2012-05-21 Gavin Barraclough Disable private names by default in WebCore https://bugs.webkit.org/show_bug.cgi?id=87088 Reviewed by Geoff Garen. r117859 introduced a preliminary implementation of ES6-like private name objects to JSC. These are probably not yet ready to be web-facing, so disabling by default in WebCore. Opting-in for JSC & DumpRenderTree so that we can still run the fast/js/names.html test. * DumpRenderTree/mac/DumpRenderTree.mm: (resetDefaultsToConsistentValues): - Enable private name support in DRT. * WebKitTestRunner/TestController.cpp: (WTR::TestController::resetStateToConsistentValues): - Enable private name support in WKTR. 2012-05-22 Martin Robinson Fix a small syntax error in the GTK+ API test runner. Reviewed by Gustavo Noronha Silva. * gtk/run-api-tests: (TestRunner._setup_testing_environment): When reporting the error use the correct Python format string syntax. 2012-05-22 Mikhail Pozdnyakov [GTK][DRT] Wrong cache model is taken in LayoutTestController's setCacheModel() https://bugs.webkit.org/show_bug.cgi?id=86962 Reviewed by Gustavo Noronha Silva. * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: (LayoutTestController::setCacheModel): Sets WEBKIT_CACHE_MODEL_WEB_BROWSER if argument is 2. 2012-05-22 Mikhail Pozdnyakov [EFL][DRT] Key modifier bit flags have enum type. https://bugs.webkit.org/show_bug.cgi?id=86838 Reviewed by Antonio Gomes. Refactoring. Key modifier flags are unsigned int. Key modifier bit flags used have enum type. It caused problems as a key can have several modifiers at the same time and the union of modifiers (i.e. EvasKeyModifierControl | EvasKeyModifierShift) would have been out of EvasKeyModifier enum. * DumpRenderTree/efl/EventSender.cpp: (KeyEventInfo::KeyEventInfo): (KeyEventInfo): (MouseEventInfo::MouseEventInfo): (MouseEventInfo): (setEvasModifiers): (modifiersFromJSValue): (mouseDownCallback): (mouseUpCallback): (keyPadNameFromJSValue): (keyNameFromJSValue): (createKeyEventInfo): (sendKeyDown): 2012-05-22 Takashi Sakamoto Implement DOM_KEY_LOCATION_LEFT and RIGHT of KeyboardEvent's location property https://bugs.webkit.org/show_bug.cgi?id=86694 Modify keyDown to create left- and right- control, shift, and alt keyevent for a test to check that keyLocation property can have DOM_KEY_LOCATION_LEFT for left-control, shift, and alt, and DOM_KEY_LOCATION_RIGHT for right-control, shift, and alt. Reviewed by Alexey Proskuryakov. * DumpRenderTree/chromium/EventSender.cpp: (EventSender::keyDown): * DumpRenderTree/mac/EventSendingController.mm: (-[EventSendingController keyDown:withModifiers:withLocation:]): 2012-05-22 Christophe Dumez [EFL] EFL's DRT needs to print information about received Web Intents https://bugs.webkit.org/show_bug.cgi?id=86864 Reviewed by Adam Barth. Catch "intent,new" signal in EFL's DumpRenderTree and dump information about the intent. This output is expected by several webintents test cases. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::createView): (DumpRenderTreeChrome::onFrameCreated): (DumpRenderTreeChrome::onFrameIntentNew): * DumpRenderTree/efl/DumpRenderTreeChrome.h: (DumpRenderTreeChrome): 2012-05-22 Sudarsana Nagineni [EFL] Add support to enable/disable DOM paste https://bugs.webkit.org/show_bug.cgi?id=85040 Reviewed by Gustavo Noronha Silva. Enable JavascriptCanAccessClipboard setting on each test startup so that test cases which rely on clipboard can be unskipped. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): 2012-05-22 Hugo Parente Lima [Qt][WK2] EventSenderProxy::keyDown doesn't use the helper functions to send events https://bugs.webkit.org/show_bug.cgi?id=87044 Reviewed by Alexis Menard. Use sendOrQueueEvent instead of m_testController->mainWebView()->sendEvent * WebKitTestRunner/qt/EventSenderProxyQt.cpp: (WTR::EventSenderProxy::keyDown): 2012-05-22 Zan Dobersek [Gtk][LayoutTests] Repaint the complete WebKitWebView before dumping pixel results https://bugs.webkit.org/show_bug.cgi?id=86284 Reviewed by Martin Robinson. Paint the current dirty regions of the web view once more before dumping pixel results. * DumpRenderTree/gtk/DumpRenderTree.cpp: (dump): 2012-05-22 Allan Sandfeld Jensen GCC 4.7 and C++11 support. https://bugs.webkit.org/show_bug.cgi?id=86465 Reviewed by Darin Adler. Update detection of C++11 mode in GCC. * qmake/mkspecs/features/unix/default_post.prf: 2012-05-22 Hao Zheng [chromium] WebViewHost should use ENABLE guards for some features. https://bugs.webkit.org/show_bug.cgi?id=87087 Reviewed by Kent Tamura. These features are disabled on OS(ANDROID) and so shouldn't be compiled when the feature is off. * DumpRenderTree/chromium/WebViewHost.cpp: (WebViewHost::reset): * DumpRenderTree/chromium/WebViewHost.h: (WebViewHost): 2012-05-21 Christophe Dumez [EFL] EFL's DumpRenderTree does not print didReceiveTitle messages https://bugs.webkit.org/show_bug.cgi?id=85971 Reviewed by Antonio Gomes. Listen for the "title,changed" signal on the frames instead of the view so that we get notified for other frame than the main one. Print out the didReceiveTitle messages which are expected if LayoutTestController's dumpFrameLoadCallbacks() returns true. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::createView): (DumpRenderTreeChrome::onFrameTitleChanged): (DumpRenderTreeChrome::onFrameCreated): * DumpRenderTree/efl/DumpRenderTreeChrome.h: (DumpRenderTreeChrome): 2012-05-21 Raphael Kubo da Costa [EFL] Unreviewed, reverting r116461. r116461 caused a regression in the pixel tests, as scrollbars stopped being rendered in the PNGs. * DumpRenderTree/efl/PixelDumpSupportEfl.cpp: (createBitmapContextFromWebView): 2012-05-21 Brady Eidson and https://bugs.webkit.org/show_bug.cgi?id=87061 WKPageGetScaleFactor can return 0.0 after a session is restored Reviewed by Beth Dakin. Loads a simple page, creates a session from it, restores that session, then queries the page scale factor for the appropriate value of 1.0. * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: * TestWebKitAPI/Tests/WebKit2/WKPageGetScaleFactorNotZero.cpp: Added. (TestWebKitAPI): (TestWebKitAPI::didFinishLoadForFrame): (TestWebKitAPI::setPageLoaderClient): (TestWebKitAPI::createSessionState): (TestWebKitAPI::TEST): 2012-05-21 Hugo Parente Lima [WK2] Remove a useless #if on Tools/WebKitTestRunner/TestController.cpp https://bugs.webkit.org/show_bug.cgi?id=87047 Reviewed by Alexis Menard. * WebKitTestRunner/TestController.cpp: (WTR::TestController::didReceiveSynchronousMessageFromInjectedBundle): 2012-05-21 Dirk Pranke webkitpy: clean up the TestExpectations constructor https://bugs.webkit.org/show_bug.cgi?id=86926 Reviewed by Ojan Vafai. The TestExpectations constructor was attempting to pretend it didn't need to get stuff from the Port, and as a result we had a complicated constructor with a bunch of arguments, and calling it was too complicated (although it made testing a little easier and simplified the style checker). This patch has the constructor pull all the data it needs from the port directly, and allows us to delete a bunch of code. * Scripts/webkitpy/layout_tests/controllers/manager.py: (Manager.parse_expectations): * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py: (ManagerTest.test_update_summary_with_result): (ResultSummaryTest.get_result_summary): * Scripts/webkitpy/layout_tests/models/test_expectations.py: (TestExpectations.__init__): * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: (parse_exp): (SkippedTests.check): (RemoveConfigurationsTest.test_remove): (test_remove_line): * Scripts/webkitpy/layout_tests/run_webkit_tests.py: (lint): (run): * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py: (LintTest.test_all_configurations.FakePort.__init__): (LintTest.test_all_configurations.FakePort.test_expectations): (LintTest.test_all_configurations.FakePort.skipped_layout_tests): (LintTest.test_all_configurations.FakePort): (LintTest.test_all_configurations.FakePort.all_test_configurations): (LintTest.test_all_configurations.FakePort.configuration_specifier_macros): (LintTest.test_all_configurations.FakePort.path_from_webkit_base): (LintTest.test_all_configurations.FakePort.get_option): (LintTest.test_all_configurations.FakeFactory.__init__): (LintTest.test_all_configurations.FakeFactory.all_port_names): (LintTest.test_all_configurations): * Scripts/webkitpy/layout_tests/views/printing_unittest.py: (Testprinter.get_result_summary): * Scripts/webkitpy/style/checkers/test_expectations.py: (TestExpectationsChecker.check_test_expectations): * Scripts/webkitpy/tool/commands/queries.py: (PrintExpectations._model): * Scripts/webkitpy/tool/commands/rebaseline.py: (RebaselineTest._update_expectations_file): (RebaselineExpectations._expectations): 2012-05-21 Sudarsana Nagineni [GTK] DRT support for layoutTestController.setSerializeHTTPLoads https://bugs.webkit.org/show_bug.cgi?id=86653 Reviewed by Gustavo Noronha Silva. Add missing implementation setSerializeHTTPLoads to allow testing resource load order. * DumpRenderTree/gtk/DumpRenderTree.cpp: (resetDefaultsToConsistentValues): * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: (LayoutTestController::setSerializeHTTPLoads): 2012-05-21 Elliot Poger add skia_test_expectations override file to chromium NRWT https://bugs.webkit.org/show_bug.cgi?id=86749 Reviewed by Dirk Pranke * Scripts/webkitpy/layout_tests/port/base.py: (Port.test_expectations_overrides): * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumPort.test_expectations_overrides): * Scripts/webkitpy/layout_tests/port/chromium_unittest.py: (ChromiumPortTest.test_overrides_and_builder_names): 2012-05-21 Thiago Marcos P. Santos webkitdirs.pm: Fix undefined CXXFLAGS variable https://bugs.webkit.org/show_bug.cgi?id=87000 Reviewed by Martin Robinson. Fixes a warning on Autotools and CMake buildsystem when CXXFLAGS is not defined. * Scripts/webkitdirs.pm: (runAutogenForAutotoolsProjectIfNecessary): (generateBuildSystemFromCMakeProject): 2012-05-21 Caio Marcelo de Oliveira Filho Move setEditingBehavior() from layoutTestController to window.internals https://bugs.webkit.org/show_bug.cgi?id=42689 Reviewed by Hajime Morita and Kenneth Rohde Christiansen. Remove code that implements setEditingBehavior() for each platform. Note that During the InternalsObject reset the behavior is changed back to the platform default. Qt and Chromium ports set a default editing behavior for running tests different from the platform default, so they still keep code for resetting to the desired editing behavior. Other ports use the their defaults, that are restored by internals object. * DumpRenderTree/LayoutTestController.cpp: (LayoutTestController::staticFunctions): * DumpRenderTree/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/blackberry/DumpRenderTree.cpp: (BlackBerry::WebKit::DumpRenderTree::resetToConsistentStateBeforeTesting): * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp: * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::LayoutTestController): * DumpRenderTree/chromium/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: * DumpRenderTree/mac/DumpRenderTree.mm: (resetDefaultsToConsistentValues): * DumpRenderTree/mac/LayoutTestControllerMac.mm: * DumpRenderTree/qt/DumpRenderTreeQt.cpp: (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting): * DumpRenderTree/qt/LayoutTestControllerQt.cpp: * DumpRenderTree/qt/LayoutTestControllerQt.h: (LayoutTestController): * DumpRenderTree/win/DumpRenderTree.cpp: (resetDefaultsToConsistentValues): * DumpRenderTree/win/LayoutTestControllerWin.cpp: * DumpRenderTree/wx/LayoutTestControllerWx.cpp: 2012-05-21 Mikhail Pozdnyakov [EFL] PlatformKeyboardEventEfl is missing ESC key handling. https://bugs.webkit.org/show_bug.cgi?id=86837 Reviewed by Kenneth Rohde Christiansen. * DumpRenderTree/efl/EventSender.cpp: (keyNameFromJSValue): Added ESC key handling. 2012-05-20 Carlos Garcia Campos [GTK] Add GCancellable parameter to all methods using gio async pattern https://bugs.webkit.org/show_bug.cgi?id=86843 Reviewed by Gustavo Noronha Silva. * MiniBrowser/gtk/BrowserWindow.c: (browser_window_load_uri): 2012-05-20 Daniel Bates svn-apply cannot apply patches to files that contain space characters in their path https://bugs.webkit.org/show_bug.cgi?id=85742 Reviewed by Eric Seidel. Fixes an issue where svn-apply cannot apply a patch to a file if there is a space in its file path. The regular expression we were using to fix up +++/--- lines was too strict; it only matched file paths that contained non- whitespace characters. Instead, it's sufficient to match file paths whose characters aren't in the set {\t, \n, \r}. * Scripts/VCSUtils.pm: (parseSvnDiffHeader): (runCommand): Added. * Scripts/svn-apply: (scmWillDeleteFile): Modified to call runCommand() so as to handle querying git for a file whose path may contain a space character. * Scripts/webkitperl/VCSUtils_unittest/parseSvnDiffHeader.pl: - Added test case "new file with spaces in its name". * Scripts/webkitperl/VCSUtils_unittest/runCommand.pl: Added. 2012-05-20 Daniel Bates svn-apply fails to apply a patch that moves files from directory A to A/B https://bugs.webkit.org/show_bug.cgi?id=86973 Reviewed by Eric Seidel. Fixes an issue where svn-apply fails to apply a patch that moves files in a directory A to some sub-directory B in A with a Git checkout of WebKit. Currently, svn-apply only creates new directories along a file system path that it hasn't already processed. That is, if svn-apply creates/traverses all the intermediate directories along the path A/B then it will assume the directory A/B exists for all subsequent requests to create sub-directories in A/B (e.g. A/B/C). When moving a file F in directory A to directory A/B using a Git checkout, Git may remove directory A if F is the last file in A. Therefore, svn-apply will fail to create sub-directory B in A (since A no longer exists). * Scripts/svn-apply: (addDirectoriesIfNeeded): 2012-05-20 Gyuyoung Kim [Chromium] Print layout test result on buildbot https://bugs.webkit.org/show_bug.cgi?id=86965 Reviewed by Eric Seidel. * Scripts/run-webkit-tests: Add chrome-bot to buildbot user list. (runningOnBuildBot): 2012-05-19 Kevin Ollivier [wx] Unreviewed build fix. Don't stop the build if a build system update check fails. * waf/build/build_utils.py: (download_if_newer): 2012-05-19 Ryosuke Niwa Build fix. Sometimes --suffixes isn't supplied. * Scripts/webkitpy/tool/servers/gardeningserver.py: (GardeningHTTPRequestHandler.rebaseline): (GardeningHTTPRequestHandler.optimizebaselines): 2012-05-19 Ryosuke Niwa Add "Lion" -> "mac-lion" to the builder port map. * Scripts/webkitpy/layout_tests/port/builders.py: 2012-05-19 Sheriff Bot Unreviewed, rolling out r117675. http://trac.webkit.org/changeset/117675 https://bugs.webkit.org/show_bug.cgi?id=86938 Broke builds. You can't initialize members by members themselves (Requested by rniwa on #webkit). * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::numberOfPages): * DumpRenderTree/chromium/WebViewHost.cpp: (WebViewHost::printPage): 2012-05-18 Kausalya Madhusudhanan [chromium] User overloaded printBegin() webkit API to support auto fit to page functionality. https://bugs.webkit.org/show_bug.cgi?id=86684 Reviewed by Darin Fisher. * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::numberOfPages): * DumpRenderTree/chromium/WebViewHost.cpp: (WebViewHost::printPage): 2012-05-18 Raphael Kubo da Costa [EFL][DRT] Make it possible to enable CSS Grid Layout. https://bugs.webkit.org/show_bug.cgi?id=86921 Reviewed by Tony Chang. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::overridePreference): 2012-05-18 Dirk Pranke scm.add() doesn't work properly with svn 1.7 https://bugs.webkit.org/show_bug.cgi?id=86779 Reviewed by Eric Seidel. Re-land the change in r117526 with a fix to maintain compatibility with SVN 1.4.4 (yay Leopard!); turns out that "svn info foo" works correctly in 1.4.4 but "svn info" from the directory "foo" doesn't if foo is not part of a checkout. in_working_directory() and value_from_svn_info() were using the arguments inconsistently, leading to weird errors. * Scripts/webkitpy/common/checkout/scm/scm.py: (SCM): (SCM.in_working_directory): * Scripts/webkitpy/common/checkout/scm/svn.py: (SVN): (SVN.in_working_directory): 2012-05-18 Levi Weintraub Unreviewed. Moving myself from committer to reviewer. * Scripts/webkitpy/common/config/committers.py: 2012-05-18 Thiago Marcos P. Santos [EFL] Buildbot output does not print result stats anymore https://bugs.webkit.org/show_bug.cgi?id=85366 Reviewed by Ryosuke Niwa. Make run-webkit-tests pass --verbose to new-run-webkit-tests on EFL Debug bot. * Scripts/run-webkit-tests: (runningOnBuildBot): 2012-05-18 Tony Chang remove the CSS_GRID_LAYOUT compiler define, but default grid layout to off https://bugs.webkit.org/show_bug.cgi?id=86767 Reviewed by Ojan Vafai. * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::overridePreference): * DumpRenderTree/chromium/WebPreferences.cpp: (WebPreferences::reset): (WebPreferences::applyTo): * DumpRenderTree/chromium/WebPreferences.h: (WebPreferences): * Scripts/webkitperl/FeatureList.pm: 2012-05-18 Hugo Parente Lima [Qt][WK2] fast/events/page-visibility-iframe-move-test.html fails https://bugs.webkit.org/show_bug.cgi?id=86731 Reviewed by Simon Hausmann. Set the page visibility on a single page instead of on all pages from the page group. * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: (WTR::LayoutTestController::setPageVisibility): (WTR::LayoutTestController::resetPageVisibility): 2012-05-18 Yi Shen REGRESSION (r102553): Smart links do not work https://bugs.webkit.org/show_bug.cgi?id=85463 Reviewed by Ryosuke Niwa. Add LayoutTestController::setAutomaticLinkDetectionEnabled(). * DumpRenderTree/LayoutTestController.cpp: (setAutomaticLinkDetectionEnabledCallback): (LayoutTestController::staticFunctions): * DumpRenderTree/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp: (LayoutTestController::setAutomaticLinkDetectionEnabled): Not implemented. * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::setAutomaticLinkDetectionEnabled): Not implemented. * DumpRenderTree/chromium/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::setAutomaticLinkDetectionEnabled): Not implemented. * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: (LayoutTestController::setAutomaticLinkDetectionEnabled): Not implemented. * DumpRenderTree/mac/LayoutTestControllerMac.mm: (LayoutTestController::setAutomaticLinkDetectionEnabled): Set automatic link detection enabled. * DumpRenderTree/qt/LayoutTestControllerQt.cpp: (LayoutTestController::setAutomaticLinkDetectionEnabled): Not implemented. * DumpRenderTree/qt/LayoutTestControllerQt.h: (LayoutTestController): * DumpRenderTree/win/LayoutTestControllerWin.cpp: (LayoutTestController::setAutomaticLinkDetectionEnabled): Not implemented. * DumpRenderTree/wx/LayoutTestControllerWx.cpp: (LayoutTestController::setAutomaticLinkDetectionEnabled): Not implemented. 2012-05-18 Thiago Marcos P. Santos [EFL] Add shortcut to dump Security Origin and Web Database information https://bugs.webkit.org/show_bug.cgi?id=86521 Reviewed by Gustavo Noronha Silva. Added key combination (CTRL+s) to dump information about the security origin and databases of the main frame. Useful for API testing and debugging. * EWebLauncher/main.c: (on_key_down): 2012-05-18 Sudarsana Nagineni [EFL] DRT needs an implementation of layoutTestController.setSerializeHTTPLoads https://bugs.webkit.org/show_bug.cgi?id=86101 Reviewed by Ryosuke Niwa. Add missing implementation setSerializeHTTPLoads to allow testing resource load order. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::setSerializeHTTPLoads): 2012-05-18 Jochen Eisinger [chromium] plumb the frame for which a drag was initiated to the WebViewClient https://bugs.webkit.org/show_bug.cgi?id=86670 Reviewed by Kent Tamura. * DumpRenderTree/chromium/WebViewHost.cpp: (WebViewHost::startDragging): * DumpRenderTree/chromium/WebViewHost.h: (WebViewHost): 2012-05-18 Christophe Dumez [EFL] appcache tests are flaky https://bugs.webkit.org/show_bug.cgi?id=86619 Reviewed by Ryosuke Niwa. Correctly set the application cache database path in EFL's DumpRenderTree initialization code. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::initialize): 2012-05-17 Raphael Kubo da Costa [EFL][DRT] Call resetInternalsObject() after each test. https://bugs.webkit.org/show_bug.cgi?id=86793 Reviewed by Antonio Gomes. We need to reset the internals object's settings after running each test so that there are no interferences between each test's results. * DumpRenderTree/efl/DumpRenderTree.cpp: (runTest): 2012-05-17 Sheriff Bot Unreviewed, rolling out r117526. http://trac.webkit.org/changeset/117526 https://bugs.webkit.org/show_bug.cgi?id=86801 broke Chromium Mac 10.5 layout test canaries (Requested by pkasting on #webkit). * Scripts/webkitpy/common/checkout/scm/scm.py: (SCM): (SCM.in_working_directory): * Scripts/webkitpy/common/checkout/scm/svn.py: (SVN): (SVN.in_working_directory): 2012-05-17 Dirk Pranke scm.add() doesn't work properly with svn 1.7 https://bugs.webkit.org/show_bug.cgi?id=86779 Reviewed by Ojan Vafai. The code relied on '.svn' directories being present in order to tell if a directory had already been added to the repository; this is no longer true in SVN 1.7. * Scripts/webkitpy/common/checkout/scm/scm.py: (SCM): (SCM.in_working_directory): * Scripts/webkitpy/common/checkout/scm/svn.py: (SVN): (SVN.in_working_directory): 2012-05-17 Jon Lee Update Apple buildbots to prioritize latest changelists https://bugs.webkit.org/show_bug.cgi?id=86766 Reviewed by Brady Eidson. * BuildSlaveSupport/build.webkit.org-config/master.cfg: (pickLatestBuild): Function to get the most pick the most recent build request. (loadBuilderConfig): Configure builders in the AppleMac and AppleWin categories to use the pickLatestBuild function when choosing which request to build next. 2012-05-17 Caio Marcelo de Oliveira Filho [Qt] Fix MiniBrowser by using correct syntax for qrc URLs https://bugs.webkit.org/show_bug.cgi?id=86761 Reviewed by Alexis Menard. * MiniBrowser/qt/BrowserWindow.cpp: (BrowserWindow::BrowserWindow): (BrowserWindow::updateVisualMockTouchPoints): 2012-05-17 Thiago Marcos P. Santos [EFL] Add API for Web Database handling https://bugs.webkit.org/show_bug.cgi?id=85178 Reviewed by Antonio Gomes. Use the newly introduced Web Database API in EFL's DRT. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::clearAllDatabases): 2012-05-17 Dan Bernstein REGRESSION (r117428): WebKit API/SPI was removed https://bugs.webkit.org/show_bug.cgi?id=86748 Reverted r117428. * DumpRenderTree/LayoutTestController.cpp: (setDomainRelaxationForbiddenForURLSchemeCallback): (LayoutTestController::staticFunctions): * DumpRenderTree/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp: (LayoutTestController::setDomainRelaxationForbiddenForURLScheme): * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::LayoutTestController): (LayoutTestController::setDomainRelaxationForbiddenForURLScheme): * DumpRenderTree/chromium/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::setDomainRelaxationForbiddenForURLScheme): * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: (LayoutTestController::setDomainRelaxationForbiddenForURLScheme): * DumpRenderTree/mac/LayoutTestControllerMac.mm: (LayoutTestController::setDomainRelaxationForbiddenForURLScheme): * DumpRenderTree/qt/LayoutTestControllerQt.cpp: (LayoutTestController::setDomainRelaxationForbiddenForURLScheme): * DumpRenderTree/qt/LayoutTestControllerQt.h: (LayoutTestController): * DumpRenderTree/win/LayoutTestControllerWin.cpp: (LayoutTestController::setDomainRelaxationForbiddenForURLScheme): * DumpRenderTree/wx/LayoutTestControllerWx.cpp: (LayoutTestController::setDomainRelaxationForbiddenForURLScheme): 2012-05-17 Li Yin [chromium] run-webkit-tests can't work on ubuntu 12.04 https://bugs.webkit.org/show_bug.cgi?id=86565 Reviewed by Kent Tamura. On ubuntu 12.04, the package ttf-thai-tlwg install the Garuda.ttf into /usr/share/fonts/truetype/tlwg/Garuda.ttf, not /usr/share/fonts/truetype/thai/Garuda.ttf * DumpRenderTree/chromium/TestShellLinux.cpp: (checkAndLoadFontFile): (setupFontconfig): 2012-05-17 Elliot Poger Add setting of additional_expectations option to chromium_unittest.test_overrides_and_builder_names(), so we can test order in which the base and chromium expectations files are combined. https://bugs.webkit.org/show_bug.cgi?id=86642 Reviewed by Dirk Pranke. * Scripts/webkitpy/layout_tests/port/chromium_unittest.py: (ChromiumPortTest.test_overrides_and_builder_names): 2012-05-17 Balazs Kelemen [Qt] Web process hangs while waiting for plugin process connection during layouttests https://bugs.webkit.org/show_bug.cgi?id=86620 Reviewed by Csaba Osztrogonác. Reenable TestNetscapePlugin with Qt 4. * Tools.pro: 2012-05-17 Hugo Parente Lima [Qt][WK2] fast/events/page-visibility-* tests fail https://bugs.webkit.org/show_bug.cgi?id=83263 Reviewed by Kenneth Rohde Christiansen. Handle "preview" state and avoid string conversion. * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: (WTR::LayoutTestController::setPageVisibility): 2012-05-17 Gyuyoung Kim Convert setDomainRelaxationForbiddenForURLScheme to use InternalSettings interface https://bugs.webkit.org/show_bug.cgi?id=86704 Reviewed by Eric Seidel. Remove setDomainRelaxationForbiddenForURLScheme functions, because it is able to work in the cross-port way through the InternalSettings interface. * DumpRenderTree/LayoutTestController.cpp: (LayoutTestController::staticFunctions): * DumpRenderTree/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp: * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::LayoutTestController): * DumpRenderTree/chromium/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: * DumpRenderTree/mac/LayoutTestControllerMac.mm: * DumpRenderTree/qt/LayoutTestControllerQt.cpp: * DumpRenderTree/qt/LayoutTestControllerQt.h: (LayoutTestController): * DumpRenderTree/win/LayoutTestControllerWin.cpp: * DumpRenderTree/wx/LayoutTestControllerWx.cpp: 2012-05-17 Mariusz Grzegorczyk [EFL] [DRT] EFL's DumpRenderTree should support LayoutTestController's dumpEditingCallbacks() https://bugs.webkit.org/show_bug.cgi?id=84835 Reviewed by Eric Seidel. Added editing callbacks implementation. * DumpRenderTree/efl/CMakeLists.txt: Add EditingCallbacks.cpp to compilation. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: Invoke connectEditingCallbacks. (DumpRenderTreeChrome::createView): * DumpRenderTree/efl/EditingCallbacks.cpp: Added. Dump editing callbacks. (dumpPath): (dumpRange): (insertActionString): (selectionAffinityString): (shouldBeginEditing): (shouldEndEditing): (shouldInsertNode): (shouldInsertText): (shouldDeleteRange): (shouldChangeSelectedRange): (shouldApplyStyle): (editingBegan): (userChangedContents): (editingEnded): (selectionChanged): (connectEditingCallbacks): * DumpRenderTree/efl/EditingCallbacks.h: Added. 2012-05-17 Gyuyoung Kim Unreviewed. Roll out r117382 because of layout test crashes on EFL debug buildbot. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::initialize): 2012-05-17 Kentaro Hara [Performance test] In perftest.py, do not include 'description' in 'results' https://bugs.webkit.org/show_bug.cgi?id=86696 Reviewed by Csaba Osztrogoná. 'results' in perftest.py should not include any data that is not expected by the perf bot server. We do not need to send 'description' to the perf bot server since 'description' is always the same. Thus this patch just removes 'description' from 'results'. * Scripts/webkitpy/performance_tests/perftest.py: (PerfTest.parse_output): (PerfTest.output_statistics): (ChromiumStylePerfTest.parse_output): (PageLoadingPerfTest.run): 2012-05-17 Kentaro Hara [Performance test] Always fill results['description'] in perftest.py https://bugs.webkit.org/show_bug.cgi?id=86696 Reviewed by Ryosuke Niwa. This is a follow-up patch for r117397 to fix perf bot failures. This patch fills results['description'] always. * Scripts/webkitpy/performance_tests/perftest.py: (PageLoadingPerfTest.run): 2012-05-16 Kentaro Hara [Performance test] Support "description" for PerfTestRunner.run and PerfTestRunner.runPerSecond https://bugs.webkit.org/show_bug.cgi?id=86696 Reviewed by Ryosuke Niwa. We want to add some description for each PerfTestRunner.run and PerfTestRunner.runPerSecond, so that people can know what each run is testing. Output example: $ ./Tools/Scripts/run-perf-tests Bindings/first-child.html Running Bindings/first-child.html (1 of 1) DESCRIPTION: Benchmark for DOM attributes that return a Node object. RESULT Bindings: first-child= 788.359076534 runs/s median= 797.508097751 runs/s, stdev= 19.0972905207 runs/s, min= 746.666666667 runs/s, max= 801.001251564 runs/s * Scripts/webkitpy/performance_tests/perftest.py: (PerfTest): (PerfTest.parse_output): (PerfTest.output_statistics): 2012-05-16 Christophe Dumez [EFL] appcache tests are flaky https://bugs.webkit.org/show_bug.cgi?id=86619 Reviewed by Gustavo Noronha Silva. Correctly set the application cache database path in EFL's DumpRenderTree initialization code. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::initialize): 2012-05-16 Gyuyoung Kim Add new build slave for EFL port. https://bugs.webkit.org/show_bug.cgi?id=86567 Reviewed by Ryosuke Niwa. Add a build slave for EFL 32-bit linux build. * BuildSlaveSupport/build.webkit.org-config/config.json: 2012-05-16 Kihong Kwon [EFL] Support for Battery Status API on the WebKit-Efl https://bugs.webkit.org/show_bug.cgi?id=83254 Reviewed by Chang Shu. Add e_dbus libraries to jhbuild's modules list. * efl/jhbuild.modules: 2012-05-16 Christophe Dumez [EFL] Rounding errors on 32-bit machines causes tests to fail https://bugs.webkit.org/show_bug.cgi?id=86194 Reviewed by Martin Robinson. Pass additional CXXFLAGS to CMake in an attempt to harmonize floating point values between 32-bit and 64-bit architectures. Same flags are already used by Qt and GTK ports. * Scripts/webkitdirs.pm: (determineArchitecture): (generateBuildSystemFromCMakeProject): 2012-05-16 Balazs Kelemen [Qt] Web process hangs while waiting for plugin process connection during layouttests https://bugs.webkit.org/show_bug.cgi?id=86620 Reviewed by Csaba Osztrogonác. Temporary workaround not to fail layout tests. * Tools.pro: Disable TestNetscapePlugin. 2012-05-16 Dinu Jacob Add didFinishProgress BundleUIClient callback https://bugs.webkit.org/show_bug.cgi?id=86541 Reviewed by Kenneth Rohde Christiansen. Added support for dumpProgressFinishedCallback. * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl: * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp: (WTR::InjectedBundlePage::InjectedBundlePage): (WTR::InjectedBundlePage::didFinishProgress): (WTR): * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h: (InjectedBundlePage): * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: (WTR::LayoutTestController::LayoutTestController): * WebKitTestRunner/InjectedBundle/LayoutTestController.h: (WTR::LayoutTestController::dumpProgressFinishedCallback): (WTR::LayoutTestController::setShouldDumpProgressFinishedCallback): (WTR::LayoutTestController::shouldDumpProgressFinishedCallback): (LayoutTestController): 2012-05-15 Pierre Rossi [Qt] Enable SVG Fonts by default https://bugs.webkit.org/show_bug.cgi?id=86463 Reviewed by Nikolas Zimmermann. * qmake/mkspecs/features/features.pri: 2012-05-16 Christophe Dumez [EFL] EFL's LayoutTestController does not implement setApplicationCacheOriginQuota https://bugs.webkit.org/show_bug.cgi?id=85585 Reviewed by Gustavo Noronha Silva. Implement setApplicationCacheOriginQuota in EFL's LayoutTestController. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): * DumpRenderTree/efl/DumpRenderTreeView.cpp: (onExceededApplicationCacheQuota): (drtViewAdd): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::setApplicationCacheOriginQuota): 2012-05-16 Mikhail Pozdnyakov [EFL] EFL's LayoutTestController evaluateScriptInIsolatedWorld implementation https://bugs.webkit.org/show_bug.cgi?id=83776 Reviewed by Tor Arne Vestbø. Added missing implementation of LayoutTestController::evaluateScriptInIsolatedWorld(). * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::evaluateScriptInIsolatedWorld): 2012-05-16 Grzegorz Czajkowski Unreviewed. Add myself as a committer. * Scripts/webkitpy/common/config/committers.py: 2012-05-15 Kangil Han [EFL][DRT] Fix eventSender key providing https://bugs.webkit.org/show_bug.cgi?id=84710 Reviewed by Andreas Kling. Currently eventSender doesn't provide key event correctly. For example, 'Right' will be visible when 'right arrow' key event is fired. Given circumstance, nothing has to be shown. Therefore, this patch fixed incorrectly provided key event. * DumpRenderTree/efl/EventSender.cpp: (KeyEventInfo::KeyEventInfo): (KeyEventInfo): (keyPadNameFromJSValue): (keyNameFromJSValue): (createKeyEventInfo): (sendKeyDown): 2012-05-15 Dan Bernstein Fixed a typo. * Scripts/copy-webkitlibraries-to-product-directory: 2012-05-15 Kihong Kwon [EFL] Enable Fullscreen API https://bugs.webkit.org/show_bug.cgi?id=85870 Reviewed by Antonio Gomes. Enable full screen api feature on the EFL port. There are two failed test cases under LayoutTests/fullscreen. They are going to be fixed with other patch. * Scripts/webkitperl/FeatureList.pm: 2012-05-15 James Robinson [chromium] Chromium port never sets USE(CG) so code behind it is dead https://bugs.webkit.org/show_bug.cgi?id=86537 Reviewed by Adrienne Walker. Deletes code behind USE(CG) in chromium-specific code and removes unneeded USE(SKIA) guards. * DumpRenderTree/chromium/WebViewHost.cpp: (WebViewHost::paintRect): 2012-05-15 James Robinson [chromium] Remove WEBKIT_USING_(SKIA|CG) #defines https://bugs.webkit.org/show_bug.cgi?id=86536 Reviewed by Adam Barth. Deletes dead WEBKIT_USING_CG code and #ifdefs guarding it. * DumpRenderTree/chromium/WebThemeEngineDRTMac.mm: (WebThemeEngineDRTMac::paintHIThemeScrollbarThumb): (WebThemeEngineDRTMac::paintNSScrollerScrollbarThumb): * DumpRenderTree/chromium/WebViewHost.cpp: (WebViewHost::paintPagesWithBoundaries): 2012-05-15 James Robinson [chromium] Unreviewed build fix. Partial revert of r117166 - it appears making DumpRenderTree depend on webkit_platform causes crashes in the component build, at least on linux. This restores Tools/DumpRenderTree to its pre-r117166 state. * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp: * DumpRenderTree/chromium/TestWebPlugin.cpp: (TestWebPlugin::initialize): 2012-05-15 Elliot Poger add unit test for NRWT's --additional-expectations option https://bugs.webkit.org/show_bug.cgi?id=86501 Reviewed by Ojan Vafai. * Scripts/webkitpy/layout_tests/port/base_unittest.py: (PortTest.test_additional_expectations): 2012-05-15 James Robinson [chromium] Move createOffscreenGraphicsContext3D() from WebKitPlatformSupport to Platform https://bugs.webkit.org/show_bug.cgi?id=86524 Reviewed by Adam Barth. Updates TestWebPlugin to use the Platform headers and updates the build dependencies so it can. * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp: * DumpRenderTree/chromium/TestWebPlugin.cpp: (TestWebPlugin::initialize): 2012-05-15 Jessie Berlin WKContextSetCacheModel(contextref, kWKCacheModelDocumentViewer) doesn't prevent pages from going into the PageCache. https://bugs.webkit.org/show_bug.cgi?id=85891 Reviewed by Brady Eidson. Set the cache model in DOMWindowExtensionNoCache so that the page cache is not used. This is better than using pages with unload handlers because we may find a way to make pages with unload handlers cacheable in the future. * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Remove simple-iframe-unload.html and simple-unload.html, since they are no longer used. * TestWebKitAPI/Tests/WebKit2/DOMWindowExtensionNoCache.cpp: (TestWebKitAPI::TEST): Set the cache model to kWKCacheModelDocumentViewer. * TestWebKitAPI/Tests/WebKit2/simple-iframe-unload.html: Removed. * TestWebKitAPI/Tests/WebKit2/simple-unload.html: Removed. 2012-05-15 Kenneth Rohde Christiansen [Qt] Add infra for testing double-tap to zoom functionality etc https://bugs.webkit.org/show_bug.cgi?id=86474 Reviewed by Simon Hausmann. The WebViewportInfo class was renamed to QWebKitTest in preparation of becoming a separate module. Adapt the use API. * MiniBrowser/qt/qml/BrowserWindow.qml: * MiniBrowser/qt/qml/ViewportInfoItem.qml: 2012-05-15 Csaba Osztrogonác Introduce ENABLE_CSS_VARIABLES compile flag. https://bugs.webkit.org/show_bug.cgi?id=86338 Reviewed by Hajime Morita. * Scripts/webkitpy/layout_tests/port/webkit_unittest.py: Unittest fix after r117028. (WebKitPortTest.test_skipped_directories_for_symbols): 2012-05-15 Tor Arne Vestbø Teach webkit-patch to gracefully handle writes of raw utf-8 encoded strings to sys.stdout https://bugs.webkit.org/show_bug.cgi?id=85809 webkit-patch overrides sys.stdout with a StreamWriter that ensures we write utf8 output to stdout, but StreamWriter will by default always encode the object that's passed to write(). This blows up when the object is a raw string, already encoded as utf8, and does not match the default behavior of sys.stdout. To solve this in one place, instead of every single call-site that might write raw strings, we extend the stdout-wrapper to assume raw strings are utf8 before passing them on to StreamWriter.write(). This has the benefit of also making the stacktrace easy to read in the case where this assumption does not hold. The source of this particular bug was that Executive.run_and_throw_if_fail() will tee the output of the subprocess, and in the case of running prepare-ChangeLog this might produce non-ascii output, which gets written directly to sys.stdout as raw utf8-encoded strings. Reviewed by Dirk Pranke. * Scripts/webkit-patch: 2012-05-15 Tomasz Morawski [EFL][DRT] Split ewk_private.h file to multiple private files. https://bugs.webkit.org/show_bug.cgi?id=84572 Reviewed by Antonio Gomes. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: 2012-05-15 Michael Brüning [Qt][WK2] http/tests/navigation/https-in-page-cache.html fails with timeout https://bugs.webkit.org/show_bug.cgi?id=83576 Reviewed by Simon Hausmann. Set the PlatformWebView for the Qt port to accept all SSL certificates. * WebKitTestRunner/qt/PlatformWebViewQt.cpp: (WTR::PlatformWebView::PlatformWebView): 2012-05-14 Luke Macpherson Introduce ENABLE_CSS_VARIABLES compile flag. https://bugs.webkit.org/show_bug.cgi?id=86338 Reviewed by Dimitri Glazkov. Add a configuration option for CSS Variables support, disabling it by default. * Scripts/webkitperl/FeatureList.pm: * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitPort._missing_symbol_to_skipped_tests): * qmake/mkspecs/features/features.pri: 2012-05-14 Nate Chapin Add end-of-life test cases for https://bugs.webkit.org/show_bug.cgi?id=86131. Reviewed by Adam Barth. * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp: (callDeletePlugin): (pluginHasProperty): (pluginHasMethod): (pluginGetProperty): (pluginSetProperty): (pluginInvoke): 2012-05-14 Dirk Pranke Re-enable "drt mode" on chromium-mac-leopard https://bugs.webkit.org/show_bug.cgi?id=83076 Unreviewed, build fix (change?). This change is another attempt to gather data on the flakiness and timeouts we're seeing on Mac SL now that we have another bot. * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumDriver.__init__): 2012-05-14 Sheriff Bot Unreviewed, rolling out r116981. http://trac.webkit.org/changeset/116981 https://bugs.webkit.org/show_bug.cgi?id=86402 Caused compositing and mathml tests to fail. (Requested by rakuco on #webkit). * Scripts/webkitpy/layout_tests/port/efl.py: (EflPort._path_to_webcore_library): 2012-05-14 Dirk Pranke Fix regressions in garden-o-matic after r116988. Unreviewed, build fix. I had a typo in a field name in checkout.js, and unfortunately the unittests had the same typo :(. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/checkout.js: (.): * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/checkout_unittests.js: 2012-05-14 Dirk Pranke garden-o-matic should specify which types of baselines to rebaseline https://bugs.webkit.org/show_bug.cgi?id=86242 Reviewed by Adam Barth. This change modifies garden-o-matic so that we are explicit about whether to rebaseline the text file, the image file, or both. This should prevent us from pulling incorrect or unexpected baselines off the bots. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/checkout.js: (.): * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/checkout_unittests.js: (.): * Scripts/webkitpy/tool/commands/rebaseline.py: (AbstractRebaseliningCommand): (AbstractRebaseliningCommand.__init__): (RebaselineTest): (RebaselineTest._rebaseline_test_and_update_expectations): (RebaselineTest.execute): (OptimizeBaselines): (OptimizeBaselines._optimize_baseline): (OptimizeBaselines.execute): (AnalyzeBaselines): (AnalyzeBaselines._analyze_baseline): (AnalyzeBaselines.execute): (RebaselineExpectations._rebaseline_port): * Scripts/webkitpy/tool/servers/gardeningserver.py: (GardeningHTTPRequestHandler.rebaseline): (GardeningHTTPRequestHandler.optimizebaselines): * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py: (GardeningServerTest.test_rebaseline): (GardeningServerTest.test_rebaseline_new_port): (GardeningServerTest.test_optimizebaselines): 2012-05-14 Kangil Han [EFL][DRT] Fix WebCore library path https://bugs.webkit.org/show_bug.cgi?id=86355 Reviewed by Dirk Pranke. Currently EFL DRT uses wrong WebCore library path when running DRT. This patch simply adjusted it. * Scripts/webkitpy/layout_tests/port/efl.py: (EflPort._path_to_webcore_library): 2012-05-14 Wei James [Chromium] ImageDiff should be build for host on Android https://bugs.webkit.org/show_bug.cgi?id=82039 Reviewed by Adam Barth. * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp: 2012-05-14 Joone Hur Unreviewed. Updating email for committers.py script. * Scripts/webkitpy/common/config/committers.py: 2012-05-14 Carlos Garcia Campos Unreviewed. Fix make distcheck. * GNUmakefile.am: Add run-api-tests script to EXTRA_DIST. * gtk/run-api-tests: (TestRunner._lookup_atspi2_binary): Handle the case when pkg_config_file_variable() return None. 2012-05-14 Sheriff Bot Unreviewed, rolling out r116935. http://trac.webkit.org/changeset/116935 https://bugs.webkit.org/show_bug.cgi?id=86373 Something is still broken (Requested by Ossy on #webkit). * Scripts/webkitpy/common/checksvnconfigfile.py: Removed. * Scripts/webkitpy/style/checkers/png.py: (PNGChecker.check): (PNGChecker): (PNGChecker._config_file_path): * Scripts/webkitpy/tool/commands/download.py: (Land): * Scripts/webkitpy/tool/steps/__init__.py: * Scripts/webkitpy/tool/steps/addsvnmimetypeforpng.py: Removed. * Scripts/webkitpy/tool/steps/addsvnmimetypeforpng_unittest.py: Removed. 2012-05-14 Allan Sandfeld Jensen [Qt] Doesn't build with ENABLE_INSPECTOR=0 https://bugs.webkit.org/show_bug.cgi?id=85056 Reviewed by Tor Arne Vestbø. * WebKitTestRunner/TestInvocation.cpp: Ifdef otherwise unused static function. * qmake/mkspecs/features/features.prf: * qmake/mkspecs/features/features.pri: 2012-05-11 Tor Arne Vestbø [Qt] Make sure dialogs have their parent and attached properties set before Component.onCompleted https://bugs.webkit.org/show_bug.cgi?id=86222 The WebView QML api for dialogs is based on delegates for each dialog, which are implemented as components. Since there's no explicit API to show the dialog, the natural place to allow the dialog to choose how to display itself is in Component.onCompleted. The issue was that we were using QQmlComponent::create() when instantiating the component, at which point the component did not have a parent or it's attached WebView property set. That meant that we had to do some ugly workaround in our test, using a 1ms timer, to get a callback once the parent was there. We now use the QQmlComponent::beginCreate() -> QQmlComponent::completeCreate() pair, which allows us to set the relevant properties before triggering the Component.onCompleted callback. We were already doing this for the context menus (except for setting the parent, which we did afterwards). Reviewed by Simon Hausmann. * MiniBrowser/qt/qml/BrowserWindow.qml: * MiniBrowser/qt/qml/Dialog.qml: 2012-05-14 Balazs Ankes webkit-patch land should automatically add svn:mime-type for .png files https://bugs.webkit.org/show_bug.cgi?id=75825 I refactored the png.py to avoid code duplication. Reviewed by Dirk Pranke. * Scripts/webkitpy/common/checksvnconfigfile.py: Added. (check): (_config_file_path): (errorstr_autoprop): (errorstr_png): * Scripts/webkitpy/style/checkers/png.py: (PNGChecker.check): * Scripts/webkitpy/tool/commands/download.py: (Land): * Scripts/webkitpy/tool/steps/__init__.py: * Scripts/webkitpy/tool/steps/addsvnmimetypeforpng.py: Added. (AddSvnMimetypeForPng): (AddSvnMimetypeForPng.__init__): (AddSvnMimetypeForPng.run): (AddSvnMimetypeForPng._check_pngs): * Scripts/webkitpy/tool/steps/addsvnmimetypeforpng_unittest.py: Added. (MockSCMDetector): (MockSCMDetector.