by Csaba Osztrogonác.
+
+        This patch will:
+        - Add a verbose keyword argument to the BindingsTests constructor,
+        - Add --no-verbose option to the run-bindigs-test.
+        Verbose mode, which is on by default, prints positive results to stdout as before.
+        When the --no-verbose option added to run-bindings-tests, positive results are not printed at all.
+
+        Original patch from Matthew Hanson <matthew_hanson@apple.com>
+
+        * Scripts/run-bindings-tests:
+        (main):
+        * Scripts/webkitpy/bindings/main.py:
+        (BindingsTests.__init__):
+        (BindingsTests.detect_changes):
+
+2014-06-01  Ryosuke Niwa  <rniwa@webkit.org>
+
+        Rename DoYouEvenBench 0.17 to Speedometer 1.0 and add a new look.
+        https://bugs.webkit.org/show_bug.cgi?id=133455
+
+        Reviewed by Timothy Hatcher.
+
+        Since the perf. dashboard doesn't support following renamed benchmarks,
+        keep reporting Speedometer results as DoYouEvenBench results.
+
+        * Scripts/webkitpy/performance_tests/perftestsrunner.py:
+        (_generate_results_dict):
+
+2014-05-31  Darin Adler  <darin@apple.com>
+
+        Viewport percentage tests that resize the viewport are flaky
+        https://bugs.webkit.org/show_bug.cgi?id=133351
+
+        Reviewed by Anders Carlsson.
+
+        * DumpRenderTree/mac/DumpRenderTree.mm:
+        (sizeWebViewForCurrentTest): Added a call to reset the window origin, which
+        happens to be stored in the UI delegate object. This is part of resetting the
+        window position.
+
+        * DumpRenderTree/mac/UIDelegate.h: Store the window origin rather than storing
+        a frame rectangle. The size is stored in the size of the web view itself.
+        * DumpRenderTree/mac/UIDelegate.mm:
+        (-[UIDelegate resetWindowOrigin]): Added.
+        (-[UIDelegate webView:setFrame:]): Changed to set the size of the web view. The
+        old version would store the frame, but not resize the web view!
+        (-[UIDelegate webViewFrame:]): Get the size from the web view.
+
+        * WebKitTestRunner/efl/PlatformWebViewEfl.cpp:
+        (WTR::PlatformWebView::setWindowFrame): Added code to resize the web view.
+
+        * WebKitTestRunner/gtk/PlatformWebViewGtk.cpp:
+        (WTR::PlatformWebView::resizeTo): Added a comment about missing code to resize
+        the window.
+
+        * WebKitTestRunner/ios/PlatformWebViewIOS.mm:
+        (WTR::PlatformWebView::resizeTo): Changed this to call setWindowFrame so the
+        window gets resized too, not just the web view.
+        (WTR::PlatformWebView::setWindowFrame): Added code to set the size of the web view.
+
+        * WebKitTestRunner/mac/PlatformWebViewMac.mm:
+        (WTR::PlatformWebView::resizeTo): Changed this to call setWindowFrame so the
+        window gets resized too, not just the web view.
+        (WTR::PlatformWebView::setWindowFrame): Added code to set the size of the web view.
+
+2014-05-30  Michael Saboff  <msaboff@apple.com>
+
+        Change run-jsc-stress-tests to not resolve jsc path locally when creating a bundle
+        https://bugs.webkit.org/show_bug.cgi?id=133409
+
+        Reviewed by Filip Pizlo.
+
+        Handle creating a tarball the same way as running remote when processing
+        paths in prepareBundle.
+
+        * Scripts/run-jsc-stress-tests:
+
+2014-05-30  Mario Sanchez Prada  <mario.prada@samsung.com>
+
+        [ATK] Deprecate usage of logAccessibilityEvents() in layout tests
+        https://bugs.webkit.org/show_bug.cgi?id=132280
+
+        Reviewed by Chris Fleizach.
+
+        Remove implementation of logAccessibilityEvents() for ATK based
+        ports, as this has proved to be a flaky method of checking that
+        ATK events are being properly emitted. Instead, it's better to use
+        the addNotificationListener() functions and provide a JS callback
+        to check that the right events are being emmitted from the tests.
