2015-11-20 Said Abou-Hallawa Unreviewed, add watchlist for PerformanceTests/Animometer and add myself. * Scripts/webkitpy/common/config/watchlist: 2015-11-20 Alexey Proskuryakov Went a bit too far, revert part of the previous patch. * BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Images/Mavericks.png: Added. * BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Images/Mavericks@2x.png: Added. * BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/Dashboard.js: * BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Styles/Main.css: 2015-11-20 Alexey Proskuryakov Remove Mavericks bots. Rubber-stamped by Lucas Forschler. * BuildSlaveSupport/build.webkit.org-config/config.json: * BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py: * BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Images/Mavericks.png: Removed. * BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Images/Mavericks@2x.png: Removed. * BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/BubbleQueueServer.js: * BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/Dashboard.js: * BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/WebKitBuildbot.js: * BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Styles/Main.css: * BuildSlaveSupport/build.webkit.org-config/templates/root.html: * BuildSlaveSupport/build.webkit.org-config/wkbuild.py: * BuildSlaveSupport/build.webkit.org-config/wkbuild_unittest.py: * Scripts/copy-webkitlibraries-to-product-directory: 2015-11-20 Said Abou-Hallawa Update the patch and the plan files of the graphics benchmark with the latest revision https://bugs.webkit.org/show_bug.cgi?id=151503 Reviewed by Ryosuke Niwa. The time for each graphics test has been shortened from 30 seconds to 10 seconds. We need to update the graphics benchmark patch and plan files so we can get the results faster. Also this will allow running the benchmark also on the iOS performance bots. * Scripts/webkitpy/benchmark_runner/data/patches/Animometer.patch: * Scripts/webkitpy/benchmark_runner/data/plans/animometer.plan: 2015-11-20 Michael Catanzaro Unreviewed, fix installation of ninja-build after the previous commit * gtk/install-dependencies: 2015-11-20 Michael Catanzaro [GTK] install-dependencies should run installer exactly once https://bugs.webkit.org/show_bug.cgi?id=151477 Reviewed by Carlos Garcia Campos. * gtk/install-dependencies: 2015-11-20 Csaba Osztrogonác [EFL] Enable FTL JIT by default on X86_64 https://bugs.webkit.org/show_bug.cgi?id=143822 Reviewed by Carlos Garcia Campos. * Scripts/build-jsc: * Scripts/run-javascriptcore-tests: * Scripts/webkitperl/FeatureList.pm: * efl/jhbuild.modules: * gtk/jhbuildrc: * jhbuild/jhbuildrc_common.py: (init): 2015-11-19 Michael Catanzaro Unreviewed, install llvm-static on Fedora to fix build now FTL is enabled * gtk/install-dependencies: 2015-11-19 Commit Queue Unreviewed, rolling out r192667 and r192668. https://bugs.webkit.org/show_bug.cgi?id=151476 broke api tests (Requested by alexchristensen on #webkit). Reverted changesets: "Remove the non-NetworkProcess configurations" https://bugs.webkit.org/show_bug.cgi?id=151418 http://trac.webkit.org/changeset/192667 "Fix GTK Build after r192667." http://trac.webkit.org/changeset/192668 2015-11-19 Alex Christensen Remove the non-NetworkProcess configurations https://bugs.webkit.org/show_bug.cgi?id=151418 Reviewed by Geoffrey Garen. * TestWebKitAPI/Tests/WebKit2/mac/GetPIDAfterAbortedProcessLaunch.cpp: (TestWebKitAPI::TEST): * WebKitTestRunner/TestController.cpp: (WTR::TestController::generatePageConfiguration): 2015-11-19 Alexey Proskuryakov LayoutTestRelay processes leak on iOS testers https://bugs.webkit.org/show_bug.cgi?id=151447 Reviewed by Tim Horton. * BuildSlaveSupport/kill-old-processes: Added LayoutTestRelay. 2015-11-19 Brent Fulgham [Win] webkitpy is applying abspath to DOS paths, yielding invalid paths https://bugs.webkit.org/show_bug.cgi?id=151156 Unreviewed test fix. I tracked down the cause of the test regression on the Windows builders. Once I fixed that, I realized that the new 'abs_results_directory' value was not needed, so this change reverts that code. * Scripts/webkitpy/port/base.py: (Port.to.start_websocket_server): We don't need a special directory variable on Cygwin anymore. (Port.abs_results_directory): Deleted. * Scripts/webkitpy/port/win.py: (WinPort.__init__): Deleted. (WinPort.abs_results_directory): Deleted. (WinPort.results_directory): Deleted. 2015-11-19 Alexey Proskuryakov build.webkit.org/dashboard: Move layoutTestResultsDirectoryURLForIteration implementation to base class https://bugs.webkit.org/show_bug.cgi?id=151425 Reviewed by Tim Horton. * BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/Buildbot.js: (Buildbot.prototype.layoutTestResultsDirectoryURLForIteration): * BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/WebKitBuildbot.js: (WebKitBuildbot.prototype.layoutTestResultsDirectoryURLForIteration): Deleted. 