__init__): (MockSCMDetector.display_name): (AddSvnMimetypeForPngTest): (AddSvnMimetypeForPngTest.test_run): 2012-05-14 David Kilzer webkit-patch apply-attachment is very slow for big patches https://bugs.webkit.org/show_bug.cgi?id=74929 Add "excludefield=attachmentdata" to the URL everywhere "ctype=xml" is used to exclude this field from the XML. Reviewed by Adam Barth. * Scripts/prepare-ChangeLog: (fetchBugDescriptionFromURL): * Scripts/webkitpy/common/config/urls.py: * Scripts/webkitpy/common/config/urls_unittest.py: (URLsTest.test_parse_bug_id): * Scripts/webkitpy/common/net/bugzilla/bugzilla.py: (Bugzilla.bug_url_for_bug_id): 2012-05-13 Ryosuke Niwa performance tests should be able to measure runs/sec rather than time https://bugs.webkit.org/show_bug.cgi?id=86021 Reviewed by Ojan Vafai. Allow " runs/s" or " ms" to appear after numerical values in tests. * Scripts/webkitpy/performance_tests/perftest.py: (PerfTest): 2012-05-12 Tim Horton run-safari and co. should support --guard-malloc command line argument https://bugs.webkit.org/show_bug.cgi?id=86287 Reviewed by Dan Bernstein. Add setUpGuardMallocIfNeeded(), which consults the argument --guard-malloc on OS X, setting DYLD_INSERT_LIBRARIES="/usr/lib/libgmalloc.dylib" if it's enabled. Use setUpGuardMallocIfNeeded() before running a WebKit app (whether standalone or in the debugger). This will cause gmalloc to be loaded for gdb as well, but this tends to be harmless. Change "malloc guard" to the more common "Guard Malloc" in run-webkit-tests. * Scripts/old-run-webkit-tests: * Scripts/webkitdirs.pm: (appendToEnvironmentVariableList): (setUpGuardMallocIfNeeded): (printHelpAndExitForRunAndDebugWebKitAppIfNeeded): (runMacWebKitApp): (execMacWebKitAppForDebugging): * Scripts/webkitpy/layout_tests/run_webkit_tests.py: (parse_args): 2012-05-11 Ojan Vafai Add previous/next buttons the garden-o-matics perf panel https://bugs.webkit.org/show_bug.cgi?id=86280 Reviewed by Dimitri Glazkov. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/perf.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/perf_unittests.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/styles/perf.css: 2012-05-11 Kevin Ollivier [wx] Unreviewed build fix. Fix memory exhausted errors from 32-bit ld on *nix. * waf/build/settings.py: (common_configure): 2012-05-11 Dirk Pranke webkit-patch rebaseline should fetch stale expectations :) https://bugs.webkit.org/show_bug.cgi?id=86231 Unreviewed, build fix. This change reverts r115492 and goes back to fetching individual baselines from the layout tests directory, rather than fetching entire zip files. The performance hit over networks for people far away from the bots has turned out to be severe, so we need a better solution to this. * Scripts/webkitpy/tool/commands/rebaseline.py: (RebaselineTest._results_url): (RebaselineTest._file_name_for_expected_result): (RebaselineTest._rebaseline_test): * Scripts/webkitpy/tool/commands/rebaseline_unittest.py: (TestRebaseline.test_rebaseline_updates_expectations_file_noop): (test_rebaseline_updates_expectations_file): (test_rebaseline_test): (test_rebaseline_and_copy_test): (test_rebaseline_and_copy_test_no_existing_result): (test_rebaseline_and_copy_test_with_lion_result): (test_rebaseline_and_copy_no_overwrite_test): 2012-05-11 Christophe Dumez [GTK] LayoutTests/svg/batik/text/verticalText.svg fails to display Chinese characters https://bugs.webkit.org/show_bug.cgi?id=86198 Reviewed by Martin Robinson. Update webkitgtk-fonts package to v0.0.3 in order to get new Kochi Mincho / Gothic fonts to display Japanese and Chinese characters. * gtk/jhbuild.modules: 2012-05-11 Christophe Dumez [EFL][DRT] SVG Batik test fails to display Chinese characters https://bugs.webkit.org/show_bug.cgi?id=84461 Reviewed by Martin Robinson. Update webkitgtk-fonts package from v0.0.1 to v0.0.3 in order to get new Kochi Mincho / Gothic fonts to display Japanese and Chinese characters. * DumpRenderTree/efl/CMakeLists.txt: * efl/jhbuild.modules: 2012-05-11 Philippe Normand Unreviewed, GTK 64-bit Debug bot fix attempt. * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkPort._get_gdb_output): Ignore unicode decoding errors when processing gdb output. 2012-05-11 Gustavo Noronha Silva REGRESSION(r116205): [GTK]: build no longer supports thin archives https://bugs.webkit.org/show_bug.cgi?id=86207 * GNUmakefile.am: link libWebCoreModules to DRT. 2012-05-10 Ojan Vafai Add a perf tab to garden-o-matic https://bugs.webkit.org/show_bug.cgi?id=86169 Reviewed by Dimitri Glazkov. This is a first stab at making it easier to view the Chromium WebKit canary perf graphs. For now, it's just a couple select elements and an iframe. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/garden-o-matic.html: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/run-unittests.html: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/builders.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/builders_unittests.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/garden-o-matic.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/perf.js: Added. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/perf_unittests.js: Added. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui_unittests.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/styles/perf.css: Added. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/styles/results.css: 2012-05-10 Gyuyoung Kim Move resumeAnimations to use Internals interface https://bugs.webkit.org/show_bug.cgi?id=86063 Reviewed by Alexey Proskuryakov. Remove resumeAnimations functions, because it is able to work in the cross-port way through the Internals interface. * DumpRenderTree/LayoutTestController.cpp: (LayoutTestController::staticFunctions): * DumpRenderTree/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp: * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::LayoutTestController): * DumpRenderTree/chromium/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: * DumpRenderTree/mac/LayoutTestControllerMac.mm: * DumpRenderTree/qt/LayoutTestControllerQt.cpp: * DumpRenderTree/qt/LayoutTestControllerQt.h: (LayoutTestController): * DumpRenderTree/win/LayoutTestControllerWin.cpp: * DumpRenderTree/wx/LayoutTestControllerWx.cpp: * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl: * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: * WebKitTestRunner/InjectedBundle/LayoutTestController.h: (LayoutTestController): 2012-05-10 Alexey Proskuryakov Crash in 3rd party WebKit apps that disable cache at a wrong time https://bugs.webkit.org/show_bug.cgi?id=86027 Reviewed by Antti Koivisto. Added a test that's very similar to MemoryCachePruneWithinResourceLoadDelegate, but for disabling the cache instead of triggering a prune. * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: * TestWebKitAPI/Tests/mac/MemoryCacheDisableWithinResourceLoadDelegate.html: Added. * TestWebKitAPI/Tests/mac/MemoryCacheDisableWithinResourceLoadDelegate.mm: Added. (-[MemoryCacheDisableTestResourceLoadDelegate webView:identifierForInitialRequest:fromDataSource:]): (-[MemoryCacheDisableTestResourceLoadDelegate webView:resource:willSendRequest:redirectResponse:fromDataSource:]): (-[MemoryCacheDisableTestResourceLoadDelegate webView:resource:didFinishLoadingFromDataSource:]): (-[MemoryCacheDisableTestResourceLoadDelegate webView:resource:didFailLoadingWithError:fromDataSource:]): (TestWebKitAPI::TEST): 2012-05-10 Anders Carlsson WebKit2: Add a way to blacklist specific plug-ins/plug-in versions https://bugs.webkit.org/show_bug.cgi?id=86164 Reviewed by Sam Weinig. Update for WK2 API changes. * MiniBrowser/mac/BrowserWindowController.m: (-[BrowserWindowController awakeFromNib]): * WebKitTestRunner/TestController.cpp: (WTR::TestController::createOtherPage): (WTR::TestController::initialize): 2012-05-10 Dirk Pranke nrwt: [chromium] fix path to asan symbolize script https://bugs.webkit.org/show_bug.cgi?id=86161 Reviewed by Abhishek Arya. * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumPort._get_crash_log): 2012-05-10 Lei Zhang [chromium] Set Valgrind environment variables for layout tests. https://bugs.webkit.org/show_bug.cgi?id=86156 Reviewed by Dirk Pranke. * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumPort.setup_environ_for_server): 2012-05-10 Raphael Kubo da Costa [webkitpy][EFL] It should be possible to use a wrapper command for ImageDiff. https://bugs.webkit.org/show_bug.cgi?id=86145 Reviewed by Dirk Pranke. The GTK+ and EFL ports use jhbuild to build a few base libraries which are then used when building the port itself. Programs such as DRT and ImageDiff are then supposed to be run with the `run-with-jhbuild' wrapper so that the libraries are properly found and loaded. This was already the case for DumpRenderTree (EflPort defines the 'wrapper' option for that), but it was not possible to do the same with ImageDiff so far. The code which created the ImageDiff command-line call has been moved to a separate method, _image_diff_command, which is then overridden by EflPort to add the jhbuild wrapper. * Scripts/webkitpy/layout_tests/port/efl.py: (EflPort.__init__): (EflPort._image_diff_command): * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitPort): (WebKitPort._image_diff_command): (WebKitPort._start_image_diff_process): 2012-05-10 Zan Dobersek [GTK] ENABLE_IFRAME_SEAMLESS support https://bugs.webkit.org/show_bug.cgi?id=85843 Reviewed by Eric Seidel. Add the iframe-seamless option to the options list. This makes it possible to enable the iframe seamless support through build-webkit script. Enabled by default for all the ports. * Scripts/webkitperl/FeatureList.pm: 2012-05-10 Raphael Kubo da Costa [EFL][webkitpy] Do not redefine check_build() in EflPort. https://bugs.webkit.org/show_bug.cgi?id=86124 Reviewed by Dirk Pranke. Remove an old FIXME and do not redefine check_build in EflPort; once build-dumprendertree recognizes --efl, we can use the check_build implementation in WebKitPort. * Scripts/build-dumprendertree: * Scripts/webkitpy/layout_tests/port/efl.py: (EflPort.check_build): 2012-05-10 Tommy Widenflycht [chromium] MediaStream API: Fix the ExtraData functionality in WebMediaStreamDescriptor https://bugs.webkit.org/show_bug.cgi?id=86087 Reviewed by Adam Barth. * DumpRenderTree/chromium/WebUserMediaClientMock.cpp: (MockExtraData): (WebKit): (WebKit::WebUserMediaClientMock::requestUserMedia): 2012-05-10 Varun Jain [chromium] Trigger context menu for long press gesture https://bugs.webkit.org/show_bug.cgi?id=85919 Reviewed by Adam Barth. * DumpRenderTree/chromium/EventSender.cpp: (EventSender::EventSender): (EventSender::gestureLongPress): (EventSender::gestureEvent): * DumpRenderTree/chromium/EventSender.h: (EventSender): 2012-05-10 Takashi Toyoshima don't use autoinstall to import pywebsocket but check it in WebKit directly. https://bugs.webkit.org/show_bug.cgi?id=86107 Reviewed by Adam Barth. This change removes pywebsocket from webkitpy's autoinstall list and imports pywebsocket 0.7.4 directly into webkitpy/thirparty. * Scripts/webkitpy/layout_tests/servers/websocket_server.py: (PyWebSocket._prepare_config): * Scripts/webkitpy/thirdparty/__init__.py: (AutoinstallImportHook.find_module): (AutoinstallImportHook._install_irc): * Scripts/webkitpy/thirdparty/__init___unittest.py: (ThirdpartyTest.test_import_hook.MockImportHook.__init__): (ThirdpartyTest.test_import_hook.MockImportHook._install_eliza): (ThirdpartyTest): (ThirdpartyTest.test_import_hook): * Scripts/webkitpy/thirdparty/mod_pywebsocket/COPYING: Added. * Scripts/webkitpy/thirdparty/mod_pywebsocket/__init__.py: Added. * Scripts/webkitpy/thirdparty/mod_pywebsocket/_stream_base.py: Added. (ConnectionTerminatedException): (InvalidFrameException): (BadOperationException): (UnsupportedFrameException): (InvalidUTF8Exception): (StreamBase): (StreamBase.__init__): (StreamBase._read): (StreamBase._write): (StreamBase.receive_bytes): (StreamBase._read_until): * Scripts/webkitpy/thirdparty/mod_pywebsocket/_stream_hixie75.py: Added. (for): (StreamHixie75): (StreamHixie75.for): (StreamHixie75.__init__): (StreamHixie75.send_message): (StreamHixie75._read_payload_length_hixie75): (StreamHixie75.receive_message): (StreamHixie75._send_closing_handshake): (StreamHixie75.close_connection): (StreamHixie75.send_ping): * Scripts/webkitpy/thirdparty/mod_pywebsocket/_stream_hybi.py: Added. (Frame): (Frame.__init__): (create_length_header): (create_header): (_build_frame): (_filter_and_format_frame_object): (create_binary_frame): (create_text_frame): (FragmentedFrameBuilder): (FragmentedFrameBuilder.to): (FragmentedFrameBuilder.__init__): (FragmentedFrameBuilder.build): (_create_control_frame): (create_ping_frame): (create_pong_frame): (create_close_frame): (StreamOptions): (StreamOptions.__init__): (Stream): (Stream.for): (Stream.__init__): (Stream._receive_frame): (Stream._receive_frame_as_frame_object): (Stream.send_message): (Stream.receive_message): (Stream._send_closing_handshake): (Stream.close_connection): (Stream.send_ping): (Stream._send_pong): (Stream._drain_received_data): * Scripts/webkitpy/thirdparty/mod_pywebsocket/common.py: Added. (is_control_opcode): (ExtensionParameter): (ExtensionParameter.__init__): (ExtensionParameter.name): (ExtensionParameter.add_parameter): (ExtensionParameter.get_parameters): (ExtensionParameter.get_parameter_names): (ExtensionParameter.has_parameter): (ExtensionParameter.get_parameter_value): * Scripts/webkitpy/thirdparty/mod_pywebsocket/dispatch.py: Added. (DispatchException): (DispatchException.__init__): (_default_passive_closing_handshake_handler): (_normalize_path): (_create_path_to_resource_converter): (_create_path_to_resource_converter.converter): (_enumerate_handler_file_paths): (_HandlerSuite): (_HandlerSuite.__init__): (_source_handler_file): (_extract_handler): (Dispatcher): (Dispatcher.maintains): (Dispatcher.__init__): (Dispatcher.add_resource_path_alias): (Dispatcher.source_warnings): (Dispatcher.do_extra_handshake): (Dispatcher.transfer_data): (Dispatcher.passive_closing_handshake): (Dispatcher.get_handler_suite): (Dispatcher._source_handler_files_in_dir): * Scripts/webkitpy/thirdparty/mod_pywebsocket/extensions.py: Added. (ExtensionProcessorInterface): (ExtensionProcessorInterface.get_extension_response): (ExtensionProcessorInterface.setup_stream_options): (DeflateStreamExtensionProcessor): (DeflateStreamExtensionProcessor.__init__): (DeflateStreamExtensionProcessor.get_extension_response): (DeflateStreamExtensionProcessor.setup_stream_options): (DeflateFrameExtensionProcessor): (DeflateFrameExtensionProcessor.__init__): (DeflateFrameExtensionProcessor.get_extension_response): (DeflateFrameExtensionProcessor.setup_stream_options): (DeflateFrameExtensionProcessor.setup_stream_options._OutgoingFilter): (DeflateFrameExtensionProcessor.setup_stream_options._OutgoingFilter.__init__): (DeflateFrameExtensionProcessor.setup_stream_options._OutgoingFilter.filter): (DeflateFrameExtensionProcessor.setup_stream_options._IncomingFilter): (DeflateFrameExtensionProcessor.setup_stream_options._IncomingFilter.__init__): (DeflateFrameExtensionProcessor.setup_stream_options._IncomingFilter.filter): (DeflateFrameExtensionProcessor.set_response_window_bits): (DeflateFrameExtensionProcessor.set_response_no_context_takeover): (DeflateFrameExtensionProcessor.enable_outgoing_compression): (DeflateFrameExtensionProcessor.disable_outgoing_compression): (DeflateFrameExtensionProcessor._outgoing_filter): (DeflateFrameExtensionProcessor._incoming_filter): (get_extension_processor): (get_extension_processor.is): * Scripts/webkitpy/thirdparty/mod_pywebsocket/handshake/__init__.py: Added. (try): (do_handshake): * Scripts/webkitpy/thirdparty/mod_pywebsocket/handshake/_base.py: Added. (AbortedByUserException): (HandshakeException): (HandshakeException.__init__): (VersionException): (VersionException.__init__): (get_default_port): (validate_subprotocol): (parse_host_header): (format_header): (build_location): (get_mandatory_header): (validate_mandatory_header): (check_request_line): (check_header_lines): (parse_token_list): (_parse_extension_param): (_parse_extension): (parse_extensions): (format_extensions): * Scripts/webkitpy/thirdparty/mod_pywebsocket/handshake/draft75.py: Added. (Handshaker): (Handshaker.performs): (Handshaker.__init__): (Handshaker.do_handshake): (Handshaker._set_resource): (Handshaker._set_origin): (Handshaker._set_location): (Handshaker._set_subprotocol): (Handshaker._set_protocol_version): (Handshaker._sendall): (Handshaker._send_handshake): (Handshaker._check_header_lines): (Handshaker._check_first_lines): * Scripts/webkitpy/thirdparty/mod_pywebsocket/handshake/hybi.py: Added. (compute_accept): (Handshaker): (Handshaker.__init__): (Handshaker._validate_connection_header): (Handshaker.do_handshake): (Handshaker._get_origin): (Handshaker._check_version): (Handshaker._set_protocol): (Handshaker._parse_extensions): (Handshaker._validate_key): (Handshaker._get_key): (Handshaker._send_handshake): * Scripts/webkitpy/thirdparty/mod_pywebsocket/handshake/hybi00.py: Added. (Handshaker): (Handshaker.__init__): (Handshaker.do_handshake): (Handshaker._set_resource): (Handshaker._set_subprotocol): (Handshaker._set_location): (Handshaker._set_origin): (Handshaker._set_protocol_version): (Handshaker._set_challenge_response): (Handshaker._get_key_value): (Handshaker._get_challenge): (Handshaker._send_handshake): * Scripts/webkitpy/thirdparty/mod_pywebsocket/headerparserhandler.py: Added. (ApacheLogHandler): (ApacheLogHandler.__init__): (ApacheLogHandler.emit): (_configure_logging): (_parse_option): (_create_dispatcher): (headerparserhandler): * Scripts/webkitpy/thirdparty/mod_pywebsocket/http_header_util.py: Added. (_is_char): (_is_ctl): (ParsingState): (ParsingState.__init__): (peek): (consume): (consume_string): (consume_lws): (consume_lwses): (consume_token): (consume_token_or_quoted_string): (quote_if_necessary): (parse_uri): * Scripts/webkitpy/thirdparty/mod_pywebsocket/memorizingfile.py: Added. (MemorizingFile): (MemorizingFile.__init__): (MemorizingFile.__getattribute__): (MemorizingFile.readline): (MemorizingFile.get_memorized_lines): * Scripts/webkitpy/thirdparty/mod_pywebsocket/msgutil.py: Added. (close_connection): (send_message): (receive_message): (send_ping): (MessageReceiver): (MessageReceiver.receives): (MessageReceiver.provides): (MessageReceiver.should): (MessageReceiver.__init__): (MessageReceiver.run): (MessageReceiver.receive): (MessageReceiver.receive_nowait): (MessageReceiver.stop): (MessageSender): (MessageSender.sends): (MessageSender.provides): (MessageSender.should): (MessageSender.__init__): (MessageSender.run): (MessageSender.send): (MessageSender.send_nowait): * Scripts/webkitpy/thirdparty/mod_pywebsocket/standalone.py: Added. (_StandaloneConnection): (_StandaloneConnection.__init__): (_StandaloneConnection.get_local_addr): (_StandaloneConnection.get_remote_addr): (_StandaloneConnection.write): (_StandaloneConnection.read): (_StandaloneConnection.get_memorized_lines): (_StandaloneRequest): (_StandaloneRequest.__init__): (_StandaloneRequest.get_uri): (_StandaloneRequest.get_method): (_StandaloneRequest.is_https): (_StandaloneRequest._drain_received_data): (_StandaloneSSLConnection): (_StandaloneSSLConnection.for): (_StandaloneSSLConnection.__init__): (_StandaloneSSLConnection.__getattribute__): (_StandaloneSSLConnection.__setattr__): (_StandaloneSSLConnection.makefile): (WebSocketServer): (WebSocketServer.__init__): (WebSocketServer._create_sockets): (WebSocketServer.server_bind): (WebSocketServer.server_activate): (WebSocketServer.server_close): (WebSocketServer.fileno): (WebSocketServer.handle_error): (WebSocketServer.get_request): (WebSocketServer.serve_forever): (WebSocketServer.shutdown): (WebSocketRequestHandler): (WebSocketRequestHandler.setup): (WebSocketRequestHandler.setup.SocketServer): (WebSocketRequestHandler.__init__): (WebSocketRequestHandler.parse_request): (WebSocketRequestHandler.log_request): (WebSocketRequestHandler.log_error): (WebSocketRequestHandler.is_cgi): (_configure_logging): (_alias_handlers): (_build_option_parser): (ThreadMonitor): (ThreadMonitor.__init__): (ThreadMonitor.run): (_parse_args_and_config): (_main): (_main.if): (_main.if.__check_script): * Scripts/webkitpy/thirdparty/mod_pywebsocket/stream.py: Copied from Tools/Scripts/webkitpy/thirdparty/__init___unittest.py. * Scripts/webkitpy/thirdparty/mod_pywebsocket/util.py: Added. (get_stack_trace): (prepend_message_to_exception): (__translate_interp): (get_script_interp): (wrap_popen3_for_win): (wrap_popen3_for_win.__wrap_popen3): (hexify): (get_class_logger): (NoopMasker): (NoopMasker.__init__): (NoopMasker.mask): (RepeatedXorMasker): (RepeatedXorMasker.__init__): (RepeatedXorMasker.mask): (DeflateRequest): (DeflateRequest.for): (DeflateRequest.__init__): (DeflateRequest.__getattribute__): (DeflateRequest.__setattr__): (_Deflater): (_Deflater.__init__): (_Deflater.compress_and_flush): (_Inflater): (_Inflater.__init__): (_Inflater.decompress): (_Inflater.append): (_Inflater.reset): (_RFC1979Deflater): (_RFC1979Deflater.that): (_RFC1979Deflater.__init__): (_RFC1979Deflater.filter): (_RFC1979Inflater): (_RFC1979Inflater.for): (_RFC1979Inflater.__init__): (_RFC1979Inflater.filter): (DeflateSocket): (DeflateSocket.for): (DeflateSocket.__init__): (DeflateSocket.recv): (DeflateSocket.sendall): (DeflateSocket.send): (DeflateConnection): (DeflateConnection.for): (DeflateConnection.__init__): (DeflateConnection.get_remote_addr): (DeflateConnection.put_bytes): (DeflateConnection.read): (DeflateConnection.write): (_is_ewouldblock_errno): (drain_received_data): 2012-05-10 Raphael Kubo da Costa [EFL][webkitpy] Define the 'wrapper' option in EflPort instead of creating a EflDriver. https://bugs.webkit.org/show_bug.cgi?id=86117 Reviewed by Adam Barth. Simplify the code used to wrap driver calls with jhbuild: we can just define the 'wrapper' option instead of having to create a Driver class just for that. * Scripts/webkitpy/layout_tests/port/efl.py: (EflPort.__init__): (EflPort._port_flag_for_scripts): 2012-05-10 Sheriff Bot Unreviewed, rolling out r116633. http://trac.webkit.org/changeset/116633 https://bugs.webkit.org/show_bug.cgi?id=86093 Breaks EFL tests (Requested by kov on #webkit). * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): * DumpRenderTree/efl/EventSender.cpp: (keyDownCallback): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::setTabKeyCyclesThroughElements): 2012-05-10 Sudarsana Nagineni [EFL] Memory leaks found in EFL DRT https://bugs.webkit.org/show_bug.cgi?id=85889 Reviewed by Gustavo Noronha Silva. Fix two memory leaks in EFL's DRT by using JSRetainPtr for JSString created with JSStringCreateWithUTF8CString. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::onWindowObjectCleared): * DumpRenderTree/efl/EventSender.cpp: (modifiersFromJSValue): 2012-05-10 Alexander Shalamov [EFL] LayoutTestController needs implementation of setTabKeyCyclesThroughElements https://bugs.webkit.org/show_bug.cgi?id=82864 Reviewed by Gustavo Noronha Silva. Add missing implementation for setTabKeyCyclesThroughElements to EFL's LayoutTestController in order to unskip fast/events/keypress-insert-tab.html * DumpRenderTree/efl/EventSender.cpp: (keyDownCallback): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::setTabKeyCyclesThroughElements): 2012-05-10 Alexander Shalamov [EFL][DRT] EventSender - set delayed event information if it was not set before https://bugs.webkit.org/show_bug.cgi?id=85972 Reviewed by Gustavo Noronha Silva. Reset delayed event information if it is missing. * DumpRenderTree/efl/EventSender.cpp: (feedOrQueueMouseEvent): 2012-05-07 Tor Arne Vestbø WebPageProxy::activeURL() should return the pending API request, even when there's no main frame https://bugs.webkit.org/show_bug.cgi?id=85806 The m_pendingAPIRequestURL member is used (presumably) to mask over the async nature of WebKit2, so that starting a load of a URL will reflect that URL immedeatly from activeURL, even if the request has not been passed over to the web process yet and reflected there. This works well, except in the case of the initial request, where the main frame creation happens on the web process side and is notified back to the UI process. Until we've recived the notification we don't know about the main frame, and this race condition will potentially give us an empty url instead of the pending request. To solve this we always return the pending API request if it's set, even when there's no mainframe yet (that we known about). Reviewed by Simon Hausmann. * TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp: 2012-05-10 Mikhail Pozdnyakov [EFL][DRT] Clear added user style sheets before a new testcase execution. https://bugs.webkit.org/show_bug.cgi?id=84792 Reviewed Gustavo Noronha Silva. The added user style sheets from the previous testcases are cleared in order to avoid flaky results. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): 2012-05-09 Gyuyoung Kim Move suspendAnimations to use Internals interface. https://bugs.webkit.org/show_bug.cgi?id=85986 Reviewed by Ryosuke Niwa. Remove suspendAnimations functions, because it is able to work in the cross-port way through the Internals interface * DumpRenderTree/LayoutTestController.cpp: (LayoutTestController::staticFunctions): * DumpRenderTree/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp: * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::LayoutTestController): * DumpRenderTree/chromium/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: * DumpRenderTree/mac/LayoutTestControllerMac.mm: * DumpRenderTree/qt/LayoutTestControllerQt.cpp: * DumpRenderTree/qt/LayoutTestControllerQt.h: (LayoutTestController): * DumpRenderTree/win/LayoutTestControllerWin.cpp: * DumpRenderTree/wx/LayoutTestControllerWx.cpp: 2012-05-09 Kent Tamura [Chromium] Attempt to fix an Android build error. https://bugs.webkit.org/show_bug.cgi?id=86045 * DumpRenderTree/DumpRenderTree.gypi: 2012-05-09 Ojan Vafai Don't show partytime if there are non-layout test failures. https://bugs.webkit.org/show_bug.cgi?id=86056 Reviewed by Dimitri Glazkov. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/controllers.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/controllers_unittests.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/garden-o-matic.js: 2012-05-09 Hayato Ito Improve a warning message for a reftest having an unused expectation file. https://bugs.webkit.org/show_bug.cgi?id=85975 Reviewed by Adam Barth. The current warning message is not intuitive. It might be better to update the message so that it encourage users to remove unused expectation files. * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py: (SingleTestRunner.__init__): 2012-05-09 Ojan Vafai Change "losing test coverage" to cover all non-layout test problems in garden-o-matic https://bugs.webkit.org/show_bug.cgi?id=86043 Reviewed by Adam Barth. -Get rid of the gtest iframe. -Cover all non-layout test problems. Before we wouldn't catch compile failures or gtest failures since they were not on the webkit testing bots. -Include the list of failing steps next to the builder name. This is a lot less cluttered than you'd expect and already helped me find a compile failure early. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/builders.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/builders_unittests.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/config.js: Delete unused map. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/controllers.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/garden-o-matic.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/failures.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/failures_unittests.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/notifications.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/notifications_unittests.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui_unittests.js: 2012-05-09 Jessie Berlin Crash using the new WKBundleDOMWindowExtensions APIs. https://bugs.webkit.org/show_bug.cgi?id=85888 Reviewed by Brady Eidson. Cached frames can live slightly longer than the page, but most clients unregister themselves and do other cleanup in the willDestroyPage callback, making them miss the willDestroyGlobalObjectForDOMWindowExtension callbacks. The calls to willDestroyGlobalObjectForDOMWindowExtension in the DOMWindowExtensionBasic test were all being invoked underneath WebPage::close. This is unrealistic. Update that test to destroy the BundleDOMWindowExtensions in response to the willDestroyPage callback. Add a test to verify that willDestroyGlobalObjectForDOMWindowExtension is being called for pages that don't go into the page cache. * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Add DOMWindowExtensionNoCache.cpp, DOMWindowExtensionNoCache_Bundle.cpp, simple-unload.html and simple-iframe-unload.html * TestWebKitAPI/Tests/WebKit2/DOMWindowExtensionBasic.cpp: (TestWebKitAPI): Remove the expected messages for willDestroyGlobalObjectForDOMWindowExtension. (TestWebKitAPI::didReceiveMessageFromInjectedBundle): Do not bother to keep track of the live extension count - all of them are expected to be live until the test completes. (TestWebKitAPI::TEST): Fix the calls to EXPECT to pass the expected value first, and use EXPECT_WK_STREQ so that message failures will be clearer. * TestWebKitAPI/Tests/WebKit2/DOMWindowExtensionBasic_Bundle.cpp: (TestWebKitAPI::DOMWindowExtensionBasic::willDestroyPage): Clean up the BundleDOMWindowExtensions. (TestWebKitAPI::DOMWindowExtensionBasic::willDestroyGlobalObjectForDOMWindowExtension): Add an ASSERT_NOT_REACHED. * TestWebKitAPI/Tests/WebKit2/DOMWindowExtensionNoCache.cpp: Added. (TestWebKitAPI::didReceiveMessageFromInjectedBundle): Keep track of the messages received so they can be checked at the end of the test. (TestWebKitAPI::TEST): Navigate to uncacheable pages and back. * TestWebKitAPI/Tests/WebKit2/DOMWindowExtensionNoCache_Bundle.cpp: Copied from Tools/TestWebKitAPI/Tests/WebKit2/DOMWindowExtensionBasic_Bundle.cpp. (DOMWindowExtensionNoCache): (TestWebKitAPI::DOMWindowExtensionNoCache::DOMWindowExtensionNoCache): Set up all the states for each BundleDOMWindowExtension. (TestWebKitAPI::DOMWindowExtensionNoCache::frameLoadFinished): Tell the UI Process about the states of the BundleDOMWindowExtensions. (TestWebKitAPI::DOMWindowExtensionNoCache::sendExtensionStateMessage): (TestWebKitAPI::DOMWindowExtensionNoCache::initialize): (TestWebKitAPI::DOMWindowExtensionNoCache::didCreatePage): (TestWebKitAPI::DOMWindowExtensionNoCache::willDestroyPage): Remvoe the remaining BundleDOMWindowExtensions, send the updated state, and finish the test. (TestWebKitAPI::DOMWindowExtensionNoCache::updateExtensionStateRecord): (TestWebKitAPI::DOMWindowExtensionNoCache::sendBundleMessage): (TestWebKitAPI::DOMWindowExtensionNoCache::globalObjectIsAvailableForFrame): (TestWebKitAPI::DOMWindowExtensionNoCache::willDisconnectDOMWindowExtensionFromGlobalObject): ASSERT that these pages not going into the page cache are not getting disconnected to go into the page cache. (TestWebKitAPI::DOMWindowExtensionNoCache::didReconnectDOMWindowExtensionToGlobalObject): Ditto about getting reconnected when coming out of the page cache. (TestWebKitAPI::DOMWindowExtensionNoCache::willDestroyGlobalObjectForDOMWindowExtension): Tell the UI Process, update the state, and get rid of the BundleDOMWindowExtension. (TestWebKitAPI::didFinishLoadForFrameCallback): (TestWebKitAPI::globalObjectIsAvailableForFrameCallback): (TestWebKitAPI::willDisconnectDOMWindowExtensionFromGlobalObjectCallback): (TestWebKitAPI::didReconnectDOMWindowExtensionToGlobalObjectCallback): (TestWebKitAPI::willDestroyGlobalObjectForDOMWindowExtensionCallback): * TestWebKitAPI/Tests/WebKit2/simple-iframe-unload.html: Added. * TestWebKitAPI/Tests/WebKit2/simple-unload.html: Added. 2012-05-09 Adam Barth [Chromium] Android's DRT doesn't want to build WebUserMediaClientMock https://bugs.webkit.org/show_bug.cgi?id=86045 Reviewed by Eric Seidel. * DumpRenderTree/DumpRenderTree.gypi: 2012-05-09 Adam Barth [Chromium] MockWebSpeechInputController.cpp should be guarded by ENABLE(INPUT_SPEECH) https://bugs.webkit.org/show_bug.cgi?id=86041 Reviewed by Eric Seidel. * DumpRenderTree/chromium/MockWebSpeechInputController.cpp: 2012-05-09 Adam Barth [Chromium] Android needs to swizzle pixel dumps in order to get the same hash as other platforms https://bugs.webkit.org/show_bug.cgi?id=86040 Reviewed by Kent Tamura. * DumpRenderTree/chromium/TestShell.cpp: (TestShell::dumpImage): 2012-05-09 Raphael Kubo da Costa [EFL] Build fix after r116464. `ulong' is a type defined by glibc, so it is not very portable. Use `unsigned long' instead. * DumpRenderTree/efl/EventSender.cpp: (DelayedEvent::DelayedEvent): (DelayedEvent): (leapForwardCallback): 2012-05-09 Adam Barth [Chromium] LayoutTestController should use ENABLE guards for some features https://bugs.webkit.org/show_bug.cgi?id=86030 Reviewed by Eric Seidel. These features are disabled on OS(ANDROID) and so shouldn't be compiled when the feature is off. * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::LayoutTestController): * DumpRenderTree/chromium/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/chromium/TestShell.h: (TestShell): 2012-05-09 Tom Zakrajsek Fix overzealous re-opened since blocked by... message https://bugs.webkit.org/show_bug.cgi?id=86020 Reviewed by Adam Barth. Changed CreateBug behavior to only re-open a blocked bug if it was RESOLVED. Also had to update the test infrastructure slightly to allow for distinct mock commit revisions, so one could point to a RESOLVED bug for testing while another pointed to an OPEN bug. Also moved default created mock bug to be 60001 instead of 50004 which already exists as a reference mock bug. * Scripts/webkitpy/common/checkout/checkout_mock.py: (MockCheckout.__init__): (MockCheckout.commit_info_for_revision): * Scripts/webkitpy/common/net/bugzilla/bugzilla_mock.py: (MockBugzilla.create_bug): * Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py: * Scripts/webkitpy/tool/commands/download_unittest.py: (test_create_rollout_resolved): * Scripts/webkitpy/tool/commands/queries_unittest.py: (FailureReasonTest.test_blame_line_for_revision): * Scripts/webkitpy/tool/steps/createbug.py: (CreateBug.run): 2012-05-09 Hugo Parente Lima Use suitable viewport values on XHTML-MP pages. https://bugs.webkit.org/show_bug.cgi?id=85425 Reviewed by Kenneth Rohde Christiansen. Enable LEGACY_VIEWPORT_ADAPTION on Qt port. * qmake/mkspecs/features/features.prf: 2012-05-09 Jochen Eisinger When creating a new page during a navigation, prime the initial document with the correct referrer policy https://bugs.webkit.org/show_bug.cgi?id=86001 Reviewed by Adam Barth. Store the referrer policy in the request, so the network stack can comply with it. * DumpRenderTree/chromium/WebViewHost.cpp: (WebViewHost::willSendRequest): 2012-05-09 Vivek Galatage Remove CYGWIN=tty from environment variable as its no longer supported https://bugs.webkit.org/show_bug.cgi?id=85791 The CYGWIN=tty environment variable is no longer supported from Cygwin version 1.7.10. This fix avoids addition of this environment variable for Cygwin versions greater than 1.7.9. For supporting older versions i.e. 1.7.9 or older, the script takes care of retaining the environment variable. Reviewed by Adam Roben. * Scripts/webkitdirs.pm: (setupAppleWinEnv): 2012-05-09 Dirk Pranke nrwt: --skip-failing-tests should also skip flaky tests https://bugs.webkit.org/show_bug.cgi?id=85941 Reviewed by Ojan Vafai. This changes --skip-failing-tests so that we skip tests that are believed to pass sometimes as well as tests that are believed to always fail. In addition, in testing this I discovered that we had lousy test coverage of the details of retrying test failures and --clobber-old-results, so I wrote some more tests. * Scripts/webkitpy/layout_tests/controllers/manager.py: (Manager.prepare_lists_and_print_output): * Scripts/webkitpy/layout_tests/port/test.py: (unit_test_list): (TestPort.__init__): (TestDriver.run_test): * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py: (MainTest.test_skip_failing_tests): (MainTest.test_retrying_and_flaky_tests): 2012-05-09 Christophe Dumez [EFL][DRT] EFL's DRT needs to support LayoutTestController.dumpIconChanges() https://bugs.webkit.org/show_bug.cgi?id=84435 Reviewed by Antonio Gomes. Catch the "icon,changed" signal on the frames in EFL's DumpRenderTree to support LayoutTestController's dumpIconChanges(). * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::createView): (DumpRenderTreeChrome::onFrameIconChanged): (DumpRenderTreeChrome::onFrameCreated): * DumpRenderTree/efl/DumpRenderTreeChrome.h: (DumpRenderTreeChrome): 2012-05-09 Dominik Röttsches webkit-patch upload Fails Behind Load-balancing Proxies https://bugs.webkit.org/show_bug.cgi?id=85980 Reviewed by Adam Barth. Unchecking the login restriction to IP to allow the tool to function behind load-balancing proxies that use multiple outgoing IPs. * Scripts/webkitpy/common/net/bugzilla/bugzilla.py: (Bugzilla.authenticate): 2012-05-09 Hao Zheng [Chromium-android] Add unit test for Android layout test script (chromium_android.py) https://bugs.webkit.org/show_bug.cgi?id=80852 Reviewed by Adam Barth. Add some tests to ensure that chromium android port works to some extent. Some methods is hard to test, as they rely on the actual behavior on device, e.g. pushing files to device. * Scripts/webkitpy/layout_tests/port/chromium_android.py: (ChromiumAndroidPort.get_last_stacktrace): Check if dir exists. * Scripts/webkitpy/layout_tests/port/chromium_android_unittest.py: (ChromiumAndroidPortTest): (ChromiumAndroidPortTest.test_attributes): (ChromiumAndroidPortTest.test_get_last_stacktrace): (ChromiumAndroidPortTest.test_get_last_stacktrace.mock_run_command_no_dir): (ChromiumAndroidPortTest.test_get_last_stacktrace.mock_run_command_no_file): (ChromiumAndroidPortTest.test_get_last_stacktrace.mock_run_command_non_empty): (ChromiumAndroidDriverTest): (ChromiumAndroidDriverTest.setUp): (ChromiumAndroidDriverTest.test_get_drt_return_value): (ChromiumAndroidDriverTest.test_has_crash_hint): (ChromiumAndroidDriverTest.test_read_prompt): (ChromiumAndroidDriverTest.test_test_shell_command): (ChromiumAndroidDriverTest.test_write_command_and_read_line): 2012-05-09 Allan Sandfeld Jensen [Qt] ORWT fails with Qt 5 and WK2. https://bugs.webkit.org/show_bug.cgi?id=85968 Reviewed by Csaba Osztrogonác. The Qt5 binary does not have a -style argument, so avoid using it. * Scripts/old-run-webkit-tests: 2012-05-09 Mikhail Pozdnyakov [EFL][DRT] LayoutTestController addUserScript implementation https://bugs.webkit.org/show_bug.cgi?id=84718 Reviewed by Antonio Gomes. Added implementation of addUserScript() to LayoutTestController. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): clear added user scripts * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::addUserScript): added implementation 2012-05-09 Dan Bernstein Made build-webkit copy WebKitLibraries to the right location again after r116257 Reviewed by Mark Rowe. * Scripts/build-webkit: Pass the built product directory to copy-webkitlibraries-to-product-directory. * Scripts/copy-webkitlibraries-to-product-directory: Use the first argument, if present, as the product directory. 2012-05-08 Ojan Vafai Show the chromium non-layout test bots in garden-o-matic https://bugs.webkit.org/show_bug.cgi?id=85903 Reviewed by Eric Seidel. This is just a single-line high-level overview of the chromium non-layout test bots. This means that Chromium WebKit gardeners only need to keep garden-o-matic open in order to do their job. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/garden-o-matic.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui_unittests.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/styles/onebar.css: (#chromium-gtests iframe): (#chromium-gtests): (#gtest-label): 2012-05-08 Kangil Han [EFL][DRT] Implement TextInputController. https://bugs.webkit.org/show_bug.cgi?id=78559 Reviewed by Martin Robinson. TextInputController has been created for an input emulation purpose in DRT. However, not like other ports, EFL doesn't have it. Therefore, this patch will enable EFL port to have an input emulation in DRT. * DumpRenderTree/efl/CMakeLists.txt: * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::onWindowObjectCleared): * DumpRenderTree/efl/TextInputController.cpp: Added. (setMarkedTextCallback): (hasMarkedTextCallback): (markedRangeCallback): (insertTextCallback): (unmarkTextCallback): (firstRectForCharacterRangeCallback): (selectedRangeCallback): (makeTextInputController): * DumpRenderTree/efl/TextInputController.h: Added. 2012-04-23 Raphael Kubo da Costa [EFL] Pass a WTF::String directly from DRT to ewebkit when possible https://bugs.webkit.org/show_bug.cgi?id=84634 Reviewed by Gustavo Noronha Silva. Drop unnecessary const char* -> WTF::String conversions by creating String objects directly in LayoutTestController out of JSStringRef's and passing them to DumpRenderTreeSupportEfl. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::queueLoad): (LayoutTestController::setValueForUser): (LayoutTestController::findString): (LayoutTestController::addUserStyleSheet): 2012-05-08 Alexander Shalamov [EFL][DRT] EventSender needs implementation of leapForward function https://bugs.webkit.org/show_bug.cgi?id=83486 Reviewed by Eric Seidel. Implemented leapForward function in Efl's DRT. This function is required to delay and queue event execution and simulate real user behavior. * DumpRenderTree/efl/EventSender.cpp: (DelayedEvent): (MouseEventInfo::MouseEventInfo): (MouseEventInfo): (sendClick): (mouseDownCallback): (mouseUpCallback): (mouseMoveToCallback): (leapForwardCallback): (mouseScrollByCallback): (feedOrQueueMouseEvent): (feedMouseEvent): (feedQueuedMouseEvents): 2012-05-08 Sudarsana Nagineni [GTK] Memory leak in DRT EventSender.cpp https://bugs.webkit.org/show_bug.cgi?id=85886 Reviewed by Martin Robinson. Fix a memory leak in gdkModifersFromJSValue() by using JSRetainPtr for JSString created with JSStringCreateWithUTF8CString. * DumpRenderTree/gtk/EventSender.cpp: (gdkModifersFromJSValue): 2012-05-08 Christophe Dumez [EFL] glib dependency needs to be updated in jhbuild https://bugs.webkit.org/show_bug.cgi?id=85878 Reviewed by Martin Robinson. Bump glib version from 2.31.8 to 2.32.2 in jhbuild.modules. This addresses flakiness in media tests due to possible deadlock with the following output on stderr: GStreamer-WARNING **: wrong STREAM_LOCK count 0 The bug was fixed upstream by: http://git.gnome.org/browse/glib/commit/?id=265f265c677bc5399dbb1cfa449cb4b915a6459f * efl/jhbuild.modules: 2012-05-07 Alexis Menard [Qt] Unbreak debugging of WebKit2. https://bugs.webkit.org/show_bug.cgi?id=85839 Reviewed by Simon Hausmann. When you attach GDB to a running process, it stops it. http://trac.webkit.org/changeset/115958 introduced a pause() call to wait the debugger to be attached to then continue the execution of the WebProcess. Unfortunately the pause() function does not return unless a signal handler is called. This patch introduce an event handler to exit from the paused state when the debugger send the signal SIGCONT. The old code works with older version of GDB (<7.0) but not with newer versions where the behavior of pause() is correct. * WebKitTestRunner/qt/main.cpp: (sigcontHandler): (main): 2012-05-08 Christophe Dumez [EFL][DRT] EFL's DRT does not fully support page visibility https://bugs.webkit.org/show_bug.cgi?id=85032 Reviewed by Antonio Gomes. * DumpRenderTree/LayoutTestController.cpp: (resetPageVisibilityCallback): Fix incorrect argumentCount check. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): Properly reset page visibility setting between tests. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::setPageVisibility): Add support for "preview" visibility state. 2012-05-08 Gavin Peters [webkit-patch] Return meaningful results from single-result queries. https://bugs.webkit.org/show_bug.cgi?id=85367 Reviewed by Dirk Pranke. Sometimes a quicksearch returns a single result by just redirecting to the page for that single bug. This patch hacks around that by noticing the redirect target is a single page, and constructing a result list consisting of that one bug. * Scripts/webkitpy/common/net/bugzilla/bugzilla.py: (BugzillaQueries._fetch_bugs_from_advanced_query): 2012-05-08 Gavin Peters [webkit-patch] Add new chrome-channels command to track down current chrome release channels for a committed bug. https://bugs.webkit.org/show_bug.cgi?id=85368 Reviewed by Eric Seidel. * Scripts/webkitpy/common/config/urls.py: * Scripts/webkitpy/common/net/bugzilla/bugzilla_mock.py: (MockBugzillaQueries.fetch_bugs_matching_quicksearch): * Scripts/webkitpy/common/net/omahaproxy.py: Added. (OmahaProxy): (OmahaProxy.__init__): (OmahaProxy.set_url): (OmahaProxy._json_url): (OmahaProxy._get_json): (OmahaProxy.get_revisions): * Scripts/webkitpy/common/net/omahaproxy_unittest.py: Added. (MockOmahaProxy): (MockOmahaProxy.__init__): (OmahaProxyTest): (OmahaProxyTest.test_get_revisions): * Scripts/webkitpy/tool/commands/__init__.py: * Scripts/webkitpy/tool/commands/bugsearch.py: (BugSearch): * Scripts/webkitpy/tool/commands/chromechannels.py: Added. (ChromeChannels): (__init__): (execute): * Scripts/webkitpy/tool/commands/chromechannels_unittest.py: Added. (MockOmahaProxy): (MockOmahaProxy.get_revisions): (TestableChromeChannels): (TestableChromeChannels.