+
+        * WebKitTestRunner/InjectedBundle/atk/AccessibilityControllerAtk.cpp:
+        (WTR::AccessibilityController::logAccessibilityEvents): Replaced
+        with dummy implementation.
+        * WebKitTestRunner/InjectedBundle/atk/AccessibilityNotificationHandlerAtk.cpp:
+        (WTR::AccessibilityNotificationHandler::logAccessibilityEvents): Removed
+
+2014-05-30  Mario Sanchez Prada  <mario.prada@samsung.com>
+
+        [ATK] Deprecate usage of logAccessibilityEvents() in layout tests
+        https://bugs.webkit.org/show_bug.cgi?id=132280
+
+        Rubber stamped by Carlos Garcia Campos.
+
+        Fix two small issues with previous patch (r169483), that caused some tests to break.
+
+        * WebKitTestRunner/InjectedBundle/atk/AccessibilityNotificationHandlerAtk.cpp:
+        (WTR::AccessibilityNotificationHandler::connectAccessibilityCallbacks): Removed
+        duplicated entry.
+        * WebKitTestRunner/InjectedBundle/atk/AccessibilityUIElementAtk.cpp:
+        (WTR::AccessibilityUIElement::setSelectedTextRange): Fix boolean condition.
+
+2014-05-30  Jarek Czekalski  <jarekczek@poczta.onet.pl>
+
+        [ATK] Added new accessibility signal: text-caret-moved.
+        Make AccessibilityUIElement.setSelectedTextRange work
+        as atk_text_set_caret_offset when given offset == 0.
+        https://bugs.webkit.org/show_bug.cgi?id=132527
+
+        Reviewed by Mario Sanchez Prada.
+
+        text-caret-moved signal in accessibilityController is needed to
+        implement tests from Tools/TestWebKitAPI/Tests/WebKitGtk/testatk.c
+        as LayoutTests. Add the signal to the controller.
+
+        * WebKitTestRunner/InjectedBundle/Bindings/AccessibilityUIElement.idl:
+        * WebKitTestRunner/InjectedBundle/AccessibilityUIElement.cpp:
+        * WebKitTestRunner/InjectedBundle/AccessibilityUIElement.h:
+        * WebKitTestRunner/InjectedBundle/ios/AccessibilityUIElementIOS.mm:
+        * WebKitTestRunner/InjectedBundle/mac/AccessibilityUIElementMac.mm:
+        (setSelectedTextRange): Change return type to bool.
+        * WebKitTestRunner/InjectedBundle/atk/AccessibilityUIElementAtk.cpp:
+        (setSelectedTextRange): Call atk_text_set_caret_offset and return bool.
+        * WebKitTestRunner/InjectedBundle/atk/AccessibilityNotificationHandlerAtk.cpp:
+        (axObjectEventListener): Add notification name AXTextCaretMoved, allow
+        for extra arguments passed to handler and use it for text-caret-moved offset.
+        (WTR::AccessibilityNotificationHandler::connectAccessibilityCallbacks):
+        Connect callback to new listener for ATK:AtkText:text-caret-moved.
+        Refactor so that failures be reported.
+        Initialize atk interfaces before trying to connect to them, see
+        https://bugzilla.gnome.org/show_bug.cgi?id=729922
+        (WTR::AccessibilityNotificationHandler::disconnectAccessibilityCallbacks):
+        Simplify code by using a vector in place of dozen variables.
+        * DumpRenderTree/atk/AccessibilityCallbacksAtk.cpp:
+        Comment about not synchronizing with accessibilityController.
+
+2014-05-29  Carlos Alberto Lopez Perez  <clopez@igalia.com>
+
+        [GTK] Add a WebKit2 Performance bot.
+        https://bugs.webkit.org/show_bug.cgi?id=132862
+
+        Reviewed by Csaba Osztrogonác.
+
+        This adds a GTK WebKit2 perf bot.
+
+        * BuildSlaveSupport/build.webkit.org-config/config.json:
+
+2014-05-29  Dan Bernstein  <mitz@apple.com>
+
+        Fixed the WebKitTestRunner build after r169452.