2015-11-19 Youenn Fablet WPT server should use its own testharness.js file and generate a warning when it does not match WebKit version https://bugs.webkit.org/show_bug.cgi?id=150332 Reviewed by Darin Adler and Ryosuke Niwa. Removing testharness.js overwriting from LayoutTests/resources to LayoutTests/imported/w3c/web-platform-tests/resources. Adding support for file comparison in python filesystem. #NousSommesUnis. * Scripts/webkitpy/common/system/filesystem.py: (FileSystem.compare): Adding compare as a wrapper around filecmp.cmp. * Scripts/webkitpy/common/system/filesystem_mock.py: (MockFileSystem.compare): Adding a mock compare. * Scripts/webkitpy/layout_tests/servers/web_platform_test_server.py: (WebPlatformTestServer.__init__): Removed testharness.js as file to copy from WK to WPT. (WebPlatformTestServer._copy_webkit_test_files): Added warning generation when WK testharness.js is not matching WPT version. * Scripts/webkitpy/layout_tests/servers/web_platform_test_server_unittest.py: (TestWebPlatformTestServer.test_previously_spawned_instance): Updated tests to add testharness.js to mock filesystem. (TestWebPlatformTestServer.test_corrupted_subserver_files): Ditto. 2015-11-18 Jiewen Tan [WK1] Crash loading Blink layout test fast/dom/Window/property-access-on-cached-window-after-frame-removed.html https://bugs.webkit.org/show_bug.cgi?id=150198 Reviewed by Brent Fulgham. * WebKitTestRunner/InjectedBundle/mac/TestRunnerMac.mm: (WTR::TestRunner::inspectorTestStubURL): Since WebInspectorUI.framework is not available for iOS, the framework and corresponding functions are disabled in iOS. 2015-11-18 Alexey Proskuryakov [Mac] Swipe tests depend on user preferences https://bugs.webkit.org/show_bug.cgi?id=151403 Reviewed by Tim Horton. * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm: (WTR::InjectedBundle::platformInitialize): * WebKitTestRunner/mac/main.mm: (setDefaultsToConsistentValuesForTesting): Override these preferences to match the default for a new OS X installation. * DumpRenderTree/mac/DumpRenderTree.mm: (setDefaultsToConsistentValuesForTesting): Swipe tests are not enabled in WK1, but let's override the default anyway. 2015-11-18 Carlos Garcia Campos [GTK] Update xserver version in jhbuild https://bugs.webkit.org/show_bug.cgi?id=151383 Reviewed by Alejandro G. Castro. In r192568 I assumed the bots would have a recent enough Xvfb to use the -displayfd option, but I forgot we build our own xserver just for Xvfb. The version we build is too old, so we could just use an updated version (the same bots have installed, for example) to use the -displayfd option. * gtk/jhbuild.modules: Update xserver to 1.16.4. * gtk/patches/xserver-fix-glx-init.patch: Deleted. * gtk/patches/xserver-remove-bogus-dependencies.patch: Updated to apply on 1.16.4. 2015-11-18 Carlos Garcia Campos Unreviewed. Fix webkitpy tests after r192568. Instead of ignoring exception when closing mock fds, use a helper function that the unit test can override to do nothing. * Scripts/webkitpy/port/xvfbdriver.py: (XvfbDriver._xvfb_close_pipe): (XvfbDriver._xvfb_run): * Scripts/webkitpy/port/xvfbdriver_unittest.py: (XvfbDriverTest.make_driver): 2015-11-18 Carlos Garcia Campos [XvfbDriver] Fail to run all layout tests when X server started with -displayfd option https://bugs.webkit.org/show_bug.cgi?id=151135 Reviewed by Darin Adler. The XvfbDriver uses the x server command line to check the displays that are currently in use. This doesn't work when X server was started with -displayfd option. This option is used to let the server find the display id available that is written to the given file descriptor. With this option xorg doesn't need to create the lock files in tmp either. The -displayfd option is also available in Xvfb, so we could use it too. That would simplify the code, fixing also race conditions between the check for available displays and Xvfb opening the connection, we wouldn't need to wait for 4 seconds after launching Xvfb, and all lock files we are using won't be needed either. * Scripts/webkitpy/port/xvfbdriver.py: (XvfbDriver._xvfb_pipe): Helper function to create the pipe, only needed to be overriden by unit tests. (XvfbDriver._xvfb_read_display_id): Helper function to read from the pipe, only needed to be overriden by unit tests. (XvfbDriver._xvfb_run): Run Xvfb with -displayfd option, using a pipe to read the display id. (XvfbDriver._start): Call _xvfb_run() and remove the code to run Xvfb for a given display. (XvfbDriver.stop): Remove code to release and delete file locks. * Scripts/webkitpy/port/xvfbdriver_unittest.py: (XvfbDriverTest.make_driver): (XvfbDriverTest.test_start): (XvfbDriverTest.test_start_arbitrary_worker_number): (XvfbDriverTest.test_stop): (XvfbDriverTest.assertDriverStartSuccessful): Deleted. (XvfbDriverTest): Deleted. (XvfbDriverTest.test_stop.FakeXvfbProcess): Deleted. 