__init__): (ChromeChannelsTest): (ChromeChannelsTest.test_single_bug): (ChromeChannelsTest.test_with_query): 2012-05-08 Carlos Garcia Campos Unreviewed. Fix GTK+ unit tests. * gtk/common.py: (pkg_config_file_variable): 2012-05-08 Carlos Garcia Campos [GTK] Google tests that time out are leaked https://bugs.webkit.org/show_bug.cgi?id=84973 Reviewed by Philippe Normand. Move Tools/Scripts/run-gtk-tests to Tools/gtk/run-api-tests leaving run-gtk-tests as simple wrapper that runs run-api-tests in a jhbuild environment when jhbuild is available. * Scripts/run-gtk-tests: Run Tools/gtk/run-api-tests directly or with jhbuild if present. * gtk/common.py: (pkg_config_file_variable): Generic function to get a variable from a pkg-config file. (prefix_of_pkg_config_file): Use pkg_config_file_variable with 'prefix'. * gtk/run-api-tests: Copied from Tools/Scripts/run-gtk-tests. (TestRunner.__init__): Use common.build_path() to get the programs path. (TestRunner._lookup_atspi2_binary): Use common.pkg_config_file_variable() to get the 'exec_prefix'. (TestRunner._start_accessibility_daemons): Use subprocess.Popen() directly instead of helper _create_process that has been removed. (TestRunner._setup_testing_environment): Use subprocess.Popen() directly instead of helper _create_process that has been removed. Use common.top_level_path() and common.build_path() instead of custom methods that have been removed too. (TestRunner._run_test_command): Use subprocess.Popen() directly instead of helper _create_process that has been removed. 2012-05-08 Thiago Marcos P. Santos [EFL] [DRT] Clear extra views before setting a blank page https://bugs.webkit.org/show_bug.cgi?id=85608 Reviewed by Kenneth Rohde Christiansen. We need to clear the extra views before setting a blank page, otherwise it will cause more events interactions and sometimes trigger a unload(). This was causing frame-crash-with-page-cache.html to leak events to the next test causing any subsequent test to fail. * DumpRenderTree/efl/DumpRenderTree.cpp: (runTest): * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::clearExtraViews): (DumpRenderTreeChrome::resetDefaultsToConsistentValues): * DumpRenderTree/efl/DumpRenderTreeChrome.h: (DumpRenderTreeChrome): 2012-05-08 Balazs Kelemen [Qt] X11 plugins need to be reworked for Qt5+WK1 https://bugs.webkit.org/show_bug.cgi?id=80691 Reviewed by Simon Hausmann. Implement basic windowless plugin support with Qt5. * qmake/mkspecs/features/features.prf: Enable NPAPI plugins if X11 libraries are available. * qmake/mkspecs/features/functions.prf: Added a convenience function to determine availability of X11 libraries. 2012-05-07 Raphael Kubo da Costa [webkitpy] Remove Python 2.5-specific workaround from http_server_base.py. https://bugs.webkit.org/show_bug.cgi?id=85852 Reviewed by Dirk Pranke. * Scripts/webkitpy/layout_tests/servers/http_server_base.py: (HttpServerBase._is_server_running_on_all_ports): Revert the part of r89412 which was a workaround for an issue with Python 2.5 on Mac 10.5. 2012-05-07 Raphael Kubo da Costa webkitdirs.pm: Make determineNumberOfCPUs work on FreeBSD. https://bugs.webkit.org/show_bug.cgi?id=81481 Reviewed by Dirk Pranke. FreeBSD has the same sysctl interface present in Darwin to query the number of CPUs available, so use that too. * Scripts/webkitdirs.pm: (determineNumberOfCPUs): Share the sysctl call with Darwin. (isFreeBSD): Added. 2012-05-07 Alexis Menard Unreviewed Qt build fix in Debug. We need some system includes here for getpid() to be recognized. * WebKitTestRunner/qt/main.cpp: 2012-05-07 Ojan Vafai When embedding the flakiness dashboard, hide popups when the frame is blurred. https://bugs.webkit.org/show_bug.cgi?id=85684 Reviewed by Dirk Pranke. If you show a popup, then click elsewhere in garden-o-matic, it will now properly hide it. * TestResultServer/static-dashboards/dashboard_base.js: (showPopup): * TestResultServer/static-dashboards/flakiness_dashboard.js: (showPopupForBuild): (generatePageForIndividualTests.if): (generatePageForIndividualTests): * TestResultServer/static-dashboards/flakiness_dashboard_embedded_unittests.js: Added. * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js: * TestResultServer/static-dashboards/run-embedded-unittests.html: Added. 2012-05-07 Ian Vollick [chromium] WebViewHost should setCompositorSurfaceReady https://bugs.webkit.org/show_bug.cgi?id=85418 Reviewed by James Robinson. * DumpRenderTree/chromium/WebViewHost.cpp: (WebViewHost::setWebWidget): 2012-05-07 Jon Lee WebKitTestRunner needs layoutTestController.dumpFrameLoadCallbacks https://bugs.webkit.org/show_bug.cgi?id=42331 Reviewed by Darin Adler. This patch sets up the framework for supporting dumpFrameLoadCallbacks, as well as outputting some of the required strings. As tests become unskipped the other callbacks can be supported. * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl: Add dumpFrameLoadCallbacks() * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp: (WTR::InjectedBundle::didReceiveMessage): pass in the dictionary containing keys needed to initialize dump callbacks for the layoutTestController. (WTR::InjectedBundle::booleanForKey): Helper function to extract boolean value from key in WKDictionaryRef. Output some warning if we are trying to extract a value that is not a boolean. (WTR::InjectedBundle::beginTesting): Update to take in the dictionary passed in by the TestController. We cannot set the bits on the layoutTestController until it has been created in beginTesting(). * WebKitTestRunner/InjectedBundle/InjectedBundle.h: * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp: (WTR::InjectedBundle::didReceiveMessage): (WTR): (WTR::InjectedBundle::beginTesting): * WebKitTestRunner/InjectedBundle/InjectedBundle.h: (InjectedBundle): * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp: (WTR): (WTR::dumpFrameDescriptionSuitableForTestResult): Static function similar to -[WebFrame _drt_descriptionSuitableForTestResult] in DumpRenderTree. (WTR::InjectedBundlePage::didStartProvisionalLoadForFrame): Output string. (WTR::InjectedBundlePage::didCommitLoadForFrame): Output string. (WTR::InjectedBundlePage::didFinishLoadForFrame): Output string. (WTR::InjectedBundlePage::didReceiveTitleForFrame): Output string. (WTR::InjectedBundlePage::didCancelClientRedirectForFrame): Output string. (WTR::InjectedBundlePage::willPerformClientRedirectForFrame): Output string. In WK1 it's "willPerformClientRedirectToURL" so we will have to use that value here. (WTR::InjectedBundlePage::didFinishDocumentLoadForFrame): Output string. (WTR::InjectedBundlePage::didHandleOnloadEventsForFrame): Output string. * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: (WTR::LayoutTestController::LayoutTestController): Initialize to not dump frame load callbacks. * WebKitTestRunner/InjectedBundle/LayoutTestController.h: (WTR::LayoutTestController::dumpFrameLoadCallbacks): Change to use setShouldDumpFrameLoadCallbacks. (WTR::LayoutTestController::setShouldDumpFrameLoadCallbacks): Set bit. (WTR::LayoutTestController::shouldDumpFrameLoadCallbacks): Return bit. (LayoutTestController): Added m_dumpFrameLoadCallbacks to track whether to dump the output. * WebKitTestRunner/TestInvocation.cpp: (WTR::shouldLogFrameLoadDelegates): Similar to DRT. (WTR::TestInvocation::invoke): Based on the URL loaded, set the bit in the dictionary that gets passed to the injected bundle. 2012-05-07 Tommy Widenflycht [chromium] MediaStream API: Enhance WebUserMediaClientMock and add a test for it https://bugs.webkit.org/show_bug.cgi?id=85618 Reviewed by Adam Barth. Adding the capability to control which MediaStreamTracks actually get created, and adding a test that exercises the JS api using that capability. * DumpRenderTree/chromium/WebUserMediaClientMock.cpp: (WebKit::WebUserMediaClientMock::requestUserMedia): * DumpRenderTree/chromium/WebUserMediaClientMock.h: * DumpRenderTree/chromium/WebViewHost.cpp: (WebViewHost::testMediaStreamClient): * DumpRenderTree/chromium/WebViewHost.h: (WebViewHost): 2012-05-07 Zan Dobersek [Gtk] WebGL feature is not built anymore through build-webkit after r116251 https://bugs.webkit.org/show_bug.cgi?id=85788 Reviewed by Martin Robinson. Enable WebGL by default for the Gtk port. The feature is already enabled by default in configure.ac, so there's no reason it shouldn't be enabled when building through build-webkit. * Scripts/webkitperl/FeatureList.pm: 2012-05-07 Csaba Osztrogonác Unittest fix after r116288, because now there are more than 5 hit for the 'Eric' word. We have only 2 'Gavin', so it is a better choice than 'Eric' for unit testing. Reviewed by Tor Arne Vestbø. * Scripts/webkitpy/tool/bot/irc_command_unittest.py: (IRCCommandTest.test_whois): 2012-05-07 Sheriff Bot Unreviewed, rolling out r116299, r116301, and r116303. http://trac.webkit.org/changeset/116299 http://trac.webkit.org/changeset/116301 http://trac.webkit.org/changeset/116303 https://bugs.webkit.org/show_bug.cgi?id=85795 Build is still broken (Requested by Ossy on #webkit). * DumpRenderTree/qt/DumpRenderTree.pro: * DumpRenderTree/qt/main.cpp: (main): * QtTestPlatformPlugin/QtTestPlatformPlugin.pro: Removed. * QtTestPlatformPlugin/TestIntegration.cpp: Removed. * QtTestPlatformPlugin/TestIntegration.h: Removed. * QtTestPlatformPlugin/mac/TestFontDatabase.h: Removed. * QtTestPlatformPlugin/mac/TestFontDatabase.mm: Removed. * QtTestPlatformPlugin/mac/TestIntegrationMac.mm: Removed. * QtTestPlatformPlugin/main.cpp: Removed. * QtTestPlatformPlugin/testplatform.json: Removed. * Tools.pro: * WebKitTestRunner/Target.pri: * WebKitTestRunner/qt/main.cpp: (main): 2012-05-07 Balazs Kelemen Unreviewed Qt build fix. One more try to learn how to use const with a pointer array. * DumpRenderTree/qt/main.cpp: (initializeTestPlatformPlugin): 2012-05-07 Balazs Kelemen Unreviewed build fix after r116299. * DumpRenderTree/qt/main.cpp: (initializeTestPlatformPlugin): 2012-05-07 Balazs Kelemen [Qt] Add test specific platform plugin to achieve unified layout test results https://bugs.webkit.org/show_bug.cgi?id=80996 Reviewed by Simon Hausmann. Added QtTestPlatformPlugin as a new project under Tools. This is a Qt5-ish platform plugin that can be used to tweak the platform support interfaces in order to unify layout test results. For now it only overrides the font database on Mac and redirects everything else to the real platform plugin. The font database it provides mimics the way how we set up test fonts with fontconfig on Linux. Make DumpRenderTree and WebKitTestRunner use this platform plugin. * DumpRenderTree/qt/DumpRenderTree.pro: * DumpRenderTree/qt/main.cpp: (initializeTestPlatformPlugin): (main): * QtTestPlatformPlugin/QtTestPlatformPlugin.pro: Added. * QtTestPlatformPlugin/TestIntegration.cpp: Added. (TestIntegration::TestIntegration): (TestIntegration::fontDatabase): * QtTestPlatformPlugin/TestIntegration.h: Added. (TestIntegration): (TestIntegration::hasCapability): (TestIntegration::createPlatformPixmap): (TestIntegration::createPlatformWindow): (TestIntegration::createPlatformBackingStore): (TestIntegration::createPlatformOpenGLContext): (TestIntegration::createPlatformSharedGraphicsCache): (TestIntegration::guiThreadEventDispatcher): (TestIntegration::clipboard): (TestIntegration::drag): (TestIntegration::inputContext): (TestIntegration::accessibility): (TestIntegration::nativeInterface): (TestIntegration::services): (TestIntegration::styleHint): (TestIntegration::platformTheme): * QtTestPlatformPlugin/mac/TestFontDatabase.h: Added. (TestFontDatabase): * QtTestPlatformPlugin/mac/TestFontDatabase.mm: Added. (TestFontDatabase::populateFontDatabase): * QtTestPlatformPlugin/mac/TestIntegrationMac.mm: Added. (TestIntegration::fontDatabase): * QtTestPlatformPlugin/main.cpp: Added. (TestIntegrationPlugin::keys): (TestIntegrationPlugin::create): (TestIntegrationPlugin::initialize): * QtTestPlatformPlugin/testplatform.json: Added. * Tools.pro: * WebKitTestRunner/Target.pri: * WebKitTestRunner/qt/main.cpp: (main): 2012-05-07 Christophe Dumez [EFL] media/video-poster-blocked-by-willsendrequest.html is flaky https://bugs.webkit.org/show_bug.cgi?id=85609 Reviewed by Simon Hausmann. Flakiness was caused by the fact that media/crash-closing-page-with-media-as-plugin-fallback.html and media/video-poster-blocked-by-willsendrequest.html use the same PNG file as poster attribute in the video tag. The PNG file gets cached when the first test is executed and this somehow causes the second test to fail. The flakiness was addressed by clearing the memory and page caches between tests. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): 2012-05-07 Eric Seidel Add values for all features to Qt's features.pri https://bugs.webkit.org/show_bug.cgi?id=85746 Reviewed by Tor Arne Vestbø. When generate-feature-files generates features.pri it provides a value for every possible ENABLE_. In prepration for landing generate-feature-files it seems appropriate to add a value for all possible ENABLEs to the existing features.pri. This lets me make sure that I'm getting them all right! * qmake/mkspecs/features/features.pri: 2012-05-07 Adam Bergkvist Added myself as a committer. Not reviewed. * Scripts/webkitpy/common/config/committers.py: 2012-05-07 Sudarsana Nagineni [EFL][DRT] Implement LayoutTestController::execCommand https://bugs.webkit.org/show_bug.cgi?id=82286 Reviewed by Antonio Gomes. Add missing implementation execCommand and isCommandEnabled to EFL's LayoutTestController. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::execCommand): (LayoutTestController::isCommandEnabled): 2012-05-06 Hayato Ito [prepare-ChangeLog] Excludes a reference file used by reftests from test files. https://bugs.webkit.org/show_bug.cgi?id=74268 Reviewed by Ryosuke Niwa. * Scripts/prepare-ChangeLog: (generateFileList): 2012-05-06 Mary Wu Add a contributor to committers.py https://bugs.webkit.org/show_bug.cgi?id=85761 Reviewed by Unreviewed. Add myself to committers.py contributor. * Scripts/webkitpy/common/config/committers.py: 2012-05-06 Dan Bernstein Tools part of: Building and debugging WebKit in the Xcode IDE requires a lot of setup https://bugs.webkit.org/show_bug.cgi?id=85739 Reviewed by Daniel Bates. * Scripts/build-webkit: Moved code that copies from WebKitLibraries to the product directory from here... * Scripts/copy-webkitlibraries-to-product-directory: ...to this new script. 2012-05-06 Andy Estes Fix the build after r116246. * DumpRenderTree/LayoutTestController.cpp: Remove an unused function. 2012-05-06 Eric Seidel Update several build-webkit options to match the #defines they toggle https://bugs.webkit.org/show_bug.cgi?id=85744 Reviewed by Adam Barth. Most notable here is fixing --3d-canvas to be --webgl these days -- the define was renamed 15 months ago! * Scripts/webkitperl/FeatureList.pm: 2012-05-06 Mikhail Pozdnyakov [EFL] EFL's LayoutTestController setJavaScriptCanAccessClipboard implementation https://bugs.webkit.org/show_bug.cgi?id=83687 Reviewed by Antonio Gomes. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::setJavaScriptCanAccessClipboard): 2012-05-06 Gyuyoung Kim Convert isPageBoxVisible to use Internals interface. https://bugs.webkit.org/show_bug.cgi?id=85692 Reviewed by Darin Adler. Remove isPageBoxVisible functions, because it is able to work in the cross-port way through the Internals interface. * DumpRenderTree/LayoutTestController.cpp: (LayoutTestController::staticFunctions): * DumpRenderTree/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp: * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::LayoutTestController): * DumpRenderTree/chromium/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: * DumpRenderTree/mac/LayoutTestControllerMac.mm: * DumpRenderTree/qt/LayoutTestControllerQt.cpp: * DumpRenderTree/qt/LayoutTestControllerQt.h: (LayoutTestController): * DumpRenderTree/win/LayoutTestControllerWin.cpp: * DumpRenderTree/wx/LayoutTestControllerWx.cpp: 2012-05-06 Tom Zakrajsek sheriffbot isn't reopening patches after it lands rollouts https://bugs.webkit.org/show_bug.cgi?id=64418 Reviewed by Adam Barth. * Scripts/webkitpy/tool/commands/download_unittest.py: * Scripts/webkitpy/tool/steps/createbug.py: (CreateBug.run): 2012-05-04 Raphael Kubo da Costa webkitpy: Use os.pathsep instead of manually finding out the path separator in NRWT. https://bugs.webkit.org/show_bug.cgi?id=85697 Reviewed by Eric Seidel. * Scripts/new-run-webkit-tests: Instead of manually detecting whether the path separator should be ':' or ';' by checking sys.platform, rely on os.pathsep which has exactly the same information (os.pathsep is ':' for Cygwin, from what I could check in Python's sources). 2012-05-05 Sheriff Bot Unreviewed, rolling out r116223. http://trac.webkit.org/changeset/116223 https://bugs.webkit.org/show_bug.cgi?id=85718 The bots need to set XDG_DATA_DIRS (Requested by rakuco on #webkit). * efl/jhbuildrc: * gtk/jhbuildrc: 2012-05-05 Raphael Kubo da Costa [jhbuild] Do not set XDG_DATA_DIRS in jhbuildrc. https://bugs.webkit.org/show_bug.cgi?id=85717 Reviewed by Martin Robinson. After r116209, the XDG_DATA_DIRS environment variable is now passed by webkitpy's layout_tests code to run-with-jhbuild, so instead of setting a few defaults in jhbuildrc we can now rely on the values set and really used by each system. * efl/jhbuildrc: * gtk/jhbuildrc: 2012-05-05 Tor Arne Vestbø [Qt] Buildfix for --minimal build after r116107. Reviewed by Csaba Osztrogonác. * qmake/mkspecs/features/features.prf: 2012-05-05 Zan Dobersek [Gtk] GtkDriver is not required anymore after r116134 https://bugs.webkit.org/show_bug.cgi?id=85699 Reviewed by Csaba Osztrogonác. Remove GtkDriver as it was replaced by the XvfbDriver and is no longer needed. * Scripts/webkitpy/layout_tests/port/gtk.py: 2012-05-05 Zan Dobersek Increase the display ID for the Xvfb process to use if pixel_tests argument is true rather than if pixel_tests option is present. This fixes the XvfbDriver in the same way r115825 fixed the now-redundant GtkDriver. Rubber-stamped by Csaba Osztrogonác. * Scripts/webkitpy/layout_tests/port/xvfbdriver.py: (XvfbDriver._start): 2012-05-04 Raphael Kubo da Costa webkitpy: Preserve XDG_DATA_DIRS in the environment passed to ServerProcess. https://bugs.webkit.org/show_bug.cgi?id=85694 Reviewed by Martin Robinson. This is sort of a follow-up to r100674: $XDG_DATA_DIRS should be preserved, as it is particularly useful for Linux/Unix environments. As a real use case, if shared-mime-info is installed into a non-trivial prefix, Glib and other libraries will have trouble to properly identify mimetypes (in ResourceResponses, for example) as the base freedesktop.org mimetype file will not be found. * Scripts/webkitpy/layout_tests/port/base.py: (Port.to.setup_environ_for_server): 2012-05-04 Gustavo Noronha Silva [GTK] Simplify how libWebCoreModules is linked in, and fix WebKit2 build https://bugs.webkit.org/show_bug.cgi?id=85691 * GNUmakefile.am: no longer link libWebCoreModules to DumpRenderTree. 2012-05-04 Eric Seidel Make capitalization and descriptions in FeatureList.pm more consistent to make easier to autogenerate https://bugs.webkit.org/show_bug.cgi?id=85583 Reviewed by Daniel Bates. * Scripts/webkitperl/FeatureList.pm: 2012-05-04 Eric Seidel Sort entries in FeaturesList.pm to make them easier to autogenerate https://bugs.webkit.org/show_bug.cgi?id=85584 Reviewed by Daniel Bates. No functional changes, only sorting. * Scripts/webkitperl/FeatureList.pm: 2012-05-04 Andy Estes REGRESSION (r115752): WebKit2.NewFirstVisuallyNonEmptyLayout API test times out https://bugs.webkit.org/show_bug.cgi?id=85675 * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayout.cpp: (TestWebKitAPI::TEST): Disable the test. 2012-04-30 Jer Noble Full screen will exit during a provisional load of a non-ancestor iframe. https://bugs.webkit.org/show_bug.cgi?id=85230 Reviewed by Maciej Stachowiak . Add support for the new UIDelegate webView:closeFullScreenWithListener: method: * DumpRenderTree/mac/UIDelegate.mm: (-[UIDelegate webView:closeFullScreenWithListener:]): Add support for the new closeFullScreen injected bundle method: * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp: (WTR::InjectedBundlePage::InjectedBundlePage): (WTR::InjectedBundlePage::closeFullScreen): * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h: 2012-05-04 Ojan Vafai Fix test naming to be camel-case like the methods they are testing. * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js: (test): 2012-05-04 Ojan Vafai Show the results links in the embedded flakiness dashboard https://bugs.webkit.org/show_bug.cgi?id=85660 Reviewed by Adam Barth. Turns out these links don't add much visual noise and it's annoying to have to go to the dashboard from garden-o-matic just to view the results for a test. This is useful for cases where garden-o-matic doesn't properly show the results (e.g. missing tests). If we address all those cases, then we could easily hide the results links again if we wanted to. * TestResultServer/static-dashboards/flakiness_dashboard.js: (showUpdateInfoForTest): (htmlForIndividualTestOnAllBuildersWithResultsLinks): * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js: (test): 2012-05-04 Dirk Pranke Revert to using chromium --test-shell mode on SL in NRWT https://bugs.webkit.org/show_bug.cgi?id=83076 Unreviewed, build fix. We seem to be seeing a lot more flakiness on the bot since I switched NRWT to "DRT" mode. We'll try reverting back and see if this helps things. * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumDriver.__init__): 2012-05-04 Christophe Dumez [EFL] Implement layoutTestController.dumpResourceResponseMIMETypes https://bugs.webkit.org/show_bug.cgi?id=84941 Reviewed by Martin Robinson. Print MIME type detection messages in EFL's DumpRenderTree if LayoutTestController's dumpResourceResponseMIMETypes() returns true. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::onResponseReceived): 2012-05-04 Sudarsana Nagineni [GTK] DRT needs an implementation of LayoutTestController::setDefersLoading and ::goBack https://bugs.webkit.org/show_bug.cgi?id=85134 Reviewed by Martin Robinson. Add missing implementation setDefersLoading and goBack to GTK's LayoutTestController so that the test case rely on these methods can be enabled. * DumpRenderTree/gtk/DumpRenderTree.cpp: (resetDefaultsToConsistentValues): * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: (LayoutTestController::goBack): (LayoutTestController::setDefersLoading): 2012-05-04 Kristóf Kosztyó [Qt][NRWT] Run each DRT in it's own xvfb https://bugs.webkit.org/show_bug.cgi?id=77335 Reviewed by Dirk Pranke. * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkPort._driver_class): * Scripts/webkitpy/layout_tests/port/qt.py: (QtPort._driver_class): * Scripts/webkitpy/layout_tests/port/xvfbdriver.py: Added. (XvfbDriver): (XvfbDriver._start): (XvfbDriver._start.x_filter): (XvfbDriver.stop): 2012-05-04 Zan Dobersek Use destructuring assingment for the return values of the port.diff_image() method call. Rubber-stamped by Dirk Pranke. * Scripts/webkitpy/layout_tests/controllers/test_result_writer.py: (write_test_result): 2012-05-03 Ojan Vafai Store fewer runs by default in the flakiness dashboard https://bugs.webkit.org/show_bug.cgi?id=85568 Reviewed by Dirk Pranke. This will greatly shrink the size of the json files we load, resulting in a much faster initial load. People rarely care about more than 100 runs. For the cases they do, they can still click the show all runs checkbox to show 500 runs. * TestResultServer/model/jsonresults.py: 2012-05-04 Christophe Dumez [EFL] Frame load callbacks output is missing for redirections https://bugs.webkit.org/show_bug.cgi?id=85173 Reviewed by Antonio Gomes. Catch the new ewk_frame signals to print the expected output in case of redirections, when LayoutTestController's dumpFrameLoadCallbacks() returns true. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::createView): (DumpRenderTreeChrome::onWillSendRequest): (DumpRenderTreeChrome::onFrameCreated): (DumpRenderTreeChrome::onFrameRedirectForProvisionalLoad): (DumpRenderTreeChrome::onFrameRedirectRequested): (DumpRenderTreeChrome::onResponseReceived): * DumpRenderTree/efl/DumpRenderTreeChrome.h: (DumpRenderTreeChrome): 2012-05-04 Tor Arne Vestbø [Qt] Fix up warning about missing GLib/Gio/GStreamer We don't actually fall back. Rubber-stamped by Csaba Osztrogonác. * qmake/mkspecs/features/features.prf: 2012-05-04 Tor Arne Vestbø [Qt] Clarify warning about missing GLib/Gio/GStreamer for media support And only print it once. Reviewed by Csaba Osztrogonác. * qmake/mkspecs/features/features.prf: 2012-05-04 Tor Arne Vestbø [Qt] Clean up and split features.prf into a static list of defaults The static list of feature defaults is used as a fallback for any feature that's not dynamically detected or overriden on the command line (though build-webkit or passing DEFINES+= to qmake). The static list is complete, which allows for auto-generation based on Features.py (see bug https://bugs.webkit.org/show_bug.cgi?id=85456) https://bugs.webkit.org/show_bug.cgi?id=85611 Reviewed by Simon Hausmann. * Scripts/build-webkit: * Tools.pro: * qmake/mkspecs/features/features.prf: * qmake/mkspecs/features/features.pri: Added. 2012-05-04 Sheriff Bot Unreviewed, rolling out r116085, r116091, and r116095. http://trac.webkit.org/changeset/116085 http://trac.webkit.org/changeset/116091 http://trac.webkit.org/changeset/116095 https://bugs.webkit.org/show_bug.cgi?id=85628 We are not ready with dependencies on all platform yet (mac) + problems with debug builds. (Requested by Zoltan on #webkit). * qmake/mkspecs/features/features.prf: 2012-05-04 Christophe Dumez [EFL] LayoutTestController does not support overriding WebKitLoadSiteIconsKey https://bugs.webkit.org/show_bug.cgi?id=85171 Reviewed by Andreas Kling. Implement support for overriding "WebKitLoadSiteIconsKey" setting in EFL's LayoutTestController. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::overridePreference): 2012-05-03 Alexis Menard [Qt] Enable fullscreen API for WebKit2. https://bugs.webkit.org/show_bug.cgi?id=85498 Reviewed by Simon Hausmann. Enable the fullscreen API on the MiniBrowser and turn it default on WebKit2, disable it on WK1. * MiniBrowser/qt/BrowserWindow.cpp: (BrowserWindow::BrowserWindow): * MiniBrowser/qt/qml/BrowserWindow.qml: * Scripts/webkitperl/FeatureList.pm: * qmake/mkspecs/features/features.prf: 2012-05-04 Allan Sandfeld Jensen REGRESSION(r116008) old-run-webkit-tests broken https://bugs.webkit.org/show_bug.cgi?id=85597 Reviewed by Simon Hausmann. Reinstate libraryContainsSymbol. * Scripts/webkitperl/features.pm: (libraryContainsSymbol): 2012-05-04 Zoltan Horvath [Qt] Set WebCore imagedecoders as default and add fallback to QImageDecoder https://bugs.webkit.org/show_bug.cgi?id=80400 Remove WTF_USE_QT_IMAGE_DECODER macro. Reviewed by Simon Hausmann. * qmake/mkspecs/features/features.prf: 2012-05-03 Christophe Dumez [EFL] DRT should obey layoutTestController's addURLToRedirect() https://bugs.webkit.org/show_bug.cgi?id=82722 Reviewed by Maciej Stachowiak. Implement support for layoutTestController's addURLToRedirect() in EFL's DRT so that it redirects properly when indicated. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::onWillSendRequest): 2012-05-03 Christophe Dumez [EFL] DRT should support LayoutTestController's willSendRequestReturnsNullOnRedirect() https://bugs.webkit.org/show_bug.cgi?id=82704 Reviewed by Antonio Gomes. EFL's DRT needs to obey LayoutTestController's willSendRequestReturnsNullOnRedirect(). * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::onWillSendRequest): 2012-05-03 Raphael Kubo da Costa git-add-reviewer: Do not put the bug title and its URL in the same line in the commit message. https://bugs.webkit.org/show_bug.cgi?id=85570 Reviewed by Adam Roben. The standard format of commit messages and ChangeLog entries has a short summary in the first line, followed by the bug URL in Bugzilla. Passing "%s" in the pretty format line to git-rev-list merges these two lines into a single one separated by a space character, which is not what we want. Since "%s%n%n%b" is in practice "%B" (raw message), use it instead, as the first two lines do not end up being merged together. * Scripts/git-add-reviewer: (writeCommitMessageToFile): 2012-05-03 Ojan Vafai Improve UI for garden-o-matic examine view when there are fewer than 5 tests. https://bugs.webkit.org/show_bug.cgi?id=85566 Reviewed by Adam Barth. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results.js: (.): * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results_unittests.js: (.): * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/styles/results.css: 2012-05-03 Raphael Kubo da Costa webkitpy: Recognize FreeBSD as a valid platform. https://bugs.webkit.org/show_bug.cgi?id=81467 Reviewed by Adam Barth. Add PlatformInfo.is_freebsd() so that trying to use webkit-patch on FreeBSD does not raise an "unrecognized platform" exception. Most of this patch contains some unit tests for the mentioned method, as currently platform-specific code is needed only on Windows. * Scripts/webkitpy/common/system/platforminfo.py: (PlatformInfo.__init__): (PlatformInfo.is_freebsd): (PlatformInfo._determine_os_name): * Scripts/webkitpy/common/system/platforminfo_mock.py: (MockPlatformInfo.is_freebsd): * Scripts/webkitpy/common/system/platforminfo_unittest.py: (fake_platform): (fake_platform.FakePlatformModule.release): (TestPlatformInfo.test_real_code): (TestPlatformInfo.test_os_name_and_wrappers): (TestPlatformInfo.test_os_version): (TestPlatformInfo.test_display_name): (TestPlatformInfo.test_total_bytes_memory): (TestPlatformInfo.test_free_bytes_memory): * Scripts/webkitpy/layout_tests/port/factory.py: (PortFactory._default_port): 2012-05-03 Ojan Vafai Make the table headers white-space:nowrap. This decreases the vertical size, which makes for more minimal embedding in garden-o-matic. * TestResultServer/static-dashboards/flakiness_dashboard_tests.css: (.table-header-content *): 2012-05-03 Eric Seidel Split build-webkit's feature option list off into its own module in preparation for autogeneration https://bugs.webkit.org/show_bug.cgi?id=85548 Reviewed by Daniel Bates. My Perl-fu is very weak. This moves the feature option list off into a separate module file so that I can easily autogenerate just that file with the fancy new generate-feature-files command. It's a bit odd to have an array which we return out to build-webkit to modify, but it turns out not to matter. Also, build-webkit doesn't actually use any of the variables we're using to store the results from the option parse, so those end up inaccessible from build-webkit (except by reference) which is actually cleaner. Callers can't assume that they would get a new array every time, but someone with more Perl-fu than I could create something fancier than this if we need in the future. * Scripts/build-webkit: * Scripts/webkitperl/FeatureList.pm: Added. (getFeatureOptionList): 2012-05-03 Dirk Pranke REGRESSION: run-webkit-tests failing on Chromium Mac https://bugs.webkit.org/show_bug.cgi?id=85459 Unreviewed, build fix. Turns out select doesn't throw IOErrors, it has it's own kind of error :(. Trap that instead. * Scripts/webkitpy/layout_tests/port/server_process.py: (ServerProcess._wait_for_data_and_update_buffers_using_select): 2012-05-03 Raphael Kubo da Costa [CMake] Rewrite FindCairo.cmake. https://bugs.webkit.org/show_bug.cgi?id=84895 Reviewed by Daniel Bates. The old approach relied on pkg-config for finding Cairo (which introduced a dependency on pkg-config that could be avoided), used the LibFindMacros code that we should probably remove in the future and did not use the FindPackageHandleStandardArguments module. Change all that by rewriting the module. - Use the pkg-config output optionally instead of requiring it like LibFindMacros did. - Remove the implicit dependency on FreeType which often found it the wrong way via pkg-config and without considering CMAKE_PREFIX_PATH. - Retrieve the Cairo version by looking at cairo-version.h instead of relying on pkg-config. It requires some additional code for checking if the desired version has been found, but that will not be needed once we start depending on CMake 2.8.3 or later. The only downside is that FPHSA sets _FOUND instead of _FOUND, and to keep things consistent Cairo_LIBRARIES and Cairo_INCLUDE_DIRS have become CAIRO_LIBRARIES and CAIRO_INCLUDE_DIRS. * DumpRenderTree/efl/CMakeLists.txt: Use CAIRO_FOO instead of Cairo_FOO. * EWebLauncher/CMakeLists.txt: Ditto. 2012-05-03 Ojan Vafai Fix modifier and bug sorting on the flakiness dashboard https://bugs.webkit.org/show_bug.cgi?id=85554 Reviewed by Dirk Pranke. The sorting assumes that the column header name matches a field in the testResults object and uses that field for sorting. Filter the bugs out of the modifiers list and put them in a bugs field and only put the non-bug modifiers in the modifiers list. * TestResultServer/static-dashboards/flakiness_dashboard.js: (populateExpectationsData): (htmlForSingleTestRow): * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js: 2012-05-03 Ojan Vafai Fix some state resetting when running unittests. HtmlForSingleTestRow was flaky. * TestResultServer/static-dashboards/flakiness_dashboard.js: (htmlForSingleTestRow): * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js: (resetGlobals): 2012-05-03 Raphael Kubo da Costa webkitpy: Use PlatformInfo wherever possible in the User class. https://bugs.webkit.org/show_bug.cgi?id=85549 Reviewed by Dirk Pranke. Instead of relying on sys.platform to detect whether the current platform is a Mac, use the PlatformInfo class instead. * Scripts/webkitpy/common/system/user.py: (User.__init__): (User.edit_changelog): 2012-05-03 Ojan Vafai Remove the percent failed column from the flakiness dashboard https://bugs.webkit.org/show_bug.cgi?id=85552 Reviewed by Dirk Pranke. This column isn't useful enough to justify the visual noise. * TestResultServer/static-dashboards/flakiness_dashboard.js: (processMissingAndExtraExpectations): (tableHeaders): (htmlForSingleTestRow): (sortTests): * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js: (test): 2012-05-03 Eric Seidel --coverage should not not use the "feature" options system in build-webkit, similarly remove special-casing for WTF_URL https://bugs.webkit.org/show_bug.cgi?id=85542 Reviewed by Daniel Bates. Remove these last two odd-men-out, so that we can generate the default feature options for build-webkit using the new generate-feature-files command. The Mac build system has some special casing for ENABLE_ defines which gets them defined as -D on the build command, even though initially they start out as environment variables ENABLE_FOO = ENABLE_FOO. The Feature options in build-webkit depend on this support, but it doesn't work for WTF_USE_ defines like WTF_USE_WTF_URL. I'm told that WTF_URL is not actively being worked on, so just removing it for now. It should probably come back to life as an ENABLE_, or build-webkit or the Xcode projects should be made generically smarter about these USE_WTF defines. (There is still WTF_USE_TILED_BACKING_STORE in the feature list, but it only seems used by Qt. If someone tried build-webkit --tiled-backing-store on Mac today it would not work!) * Scripts/build-webkit: 2012-05-03 Eric Seidel Remove the last bits of Leopard support from our perl scripts https://bugs.webkit.org/show_bug.cgi?id=85545 Reviewed by Daniel Bates. Chromium still supports Leopard (for a little while yet), but AppleMacWebKit dropped Leopard support several months ago. Remove the last bits of isLeopard() branching in our Perl scripts, since no one is using this (Chromium has never used this Perl code). * Scripts/build-webkit: * Scripts/old-run-webkit-tests: (countAndPrintLeaks): (captureSavedCrashLog): * Scripts/webkitdirs.pm: (determineArchitecture): (argumentsForRunAndDebugMacWebKitApp): 2012-05-03 Ojan Vafai [Chromium] Don't show the WebKit Linux ASAN bot on the flakiness dashboard. https://bugs.webkit.org/show_bug.cgi?id=85551 Reviewed by Dirk Pranke. It's not uploading results and in the short term we don't plan to have it start. * TestResultServer/static-dashboards/builders.js: (isChromiumWebkitTipOfTreeTestRunner): * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js: 2012-05-03 Ojan Vafai Show the test name in the dashboard if showChrome==false and we're showing multiple tests https://bugs.webkit.org/show_bug.cgi?id=85243 Reviewed by Dirk Pranke. This happens for cases where a test name matches multiple tests. In practice, this only happens for virtual test suites (e.g. the canvas tests). * TestResultServer/static-dashboards/flakiness_dashboard.js: (htmlForIndividualTestOnAllBuildersWithChrome): * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js: (testHtmlForIndividualTestOnAllBuildersWithChromeNonexistant): (testHtmlForIndividualTestOnAllBuildersWithChrome): (testHtmlForIndividualTestOnAllBuildersWithChromeWebkitMaster): (testHtmlForIndividualTests): 2012-05-03 Ojan Vafai Port the flakiness dashboard unittests to qunit https://bugs.webkit.org/show_bug.cgi?id=85544 Reviewed by Adam Barth. * TestResultServer/static-dashboards/dashboard_base.js: (appendJSONScriptElements): * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js: (resetGlobals): (runExpectationsTest): (test): * TestResultServer/static-dashboards/run-unittests.html: Added. 2012-05-03 Raphael Kubo da Costa webkitpy: Remove check for readline on Mac. https://bugs.webkit.org/show_bug.cgi?id=85547 Reviewed by Eric Seidel. Python 2.5 is not supported by the WebKit project anymore, and Python's readline documentation does not mention the module not being available in the supported versions. * Scripts/webkitpy/common/system/user.py: 2012-05-03 Thiago Marcos P. Santos [NRWT] Run performance tests with lock https://bugs.webkit.org/show_bug.cgi?id=78628 Reviewed by Tony Chang. Locking performance tests (like we do for http tests) will force them to run in serial. This reduces the load of the machine when running perf tests and minimizes the chances of the tests to fail, specially timeouts. * Scripts/webkitpy/layout_tests/controllers/manager.py: (Manager.__init__): (Manager._is_perf_test): (Manager._test_requires_lock): * Scripts/webkitpy/layout_tests/port/http_lock.py: 2012-05-03 Dirk Pranke Attempt to fix a crash on SL ... Unreviewed, build fix. It looks like we may be getting an IOError raised during select() on chromium mac, and that error isn't being caught. Pushing it into the try/catch block will hopefully catch this and allow us to see why the subprocess is crashing. * Scripts/webkitpy/layout_tests/port/server_process.py: (ServerProcess._wait_for_data_and_update_buffers_using_select): 2012-05-02 Jon Honeycutt WebFrameLoaderClient::dispatchWillSendSubmitEvent() needs to be implemented for WebKit2 https://bugs.webkit.org/show_bug.cgi?id=84304 Reviewed by Jessie Berlin. * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Added new files to project. * TestWebKitAPI/Tests/WebKit2/WillSendSubmitEvent.cpp: Added. (TestWebKitAPI::didReceiveMessageFromInjectedBundle): Look for the "DidReceiveWillSendSubmitEvent" message. The message body will be the dictionary of form control names and values passed to the injected bundle's willSendSubmitEvent handler; look for the pairs "textField" -> "text field" and "passwordField" -> "password field". Check that "hiddenField" is not included. (TestWebKitAPI::setInjectedBundleClient): (TestWebKitAPI::TEST): * TestWebKitAPI/Tests/WebKit2/WillSendSubmitEvent_Bundle.cpp: Added. (TestWebKitAPI::willSendSubmitEvent): Post a message to the TestWebKitAPI process with the dictionary of form control names and values that we received. (TestWebKitAPI::WillSendSubmitEventTest::WillSendSubmitEventTest): (TestWebKitAPI::WillSendSubmitEventTest::didCreatePage): Set up the page's form client. * TestWebKitAPI/Tests/WebKit2/auto-submitting-form.html: Added. 2012-05-03 Eric Seidel Remove dead code from build-webkit for working around old XCode dependency bug https://bugs.webkit.org/show_bug.cgi?id=85535 Reviewed by Adam Barth. Xcode used to have a bug whereby it didn't notice if you changed the #defines passed on the command line to xcodebuild. I wrote code to work around this while bringing up SVG support 5-6 years ago. I think we can finally remove this code from build-webkit, since SVG is the only feature flag to ever use it (modern in-progress features like MathML don't seem to be using the support) an I suspect the XCode bug has long been fixed. * Scripts/build-webkit: * Scripts/webkitperl/features.pm: 2012-05-03 Andy Estes InjectedBundleControllerMac.mm needs to include config.h https://bugs.webkit.org/show_bug.cgi?id=85531 Reviewed by Andreas Kling. * TestWebKitAPI/mac/InjectedBundleControllerMac.mm: 2012-05-03 Raphael Kubo da Costa webkitpy: Use PlatformInfo whenever possible in server_process.py. https://bugs.webkit.org/show_bug.cgi?id=85516 Reviewed by Eric Seidel. Eliminate direct access to sys.platform wherever possible by using PlatformInfo.is_win() instead. * Scripts/webkitpy/layout_tests/port/server_process.py: (ServerProcess.__init__): Remove the executive parameter as it was not passed by any caller, obtain a SystemHost from port_obj instead. (ServerProcess._log): Style, remove extra empty line. (ServerProcess._start): (ServerProcess.stop): (ServerProcess.kill): * Scripts/webkitpy/layout_tests/port/server_process_unittest.py: (TrivialMockPort.__init__): Create a MockSystemHost used by ServerProcess. (TestServerProcess.test_broken_pipe): Test different platforms. 2012-05-03 Thiago Marcos P. Santos webkitpy: Style checker broken when committing a PNG https://bugs.webkit.org/show_bug.cgi?id=85504 Reviewed by Dirk Pranke. * Scripts/webkitpy/style/checkers/png.py: (PNGChecker.__init__): 2012-05-03 Philippe Normand Unreviewed, fix another call to join(), similar to the ones landed in r115975. * Scripts/webkitdirs.pm: (jhbuildConfigurationChanged): 2012-05-03 Raphael Kubo da Costa webkitpy: Remove sys.platform accesses in CrashLogs. https://bugs.webkit.org/show_bug.cgi?id=85436 Reviewed by Dirk Pranke. Instead of directly checking for sys.platform == 'darwin', use a SystemHost and call PlatformInfo.is_mac(). * Scripts/webkitpy/common/system/crashlogs.py: (CrashLogs.__init__): (CrashLogs.find_newest_log): (CrashLogs._log_directory_darwin): (CrashLogs._find_newest_log_darwin): * Scripts/webkitpy/common/system/crashlogs_unittest.py: (CrashLogsTest.test_find_log_darwin): * Scripts/webkitpy/layout_tests/port/mac.py: (MacPort._get_crash_log): * Scripts/webkitpy/layout_tests/port/test.py: (TestDriver.run_test): * Scripts/webkitpy/tool/commands/queries.py: (execute): 2012-05-03 Philippe Normand [GTK] join() is wrongly used in webkitdirs.pm since r115532 https://bugs.webkit.org/show_bug.cgi?id=85501 Reviewed by Martin Robinson. * Scripts/webkitdirs.pm: (getJhbuildPath): (buildAutotoolsProject): 2012-05-03 Zan Dobersek [Gtk] ImageDiff crashes when it can't open a display https://bugs.webkit.org/show_bug.cgi?id=85476 Reviewed by Martin Robinson. Replace gdk_init() initialization call with the g_type_init() call. GdkPixbufLoader doesn't actually need the current call to be made and this change will also remove crashes when ImageDiff is started in an environment without a DISPLAY env. * DumpRenderTree/gtk/ImageDiff.cpp: (main): 2012-05-03 Zan Dobersek ImageDiff should be run inside a properly established environment https://bugs.webkit.org/show_bug.cgi?id=85292 Reviewed by Martin Robinson. When creating the ImageDiff server process, pass along a freshly-set-up environment in which the process should be executed in. * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitPort._start_image_diff_process): 2012-05-03 Tor Arne Vestbø [Qt] Allow the web process and WTR to be paused on startup Makes it easier to debug the web process or run-webkit-tests -2, as you have ample time to attach gdb to the process. Reviewed by Simon Hausmann. * Scripts/webkitpy/layout_tests/port/qt.py: (QtPort.setup_environ_for_server): * WebKitTestRunner/qt/main.cpp: (main): 2012-05-03 Csaba Osztrogonác Password protect "Stop build" button https://bugs.webkit.org/show_bug.cgi?id=81982 Reviewed by Ryosuke Niwa. Only authenticated users should be able to use stopBuild, stopAllBuilds, cancelPendingBuild features to prevent killing builds on build.webkit.org by SPAM bots. * BuildSlaveSupport/build.webkit.org-config/master.cfg: 2012-05-03 Dongwoo Im [DRT] Remove the name of parameter, visibility, from the setPageVisibility stub implementations. https://bugs.webkit.org/show_bug.cgi?id=85468 Reviewed by Nikolas Zimmermann. * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp: (LayoutTestController::setPageVisibility): Remove the name of parameter, visibility. * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: (LayoutTestController::setPageVisibility): Remove the name of parameter, visibility. * DumpRenderTree/qt/LayoutTestControllerQt.cpp: (LayoutTestController::setPageVisibility): Remove the name of parameter, visibility. * DumpRenderTree/win/LayoutTestControllerWin.cpp: (LayoutTestController::setPageVisibility): Remove the name of parameter, visibility. * DumpRenderTree/wx/LayoutTestControllerWx.cpp: (LayoutTestController::setPageVisibility): Remove the name of parameter, visibility. 