+
+        * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
+        (WTR::InjectedBundle::initialize):
+
+2014-05-28  Gwang Yoon Hwang  <yoon@igalia.com>
+
+        Unreviewed. Update my email addresses in contributors.json.
+
+        * Scripts/webkitpy/common/config/contributors.json:
+
+2014-05-27  Jon Honeycutt  <jhoneycutt@apple.com>
+
+        Need an API test for bug #133193 (r169315)
+        <https://bugs.webkit.org/show_bug.cgi?id=133324>
+
+        This adds an API test that ensures that calling "stop loading" during a
+        "did fail provisional load" callback doesn't crash.
+
+        Reviewed by Andy Estes.
+
+        * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
+        Add new files to project.
+
+        * TestWebKitAPI/Tests/WebKit2/StopLoadingDuringDidFailProvisionalLoad.cpp: Added.
+        (TestWebKitAPI::didReceiveMessageFromInjectedBundle):
+        If the message is the "test done" message, record that.
+        (TestWebKitAPI::setInjectedBundleClient):
+        Register a function for the "did receive message from injected bundle"
+        callback.
+        (TestWebKitAPI::didFailProvisionalLoadWithErrorForFrame):
+        Assert that we received the message from the injected bundle, and set
+        'done' to finish the test.
+        (TestWebKitAPI::TEST):
+        Create a WebView, register a page loader client for the "did fail
+        provisional load" callback, and load a non-existent resource.
+
+        * TestWebKitAPI/Tests/WebKit2/StopLoadingDuringDidFailProvisionalLoad_bundle.cpp: Added.
+        (TestWebKitAPI::StopLoadingDuringDidFailProvisionalLoadTest::StopLoadingDuringDidFailProvisionalLoadTest):
+        (TestWebKitAPI::didFailProvisionalLoadWithErrorForFrameCallback):
+        Tell the StopLoadingDuringDidFailProvisionalLoadTest object that the
+        provisional load failed.
+        (TestWebKitAPI::StopLoadingDuringDidFailProvisionalLoadTest::didCreatePage):
+        Register a page loader client for the "did fail provisional load"
+        callback.
+        (TestWebKitAPI::StopLoadingDuringDidFailProvisionalLoadTest::didFailProvisionalLoad):
+        Call WKBundlePageStopLoading() during the callback. This not crashing
+        is what we're ensuring with this test.
+
+2014-05-28  Carlos Garcia Campos  <cgarcia@igalia.com>
+
+        [GTK] Make it possible to build with FTL enabled
+        https://bugs.webkit.org/show_bug.cgi?id=133219
+
+        Reviewed by Philippe Normand.
+
+        * gtk/jhbuild-optional.modules: Add llvm from svn and build the
+        same revision than mac.
+
+2014-05-27  Michael Saboff  <msaboff@apple.com>
+
+        REGRESSION(r169265): sh: line 0: cd: ../.vm
+        https://bugs.webkit.org/show_bug.cgi?id=133242
+
+        Rubber stamped by Oliver Hunt.
+
+        Changed runAndMonitorTestRunnerCommand() to pass mysys as a string instead of an
+        array since the command may contain character escapes that don't work with the
+        array form of system().
+
+        * Scripts/run-jsc-stress-tests:
+
+2014-05-27  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
+
+        [EFL] Change download url of freetype6
+        https://bugs.webkit.org/show_bug.cgi?id=133303
+
+        Reviewed by Csaba Osztrogonác.
+
+        freetype6 tarball isn't downloaded in Korea. So, we need to
+        change the url for now.
+
+        * efl/jhbuild.modules:
+
+2014-05-27  Tibor Meszaros  <tmeszaros.u-szeged@partner.samsung.com>
+
+        Remove BLOB guards
+        https://bugs.webkit.org/show_bug.cgi?id=132863
+
+        Reviewed by Csaba Osztrogonác.
+
+        * Scripts/webkitperl/FeatureList.pm:
+
+2014-05-27  Alberto Garcia  <berto@igalia.com>
+
+        Remove leftover comment from r165182.
+
+        Rubber stamped by Carlos Garcia Campos.