2015-11-17 Alexey Proskuryakov run-webkit-tests should not truncate persistent lines https://bugs.webkit.org/show_bug.cgi?id=151376 rdar://problem/23162775 Reviewed by Antti Koivisto. It is OK to truncate progress lines, but it's not OK to truncate error (or otherwise unexpected result) messages that remain on the screen to be read. In practice, this almost always truncates the test path, which is important to see. * Scripts/webkitpy/layout_tests/views/printing.py: (Printer._test_status_line): (Printer.print_finished_test): * Scripts/webkitpy/layout_tests/views/printing_unittest.py: (Testprinter.test_test_status_line): 2015-11-16 Carlos Garcia Campos Unreviewed. Fix GTK+ unit tests after r192469. In r192469 jhbuild version was updated and coinfig lcas API changed. * jhbuild/jhbuildutils.py: (enter_jhbuild_environment_if_available): Pass an empty list for conditions to the Config constructor. 2015-11-16 Carlos Garcia Campos Unreviewed. Use commit id instead of tag name for jhbuild revision. In r192469 I updated the jhbuild version to 3.12.0, and used the tag instead of the commit id to make it easier to read and remember. But we use the commit id to check if the current revision is up to date, and when compared to the tag id it always fails and jhbuild is always updated. * jhbuild/jhbuild-wrapper: Use commit id and add a comment to remember this is the id of 3.12.0. 2015-11-16 Carlos Garcia Campos Unreviewed. Fix GTK+ layout tests running after r192469. We fail to find the Gallium llvmpipe driver because of the build directory change in r192469. * gtk/jhbuildrc: Use buildroot instead of checkoutdir to build the LLVMPIPE_LIBGL_PATH. 2015-11-16 Carlos Garcia Campos [GTK] Use FTL by default when LLVM 3.7 is available https://bugs.webkit.org/show_bug.cgi?id=142128 Reviewed by Csaba Osztrogonác. * Scripts/run-javascriptcore-tests: Enable FTL for GTK+ port when architecture is x86_64. * Scripts/update-webkit-libs-jhbuild: (cleanJhbuild): Remove also the separate Build directory used by GTK+ port now. * Scripts/webkitdirs.pm: (isX86_64): * Scripts/webkitperl/FeatureList.pm: Enable FTL by default for GTK+ port when architecture is x86_64. * gtk/jhbuild-optional.modules: Remove llvm. * gtk/jhbuild.modules: Add llvm 3.7 and include it as dependecy only when architecture is x86_64. * gtk/jhbuildrc: Set a buildroot to use a separate build directory. LLVM 3.7 fails to build if configure script is run from source dir. Since it's not possible to use a separate build dir only for llvm, we do it now globally. Modules not supporting non source build dirs are marked in the moduleset. Also set x86_64 condition when the architecture is x86_64. * gtk/patches/llvm-elf-add-stackmaps-arm64.patch: Removed. * gtk/patches/llvm-elf-add-stackmaps.patch: Removed. * gtk/patches/llvm-elf-allow-fde-references-outside-the-2gb-range-arm64.patch: Removed. * gtk/patches/llvm-elf-allow-fde-references-outside-the-2gb-range.patch: Removed. * jhbuild/jhbuild-wrapper: Use a newer jhbuild version to be able to use conditions. 2015-11-15 Commit Queue Unreviewed, rolling out r192462. https://bugs.webkit.org/show_bug.cgi?id=151295 failing python tests (Requested by youenn on #webkit). Reverted changeset: "WPT server should use its own testharness.js file and generate a warning when it does not match WebKit version" https://bugs.webkit.org/show_bug.cgi?id=150332 http://trac.webkit.org/changeset/192462 2015-11-15 Youenn Fablet WPT server should use its own testharness.js file and generate a warning when it does not match WebKit version https://bugs.webkit.org/show_bug.cgi?id=150332 Reviewed by Darin Adler. * Scripts/webkitpy/layout_tests/servers/web_platform_test_server.py: (WebPlatformTestServer.__init__): Removed testharness.js as file to copy from WK to WPT. (WebPlatformTestServer._copy_webkit_test_files): Added warning generation when WK testharness.js is not matching WPT version. 2015-11-14 Nan Wang AX: add a new trait for elements in fieldset on iOS https://bugs.webkit.org/show_bug.cgi?id=151281 Reviewed by Chris Fleizach. * DumpRenderTree/AccessibilityUIElement.cpp: (assistiveTechnologySimulatedFocusCallback): (fieldsetAncestorElementCallback): (childAtIndexCallback): (getElementTextLengthCallback): (hasContainedByFieldsetTraitCallback): (AccessibilityUIElement::getJSClass): * DumpRenderTree/AccessibilityUIElement.h: * DumpRenderTree/ios/AccessibilityUIElementIOS.mm: (AccessibilityUIElement::elementTextLength): (AccessibilityUIElement::hasContainedByFieldsetTrait): (AccessibilityUIElement::fieldsetAncestorElement): (AccessibilityUIElement::url): * WebKitTestRunner/InjectedBundle/AccessibilityUIElement.cpp: (WTR::AccessibilityUIElement::scrollPageDown): (WTR::AccessibilityUIElement::scrollPageLeft): (WTR::AccessibilityUIElement::scrollPageRight): (WTR::AccessibilityUIElement::hasContainedByFieldsetTrait): (WTR::AccessibilityUIElement::fieldsetAncestorElement): * WebKitTestRunner/InjectedBundle/AccessibilityUIElement.h: * WebKitTestRunner/InjectedBundle/Bindings/AccessibilityUIElement.idl: * WebKitTestRunner/InjectedBundle/ios/AccessibilityUIElementIOS.mm: (WTR::AccessibilityUIElement::identifier): (WTR::AccessibilityUIElement::hasContainedByFieldsetTrait): (WTR::AccessibilityUIElement::fieldsetAncestorElement): (WTR::AccessibilityUIElement::rowCount): 2015-11-13 Keith Miller Unreviewed, change JavaScriptCore watchlist to exclude inspector things. * Scripts/webkitpy/common/config/watchlist: 2015-11-13 Tim Horton Support printing in WKWebView https://bugs.webkit.org/show_bug.cgi?id=151276 Reviewed by Beth Dakin. * MiniBrowser/mac/MainMenu.xib: * MiniBrowser/mac/WK1BrowserWindowController.m: (-[WK1BrowserWindowController printWebView:]): * MiniBrowser/mac/WK2BrowserWindowController.m: (-[WK2BrowserWindowController printWebView:]): Hook up printing to Minibrowser! 2015-11-13 Tim Horton Hardware keyboard spacebar scrolls too far on iOS https://bugs.webkit.org/show_bug.cgi?id=151227 Reviewed by Simon Fraser. * WebKitTestRunner/UIScriptContext/Bindings/UIScriptController.idl: * WebKitTestRunner/UIScriptContext/UIScriptContext.h: * WebKitTestRunner/UIScriptContext/UIScriptController.cpp: (WTR::UIScriptController::setDidEndScrollingCallback): (WTR::UIScriptController::didEndScrollingCallback): (WTR::UIScriptController::platformSetDidEndScrollingCallback): * WebKitTestRunner/UIScriptContext/UIScriptController.h: * WebKitTestRunner/cocoa/TestRunnerWKWebView.h: * WebKitTestRunner/cocoa/TestRunnerWKWebView.mm: (-[TestRunnerWKWebView dealloc]): (-[TestRunnerWKWebView _didFinishScrolling]): * WebKitTestRunner/ios/UIScriptControllerIOS.mm: (WTR::UIScriptController::platformSetDidEndScrollingCallback): (WTR::UIScriptController::platformClearAllCallbacks): Expose _didFinishScrolling on WKWebView to the UIScriptController. 2015-11-13 Brent Fulgham [Win] Unreviewed test fix. * Scripts/webkitpy/port/win.py: (WinPort.results_directory): Our 'results_directory' should be an absolute path (even when not using Cygwin), but should be in DOS format for non-Cygwin clients. 2015-11-12 Daniel Bates Add my apple.com email address to contributors.json * Scripts/webkitpy/common/config/contributors.json: 2015-11-12 Daniel Bates webkitpy: Remove extraneous word "the" from the description of command attach-to-bug * Scripts/webkitpy/tool/commands/upload.py: (AttachToBug): 2015-11-12 Brent Fulgham [Win] webkitpy is applying abspath to DOS paths, yielding invalid paths https://bugs.webkit.org/show_bug.cgi?id=151156 Reviewed by Anders Carlsson. My earlier patch was incomplete. There are some Python libraries that want to receive UNIX style paths, even though all of our applications and tools use DOS paths. To handle this special case, we need to track a UNIX-style absolute path that can be given to things like the Python socket server. When we cut over to all-native Windows Python we can get rid of this special case code. * Scripts/webkitpy/common/system/filesystem.py: (FileSystem.abspath): Add special code for Cygwin to convert a DOS-style path into something Cygwin can use internally. * Scripts/webkitpy/port/base.py: (Port.results_directory): Remove custom Windows code. Instead, override the method in the windows port object. (Port.abs_results_directory): Add stub. (Port.to.start_websocket_server): For Cygwin, use special cygwin absolute path. * Scripts/webkitpy/port/win.py: (WinPort.__init__): Added to initialze new _abs_results_directory value. (WinPort.abs_results_directory): Added, (WinPort.results_directory): Modified to also create a value for _abs_results_directory 2015-11-12 Keith Miller Unreviewed, add Michael Saboff to the JavaScriptCore watchlist, per request. * Scripts/webkitpy/common/config/watchlist: 2015-11-12 Keith Miller Unreviewed, add watchlist for Source/JavaScriptCore and add myself, Saam, and Mark. * Scripts/webkitpy/common/config/watchlist: 2015-11-12 Daniel Bates Support building configuration Production of DumpRenderTree and WebKitTestRunner for iOS https://bugs.webkit.org/show_bug.cgi?id=151191 Reviewed by Andy Estes. * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Remove hardcoded SKIP_INTALL. Will override this setting in DumpRenderTree.xcconfig. * DumpRenderTree/mac/Configurations/DumpRenderTree.xcconfig: When building for iOS we now compile source file DumpRenderTreeMain.mm to produce a built product and make the Apple Internal build system happy. Do not waste time coping OpenType font files (*.otf files) and skip installation when building for iOS since we are not interested in the build product and are only building this target to make the Apple Internal build