2012-05-03 Stephanie Lewis https://bugs.webkit.org/show_bug.cgi?id=85450 unbounded growth of JSDOMWindowShells loading pages in the same window REGRESSION (r115083): PLT3 shows linear memory growth and gets slower with each run Reviewed by Brady Eidson. The API added for DOMWindowExtension, didCreateGlobalObjectForFrame, would create a global object for every world, even those that did not need the callback. This had the side effect of creating a JSDOMWindowShell that the associated world didn't necessarily know to clean up. Instead of creating unnecessary objects change the API to globalObjectIsAvailableForFrame and do not pass the global object in the API. The object can be accessed later by those worlds which require it. * TestWebKitAPI/Tests/WebKit2/DOMWindowExtensionBasic.cpp: (TestWebKitAPI): (TestWebKitAPI::didReceiveMessageFromInjectedBundle): * TestWebKitAPI/Tests/WebKit2/DOMWindowExtensionBasic_Bundle.cpp: (TestWebKitAPI): (DOMWindowExtensionBasic): (TestWebKitAPI::DOMWindowExtensionBasic::didCreatePage): (TestWebKitAPI::DOMWindowExtensionBasic::globalObjectIsAvailableForFrame): (TestWebKitAPI::globalObjectIsAvailableForFrameCallback): 2012-05-03 Nikolas Zimmermann Not reviewed. Fix mac build, it was missing setPageVisibility/resetPageVisibility stub implementations. * DumpRenderTree/mac/LayoutTestControllerMac.mm: (LayoutTestController::resetPageVisibility): (LayoutTestController::setPageVisibility): 2012-05-02 Alexander Færøy Rename deviceDPI to devicePixelRatio https://bugs.webkit.org/show_bug.cgi?id=85049 Reviewed by Kenneth Rohde Christiansen. Use 1.5 as device pixel ratio for the Qt MiniBrowser. * MiniBrowser/qt/qml/BrowserWindow.qml: 2012-05-03 Dongwoo Im [EFL][DRT] Implement the LayoutTestController's methods related to the Page Visibility API. https://bugs.webkit.org/show_bug.cgi?id=85347 Reviewed by Nikolas Zimmermann. * DumpRenderTree/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp: (LayoutTestController::resetPageVisibility): Not implemented. (LayoutTestController::setPageVisibility): Not implemented. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::resetPageVisibility): Reset the page visibility. (LayoutTestController::setPageVisibility): Set the page visibility. * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: (LayoutTestController::resetPageVisibility): Not implemented. (LayoutTestController::setPageVisibility): Not implemented. * DumpRenderTree/qt/LayoutTestControllerQt.cpp: (LayoutTestController::resetPageVisibility): Not implemented. (LayoutTestController::setPageVisibility): Not implemented. * DumpRenderTree/win/LayoutTestControllerWin.cpp: (LayoutTestController::resetPageVisibility): Not implemented. (LayoutTestController::setPageVisibility): Not implemented. * DumpRenderTree/wx/LayoutTestControllerWx.cpp: (LayoutTestController::resetPageVisibility): Not implemented. (LayoutTestController::setPageVisibility): Not implemented. 2012-05-02 Sheriff Bot Unreviewed, rolling out r115918. http://trac.webkit.org/changeset/115918 https://bugs.webkit.org/show_bug.cgi?id=85452 "Broke test-webkitpy on Apple Lion bots" (Requested by estes on #webkit). * Scripts/webkitpy/common/system/crashlogs.py: (CrashLogs.__init__): (CrashLogs.find_newest_log): (CrashLogs._log_directory_darwin): (CrashLogs._find_newest_log_darwin): * Scripts/webkitpy/common/system/crashlogs_unittest.py: (CrashLogsTest.test_find_log_darwin): * Scripts/webkitpy/layout_tests/port/mac.py: (MacPort._get_crash_log): * Scripts/webkitpy/layout_tests/port/test.py: (TestDriver.run_test): * Scripts/webkitpy/tool/commands/queries.py: (execute): 2012-05-02 Raphael Kubo da Costa webkitpy: Remove sys.platform accesses in CrashLogs. https://bugs.webkit.org/show_bug.cgi?id=85436 Reviewed by Dirk Pranke. Instead of directly checking for sys.platform == 'darwin', use a SystemHost and call PlatformInfo.is_mac(). * Scripts/webkitpy/common/system/crashlogs.py: (CrashLogs.__init__): (CrashLogs.find_newest_log): (CrashLogs._log_directory_darwin): (CrashLogs._find_newest_log_darwin): * Scripts/webkitpy/common/system/crashlogs_unittest.py: (CrashLogsTest.test_find_log_darwin): * Scripts/webkitpy/layout_tests/port/mac.py: (MacPort._get_crash_log): * Scripts/webkitpy/layout_tests/port/test.py: (TestDriver.run_test): * Scripts/webkitpy/tool/commands/queries.py: (execute): 2012-05-02 Raphael Kubo da Costa webkitpy: Special-case win32 calls in file_lock and executive. https://bugs.webkit.org/show_bug.cgi?id=85431 Reviewed by Dirk Pranke. Simplify accesses to sys.platform by only checking if sys.platform is "win32" and removing explicit checks for the else case. * Scripts/webkitpy/common/system/executive.py: (Executive.check_running_pid): * Scripts/webkitpy/common/system/file_lock.py: (FileLock._create_lock): (FileLock._remove_lock): 2012-05-02 Ojan Vafai Remove pink from flakiness dashboard https://bugs.webkit.org/show_bug.cgi?id=85422 Reviewed by Dirk Pranke. The pink indicates that the test doesn't match it's listing in test_expectations.txt. In practice, this just added noise and confused people. * TestResultServer/static-dashboards/flakiness_dashboard.js: (htmlForSingleTestRow): (showLegend): * TestResultServer/static-dashboards/flakiness_dashboard_tests.css: (#legend-contents *): * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js: (testHtmlForSingleTestRow): 2012-05-02 Gustavo Noronha Silva Unreviewed. Fix mistake I committed when applying review comments. * jhbuild/jhbuild-wrapper: dependencies_path was getting 'Root' appended to its path, so jhbuild itself was being cloned and installed in the wrong place when using WEBKITOUTPUTDIR. 2012-05-02 Dirk Pranke nrwt: make ServerProcess work on chromium win https://bugs.webkit.org/show_bug.cgi?id=85333 Reviewed by Ojan Vafai. This change implements a poor man's select() that will slow-spin doing non-blocking reads on the stdout and stderr named pipes connecting the worker to the driver. Seems to work and I have yet to see much overhead or ill effects but it probably needs more testing. * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumDriver.__init__): * Scripts/webkitpy/layout_tests/port/server_process.py: (ServerProcess._start): (ServerProcess._wait_for_data_and_update_buffers_using_select): (ServerProcess._wait_for_data_and_update_buffers_using_win32_apis): (ServerProcess): (ServerProcess._non_blocking_read_win32): (ServerProcess._read): * Scripts/webkitpy/layout_tests/port/server_process_unittest.py: (TestServerProcess.test_basic): Added. 2012-05-02 Raphael Kubo da Costa webkitpy: Use a SystemHost instead of PlatformInfo in the PNG checker. https://bugs.webkit.org/show_bug.cgi?id=85426 Reviewed by Dirk Pranke. r115880 replaced the direct use of sys.platform with a PlatformInfo object. Dirk Pranke later told me SystemHost objects should be preferred, as nothing out of webkitpy.common.system should create its own PlatformInfo instances. * Scripts/webkitpy/common/system/systemhost_mock.py: (MockSystemHost.__init__): Make it possible to specify a custom filesystem object. * Scripts/webkitpy/style/checkers/png.py: (PNGChecker.__init__): Accept a host object instead of a filesystem one, which is now obtained from host. (PNGChecker._config_file_path): Retrieve PlatformInfo from SystemHost. * Scripts/webkitpy/style/checkers/png_unittest.py: (PNGCheckerTest.test_init): Adjust calls to PNGChecker to its new constructor signature. (PNGCheckerTest.test_check): Ditto. 2012-05-02 Raphael Kubo da Costa webkitpy: Remove direct checks for sys.platform == 'darwin' in the layout test classes. https://bugs.webkit.org/show_bug.cgi?id=85424 Reviewed by Tony Chang. We should reduce direct usage of sys.platform in webkitpy and use PlatformInfo instead. * Scripts/webkitpy/layout_tests/servers/http_server.py: (Lighttpd._prepare_config): Call PlatformInfo.is_mac() instead of checking for 'darwin' directly. * Scripts/webkitpy/layout_tests/servers/http_server_base.py: (HttpServerBase.__init__): Ditto. 2012-05-02 Raphael Kubo da Costa webkitpy: Remove direct sys.platform usage in run_webkit_tests_integrationtest. https://bugs.webkit.org/show_bug.cgi?id=85419 Reviewed by Dirk Pranke. As pointed out by Dirk Pranke, a PlatformInfo is not supposed to be created directly, we should should a SystemHost instead. * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py: (MainTest.setUp): Create a SystemHost and obtain the PlatformInfo object from it instead of creating PlatformInfo directly. 2012-05-02 Raphael Kubo da Costa webkitpy: Remove direct sys.platform usage in run_webkit_tests_integrationtest. https://bugs.webkit.org/show_bug.cgi?id=85419 Rubber-stamped by Dirk Pranke. Follow-up patch to r115887 which fixes some capitalization problems pointed out in comments 2 and 4. * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py: (MainTest.setUp): (MainTest.test_child_processes_2): (MainTest.test_child_processes_min): (MainTest.test_exception_raised): (MainTest.test_keyboard_interrupt): 2012-05-02 Raphael Kubo da Costa webkitpy: Remove direct sys.platform usage in run_webkit_tests_integrationtest. https://bugs.webkit.org/show_bug.cgi?id=85419 Reviewed by Dirk Pranke. Another part of the work towards replacing direct sys.platform usage with PlatformInfo. * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py: (MainTest.setUp): Move the setting of SHOULD_TEST_PROCESSES here. Even though it is costlier, unittest.setUpClass was introduced in Python 2.7. (MainTest.test_child_processes_2): Use self.SHOULD_TEST_PROCESSES. (MainTest.test_child_processes_min): Ditto. (MainTest.test_exception_raised): Ditto. (MainTest.test_keyboard_interrupt): Ditto. (MainTest.test_crash_log): Check for PlatformInfo.is_mac(). (MainTest.test_web_process_crash_log): Ditto, and remove duplicate check at the end. (MainTest.test_no_http_tests): Remove check for (cygwin && Python 2.5.2), as we only support Python 2.6+. 2012-05-02 Raphael Kubo da Costa webkitpy: Remove direct sys.platform usage in the PNG checker. https://bugs.webkit.org/show_bug.cgi?id=85407 Reviewed by Tony Chang. Work towards replacing direct sys.platform usage with PlatformInfo. * Scripts/webkitpy/common/system/platforminfo_mock.py: (MockPlatformInfo.__init__): Accept os_name and os_version as arguments. * Scripts/webkitpy/style/checkers/png.py: (PNGChecker.__init__): Create a PlatformInfo object instead of using sys.platform when needed. (PNGChecker._config_file_path): Check for the current platform via PlatformInfo. * Scripts/webkitpy/style/checkers/png_unittest.py: (PNGCheckerTest.test_check): Update tests after the changes above. 2012-05-02 Raphael Kubo da Costa [watchlist] Update watchlist rules for EFL-related patches. http://bugs.webkit.org/show_bug.cgi?id=85160 Reviewed by Eric Seidel. * Scripts/webkitpy/common/config/watchlist: Monitor Source/WTF/wtf/efl and LayoutTests/platform/efl as well. 2012-05-02 Tor Arne Vestbø Revert r115191 - "Make the web view's url property follow the active url" It uncovered/caused issues in the icon-implementation that can't be easily fixed, so rolling out instead. https://bugs.webkit.org/show_bug.cgi?id=77554 Rubber-stamped by Simon Hausmann. * MiniBrowser/qt/qml/BrowserWindow.qml: * TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp: (TestWebKitAPI::TEST): 2012-05-02 Zan Dobersek Reftest mismatches are providing undefined diff percent in results.html https://bugs.webkit.org/show_bug.cgi?id=85283 Reviewed by Ojan Vafai. When writing test results for a reftest mismatch, also take into account the difference percent that is produced when generating image diff. This value is then used when creating test results in the form of dictionaries, stored under the 'image_diff_percent' key. It ultimately ends up in JSON results data, providing additional information about the reftest difference. * Scripts/webkitpy/layout_tests/controllers/manager.py: (interpret_test_failures): Store the diff percent for reftest failure under the 'image_diff_percent' key. * Scripts/webkitpy/layout_tests/controllers/test_result_writer.py: (write_test_result): Also set the new diff_percent attribute of the reftest mismatch failure to the value of the difference percent ImageDiff outputs. * Scripts/webkitpy/layout_tests/models/test_failures.py: (FailureReftestMismatch.__init__): Add the new diff_percent attribute with the default value of None. * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py: (EndToEndTest.test_reftest_with_two_notrefs): Update the test case expected results. 2012-05-02 Terry Anderson Addition of a contributor to committers.py https://bugs.webkit.org/show_bug.cgi?id=85324 Reviewed by Unreviewed. * Scripts/webkitpy/common/config/committers.py: 2012-05-02 Mikhail Pozdnyakov [EFL] EFL's LayoutTestController overridePreference implementation https://bugs.webkit.org/show_bug.cgi?id=83007 Reviewed by Antonio Gomes. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (toBool): aux function converting JSStringRef to bool (toInt): aux function converting JSStringRef to int (LayoutTestController::overridePreference): added implementation 2012-05-01 Raphael Kubo da Costa webkitpy: Fix sys.platform check in find_files_unittest. https://bugs.webkit.org/show_bug.cgi?id=85339 Reviewed by Ryosuke Niwa. Follow-up to r101618 -- 'win' is not a valid sys.platform value, 'win32' should be used instead. The check always returned false due to this, so assert_normalize() was never called. * Scripts/webkitpy/common/find_files_unittest.py: (TestWinNormalize.test_win): Add the missing import for the FileSystem class and fix the sys.platform check. 2012-05-02 Zan Dobersek [GTK] drag&drop related failures https://bugs.webkit.org/show_bug.cgi?id=84675 Reviewed by Philippe Normand. Rather than checking options to see whether we're running pixel tests the pixel_tests argument passed to GtkDriver._start should be used as it also covers cases when pixel tests should be run because of testing reftests. This again creates different Xvfb processes with proper display IDs for a worker that's performing both ordinary and pixel tests, as was the case before changes made in r114788. * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkDriver._start): 2012-05-02 Simon Hausmann [Qt] Enable device orientation/motion with Qt 5 Reviewed by Tor Arne Vestbø. Enable device orientation/motion features with Qt 5 if Qt Sensors is available. * qmake/mkspecs/features/features.prf: 2012-05-01 Raphael Kubo da Costa webkitpy: Remove unused platform imports. https://bugs.webkit.org/show_bug.cgi?id=85341 Reviewed by Eric Seidel. Remove a few uses of `import platform' which were either rendered unneeded after r105931 or were never needed at all. * Scripts/webkitpy/common/system/executive.py: * Scripts/webkitpy/tool/bot/flakytestreporter.py: 2012-05-01 Terry Anderson [chromium] Accept four parameters when dispatching a WebInputEvent::GestureTap in chromium DRT eventSender https://bugs.webkit.org/show_bug.cgi?id=85289 Reviewed by Eric Seidel. * DumpRenderTree/chromium/EventSender.cpp: (EventSender::gestureEvent): This change will allow for testing of WK85101. Currently no tests use the extra parameters deltaX and deltaY. Added a check for the size of |arguments|. 2012-05-01 Eric Seidel Add myself to the watchlist to watch all rendering changes, as well as sort the watchers in the list https://bugs.webkit.org/show_bug.cgi?id=85305 Reviewed by Adam Barth. * Scripts/webkitpy/common/config/watchlist: 2012-05-01 Sheriff Bot Unreviewed, rolling out r115735. http://trac.webkit.org/changeset/115735 https://bugs.webkit.org/show_bug.cgi?id=85314 cause two tests to crash due to an ASSERTION failure (Requested by zhenyao on #webkit). * DumpRenderTree/chromium/EventSender.cpp: (EventSender::gestureEvent): 2012-05-01 Terry Anderson [chromium] Accept four parameters when dispatching a WebInputEvent::GestureTap in chromium DRT eventSender https://bugs.webkit.org/show_bug.cgi?id=85289 Reviewed by Eric Seidel. * DumpRenderTree/chromium/EventSender.cpp: (EventSender::gestureEvent): This change will allow for testing of WK85101. Currently no tests use the extra parameters deltaX and deltaY. 2012-05-01 Zan Dobersek Print out stderr output of ImageDiff if it is present https://bugs.webkit.org/show_bug.cgi?id=85285 Reviewed by Philippe Normand. Log the stderr output of ImageDiff as a warning if present. This is especially helpful when ImageDiff crashes and stderr data might contain relevant information about what caused the crash. * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitPort._read_image_diff): 2012-05-01 Kenneth Rohde Christiansen [Qt] Add an experimental extension to set the min. contents width https://bugs.webkit.org/show_bug.cgi?id=85281 Reviewed by Antonio Gomes. * MiniBrowser/qt/qml/BrowserWindow.qml: Set the value of preferredMinimumContentsWidth to 980. 2012-04-30 Ojan Vafai Fix tests to not depend on the actual list of builders at build.chromium.org. * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js: (testHtmlForIndividualTestOnAllBuildersWithChrome): (testHtmlForIndividualTestOnAllBuildersWithChromeWebkitMaster): 2012-04-29 Sam Weinig Remove BlobBuilder https://bugs.webkit.org/show_bug.cgi?id=84036 Reviewed by Anders Carlsson. * Scripts/build-webkit: * qmake/mkspecs/features/features.prf: Make exposing the WebKitBlobBuilder JS object conditional on a new ENABLE_LEGACY_WEBKIT_BLOB_BUILDER flag. Don't enable this for the Mac, but do for everyone else. 2012-04-30 Amruth Raj Padmanabhuni [GTK] Add glib as a dependency to build glib-networking https://bugs.webkit.org/show_bug.cgi?id=85181 Reviewed by Martin Robinson. * gtk/jhbuild.modules: 2012-04-30 Alexis Menard Move myself to reviewers. Reviewed by Andreas Kling. * Scripts/webkitpy/common/config/committers.py: 2012-04-30 Jocelyn Turcotte Unreviewed, move myself to reviewers * Scripts/webkitpy/common/config/committers.py: 2012-04-30 Carlos Garcia Campos Unreviewed. Fix make distcheck. * TestWebKitAPI/GNUmakefile.am: Add missing headers to compilation and resources to EXTRA_DIST. 2012-04-30 Sudarsana Nagineni [GTK] DRT support for layoutTestController.removeOriginAccessWhitelistEntry https://bugs.webkit.org/show_bug.cgi?id=85156 Reviewed by Martin Robinson. Add missing implementation removeOriginAccessWhitelistEntry to GTK's LayoutTestController. * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: (LayoutTestController::removeOriginAccessWhitelistEntry): 2012-04-29 Maciej Stachowiak REGRESSION: On Lion, run-webkit-tests changes the display color profile even when not running pixel tests https://bugs.webkit.org/show_bug.cgi?id=81729 Reviewed by Dan Bernstein. * Scripts/webkitpy/layout_tests/controllers/manager.py: (Manager._set_up_run): Only launch LayoutTestHelper when running pixel tests. 2012-04-29 Sudarsana Nagineni [GTK] DRT needs an implementation of layoutTestController.setDomainRelaxationForbiddenForURLScheme https://bugs.webkit.org/show_bug.cgi?id=85131 Reviewed by Martin Robinson. Add missing implementation setDomainRelaxationForbiddenForURLScheme to GTK's LayoutTestController. * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: (LayoutTestController::setDomainRelaxationForbiddenForURLScheme): 2012-04-28 Sam Weinig Smooth scrolling needs a new key Reviewed by Dan Bernstein. * DumpRenderTree/mac/DumpRenderTree.mm: (resetDefaultsToConsistentValues): * TestWebKitAPI/mac/InjectedBundleControllerMac.mm: (TestWebKitAPI::InjectedBundleController::platformInitialize): * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm: (WTR::InjectedBundle::platformInitialize): Update for new key. 2012-04-28 Zan Dobersek [WK2] http/tests/navigation/anchor-frames-gbk.html fails https://bugs.webkit.org/show_bug.cgi?id=76896 Reviewed by Darin Adler. Only dump as text if currently dumping render tree. This ensures that calling dumpAsText in a test with dumpChildFramesAsText already being called doesn't override the first decision, possibly making tests fail. * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: (WTR::LayoutTestController::dumpAsText): 2012-04-28 Dan Winship Unreviewed; Add me to the SoupNetwork watchlist * Scripts/webkitpy/common/config/watchlist: 2012-04-27 Gustavo Noronha Silva [GTK] jhbuild cloning is not following WEBKITOUTPUTDIR. https://bugs.webkit.org/show_bug.cgi?id=76161 Reviewed by Martin Robinson. * Scripts/webkitdirs.pm: (getJhbuildPath): New method to obtain the jhbuild base directory, using the product base directory (jhbuildConfigurationChanged): Use the new method (buildAutotoolsProject): Ditto. * efl/jhbuildrc: Use WEBKITOUTPUTDIR when calculating the path. * gtk/jhbuildrc: Ditto. * jhbuild/jhbuild-wrapper: Ditto. 2012-04-27 Scott Graham [chromium] make case of npTestNetscapePlugIn.dll match case in copy rule https://bugs.webkit.org/show_bug.cgi?id=85083 Reviewed by Dirk Pranke. Make output product_name for npTestNetscapePlugIn.dll match the case of the copy_TestNetscapePlugIn rule. This is required for ninja, which is more particular about case matching in rules than the VS build. * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp: 2012-04-27 Dirk Pranke [chromium] use "drt-style" output, not "test-shell-style" output, on mac and linux DRT https://bugs.webkit.org/show_bug.cgi?id=84917 Unreviewed, build fix. Reland the change in r115453 with a fix for chromium win. The logic in ChromiumDriver was busted on windows, causing DRT to get launched without the --test-shell flag. * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumDriver): (ChromiumDriver.__init__): (ChromiumDriver._wrapper_options): (ChromiumDriver.cmd_line): (ChromiumDriver._start): (ChromiumDriver.has_crashed): (ChromiumDriver.run_test): (ChromiumDriver.stop): * Scripts/webkitpy/layout_tests/port/chromium_unittest.py: (ChromiumDriverTest.setUp): (ChromiumDriverTest.test_stop): (ChromiumDriverTest.test_two_drivers.MockDriver.__init__): (ChromiumDriverTest.test_two_drivers): 2012-04-27 Peter Beverloo Add John Grabowski as a non-committer to committers.py https://bugs.webkit.org/show_bug.cgi?id=85080 Reviewed by Dirk Pranke. He's being CC'ed often enough because of his involvement in the Chromium side of Chrome for Android upstreaming. * Scripts/webkitpy/common/config/committers.py: add jrg@chromium.org 2012-04-27 Allan Sandfeld Jensen [Qt] Fix minimal build. https://bugs.webkit.org/show_bug.cgi?id=85045 Reviewed by Tor Arne Vestbø. Do not set conflicting values of ENABLE_XSLT. * qmake/mkspecs/features/features.prf: 2012-04-27 Dominik Röttsches [EFL][DRT] EFL should sanitize pulseaudio module configuration too https://bugs.webkit.org/show_bug.cgi?id=85047 Reviewed by Dirk Pranke. PulseAudio test setup sanitization shared between EFL and GTK. * Scripts/webkitpy/layout_tests/port/efl.py: (EflPort): (EflPort.setup_test_run): (EflPort.clean_up_test_run): * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkPort): Moving pulseaudio unload/restore out to shared class. * Scripts/webkitpy/layout_tests/port/pulseaudio_sanitizer.py: Added. (PulseAudioSanitizer): (PulseAudioSanitizer._unload_pulseaudio_module): (PulseAudioSanitizer._restore_pulseaudio_module): 2012-04-27 Dirk Pranke Unreviewed, revert r115453 - broke chromium win bots * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumPort._path_to_image_diff): (ChromiumDriver): (ChromiumDriver.__init__): (ChromiumDriver._wrapper_options): (ChromiumDriver.cmd_line): (ChromiumDriver.stop): * Scripts/webkitpy/layout_tests/port/chromium_unittest.py: (ChromiumDriverTest.setUp): (ChromiumDriverTest.test_stop): (ChromiumDriverTest.test_two_drivers): (ChromiumDriverTest.test_two_drivers.MockDriver.__init__): 2012-04-27 Ryosuke Niwa Chromium Windows Perf bots timeout due to not output https://bugs.webkit.org/show_bug.cgi?id=84940 Reviewed by Dirk Pranke. Dirk and I investigated the issue on the bot but we couldn't figure out what is going wrong. Since we run-perf-tests don't need any of fancy feature printer provides, just use python's built-in logging module instead. Printing out to stdout and stderr seem to work so hopefully this will fix the issue on the bot. * Scripts/run-perf-tests: * Scripts/webkitpy/performance_tests/perftest.py: (PerfTest.run): (PerfTest.run_failed): (PerfTest.parse_output): (PerfTest.output_statistics): (ChromiumStylePerfTest.parse_output): (PageLoadingPerfTest.run): * Scripts/webkitpy/performance_tests/perftest_unittest.py: (MainTest.test_parse_output): (MainTest.test_parse_output_with_failing_line): (TestPageLoadingPerfTest): (TestPageLoadingPerfTest.test_run): (TestPageLoadingPerfTest.test_run_with_bad_output): * Scripts/webkitpy/performance_tests/perftestsrunner.py: (PerfTestsRunner.__init__): (PerfTestsRunner._parse_args): (PerfTestsRunner.run): (PerfTestsRunner._upload_json): (PerfTestsRunner._print_status): (PerfTestsRunner._run_tests_set): (PerfTestsRunner._run_single_test): * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py: (create_runner): (test_run_test_set): (test_run_test_set_kills_drt_per_run): (test_run_test_pause_before_testing): (test_run_test_set_for_parser_tests): (test_run_test_set_with_json_output): (test_run_test_set_with_json_source): (test_run_test_set_with_multiple_repositories): (test_upload_json): (test_parse_args): 2012-04-27 Dirk Pranke [chromium] use "drt-style" output, not "test-shell-style" output, on mac and linux DRT https://bugs.webkit.org/show_bug.cgi?id=84917 Reviewed by Ojan Vafai. This change changes the chromium new-run-webkit-tests implementation to use the code in WebKitDriver and ServerProcess by default on Mac and Linux, instead of using the old-style Chromium/test_shell style of output. We will still use the test_shell style on Windows for now and you can still get the old style of behavior with new-run-webkit-tests --additional-drt-flag --test-shell * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumDriver): (ChromiumDriver.__init__): (ChromiumDriver._wrapper_options): (ChromiumDriver.cmd_line): (ChromiumDriver._start): (ChromiumDriver.has_crashed): (ChromiumDriver.run_test): (ChromiumDriver.stop): * Scripts/webkitpy/layout_tests/port/chromium_unittest.py: (ChromiumDriverTest.setUp): (ChromiumDriverTest.test_stop): (ChromiumDriverTest.test_two_drivers.MockDriver.__init__): (ChromiumDriverTest.test_two_drivers): 2012-04-27 Dirk Pranke webkitpy: fix race in unit tests for manager_worker_broker https://bugs.webkit.org/show_bug.cgi?id=84925 Reviewed by Ojan Vafai. I have been intermittently seeing hangs when running test-webkitpy on Mac SL; it appears that there's a bug (?) in the multiprocessing module where queue references are getting leaked or not cleaned up somehow. By explicitly closing the queues, things seem to be happy, so this patch adds a cleanup() method to the message broker and modifies the manager (and the unit tests) to call it. This may get rid of the intermittent error on shutdown on Linux I've seen as well. * Scripts/webkitpy/layout_tests/controllers/manager.py: (Manager._run_tests): * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py: (_Broker.__del__): (_Broker): (_Broker.cleanup): (_BrokerConnection.cleanup): * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py: (_TestsMixin.test_name): (_TestsMixin.test_cancel): (_TestsMixin.test_done): (_TestsMixin.test_unknown_message): 2012-04-27 Gavin Peters Add new ENABLE_LINK_PRERENDER define to control the Prerendering API https://bugs.webkit.org/show_bug.cgi?id=84871 Reviewed by Adam Barth. Prerendering is currently covered by the ENABLE_LINK_PREFETCH macro, but the new Prerendering API separates it from prefetching. Having separate include guards lets ports enable prefetching, a relatively easy change, without needing to build the infrastructure for prerendering, which is considerably more complicated. 2012-04-27 Philippe Normand Unreviewed, GTK gardening. * Scripts/run-gtk-tests: (TestRunner): Skip flaky WebKit2 API test. 2012-04-27 Philippe Normand Unreviewed, GTK gardening. * Scripts/run-gtk-tests: (TestRunner): Fix another typo. So easy to get confused by a Test name and the actual name of the executable... 2012-04-27 Philippe Normand Unreviewed, GTK gardening. * Scripts/run-gtk-tests: (TestRunner): Skip failing WebKit2 API test. 2012-04-27 Christophe Dumez [EFL] DRT should support LayoutTestController's dumpResourceLoadCallbacks() https://bugs.webkit.org/show_bug.cgi?id=81891 Reviewed by Philippe Normand. Implement support for LayoutTestController's dumpResourceLoadCallbacks() in EFL's DumpRenderTree by catching new ewk_view signals about resource loading and print out the expected information. This allows for several test cases to be removed from the skip list. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::createView): (descriptionSuitableForTestResult): (DumpRenderTreeChrome::onResponseReceived): (DumpRenderTreeChrome::onResourceLoadFinished): (DumpRenderTreeChrome::onResourceLoadFailed): (DumpRenderTreeChrome::onNewResourceRequest): * DumpRenderTree/efl/DumpRenderTreeChrome.h: (DumpRenderTreeChrome): 2012-04-27 Philippe Normand Unreviewed, GTK typo fix. * Scripts/run-gtk-tests: (TestRunner): Fix-up skipped WebKit2 API test name TestWKConnection. 2012-04-26 Carlos Garcia Campos [SOUP] Add a way to register custom uri schemes in WebKit2 https://bugs.webkit.org/show_bug.cgi?id=84130 Reviewed by Martin Robinson. * Scripts/webkitpy/style/checkers/cpp.py: (check_identifier_name_in_declaration): Add webkit_soup prefix to the list of exceptions to the underscores in identifiers rule. 2012-04-26 Ryosuke Niwa Extract PerfTestFactory https://bugs.webkit.org/show_bug.cgi?id=83944 Reviewed by Dirk Pranke. Extracted PerfTestFactory out of PerfTest. This class creates appropriate PerfTest object given a test name and a path. Also removed dirname from the constructors of PerfTest and its subclasses. Also added simple test cases to sanity check PerfTestFactory. * Scripts/webkitpy/performance_tests/perftest.py: (PerfTest.__init__): (PerfTest.test_name): (ChromiumStylePerfTest.__init__): (PageLoadingPerfTest.__init__): (PageLoadingPerfTest.run): (PerfTestFactory): (PerfTestFactory.create_perf_test): * Scripts/webkitpy/performance_tests/perftest_unittest.py: (MainTest.test_parse_output): (MainTest.test_parse_output_with_failing_line): (TestPageLoadingPerfTest.test_run): (TestPageLoadingPerfTest.test_run_with_bad_output): (TestPerfTestFactory): (TestPerfTestFactory.test_regular_test): (TestPerfTestFactory.test_inspector_test): (TestPerfTestFactory.test_page_loading_test): * Scripts/webkitpy/performance_tests/perftestsrunner.py: (PerfTestsRunner): (PerfTestsRunner._collect_tests): * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py: (run_test): (_tests_for_runner): 2012-04-26 Dirk Pranke nrwt: clean up server process, webkit driver so chromium can use it https://bugs.webkit.org/show_bug.cgi?id=84910 Reviewed by Ojan Vafai. This change moves the "sample a process" logic out of server_process.py and into a port-specific class (where really only the mac has an implementation), and also preemptively kills DRT when a test times out in WebKitDriver (rather than waiting through the additional delays caused by calling stop() when we would want to restart the driver generically in worker.py). These changes will make it possible for the chromium port to switch over to the stock WebKitDriver implementation, at least on mac and linux. * Scripts/webkitpy/layout_tests/port/base.py: (Port.sample_process): * Scripts/webkitpy/layout_tests/port/mac.py: (MacPort.sample_process): * Scripts/webkitpy/layout_tests/port/mac_unittest.py: (test_helper_fails_to_stop): (test_sample_process): (test_sample_process.logging_run_command): (test_sample_process_throws_exception): (test_sample_process_throws_exception.throwing_run_command): * Scripts/webkitpy/layout_tests/port/server_process.py: (ServerProcess._log): (ServerProcess._handle_timeout): (ServerProcess.stop): (ServerProcess): (ServerProcess.kill): Here we add a method to immediately stop the process rather than trying to shut it down cleanly. * Scripts/webkitpy/layout_tests/port/server_process_unittest.py: (TestServerProcess.test_broken_pipe): * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitDriver.run_test): Fix an issue where we weren't passing along any per-test args (only needed for Chromium, but still). Also, kill the driver immediately if we time out a test. 2012-04-26 Benjamin Poulain ObjcClass::methodsNamed() can leak if buffer is dynamically allocated https://bugs.webkit.org/show_bug.cgi?id=84668 Reviewed by Alexey Proskuryakov. Extend ObjCPlugin to support the new layout tests of the Objective-C bridge. * DumpRenderTree/mac/ObjCPlugin.m: (+[ObjCPlugin isSelectorExcludedFromWebScript:]): (+[ObjCPlugin webScriptNameForSelector:]): (-[ObjCPlugin methodMappedToLongName]): (-[ObjCPlugin testConversionColon:]): (-[ObjCPlugin _:]): 2012-04-26 Dimitri Glazkov Unreviewed, rolling out r115340. http://trac.webkit.org/changeset/115340 https://bugs.webkit.org/show_bug.cgi?id=83048 Does not work with Windows. * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py: (SingleTestRunner._run_reftest): * Scripts/webkitpy/layout_tests/controllers/test_result_writer.py: (copy_file): * Scripts/webkitpy/layout_tests/port/base.py: (Port._parse_reftest_list): (Port.reference_files): (Port._real_tests): * Scripts/webkitpy/layout_tests/port/base_unittest.py: (PortTest.test_reference_files): * Scripts/webkitpy/layout_tests/port/chromium_unittest.py: (ChromiumDriverTest.test_write_command_and_read_line): 2012-04-26 Christophe Dumez [EFL] Enable VIDEO_TRACK feature https://bugs.webkit.org/show_bug.cgi?id=84830 Reviewed by Gustavo Noronha Silva. Enable VIDEO_TRACK by default on EFL port whenever build-webkit is used. * Scripts/build-webkit: 2012-04-08 Robert Hogan new-webkit-run-tests: handle ref tests from the CSS test suite https://bugs.webkit.org/show_bug.cgi?id=83048 Reviewed by Ryosuke Niwa. This allows new-run-webkit-tests to run reference tests created for the CSS test suite. It removes a big swathe of pixel and text results that are no longer required now that the reference results are used and adds in the support files from the CSS test suite that allow the ref tests to run. * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py: (SingleTestRunner._run_reftest): Assert that the reference result exists. Previously the test would run without it and report a pass! * Scripts/webkitpy/layout_tests/controllers/test_result_writer.py: (copy_file): ensure the output directory exists, it may not if the reference test is contained in a folder that contains nothing but reference tests. * Scripts/webkitpy/layout_tests/port/base.py: (Port.available_reference_files): Add as a helper. Handle embedded relative paths, e.g: some/../path (Port.reference_files): Only return reference results in the reftest list that actually exist on disk. Also, don't ignore potential reference results on disk that are not in the reftest list, and don't ignore them just because we have a reftest list for the test's directory. (Port._real_tests): ignore anything that looks like a test in a 'support' directory, this semantic is used by the CSS 2.1 test suite. * Scripts/webkitpy/layout_tests/port/base_unittest.py: (PortTest.test_reference_files_when_reftestlist_and_reference_file_present): Ensure that a test with a reference result on disk but no reference result listed in the directory's reftest list picks up its reference result. * Scripts/webkitpy/layout_tests/port/chromium_unittest.py: (ChromiumDriverTest.test_strip_uri): Test that some/test.html and some/other/../test.html are treated as the same file. 2012-04-26 Carlos Garcia Campos Unreviewed. Fix GTK+ build after r115314. For some reason it fails while bulding GtkVersioning for WTR, but it didn't fail neither in EWS nor in my laptop. We don't really need GtkVersioning in WTR since it already depends on GTK+3 unconditionally. * WebKitTestRunner/GNUmakefile.am: Remove GtkVersioning.c from compilation. * WebKitTestRunner/gtk/EventSenderProxyGtk.cpp: (WTR::EventSenderProxy::createMouseButtonEvent): Use GTK+ 3 API directly instead of getDefaultGDKPointerDevice() helper. (WTR::EventSenderProxy::keyDown): Ditto. (WTR::EventSenderProxy::mouseMoveTo): Ditto. 2012-04-24 Ojan Vafai Show flakiness dashboard data in garden-o-matic https://bugs.webkit.org/show_bug.cgi?id=83716 Reviewed by Dimitri Glazkov. Put an iframe below the expected/actual results in the Results view. Size the iframe to it's height. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/garden-o-matic.html: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/run-unittests.html: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results_unittests.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui_unittests.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/styles/results.css: 2012-04-26 Sudarsana Nagineni [EFL] [DRT] LayoutTestController needs implementation of setJavaScriptProfilingEnabled https://bugs.webkit.org/show_bug.cgi?id=84576 Reviewed by Martin Robinson. Add missing implementation setJavaScriptProfilingEnabled to EFL's LayoutTestController in order to unskip tests rely on this setting. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::setJavaScriptProfilingEnabled): 2012-04-26 Xianzhu Wang Add Jing Zhao, Min Qin, Tien-Ren Chen and Hao Zheng into contributors list. Not reviewed. * Scripts/webkitpy/common/config/committers.py: 2012-04-26 Carlos Garcia Campos [GTK] Test TestWebKitAPI/WebKit2/TestNewFirstVisuallyNonEmptyLayout times out https://bugs.webkit.org/show_bug.cgi?id=84957 Reviewed by Martin Robinson. * Scripts/run-gtk-tests: (TestRunner): Unskip TestNewFirstVisuallyNonEmptyLayout and TestNewFirstVisuallyNonEmptyLayoutForImages. * TestWebKitAPI/gtk/PlatformWebViewGtk.cpp: (TestWebKitAPI::PlatformWebView::PlatformWebView): Show the toplevel window containing the web view to make sure events are received. 2012-04-26 Carlos Garcia Campos Unreviewed. Add bug numbers to skipped tests. * Scripts/run-gtk-tests: (TestRunner): 2012-04-26 Carlos Garcia Campos [GTK] Build and run TestWebKitAPI WebKit2 unit tests https://bugs.webkit.org/show_bug.cgi?id=84446 Reviewed by Philippe Normand. * Scripts/run-gtk-tests: (TestRunner): Add TestWebKitAPI/WebKit2 to the list of directories containing tests. Unskip FunctionalTest.RefCountedStorage, since we are now using a custom main that initializes WTF threading. (TestRunner._setup_testing_environment): Set environment variables required to run WebKit2 tests. * TestWebKitAPI/GNUmakefile.am: * TestWebKitAPI/JavaScriptTest.cpp: Use JavaScriptCore/JSContextRef.h instead of JavaScriptCore/JavaScriptCore.h since it includes JSStringRefCF.h unconditionally. * TestWebKitAPI/PlatformWebView.h: Define PlatformWKView and PlatformWindow for GTK. * TestWebKitAPI/Tests/WebKit2/EvaluateJavaScript.cpp: Remove unneded header that breaks the build of GTK port. * TestWebKitAPI/Tests/WebKit2/PreventEmptyUserAgent.cpp: Use JavaScriptCore/JSContextRef.h instead of JavaScriptCore/JavaScriptCore.h since it includes JSStringRefCF.h unconditionally. * TestWebKitAPI/Tests/WebKit2/WKPreferences.cpp: (TestWebKitAPI::TEST): Add expected values for GTK+ port. * TestWebKitAPI/gtk/InjectedBundleControllerGtk.cpp: (TestWebKitAPI): (TestWebKitAPI::InjectedBundleController::platformInitialize): * TestWebKitAPI/gtk/PlatformUtilitiesGtk.cpp: Added. (TestWebKitAPI::Util::checkTestFinished): (TestWebKitAPI::Util::run): (TestWebKitAPI::Util::sleep): (TestWebKitAPI::Util::getFilenameFromEnvironmentVariableAsUTF8): (TestWebKitAPI::Util::createInjectedBundlePath): (TestWebKitAPI::Util::createURLForResource): (TestWebKitAPI::Util::URLForNonExistentResource): (TestWebKitAPI::Util::isKeyDown): * TestWebKitAPI/gtk/PlatformWebViewGtk.cpp: Added. (TestWebKitAPI::PlatformWebView::PlatformWebView): (TestWebKitAPI::PlatformWebView::~PlatformWebView): (TestWebKitAPI::PlatformWebView::page): (TestWebKitAPI::PlatformWebView::resizeTo): (TestWebKitAPI::doKeyStroke): (TestWebKitAPI::PlatformWebView::simulateSpacebarKeyPress): (TestWebKitAPI::PlatformWebView::simulateAltKeyPress): (TestWebKitAPI::doMouseButtonEvent): (TestWebKitAPI::PlatformWebView::simulateRightClick): (TestWebKitAPI::PlatformWebView::simulateMouseMove): * TestWebKitAPI/gtk/main.cpp: (main): 2012-04-26 Carlos Garcia Campos [GTK] run-gtk-tests: Use a timeout per test instead of a global timeout https://bugs.webkit.org/show_bug.cgi?id=84695 Reviewed by Philippe Normand. It also adds a command line option to be able to pass a custom timeout value, instead of hard-coding it. The default timeout is 10 seconds if no other value is passed to the script. * Scripts/run-gtk-tests: (TestTimeout): Exception raised when a test times out. (TestRunner._get_child_pid_from_test_output): Helper function to get the pid of the running test from gtester output. (TestRunner._kill_process): Helper funtion ot kill a process ignoring exceptions if the process is already died. (TestRunner._run_test_command): Run the test command raising TestTimeout exception if the test doesn't finish before the given timeout in seconds. (TestRunner._run_test_command.alarm_handler): Alarm handler that raises TestTimeout exception. (TestRunner._run_test_glib): Use _run_test_command() to run the test. (TestRunner._run_test_google): Ditto. (TestRunner.run_tests): Add current test to the list of timed out test if TestTimeout exception is raised. Show the list of timed out test at the end. 2012-04-26 Sudarsana Nagineni [EFL] [DRT] LayoutTestController needs implementation of addOriginAccessWhitelistEntry and removeOriginAccessWhitelistEntry https://bugs.webkit.org/show_bug.cgi?id=83880 Reviewed by Antonio Gomes. Add missing implementation addOriginAccessWhitelistEntry and removeOriginAccessWhitelistEntry to EFL's LayoutTestController in order to unskip tests which rely on this. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::addOriginAccessWhitelistEntry): (LayoutTestController::removeOriginAccessWhitelistEntry): 2012-04-26 Carlos Garcia Campos [GTK] Add downloads support to MiniBrowser https://bugs.webkit.org/show_bug.cgi?id=73662 Reviewed by Philippe Normand. * MiniBrowser/gtk/BrowserDownloadsBar.c: Added. (browserDownloadsBarChildRemoved): Close the downloads bar if there aren't active downloads. (browserDownloadsBarResponse): Destroy the downloads bar when closed. (browser_downloads_bar_init): (browser_downloads_bar_class_init): (browser_downloads_bar_new): Create a new downloads bar. (actionButtonClicked): Cancel the download if cancel button was clicked, or open the downloaded file if the download finished. (browserDownloadFinalize): (browser_download_init): (browser_download_class_init): (downloadReceivedResponse): Update download message using the response URI. (remainingTime): Helper function to get human readable remaining time. (downloadProgress): Update downloads progress. (downloadReceivedData): (downloadFinished): Update download message and turn download button into open. (downloadFailed): Show error message or destroy the download UI if it was cancelled. (browserDownloadNew): Create a new download widget for the given WebKitDownload. (browser_downloads_bar_add_download): Add download widget to the downloads bar. * MiniBrowser/gtk/BrowserDownloadsBar.h: Added. * MiniBrowser/gtk/BrowserWindow.c: (downloadStarted): Create a download widget and add it to the downloads bar when a new download is started. (browserWindowConstructed): Connect to WebKitWebContext::download-started signal. * MiniBrowser/gtk/GNUmakefile.am: Add new files to compilation. 2012-04-26 Carlos Garcia Campos Unreviewed. Skip unit test failing in debug bot. * Scripts/run-gtk-tests: (TestRunner): 2012-04-26 Christophe Dumez [EFL][DRT] "resource,request,willsend" signal needs to provide more information https://bugs.webkit.org/show_bug.cgi?id=84670 Reviewed by Andreas Kling. Handle in EFL's DumpRenderTree the new Ewk_Frame_Resource_Messages now passed with the "resource,request,willsend" signal and dump information about the network request and the possible network response if LayoutTestController's dumpResourceLoadCallbacks() returns true. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (pathSuitableForTestResult): (urlSuitableForTestResult): (descriptionSuitableForTestResult): (DumpRenderTreeChrome::onWillSendRequest): 2012-04-25 Csaba Osztrogonác [Qt][WK2] Unreviewed, rolling out r113703 because of crashing tests. http://trac.webkit.org/changeset/113703 https://bugs.webkit.org/show_bug.cgi?id=83024 * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro: * Tools.pro: * qmake/mkspecs/features/features.prf: * qmake/mkspecs/features/functions.prf: 2012-04-25 Maciej Stachowiak Teach LayoutTestHelper to correctly restore the color profile when it's a factory profile https://bugs.webkit.org/show_bug.cgi?id=84931 Reviewed by Filip Pizlo. * DumpRenderTree/mac/LayoutTestHelper.m: (installLayoutTestColorProfile): Account for the possibility of the default profile being a factory profile. 2012-04-25 Sheriff Bot Unreviewed, rolling out r115240. http://trac.webkit.org/changeset/115240 https://bugs.webkit.org/show_bug.cgi?id=84928 It broke everything (Requested by Ossy_HOME on #webkit). * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkDriver): (GtkDriver._start): (GtkDriver._start.x_filter): (GtkDriver.stop): (GtkPort._driver_class): * Scripts/webkitpy/layout_tests/port/qt.py: (QtPort._build_driver): * Scripts/webkitpy/layout_tests/port/xvfbdriver.py: Removed. 