+
+        * TestWebKitAPI/Tests/WebKit2Gtk/TestMultiprocess.cpp:
+
+2014-05-26  Tibor Meszaros  <tmeszaros.u-szeged@partner.samsung.com>
+
+        [EFL] Add LLVM as optional jhbuild module for FTL developers
+        https://bugs.webkit.org/show_bug.cgi?id=133216
+
+        Reviewed by Gyuyoung Kim.
+
+        * efl/jhbuild.modules:
+
+2014-05-26  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
+
+        Fix test failures on python test
+        https://bugs.webkit.org/show_bug.cgi?id=133297
+
+        Reviewed by Dirk Schulze.
+
+        2 tests are failing because of unbounded local variable reference in test_importer.py.
+
+        * Scripts/webkitpy/w3c/test_importer.py: Move local variable declaration to top.
+        (TestImporter.import_tests):
+
+2014-05-26  Carlos Alberto Lopez Perez  <clopez@igalia.com>
+
+        [GTK] GObject DOM bindings API break test fails in the tests bots
+        https://bugs.webkit.org/show_bug.cgi?id=133211
+
+        Reviewed by Martin Robinson.
+
+        When generating the built product archive on platform GTK, we need
+        to copy also the file gtkdoc-webkitdom.cfg and the directory
+        DerivedSources/webkitdom.
+
+        * BuildSlaveSupport/built-product-archive:
+        (archiveBuiltProduct):
+
+2014-05-26  Commit Queue  <commit-queue@webkit.org>
+
+        Unreviewed, rolling out r169122.
+        https://bugs.webkit.org/show_bug.cgi?id=133288
+
+        Broke CJK font fallback (Requested by ap on #webkit).
+
+        Reverted changeset:
+
+        "[Mac] WebProcess doesn't follow localization of UI process
+        when run as a service"
+        https://bugs.webkit.org/show_bug.cgi?id=133126
+        http://trac.webkit.org/changeset/169122
+
+2014-05-26  Carlos Garcia Campos  <cgarcia@igalia.com>
+
+        [GTK] WebProcess leaked when closing pages with network process enabled
+        https://bugs.webkit.org/show_bug.cgi?id=129684
+
+        Reviewed by Anders Carlsson.
+
+        Enable the test to check that web processes finish when the web
+        view is destroyed.
+
+        * TestWebKitAPI/Tests/WebKit2Gtk/TestMultiprocess.cpp:
+
+2014-05-26  Carlos Garcia Campos  <cgarcia@igalia.com>
+
+        REGRESSION(r164632): [GTK] Crash in generate-gtkdoc when rebasing docs
+        https://bugs.webkit.org/show_bug.cgi?id=133279
+
+        Reviewed by Philippe Normand.
+
+        generator.saw_warnings is only set by GTKDoc::generate(), so it
+        should only be used after generate.
+
+        * gtk/generate-gtkdoc:
+        (generate_documentation_for_config): Do not check
+        generator.saw_warnings again, simply return the value returned by
+        generate_doc() which is generator.saw_warnings. When rebasing,
+        returns always False since there aren't warnings.
+
+2014-05-26  Youenn Fablet  <youenn.fablet@crf.canon.fr>
+
+        W3C test importer should catch exceptions coming from HTML/CSS conversion
+        https://bugs.webkit.org/show_bug.cgi?id=133176
+
+        Reviewed by Darin Adler.
+
+        Exceptions raised when converting HTML/CSS are caught.
+        Files for which conversion failed are still copied to the target directory.
+        Each individual conversion failure is logged.
+        Total conversion failure number is logged at the end of the import process.
+
+        * Scripts/webkitpy/w3c/test_importer.py:
+        (TestImporter.import_tests):
+
+2014-05-26  Carlos Alberto Lopez Perez  <clopez@igalia.com>
+
+        [GTK] Allow to run the tests on the native X display.
+        https://bugs.webkit.org/show_bug.cgi?id=133157
+
+        Reviewed by Benjamin Poulain.
+
+        This adds a new driver (xorgdriver) that runs the tests on the
+        X display referenced by the environment variable DISPLAY.
+
+        This new driver will be activated only if the environment variable
+        USE_NATIVE_XDISPLAY is defined. This can be used both for layout
+        tests and performance tests.