system happy. * DumpRenderTree/mac/DumpRenderTreeMain.mm: (main): Added dummy implementation for iOS so that we build an actual command line tool to make the Apple Internal build system happy. * WebKitTestRunner/Configurations/Base.xcconfig: Add iOS-specific definition for WEBKIT_UMBRELLA_FRAMEWORKS_DIR. * WebKitTestRunner/Configurations/DebugRelease.xcconfig: Remove variable WEBKIT_UMBRELLA_FRAMEWORKS_DIR. * WebKitTestRunner/Configurations/WebKitTestRunner.xcconfig: Build file WebKitTestRunner/mac/main.mm on iOS Only to produce a built product and make the Apple Internal build system happy. No need to define OTHER_LDFLAGS when building this target for iOS as the default linker flags are sufficient. * WebKitTestRunner/mac/main.mm: (main): Added dummy implementation for iOS so that we build an actual command line tool to make the Apple Internal build system happy. 2015-11-12 Chris Dumez Unreviewed, add watchlist for history/ and add myself. * Scripts/webkitpy/common/config/watchlist: 2015-11-12 Chris Dumez Unreviewed, update my watchlists. * Scripts/webkitpy/common/config/watchlist: 2015-11-12 Daniel Bates Rename webkitdirs::willUseIOS{Device, Simulator}SDKWhenBuilding() to willUseIOS{Device, Simulator}SDK() https://bugs.webkit.org/show_bug.cgi?id=151207 Rubber-stamped by Alexey Proskuryakov. The functions webkitdirs::willUseIOS{Device, Simulator}SDKWhenBuilding() are not specific to building iOS software. We should rename these functions to webkitdirs::willUseIOS{Device, Simulator}SDK() to convey their general-purpose nature of determining whether we are using the iphoneos or iphonesimulator SDK. * Scripts/build-jsc: * Scripts/build-layouttestrelay: * Scripts/build-webkit: * Scripts/copy-webkitlibraries-to-product-directory: * Scripts/run-api-tests: (runTest): (prepareEnvironmentForRunningTestTool): * Scripts/run-javascriptcore-tests: * Scripts/webkitdirs.pm: (XcodeOptions): (willUseIOSDeviceSDK): Formerly named willUseIOSDeviceSDKWhenBuilding. (willUseIOSSimulatorSDK): Formerly name willUseIOSSimulatorSDKWhenBuilding. (isIOSWebKit): (runIOSWebKitApp): (willUseIOSDeviceSDKWhenBuilding): Deleted. (willUseIOSSimulatorSDKWhenBuilding): Deleted. 2015-11-12 Daniel Bates run-api-tests fails to run with public iOS SDK https://bugs.webkit.org/show_bug.cgi?id=151076 Reviewed by Alexey Proskuryakov. Use the command line tool simctl as the sim tool was removed from the public iOS 9 SDK. A side benefit of this change it is sufficient to run the tests without launching Simulator.app to boot the simulator device beforehand because simctl will boot the device for us. Additionally, use Perl pragma sigtrap to install signal handlers to catch SIG{HUP, INT, PIPE, TERM} signals and ultimately run our END block so that we shutdown the simulator device (if applicable). This makes Control-C terminate the app gracefully. * Scripts/run-api-tests: (runTest): (listAllTests): (prepareEnvironmentForRunningTestTool): When running for iOS Simulator use setupIOSWebKitEnvironment() to setup the environment variables for iOS instead of using setupMacWebKitEnvironment(). We also prefix the name of these environment variables with "SIMCTL_CHILD_" so that simctl sets these environment variables in the simulator environment. * Scripts/webkitdirs.pm: (shutDownIOSSimulatorDevice): Added. (restartIOSSimulatorDevice): Added. 2015-11-12 KwangHyuk Kim [EFL] fix EvasGL configuration error https://bugs.webkit.org/show_bug.cgi?id=151180 Reviewed by Csaba Osztrogonác. EvasGL backend and depth, stencil buffer size should be specified before elm_win_add to let EvasGL keep quiet. Additionally, duplicated elm_config_accel_preference_set() call is eleminated. * MiniBrowser/efl/main.c: (window_create): (elm_main): 2015-11-12 Csaba Osztrogonác [EFL] Fix EFL 1.16 with enabled NEON support https://bugs.webkit.org/show_bug.cgi?id=151188 Unreviewed buildfix, disable NEON temporarily after r192358. * efl/jhbuild.modules: 2015-11-12 Csaba Osztrogonác [EFL] Bump EFL version to 1.16.0 https://bugs.webkit.org/show_bug.cgi?id=150228 Reviewed by Gyuyoung Kim. * efl/jhbuild.modules: * efl/patches/efl-remove-XPrint.patch: Removed. 2015-11-12 Csaba Osztrogonác Remove ENABLE(SATURATED_LAYOUT_ARITHMETIC) guards https://bugs.webkit.org/show_bug.cgi?id=150972 Reviewed by Darin Adler. * TestWebKitAPI/Tests/WebCore/LayoutUnit.cpp: 2015-11-12 KwangHyuk Kim [EFL] fix EvasGL configuration error https://bugs.webkit.org/show_bug.cgi?id=151180 Reviewed by Gyuyoung Kim. EvasGL backend and depth, stencil buffer size should be specified before elm_win_add to let EvasGL keep quiet * MiniBrowser/efl/main.c: (window_create): 2015-11-11 Brent Fulgham [Win] webkitpy is applying abspath to DOS paths, yielding invalid paths https://bugs.webkit.org/show_bug.cgi?id=151156 Reviewed by Anders Carlsson. * Scripts/webkitpy/port/base.py: (Port.results_directory): Don't pass paths through 'abspath' (Port._build_path): Ditto. * Scripts/webkitpy/port/win.py: (WinPort._build_path): Ditto. 