2012-04-25 Dirk Pranke [chromium] change DRT's "DRT" output to match the other ports https://bugs.webkit.org/show_bug.cgi?id=84904 Reviewed by Kent Tamura. Since we run Chromium's DRT in --test-shell mode, the non-test-shell output has never been really beaten on to make sure it's right. This fixes a few issues and will be tested with upcoming NRWT changes that'll land in a separate patch. In particular we should never output framing text like "\n" or "#EOF" outside of something outside of TestEventPrinter, and we weren't handling audio output properly at all. * DumpRenderTree/chromium/TestEventPrinter.cpp: (DRTPrinter): (TestShellPrinter): (DRTPrinter::handleTextFooter): (DRTPrinter::handleAudioFooter): (DRTPrinter::handleTestFooter): (TestShellPrinter::handleAudioFooter): * DumpRenderTree/chromium/TestEventPrinter.h: (TestEventPrinter): * DumpRenderTree/chromium/TestShell.cpp: (TestShell::TestShell): (TestShell::dump): 2012-04-25 Joshua Bell Unreviewed: Adding Alec Flett as a contributor (non-committer). * Scripts/webkitpy/common/config/committers.py: 2012-04-25 Abhishek Arya Mark myself as a reviewer. https://bugs.webkit.org/show_bug.cgi?id=84902 Reviewed by Dimitri Glazkov. * Scripts/webkitpy/common/config/committers.py: 2012-04-25 Kristóf Kosztyó [Qt][NRWT] Run each DRT in it's own xvfb https://bugs.webkit.org/show_bug.cgi?id=77335 Reviewed by Dirk Pranke. * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkPort._driver_class): * Scripts/webkitpy/layout_tests/port/qt.py: (QtPort._driver_class): * Scripts/webkitpy/layout_tests/port/xvfbdriver.py: Added. (XvfbDriver): (XvfbDriver._start): (XvfbDriver._start.x_filter): (XvfbDriver.stop): 2012-04-25 Sheriff Bot Unreviewed, rolling out r115222. http://trac.webkit.org/changeset/115222 https://bugs.webkit.org/show_bug.cgi?id=84874 Caused test regressions on the EFL build (Requested by rakuco on #webkit). * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::addOriginAccessWhitelistEntry): (LayoutTestController::removeOriginAccessWhitelistEntry): 2012-04-25 Sudarsana Nagineni [EFL] [DRT] LayoutTestController needs implementation of addOriginAccessWhitelistEntry and removeOriginAccessWhitelistEntry https://bugs.webkit.org/show_bug.cgi?id=83880 Reviewed by Antonio Gomes. Add missing implementation addOriginAccessWhitelistEntry and removeOriginAccessWhitelistEntry to EFL's LayoutTestController in order to unskip tests which rely on this. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::addOriginAccessWhitelistEntry): (LayoutTestController::removeOriginAccessWhitelistEntry): 2012-04-25 Thiago Marcos P. Santos [EFL] [DRT] Reset text zoom factor before running a test https://bugs.webkit.org/show_bug.cgi?id=84861 Reviewed by Antonio Gomes. This fix eliminates most of the current flakiness on EFL build bots, although some CSS tests will now need rebasiling. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): 2012-04-25 Milian Wolff [Qt] add LayoutTestController::setPrinting support to Qt unit tests https://bugs.webkit.org/show_bug.cgi?id=84246 Reviewed by Simon Hausmann. * DumpRenderTree/qt/DumpRenderTreeQt.cpp: (WebCore::DumpRenderTree::dump): 2012-04-18 Tor Arne Vestbø [Qt] Make the web view's url property follow the active url https://bugs.webkit.org/show_bug.cgi?id=77554 Update the location bar in the minibrowser to behave a bit more like normal browsers in terms of when the url will change and how active focus is handled. Reviewed by Simon Hausmann. * MiniBrowser/qt/qml/BrowserWindow.qml: 2012-04-25 Philippe Normand Webkit build fails due to missing gstreamer include file on Kubuntu 8.04 https://bugs.webkit.org/show_bug.cgi?id=81913 Reviewed by Tor Arne Vestbø. * qmake/mkspecs/features/features.prf: Check for gstreamer-0.10 and gstreamer-plugins-base >= 0.10.30 just like the GTK port does. 0.10.27 is unfortunately not recent enough to get the media player working reliably. 2012-04-25 Xiaobo Wang [BlackBerry] Run tests in '/http/tests/local' with local URI https://bugs.webkit.org/show_bug.cgi?id=84820 Reviewed by Nikolas Zimmermann. Currently we run all tests under folder "/http/tests/" as HTTP tests (http://...). However in WebKit ORWT and NRWT scripts, tests in 'http/tests/local' are run with local path (like '/developer/LayoutTests /http/tests/local/absolute-url-strip-whitespace.html'). With this patch we can keep the same behaviour as other portings. * DumpRenderTree/blackberry/DumpRenderTree.cpp: (BlackBerry::WebKit::DumpRenderTree::isHTTPTest): 2012-04-25 Mikhail Pozdnyakov [EFL][DRT] Reset DRT settings to their default values after a testcase was executed. https://bugs.webkit.org/show_bug.cgi?id=84709 Reviewed by Eric Seidel. Not all the settings that can be changed by layout testcases through a JS API were reset in DumpRenderTreeChrome::resetDefaultsToConsistentValues(). * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (defaultEditingBehavior): returns editing behavior depending on current platform. (DumpRenderTreeChrome::resetDefaultsToConsistentValues): reset some missed DRT settings 2012-04-24 Dirk Pranke nrwt: add ASAN handling to chromium crash logs https://bugs.webkit.org/show_bug.cgi?id=84664 Reviewed by Ojan Vafai. * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumPort._get_crash_log): 2012-04-24 Dirk Pranke webkit-patch rebaseline-test is retrieving stale expectations https://bugs.webkit.org/show_bug.cgi?id=84762 Reviewed by Ojan Vafai. webkit-patch rebaseline-expectations is fetching files directly from the layout-test-results directory on the buildbots, rather than fetching them from the archive. This is problematic because we don't clobber the directory after each run and so you can end up fetching stale failing results. This change temporarily changes the code to fetch the zip file instead -- making things much slower -- until I can confer w/ abarth and ojan over the best way to fix this for the long-term. * Scripts/webkitpy/tool/commands/rebaseline.py: (RebaselineTest.__init__): (RebaselineTest._zip_file_set): (RebaselineTest): (RebaselineTest._fetch_baseline): (RebaselineTest._rebaseline_test): * Scripts/webkitpy/tool/commands/rebaseline_unittest.py: (TestRebaseline.stub_rebaseline_test_command_and_tool): (TestRebaseline.stub_rebaseline_test_command_and_tool.FakeZipFileSet): (TestRebaseline.stub_rebaseline_test_command_and_tool.FakeZipFileSet.read): (TestRebaseline.test_rebaseline_updates_expectations_file_noop): (test_rebaseline_updates_expectations_file): (test_rebaseline_test): (test_rebaseline_and_copy_test): (test_rebaseline_and_copy_test_no_existing_result): (test_rebaseline_and_copy_test_with_lion_result): (test_rebaseline_and_copy_no_overwrite_test): 2012-04-24 Ojan Vafai Improve code for embedding flakiness dashboard as an iframe https://bugs.webkit.org/show_bug.cgi?id=84782 Reviewed by Dirk Pranke. -Include the horizontal scrollbar height in the measured height of the page. -Post a message everytime the height changes instead of in response to a queryContentHeight message. This way, the embedder doesn't need to poll. No new tests. This is hard to test because it involves postMessage and iframes. Once we switch the dashboard tests over to qunit, we should be able to write an asyncTest for this. * TestResultServer/static-dashboards/flakiness_dashboard.js: (generatePage): (appendHTML): (appendExpectations): (postHeightChangedMessage): (generatePageForIndividualTests.if): (generatePageForIndividualTests): * TestResultServer/static-dashboards/flakiness_dashboard_embedded.html: Added. 2012-04-16 Ojan Vafai Replace garden-o-matic accordion UI with a panel UI https://bugs.webkit.org/show_bug.cgi?id=84081 Reviewed by Dimitri Glazkov. This is a nicer UI to work with and makes it easier to embed the flakiness dashboard, since we can reuse the same iframe without moving it in the DOM. Also, excise unnecessary jquery-isms. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results_unittests.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/styles/results.css: 2012-04-24 Peter Beverloo [Chromium] DumpRenderTree should depend on base.gyp:test_support_base https://bugs.webkit.org/show_bug.cgi?id=84729 Reviewed by Dirk Pranke. This dependency is necessary for Android since it'll soon be calling InitAndroidOSPathStub() in it's platform support. The TestWebKitAPI and webkit_unit_tests already have this dependency. * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp: 2012-04-24 Ojan Vafai Replace vague informative statement with the list of builders where the test isn't listed https://bugs.webkit.org/show_bug.cgi?id=84779 Reviewed by Dirk Pranke. This is more informative and more likely to be noticed. With the old UI, people never read the line and if they did, they didn't really understand what it meant. Hopefully have the list of builders will be more clear. * TestResultServer/static-dashboards/flakiness_dashboard.js: * TestResultServer/static-dashboards/flakiness_dashboard_tests.css: * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js: 2012-04-24 Tommy Widenflycht MediaStream API: Create a new flag for PeerConnection https://bugs.webkit.org/show_bug.cgi?id=84723 Reviewed by Dimitri Glazkov. To allow Chrome to make the GetUserMedia functionality available by default we need another flag for the PeerConnection related parts. This new flag is in addition to the general media stream flag, and both need to be enabled for PeerConnection object creation. * DumpRenderTree/chromium/TestShell.cpp: (TestShell::TestShell): 2012-04-24 Joe Mason [BlackBerry] Revert broken changes to authentication dialog https://bugs.webkit.org/show_bug.cgi?id=80135 Reviewed by Antonio Gomes. The previous patches from this bug caused an infinite loop when using digest auth; apparently they were only tested with basic. * DumpRenderTree/blackberry/DumpRenderTree.cpp: * DumpRenderTree/blackberry/DumpRenderTreeBlackBerry.h: (WebCore): 2012-04-24 Zan Dobersek [TestResultsServer] Add a unit test for changes in r114868 https://bugs.webkit.org/show_bug.cgi?id=84713 Reviewed by Ojan Vafai. Add a new unit test to test the isToTWebKit attribute of a BuilderGroup object that is created for either DEPS or ToT builder groups. The unit test covers changes made in r114868. * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js: (testBuilderGroupIsToTWebKitAttribute.g_handleBuildersListLoaded): (testBuilderGroupIsToTWebKitAttribute.assertEquals): 2012-04-24 Brady Eidson and https://bugs.webkit.org/show_bug.cgi?id=82664 Need DOMWindow mechanism to supplement UserScripts for page cache notifications Reviewed by Sam Weinig. Add a new bundle test that stresses the basics of the WKBundleDOMWindowExtension API. It makes sure new extensions can be added at the right times, and that all of the expected callbacks are made. * TestWebKitAPI/Tests/WebKit2/DOMWindowExtensionBasic.cpp: Added. (TestWebKitAPI::didReceiveMessageFromInjectedBundle): Record all messages sent from the bundle. (TestWebKitAPI::TEST): Run a page through some basic exercises, then compared messages sent from the bundle to a set of known expected messages. Listen to all the new client callbacks, create some extensions, and monitor their lifetimes: * TestWebKitAPI/Tests/WebKit2/DOMWindowExtensionBasic_Bundle.cpp: Added. (TestWebKitAPI::DOMWindowExtensionBasic::DOMWindowExtensionBasic): (TestWebKitAPI::DOMWindowExtensionBasic::frameLoadFinished): (TestWebKitAPI::DOMWindowExtensionBasic::sendExtensionStateMessage): (TestWebKitAPI::DOMWindowExtensionBasic::initialize): (TestWebKitAPI::DOMWindowExtensionBasic::didCreatePage): (TestWebKitAPI::DOMWindowExtensionBasic::updateExtensionStateRecord): (TestWebKitAPI::DOMWindowExtensionBasic::sendBundleMessage): (TestWebKitAPI::DOMWindowExtensionBasic::didCreateGlobalObjectForFrame): (TestWebKitAPI::DOMWindowExtensionBasic::willDisconnectDOMWindowExtensionFromGlobalObject): (TestWebKitAPI::DOMWindowExtensionBasic::didReconnectDOMWindowExtensionToGlobalObject): (TestWebKitAPI::DOMWindowExtensionBasic::willDestroyGlobalObjectForDOMWindowExtension): (TestWebKitAPI::didFinishLoadForFrameCallback): (TestWebKitAPI::didCreateGlobalObjectForFrameCallback): (TestWebKitAPI::willDisconnectDOMWindowExtensionFromGlobalObjectCallback): (TestWebKitAPI::didReconnectDOMWindowExtensionToGlobalObjectCallback): (TestWebKitAPI::willDestroyGlobalObjectForDOMWindowExtensionCallback): * TestWebKitAPI/Tests/WebKit2/InjectedBundleBasic.cpp: (TestWebKitAPI::didReceiveMessageFromInjectedBundle): Make this method static, as it always should have been. * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: 2012-04-24 Carlos Garcia Campos [GTK] Build and run TestWebKitAPI unit tests https://bugs.webkit.org/show_bug.cgi?id=84325 Reviewed by Philippe Normand. We only build WTF unit tests for now. WebKit2 tests will be enabled in a follow up patch. * Scripts/run-gtk-tests: (TestRunner): Add TestWebKitAPI/WTF to the list of directories containing unit tests. (TestRunner._run_test_glib): Run glib tests using gtester. (TestRunner._run_test_google): Run google tests. (TestRunner._run_test): Use _run_test_glib or _run_test_google depending on whether given test is glib or google test * TestWebKitAPI/GNUmakefile.am: Added. * TestWebKitAPI/config.h: Use #ifdef instead of # to fix a compile warning with -Wundef. Include WebKit2 headers only when building WebKit2. 2012-04-24 Allan Sandfeld Jensen [Qt][WK2] Convert touch-point area https://bugs.webkit.org/show_bug.cgi?id=84434 Reviewed by Kenneth Rohde Christiansen. Fix failing test after r114917 by also adding the new EventSender API to WebKitTestRunner. * WebKitTestRunner/EventSenderProxy.h: (EventSenderProxy): * WebKitTestRunner/InjectedBundle/Bindings/EventSendingController.idl: * WebKitTestRunner/InjectedBundle/EventSendingController.cpp: (WTR::EventSendingController::setTouchPointRadius): * WebKitTestRunner/InjectedBundle/EventSendingController.h: (EventSendingController): * WebKitTestRunner/TestController.cpp: (WTR::TestController::didReceiveSynchronousMessageFromInjectedBundle): * WebKitTestRunner/qt/EventSenderProxyQt.cpp: (WTR::EventSenderProxy::addTouchPoint): (WTR::EventSenderProxy::updateTouchPoint): (WTR::EventSenderProxy::setTouchPointRadius): (WTR::EventSenderProxy::clearTouchPoints): 2012-04-24 Mikhail Naganov [Chromium] Unreviewed test expectations update. https://bugs.webkit.org/show_bug.cgi?id=84726 Mark webkitpy.layout_tests.servers.http_server_unittest.TestHttpServer.test_start_cmd as failing on Windows. * Scripts/webkitpy/test/test_finder.py: (TestFinder._default_names): 2012-04-24 Christophe Dumez [EFL] Enable Web Timing https://bugs.webkit.org/show_bug.cgi?id=84705 Reviewed by Tony Gentilcore. Enable web timing by default whenever build-webkit is used. * Scripts/build-webkit: 2012-04-23 Eriq Augustine Return value from executed script in Chromium. https://bugs.webkit.org/show_bug.cgi?id=79851 Reviewed by Adam Barth. Adding a function to LayoutTestController that evaluates a script in an isolated wold an returns the result of the evaluation. * DumpRenderTree/LayoutTestController.cpp: (evaluateScriptInIsolatedWorldAndReturnValue): (LayoutTestController::staticFunctions): * DumpRenderTree/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp: (LayoutTestController::evaluateScriptInIsolatedWorldAndReturnValue): * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::LayoutTestController): (LayoutTestController::setXSSAuditorEnabled): (LayoutTestController::evaluateScriptInIsolatedWorldAndReturnValue): * DumpRenderTree/chromium/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::evaluateScriptInIsolatedWorldAndReturnValue): * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: (LayoutTestController::evaluateScriptInIsolatedWorldAndReturnValue): * DumpRenderTree/mac/LayoutTestControllerMac.mm: (LayoutTestController::evaluateScriptInIsolatedWorldAndReturnValue): * DumpRenderTree/qt/LayoutTestControllerQt.cpp: (LayoutTestController::evaluateScriptInIsolatedWorldAndReturnValue): * DumpRenderTree/qt/LayoutTestControllerQt.h: (LayoutTestController): * DumpRenderTree/win/LayoutTestControllerWin.cpp: (LayoutTestController::evaluateScriptInIsolatedWorldAndReturnValue): * DumpRenderTree/wx/LayoutTestControllerWx.cpp: (LayoutTestController::evaluateScriptInIsolatedWorldAndReturnValue): 2012-04-23 Dirk Pranke run-perf-tests fail intermittently with an exception https://bugs.webkit.org/show_bug.cgi?id=83532 Reviewed by Ryosuke Niwa. This bulletproofs port.get_option('child_processes') against the case where the option wasn't set. * Scripts/webkitpy/layout_tests/port/mac.py: (MacPort._get_crash_log): * Scripts/webkitpy/layout_tests/port/mac_unittest.py: (test_get_crash_log): (test_get_crash_log.fake_time_cb): 2012-04-23 Balazs Kelemen [Qt] Add desktop zooming support for QQuickWebView https://bugs.webkit.org/show_bug.cgi?id=82337 Unreviewed typo fix after r114900. * MiniBrowser/qt/BrowserWindow.cpp: (BrowserWindow::BrowserWindow): Fix the mistyped value of a zoom level. 2012-04-23 Hao Zheng [chromium-android] Get rid of the trick that replaces system fallback font with DejaVuSans. https://bugs.webkit.org/show_bug.cgi?id=84580 Reviewed by Ojan Vafai. Android's first fallback font is changed to DroidSansArabic.ttf in JB. Take the opportunity to get rid of the trick. * Scripts/webkitpy/layout_tests/port/chromium_android.py: (ChromiumAndroidPort.start_helper): (ChromiumAndroidPort._push_fonts): 2012-04-23 Dirk Pranke test-webkitpy is failing https://bugs.webkit.org/show_bug.cgi?id=84617 Reviewed by Eric Seidel. * Scripts/webkitpy/layout_tests/port/http_lock_unittest.py: (HttpLockTestWithRealFileSystem.clean_all_lockfile): 2012-04-23 Ojan Vafai Split up flakiness_dashboard.html into multiple files. This is a step in the direction of making the code more maintainable (e.g. being able to use qunit for testing). * TestResultServer/static-dashboards/dashboard_base.js: * TestResultServer/static-dashboards/flakiness_dashboard.css: Added. * TestResultServer/static-dashboards/flakiness_dashboard.html: * TestResultServer/static-dashboards/flakiness_dashboard.js: Copied from Tools/TestResultServer/static-dashboards/flakiness_dashboard.html. * TestResultServer/static-dashboards/flakiness_dashboard_tests.css: Added. * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js: Renamed from Tools/TestResultServer/static-dashboards/flakiness_dashboard_tests.js. 2012-04-23 Allan Sandfeld Jensen [Qt][WK2] Convert touch-point area. https://bugs.webkit.org/show_bug.cgi?id=84434 Reviewed by Kenneth Rohde Christiansen. Add feature to set touch-point radius for emitted touch-events. * DumpRenderTree/qt/EventSenderQt.cpp: (EventSender::addTouchPoint): (EventSender::updateTouchPoint): (EventSender::setTouchPointRadius): (EventSender::clearTouchPoints): * DumpRenderTree/qt/EventSenderQt.h: (EventSender): 2012-04-23 Christophe Dumez [EFL] EFL's LayoutTestController does not implement elementDoesAutoCompleteForElementWithId https://bugs.webkit.org/show_bug.cgi?id=84425 Reviewed by Antonio Gomes. Use the new helper in DumpRenderTreeSupportEfl to properly test whether an element performs autocompletion. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::elementDoesAutoCompleteForElementWithId): 2012-04-23 Csaba Osztrogonác master.cfg cleanup, remove dead codes after r92792. https://bugs.webkit.org/show_bug.cgi?id=84433 Relanding r114881 after typo fix in UploadBuiltProduct. Reviewed by Ryosuke Niwa. * BuildSlaveSupport/build.webkit.org-config/master.cfg: 2012-04-23 Dominik Röttsches [EFL][DRT] Need to support requestAnimationFrame https://bugs.webkit.org/show_bug.cgi?id=84585 Reviewed by Martin Robinson. Making requestAnimationFrame() support a default on the EFL build. * Scripts/build-webkit: 2012-04-23 Balazs Kelemen [Qt] Add desktop zooming support for QQuickWebView https://bugs.webkit.org/show_bug.cgi?id=82337 Reviewed by Simon Hausmann. Add zooming support for MiniBrowser to test the new API. * MiniBrowser/qt/BrowserWindow.cpp: (BrowserWindow::BrowserWindow): (BrowserWindow::zoomIn): (BrowserWindow::zoomOut): (BrowserWindow::keyPressEvent): (BrowserWindow::wheelEvent): * MiniBrowser/qt/BrowserWindow.h: (BrowserWindow): 2012-04-23 Csaba Osztrogonác Unreviewed rolling out r114881, because it broke upload buildstep. * BuildSlaveSupport/build.webkit.org-config/master.cfg: (ConfigureBuild.__init__): (ConfigureBuild.start): (CompileWebKit.start): (UploadBuiltProduct): (UploadBuiltProduct.determineExtraFeatures): (Factory.__init__): (BuildFactory.__init__): (TestFactory.__init__): (loadBuilderConfig): 2012-04-23 Csaba Osztrogonác master.cfg cleanup, remove dead codes after r92792. https://bugs.webkit.org/show_bug.cgi?id=84433 Reviewed by Ryosuke Niwa. * BuildSlaveSupport/build.webkit.org-config/master.cfg: 2012-04-23 Milian Wolff [Qt] Make DRT not dump pixel results if test is calling layoutTestController.dumpAsText() https://bugs.webkit.org/show_bug.cgi?id=84327 Reviewed by Csaba Osztrogonác. * DumpRenderTree/qt/DumpRenderTreeQt.cpp: (WebCore::DumpRenderTree::dump): 2012-04-22 Sriram Neelakandan [Gtk] Added MOZ_X11 build flag for TARGET_X11 https://bugs.webkit.org/show_bug.cgi?id=40785 Reviewed by Anders Carlsson. * GNUmakefile.am: 2012-04-22 Zan Dobersek [TestResultsServer] Impossible to inspect builds on non-Chromium builders https://bugs.webkit.org/show_bug.cgi?id=84544 Reviewed by Ryosuke Niwa. Use the groupEnum parameter that is passed through to the onBuilderListLoad function rather than an undefined attribute of the BuilderGroup object. This makes it again possible to inspect builds on non-Chromium builders in the test results server by not trying to always add Chromium revision links. * TestResultServer/static-dashboards/builders.js: 2012-04-20 Martin Robinson [GTK] Revert changes that disabled Cairo fonts metrics hinting https://bugs.webkit.org/show_bug.cgi?id=84502 Reviewed by Daniel Bates. Re-enable Cairo metrics hinting, which seems to ensure consistent results in reference and pixel test results. This makes our font metrics a bit different than the Mac port's, but it's a decent price to pay for consistent results. * DumpRenderTree/gtk/DumpRenderTree.cpp: (initializeGtkFontSettings): * WebKitTestRunner/InjectedBundle/gtk/ActivateFontsGtk.cpp: (WTR::initializeGtkSettings): 2012-04-04 Jer Noble apple.com top navigation bar appears inside video during full screen exit animation https://bugs.webkit.org/show_bug.cgi?id=83095 Reviewed by Eric Carlson. Add a new LayoutTestController function, setHasCustomFullScreenBehavior(), which allows scripts to disable the standard full screen behavior, and replace it with explicit callbacks to the relevant Document methods from within the script. This allows subtle timing bugs to be consistently reproduced from within LayoutTests. Add the new methods, hasCustomFullScreenBehavior() and setHasCustomFullScreenBehavior(): * DumpRenderTree/LayoutTestController.cpp: (setHasCustomFullScreenBehaviorCallback): (LayoutTestController::staticFunctions): * DumpRenderTree/LayoutTestController.h: (LayoutTestController::setHasCustomFullScreenBehavior): (LayoutTestController::hasCustomFullScreenBehavior): * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp: (WTR::InjectedBundlePage::enterFullScreenForElement): (WTR::InjectedBundlePage::exitFullScreenForElement): * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h: * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: * WebKitTestRunner/InjectedBundle/LayoutTestController.h: (WTR::LayoutTestController::setHasCustomFullScreenBehavior): (WTR::LayoutTestController::hasCustomFullScreenBehavior): * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::LayoutTestController): (LayoutTestController::reset): * DumpRenderTree/chromium/LayoutTestController.h: (LayoutTestController::setHasCustomFullScreenBehavior): (LayoutTestController::hasCustomFullScreenBehavior): Query the new methods before calling will/did/Enter/ExitFullScreen: * DumpRenderTree/chromium/WebViewHost.cpp: (WebViewHost::enterFullScreenNow): (WebViewHost::exitFullScreenNow): * DumpRenderTree/mac/UIDelegate.mm: (-[UIDelegate webView:enterFullScreenForElement:listener:]): (-[UIDelegate webView:exitFullScreenForElement:listener:]): 2012-04-20 Nandor Huszka NRWT: option --skip-pixel-test-if-no-baseline support on DRT https://bugs.webkit.org/show_bug.cgi?id=72841 Reviewed by Dirk Pranke This patch implements support for --skip-pixel-test-if-no-baseline generically in NRWT, instead of in every DRT and WTR implementation. We add 'should_run_pixel_test' and 'reference_files' fields to the TestInput objects, and use those flags to select the right driver to use in the DriverProxy. * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py: (SingleTestRunner.__init__): (SingleTestRunner._driver_input): (SingleTestRunner._run_reftest): * Scripts/webkitpy/layout_tests/controllers/worker.py: (Worker.handle_test_list): (Worker._update_test_input): * Scripts/webkitpy/layout_tests/models/test_input.py: (TestInput): (TestInput.__init__): (TestInput.__repr__): * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumDriver._wrapper_options): (ChromiumDriver.run_test): * Scripts/webkitpy/layout_tests/port/chromium_unittest.py: (ChromiumDriverTest.test_crash_log): * Scripts/webkitpy/layout_tests/port/driver.py: (DriverInput.__init__): (Driver.__init__): (DriverProxy.__init__): (DriverProxy.run_test): (DriverProxy.start): (DriverProxy.cmd_line): * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkDriver._start): * Scripts/webkitpy/layout_tests/port/mock_drt.py: (MockDRT.run): (MockDRT.input_from_line): (MockDRT.output_for_test): (MockDRT.write_test_output): (MockChromiumDRT.input_from_line): (MockChromiumDRT.output_for_test): (MockChromiumDRT.write_test_output): * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitDriver.cmd_line): (WebKitDriver.run_test): * WebKitTestRunner/TestController.cpp: (WTR::TestController::TestController): (WTR::TestController::initialize): (WTR::TestController::runTest): * WebKitTestRunner/TestController.h: (TestController): * WebKitTestRunner/TestInvocation.cpp: (WTR::TestInvocation::TestInvocation): (WTR::TestInvocation::setIsPixelTest): * WebKitTestRunner/TestInvocation.h: (TestInvocation): 2012-04-20 Thiago Marcos P. Santos [EFL] Make it possible to define offline storage quota per origin https://bugs.webkit.org/show_bug.cgi?id=84170 Improved LayoutTestController::setDatabaseQuota to set quota based on origin which was causing some tests to fail. Also align the default quota for testing (5MB) with others ports. When dumpDatabaseCallbacks is set to true by, also dump some database information when quota is exceed, which is also required by some tests. Reviewed by Antonio Gomes. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): * DumpRenderTree/efl/DumpRenderTreeView.cpp: (onExceededDatabaseQuota): (drtViewAdd): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::setDatabaseQuota): 2012-04-20 Sudarsana Nagineni [EFL] [DRT] Send double click event from EventSender https://bugs.webkit.org/show_bug.cgi?id=82867 Reviewed by Daniel Bates. Add support for EFL DumpRenderTree EventSender to send double click events. * DumpRenderTree/efl/EventSender.cpp: (sendMouseEvent): 2012-04-20 Sergio Villar Senin [GTK] Enable Web Timing https://bugs.webkit.org/show_bug.cgi?id=42432 Reviewed by Gustavo Noronha Silva. Build web timing by default whenever build-webkit is used. * Scripts/build-webkit: 2012-04-20 Kamil Blank [EFL] Add missing reset for setAuthorAndUserStylesEnabled inside DRT. https://bugs.webkit.org/show_bug.cgi?id=84430 Reviewed by Daniel Bates. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): 2012-04-20 Ryosuke Niwa REGRESSION(r112781): Windows build slaves are no longer including .pdb files in archives handed to testers https://bugs.webkit.org/show_bug.cgi?id=84424 Reviewed by Mark Rowe. Let Windows and Chromium have their own list of files to copy. * BuildSlaveSupport/built-product-archive: (removeDirectoryIfExists): (copyBuildFiles): (archiveBuiltProduct): 2012-04-20 Kamil Blank [EFL] Add setting API for author and user styles. https://bugs.webkit.org/show_bug.cgi?id=82860 Reviewed by Daniel Bates. Implementation of setAuthorAndUserStylesEnabled. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::setAuthorAndUserStylesEnabled): 2012-04-19 Sudarsana Nagineni [EFL] Missing keycode translation for space key https://bugs.webkit.org/show_bug.cgi?id=84153 Reviewed by Antonio Gomes. Set keyname for space key before dispatching event. * DumpRenderTree/efl/EventSender.cpp: (keyNameFromJSValue): 2012-04-19 Carlos Garcia Campos [GTK] Allow to use WebInspector in GtkLauncher https://bugs.webkit.org/show_bug.cgi?id=84143 Reviewed by Martin Robinson. * GNUmakefile.am: Add new files to compilation. * GtkLauncher/LauncherInspectorWindow.c: Added. (launcherInspectorWindowFinalize): (launcher_inspector_window_init): (launcher_inspector_window_class_init): (inspectedURIChanged): Update the window title based on current inspected URI. (showInspectorWindow): Show inspector window. (closeInspectorWindow): Hide inspector window. (launcherInspectorWindowNew): Create an inspector window. (launcherInspectorWindowGetWebView): Return the inspected web view. * GtkLauncher/LauncherInspectorWindow.h: Added. * GtkLauncher/main.c: (inspectorInspectWebViewCb): Create an inspector window. (createBrowser): Connect to inspect-web-view signal of the view inspector. 2012-04-19 Jeremy Mao Media Stream API: Piece of code cleanup for the chromium port https://bugs.webkit.org/show_bug.cgi?id=84328 Reviewed by Kent Tamura. WebUserMediaClientMock use an old requestUserMedia interface and Replace it with the new interface. * DumpRenderTree/chromium/WebUserMediaClientMock.cpp: (WebKit::WebUserMediaClientMock::requestUserMedia): Remove the old requestUserMedia and replace it with the new. * DumpRenderTree/chromium/WebUserMediaClientMock.h: (WebUserMediaClientMock): 2012-04-19 Christophe Dumez [EFL] DRT console messages should contain only the file name part of local URIs https://bugs.webkit.org/show_bug.cgi?id=84236 Reviewed by Kentaro Hara. Process console messages before printing them in EFL's DumpRenderTree them so that they contain only the file name part of local URIs. This is what is expected by test cases. * DumpRenderTree/efl/DumpRenderTreeView.cpp: (urlSuitableForTestResult): (onConsoleMessage): 2012-04-19 Thiago Marcos P. Santos [EFL] [DRT] Should create offline storage inside DUMPRENDERTREE_TEMP https://bugs.webkit.org/show_bug.cgi?id=84169 All databases were being created inside ~/.webkit which was causing flaky tests when running multiple tests in parallel, specially timeouts due to concurrency. Reviewed by Eric Seidel. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::initialize): 2012-04-19 Dominik Röttsches [GTK] 3 fast/ layout tests failing after upgrading libsoup to 2.37.92 https://bugs.webkit.org/show_bug.cgi?id=82484 Reviewed by Martin Robinson. Bumping libsoup to version 2.38.1 which fixes the above WebKit bug. * efl/jhbuild.modules: * gtk/jhbuild.modules: 2012-04-19 Lucas Forschler Move apple-xserve-10 from Lion Test to SnowLeopard Debug Build * BuildSlaveSupport/build.webkit.org-config/config.json: 2012-04-19 Christophe Dumez [EFL] dispatchDidCancelClientRedirect needs implementation https://bugs.webkit.org/show_bug.cgi?id=83669 Reviewed by Antonio Gomes. Catch the new ewk_frame "redirect,cancelled" signal in EFL's DRT and and print out information that is expected by several test cases if LayoutTestController's dumpFrameLoadCallback() returns true. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::createView): (DumpRenderTreeChrome::onFrameCreated): (DumpRenderTreeChrome::onFrameRedirectCancelled): * DumpRenderTree/efl/DumpRenderTreeChrome.h: (DumpRenderTreeChrome): 2012-04-18 Hironori Bono Split SpellChecker::didCheck() to SpellChecker::didCheckSucceeded() and SpellChecker::didCheckCanceled() https://bugs.webkit.org/show_bug.cgi?id=83748 Reviewed by Ryosuke Niwa. This change replaces a call for SpellChecker::didCheck() with one for SpellChecker::didCheckCenceled() because didCheck() has been split into didCheckSucceeded() and didCheckCanceled(). * DumpRenderTree/chromium/WebViewHost.cpp: (WebViewHost::requestCheckingOfText): Replaced didCheck with didCheckCanceled. 2012-04-18 Dirk Pranke Fix failure in test-webkitpy after r114583. Unreviewed, build fix. Updating the call site in r114583 w/o updating the test caused the test to fail. Somehow I forgot to run the unit tests :(. * Scripts/webkitpy/layout_tests/port/chromium_unittest.py: (ChromiumDriverTest.test_crash_log): 2012-04-18 Filip Pizlo JSC testing should have complete coverage over typed array types https://bugs.webkit.org/show_bug.cgi?id=84302 Reviewed by Geoff Garen. Made run-fast-jsc print something and a newline if jsc threw an exception or otherwise returned with an error code, but did not crash. * Scripts/run-fast-jsc: 2012-04-18 Dirk Pranke Fix crash in chromium.py after r114877. Unreviewed, build fix. Missed updating one call site for get_crash_log() :(. * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumDriver.run_test): 2012-04-18 Philippe Normand [GTK] NRWT bails out if more than one X server runs on the machine https://bugs.webkit.org/show_bug.cgi?id=84276 Reviewed by Dirk Pranke. Set up GTK driver's Xvfb display number based on Xorg servers currently running. * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkDriver._start): (GtkDriver._start.x_filter): 2012-04-18 Dirk Pranke run-webkit-tests picked up an old crash log https://bugs.webkit.org/show_bug.cgi?id=84196 Reviewed by Ryosuke Niwa. Modify the crash log-finding code so that we can limit our search to logs newer than X - this prevents us from finding old crash logs for a given process/pid combination even when the PIDs might've wrapped. Also fix a bug in filesystem_mock where files_under() would return paths that didn't exist. * Scripts/webkitpy/common/system/crashlogs.py: (CrashLogs.find_newest_log): (CrashLogs._find_newest_log_darwin): * Scripts/webkitpy/common/system/crashlogs_unittest.py: (CrashLogsTest.test_find_log_darwin): (CrashLogsTest.test_find_log_darwin.bad_read): (CrashLogsTest): * Scripts/webkitpy/common/system/filesystem_mock.py: (MockFileSystem.files_under): * Scripts/webkitpy/layout_tests/port/base.py: (Port._get_crash_log): * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkPort._get_crash_log): * Scripts/webkitpy/layout_tests/port/gtk_unittest.py: * Scripts/webkitpy/layout_tests/port/mac.py: (MacPort._get_crash_log): * Scripts/webkitpy/layout_tests/port/port_testcase.py: (PortTestCase.test_get_crash_log): * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitDriver.run_test): 2012-04-18 Dirk Pranke nrwt: support additional test expectations files https://bugs.webkit.org/show_bug.cgi?id=84222 Reviewed by Ojan Vafai. This patch adds an --additional-expectations flag that can be used to point to more files that will override the default set of expectations. This will allow us to remove the 'google_chrome' ports and eventually help to clean up override handling in the chromium ports and unify expectations files and Skipped files; this can also be used to specify overrides for tests that run differently on a given machine. * Scripts/webkitpy/layout_tests/port/base.py: (Port.test_expectations_overrides): * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumPort.test_expectations_overrides): * Scripts/webkitpy/layout_tests/run_webkit_tests.py: (parse_args): * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py: (MainTest.test_additional_expectations): 2012-04-18 Jer Noble Lion Production Test failing with error: "Failed to stop httpd: pid file still exists" https://bugs.webkit.org/show_bug.cgi?id=84177 Reviewed by Dirk Pranke. Try to delete the pid file left by Apache in the case it was killed forcefully. * Scripts/webkitpy/layout_tests/servers/apache_http_server.py: (LayoutTestApacheHttpd._stop_running_server): 2012-04-18 Allan Sandfeld Jensen Clean-up WheelEvent Conversion. https://bugs.webkit.org/show_bug.cgi?id=84243 Reviewed by Simon Hausmann. Calculate proper wheel-delta from pixel-delta given. * DumpRenderTree/qt/EventSenderQt.cpp: (EventSender::EventSender): (EventSender::mouseScrollBy): (EventSender::continuousMouseScrollBy): 2012-04-18 Sheriff Bot Unreviewed, rolling out r114506. http://trac.webkit.org/changeset/114506 https://bugs.webkit.org/show_bug.cgi?id=84254 Seems to be making gtk's DumpRenderTree crash occasionally (Requested by tomz on #webkit). * DumpRenderTree/gtk/DumpRenderTree.cpp: (resetDefaultsToConsistentValues): 2012-04-18 Thiago Marcos P. Santos [CMake] Use jsc target instead of ONLY_BUILD_JAVASCRIPTCORE https://bugs.webkit.org/show_bug.cgi?id=84229 CMake caches ONLY_BUILD_JAVASCRIPTCORE=1 and will reuse the setting for subsequent builds, stopping WebCore from being built ever again. Instead of trying to emulate a build target using command line options, just build the existing jsc target directly. Reviewed by Rob Buis. * Scripts/build-jsc: 2012-04-18 Simon Pena [GTK] r114021 triggered media flakyness https://bugs.webkit.org/show_bug.cgi?id=83874 Reviewed by Philippe Normand. Now that there is gtk API to update the MediaPlaybackRequiresUserGesture property, set it to false from the DumpRenderTree tool when resetting defaults to consistent values. * DumpRenderTree/gtk/DumpRenderTree.cpp: (resetDefaultsToConsistentValues): 2012-04-18 Vivek Galatage Web Inspector: Disabling Inspector causes build failure on Windows https://bugs.webkit.org/show_bug.cgi?id=83557 Reviewed by Pavel Feldman. The calls to inspector APIs are not put under ENABLE(INSPECTOR) guard. This change brings all such API calls under the guard. * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: (WTR::LayoutTestController::showWebInspector): (WTR::LayoutTestController::closeWebInspector): (WTR::LayoutTestController::evaluateInWebInspector): (WTR::LayoutTestController::setJavaScriptProfilingEnabled): * WebKitTestRunner/TestInvocation.cpp: (WTR::TestInvocation::invoke): 2012-04-18 Mario Sanchez Prada [GTK] atk/textChangedNotifications API test fails https://bugs.webkit.org/show_bug.cgi?id=84010 Unreviewed, unskipping test after bug being fixed by r114311. * Scripts/run-gtk-tests: (TestRunner): Unskip now passing test. 2012-04-17 Jacky Jiang [BlackBerry] Viewport metatag doesn't disable double-tap zoom https://bugs.webkit.org/show_bug.cgi?id=84199 Reviewed by George Staikos. Revert a part of the change master_32/SHA:fa0f8ee9 and use WebPage::setUserScalable instead. * DumpRenderTree/blackberry/DumpRenderTree.cpp: (BlackBerry::WebKit::DumpRenderTree::resetToConsistentStateBeforeTesting): 2012-04-17 Dirk Pranke build-webkit: make ninja work with chromium linux https://bugs.webkit.org/show_bug.cgi?id=84213 Reviewed by Eric Seidel. Fix typos so that build-webkit will use ninja to build chromium if a build.ninja file exists and is up-to-date. * Scripts/webkitdirs.pm: (determineIsChromiumNinja): 2012-04-17 Ojan Vafai Add caching to the flakiness dashboard json files https://bugs.webkit.org/show_bug.cgi?id=84198 Reviewed by Dirk Pranke. * TestResultServer/handlers/testfilehandler.py: (GetFile._get_file_list): (GetFile._get_file_content): (GetFile._get_file_content_from_key): (GetFile._get_test_list_json): (GetFile._serve_json): Add a Last-Modified header and handle If-Modified-Since headers. (GetFile.get): 2012-04-17 Anders Carlsson Need a client callback for when the user tries to interact with an already unresponsive page https://bugs.webkit.org/show_bug.cgi?id=84201 Reviewed by Andreas Kling. Add new client functions. * MiniBrowser/mac/BrowserWindowController.m: (-[BrowserWindowController awakeFromNib]): * WebKitTestRunner/TestController.cpp: (WTR::TestController::initialize): 2012-04-17 Dirk Pranke nrwt: handle stopping layout test helper after a ctrl-c cleanly on the chromium port https://bugs.webkit.org/show_bug.cgi?id=84190 Reviewed by Ojan Vafai. LayoutTestHelper would exit when the user typed ctrl-c, but we would attempt to write to it to shut it down, generating an IOError. * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumPort.stop_helper): 2012-04-17 Dirk Pranke rename --experimental-fully-parallel https://bugs.webkit.org/show_bug.cgi?id=84188 Reviewed by Ojan Vafai. I thought this was done long ago; rename --experimental-fully-parallel to just --fully-parallel and remove the --no- switch that AFAIK no one uses. * Scripts/webkitpy/layout_tests/controllers/manager.py: (Manager._run_tests): * Scripts/webkitpy/layout_tests/run_webkit_tests.py: (parse_args): 2012-04-17 Christophe Dumez EFL's LayoutTestController does not implement callShouldCloseOnWebView https://bugs.webkit.org/show_bug.cgi?id=82301 Reviewed by Martin Robinson. Implement callShouldCloseOnWebView in EFL's LayoutTestController by calling shouldClose() on the FrameLoader, allowing the related test to be removed from the skip list. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::callShouldCloseOnWebView): 2012-04-17 Sudarsana Nagineni [EFL] Add setting API to enable/disable XSSAuditor https://bugs.webkit.org/show_bug.cgi?id=83281 Add missing implementation setXSSAuditorEnabled to EFL's LayoutTestController in order to unskip tests in http/tests/security/xssAuditor. Also, catch the signal 'xss,detected' in DRT to enable a test, which is expecting a line containing 'didDetectXSS' in the output when reflected XSS is encountered in the page. Reviewed by Antonio Gomes. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::createView): (DumpRenderTreeChrome::onFrameCreated): (DumpRenderTreeChrome::onDidDetectXSS): * DumpRenderTree/efl/DumpRenderTreeChrome.h: (DumpRenderTreeChrome): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::setXSSAuditorEnabled): 2012-04-17 Nandor Huszka [Qt] JSC build should handle --no-webkit2 option to avoid unwanted clean-builds https://bugs.webkit.org/show_bug.cgi?id=77533 Reviewed by Simon Hausmann. Pick up the --no-webkit2 option from BUILD_WEBKIT_ARGS and pass on the corresponding parameter to the buildQMakeProjects function. * Scripts/build-jsc: 2012-04-17 Christophe Dumez [EFL] DRT should support LayoutTestController's dumpFrameLoadCallbacks() https://bugs.webkit.org/show_bug.cgi?id=83030 Reviewed by Antonio Gomes. Implement support for LayoutTestController's dumpFrameLoadCallbacks() in EFL's DumpRenderTree. Make use of the new frame argument passed to the "load,started" and "load,finished" signals. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::createView): (DumpRenderTreeChrome::onLoadStarted): (DumpRenderTreeChrome::processWork): (DumpRenderTreeChrome::topLoadingFrameLoadFinished): (DumpRenderTreeChrome::onWebViewOnloadEvent): (DumpRenderTreeChrome::onInsecureContentRun): (DumpRenderTreeChrome::onInsecureContentDisplayed): (DumpRenderTreeChrome::onFrameCreated): (DumpRenderTreeChrome::onFrameProvisionalLoad): (DumpRenderTreeChrome::onFrameLoadCommitted): (DumpRenderTreeChrome::onFrameLoadFinished): (DumpRenderTreeChrome::onFrameLoadError): * DumpRenderTree/efl/DumpRenderTreeChrome.h: (DumpRenderTreeChrome): 2012-04-17 Sheriff Bot Unreviewed, rolling out r114376. http://trac.webkit.org/changeset/114376 https://bugs.webkit.org/show_bug.cgi?id=84155 Wrong patch version committed, asked by the patch author. (Requested by rakuco on #webkit). * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::createView): (DumpRenderTreeChrome::processWork): (DumpRenderTreeChrome::onLoadFinished): (DumpRenderTreeChrome::onWillSendRequest): * DumpRenderTree/efl/DumpRenderTreeChrome.h: (DumpRenderTreeChrome): 2012-04-17 Christophe Dumez [EFL] DRT should support LayoutTestController's dumpFrameLoadCallbacks() https://bugs.webkit.org/show_bug.cgi?id=83030 Reviewed by Antonio Gomes. Implement support for LayoutTestController's dumpFrameLoadCallbacks() in EFL's DumpRenderTree. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::createView): (DumpRenderTreeChrome::processWork): (DumpRenderTreeChrome::topLoadingFrameLoadFinished): (DumpRenderTreeChrome::onWebViewOnloadEvent): (DumpRenderTreeChrome::onInsecureContentRun): (DumpRenderTreeChrome::onInsecureContentDisplayed): (DumpRenderTreeChrome::onFrameCreated): (DumpRenderTreeChrome::onFrameProvisionalLoad): (DumpRenderTreeChrome::onFrameLoadCommitted): (DumpRenderTreeChrome::onFrameLoadFinished): (DumpRenderTreeChrome::onFrameLoadError): * DumpRenderTree/efl/DumpRenderTreeChrome.h: (DumpRenderTreeChrome): 2012-04-17 Michael Brüning [Qt] Fix builds with force_static_libs_as_shared for non-linux builds. https://bugs.webkit.org/show_bug.cgi?id=83035 Reviewed by Simon Hausmann. The force_static_libs_as_shared feature for the Qt port will remove hide_symbols from Qt's CONFIG in order to make them visible to other libraries again. Due to an evaluation order problem in qmake, the flags for the C++ compiler that are set by adding hide_symbols to the CONFIG variables get applied to qmake's C++ compiler flags directly and hence trying to remove these later will not work and cause build failures when using the force_static_libs_as_shared feature of the Qt port. This is fixed by removing the flags that hide the exports from the compiler flags directly. * qmake/mkspecs/features/force_static_libs_as_shared.prf: 2012-04-16 Ryosuke Niwa Remove CreateWebKitBuildDirectory step to fix Chromium Windows Perf bot https://bugs.webkit.org/show_bug.cgi?id=84068 Reviewed by Eric Seidel. Remove CreateWebKitBuildDirectory from master.cfg since download-built-product automatically creates the build directory if needed. * BuildSlaveSupport/build.webkit.org-config/master.cfg: (DownloadAndPerfTestFactory.__init__): (DownloadAndPerfTestWebKit2Factory.