+
+        This patch also makes the script run-perf-tests to check the
+        system dependencies before starting the tests. Previously this
+        was not checked, and if the system dependencies were not met,
+        the script tried to execute the tests anyway, causing massive
+        failures. For example, if you had Xvfb not installed and you
+        wanted to use the Xvfb driver (the default on GTK and EFL),
+        run-perf-tests would not abort.
+
+        * Scripts/webkitpy/performance_tests/perftestsrunner.py:
+        (PerfTestsRunner._parse_args): Check the system dependencies of
+        the driver before starting the tests.
+        * Scripts/webkitpy/port/driver.py:
+        (Driver): Implement generic check_driver method.
+        (Driver.check_driver):
+        (Driver.check_driver.implementation):
+        * Scripts/webkitpy/port/gtk.py:
+        (GtkPort._driver_class): Check for environment variable
+        USE_NATIVE_XDISPLAY to decide if the Xorg driver should be used.
+        * Scripts/webkitpy/port/xorgdriver.py: Added.
+        (XorgDriver): Implement Xorg driver.
+        (XorgDriver.check_driver):
+        (XorgDriver._start):
+
+2014-05-26  Carlos Alberto Lopez Perez  <clopez@igalia.com>
+
+        REGRESSION(r166798): [GTK] [EFL] kill-old-processes should not kill dbus-daemon
+        https://bugs.webkit.org/show_bug.cgi?id=133215
+
+        Reviewed by Csaba Osztrogonác.
+
+        This reverts r166798.
+
+        * BuildSlaveSupport/kill-old-processes:
+        (main):
+
+2014-05-26  Michał Pakuła vel Rutka  <m.pakula@samsung.com>
+
+        [EFL] Turn on ENABLE_CSS_FILTERS in build script
+        https://bugs.webkit.org/show_bug.cgi?id=133278
+
+        Reviewed by Gyuyoung Kim.
+
+        * Scripts/webkitperl/FeatureList.pm: Add a build script condition missing in r169172.
+
+2014-05-26  Eva Balazsfalvi  <evab.u-szeged@partner.samsung.com>
+
+        run-javascriptcore-tests warning fix on non Mac platforms
+        https://bugs.webkit.org/show_bug.cgi?id=133260
+
+        Reviewed by Darin Adler.
+
+        * Scripts/webkitdirs.pm:
+        (argumentsForConfiguration):
+
+2014-05-23  Michael Saboff  <msaboff@apple.com>
+
+        REGRESSION(r169265): sh: line 0: cd: ../.vm
+        https://bugs.webkit.org/show_bug.cgi?id=133242
+
+        Rubber stamped by Geoffrey Garen.
+
+        Added missing shell escapes '\' required as part of previous fix.
+
+        * Scripts/run-jsc-stress-tests:
+
+2014-05-23  Ryosuke Niwa  <rniwa@webkit.org>
+
+        Revert the erroneous change committed in r169286.
+
+        * Scripts/webkitpy/performance_tests/perftest.py:
+        (PerfTest._run_with_driver):
+
+2014-05-23  Geoffrey Garen  <ggaren@apple.com>
+
+        run-perf-tests should ignore whitespace lines when snarfing test output
+        https://bugs.webkit.org/show_bug.cgi?id=133238
+
+        Reviewed by Ryosuke Niwa.
+
+        Required for PerformanceTests/Canvas/terrain.html.
+
+        * Scripts/webkitpy/performance_tests/perftest.py:
+        (PerfTest._run_with_driver):
+        (PerfTest):
+
+2014-05-23  Michael Saboff  <msaboff@apple.com>
+
+        REGRESSION (r169241) Crash running run-jsc-stress-tests on EFL ARM
+        https://bugs.webkit.org/show_bug.cgi?id=133223
+
+        Reviewed by Geoffrey Garen.
+
+        Change runAndMonitorTestRunnerCommand() to convert the *cmd argument to a
+        string before calling IO.popen() to work with older versions of ruby.
+        Changed the remote path in runShellTestRunner to quote the command
+        string we want to run via ssh so that it will work correctly with array
+        to string conversion needed for popen.