2015-11-11 Akos Kiss Unreviewed, add myself to the committers list. * Scripts/webkitpy/common/config/contributors.json: 2015-11-11 Anders Carlsson _WKRemoteObjectInterface should handle specifying allowed classes for reply block arguments https://bugs.webkit.org/show_bug.cgi?id=151144 Reviewed by Tim Horton. * TestWebKitAPI/Tests/WebKit2Cocoa/RemoteObjectRegistry.h: (remoteObjectInterface): * TestWebKitAPI/Tests/WebKit2Cocoa/RemoteObjectRegistry.mm: (TEST): * TestWebKitAPI/Tests/WebKit2Cocoa/RemoteObjectRegistryPlugIn.mm: (-[RemoteObjectRegistryPlugIn webProcessPlugIn:didCreateBrowserContextController:]): (-[RemoteObjectRegistryPlugIn selectionAndClickInformationForClickAtPoint:completionHandler:]): 2015-11-10 Wenson Hsieh UI-side scripts in WebKitTestRunner should wait until event handling completes before finishing https://bugs.webkit.org/show_bug.cgi?id=151101 Reviewed by Simon Fraser. WebKitTestRunner may still crash in the scenario where a marker event is dequeued and handled after uiScriptComplete has been called. This patch teaches the UI script execution context to defer script completion until all non-persistent tasks (currently tap, double tap and hardware keyboard) have been handled, so marker events will no longer bleed through tests. We accomplish this by changing the behavior of uiScriptComplete. When calling uiScriptComplete, we store that a request to complete the UI-side script for the current parent callback has been made. Subsequently, when a callback (either persistent or non-persistent) finishes invoking and a request has been made to complete the UI script (this request may have been made when running a previous callback) we check if all the non-persistent callbacks that have the same parent callback as the current one have finished. If so, we complete the callback immediately; otherwise, we wait until the in-flight non-persistent callbacks finish execution to complete the UI script. This patch also refactors some logic in UIScriptContext. It introduces a new convention for assigning IDs to callbacks: IDs 1000 and above are treated as non-persistent callbacks, whereas IDs between 0 and 999 are persistent task callbacks. This is similar to the existing convention for assigning IDs in the 100s range to parent callbacks, and allows us to easily differentiate between callbacks that are persistent and non-persistent, as well as determine when an existing persistent callback must be unregistered before a new callback function can be set. * WebKitTestRunner/UIScriptContext/UIScriptContext.cpp: (isPersistentCallbackID): (UIScriptContext::runUIScript): (UIScriptContext::nextTaskCallbackID): (UIScriptContext::prepareForAsyncTask): (UIScriptContext::asyncTaskComplete): (UIScriptContext::registerCallback): (UIScriptContext::fireCallback): (UIScriptContext::requestUIScriptCompletion): (UIScriptContext::tryToCompleteUIScriptForCurrentParentCallback): (UIScriptContext::currentParentCallbackHasOutstandingAsyncTasks): (UIScriptContext::uiScriptComplete): Deleted. * WebKitTestRunner/UIScriptContext/UIScriptContext.h: (WTR::UIScriptContext::currentParentCallbackIsPendingCompletion): * WebKitTestRunner/UIScriptContext/UIScriptController.cpp: (WTR::UIScriptController::setWillBeginZoomingCallback): (WTR::UIScriptController::willBeginZoomingCallback): (WTR::UIScriptController::setDidEndZoomingCallback): (WTR::UIScriptController::didEndZoomingCallback): (WTR::UIScriptController::setDidShowKeyboardCallback): (WTR::UIScriptController::didShowKeyboardCallback): (WTR::UIScriptController::setDidHideKeyboardCallback): (WTR::UIScriptController::didHideKeyboardCallback): (WTR::UIScriptController::uiScriptComplete): * WebKitTestRunner/UIScriptContext/UIScriptController.h: * WebKitTestRunner/ios/UIScriptControllerIOS.mm: (WTR::UIScriptController::doAsyncTask): (WTR::UIScriptController::zoomToScale): (WTR::UIScriptController::singleTapAtPoint): (WTR::UIScriptController::doubleTapAtPoint): (WTR::UIScriptController::typeCharacterUsingHardwareKeyboard): (WTR::UIScriptController::platformSetWillBeginZoomingCallback): (WTR::UIScriptController::platformSetDidEndZoomingCallback): (WTR::UIScriptController::platformSetDidShowKeyboardCallback): (WTR::UIScriptController::platformSetDidHideKeyboardCallback): (WTR::UIScriptController::platformClearAllCallbacks): Deleted. * WebKitTestRunner/mac/UIScriptControllerMac.mm: (WTR::UIScriptController::doAsyncTask): 2015-11-10 Daniel Bates Teach Makefile to build LayoutTestRelay when building for iOS Simulator https://bugs.webkit.org/show_bug.cgi?id=150849 Reviewed by Alexey Proskuryakov. Unless variable DO_NOT_BUILD_LAYOUT_TEST_RELAY is non-empty, build LayoutTestRelay when building WebKit tools. * Makefile: 2015-11-10 David Kilzer [check-webkit-style] False positive (readability/enum_casing) on CF-styled enum members Reviewed by Brent Fulgham. * Scripts/webkitpy/style/checkers/cpp.py: (_EnumState.process_clean_line): Allow 'k' to start enums. (check_enum_casing): Update error message to discuss 'k'. * Scripts/webkitpy/style/checkers/cpp_unittest.py: (NoNonVirtualDestructorsTest.test_enum_casing): Update error messages, and add new tests for enums that start with lowercase letters. 