__init__): 2012-04-16 Gustavo Noronha Silva [GTK] Use configure instead of autogen.sh for gtk+ https://bugs.webkit.org/show_bug.cgi?id=84112 Rubber-stamped by Martin Robinson. * gtk/jhbuild.modules: GTK+'s autogen requires gobject-introspection; we don't want to depend on that for now, so run configure instead - we can, since we are using a tarball 2012-04-16 Gustavo Noronha Silva Unreviewed, rolling out r114322. http://trac.webkit.org/changeset/114322 https://bugs.webkit.org/show_bug.cgi?id=84096 Didn't make the debug bot happy, so back out * gtk/jhbuild.modules: 2012-04-16 Gustavo Noronha Silva [GTK] Try to fix debug bot by adding gobject-introspection to the build https://bugs.webkit.org/show_bug.cgi?id=84096 Unreviewed speculative build fix. * gtk/jhbuild.modules: Added gobject-introspection, since modules need an up-to-date introspection m4 for running autogen. 2012-04-16 Gustavo Noronha Silva [GTK] Add pango to jhbuild https://bugs.webkit.org/show_bug.cgi?id=84086 Unreviewed build fix. * gtk/jhbuild.modules: Add pango's latest stable release, 1.30, to our jhbuild environment. It's required for latest stable GTK+ and some systems may lack it. 2012-04-16 Gustavo Noronha Silva [GTK] fonts jhbuild module needs an empty clean target https://bugs.webkit.org/show_bug.cgi?id=84085 Rubber-stamped by Martin Robinson. * gtk/jhbuild.modules: update to webkitgtk-test-fonts 0.0.2, which includes a phony clean target that should make warnings and errors go away. 2012-04-16 Gustavo Noronha Silva [GTK] Bump dependency on GTK+ 3.x to match reality https://bugs.webkit.org/show_bug.cgi?id=84060 Rubber-stamped by Martin Robinson. * gtk/jhbuild.modules: Bump GTK+ to 3.4.0, plus glib and gdk-pixbuf to their GNOME 3.4 versions. * gtk/patches/gtk+-3.0.12-gtkprinter.patch: Removed. This patch is applied upstream. 2012-04-16 Vincent Scheib Adding Contributors Brian Salomon and Tom Hudson to committers.py. Unreviewed. * Scripts/webkitpy/common/config/committers.py: 2012-04-16 Jonathan Dong [BlackBerry] Add ability for DRT to handle HTTP authentication challenges. https://bugs.webkit.org/show_bug.cgi?id=84031 Reviewed by Rob Buis. Implemented interface function didReceiveAuthenticationChallenge in class DumpRenderTree. * DumpRenderTree/blackberry/DumpRenderTree.cpp: (drtCredentialDescription): (BlackBerry::WebKit::DumpRenderTree::didReceiveAuthenticationChallenge): (WebKit): * DumpRenderTree/blackberry/DumpRenderTreeBlackBerry.h: (WebCore): (DumpRenderTree): 2012-04-16 Philippe Normand build-webkit --gtk --no-webkit2 builds WK2 https://bugs.webkit.org/show_bug.cgi?id=83977 Reviewed by Martin Robinson. * Scripts/build-webkit: Fixup ./configure option to disable WebKit2 build. 2012-04-16 Sudarsana Nagineni [EFL] [DRT] Feeding key events with invalid keyName https://bugs.webkit.org/show_bug.cgi?id=83797 Keep the keyName valid within scope of KeyEventInfo to avoid feeding key events with invalid keyName. Reviewed by Philippe Normand. * DumpRenderTree/efl/EventSender.cpp: (KeyEventInfo::KeyEventInfo): (KeyEventInfo): (keyPadNameFromJSValue): (keyNameFromJSValue): (createKeyEventInfo): (sendKeyDown): 2012-04-16 Andras Piroska [Qt] Replace QtXmlPatterns usage with libxslt dependency Reviewed by Simon Hausmann. The features.prf file was modified, in order to use libxml and libxslt by default. This modification only work at Qt5. * qmake/config.tests/libxml2/libxml2.cpp: Added. (main): * qmake/config.tests/libxml2/libxml2.pro: Added. * qmake/config.tests/libxslt/libxslt.cpp: Added. (main): * qmake/config.tests/libxslt/libxslt.pro: Added. * qmake/mkspecs/features/features.prf: * qmake/sync.profile: 2012-04-16 Seokju Kwon [EFL] EWebLaucher : Fix wrong parameter order in function call for multiple windows. https://bugs.webkit.org/show_bug.cgi?id=83862 Reviewed by Kenneth Rohde Christiansen. EWebLauncher creates multiple windows, like multiple tabs when pressing F9 key. It doesn't work properly because of wrong parameter order in function call. * EWebLauncher/main.c: 2012-04-16 Philippe Normand Unreviewed, GTK gardening. * Scripts/run-gtk-tests: (TestRunner): Skip /webkit/atk/textChangedNotifications. 2012-04-13 Ryosuke Niwa Chromium perf bots should not build https://bugs.webkit.org/show_bug.cgi?id=83641 Reviewed by Eric Seidel. Don't build on perf bots to improve the cycle time. * BuildSlaveSupport/build.webkit.org-config/config.json: 2012-04-14 Zalan Bujtas [WK2] Set WebKitTestRunner's WebProcess cache model to DocumentBrowser. https://bugs.webkit.org/show_bug.cgi?id=81945 Reviewed by Dan Bernstein. WebProcess's default cache model is DocumentViewer. When DocumentViewer model is on, the PageCache's capacity is set to 0. Tests that are enabling 'WebKitUsesPageCachePreferenceKey' expect PageCache to be functional. * WebKitTestRunner/TestController.cpp: (WTR::TestController::resetStateToConsistentValues): 2012-04-11 James Robinson [chromium] Remove unused compositeToTexture / compositeOffscreen setting https://bugs.webkit.org/show_bug.cgi?id=83733 Reviewed by Adrienne Walker. Remove compositeOffscreen / compositeToTexture setting. This was an experiment that is no longer needed. * DumpRenderTree/chromium/DumpRenderTree.cpp: (main): * DumpRenderTree/chromium/TestShell.cpp: (TestShell::TestShell): (TestShell::resetWebSettings): * DumpRenderTree/chromium/TestShell.h: (TestShell): * DumpRenderTree/chromium/WebPreferences.cpp: (WebPreferences::reset): (WebPreferences::applyTo): * DumpRenderTree/chromium/WebPreferences.h: (WebPreferences): 2012-04-13 Ryosuke Niwa webkitpy build fix. Revert the debug change (number of runs was 2 instead of 20), and also rebaseline the test results since we ignore the first run now. * Scripts/webkitpy/performance_tests/perftest.py: (PageLoadingPerfTest.run): * Scripts/webkitpy/performance_tests/perftest_unittest.py: (TestPageLoadingPerfTest.test_run): 2012-04-12 Ryosuke Niwa Enable SVG page loading performance tests https://bugs.webkit.org/show_bug.cgi?id=83861 Reviewed by Adam Barth. Add support for page loading tests. These are tests without any test running harnesses, and we use DriverOutput's test_timeout to measure the loading time. Extracted PerfTest.run and PerfTest.run_failed from PerfTestRunner._run_single_test. * Scripts/webkitpy/performance_tests/perftest.py: (PerfTest.run): (PerfTest): (PerfTest.run_failed): (PerfTest.parse_output): (PerfTest.output_statistics): (ChromiumStylePerfTest.parse_output): (PageLoadingPerfTest): Added. (PageLoadingPerfTest.__init__): (PageLoadingPerfTest.run): Run the driver 20 times to get the statistics for the loading time. * Scripts/webkitpy/performance_tests/perftest_unittest.py: Added. (MockPrinter): (MockPrinter.__init__): (MockPrinter.write): (MainTest): (MainTest.test_parse_output): (MainTest.test_parse_output_with_failing_line): (TestPageLoadingPerfTest): (TestPageLoadingPerfTest.assertWritten): (TestPageLoadingPerfTest.MockDriver): (TestPageLoadingPerfTest.MockDriver.__init__): (TestPageLoadingPerfTest.MockDriver.run_test): (TestPageLoadingPerfTest.test_run): (TestPageLoadingPerfTest.test_run_with_bad_output): * Scripts/webkitpy/performance_tests/perftestsrunner.py: (PerfTestsRunner): (PerfTestsRunner._collect_tests._is_test_file): Now recognizes svg tests. (PerfTestsRunner._collect_tests): (PerfTestsRunner._run_tests_set): (PerfTestsRunner._run_single_test): * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py: (test_collect_tests_with_page_load_svg): (test_collect_tests_with_page_load_svg.add_file): 2012-04-10 Brady Eidson and https://bugs.webkit.org/show_bug.cgi?id=83600 Need WebKit2 API to notify whether history loads are in the page cache Add a test ShouldGoToBackForwardListItem that exercises the new BundlePageLoaderClient callback as well as the new WKBundleBackForwardListItemIsInPageCache() API. Reviewed by Darin Adler. * TestWebKitAPI/Tests/WebKit2/ShouldGoToBackForwardListItem.cpp: Added. (TestWebKitAPI::didFinishLoadForFrame): (TestWebKitAPI::willGoToBackForwardListItem): (TestWebKitAPI::setPageLoaderClient): (TestWebKitAPI::TEST): * TestWebKitAPI/Tests/WebKit2/ShouldGoToBackForwardListItem_Bundle.cpp: Added. (ShouldGoToBackForwardListItemTest): (TestWebKitAPI::shouldGoToBackForwardListItemCallback): (TestWebKitAPI::ShouldGoToBackForwardListItemTest::ShouldGoToBackForwardListItemTest): (TestWebKitAPI::ShouldGoToBackForwardListItemTest::didCreatePage): * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp: (WTR::InjectedBundlePage::InjectedBundlePage): Update for the new client. 2012-04-13 Dominik Röttsches Group EFL bots into a category on the buildbot master https://bugs.webkit.org/show_bug.cgi?id=82712 Reviewed by Adam Roben. Creating a new category definition on the buildbot frontend. * BuildSlaveSupport/build.webkit.org-config/master.cfg: (loadBuilderConfig): Determine category for EFL bots * BuildSlaveSupport/build.webkit.org-config/templates/root.html: Add links to EFL category 2012-04-13 Sheriff Bot Unreviewed, rolling out r113714. http://trac.webkit.org/changeset/113714 https://bugs.webkit.org/show_bug.cgi?id=83887 It is behaving strange on the bots and needs more investigation. (Requested by jeez_ on #webkit). * WebKitTestRunner/qt/PlatformWebViewQt.cpp: (WTR::PlatformWebView::PlatformWebView): 2012-04-11 Jocelyn Turcotte webkit-tools-completion: Support bugzilla email completion for webkit-patch upload --cc https://bugs.webkit.org/show_bug.cgi?id=83690 Reviewed by Tor Arne Vestbø. Typing part of an email and pressing will bash-complete it using the list in Tools/Scripts/webkitpy/common/config/committers.py * Scripts/webkit-tools-completion.sh: 2012-04-13 Thiago Marcos P. Santos [EFL] Add API for color chooser https://bugs.webkit.org/show_bug.cgi?id=83692 Enables INPUT_TYPE_COLOR by default on the EFL port. Reviewed by Kenneth Rohde Christiansen. * Scripts/build-webkit: 2012-04-12 Ryosuke Niwa Extract PerfTest class from PerfTestRunner https://bugs.webkit.org/show_bug.cgi?id=83847 Reviewed by Hajime Morita. Extracted PerfTest and ChromiumStylePerfTest from PerfTestRunner. This class abstracts a test that was previously represented by a tuple. Also moved the logic to determine whether a given test is chromium style or not from run() to _collect_tests(). And moved the output parsing algorithms for parser style and chromium style tests from PerfTestRunner to PerfTest and ChromiumStylePerfTest respectively so that we may add new types of tests more easily. * Scripts/webkitpy/performance_tests/perftest.py: Added. (PerfTest): (PerfTest.__init__): (PerfTest.test_name): (PerfTest.dirname): (PerfTest.path_or_url): (PerfTest._should_ignore_line_in_parser_test_result): (PerfTest.parse_output): (ChromiumStylePerfTest): (ChromiumStylePerfTest.__init__): (ChromiumStylePerfTest.parse_output): * Scripts/webkitpy/performance_tests/perftestsrunner.py: (PerfTestsRunner._collect_tests): (PerfTestsRunner.run): (PerfTestsRunner._run_tests_set): (PerfTestsRunner._run_single_test): * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py: (run_test): (_tests_for_runner): (test_run_test_set_with_json_output): (test_run_test_set_with_json_source): (test_run_test_set_with_multiple_repositories): (_collect_tests_and_sort_test_name): (test_collect_tests): (test_collect_tests_with_skipped_list): 2012-04-12 Balazs Kelemen [Qt] Fix WebKit1 build with V8 https://bugs.webkit.org/show_bug.cgi?id=83322 Reviewed by Adam Barth. * DumpRenderTree/qt/DumpRenderTree.pro: * qmake/mkspecs/features/default_post.prf: * qmake/mkspecs/features/default_pre.prf: * qmake/mkspecs/features/features.prf: 2012-04-12 Dirk Pranke test-webkitpy fails on metered_stream_unittest for non-US West Coast Time Zones https://bugs.webkit.org/show_bug.cgi?id=83583 Reviewed by Philippe Normand Updated the unit test to ignore the hours and minutes of the localtime for time(0), since we can't easily guarantee which timezone we're in or what timezone rules were in effect on 1/1/1970; this should be mostly fine since we can trust that time.localtime() will work, although there is the possibility that we'll screw up the time formatting of which field is the hours and which is the minutes. * Scripts/webkitpy/layout_tests/views/metered_stream_unittest.py: (VerboseTest): (VerboseTest.test_basic): (VerboseTest.test_log_after_update): 2012-04-12 Dominik Rottsches Add an EFL Debug Buildbot to the master configuration https://bugs.webkit.org/show_bug.cgi?id=82711 Forgot to add EFL Linux Debug to scheduler. Reviewed by Philippe Normand. * BuildSlaveSupport/build.webkit.org-config/config.json: 2012-04-12 Dominik Röttsches Add an EFL Debug Buildbot to the master configuration https://bugs.webkit.org/show_bug.cgi?id=82711 Reviewed by Philippe Normand. Hooking up our EFL Linux 64bit debug configuration buildbot to the buildbot master. * BuildSlaveSupport/build.webkit.org-config/config.json: EFL Debug builder added 2012-04-12 George Staikos Enable the fullscreen API for BlackBerry. https://bugs.webkit.org/show_bug.cgi?id=83757 Reviewed by Antonio Gomes. * Scripts/build-webkit: 2012-04-12 Simon Pena [GTK] Test runner crashes without recent versions of pulseaudio-utils https://bugs.webkit.org/show_bug.cgi?id=83774 Reviewed by Philippe Normand. Besides checking that pactl is available, check also that its output is the expected one, and handle gracefully when it isn't, when trying to unload pulseaudio modules. * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkPort._unload_pulseaudio_module): (GtkPort._restore_pulseaudio_module): 2012-03-05 Pierre Rossi [Qt] Use QRawFont when building with Qt 5 https://bugs.webkit.org/show_bug.cgi?id=78001 Flip the QRawFont switch for Qt 5. Reviewed by Simon Hausmann. * qmake/mkspecs/features/features.prf: 2012-04-11 Vivek Galatage Web Inspector: Disabling Inspector causes build failure on Windows https://bugs.webkit.org/show_bug.cgi?id=83557 Reviewed by Pavel Feldman. The calls to inspector APIs are not put under ENABLE(INSPECTOR) guard. This change brings all such API calls under the guard. * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl: * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: (WTR): * WebKitTestRunner/InjectedBundle/LayoutTestController.h: (LayoutTestController): * WebKitTestRunner/TestInvocation.cpp: (WTR::TestInvocation::invoke): 2012-04-11 Dirk Pranke Fix regression introduced in r113927. Unreviewed, build fix. I failed to rename skipped_tests() to skipped_layout_tests() everywhere it needed to be renamed. The unit tests didn't catch this because the TestPort was one of the places that didn't get renamed :(. * Scripts/webkitpy/layout_tests/controllers/manager.py: (Manager.parse_expectations): * Scripts/webkitpy/layout_tests/port/test.py: (TestPort.skipped_layout_tests): * Scripts/webkitpy/tool/commands/queries.py: (PrintExpectations._model): 2012-04-11 Dirk Pranke new-run-webkit-tests: 'webkit-patch skipped-ports' introduced bad layering https://bugs.webkit.org/show_bug.cgi?id=47528 Reviewed by Adam Barth. The skipped-ports command has been replaced by 'webkit-patch print-expectations', which offers the equivalent functionality via 'webkit-patch print-expectations --platform \* -t skip' (and other features, of course). This also allows me to fix a nasty layering violation where the chromium port had to know about the TestExpectations objects in order to implement the command properly. Lastly, this allows me to rename skipped_tests() to skipped_layout_tests() to more consistent with skipped_perf_tests(). This patch removes skips_layout_test() from the Port interface, because that can't be implemented without the port knowing about Expectations objects (and skipped-ports was the only thing using it). * Scripts/webkitpy/layout_tests/port/base.py: (Port.skipped_layout_tests): (Port.skipped_perf_tests): * Scripts/webkitpy/layout_tests/port/base_unittest.py: (PortTest.test_default_configuration_notfound): * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumPort.test_expectations_overrides): * Scripts/webkitpy/layout_tests/port/chromium_unittest.py: (ChromiumPortTest.test_path_to_image_diff): * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitPort.skipped_layout_tests): * Scripts/webkitpy/tool/commands/queries.py: (execute): * Scripts/webkitpy/tool/commands/queries_unittest.py: (QueryCommandsTest.test_tree_status): 2012-04-11 Mark Rowe Remove a fprintf from LayoutTestController. This fprintf is causing 100+MB of data to be written to stderr during a single run of the tests. It was added for no apparent reason by Ryosuke in r113651. * DumpRenderTree/LayoutTestController.cpp: (LayoutTestController::redirectionDestinationForURL): 2012-04-11 Nico Weber [mac] Suppress deprecation warning for GlobalToLocal in the carbon parts of the test plugin. https://bugs.webkit.org/show_bug.cgi?id=83709 Reviewed by Anders Carlsson. * DumpRenderTree/TestNetscapePlugIn/main.cpp: (handleEventCarbon): 2012-04-11 Rafael Brandao Adding myself to committers.py https://bugs.webkit.org/show_bug.cgi?id=83711 Unreviewed update to committers.py. * Scripts/webkitpy/common/config/committers.py: 2012-04-11 Sudarsana Nagineni [EFL] LayoutTestController needs implementation of pageProperty https://bugs.webkit.org/show_bug.cgi?id=82449 Add missing implementation pageProperty to EFL's LayoutTestController in order to unskip printing/page-rule-selection.html Reviewed by Antonio Gomes. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::pageProperty): 2012-04-11 Sheriff Bot Unreviewed, rolling out r113836. http://trac.webkit.org/changeset/113836 https://bugs.webkit.org/show_bug.cgi?id=83705 Broke all the WebKit2 inspector tests (Requested by andersca on #webkit). * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl: * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: (WTR::LayoutTestController::showWebInspector): (WTR::LayoutTestController::closeWebInspector): (WTR::LayoutTestController::evaluateInWebInspector): (WTR::LayoutTestController::setJavaScriptProfilingEnabled): * WebKitTestRunner/InjectedBundle/LayoutTestController.h: (LayoutTestController): * WebKitTestRunner/TestInvocation.cpp: (WTR::TestInvocation::invoke): 2012-04-11 Vivek Galatage Web Inspector: Disabling Inspector causes build failure on Windows https://bugs.webkit.org/show_bug.cgi?id=83557 Reviewed by Pavel Feldman. The calls to inspector APIs are not put under ENABLE(INSPECTOR) guard. This change brings all such API calls under the guard. * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl: * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: (WTR): * WebKitTestRunner/InjectedBundle/LayoutTestController.h: (LayoutTestController): * WebKitTestRunner/TestInvocation.cpp: (WTR::TestInvocation::invoke): 2012-04-11 No'am Rosenthal [Qt][WK2] Manage graphics buffers in the web process https://bugs.webkit.org/show_bug.cgi?id=78675 Make graphics surface available only for Qt 5, attempting to fix the Qt-mac build. Unreviewed build fix. * qmake/mkspecs/features/features.prf: 2012-04-11 Philippe Normand Unreviewed, GTK NRWT follow-up fix after r113849. * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkPort._unload_pulseaudio_module): Exit early if pactl is not found. 2012-04-11 Simon Pena [GTK] media/event-attributes.html fails https://bugs.webkit.org/show_bug.cgi?id=71662 Reviewed by Philippe Normand. PulseAudio's module "module-stream-restore" allows saving the volume of a stream, restoring it the next time it runs. This affects the tests, since DumpRenderTree's volume settings are saved between test runs, and tests relying on specific volume values would miss some volumeChange events (or get additional ones). This patch hooks on the existing GtkPort setup_test_run method and creates a new clean_up_test_run method, so PulseAudio's module is unloaded (if found) before running the tests, and restored (if it was there previously) after they finished, ensuring the tests run in the right environment. * Scripts/webkitpy/layout_tests/controllers/manager.py: (Manager._clean_up_run): Invoke the port implementation of clean_up_test_run. * Scripts/webkitpy/layout_tests/port/base.py: (Port.clean_up_test_run): Add an empty implementation of clean_up_test_run. * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkPort._unload_pulseaudio_module): Unloads the offending pulseaudio module, if found. (GtkPort): (GtkPort._restore_pulseaudio_module): Restores the offending pulseaudio module, if it was there previously. (GtkPort.setup_test_run): Calls _unload_pulseaudio_module. (GtkPort.clean_up_test_run): Calls _restore_pulseaudio_module. 2012-04-11 Andras Becsi Fix the build with gcc 4.7.0 https://bugs.webkit.org/show_bug.cgi?id=83584 Reviewed by Alexey Proskuryakov. * qmake/mkspecs/features/unix/default_post.prf: The -fuse-ld=gold option is not supported with upstream gcc. 2012-04-11 Ryosuke Niwa Refactor PerfTestRunner to include file extensions in test names https://bugs.webkit.org/show_bug.cgi?id=83677 Reviewed by Hajime Morita. Move the logic that extracts the test name from _process_parser_test_result to _collect_tests. * Scripts/webkitpy/performance_tests/perftestsrunner.py: (PerfTestsRunner._collect_tests): (PerfTestsRunner._run_tests_set): (PerfTestsRunner._process_parser_test_result): Still replace "/" by ": " to keep the output format. We can make the said behavior change by simply removing "test_name = re.sub(r'\.\w+$', '', test_name)" here. (PerfTestsRunner._run_single_test): * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py: (run_test): (_tests_for_runner): (test_run_test_set): (test_run_test_set_kills_drt_per_run): (test_run_test_pause_before_testing): (test_run_test_set_for_parser_tests): (test_collect_tests): (test_collect_tests_with_skipped_list): 2012-04-11 Vivek Galatage Web Inspector: Disabling Inspector causes build failure on Windows https://bugs.webkit.org/show_bug.cgi?id=83557 Reviewed by Pavel Feldman. The calls to inspector APIs are not put under ENABLE(INSPECTOR) guard. This change brings all such API calls under the guard. * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl: * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: (WTR): * WebKitTestRunner/InjectedBundle/LayoutTestController.h: (LayoutTestController): * WebKitTestRunner/TestInvocation.cpp: (WTR::TestInvocation::invoke): 2012-04-10 Kent Tamura Unreviewed. Using tkent+wkapi@ for ChromiumPublicApi notifications. * Scripts/webkitpy/common/config/watchlist: * Scripts/webkitpy/common/config/committers.py: Add tkent+wkapi@ to pass the style checker. 2012-04-10 Dirk Pranke If NRWT gets killed halfway through a run, it incorrectly reports tests that weren't run as passes https://bugs.webkit.org/show_bug.cgi?id=82799 Reviewed by Ojan Vafai. The tests we were skipping in _mark_interrupted_tests_as_skipped had incorrectly-constructed TestResults, and so they were not being treated as failures. The JSON generator doesn't record tests that are passed or skiped in incremental_results, so you couldn't tell these tests were being skipped at all. This change changes _mark_interrupted_tests_as_skipped() to treat the skipped tests as failures; this is arguably still incorrect but is minimally invasive for now. We probably need to revisit how we are uploading information about tests that are skipped and passed altogether in a separate change. This change also cleans up a couple of other nits that were unnecessary or broken in this same code path. * Scripts/webkitpy/layout_tests/controllers/manager.py: (Manager._mark_interrupted_tests_as_skipped): (Manager._upload_json_files): Deletes the check for chromium-mac-leopard (as it is no longer necessary). * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py: (JSONLayoutResultsGenerator.__init__): Here we were recalculating the failure type unnecessarily. * Scripts/webkitpy/layout_tests/models/test_failures.py: (determine_result_type): (FailureEarlyExit): Adds a new failure type so we can tell what happened. (FailureEarlyExit.message): * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py: (MainTest.test_exit_after_n_failures_upload): Add more checks to ensure that we are recording the SKIP correctly; this is less than ideal, because we can't actually test the contents of incremental_results.json. We should fix that at some point. 2012-04-10 Dirk Pranke make chromium layout tests not depend on platform/win, platform/mac-leopard, or platform/mac-snowleopard https://bugs.webkit.org/show_bug.cgi?id=83401 Reviewed by Adam Barth. Last patch in the series - no longer look in mac-leopard, mac-snowleopard, or mac-lion. * Scripts/webkitpy/layout_tests/port/chromium_mac.py: (ChromiumMacPort): * TestResultServer/static-dashboards/flakiness_dashboard.html: 2012-04-10 Dirk Pranke webkit-patch: add a print-expectations command https://bugs.webkit.org/show_bug.cgi?id=83347 Reviewed by Adam Barth. This command will print the expected results for a given set of tests on a given set of ports matching a given set of keywords. Example output (printing all the tests not expected to pass, and their expected results, on a Mac): $ webkit-patch print-expectations -x pass fast/html // For mac-snowleopard fast/html/details-open4.html = TEXT fast/html/details-open2.html = TEXT fast/html/details-no-summary4.html = TEXT fast/html/details-open-javascript.html = TEXT $ You can also print full test-expectatons.txt lines using --full and a CSV-style report (which can be useful for post-processing) using --csv. It will replace the 'skipped-files' command (which will be removed in a subsequent patch) and is a more general (and cleaner and properly layered) solution. Also add an update() command to MockOptions() to make overriding keywords easier. * Scripts/webkitpy/layout_tests/models/test_expectations.py: (TestExpectationSerializer.to_string): (TestExpectationSerializer): (TestExpectationSerializer.to_csv): (TestExpectationSerializer._format_result): (TestExpectationLine.create_passing_expectation): (TestExpectationsModel.get_test_set_for_keyword): (TestExpectationsModel.has_keyword): (TestExpectations.model): * Scripts/webkitpy/tool/commands/queries.py: (execute): (PrintExpectations): (PrintExpectations.__init__): (PrintExpectations.execute): (PrintExpectations._filter_tests): (PrintExpectations._format_lines): (PrintExpectations._model): * Scripts/webkitpy/tool/commands/queries_unittest.py: (TestPrintExpectations): Added. * Scripts/webkitpy/tool/mocktool.py: (MockOptions): (MockOptions.__init__): (MockOptions.update): Added. 2012-04-10 Dirk Pranke webkitpy: refactor handling of --platform and related options https://bugs.webkit.org/show_bug.cgi?id=83525 Reviewed by Adam Barth. This change moves to centralize handling of --platform, --debug, --gtk, and other similar flags into a central place next to the code that actually uses those flags in PortFactory to get the right Port object. * Scripts/webkitpy/layout_tests/port/__init__.py: * Scripts/webkitpy/layout_tests/port/factory.py: (port_options): * Scripts/webkitpy/layout_tests/run_webkit_tests.py: (parse_args): * Scripts/webkitpy/tool/commands/queries.py: (PrintBaselines.__init__): 2012-04-10 Ojan Vafai Add a chromeless view to the individual tests view https://bugs.webkit.org/show_bug.cgi?id=83541 Reviewed by Daniel Bates. This is the view suitable to inlining in other contexts (e.g. garden-o-matic). * TestResultServer/static-dashboards/flakiness_dashboard.html: * TestResultServer/static-dashboards/flakiness_dashboard_tests.js: (testHtmlForIndividualTestOnAllBuilders): (testHtmlForIndividualTestOnAllBuildersWithChromeNonexistant): (testHtmlForIndividualTestOnAllBuildersWithChrome): (testHtmlForIndividualTestOnAllBuildersWithChromeWebkitMaster): (testHtmlForIndividualTests): (htmlEscape): (runTests): 2012-04-10 Dirk Pranke remove 'win' from chromium-win, chromium-linux fallback paths https://bugs.webkit.org/show_bug.cgi?id=83613 Reviewed by Ojan Vafai. The chromium ports no longer need to look at the apple 'win' directory for baseline results. Also remove the old '-gpu-' entries from the flakiness dashboard map since I'm touching lines nearby. * Scripts/webkitpy/layout_tests/port/chromium_linux.py: (ChromiumLinuxPort): * Scripts/webkitpy/layout_tests/port/chromium_win.py: (ChromiumWinPort): * TestResultServer/static-dashboards/flakiness_dashboard.html: 2012-04-10 No'am Rosenthal [Qt][WK2] Manage graphics buffers in the web process https://bugs.webkit.org/show_bug.cgi?id=78675 Enable graphics surfaces on Mac for Qt. Reviewed by Kenneth Rohde Christiansen. * qmake/mkspecs/features/features.prf: 2012-04-10 Patrick Gansterer [CMake] Enable USE_FOLDERS property https://bugs.webkit.org/show_bug.cgi?id=83571 Reviewed by Daniel Bates. Setting the FOLDER property on targets gives more structure to the generated Visual Studio solutions. This does not affect other CMake generators. * DumpRenderTree/efl/CMakeLists.txt: * EWebLauncher/CMakeLists.txt: * WinCELauncher/CMakeLists.txt: 2012-04-10 David Dorwin Add Encrypted Media Extensions methods to HTMLMediaElement https://bugs.webkit.org/show_bug.cgi?id=82971 Reviewed by Adam Barth. The extensions are behind the ENABLE(ENCRYPTED_MEDIA) feature define. Implementation is based on v0.1 of the draft proposal at http://dvcs.w3.org/hg/html-media/raw-file/tip/encrypted-media/encrypted-media.html. * DumpRenderTree/chromium/TestShell.cpp: (TestShell::TestShell): 2012-04-10 Christophe Dumez [EFL][DRT] Catch the "title,changed" signal https://bugs.webkit.org/show_bug.cgi?id=82174 Reviewed by Gustavo Noronha Silva. EFL's DumpRenderTree now catches the "title,changed" signal, and prints out the needed information. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::onTitleChanged): 2012-04-10 Christophe Dumez [EFL] DRT should support LayoutTestController's willSendRequestReturnsNull() https://bugs.webkit.org/show_bug.cgi?id=82443 Reviewed by Philippe Normand. EFL's DRT needs to support LayoutTestController's willSendRequestReturnsNull() and return NULL when expected. This allows for several tests to be removed from the skip list. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::onWillSendRequest): 2012-04-10 Sudarsana Nagineni [EFL] LayoutTestController needs implementation of addUserStyleSheet https://bugs.webkit.org/show_bug.cgi?id=82446 Adding missing implementation addUserStyleSheet to EFL's LayoutTestController so that we can unskip related tests from the skip list. Reviewed by Antonio Gomes. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::addUserStyleSheet): Implemented. 2012-04-10 Philip Rogers Adding myself to committers.py! Unreviewed update to committers.py. * Scripts/webkitpy/common/config/committers.py: 2012-04-10 Dinu Jacob [Qt][WK2] Title in MiniBrowser is not updated for a page with no title https://bugs.webkit.org/show_bug.cgi?id=82483 Reviewed by Kenneth Rohde Christiansen. Set window title to default if there is no page title. * MiniBrowser/qt/BrowserWindow.cpp: (BrowserWindow::BrowserWindow): (BrowserWindow::onTitleChanged): * MiniBrowser/qt/BrowserWindow.h: (BrowserWindow): 2012-03-23 Jesus Sanchez-Palencia [Qt][WK2] Implement PageClient::isViewWindowActive() https://bugs.webkit.org/show_bug.cgi?id=81143 Reviewed by Kenneth Rohde Christiansen. Fix the WrapperWindow from our PlatformWebView so it correctly creates the platform related stuff for QWindow, which is not created unless QWindow::setVisible() or QWindow::show() are called. * WebKitTestRunner/qt/PlatformWebViewQt.cpp: (WTR::PlatformWebView::PlatformWebView): 2012-04-10 Balazs Kelemen [Qt][WK2] TestNetscapePlugin is broken https://bugs.webkit.org/show_bug.cgi?id=83024 Reviewed by Csaba Osztrogonác. * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro: Fix X11 availability check. embedded is always true with QPA. * Tools.pro: Build the test plugin if plugins are enabled for WebKit2. * qmake/mkspecs/features/features.prf: * qmake/mkspecs/features/functions.prf: Added a convenience function to determine availability of X11 libraries. 2012-04-10 Patrick Gansterer [WIN] Fix build without precompiled header. * DumpRenderTree/cg/ImageDiffCG.cpp: Define max as max as we do for min. 2012-04-10 Arko Saha Unreviewed: adding myself to committers.py. * Scripts/webkitpy/common/config/committers.py: 2012-04-09 Kaustubh Atrawalkar Unreviewed: adding myself to committers.py. * Scripts/webkitpy/common/config/committers.py: 2012-04-09 Sheriff Bot Unreviewed, rolling out r113656. http://trac.webkit.org/changeset/113656 https://bugs.webkit.org/show_bug.cgi?id=83542 test is broken on chromium-win (Requested by simonjam on #webkit). * DumpRenderTree/chromium/WebViewHost.cpp: (WebViewHost::dispatchIntent): 2012-04-09 Takashi Toyoshima Update pywebsocket to 0.7.4. https://bugs.webkit.org/show_bug.cgi?id=83461 Reviewed by Kent Tamura. This version supports TLS by using ssl in addition to pyOpenSSL. Other functional difrerences from 0.7.2 are: - strict Sec-WebSocket-Key validation conforming RFC 4648 - compression ratio logging in debug * Scripts/webkitpy/thirdparty/__init__.py: (AutoinstallImportHook._install_pywebsocket): 2012-04-09 Dirk Pranke add a webkit-patch print-baselines command https://bugs.webkit.org/show_bug.cgi?id=83348 Reviewed by Adam Barth. This adds a simple 'print-baselines' command to webkit-patch that can be used in one of two forms: to print where the baselines for a given test (or set of tests) on a single port will be found, or to generate a report for multiple ports. Example of the former: $ webkit-patch print-baselines fast/html/keygen.html // For mac-snowleopard platform/mac/fast/html/keygen-expected.txt platform/mac/fast/html/keygen-expected.png $ The latter is most useful for finding out which fallback dirs are used by which ports for which tests, for example, 'chromium-mac-leopard' uses 100 results from 'platform/mac/leopard'. No example output is given because the reports are pretty wordy :). * Scripts/webkitpy/layout_tests/port/base.py: (Port.baseline_extensions): Added. (Port.expected_baseline_dict): Added. * Scripts/webkitpy/tool/commands/queries.py: (PrintBaselines): (PrintBaselines.__init__): (PrintBaselines.execute): (PrintBaselines._print_baselines): (PrintBaselines._platform_for_path): 2012-04-09 Florin Malita Unreviewed: adding myself to committers.py. * Scripts/webkitpy/common/config/committers.py: 2012-04-09 Ming Xie [BlackBerry] Allow simulator to use GLES2 acceleration https://bugs.webkit.org/show_bug.cgi?id=81672 Reviewed by Rob Buis. * Scripts/webkitdirs.pm: (blackberryCMakeArguments): 2012-04-09 Dana Jansens Mark myself as a committer https://bugs.webkit.org/show_bug.cgi?id=83510 Reviewed by Adrienne Walker. * Scripts/webkitpy/common/config/committers.py: 2012-04-09 Dirk Pranke nrwt is failing to upload test results on the chromium-mac-leopard bots https://bugs.webkit.org/show_bug.cgi?id=83230 Unreviewed, build fix. It looks like the change in r113399 to use the per-request timeout variable didn't actually work, so I am removing the timeout code altogether and counting on the idea that upload just isn't likely to take that long and the timeouts and retransmits are largely unused and largely unnecessary. We'll see if this breaks anywhere else. * Scripts/webkitpy/common/net/file_uploader.py: (FileUploader.__init__): (FileUploader._upload_data.callback): (FileUploader): 2012-04-09 Rob Buis [BlackBerry] Cleanup LayoutTestControllerBlackBerry.cpp https://bugs.webkit.org/show_bug.cgi?id=83469 Reviewed by Dirk Schulze. Remove WTF:: prefixes and fix a warning with the length variable. * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp: (jsStringRefToWebCoreString): (LayoutTestController::execCommand): (LayoutTestController::isCommandEnabled): (LayoutTestController::setUserStyleSheetLocation): (LayoutTestController::counterValueForElementById): (LayoutTestController::overridePreference): (LayoutTestController::setMockGeolocationError): (LayoutTestController::findString): 2012-04-09 Patrick Gansterer Build fix for WinCE after r113570. * Scripts/build-webkit: Do not use FastMalloc for WinCE. 2012-04-09 Eric Carlson Subscribe eric.carlson@apple.com to Media-related changes https://bugs.webkit.org/show_bug.cgi?id=83421 Reviewed by Dan Bernstein. * Scripts/webkitpy/common/config/watchlist: 2012-04-09 Martin Robinson check-webkit-style should ignore NULL usage in gtk_style_context_get_style https://bugs.webkit.org/show_bug.cgi?id=83412 Reviewed by David Levin. Accept NULL as an argument for gtk_style_context_get_style, because it's used as a sentinel in a variable argument list. * Scripts/webkitpy/style/checkers/cpp.py: (check_for_null): Add an exception. * Scripts/webkitpy/style/checkers/cpp_unittest.py: (WebKitStyleTest.test_null_false_zero): Add a test for the exception. 2012-04-09 Dirk Pranke webkitpy: use the filesystem object properly in fileuploader https://bugs.webkit.org/show_bug.cgi?id=83326 Reviewed by Adam Barth. We should be using the filesystem object rather than the codecs object. * Scripts/webkitpy/common/net/file_uploader.py: (FileUploader.upload_as_multipart_form_data): 2012-04-09 Eric Seidel Enable webkit_unit_tests for commit queue and EWS while tracking failures https://bugs.webkit.org/show_bug.cgi?id=83329 Unreviewed. Add a missing import to unbreak all the Chromium EWS/CQ bots. :) line 62, in _create_unit_test_results return UnitTestResults.results_from_string(results_xml) NameError: global name 'UnitTestResults' is not defined * Scripts/webkitpy/tool/bot/layouttestresultsreader.py: * Scripts/webkitpy/tool/bot/layouttestresultsreader_unittest.py: (LayoutTestResultsReaderTest.test_create_unit_test_results): 2012-04-08 Adam Barth The port object is missing unit_tests_results_path https://bugs.webkit.org/show_bug.cgi?id=83448 Unreviewed. This is causing the cr-linux-ews to throw an exception. * Scripts/webkitpy/common/config/ports.py: (DeprecatedPort.unit_tests_results_path): 2012-04-07 Patrick Gansterer [CMake] Cleanup WTF include directories https://bugs.webkit.org/show_bug.cgi?id=82716 Reviewed by Eric Seidel. * DumpRenderTree/efl/CMakeLists.txt: * WinCELauncher/CMakeLists.txt: 2012-04-07 Rob Buis [BlackBerry] Add Battery Status API support https://bugs.webkit.org/show_bug.cgi?id=82615 Reviewed by George Staikos. Enable BATTERY_STATUS. * Scripts/build-webkit: 2012-04-06 Ami Fischman Subscribe feature-media-reviews@c.o to Media-related changes. https://bugs.webkit.org/show_bug.cgi?id=83397 Reviewed by David Levin. * Scripts/webkitpy/common/config/committers.py: * Scripts/webkitpy/common/config/watchlist: 2012-04-06 Adrienne Walker [chromium] Add virtual test suite for threaded compositing https://bugs.webkit.org/show_bug.cgi?id=82263 Reviewed by James Robinson. Add compositing/visibility as a virtual test suites to test threaded compositing with. This is a reasonable smoke test of 9 tests, none of which have any expectations in the non-threaded case. To fix many of the remaining tests in this virtual test suite, there will need to be some serious refactoring to merge the conflict between the way CCScheduler wants to draw frames and the way DRT also wants to control things. So, in the short term, enable a few simple tests that are known to be passing. * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumPort.virtual_test_suites): 2012-04-06 Enrica Casucci Provide Obj-C private API to simplify markup. https://bugs.webkit.org/show_bug.cgi?id=83334 Reviewed by Sam Weinig. * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: * TestWebKitAPI/Tests/mac/SimplifyMarkup.mm: Added. (-[SimplifyMarkupTest webView:didFinishLoadForFrame:]): * TestWebKitAPI/Tests/mac/verboseMarkup.html: Added. 2012-04-06 Dirk Pranke new-run-webkit-tests: fix bugs in 'mock' drt implementation https://bugs.webkit.org/show_bug.cgi?id=83341 Reviewed by Eric Seidel. The current implementation crashed in some circumstances for reference tests and files that were missing expected results. This patch fixes those cases and adds more unit tests; it is a precursor to adding unit tests each port that actually uses this code, in order to get better coverage of the port/* implementations. * Scripts/webkitpy/layout_tests/port/mock_drt.py: (parse_options): (MockDRT.input_from_line): (MockDRT.output_for_test): (MockDRT.run_one_test): (MockChromiumDRT.input_from_line): (MockChromiumDRT.output_for_test): (MockChromiumDRT.run_one_test): * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py: (MockDRTTest.make_input_output): (MockDRTTest.expected_output): (MockDRTTest): (MockDRTTest.assertTest): (MockDRTTest.test_missing_image): (MockDRTTest.test_missing_text): (MockDRTTest.test_reftest_match): (MockDRTTest.test_reftest_mismatch): (MockChromiumDRTTest.expected_output): 2012-04-06 James Robinson Enable webkit_unit_tests for commit queue and EWS while tracking failures https://bugs.webkit.org/show_bug.cgi?id=83329 Reviewed by Adam Barth. This adds a step to parse XML output from webkit_unit_tests and consider regressions in the PatchAnalysisTask so the commit queue and EWS can reject patches that cause regressions. * Scripts/webkitpy/common/config/ports.py: (ChromiumPort.run_webkit_unit_tests_command): * Scripts/webkitpy/common/config/ports_mock.py: (MockPort.unit_tests_results_path): * Scripts/webkitpy/common/net/layouttestresults.py: (LayoutTestResults.__init__): (LayoutTestResults.failing_tests): (LayoutTestResults): (LayoutTestResults.add_unit_test_failures): * Scripts/webkitpy/common/net/unittestresults.py: Copied from Tools/Scripts/webkitpy/tool/steps/runtests_unittest.py. (UnitTestResults): (UnitTestResults.results_from_string): * Scripts/webkitpy/common/net/unittestresults_unittest.py: Added. (UnitTestResultsTest): (UnitTestResultsTest.test_nostring): (UnitTestResultsTest.test_emptystring): (UnitTestResultsTest.test_nofailures): (test_onefailure): (test_multiple_failures_per_test): * Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py: (MockCommitQueue.test_results): (MockCommitQueue.archive_last_test_results): (FailingTestCommitQueue.test_results): (test_flaky_test_failure): (test_failed_archive): * Scripts/webkitpy/tool/bot/layouttestresultsreader.py: (LayoutTestResultsReader._create_unit_test_results): (LayoutTestResultsReader.results): * Scripts/webkitpy/tool/bot/layouttestresultsreader_unittest.py: (LayoutTestResultsReaderTest.test_missing_layout_test_results): * Scripts/webkitpy/tool/bot/patchanalysistask.py: (PatchAnalysisTaskDelegate.test_results): (PatchAnalysisTaskDelegate.archive_last_test_results): (PatchAnalysisTask._test): (PatchAnalysisTask._build_and_test_without_patch): (PatchAnalysisTask._test_patch): * Scripts/webkitpy/tool/commands/earlywarningsystem.py: (AbstractEarlyWarningSystem.test_results): (AbstractEarlyWarningSystem.archive_last_test_results): * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py: (EarlyWarningSytemTest._test_testing_ews): * Scripts/webkitpy/tool/commands/queues.py: (CommitQueue.test_results): (CommitQueue.archive_last_test_results): * Scripts/webkitpy/tool/commands/queues_unittest.py: (CommitQueueTest.test_commit_queue): (test_rollout): (test_manual_reject_during_processing): * Scripts/webkitpy/tool/steps/runtests.py: (RunTests.run): * Scripts/webkitpy/tool/steps/runtests_unittest.py: (RunTestsTest.test_webkit_run_unit_tests): 2012-04-06 Aaron Colwell Enable MediaSource feature for DumpRenderTree so MediaSource LayoutTests will pass for Chromium. https://bugs.webkit.org/show_bug.cgi?id=83053 Reviewed by Dimitri Glazkov. * DumpRenderTree/chromium/TestShell.cpp: (TestShell::TestShell): 2012-04-06 George Staikos Turn on vibration support for BlackBerry. Reviewed by Rob Buis. * Scripts/build-webkit: 2012-04-06 Dan Bernstein HiDPI: Have canvas use a hidpi backing store, but downsample upon access Reviewed by Sam Weinig. * Scripts/build-webkit: Added high-dpi-canvas option to control ENABLE_HIGH_DPI_CANVAS. 2012-04-05 Dirk Pranke nrwt is failing to upload test results on the chromium-mac-leopard bots https://bugs.webkit.org/show_bug.cgi?id=83230 Reviewed by Ojan Vafai. This should fix things properly; FileUploader() was setting the socket default timeout value, and apparently that doesn't work properly with urllib. Also, the class had a bad try/finally block that was causing the exceptions to be swallowed :(. * Scripts/webkitpy/common/net/file_uploader.py: (FileUploader.__init__): (FileUploader._upload_data.callback): (FileUploader): (FileUploader._upload_data): * Scripts/webkitpy/common/net/networktransaction.py: (NetworkTimeout.