+
+        * Scripts/run-jsc-stress-tests:
+
+2014-05-23  Zsolt Borbely  <zsborbely.u-szeged@partner.samsung.com>
+
+        Remove 'css-shaders' option from the build-webkit
+        https://bugs.webkit.org/show_bug.cgi?id=133212
+
+        Reviewed by Daniel Bates.
+
+        Original CSS shaders removed in r162644.
+
+        * Scripts/webkitperl/FeatureList.pm:
+
+2014-05-23  Carlos Garcia Campos  <cgarcia@igalia.com>
+
+        [GTK] GObject DOM bindings API break test should fail in the bots when the command fails to run
+        https://bugs.webkit.org/show_bug.cgi?id=133209
+
+        Reviewed by Sergio Villar Senin.
+
+        Return FAILURE when the command failed to run.
+
+        * BuildSlaveSupport/build.webkit.org-config/master.cfg:
+        (RunGtkWebKitGObjectDOMBindingsAPIBreakTests.evaluateCommand):
+
+2014-05-23  Alberto Garcia  <berto@igalia.com>
+
+        [GTK] Install libharfbuzz-dev in Tools/gtk/install-dependencies
+        https://bugs.webkit.org/show_bug.cgi?id=133210
+
+        Reviewed by Carlos Garcia Campos.
+
+        * gtk/install-dependencies:
+
+2014-05-22  Benjamin Poulain  <bpoulain@apple.com>
+
+        Add Yusuke Suzuki to the list of contributors (for real)
+
+        Wrong section...
+
+        * Scripts/webkitpy/common/config/contributors.json:
+
+2014-05-22  Benjamin Poulain  <bpoulain@apple.com>
+
+        Add Yusuke Suzuki to the list of contributors
+
+        * Scripts/webkitpy/common/config/contributors.json:
+
+2014-05-22  Michael Saboff  <msaboff@apple.com>
+
+        Eliminate n/total progress update from run-jsc-stress-tests output to file
+        https://bugs.webkit.org/show_bug.cgi?id=133191
+
+        Reviewed by Geoffrey Garen.
+
+        Changed progressMeter setting to be based on stdout being a tty instead of
+        stdin.  Unified the processing of output from the shell runner to match the
+        same processing as is used by the makefile runner.  As part of this, the 
+        shell runner script was simplified.  It now forwards the output of each
+        test_script just like the Makefile and doesn't provide its own progress.
+
+        * Scripts/jsc-stress-test-helpers/shell-runner.sh:
+        * Scripts/run-jsc-stress-tests:
+
+2014-05-22  Michael Saboff  <msaboff@apple.com>
+
+        Add option to run-jsc-stress-tests to use installed jsc
+        https://bugs.webkit.org/show_bug.cgi?id=133102
+
+        Reviewed by Geoffrey Garen.
+
+        Added --no-copy option to not copy a JavaScriptCore framework, but use the one
+        provied with the --jsc option.
+
+        * Scripts/run-jsc-stress-tests:
+
+2014-05-21  Zsolt Borbely  <zsborbely.u-szeged@partner.samsung.com>
+
+        [EFL] Add include path of compact_unwind_encoding.h if FTL JIT is enabled
+        https://bugs.webkit.org/show_bug.cgi?id=132907
+
+        Reviewed by Gyuyoung Kim.
+
+        We need the mach-o/compact_unwind_encoding.h header from libc++abi-dev package
+        in JavaScriptCore/ftl/FTLUnwindInfo.cpp when FTL JIT is enabled.
+
+        * efl/install-dependencies:
+
+2014-05-21  Eva Balazsfalvi  <evab.u-szeged@partner.samsung.com>
+
+        REGRESSION(r169092): It broke run-jsc-stress tests on non Mac platforms
+        https://bugs.webkit.org/show_bug.cgi?id=133145
+
+        Reviewed by Csaba Osztrogonác.
+
+        * Scripts/run-jsc-stress-tests:
+
+2014-05-20  Alexey Proskuryakov  <ap@apple.com>
+
+        [Mac] WebProcess doesn't follow localization of UI process when run as a service
+        https://bugs.webkit.org/show_bug.cgi?id=133126
+        <rdar://problem/13396515>
+        <rdar://problem/16561084>
+
+        Reviewed by Sam Weinig and Tim Horton.