2015-11-10 Geoffrey Garen run-jsc-benchmarks doesn't work with internal benchmarks https://bugs.webkit.org/show_bug.cgi?id=151047 Reviewed by Filip Pizlo. * Scripts/run-jsc-benchmarks: Use Pathname so we get a "/" inserted when we concatenate. 2015-11-10 Carlos Garcia Campos [GTK] Use CROSS_PLATFORM_CONTEXT_MENUS https://bugs.webkit.org/show_bug.cgi?id=150642 Reviewed by Martin Robinson. Remove test for input methods submenu. * TestWebKitAPI/Tests/WebKit2Gtk/TestContextMenu.cpp: 2015-11-09 Joonghun Park Unreviewed, add myself to the committers list. * Scripts/webkitpy/common/config/contributors.json: 2015-11-09 Ryuan Choi [EFL] Crash while opening child webview with EWK_PROCESS_MODEL_MULTIPLE_SECONDARY https://bugs.webkit.org/show_bug.cgi?id=145924 Reviewed by Gyuyoung Kim. * MiniBrowser/efl/main.c: (on_window_create): (window_create): * TestWebKitAPI/Tests/WebKit2/CoordinatedGraphics/WKViewUserViewportToContents.cpp: (TestWebKitAPI::TEST): * TestWebKitAPI/efl/PlatformWebView.cpp: (TestWebKitAPI::PlatformWebView::PlatformWebView): * WebKitTestRunner/efl/PlatformWebViewEfl.cpp: (WTR::PlatformWebView::PlatformWebView): 2015-11-09 Anders Carlsson Add reply blocks to _WKRemoteObjectInterface similar to NSXPCConnection https://bugs.webkit.org/show_bug.cgi?id=151056 rdar://problem/23222609 Reviewed by Tim Horton. Update test. * TestWebKitAPI/Tests/WebKit2Cocoa/RemoteObjectRegistry.mm: (TEST): * TestWebKitAPI/Tests/WebKit2Cocoa/RemoteObjectRegistryPlugIn.mm: (-[RemoteObjectRegistryPlugIn sayHello:completionHandler:]): 2015-11-09 Tim Horton Add drawsBackground SPI to WKWebView, and get rid of drawsTransparentBackground from WebKit2 https://bugs.webkit.org/show_bug.cgi?id=151054 Reviewed by Simon Fraser. * MiniBrowser/mac/WK1BrowserWindowController.m: (-[WK1BrowserWindowController didChangeSettings]): * MiniBrowser/mac/WK2BrowserWindowController.m: (-[WK2BrowserWindowController didChangeSettings]): Use drawsBackground instead, and make sure to set the window background color, otherwise it might end up being gray anyway! WebKit1 still doesn't work unless you turn off toolbar blurring, but at least WebKit2 is working now! 2015-11-09 Isaac Devine Fresh checkout fails to build on windows, DumpRenderTree can't find cairo_win.h https://bugs.webkit.org/show_bug.cgi?id=151013 Use the variable defined in the CMake scripts to determine the cairo include location, rather than relying on a environment variable to be set correctly. Otherwise the DumpRenderTreeLib.vcxproj will contain "\include\cairo" rather than the fully qualified path to the cairo include location. Reviewed by Darin Adler. * DumpRenderTree/PlatformWin.cmake: 2015-11-09 Csaba Osztrogonác [EFL] Fix the gst-plugins-bad jhbuild module build on Ubuntu 15.10 https://bugs.webkit.org/show_bug.cgi?id=150928 Reviewed by Gyuyoung Kim. * efl/jhbuild.modules: 2015-11-07 Daniel Bates Attempt to fix the Windows EWS bot after r192087 (https://bugs.webkit.org/show_bug.cgi?id=150859) Following r192087 Port._build_path() stores the computed root directory path in option _cached_root instead of overwriting option root. We need to teach the Windows-specific logic to make use of option _cached_root instead of option root when computing its adjusted path and to cache its adjusted in _cached_root instead of root. * Scripts/webkitpy/port/win.py: (WinPort._build_path): 2015-11-06 Daniel Bates Do not build LayoutTestRelay when --root is specified to run-webkit-tests https://bugs.webkit.org/show_bug.cgi?id=150989 Reviewed by Alexey Proskuryakov. The script run-webkit-tests should only check if LayoutTestRelay exists when invoked with --root and exit with an error if it does not exist. That is, we should not build LayoutTestRelay when it does not exist and an explicit directory of built executables was specified via --root. This will make the criterion for building LayoutTestRelay match the criterion for building DumpRenderTree/WebKitTestRunner. * Scripts/webkitpy/port/ios.py: (IOSSimulatorPort._check_port_build): Moved logic from IOSSimulatorPort.{_check_build_relay, check_build} to here. (IOSSimulatorPort._check_build_relay): Deleted. (IOSSimulatorPort.check_build): Deleted. 