__str__): (NetworkTransaction.run): * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py: (JSONResultsGeneratorBase.upload_json_files): 2012-04-05 Patrick Gansterer [Qt] Correct include paths. https://bugs.webkit.org/show_bug.cgi?id=83270 Reviewed by Eric Seidel. Modify the #include declerations so that the wtf types are included using the full path. * WebKitTestRunner/qt/TestInvocationQt.cpp: 2012-04-05 Philippe Normand gtk_unittest.GtkPortTest.test_get_crash_log failing on windows https://bugs.webkit.org/show_bug.cgi?id=83176 Reviewed by Tony Chang. Refactored the test to use os.path.join() to specify the core_pattern value of the empty crash log mock. * Scripts/webkitpy/layout_tests/port/gtk_unittest.py: (GtkPortTest): (GtkPortTest.assertLinesEqual): (GtkPortTest.test_get_crash_log): (_mock_gdb_output): 2012-04-05 Csaba Osztrogonác [Qt] Fix includes after QtDeclarative -> QtQML renaming https://bugs.webkit.org/show_bug.cgi?id=82195 Relanding r112651, because Qt5 is updated everywhere. * MiniBrowser/qt/BrowserWindow.cpp: (BrowserWindow::updateVisualMockTouchPoints): * MiniBrowser/qt/main.cpp: * WebKitTestRunner/qt/PlatformWebViewQt.cpp: (WTR::WrapperWindow::handleStatusChanged): * qmake/mkspecs/features/unix/default_post.prf: 2012-04-05 Pierre Rossi [Qt] WK1/WK2 layout test results are diverging https://bugs.webkit.org/show_bug.cgi?id=82720 The resolving of default font families was altered in Qt5. This directly affects the font picked by DRT for layout tests. Since we want to keep the Qt4 and Qt5 results in line, we can trick DRT in picking the same font as before for now. Reviewed by Csaba Osztrogonác. * DumpRenderTree/qt/QtInitializeTestFonts.cpp: (WebKit::initializeTestFonts): 2012-04-05 Kinuko Yasuda Expose DataTransferItem.getAsEntry() to allow users access dropped files as FileEntry https://bugs.webkit.org/show_bug.cgi?id=82592 Reviewed by David Levin. Added filesystem hook support in EventSender.beginDragWithFiles() for DumpRenderTree. * DumpRenderTree/chromium/EventSender.cpp: (EventSender::beginDragWithFiles): 2012-04-04 Tim Horton [mac] WKTR should always keep its windows offscreen https://bugs.webkit.org/show_bug.cgi?id=83244 Reviewed by Simon Fraser. We don't want to allow windows onto the main display; intercept attempts to change WebKitTestRunnerWindow's frame origin and prevent it from happening. Also keep around and respond with a new "fake" origin, in case someone is expecting it to change (I couldn't find any tests that expected this, however). * WebKitTestRunner/mac/PlatformWebViewMac.mm: (-[WebKitTestRunnerWindow setFrameOrigin:]): (-[WebKitTestRunnerWindow setFrame:display:animate:]): (-[WebKitTestRunnerWindow setFrame:display:]): (-[WebKitTestRunnerWindow frame]): 2012-04-04 Dirk Pranke Fix typo introduced in r113271. Unreviewed, build fix. * Scripts/webkitpy/common/net/file_uploader.py: (FileUploader._upload_data.callback): (FileUploader): 2012-04-04 Scott Graham [Chromium, DRT] Bounds check indices on gamepadController https://bugs.webkit.org/show_bug.cgi?id=83192 Invalid test data would cause bad array accesses. Add simple checks to ignore such invalid data. Reviewed by Kent Tamura. * DumpRenderTree/chromium/GamepadController.cpp: (GamepadController::connect): (GamepadController::disconnect): (GamepadController::setId): (GamepadController::setButtonCount): (GamepadController::setButtonData): (GamepadController::setAxisCount): (GamepadController::setAxisData): 2012-04-04 Dirk Pranke nrwt is failing to upload test results on the chromium-mac-leopard bots https://bugs.webkit.org/show_bug.cgi?id=83230 Reviewed by Ojan Vafai. More debugging info and a possible fix - stop messing with the default network timeout, and don't swallow URLErrors. * Scripts/webkitpy/common/net/file_uploader.py: (FileUploader._upload_data): * Scripts/webkitpy/common/net/networktransaction.py: (NetworkTransaction.run): * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py: (JSONResultsGeneratorBase.upload_json_files): 2012-04-04 Dirk Pranke nrwt is failing to upload test results on the chromium-mac-leopard bots https://bugs.webkit.org/show_bug.cgi?id=83230 Reviewed by Ojan Vafai. * Scripts/webkitpy/common/net/file_uploader.py: (FileUploader.__init__): (FileUploader._upload_data.callback): (FileUploader): * Scripts/webkitpy/layout_tests/controllers/manager.py: (Manager._upload_json_files): * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py: (JSONResultsGeneratorBase.upload_json_files): 2012-04-04 Jesus Sanchez-Palencia WKTR needs to implement layoutTestController.setPageVisibility() https://bugs.webkit.org/show_bug.cgi?id=69554 Reviewed by Simon Fraser. This patch implements the layoutTestController.setPageVisibility so WKTR supports the Page Visibility API implementation by using InjectedBundle to set the visibility state directly to Page. * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl: * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: (WTR::LayoutTestController::setPageVisibility): (WTR): (WTR::LayoutTestController::resetPageVisibility): * WebKitTestRunner/InjectedBundle/LayoutTestController.h: (LayoutTestController): 2012-04-04 Adam Barth [Chromium] TestWebKitAPI links in WebCore twice https://bugs.webkit.org/show_bug.cgi?id=83177 Reviewed by Dimitri Glazkov. This is causing link errors in the Windows component build. This patch moves the dependencies a bit lower down in the dependency diagram in the hopes of removing the duplicate symbols. * TestWebKitAPI/TestWebKitAPI.gyp/TestWebKitAPI.gyp: 2012-04-04 Christophe Dumez [EFL][DRT] Catch the "resource,request,willsend" signal https://bugs.webkit.org/show_bug.cgi?id=82192 Reviewed by Gustavo Noronha Silva. EFL's DumpRenderTree now catches the "resource,request,willsend" signal in order to implement the "Blocked access to external URL" behavior and avoid potential network-related timeouts in the tests. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::createView): (DumpRenderTreeChrome::onWillSendRequest): * DumpRenderTree/efl/DumpRenderTreeChrome.h: (DumpRenderTreeChrome): 2012-04-04 Patrick Gansterer [EFL][DRT] Correct include paths. https://bugs.webkit.org/show_bug.cgi?id=83168 Reviewed by Eric Seidel. Modify the #include declarations so that the wtf types are included using the full path. * DumpRenderTree/efl/DumpRenderTree.cpp: 2012-04-04 Philippe Normand gtk_unittest.GtkPortTest.test_get_crash_log failing on windows https://bugs.webkit.org/show_bug.cgi?id=83176 Reviewed by Tony Chang. * Scripts/webkitpy/layout_tests/port/gtk_unittest.py: (test_get_crash_log): Skip the test on non-linux platforms. 2012-04-04 Sudarsana Nagineni [EFL] DRT support for computedStyleIncludingVisitedInfo https://bugs.webkit.org/show_bug.cgi?id=82190 Adding missing implementation computedStyleIncludingVisitedInfo to EFL's LayoutTestController so that we can unskip related tests from the skip list. Reviewed by Antonio Gomes. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::computedStyleIncludingVisitedInfo): 2012-04-04 Raphael Kubo da Costa Unreviewed; update my email address once again in the committers and watch list. * Scripts/webkitpy/common/config/committers.py: Prefer my webkit.org email over my FreeBSD.org one. * Scripts/webkitpy/common/config/watchlist: Ditto. 2012-04-04 Jesus Sanchez-Palencia [Qt][WK2] QtWebKit2 should support Page Visibility API https://bugs.webkit.org/show_bug.cgi?id=81164 Reviewed by Kenneth Rohde Christiansen. Enabling Page Visiblity API for Qt WebKit2. * qmake/mkspecs/features/features.prf: 2012-04-04 Andras Becsi [Qt][WK2] Make the WebView a subclass of Flickable https://bugs.webkit.org/show_bug.cgi?id=83033 Reviewed by Kenneth Rohde Christiansen. Remove workaround in MiniBrowser and the ScrollIndicator now that the WebView is a Flickable. * MiniBrowser/qt/MiniBrowser.pro: * MiniBrowser/qt/qml/BrowserWindow.qml: * MiniBrowser/qt/qml/ScrollIndicator.qml: * WebKitTestRunner/Target.pri: 2012-04-04 Csaba Osztrogonác REGRESSION(r113144): It made run-webkit-tests always fail https://bugs.webkit.org/show_bug.cgi?id=83142 Unreviewed trivial fix after r113144. * Scripts/run-webkit-tests: 2012-04-03 Ryosuke Niwa I really hope this is the last build fix. Apparently exec doesn't quite arguments with double quotations on Windows. Use system instead. Hopefully, this won't break anything. * Scripts/run-webkit-tests: 2012-04-03 Keishi Hattori Disable ENABLE_DATALIST for now https://bugs.webkit.org/show_bug.cgi?id=82871 Reviewed by Kent Tamura. * Scripts/build-webkit: Disabled ENABLE_DATALIST. 2012-04-03 Ryosuke Niwa Sigh... this isn't going well. I can't even reproduce this issue locally. Another speculative fix. * Scripts/run-webkit-tests: 2012-04-03 Ryosuke Niwa I don't know why I'm so sloppy today. Another build fix. * Scripts/run-webkit-tests: 2012-04-03 Ryosuke Niwa Hopefully the last build fix for Chromium Windows. Explicitly execute new-run-webkit-tests and old-run-webkit-tests by python and perl. * Scripts/run-webkit-tests: 2012-04-03 Dirk Pranke [Chromium] Lots of timeouts causing Mac10.6 to exit early. https://bugs.webkit.org/show_bug.cgi?id=83076 Unreviewed, build fix (slightly reviewed by Simon Fraser and Eric Seidel, but not approved). Add logic to the apple mac and chromium mac code to not use too many workers; it looks like the xserves (and possibly mac pros) count hyperthreaded cores when they really shouldn't and we end up using too many workers at a time; this leads to tests thrashing and timing out. This change is a temporary fix to make the bots happy while I look into more profiling and other fixes. * Scripts/webkitpy/layout_tests/port/mac.py: (ChromiumMacPort.default_child_processes): * Scripts/webkitpy/layout_tests/port/chromium_mac.py: (ChromiumMacPort.default_child_processes): 2012-04-03 Simon Fraser WebKitTestRunner's EventSender is leaky https://bugs.webkit.org/show_bug.cgi?id=83099 Reviewed by Beth Dakin. Fix leaks of WKNumbers by using the adoptWK idiom, and make all the code consistent. * WebKitTestRunner/InjectedBundle/EventSendingController.cpp: (WTR::EventSendingController::mouseDown): (WTR::EventSendingController::mouseUp): (WTR::EventSendingController::mouseMoveTo): (WTR::EventSendingController::leapForward): (WTR::EventSendingController::keyDown): (WTR::EventSendingController::mouseScrollBy): (WTR::EventSendingController::addTouchPoint): (WTR::EventSendingController::updateTouchPoint): (WTR::EventSendingController::setTouchModifier): (WTR::EventSendingController::releaseTouchPoint): (WTR::EventSendingController::cancelTouchPoint): 2012-04-03 Ryosuke Niwa Another build fix. Don't use old-run-webkit-tests on Chromium Windows bots. * Scripts/run-webkit-tests: (useNewRunWebKitTests): 2012-04-03 Ryosuke Niwa kill-old-processes should kill apache2 and httpd https://bugs.webkit.org/show_bug.cgi?id=83065 Reviewed by Tony Chang. Kill apache2 and httpd on Mac and Linux (Windows uses lighttpd) to free up port 8000. * BuildSlaveSupport/kill-old-processes: 2012-04-03 Tony Chang run the same test steps on test only bots as on build-and-test bots https://bugs.webkit.org/show_bug.cgi?id=83090 Reviewed by Ryosuke Niwa. * BuildSlaveSupport/build.webkit.org-config/master.cfg: (TestFactory.__init__): 2012-04-03 Ami Fischman Add a Media watchlist. https://bugs.webkit.org/show_bug.cgi?id=83071 Reviewed by David Levin. * Scripts/webkitpy/common/config/watchlist: 2012-04-03 Simon Fraser Fix large leak in WebKitTestRunner https://bugs.webkit.org/show_bug.cgi?id=83084 Reviewed by Beth Dakin. Fix a leak of the bimap backing store created when doing pixel and ref tests. This leak was causing serious thrash on the test bots. We don't need to allocate memory for CGBitmapContextCreate(); if we pass NULL, it will allocate and manage its own backing store. * WebKitTestRunner/cg/TestInvocationCG.cpp: (WTR::createCGContextFromImage): 2012-04-03 Ryosuke Niwa Another build fix after r113067. Close the zipfile after extracting all files so that we can remove it. * BuildSlaveSupport/built-product-archive: (unzipArchive): 2012-04-03 Ojan Vafai Give more human friendly error message when builders fail to load or have stale data. https://bugs.webkit.org/show_bug.cgi?id=83058 Reviewed by Eric Seidel. * TestResultServer/static-dashboards/dashboard_base.js: (addError): (addBuilderLoadErrors): (handleLocationChange): * TestResultServer/static-dashboards/flakiness_dashboard_tests.js: 2012-04-03 Ryosuke Niwa Again, a build fix after r113067. Don't call r113067 twice on the configuration build directory. * BuildSlaveSupport/built-product-archive: (extractBuiltProduct): 2012-04-03 Ryosuke Niwa Yet another build fix after r113067. Don't trigger Chromium testers right away since they're now triggered by builders. * BuildSlaveSupport/build.webkit.org-config/config.json: 2012-04-03 Ryosuke Niwa Another build fix after r113067 for Mac port. It turned out that archiving the entire build directory doesn't work. * BuildSlaveSupport/built-product-archive: (createZipManually): (createZip): (archiveBuiltProduct): 2012-04-03 Mike Lattanzio [BlackBerry] Expose CaseSensitive, Wrap, and HighlightAllMatches in WebPage::findNextString() https://bugs.webkit.org/show_bug.cgi?id=82643 Update LayoutTestController to accomodate the new find API. It now provides caseSensitive functionality to DRT. Internal Review by Andy Chen. Reviewed by Rob Buis. * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp: (LayoutTestController::findString): 2012-04-03 Ryosuke Niwa Another build fix after r113067. CreateWebKitBuildDirectory step is no longer needed because download-built-product creates the build directory as needed. This step fails on Chromium Windows due to -p option not supported by Windows' native mkdir. * BuildSlaveSupport/build.webkit.org-config/master.cfg: (TestFactory.__init__): 2012-04-03 Ryosuke Niwa Build fix after r113067. Don't delete the build directory. * BuildSlaveSupport/built-product-archive: (extractBuiltProduct): 2012-04-02 Ojan Vafai Load all builder lists from the buildbot json https://bugs.webkit.org/show_bug.cgi?id=82998 Reviewed by Adam Barth. Also, remove an unnecessary list of test types. Now we don't hard-code builder names anywhere and we only have a single hard-coded list of test types. * TestResultServer/static-dashboards/builders.js: (BuilderMaster.prototype.logPath): (BuilderMaster.prototype.builderJsonPath): (requestBuilderList.xhr.onload): (requestBuilderList.xhr.onerror): (isChromiumDepsFyiGpuTestRunner): (isChromiumTipOfTreeGpuTestRunner): (isChromiumDepsGTestRunner): (isChromiumDepsCrosGTestRunner): (isChromiumTipOfTreeGTestRunner): (onBuilderListLoad): (loadBuildersList): * TestResultServer/static-dashboards/dashboard_base.js: (parseCrossDashboardParameters): (currentBuilderGroupCategory): * TestResultServer/static-dashboards/flakiness_dashboard.html: * TestResultServer/static-dashboards/flakiness_dashboard_tests.js: (testHtmlForTestsWithExpectationsButNoFailures): (testGenerateChromiumTipOfTreeGpuBuildersFromBuilderList): (testGenerateChromiumDepsGTestBuildersFromBuilderList): (testGenerateChromiumDepsCrosGTestBuildersFromBuilderList): (testGenerateChromiumTipOfTreeGTestBuildersFromBuilderList): * TestResultServer/static-dashboards/timeline_explorer.html: 2012-04-02 Ryosuke Niwa Chromium testers should extract builds instead of building on their own https://bugs.webkit.org/show_bug.cgi?id=82996 Reviewed by Tony Chang. Make Chromium builders triggger Chromium testers and make testers download and extract builds from the buildbot master instead of building binaries on their own. Add download-built-product to wrap curl used in the download-built-product step since Windows does not provide "curl" natively. * BuildSlaveSupport/build.webkit.org-config/config.json: Add new triggerables for Chromium testers and make Chromium builders trigger them. Also change the type of Chromium testers from NewBuildAndTest to Test. * BuildSlaveSupport/build.webkit.org-config/master.cfg: (DownloadBuiltProduct): Call download-built-product instead of curl. * BuildSlaveSupport/built-product-archive: (createZip): Don't zip the parent configuration build directory like "release" and "debug". While Mac port needs this behavior for compatibility reasons, Chromium Mac port doesn't want this behavior. (archiveBuiltProduct): Zip the configuration build directory on Mac port. (unzipArchive): Extracted. Use ditto on Mac, unzip on linux and cygwin, and zipfile package on Windows. (extractBuiltProduct): Refactor the code to use removeDirectoryIfExists and unzipArchive. Support Chromium. * BuildSlaveSupport/download-built-product: Added to wrap curl which isn't available on Windows. 2012-04-03 Martin Robinson [GTK] Tools/Scripts/run-gtk-tests should not force you to use the WebKit jhbuild https://bugs.webkit.org/show_bug.cgi?id=82473 Reviewed by Philippe Normand. Only use jhbuild if WebKitBuild/Dependencies exists (if update-webkitgtk-libs) was ever run. * gtk/run-with-jhbuild: Instead of using jhbuild unconditionally, first check if the user has ever run update-gtk-libs and, if so, then use jhbuild. 2012-04-03 Sudarsana Nagineni [EFL] DRT support for setInteractiveFormValidationEnabled https://bugs.webkit.org/show_bug.cgi?id=82050 Enable interactive form validation and unskip tests from the skip list. Reviewed by Antonio Gomes. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): 2012-04-03 Philippe Normand Unreviewed, another test-webkitpy fix-up after r113037. * Scripts/webkitpy/layout_tests/port/gtk_unittest.py: Adapt mock crash dump depending on current environment. 2012-04-03 Philippe Normand Unreviewed, test-webkitpy build fix after r113037. * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkPort._get_crash_log): 2012-04-03 Philippe Normand [GTK] crash log reports support https://bugs.webkit.org/show_bug.cgi?id=81659 Reviewed by Martin Robinson. Removed the daemontools crashmon/xvfb scripts and implemented the crash log reporting in the NRWT Gtk port. To get proper crash logs one needs to set the core pattern like this: echo "/path/to/cores/core-pid_%p-_-process_%e" > /proc/sys/kernel/core_pattern Then enable coredumps with "ulimit -c unlimited" and set the WEBKIT_CORE_DUMPS_DIRECTORY environment variable. * BuildSlaveSupport/gtk/README: * BuildSlaveSupport/gtk/crashmon/crashmon: Removed. * BuildSlaveSupport/gtk/crashmon/log/run: Removed. * BuildSlaveSupport/gtk/crashmon/run: Removed. * BuildSlaveSupport/gtk/daemontools-buildbot.conf: * BuildSlaveSupport/gtk/xvfb/log/run: Removed. * BuildSlaveSupport/gtk/xvfb/run: Removed. * Scripts/new-run-webkit-tests: * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkDriver.stop): (GtkPort.show_results_html_file): (GtkPort): (GtkPort._get_gdb_output): (GtkPort._get_crash_log): (GtkPort._get_crash_log.match_filename): * Scripts/webkitpy/layout_tests/port/gtk_unittest.py: (GtkPortTest): (test_show_results_html_file): (assertLinesEqual): (_mock_gdb_output): (test_get_crash_log): 2012-04-02 Jocelyn Turcotte Enable and connect the WebInspectorServer with WebKit2 pages. https://bugs.webkit.org/show_bug.cgi?id=73094 Reviewed by Simon Hausmann. Enable developer extras on pages in MiniBrowser for Qt. * MiniBrowser/qt/qml/BrowserWindow.qml: 2012-04-03 Christophe Dumez [EFL] LayoutTestController needs implementation of setDefersLoading https://bugs.webkit.org/show_bug.cgi?id=82890 Reviewed by Philippe Normand. Implement setDefersLoading() in EFL's LayoutTestController so that the corresponding test case can be removed from the skip list. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::setDefersLoading): 2012-04-03 Christophe Dumez [EFL] LayoutTestController needs implementation of goBack https://bugs.webkit.org/show_bug.cgi?id=81914 Reviewed by Philippe Normand. Implement goBack() in EFL's LayoutTestController by calling ewk_frame_back(). * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::goBack): 2012-04-03 Alexander Shalamov [EFL] LayoutTestController needs implementation of pageSizeAndMarginsInPixels https://bugs.webkit.org/show_bug.cgi?id=82589 Add missing implementation for pageSizeAndMarginsInPixels to EFL's LayoutTestController in order to unskip printing/page-format-data.html Reviewed by Philippe Normand. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::pageSizeAndMarginsInPixels): 2012-04-03 Simon Pena [GTK] DRT missing didRunInsecureContent notification https://bugs.webkit.org/show_bug.cgi?id=59367 Reviewed by Philippe Normand. Connect DumpRenderTree to WebFrame's "insecure-content-run", in both the main WebFrame or any other WebFrame created later. Added "didRunInsecureContent" notification in the callback in order to get the LayoutTests passing. * DumpRenderTree/gtk/DumpRenderTree.cpp: Connect to WebFrame's "insecure-content-run" signal and add didRunInsecureContent notification in the callback to get the LayoutTests passing. (didRunInsecureContent): (createWebView): 2012-04-03 Raphael Kubo da Costa rebaseline-server: Make it possible to not launch a browser with the server. https://bugs.webkit.org/show_bug.cgi?id=82999 Reviewed by Adam Barth. Add the `--no-show-results' option, which lets one decide not to launch a web browser when running the rebaseline server. This is especially useful if the default browser is not the wanted one, or if the rebaseline server page is already open. * Scripts/webkitpy/tool/commands/abstractlocalservercommand.py: (AbstractLocalServerCommand.__init__): (AbstractLocalServerCommand.execute): 2012-04-02 Mikhail Pozdnyakov EFL's LayoutTestController disableImageLoading implementation. https://bugs.webkit.org/show_bug.cgi?id=82848 Reviewed by Hajime Morita. * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::resetDefaultsToConsistentValues): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::disableImageLoading): 2012-04-02 Alexander Shalamov [EFL] LayoutTestController needs implementation of isPageBoxVisible https://bugs.webkit.org/show_bug.cgi?id=82591 Add missing implementation to isPageBoxVisible to EFL's LayoutTestController in order to unskip printing/page-format-data.html Reviewed by Hajime Morita. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::isPageBoxVisible): 2012-04-02 Simon Fraser run-webkit-tests with a relative --root causes tests to fail because DYLD_LIBRARY_PATH is not set https://bugs.webkit.org/show_bug.cgi?id=82962 Reviewed by Dirk Pranke. Ensure that _build_path() returns an absolute path. Eric Seidel also had to deploy MockConfig in a bunch of places in order to correct previous testing errors where we were pretending that "Mock Output from child process" (returned by MockExecutive.run_command) was a real path. The real Config object calls run_command("webkit-build-directory") to read the WebKit build directory from the webkitdirs.pm perl code. MockConfig abstracts this away and always returns "/mock-build" during testing. This change is much larger than one would think necessary because of needing to deploy this MockConfig class. * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitPort._build_path): 2012-04-02 Ojan Vafai Fix snafu in r112971. We were never calling parseParameter for builder. * TestResultServer/static-dashboards/dashboard_base.js: 2012-04-02 Ojan Vafai Generate the lists of all layout test builders from the buildbot json https://bugs.webkit.org/show_bug.cgi?id=82924 Reviewed by Mihai Parparita. In order to make this work, cleaned up a lot of existing technical debt. -Got rid of expectations builder. This concept is outdated and just dead code. -Split hash parameter parsing into two functions. One for dashboard_base (crossDashboardParameters) and one for the specific dashboard html file (dashboardSpecificParameters). In the old world, parseParameters needed to be called twice and depended on it's own output the first time through. Now we only need to parse crossDashboardParameters first and crossDashboardParameters doesn't depend on the output of crossDashboardParameters. -Lots of variable/method renames due to the above. -g_defaultDashboardSpecificStateValues now has to list all possible hash parameters for that dashboard. * TestResultServer/static-dashboards/aggregate_results.html: * TestResultServer/static-dashboards/builders.js: (BuilderGroup): (BuilderGroup.prototype.setup): (jsonRequest.xhr.onload): (jsonRequest.xhr.onerror): (isWebkitTestRunner): (isChromiumWebkitTipOfTreeTestRunner): (isChromiumWebkitDepsTestRunner): (generateBuildersFromBuilderList): (onLayoutTestBuilderListLoad): (onErrorLoadingBuilderList): (loadBuildersList): * TestResultServer/static-dashboards/dashboard_base.js: (handleValidHashParameterWrapper): (queryHashAsMap): (parseParameter): (parseCrossDashboardParameters): (parseDashboardSpecificParameters): (parseParameters): (diffStates): (defaultValue): (isLayoutTestResults): (isGPUTestResults): (currentBuilderGroupCategory): (currentBuilderGroup): (initBuilders): Now that we've split parameter parseing, these methods no longer need to take an optional state. They can always just use the global cross-dashboard state. (pathToBuilderResultsFile): (appendJSONScriptElementFor): (appendJSONScriptElements): (handleResourceLoadError): (haveJsonFilesLoaded): (combinedDashboardState): (setQueryParameter): (permaLinkURLHash): (toggleQueryParameter): (queryParameterValue): (selectHTML): (htmlForTestTypeSwitcher): (g_handleBuildersListLoaded): * TestResultServer/static-dashboards/flakiness_dashboard.html: * TestResultServer/static-dashboards/flakiness_dashboard_tests.js: (testPlatformAndBuildType): (testSubstringList): (testHtmlForTestTypeSwitcherGroup): (testLookupVirtualTestSuite): (testBaseTest): (generateBuildersFromBuilderListHelper): (testGenerateChromiumWebkitTipOfTreeBuildersFromBuilderList): (testGenerateChromiumWebkitDepsBuildersFromBuilderList): (assertObjectsDeepEqual): (testQueryHashAsMap): (testDiffStates): * TestResultServer/static-dashboards/timeline_explorer.html: * TestResultServer/static-dashboards/treemap.html: 2012-04-02 Tony Chang check-webkit-style errors when removing .png files https://bugs.webkit.org/show_bug.cgi?id=82933 Reviewed by David Levin. * Scripts/webkitpy/style/patchreader.py: (PatchReader.check): Make sure the file exists and pass in a FileSystem() object (for mocking). * Scripts/webkitpy/style/patchreader_unittest.py: (test_check_patch_with_png_deletion): 2012-04-02 Sheriff Bot Unreviewed, rolling out r112948. http://trac.webkit.org/changeset/112948 https://bugs.webkit.org/show_bug.cgi?id=82961 Someone else already checked in a similar change (Requested by sundiamonde on #webkit). * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitDriver._start): 2012-04-02 Adam Barth garden-o-matic isn't able to rebaseline audio failures https://bugs.webkit.org/show_bug.cgi?id=82957 Reviewed by Chris Rogers. This patch just adds "wav" to the list of test suffixes and updates the unit tests to show that we're rebaselining audio tests results now too. * Scripts/webkitpy/tool/commands/rebaseline.py: * Scripts/webkitpy/tool/commands/rebaseline_unittest.py: (test_rebaseline_updates_expectations_file): 2012-04-02 Adam Barth garden-o-matic should let you listen to audio failures https://bugs.webkit.org/show_bug.cgi?id=82953 Reviewed by Chris Rogers. Now that we've actually got an audio failure on the bots, we can clean up the last stray bugs. This patch doesn't have any tests because I'm lame. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/garden-o-matic.html: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/results.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results.js: (.): 2012-04-02 Stephanie Lewis run-webkit-tests --root fails if /usr/local/lib/libWebCoreTestSupport.dylib is not installed. https://bugs.webkit.org/show_bug.cgi?id=82552 Reviewed by Dirk Pranke. Use DYLD_LIBRARY_PATH so we pick up the libWebCoreTestSupport.dylib located in the root. * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitDriver._start): 2012-04-02 Dirk Pranke NRWT is not printing out the builder it's uploading JSON files for https://bugs.webkit.org/show_bug.cgi?id=82834 Reviewed by Ojan Vafai. Handle log messages of the form log.info("%s", arg) properly ... I didn't even know you could do that :). * Scripts/webkitpy/layout_tests/views/metered_stream.py: (_LogHandler.emit): * Scripts/webkitpy/layout_tests/views/metered_stream_unittest.py: (RegularTest.test_log_args): (VerboseTest.test_log_args): 2012-04-02 Tim Horton [mac] LayoutTestHelper crashes if there is no main display https://bugs.webkit.org/show_bug.cgi?id=82944 Reviewed by Simon Fraser. If there's no main display attached, ColorSyncDeviceCopyDeviceInfo returns a null dictionary, so we shouldn't go ahead and try to read from it. * DumpRenderTree/mac/LayoutTestHelper.m: (installLayoutTestColorProfile): 2012-04-02 Balazs Kelemen [Qt][WK2] Set up plugin process on Unix https://bugs.webkit.org/show_bug.cgi?id=72121 Reviewed by Simon Hausmann. * qmake/mkspecs/features/features.prf: Reenable plugins and turn on plugin process. 2012-04-02 Sheriff Bot Unreviewed, rolling out r112868, r112879, and r112881. http://trac.webkit.org/changeset/112868 http://trac.webkit.org/changeset/112879 http://trac.webkit.org/changeset/112881 https://bugs.webkit.org/show_bug.cgi?id=82901 "Build fail on bots." (Requested by kbalazs on #webkit). * MiniBrowser/gtk/GNUmakefile.am: * qmake/mkspecs/features/features.prf: 2012-04-02 Philippe Normand [GTK] Enable shadow-dom in build-webkit. Rubber-stamped by Gustavo Noronha Silva. * Scripts/build-webkit: Enable shadow-dom build, this is need to have a working build currently. 2012-04-02 Andreas Kling Adding Zalan Bujtas to committers list. * Scripts/webkitpy/common/config/committers.py: 2012-04-02 Balazs Kelemen [Qt][WK2] Set up plugin process on Unix https://bugs.webkit.org/show_bug.cgi?id=72121 Reviewed by Simon Hausmann. * qmake/mkspecs/features/features.prf: Reenable plugins and turn on plugin process. 2012-04-02 Sheriff Bot Unreviewed, rolling out r112651. http://trac.webkit.org/changeset/112651 https://bugs.webkit.org/show_bug.cgi?id=82887 It doesn't work with older Qt5 (Requested by Ossy on #webkit). * MiniBrowser/qt/BrowserWindow.cpp: (BrowserWindow::updateVisualMockTouchPoints): * MiniBrowser/qt/main.cpp: * WebKitTestRunner/qt/PlatformWebViewQt.cpp: (WTR::WrapperWindow::handleStatusChanged): * qmake/mkspecs/features/unix/default_post.prf: 2012-04-02 Carlos Garcia Campos [GTK] Fix names of failed unit tests in Tools/Scripts/run-gtk-tests https://bugs.webkit.org/show_bug.cgi?id=82877 Reviewed by Philippe Normand. * Scripts/run-gtk-tests: (TestRunner.run_tests): Use replace instead of lstrip to remove the programs_path from the full path of unit tests. 2012-04-02 Carlos Garcia Campos Unreviewed. Skip flaky unit tests. * Scripts/run-gtk-tests: (TestRunner): 2012-04-02 Ryosuke Niwa webkitpy rebaseline. * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py: (test_run_test_set_with_json_output): (test_run_test_set_with_json_source): 2012-04-01 Ryosuke Niwa perf-o-matic should store test results' units https://bugs.webkit.org/show_bug.cgi?id=82852 Reviewed by Kentaro Hara. Include units in the results JSON. * Scripts/webkitpy/performance_tests/perftestsrunner.py: (PerfTestsRunner._process_chromium_style_test_result): (PerfTestsRunner._process_parser_test_result): 2012-04-01 Tony Tseung Composite Font References is a new established standard (ISO/IEC 14496-28:2012) for specifying composite fonts from existing physical fonts. https://bugs.webkit.org/show_bug.cgi?id=82810 A new test font of this kind as been added to the test tools for running the webkit-tests Reviewed by Dan Bernstein. * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Added new dependency SampleFont.sfont * DumpRenderTree/fonts/SampleFont.sfont: Added. * DumpRenderTree/mac/Configurations/DumpRenderTree.xcconfig: * DumpRenderTree/mac/DumpRenderTree.mm: (allowedFontFamilySet): Added the Composite Font Referene sample font "Hiragino Maru Gothic Monospaced" entry to the fonts white-list (activateTestingFonts): Added the registration of the Composite Font Referene sample font * WebKitTestRunner/InjectedBundle/mac/ActivateFonts.mm: (WTR::allowedFontFamilySet): Added the Composite Font Referene sample font "Hiragino Maru Gothic Monospaced" entry to the fonts white-list (WTR::activateFonts): Added the registration of the Composite Font Referene sample font * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj: Added new dependency SampleFont.sfont * WebKitTestRunner/fonts/SampleFont.sfont: Added. 2012-04-01 Gyuyoung Kim Support the Network Information API https://bugs.webkit.org/show_bug.cgi?id=73528 Reviewed by Adam Barth. Add network information API feature to build script. * Scripts/build-webkit: 2012-03-31 Ojan Vafai Generate the flakiness dashboard's list of webkit.org builders from the buildbot JSON https://bugs.webkit.org/show_bug.cgi?id=82839 Reviewed by Adam Barth. Also, update various hard-coded lists for WIN7 and Lion ports. This is necessary since the new list of builders grabbed off the buildbot includes Lion. * TestResultServer/static-dashboards/builders.js: (generateWebkitBuildersFromBuilderList): (xhr.onreadystatechange): Don't use dashboard_base's request method to avoid layering violation. * TestResultServer/static-dashboards/dashboard_base.js: (parseParameters): (initBuilders): (haveJsonFilesLoaded): (g_handleBuildersListLoaded): Block loading the JSON files for each builder until we actually have a list of builders. * TestResultServer/static-dashboards/flakiness_dashboard.html: * TestResultServer/static-dashboards/flakiness_dashboard_tests.js: (testPlatformAndBuildType): (testGenerateWebkitBuildersFromBuilderList): * TestResultServer/static-dashboards/timeline_explorer.html: * TestResultServer/static-dashboards/treemap.html: Delay generating the page until the builder list has loaded. 2012-03-31 Ojan Vafai If NRWT gets killed halfway through a run, it incorrectly reports tests that weren't run as passes https://bugs.webkit.org/show_bug.cgi?id=82799 Reviewed by Eric Seidel. If we don't run a test, mark it as skipped. * Scripts/webkitpy/layout_tests/controllers/manager.py: (Manager._mark_interrupted_tests_as_skipped): (Manager._interrupt_if_at_failure_limits.interrupt_if_at_failure_limit): * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py: (ManagerTest.test_interrupt_if_at_failure_limits): 2012-03-31 Ryosuke Niwa One more build fix after r112781 for Chromium Windows. Don't copy zip .idb files. * BuildSlaveSupport/built-product-archive: (copyBuildFiles): 2012-03-31 Ryosuke Niwa Yet another build fix after r112781 for Chromium Windows. Run webkit-build-directory by perl explicitly. * BuildSlaveSupport/built-product-archive: (determineWebKitBuildDirectory): 2012-03-31 Ryosuke Niwa Fix the syntax error in master.cfg after r112734. * BuildSlaveSupport/build.webkit.org-config/master.cfg: (unitTestsSupported): 2012-03-31 Ryosuke Niwa Another build fix after r112781. * BuildSlaveSupport/built-product-archive: (createZip): 2012-03-31 Ryosuke Niwa Build fix after r112781. * BuildSlaveSupport/built-product-archive: (createZipManually): (createZip): 2012-03-30 Ryosuke Niwa Chromium bots should upload archived built files https://bugs.webkit.org/show_bug.cgi?id=82666 Reviewed by Tony Chang. * BuildSlaveSupport/build.webkit.org-config/config.json: * BuildSlaveSupport/built-product-archive: (main): (determineWebKitBuildDirectory): Instead of hard-coding WebKitBuild as the build outout directory, call out to webkit-build-directory. (removeDirectoryIfExists): Extracted. (copyBuildFiles): Added to avoid archiving useless intermedinate files. (createZipManually): Used in Chromium Windows where we don't execute python scripts inside cygwin. (createZipManually.addToArchive): (createZip): Extracted. Calls out to appropraite command line scripts or createZipManually. (archiveBuiltProduct): Add support for Chromium port. * Scripts/webkit-build-directory: Add support for --platform options. 2012-03-30 Dan Bernstein Reverted r112767, because it caused many vertical text tests to fail. * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: * DumpRenderTree/fonts/SampleFont.sfont: Removed. * DumpRenderTree/mac/Configurations/DumpRenderTree.xcconfig: * DumpRenderTree/mac/DumpRenderTree.mm: (allowedFontFamilySet): (activateTestingFonts): * WebKitTestRunner/InjectedBundle/mac/ActivateFonts.mm: (WTR::allowedFontFamilySet): (WTR::activateFonts): * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj: * WebKitTestRunner/fonts/SampleFont.sfont: Removed. 2012-03-30 Malcolm MacLeod [wx] Move wxWebKit API into WebKit namespace. https://bugs.webkit.org/show_bug.cgi?id=82740 Reviewed by Kevin Ollivier. * DumpRenderTree/wx/DumpRenderTreeWx.cpp: (LayoutWebViewEventHandler::LayoutWebViewEventHandler): (LayoutWebViewEventHandler::bindEvents): (LayoutWebViewEventHandler::OnLoadEvent): (LayoutWebViewEventHandler::OnAlertEvent): (LayoutWebViewEventHandler::OnConfirmEvent): (LayoutWebViewEventHandler::OnPromptEvent): (LayoutWebViewEventHandler::OnConsoleMessageEvent): (LayoutWebViewEventHandler::OnReceivedTitleEvent): (LayoutWebViewEventHandler::OnWindowObjectClearedEvent): (LayoutWebViewEventHandler): (dumpFramesAsText): (dump): (MyApp::OnInit): * wx/browser/browser.cpp: (MyApp::OnInit): 2012-03-30 Eli Fidler Enable OpenType Sanitizer for BlackBerry port. https://bugs.webkit.org/show_bug.cgi?id=82782 Reviewed by Eric Seidel. * Scripts/webkitdirs.pm: (blackberryCMakeArguments): add OTS to include path 2012-03-30 Simon Fraser run-webkit-tests needs to set DYLD_LIBRARY_PATH as well as DYLD_FRAMEWORK_PATH, so that libWebCoreTestSupport.dylib is found. Reviewed by Mark Rowe. * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitDriver._start): 2012-03-30 Simon Fraser Disable run-api-tests on release mac builds https://bugs.webkit.org/show_bug.cgi?id=82788 Reviewed by Ryosuke Niwa. TestWebKitAPI crashes every time in release builds currently (https://bugs.webkit.org/show_bug.cgi?id=82652) so disable run-api-tests on mac release builders. * BuildSlaveSupport/build.webkit.org-config/master.cfg: (unitTestsSupported): 2012-03-30 Chang Shu Unreviewed. Update my email. * Scripts/webkitpy/common/config/committers.py: 2012-03-30 Chris Fleizach AX: Crash at WebCore::renderObjectContainsPosition(WebCore::RenderObject*, WebCore::Position const&) https://bugs.webkit.org/show_bug.cgi?id=82745 Reviewed by Simon Fraser. * DumpRenderTree/AccessibilityUIElement.cpp: (rangeForPositionCallback): (AccessibilityUIElement::rangeForPosition): (AccessibilityUIElement::getJSClass): * DumpRenderTree/AccessibilityUIElement.h: (AccessibilityUIElement): * DumpRenderTree/mac/AccessibilityUIElementMac.mm: (AccessibilityUIElement::rangeForPosition): * WebKitTestRunner/InjectedBundle/AccessibilityUIElement.cpp: (WTR::AccessibilityUIElement::rangeForPosition): * WebKitTestRunner/InjectedBundle/AccessibilityUIElement.h: (AccessibilityUIElement): * WebKitTestRunner/InjectedBundle/Bindings/AccessibilityUIElement.idl: * WebKitTestRunner/InjectedBundle/mac/AccessibilityUIElementMac.mm: (WTR::AccessibilityUIElement::rangeForPosition): (WTR): 2012-03-30 David Barr Split up top-level .gitignore and .gitattributes https://bugs.webkit.org/show_bug.cgi?id=82687 Reviewed by Tor Arne Vestbø. * .gitattributes: Added. * .gitignore: Added. * DumpRenderTree/DumpRenderTree.gyp/.gitignore: Added. * TestWebKitAPI/TestWebKitAPI.gyp/.gitignore: Added. 2012-03-30 Mihai Parparita Actually remove the ChromiumOS GTK builder, like r112673 said it would. * TestResultServer/static-dashboards/builders.js: 2012-03-30 Mihai Parparita Update ChromiumOS bot names in builders.js to reflect changes made by http://crrev.com/129835 Also removes the ChromiumOS GTK builder, which was removed by http://crrev.com/129835 * TestResultServer/static-dashboards/builders.js: 2012-03-30 Vineet Chaudhary [GTK] Disable introspection build in the at-spi2-core module. https://bugs.webkit.org/show_bug.cgi?id=82729 Reviewed by Philippe Normand. This change is a follow-up of bug 82395 which initially disabled for other modules of the set. * gtk/jhbuild.modules: Disable introspection. 2012-03-30 Jesus Sanchez-Palencia Add a "preview" state to Page Visibility API implementation https://bugs.webkit.org/show_bug.cgi?id=81355 Reviewed by Adam Barth. Updating the Page Visibility API implementation to the current spec version. * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::setPageVisibility): 2012-03-30 Keishi Hattori Change ENABLE_INPUT_COLOR to ENABLE_INPUT_TYPE_COLOR and enable it for chromium https://bugs.webkit.org/show_bug.cgi?id=80972 Reviewed by Kent Tamura. * Scripts/build-webkit: * qmake/mkspecs/features/features.prf: 2012-03-29 Ádám Kallai [Qt] Build fix by renameing QtDeclarative to QtQml in header calls. https://bugs.webkit.org/show_bug.cgi?id=82195 Reviewed by Simon Hausmann. * MiniBrowser/qt/BrowserWindow.cpp: (BrowserWindow::updateVisualMockTouchPoints): * MiniBrowser/qt/main.cpp: * WebKitTestRunner/qt/PlatformWebViewQt.cpp: (WTR::WrapperWindow::handleStatusChanged): * qmake/mkspecs/features/unix/default_post.prf: 2012-03-30 Sudarsana Nagineni [EFL] Implement LayoutTestController::setMinimumTimerInterval https://bugs.webkit.org/show_bug.cgi?id=81220 Add missing implementation setMinimumTimerInterval to EFL's LayoutTestController so that we can unskip related tests from the skip list. Reviewed by Philippe Normand. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::setMinimumTimerInterval): 2012-03-29 Dan Bernstein Removed “Intel” from the Lion builders’ names. Rubber-stamped by Mark Rowe. * BuildSlaveSupport/build.webkit.org-config/config.json: * BuildSlaveSupport/build.webkit.org-config/public_html/LeaksViewer/LeaksViewer.js: (LeaksViewer._displayURLPrompt): 2012-03-29 Mihai Parparita [Chromium] Add sharded ChromiumOS debug bots to builders.js https://bugs.webkit.org/show_bug.cgi?id=82639 Reviewed by Eric Seidel. The bots were sharded by http://crrev.com/129613, update builders.js to reflect this. * TestResultServer/static-dashboards/builders.js: 2012-03-29 Dirk Pranke remove support for junit-style xml output from test-webkitpy https://bugs.webkit.org/show_bug.cgi?id=82279 Reviewed by Eric Seidel. This was added when we were looking into integrating w/ Jenkins rather than buildbot, but I believe that project got shelved, so this is unused. We can always add it back in later as necessary. * Scripts/webkitpy/test/main.py: (Tester._parse_args): (Tester._run_tests): * Scritps/webkitpy/thirdparty/__init__.py: 2012-03-29 Dirk Pranke new-run-webkit-tests: crashes when it fails to decode a stack trace https://bugs.webkit.org/show_bug.cgi?id=82673 Unreviewed, build fix. We are assuming the stdout/stderr output from the driver is utf-8 encoded when we get stack traces; this may not be a valid assumption generally, but if we do get strings that aren't valid utf-8, we would crash. Now we will ignore any decoding errors. * Scripts/webkitpy/layout_tests/port/base.py: (Port._get_crash_log): * Scripts/webkitpy/layout_tests/port/port_testcase.py: (PortTestCase.test_get_crash_log): 2012-03-29 Simon Fraser Scripts fail to detect when a tool crashes https://bugs.webkit.org/show_bug.cgi?id=82659 Reviewed by Mark Rowe. Have exitStatus() detect when the process fail to exit cleanly (possibly because of a crash), and return a non-zero exit code in that case. * Scripts/VCSUtils.pm: (exitStatus): 2012-03-29 Dominik Röttsches [GTK][EFL] run-javascriptcore-tests should be run through jhbuild https://bugs.webkit.org/show_bug.cgi?id=82581 Reviewed by Martin Robinson. Running Javascriptcore tests through jhbuild for consistency with run-webkit-tests and in order to avoid confusing libraries when facing regressions. * Scripts/run-javascriptcore-tests: 2012-03-29 Dirk Pranke test-webkitpy: add --timing https://bugs.webkit.org/show_bug.cgi?id=82550 Reviewed by Eric Seidel. This patch adds a --timing option that will display the time each test takes. It also removes the --silent option, since probably no one ever used it, and cleans up the logging option parsing code to be easier to follow. * Scripts/webkitpy/test/main.py: (Tester._parse_args): (Tester._configure): * Scripts/webkitpy/test/runner.py: (TestRunner.write_result): * Scripts/webkitpy/test/runner_unittest.py: (RunnerTest.test_regular): (RunnerTest.test_verbose): (RunnerTest): (RunnerTest.test_timing): 2012-03-29 Caio Marcelo de Oliveira Filho HashMap<>::add should return a more descriptive object https://bugs.webkit.org/show_bug.cgi?id=71063 Reviewed by Ryosuke Niwa. Update code to use AddResult instead of a pair. * DumpRenderTree/mac/LayoutTestControllerMac.mm: (LayoutTestController::evaluateScriptInIsolatedWorld): * DumpRenderTree/win/LayoutTestControllerWin.cpp: (LayoutTestController::evaluateScriptInIsolatedWorld): * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: (WTR::LayoutTestController::evaluateScriptInIsolatedWorld): 2012-03-29 Ilya Tikhonovsky Web Inspector: chromium: DRT --no-timeout option doesn't work. https://bugs.webkit.org/show_bug.cgi?id=82608 Initial value for m_timeout was initialized in constructor and was overwritten in DRT::main. This was broken by r112354 and the sequence became opposite. Reviewed by Yury Semikhatsky. * DumpRenderTree/chromium/TestShell.cpp: (TestShell::TestShell): (TestShell::initialize): 2012-03-29 Vineet Chaudhary Remove custom bindings form Internals.idl of attribute type Array. https://bugs.webkit.org/show_bug.cgi?id=82319 Reviewed by Kentaro Hara. Remove JSInternalsCustom.cpp and V8InternalsCustom.cpp as no longer required. * GNUmakefile.am: 2012-03-28 Mark Rowe Fix --no-saved-state. Ignore the flag before Lion, rather than making it specific to Lion. Rubber-stamped by Dan Bernstein. * Scripts/webkitdirs.pm: (argumentsForRunAndDebugMacWebKitApp): 2012-03-28 Dave Tharp Add dtharp to contributors list https://bugs.webkit.org/show_bug.cgi?id=82545 Reviewed by Adam Barth. Adding dtharp as Contributor. * Scripts/webkitpy/common/config/committers.py: 2012-03-28 Stuart Eichert build-webkit : the word "pass" is spelled incorrectly as "passs" https://bugs.webkit.org/show_bug.cgi?id=82532 Reviewed by Julien Chaffraix. Corrected spelling of the word "pass" in build-webkit. * Scripts/build-webkit: 2012-03-28 Dirk Pranke new-run-webkit-tests doesn't log correctly after retrying failures https://bugs.webkit.org/show_bug.cgi?id=82541 Unreviewed, build fix. r112189 introduced a regression where we would accidentally delete the log handler after we retried any failing tests - that meant that any log messages around uploading the results got dropped. * Scripts/webkitpy/layout_tests/controllers/manager.py: (Manager._run_tests): 2012-03-28 Kevin Ollivier [wx] Unreviewed. Build fix, move WTF back into JSCore target until issues with JSCore not linking in all WTF symbols are resolved. * waf/build/settings.py: 2012-03-28 Simon Fraser download-built-product build step should detect 404 errors https://bugs.webkit.org/show_bug.cgi?id=82491 Reviewed by Ryosuke Niwa. Pass --fail to curl so that it reports 404 errors. * BuildSlaveSupport/build.webkit.org-config/master.cfg: (DownloadBuiltProduct): Pass --fail to curl so that it reports errors. * BuildSlaveSupport/built-product-archive: (extractBuiltProduct): Remove trailing whitespace. 