+
+        * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm:
+        (WTR::InjectedBundle::platformInitialize): It was too late to override language
+        here, and it's not necessary any more. Keeping the code for older OS versions in
+        case it did have effect in some code paths.
+
+2014-05-19  David Farler  <dfarler@apple.com>
+
+        Move WebKitTestRunner to std::unique_ptr
+        https://bugs.webkit.org/show_bug.cgi?id=133081
+
+        Reviewed by Daniel Bates.
+
+        * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
+        (WTR::InjectedBundle::didCreatePage):
+        * WebKitTestRunner/InjectedBundle/InjectedBundle.h:
+        * WebKitTestRunner/TestController.cpp:
+        (WTR::TestController::initialize):
+        (WTR::TestController::createWebViewWithOptions):
+        (WTR::TestController::resetStateToConsistentValues):
+        (WTR::TestController::runTest):
+        * WebKitTestRunner/TestController.h:
+        * WebKitTestRunner/TestInvocation.cpp:
+        * WebKitTestRunner/TestInvocation.h:
+        * WebKitTestRunner/WorkQueueManager.cpp:
+        (WTR::WorkQueueManager::processWorkQueue):
+        (WTR::WorkQueueManager::enqueue):
+        * WebKitTestRunner/WorkQueueManager.h:
+
+2014-05-20  Carlos Garcia Campos  <cgarcia@igalia.com>
+
+        [GTK] WebKitWebPage::send-request always pass a valid pointer for redirected response
+        https://bugs.webkit.org/show_bug.cgi?id=133119
+
+        Reviewed by Sergio Villar Senin.
+
+        Check that redirect response parameter of
+        WebKitWebPage::send-request signal is NULL when not redirecting
+        and a valid WebKitURIResponse object when redirecting.
+
+        * TestWebKitAPI/Tests/WebKit2Gtk/TestResources.cpp:
+        (testWebResourceSendRequest):
+        (serverCallback):
+        * TestWebKitAPI/Tests/WebKit2Gtk/WebExtensionTest.cpp:
+        (sendRequestCallback):
+
+2014-05-20  Carlos Garcia Campos  <cgarcia@igalia.com>
+
+        REGRESSION(CMAKE): [GTK] Several WebKit2GTK+ unit tests are no longer built
+        https://bugs.webkit.org/show_bug.cgi?id=133117
+
+        Reviewed by Philippe Normand.
+
+        Bring back missing tests to the build.
+
+        * TestWebKitAPI/Tests/WebKit2Gtk/CMakeLists.txt:
+
+2014-05-19  Filip Pizlo  <fpizlo@apple.com>
+
+        Take care of some ARM64 test failures
+        https://bugs.webkit.org/show_bug.cgi?id=133090
+
+        Reviewed by Geoffrey Garen.
+
+        * Scripts/run-jsc-stress-tests: Make it easier to conditionally skip things.
+
+2014-05-19  David Kilzer  <ddkilzer@apple.com>
+
+        [iOS] REGRESSION (r168910): webkitdirs.pm: Be more forgiving when parsing --sdk, --device and --sim
+        <http://webkit.org/b/133089>
+
+        Reviewed by Darin Adler.
+
+        * Scripts/webkitdirs.pm:
+        (determineXcodeSDK): If both '--sdk SDKROOT' and '--sim' are
+        specified on the command line, we would remove the '--sdk'
+        arguments, but leave '--sim', potentially passing it along to
+        xcodebuild causing a build failure.  This happens when
+        'old-run-webkit-tests --release --sim' invokes
+        build-dumprendertree when DumpRenderTree hasn't been built yet.
+
+2014-05-19  David Farler  <dfarler@apple.com>
+
+        Dashboard: OpenSource revision key changed from WebKitOpenSource to WebKit
+        https://bugs.webkit.org/show_bug.cgi?id=133086
+
+        Reviewed by Daniel Bates.
+
+        * BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/BuildbotIteration.js:
+
 2014-05-19  David Kilzer  <ddkilzer@apple.com>
 
         REGRESSION (r169042): Fix run-api-tests on Mountain Lion