2015-11-06 Tim Horton Add preliminary (SPI) support for NSTextFinder on WKWebView https://bugs.webkit.org/show_bug.cgi?id=150907 Reviewed by Darin Adler. * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: * TestWebKitAPI/Tests/WebKit2Cocoa/FindInPage.mm: Added. (-[FindInPageNavigationDelegate webView:didFinishNavigation:]): (TEST): Add an API test. * WebEditingTester/WK1WebDocumentController.m: * WebEditingTester/WK2WebDocumentController.m: (-[WK2WebDocumentController awakeFromNib]): (-[WK2WebDocumentController contentView]): Use incremental find, and show the overlay/indicator. 2015-11-06 Alexey Proskuryakov iOS test results are not visible on the flakiness dashboard https://bugs.webkit.org/show_bug.cgi?id=150884 Reviewed by Darin Adler. One more change was needed for the minimal fix. * TestResultServer/static-dashboards/flakiness_dashboard.js: 2015-11-06 Daniel Bates Teach Makefile to build LayoutTestRelay when building for iOS Simulator https://bugs.webkit.org/show_bug.cgi?id=150849 Reviewed by Alexey Proskuryakov. Override the user-specified arguments SDKROOT and ARCHS to use the default SDK and default ARCHS since LayoutTestRelay is a Mac command line tool. * LayoutTestRelay/Makefile: Define OVERRIDE_SDKROOT and OVERRIDE_ARCHS. 2015-11-05 Nikos Andronikos Add runtime and compile time flags for enabling Web Animations API and model. https://bugs.webkit.org/show_bug.cgi?id=150914 Reviewed by Benjamin Poulain. Add ENABLE_WEB_ANIMATIONS compile time flag, runtime flag webAnimationsEnabled and Expose WK2 preference for runtime flag. * Scripts/webkitperl/FeatureList.pm: 2015-11-05 Ryuan Choi [EFL] Separate beforeunload confirm callback from confirm callback https://bugs.webkit.org/show_bug.cgi?id=150964 Reviewed by Gyuyoung Kim. * MiniBrowser/efl/main.c: (on_javascript_before_unload_confirm): (window_create): 2015-11-05 Jason Marcell run-webkit-test should look in --root directory for LayoutTestRelay https://bugs.webkit.org/show_bug.cgi?id=150859 Reviewed by Daniel Bates. For iOS run-webkit-tests, use LayoutTestRelay specified by --root; otherwise find LayoutTestRelay in the Mac build directory when --root is unspecified. * Scripts/webkitpy/port/base.py: (Port._build_path): Use '_cached_root' instead of 'root' so that we don't overwrite the argument that was passed in via the --root argument. * Scripts/webkitpy/port/ios.py: (IOSSimulatorPort.relay_path): 2015-11-05 David Kilzer TestWebKitAPI crashed in TestWebKitAPI: TestWebKitAPI::SharedBufferTest_copyBufferCreatedWithContentsOfExistingFile_Test::TestBody Reviewed by Youenn Fablet. * TestWebKitAPI/Tests/WebCore/SharedBuffer.cpp: (TestWebKitAPI::TEST_F(SharedBufferTest, copyBufferCreatedWithContentsOfExistingFile)): - Switch from strnstr() to memcmp() since strings are not guaranteed to be NULL-terminated. - Add another expectation that the size is greater than zero since memcmp() returns 0 (matching) if the length argument is zero. (TestWebKitAPI::TEST_F(SharedBufferTest, appendBufferCreatedWithContentsOfExistingFile)): - Same change to keep tests consistent, although the second string is NULL-terminated in this case. 2015-11-05 Commit Queue Unreviewed, rolling out r192073. https://bugs.webkit.org/show_bug.cgi?id=150962 "Broke the internal iOS build; will investigate offline" (Requested by dydz on #webkit). Reverted changeset: "Teach Makefile to build LayoutTestRelay when building for iOS Simulator" https://bugs.webkit.org/show_bug.cgi?id=150849 http://trac.webkit.org/changeset/192073 2015-11-05 Daniel Bates Teach Makefile to build LayoutTestRelay when building for iOS Simulator https://bugs.webkit.org/show_bug.cgi?id=150849 Reviewed by Alexey Proskuryakov. * LayoutTestRelay/Makefile: Temporarily override SDKROOT when it is iphonesimulator so that we build with the default SDK. We assume that the default SDK is the OS X SDK. * Makefile: Append LayoutTestRelay to the list of modules to build when building with SDK iphonesimulator. 2015-11-05 Ryuan Choi [EFL] Add try_close API to handle beforeunload event https://bugs.webkit.org/show_bug.cgi?id=150705 Reviewed by Gyuyoung Kim. * MiniBrowser/efl/main.c: (on_window_deletion): Call ewk_view_try_close instead of removing object directly. (window_create): Fixed that passes wrong data. 2015-11-04 Wenson Hsieh Fix crashing and flaky UIScriptController-based tests in fast/events/ios https://bugs.webkit.org/show_bug.cgi?id=150901 Reviewed by Simon Fraser. Clear out zooming and keyboard-related callbacks after uiScriptComplete has been invoked. This prevents later tests from calling into code which previous tests attached to these callbacks and causing WebKitTestRunner to crash due to accessing a deallocated UIScriptController. * WebKitTestRunner/UIScriptContext/UIScriptController.cpp: (WTR::UIScriptController::platformClearAllCallbacks): (WTR::UIScriptController::uiScriptComplete): * WebKitTestRunner/UIScriptContext/UIScriptController.h: * WebKitTestRunner/ios/UIScriptControllerIOS.mm: (WTR::UIScriptController::platformClearAllCallbacks): 2015-11-04 Jer Noble [iOS]