2012-03-28 Dirk Pranke flakiness dashboard should only keep the last 500 runs https://bugs.webkit.org/show_bug.cgi?id=82530 Reviewed by Ojan Vafai. If a bot has an overly large amount of failing tests, then storing 750 runs' worth takes a lot of memory and CPU time to process and we end up getting killed by AppEngine; this cases data corruption and confusion down the road when we have to retry the upload of the data and the tests show up as multiple runs. * TestResultServer/model/jsonresults.py: (JsonResults.update_file): 2012-03-28 Ryosuke Niwa extract-built-product build step should fail if unzipping fails https://bugs.webkit.org/show_bug.cgi?id=82493 Reviewed by Simon Fraser. Forward non-zero exit code from subprocesses so that the master can detect any failures and bail out early. * BuildSlaveSupport/built-product-archive: (main): 2012-03-28 Dirk Pranke test-webkitpy should support files, directories, and packages as command line args https://bugs.webkit.org/show_bug.cgi?id=76765 Reviewed by Adam Barth. This patch adds support for specifying files, directories, and packages to test-webkitpy along with the already existing support for modules, test classes, and individual test names. Also, fix a bug in filesystem_mock where we wouldn't normalize a path containing a reference to the current directory properly, for example, '/foo/./bar.py'. * Scripts/webkitpy/common/system/filesystem_mock.py: (MockFileSystem.normpath): * Scripts/webkitpy/test/main.py: (Tester._configure_logging): (Tester._run_tests): * Scripts/webkitpy/test/test_finder.py: (TestDirectoryTree.find_modules): (TestDirectoryTree.subpath): (TestFinder.is_dotted_name): (TestFinder.find_names): (TestFinder): (TestFinder._find_names_for_arg): (TestFinder._find_in_trees): (TestFinder._default_names): * Scripts/webkitpy/test/test_finder_unittest.py: (TestFinderTest.setUp): (TestFinderTest.tearDown): (TestFinderTest.check_names): (TestFinderTest.test_default_names): (TestFinderTest): (TestFinderTest.test_paths): 2012-03-28 Simon Fraser Fix inspector tests when running in internal builds. Reviewed by Tim Hatcher. * DumpRenderTree/mac/DumpRenderTree.mm: (resetDefaultsToConsistentValues): * WebKitTestRunner/TestController.cpp: (WTR::TestController::resetStateToConsistentValues): 2012-03-28 David Kilzer Use xcrun to find nm and ranlib on OS X builds Reviewed by Dan Bernstein. * Scripts/build-webkit: Use xcrun to find ranlib. * Scripts/webkitdirs.pm: (determineNmPath): Added. Use xcrun to find nm on OS X. (nmPath): Added. * Scripts/webkitperl/features.pm: (libraryContainsSymbol): Use nmPath(). 2012-03-28 Ojan Vafai Add history navigation to garden-o-matic https://bugs.webkit.org/show_bug.cgi?id=82495 Reviewed by Dimitri Glazkov. Also, maintain scroll offsets when returning to a tab. This makes addressing expected failures considerably easier when dealing with below-the-fold tests. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui.js: -Store the scrollTop before switching to a new tab and restore the appropriate scrollTop after switching. -Modify window.location with the tabName in the hash. Can't use pushState because this is served from a file URL and Chrome puts each different file URL in it's own origin. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results.js: Remove the href so that clicking on the accordion item does not modify window.location.hash. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui_unittests.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/styles/results.css: 2012-03-28 Philippe Normand [GTK] generate-gtk-doc doesn't cope with custom build directory https://bugs.webkit.org/show_bug.cgi?id=82448 Reviewed by Martin Robinson. Pass the product directory in generate-gtkdoc arguments from build-webkit. * Scripts/webkitdirs.pm: (buildAutotoolsProject): * gtk/common.py: (get_build_path): 2012-03-28 Carlos Garcia Campos [GTK] Add --skipped command line option to Tools/Scripts/run-gtk-tests https://bugs.webkit.org/show_bug.cgi?id=82341 Reviewed by Martin Robinson. Add --skipped=skip|ignore|only command line option. * Scripts/run-gtk-tests: (TestRunner._test_cases_to_skip): Return the list of test cases to skip for the given test depending on the skipped action. (TestRunner._should_run_test): Decide whether to run the test or not depending on the skipped action. (TestRunner.run_tests): Only show the list of skipped tests when skipped action is 'skip'. 2012-03-28 Carlos Garcia Campos [GTK] Use gtester -s to skip individual test cases instead of unit tests as a whole https://bugs.webkit.org/show_bug.cgi?id=82333 Reviewed by Martin Robinson. * Scripts/run-gtk-tests: (TestRunner.SkippedTest): Add SkippedTest class to store more information about skipped tests. (TestRunner._setup_testing_environment): Use SkippedTest to add TestWebKitAccessibility to the skipped list. (TestRunner._find_skipped_test): Return the SkippedTest object for a given test path. (TestRunner._test_cases_to_skip): Returns the list of test cases that should be skipped for a given test path. (TestRunner._should_run_test): Helper funtion to decide whether the given test path should be run or not. Tests in the skipped list that don't have a list of failing test cases are not run at all. Tests in the skipped list that contain a list of test cases are run with gtester -s to skip the individual test cases. (TestRunner._run_test): Helper funtion to run a given test. (TestRunner.run_tests): Build the list of tests to run based on _should_run_test() and use helper function _run_test() to run every test. 2012-03-28 Csaba Osztrogonác [Qt] Couple of tests have different results on 64 bit and/or in debug mode compared to 32 bit and/or release mode https://bugs.webkit.org/show_bug.cgi?id=52810 Reviewed by Simon Hausmann. Use SSE2 floating point math on 32 bit instead of the default 387 to make layout test result same on 32 and on 64 bit builds. * qmake/mkspecs/features/unix/default_post.prf: 2012-03-28 Sergio Villar Senin [Soup] DNS prefetching spams resolver, shoots self in the foot https://bugs.webkit.org/show_bug.cgi?id=41630 Reviewed by Martin Robinson. Bump libsoup and glib dependencies. * efl/jhbuild.modules: * gtk/jhbuild.modules: 2012-03-28 Gustavo Noronha Silva [GTK] Disable introspection for jhbuild modules https://bugs.webkit.org/show_bug.cgi?id=82395 Reviewed by Philippe Normand. * efl/jhbuild.modules: add --disable-introspection to all modules that support it * gtk/jhbuild.modules: ditto. 2012-03-28 Csaba Osztrogonác [Qt] Unreviewed, fix API tests after adding full platforms. (r111658) * BuildSlaveSupport/build.webkit.org-config/master.cfg: (BuildAndTestFactory.__init__): 2012-03-28 Csaba Osztrogonác [Qt] Unreviewed typo fix after r112383. * Scripts/run-qtwebkit-tests: (Main.check_crash_occurences): 2012-03-28 Ádám Kallai [Qt] run-qtwebkit-tests should report crashes https://bugs.webkit.org/show_bug.cgi?id=76905 Added checking of crashing tests. Now final summary includes the total number of crashing tests too. Made unit test for this modification by János Badics . Reviewed by Csaba Osztrogonác. * BuildSlaveSupport/build.webkit.org-config/master.cfg: (RunQtAPITests.commandComplete): (RunQtAPITests.evaluateCommand): * BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py: (RunQtAPITestsTest): (test_failure): (test_crash): * Scripts/run-qtwebkit-tests: (Main.announce_results_html): (Main): (Main.check_crash_occurences): (Main.convert_to_stdout): (Main.convert_to_html): 2012-03-28 Kwonjin Jeong Add support for Intel compiler on Linux QT port. https://bugs.webkit.org/show_bug.cgi?id=81077 Reviewed by Tor Arne Vestbø. * qmake/mkspecs/features/default_post.prf: Don't hide symbols when building with Intel compiler. The symbols are needed when building QtWebKit tests. * qmake/mkspecs/features/unix/icc.prf: Added to ignore icc-specific warnings. This feature is used with command-line arguments for build-webkit script. 2012-03-28 Philippe Normand Unreviewed, trivial update-webkit-libs-jhbuild system call fix. * Scripts/update-webkitgtk-libs: This script is written in Perl. Not Python. So use perl vocabulary to refer to command line arguments. 2012-03-28 Ryosuke Niwa kill-old-processes doesn't do anything on linux https://bugs.webkit.org/show_bug.cgi?id=82388 Reviewed by Tony Chang. Kill cc1plus, ld, etc... on linux. Also ps aux + grep to kill old run-webkit-tests processes. This should help stabilizing Chromium linux bots. * BuildSlaveSupport/kill-old-processes: 2012-03-28 Philippe Normand [GTK] jhbuild wrongly tricked into clean-up https://bugs.webkit.org/show_bug.cgi?id=82216 Reviewed by Gustavo Noronha Silva. Save jhbuild{rc,.modules}.md5sum in Dependencies instead of the build tree. Also when checking if the jhbuild configuration changed use absolute md5sum paths. * Scripts/webkitdirs.pm: (runAutogenForAutotoolsProjectIfNecessary): (jhbuildConfigurationChanged): (buildAutotoolsProject): 2012-03-28 Csaba Osztrogonác kill-old-processes doesn't do anything on linux https://bugs.webkit.org/show_bug.cgi?id=82388 One more attempt to disable kill-old-processes on the Qt bots not to kill each others. The platform in Factory comes from config.json, accordingly platform is fullplatform, for example: qt-4.8, qt-5.0-wk2. We have to check only the prefix of platform. Reviewed by Ryosuke Niwa. * BuildSlaveSupport/build.webkit.org-config/master.cfg: (Factory.__init__): 2012-03-28 Hao Zheng Make Chromium port for Android use hardware gpu path default. https://bugs.webkit.org/show_bug.cgi?id=82033 Reviewed by Dirk Pranke. Skip canvas tests, which are run in virtual gpu test suites. * Scripts/webkitpy/layout_tests/port/chromium_android.py: (ChromiumAndroidPort.__init__): (ChromiumAndroidPort.skipped_tests): 2012-03-27 Csaba Osztrogonác kill-old-processes doesn't do anything on linux https://bugs.webkit.org/show_bug.cgi?id=82388 Disable kill-old-processes on the Qt bots not to kill each others. Reviewed by Ryosuke Niwa. * BuildSlaveSupport/build.webkit.org-config/master.cfg: (Factory.__init__): 2012-03-27 Hao Zheng Handle DRT crash caused by Android OOM. https://bugs.webkit.org/show_bug.cgi?id=82310 Reviewed by Tony Chang. When Android is OOM, it sends a SIGKILL (137) signal to DRT. DRT is stopped silently and regarded as crashed. Re-run the test for such crash. Add some debug log to NRWT. * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumDriver.run_test): * Scripts/webkitpy/layout_tests/port/chromium_android.py: (ChromiumAndroidPort.__init__): (ChromiumAndroidPort.get_last_stacktrace): (ChromiumAndroidDriver.__init__): (ChromiumAndroidDriver._start): (ChromiumAndroidDriver.run_test): (ChromiumAndroidDriver._get_drt_return_value): 2012-03-27 Ojan Vafai Fix expected results for some unittest failures. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/failures_unittests.js: 2012-03-27 Adrienne Walker [chromium] Fix --enabled-threaded-compositing flag in DRT https://bugs.webkit.org/show_bug.cgi?id=82405 Reviewed by James Robinson. --enable-threaded-compositing wasn't working because there was no way to set the value before it was used in the TestShell constructor. To avoid adding yet another boolean parameter to TestShell's constructor, add an explicit initialize function that can operate on any boolean settings that are set post-constructor. * DumpRenderTree/chromium/DumpRenderTree.cpp: (main): * DumpRenderTree/chromium/TestShell.cpp: (TestShell::TestShell): (TestShell::initialize): * DumpRenderTree/chromium/TestShell.h: (TestShell): (TestShell::setTestShellMode): 2012-03-27 Dirk Pranke new-run-webkit-tests: suppress debug output from run-safari on results file https://bugs.webkit.org/show_bug.cgi?id=82400 Reviewed by Ryosuke Niwa. Send the run-safari output to /dev/null rather than the console. * Scripts/webkitpy/layout_tests/port/mac.py: (MacPort.show_results_html_file): 2012-03-27 Simon Fraser Improve error reporting in run-javascriptcore-tests https://bugs.webkit.org/show_bug.cgi?id=82379 Reviewed by Jessie Berlin. When the script dies, have it print out the reason for dying. * Scripts/run-javascriptcore-tests: 2012-03-27 Eric Seidel Add Dave Barton to the Contributors list so I can more easily CC him on bugs https://bugs.webkit.org/show_bug.cgi?id=82373 Reviewed by Ryosuke Niwa. Dave's committer-status in in-progress, so until then list him as a Contributor so at least I can more easily CC him on bugs. * Scripts/webkitpy/common/config/committers.py: 2012-03-26 Raphael Kubo da Costa [jhbuild] Prepend jhbuildrc's directory to sys.path instead of appending. https://bugs.webkit.org/show_bug.cgi?id=82252 Reviewed by Gustavo Noronha Silva. Appending the directory containing common.py may be problematic if another path in sys.path also contains either a common.py or a common/ module, so prepend to select our version instead. This is the case with fail2ban, which has a common/ directory which may end up being in Python's default path. * efl/jhbuildrc: * gtk/jhbuildrc: 2012-03-27 Dirk Pranke test-webkitpy: prepare for better test run output https://bugs.webkit.org/show_bug.cgi?id=82290 Reviewed by Adam Barth. This code basically re-implements the output of the TextTestRunner default runner code from unittest, although the implementation is quite different, in preparation for changing the test output to be metered and possibly running in parallel. The output is almost identical to before, except that instead of logging "test_regular (webkitpy.main.RunnerTest) passed" we log "webkitpy.main.RunnerTest.test_regular passed". It has always annoyed me that they invert the names to be harder to read and so that you can't copy & paste back to the input for test-webkitpy. This patch is provided to add a skeleton for unit tests and for comparison to an upcoming patch that will actually add new functionality. * Scripts/webkitpy/test/main.py: (Tester.__init__): (Tester._configure_logging): (Tester._run_tests): * Scripts/webkitpy/test/runner.py: Added. (TestRunner): (TestRunner.__init__): (TestRunner.test_name): (TestRunner.all_test_names): (TestRunner.run): (TestRunner.write_result): (TestRunner.write_summary): * Scripts/webkitpy/test/runner_unittest.py: Added. (FakeModuleSuite): (FakeModuleSuite.__init__): (FakeModuleSuite.__str__): (FakeModuleSuite.run): (FakeTopSuite): (FakeTopSuite.__init__): (FakeLoader): (FakeLoader.__init__): (FakeLoader.top_suite): (FakeLoader.loadTestsFromName): (RunnerTest): (RunnerTest.test_regular): (RunnerTest.test_verbose): 2012-03-27 Gustavo Noronha Silva [GTK] Build gnutls without p11-kit support https://bugs.webkit.org/show_bug.cgi?id=82212 Reviewed by Martin Robinson. * gtk/jhbuild.modules: pass --without-p11-kit to gnutls' configure, to avoid requiring that dependency, which is not used by libsoup and WebKit anyway 2012-03-27 Csaba Osztrogonác [Qt] Make debug build work on 32 bit https://bugs.webkit.org/show_bug.cgi?id=82331 Use --no-keep-memory option of ld not to cache the symbol tables of input files in memory to avoid memory exhaustion during the linking phase for x86-debug builds. Reviewed by Tor Arne Vestbø. * qmake/mkspecs/features/unix/default_post.prf: * qmake/mkspecs/features/unix/default_pre.prf: 2012-03-23 Balazs Ankes Add png checking to check-webkit-style https://bugs.webkit.org/show_bug.cgi?id=75824 Reviewed by NOBODY Tony Chang. * Scripts/webkitpy/style/checker.py: (_all_categories): (FileType): (CheckerDispatcher._should_skip_file_path): (CheckerDispatcher._file_type): (CheckerDispatcher._create_checker): * Scripts/webkitpy/style/checker_unittest.py: (CheckerDispatcherDispatchTest.test_none_paths): * Scripts/webkitpy/style/checkers/png.py: Added. (PNGChecker): (PNGChecker.__init__): (PNGChecker.check): (PNGChecker._config_file_path): * Scripts/webkitpy/style/checkers/png_unittest.py: Added. (MockSCMDetector): (MockSCMDetector.__init__): (MockSCMDetector.display_name): (MockSCMDetector.propget): (PNGCheckerTest): (PNGCheckerTest.test_init): (PNGCheckerTest.test_init.mock_handle_style_error): (PNGCheckerTest.test_check): (PNGCheckerTest.test_check.mock_handle_style_error): * Scripts/webkitpy/style/patchreader.py: (PatchReader.check): 2012-03-27 Carlos Garcia Campos Unreviewed. Skip GTK+ unit tests that fail in the bots * Scripts/run-gtk-tests: (TestRunner): Add failing tests to the skipped list. 2012-03-27 Carlos Garcia Campos Unreviewed. Unskip all GTK+ unit test. Currently, skipped tests contains the list of unit tests that have failed in one or more bots. In this moment it's not possible to skip just the test cases that fail, so the complete unit test is skipped when a single test case fails. For that reason, we don't know which test cases are actually failing for every skipped test. We are planning to land a patch to allow skipping individual test cases, but first we need to know which test cases are currently failing in the bots to skip them. So, this patch simply unskips all unit tests in order to watch the bots and get the list of test cases failing. Failing tests will be skipped again as soon as we have the new list of failing test cases. * Scripts/run-gtk-tests: (TestRunner): Reset the list of skipped tests. 2012-03-26 Scott Byer Enable layout testing of the scroll animator. https://bugs.webkit.org/show_bug.cgi?id=81858 Add a call to the InternalSettings that layout tests can use to turn on scroll animation. Enable animation updates for the Chromium platform DRT when scroll animation has been turned on in a test. This should be a no-op for all current layout tests. Reviewed by James Robinson. * DumpRenderTree/chromium/WebViewHost.cpp: (WebViewHost::serviceAnimation): (WebViewHost::scheduleAnimation): * DumpRenderTree/chromium/WebViewHost.h: (WebViewHost): 2012-03-26 Dinu Jacob [Qt][WK2] Support multi-file upload https://bugs.webkit.org/show_bug.cgi?id=81589 Reviewed by Simon Hausmann. Modified filePicker to support multi-file upload. * MiniBrowser/qt/MiniBrowser.qrc: * MiniBrowser/qt/icons/checkbox_checked.png: Added. * MiniBrowser/qt/icons/checkbox_unchecked.png: Added. * MiniBrowser/qt/js/MultiSelect.js: Added. (values): (isSelected): (removeValue): * MiniBrowser/qt/qml/CheckBox.qml: Added. * MiniBrowser/qt/qml/FilePicker.qml: 2012-03-26 Mark Hahnenberg Retry crashing tests serially at the end of NRWT on Apple Mac https://bugs.webkit.org/show_bug.cgi?id=82233 Reviewed by Dirk Pranke. * Scripts/webkitpy/layout_tests/controllers/manager.py: (Manager.run): * Scripts/webkitpy/layout_tests/port/base.py: (Port.should_retry_crashes): * Scripts/webkitpy/layout_tests/port/mac.py: (MacPort.should_retry_crashes): 2012-03-26 Mark Hahnenberg NRWT should retry failures serially https://bugs.webkit.org/show_bug.cgi?id=82241 Reviewed by Dirk Pranke. * Scripts/webkitpy/layout_tests/controllers/manager.py: (Manager._run_tests): Add another parameter to specify the number of workers to use. (Manager.run): Run the failures serially. 2012-03-26 Adrienne Walker Allow expectations on directories in virtual test suites https://bugs.webkit.org/show_bug.cgi?id=82250 Reviewed by Dirk Pranke. lookup_virtual_test_base was only returning valid results for files and not directories. Fix by falling back to the virtual test suite mapping if possible. * Scripts/webkitpy/layout_tests/port/base.py: (Port.lookup_virtual_test_base): * Scripts/webkitpy/layout_tests/port/test.py: (TestPort.skipped_tests): (TestPort.virtual_test_suites): 2012-03-26 Dirk Pranke Fix duplicated ChangeLog entry from r112171. Unreviewed, build fix. 2012-03-26 Dirk Pranke nrwt: remove --worker-model flag https://bugs.webkit.org/show_bug.cgi?id=82112 Reviewed by Adam Barth. The --worker-model=inline flag is basically redundant with the --child-processes=1 flag (technically, you could one run child process, but this is only useful for testing); this patch removes this flag, which simplifies things and enables more cleanup down the road. * Scripts/webkitpy/layout_tests/controllers/manager.py: (Manager._run_tests): (Manager.print_config): * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py: (get): * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py: (make_broker): (FunctionTests.test_get__inline): (FunctionTests.test_get__processes): (_TestsMixin.setUp): (_TestsMixin.make_broker): (InlineBrokerTests.setUp): (MultiProcessBrokerTests.setUp): * Scripts/webkitpy/layout_tests/port/base.py: (Port.default_child_processes): * Scripts/webkitpy/layout_tests/port/chromium_android.py: (ChromiumAndroidPort.check_sys_deps): * Scripts/webkitpy/layout_tests/port/chromium_android_unittest.py: (ChromiumAndroidPortTest): * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py: (MockDRTPortTest.make_port): * Scripts/webkitpy/layout_tests/port/port_testcase.py: (PortTestCase): (PortTestCase.make_port): * Scripts/webkitpy/layout_tests/port/test.py: (TestPort.default_child_processes): * Scripts/webkitpy/layout_tests/run_webkit_tests.py: (_set_up_derived_options): (parse_args): * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py: (parse_args): (MainTest.test_batch_size): (MainTest.test_child_processes_2): (MainTest.test_child_processes_min): (MainTest.test_exception_raised): (MainTest.test_keyboard_interrupt): (MainTest.test_exit_after_n_crashes): (MainTest.assert_run_order): (MainTest.test_run_order__inline): (MainTest.test_virtual): 2012-03-26 Dirk Pranke test-webkitpy: split test-finding code into a different module https://bugs.webkit.org/show_bug.cgi?id=82253 Reviewed by Adam Barth. Per suggestion from abarth, this change splits all the filesystem-crawling, test-finding code in test-webkitpy out into a separate module and switches to using a FileSystem object. This makes things much more testable, so we also add tests :). We also add a realpath() method to the Filesystem object, since test-webkitpy needs that in order to be able to resolve symlinks properly to determine whether a file is under a particular tree or not. * Scripts/webkitpy/common/system/filesystem.py: (FileSystem.realpath): * Scripts/webkitpy/common/system/filesystem_mock.py: (MockFileSystem.realpath): * Scripts/webkitpy/test/main.py: (Tester.__init__): (Tester.add_tree): (Tester.run): (Tester._run_tests): (Tester._log_exception): * Scripts/webkitpy/test/test_finder.py: Added. (TestDirectoryTree): (TestDirectoryTree.__init__): (TestDirectoryTree.find_modules): (TestDirectoryTree.find_modules.file_filter): (TestDirectoryTree.to_module): (TestDirectoryTree.clean): (TestFinder): (TestFinder.__init__): (TestFinder.add_tree): (TestFinder.additional_paths): (TestFinder.clean_trees): (TestFinder.is_module): (TestFinder.to_module): (TestFinder.find_names): (TestFinder._exclude): * Scripts/webkitpy/test/test_finder_unittest.py: Added. (TestFinderTest): (TestFinderTest.setUp): (TestFinderTest.tearDown): (TestFinderTest.test_additional_system_paths): (TestFinderTest.test_is_module): (TestFinderTest.test_to_module): (TestFinderTest.test_clean): (TestFinderTest.test_find_names): 2012-03-26 Raphael Kubo da Costa [jhbuild] Use $MAKE if it is defined to build jhbuild itself. https://bugs.webkit.org/show_bug.cgi?id=82209 Reviewed by Martin Robinson. Respecting the $MAKE environment variable when it is defined makes building jhbuild more portable, as some platforms may have GNU make as gmake, for example. * jhbuild/jhbuild-wrapper: (install_jhbuild): 2012-03-26 Dirk Pranke add a Tree abstraction to test-webkitpy to better encapsulate things https://bugs.webkit.org/show_bug.cgi?id=82158 Reviewed by Adam Barth. There are no functional changes in this patch, but adding a basic abstraction for the trees we are looking in for python tests will allow me to add some features later on more easily (like ignoring certain directories like webkitpy/thirdparty), and it makes things slightly less hard-coded. * Scripts/test-webkitpy: * Scripts/webkitpy/test/main.py: (Tester.__init__): (Tester): (Tester.add_tree): (Tester._parse_args): (Tester._configure): (Tester.run): (Tester._find_modules): (Tester._run_tests): (Tester._is_module): (Tester._log_exception): (TestDirectoryTree): (TestDirectoryTree.__init__): (TestDirectoryTree.find_modules): (TestDirectoryTree.clean): 2012-03-26 Ojan Vafai garden-o-matic should special case reftests https://bugs.webkit.org/show_bug.cgi?id=81812 Reviewed by Adam Barth. On the examine page, don't show a rebaseline button for reftests. We still show it on the failure summary page because tests are grouped together. Long-term, we should make sure webkit-patch rebaseline-test never tries to rebaseline reftests. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/controllers.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/styles/results.css: 2012-03-26 Dirk Pranke clean up NRWT logging/metering, be less verbose https://bugs.webkit.org/show_bug.cgi?id=73846 Reviewed by Ojan Vafai. Re-land r112014 with a build fix for Python 2.6 compatibility. * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py: (ManagerTest.test_http_locking): * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py: (AbstractWorker.run): * Scripts/webkitpy/layout_tests/controllers/worker.py: (Worker.__init__): (Worker._set_up_logging): (Worker): (Worker._set_up_host_and_port): (Worker.run): (Worker.cleanup): * Scripts/webkitpy/layout_tests/controllers/worker_unittest.py: (WorkerTest.test_default_platform_in_worker): * Scripts/webkitpy/layout_tests/run_webkit_tests.py: (run): (main): * Scripts/webkitpy/layout_tests/views/metered_stream.py: (MeteredStream): (MeteredStream.implements): (MeteredStream._erasure): (MeteredStream._ensure_newline): (MeteredStream.__init__): (MeteredStream.__del__): (MeteredStream.cleanup): (MeteredStream.write_throttled_update): (MeteredStream.write_update): (MeteredStream.write): (MeteredStream.writeln): (MeteredStream._erase_last_partial_line): (_LogHandler): (_LogHandler.__init__): (_LogHandler.emit): * Scripts/webkitpy/layout_tests/views/metered_stream_unittest.py: (RegularTest): (RegularTest.setUp): (RegularTest.tearDown): (RegularTest.time_fn): (RegularTest.test_logging_not_included): (RegularTest._basic): (RegularTest.test_basic): (RegularTest._log_after_update): (RegularTest.test_log_after_update): (TtyTest): (TtyTest.test_basic): (TtyTest.test_log_after_update): (VerboseTest): (VerboseTest.test_basic): (VerboseTest.test_log_after_update): * Scripts/webkitpy/layout_tests/views/printing.py: (parse_print_options): (Printer.__init__): (Printer.cleanup): (Printer.print_progress): (Printer.print_update): (Printer._write): * Scripts/webkitpy/layout_tests/views/printing_unittest.py: (Testprinter.get_printer): (Testprinter.test_print_progress): * Scripts/webkitpy/performance_tests/perftestsrunner.py: (PerfTestsRunner.__init__): * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py: (test_run_test_pause_before_testing): 2012-03-26 Raphael Kubo da Costa [jhbuild] Revert $MAKE environment hack introduced in r101929. https://bugs.webkit.org/show_bug.cgi?id=82234 Reviewed by Martin Robinson. The problem it was supposedly fixing was fixed in jhbuild itself in git revision da8944d8ca987ca48c91b350257a530050406092, so we can remove the workaround added to the script. * jhbuild/jhbuild-wrapper: (install_jhbuild): 2012-03-26 Leo Yang [BlackBerry] Skeleton code of AsyncFileSystemBlackBerry.cpp https://bugs.webkit.org/show_bug.cgi?id=82157 Reviewed by Rob Buis. * Scripts/build-webkit: Make ENABLE_FILE_SYSTEM default on for blackberry. 2012-03-26 Raphael Kubo da Costa [jhbuild] Explicitly disable DTrace when building glib. https://bugs.webkit.org/show_bug.cgi?id=82145 Reviewed by Gustavo Noronha Silva. DTrace support is not really needed by WebKit, so disabling it helps avoid some build errors on more exotic platforms (such as FreeBSD). * efl/jhbuild.modules: * gtk/jhbuild.modules: 2012-03-26 Sheriff Bot Unreviewed, rolling out r111993. http://trac.webkit.org/changeset/111993 https://bugs.webkit.org/show_bug.cgi?id=82184 It broke layout and API tests and made WTR crash (Requested by Ossy on #webkit). * WebKitTestRunner/qt/PlatformWebViewQt.cpp: (WTR::PlatformWebView::PlatformWebView): 2012-03-25 Kevin Ollivier [wx] Unreviewed build fix. Move WTF to its own static lib build. * waf/build/settings.py: (common_configure): 2012-03-24 Chris Fleizach AX: Support solution to handle invalid ax text marker https://bugs.webkit.org/show_bug.cgi?id=82023 Reviewed by Oliver Hunt. * DumpRenderTree/AccessibilityUIElement.cpp: (indexForTextMarkerCallback): (isTextMarkerValidCallback): (textMarkerForIndexCallback): (AccessibilityUIElement::indexForTextMarker): (AccessibilityUIElement::isTextMarkerValid): (AccessibilityUIElement::textMarkerForIndex): (AccessibilityUIElement::getJSClass): * DumpRenderTree/AccessibilityUIElement.h: (AccessibilityUIElement): * DumpRenderTree/mac/AccessibilityUIElementMac.mm: (AccessibilityUIElement::indexForTextMarker): (AccessibilityUIElement::textMarkerForIndex): (AccessibilityUIElement::isTextMarkerValid): * WebKitTestRunner/InjectedBundle/AccessibilityUIElement.cpp: (WTR::AccessibilityUIElement::indexForTextMarker): (WTR::AccessibilityUIElement::isTextMarkerValid): (WTR::AccessibilityUIElement::textMarkerForIndex): (WTR): * WebKitTestRunner/InjectedBundle/AccessibilityUIElement.h: (AccessibilityUIElement): * WebKitTestRunner/InjectedBundle/Bindings/AccessibilityUIElement.idl: * WebKitTestRunner/InjectedBundle/mac/AccessibilityUIElementMac.mm: (WTR): (WTR::AccessibilityUIElement::indexForTextMarker): (WTR::AccessibilityUIElement::isTextMarkerValid): (WTR::AccessibilityUIElement::textMarkerForIndex): 2012-03-24 Sheriff Bot Unreviewed, rolling out r112014. http://trac.webkit.org/changeset/112014 https://bugs.webkit.org/show_bug.cgi?id=82138 Broke SnowLeopard, Chromium and GTK+ bots (Requested by kov on #webkit). * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py: (ManagerTest.test_http_locking): * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py: (AbstractWorker.run): * Scripts/webkitpy/layout_tests/controllers/worker.py: (Worker.__init__): (Worker.safe_init): (Worker.run): (Worker.cleanup): * Scripts/webkitpy/layout_tests/controllers/worker_unittest.py: (WorkerTest.test_default_platform_in_worker): * Scripts/webkitpy/layout_tests/run_webkit_tests.py: (run): (main): * Scripts/webkitpy/layout_tests/views/metered_stream.py: (MeteredStream): (MeteredStream.is): (MeteredStream.__init__): (MeteredStream.write): (MeteredStream.update): (MeteredStream._overwrite): (MeteredStream._reset): * Scripts/webkitpy/layout_tests/views/metered_stream_unittest.py: (TestMeteredStream): (TestMeteredStream.test_regular): * Scripts/webkitpy/layout_tests/views/printing.py: (_configure_logging): (_restore_logging): (Printer.__init__): (Printer.cleanup): (Printer.print_progress): (Printer.print_update): (Printer._write): (Printer): (Printer._update): * Scripts/webkitpy/layout_tests/views/printing_unittest.py: (TestUtilityFunctions.assertEmpty): (TestUtilityFunctions): (TestUtilityFunctions.assertNotEmpty): (TestUtilityFunctions.assertWritten): (TestUtilityFunctions.test_configure_logging): (Testprinter.get_printer): (Testprinter.test_print_progress): (Testprinter.test_print_progress.mock_time): * Scripts/webkitpy/performance_tests/perftestsrunner.py: (PerfTestsRunner.__init__): * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py: (test_run_test_pause_before_testing): 2012-03-24 Kevin Ollivier [wx] Unreviewed. Fix the build after WTF move. * waf/build/build_utils.py: (get_excludes): (get_excludes_in_dirs): * waf/build/settings.py: (get_port_excludes): (common_configure): 2012-03-24 Dirk Pranke clean up NRWT logging/metering, be less verbose https://bugs.webkit.org/show_bug.cgi?id=73846 Reviewed by Ojan Vafai. This patch totally rewrites the metered_stream module and changes the way it and the printing module now interact with the logging subsystem. The printing module becomes a dumb layout_tests-specific wrapper around the metered stream. All of the intelligence of how and when to actually write (or erase) lines and how things interact with logging moves into the metered stream. As of now, anything written to the metered stream *bypasses the logging system*. If a logger is passed to the meter constructor, the metered stream will also attach a log handler so that all messages fed to the logger will also be written to the metered stream. This detangles the logic that used to exist where sometimes the meter would write to the logger and sometimes the logger would write to the meter. All of this makes the logic generic and reusable for test-webkitpy (coming in another patch). It also helps clarify what the mysterious "configure_logging" parameter to the printer constructor used to do (that parameter itself is gone now, and by default the printer and metered_stream will not ever touch the logging system; this makes a bunch of tests less fragile). All of the above should produce no user-visible changes. However, to also fix the bug in question, we change the message format in --verbose mode to %H:%M:%S.%frac %pid %message, eliminating the date, file:lineno, and level parameters. I am making this change at the same time because it would be difficult to write things consistently between the logs and the metered stream if we were still logging the file:lineno. * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py: Remove the configure_logging parameter * Scripts/webkitpy/layout_tests/controllers/worker.py: Use metered stream to configure the logging (we don't need a printer) and be careful about when and how we manipulate the log level and handlers on the root logger. * Scripts/webkitpy/layout_tests/run_webkit_tests.py: Pass in the root logger so that log messages are formatted properly, and set the log level outside of any routines called by unit tests so as to not interfere. * Scripts/webkitpy/layout_tests/views/manager_worker_broker.py: remove a duplicate log message. * Scripts/webkitpy/layout_tests/views/metered_stream.py: complete rewrite :). * Scripts/webkitpy/layout_tests/views/metered_stream_unittest.py: complete rewrite :). * Scripts/webkitpy/layout_tests/views/printing.py: move the logging configuration into metered_stream, and also move the "when should I log progress messages" into metered stream. * Scripts/webkitpy/layout_tests/views/printing_unittest.py: remove tests that are no longer necessary. * Scripts/webkitpy/performance_tests/perftestsrunner.py: remove the configure_logging parameter. * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py: update the test; since the printer no longer writes to the log, we need to check what is written to the regular_output instead. (test_run_test_pause_before_testing): 2012-03-24 Gustavo Noronha Silva Add myself and Martin Robinson to: * GStreamerGraphics * GtkWebKit2PublicAPI * SoupNetwork and to the newly created WebKitGTKTranslations. Rubber-stamped by Martin Robinson. * Scripts/webkitpy/common/config/watchlist: 2012-03-24 Jesus Sanchez-Palencia [Qt][WK2] Implement PageClient::isViewWindowActive() https://bugs.webkit.org/show_bug.cgi?id=81143 Reviewed by Kenneth Rohde Christiansen. Fix the WrapperWindow from our PlatformWebView so it correctly creates the platform related stuff for QWindow, which is not created unless QWindow::setVisible() or QWindow::show() are called. * WebKitTestRunner/qt/PlatformWebViewQt.cpp: (WTR::PlatformWebView::PlatformWebView): 2012-03-23 Dirk Pranke nrwt: don't wait for safari to exit before exiting after showing the results file https://bugs.webkit.org/show_bug.cgi?id=81845 Reviewed by Ryosuke Niwa. Prior to this change, if you ran new-run-webkit-tests and displayed the HTML results file at the end, we would block waiting for the user to quit the browser. There doesn't seem to be a need for that, and the Chromium ports don't do this. Also, update the mac tests to capture the output and be quiet :). * Scripts/webkitpy/common/system/executive_mock.py: Make popen() testable, implement should_log for it. * Scripts/webkitpy/layout_tests/port/mac.py: (MacPort.show_results_html_file): * Scripts/webkitpy/layout_tests/port/mac_unittest.py: Silence some of the tests via outputcapture. 2012-03-23 Gustavo Noronha Silva [GTK] libgcrypt and p11-kit should not be in jhbuild modules https://bugs.webkit.org/show_bug.cgi?id=82073 Reviewed by Martin Robinson. * Scripts/webkitdirs.pm: (jhbuildConfigurationChanged): new function that isolates the jhbuild configuration change checking from the autogen logic. (mustReRunAutogen): code for the new function lived here. (buildAutotoolsProject): now calls jhbuildConfigurationChanged itself, and cleans up the jhbuild root if configuration has changed; it also forces autogen to be run in that case, to make sure the new libraries are used. * gtk/jhbuild.modules: removed libgcrypt and p11-kit. * jhbuild/jhbuild-wrapper: (ensure_jhbuild): this might be run inside a jhbuild environment, in which case aclocal fails when trying to use the now deleted aclocal directory in the jhbuild prefix, so work around that. 2012-03-23 Ojan Vafai garden-o-matic scrolls to the top of the page when you click an accordion item https://bugs.webkit.org/show_bug.cgi?id=82088 Reviewed by Adam Barth. Check the scroll position and make sure that the clicked accordion item is visible. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results.js: 2012-03-23 Dirk Pranke LayoutTestHelper could get torn down earlier (mostly to reset color profile) https://bugs.webkit.org/show_bug.cgi?id=80567 Reviewed by Ryosuke Niwa. Move clean_up_run to occur right after we finish testing and before we start processing the results. Also makes set_up_run() and clean_up_run() private since there's no real need for callers to be aware of them. * Scripts/webkitpy/layout_tests/controllers/manager.py: (Manager._set_up_run): (Manager.run): (Manager._clean_up_run): * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py: (ManagerTest.test_http_locking): * Scripts/webkitpy/layout_tests/run_webkit_tests.py: (run): 2012-03-23 Dean Jackson Disable CSS_SHADERS in Apple builds https://bugs.webkit.org/show_bug.cgi?id=81996 Reviewed by Simon Fraser. Default enableCSSShaders to be 0 on all builds that use this script. Use the flag to turn it back on. * Scripts/build-webkit: 2012-03-23 Tony Chang [chromium] rename newwtf target back to wtf https://bugs.webkit.org/show_bug.cgi?id=82064 Reviewed by Adam Barth. * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp: 2012-03-23 Jessie Berlin Assertion failure in PlatformPasteboardMac constructor causing multiple tests to "crash" on the Lion Intel WebKit2 Debug bots. https://bugs.webkit.org/show_bug.cgi?id=81012. Rubber-stamped by Dan Bernstein. Add a step to kill the pasteboard server while we investigate what is causing these assertion failures (to get the bots green). * BuildSlaveSupport/kill-old-processes: 2012-03-22 Martin Robinson [check-webkit-style] Alphabetical sorting errors in headers are reported for the line after the first out of order header https://bugs.webkit.org/show_bug.cgi?id=81986 Reviewed by David Levin. Track whether or not errors are filtered in error handlers. If an alphabetical sorting error is reported for a header and filtered, try reporting it for the other potential source of the error. This ensures that the style bot can still find new sorting errors in both potential situations: #include <-- 1. edited line #include <-- 2. edited line * Scripts/webkitpy/style/checkers/cpp.py: (check_include_line): Potentially report the error for both lines, if the first error is filtered. * Scripts/webkitpy/style/checkers/cpp_unittest.py: Implement a new style of test that allows only reporting errors for certain lines. This verifies that when one line of an ordering error is filtered, the error is reported on the other line. (ErrorCollector.__init__): (ErrorCollector.__call__): (CppStyleTestBase.perform_lint): (CppStyleTestBase.perform_language_rules_check): (CppStyleTestBase.assert_language_rules_check): (OrderOfIncludesTest.test_check_alphabetical_include_order_errors_reported_for_both_lines): Added a new test that verifies that when one line of the two lines of an ordering error occur, the error is reported on the other line. * Scripts/webkitpy/style/checkers/jsonchecker_unittest.py: (MockErrorHandler.__call__): Report True because the error is handled. * Scripts/webkitpy/style/checkers/test_expectations_unittest.py: (ErrorCollector.__call__): Ditto. * Scripts/webkitpy/style/checkers/watchlist_unittest.py: (MockErrorHandler.__call__): Ditto. * Scripts/webkitpy/style/checkers/xcodeproj_unittest.py: (TestErrorHandler.__call__): Ditto. * Scripts/webkitpy/style/checkers/xml_unittest.py: (MockErrorHandler.__call__): Ditto. * Scripts/webkitpy/style/error_handlers.py: (DefaultStyleErrorHandler.__call__): Report True if the error is handled and False if it is filtered. 2012-03-23 Patrick Gansterer Build fix for WinCE after r111778. * WinCELauncher/CMakeLists.txt: 2012-03-23 Peter Beverloo Unreviewed buildfix for the Chromium Android builder. Disable the ImageDiff target for now. Android needs to build this for host rather than for target (as other configurations do), but since the WTF move the dependency chain got significantly more complicated. I'll work out a proper fix in the following bug, but since several things are broken due to this, this buildfix should be landed first. https://bugs.webkit.org/show_bug.cgi?id=82039 * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp: 2012-03-23 Grzegorz Czajkowski [EFL][DRT] Implement LayoutTestController's methods related with editing. https://bugs.webkit.org/show_bug.cgi?id=80136 Reviewed by Ryosuke Niwa. Implements EFL's LayoutTestController methods e.i. setSmartInsertDeleteEnabled, setSelectTrailingWhitespaceEnabled. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::setSmartInsertDeleteEnabled): (LayoutTestController::setSelectTrailingWhitespaceEnabled): 2012-03-22 Gyuyoung Kim Convert hasGrammarMarker to use Internals interface https://bugs.webkit.org/show_bug.cgi?id=82004 Reviewed by Ryosuke Niwa. Remove hasGrammarMarker functions, because it is able to work in the cross-port way through the Internals interface. * DumpRenderTree/LayoutTestController.cpp: (LayoutTestController::staticFunctions): * DumpRenderTree/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp: * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: * DumpRenderTree/mac/LayoutTestControllerMac.mm: * DumpRenderTree/win/LayoutTestControllerWin.cpp: * DumpRenderTree/wx/LayoutTestControllerWx.cpp: 2012-03-22 Adam Barth sync-master-with-upstream creates non-linear git history https://bugs.webkit.org/show_bug.cgi?id=82006 Reviewed by Julien Chaffraix. Previously this script would create a merge commit at the HEAD of origin/master, causing your fork of WebKit to diverge from upstream. The new version of this script pushes directly from upstream to origin (without round-tripping through any local branches), avoiding this problem. * Scripts/sync-master-with-upstream: 2012-03-22 Gustavo Noronha Silva Unreviewed build fix. Add knowledge of the WTF project to the autotools build system scripts so that build-jsc works correctly for GTK+. * Scripts/webkitdirs.pm: (buildAutotoolsProject): handle the WTF project, build the libWTF.la target. (buildGtkProject): add WTF to the list of supported projects. 2012-03-22 Eric Seidel Make svn-apply re-write JavaScriptCore/wtf to WTF/wtf https://bugs.webkit.org/show_bug.cgi?id=81984 Reviewed by Adam Barth. * Scripts/VCSUtils.pm: (adjustPathForRecentRenamings): 2012-03-22 Raphael Kubo da Costa [CMake] Fix build-jsc after r111778 and r111797. https://bugs.webkit.org/show_bug.cgi?id=82001 Reviewed by Filip Pizlo. Contrary to the comment in build-jsc, buildCMakeProjectOrExit() does return if no error occurred, and build-jsc failed with "building not defined for this platform". * Scripts/build-jsc: Always exit after caling buildCMakeProjectOrExit(). * Scripts/webkitdirs.pm: (buildCMakeProjectOrExit): Return 0 if no error occurred. 2012-03-22 Raphael Kubo da Costa [CMake] Unreviewed build fix after r111778. * DumpRenderTree/efl/CMakeLists.txt: Replace ${JAVASCRIPTCORE_DIR}/wtf with ${WTF_DIR}/wtf in the include paths. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: Use the right wtf include. * DumpRenderTree/efl/WorkQueueItemEfl.cpp: Ditto. * WinCELauncher/CMakeLists.txt: Replace ${JAVASCRIPTCORE_DIR}/wtf with ${WTF_DIR}/wtf in the include paths. 2012-03-22 Filip Pizlo Tools/Scripts/build-jsc doesn't work after WTF move https://bugs.webkit.org/show_bug.cgi?id=81999 Reviewed by Eric Seidel. For ports that require WTF and JSC to be built as two projects and where the machinery to build both does not already exist, teach the tool to do so for us. * Scripts/build-jsc: (buildMyProject): 2012-03-22 Ojan Vafai Fix accidental missing line in r111793. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/base.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/base_unittests.js: Remove base.values. It was unused and duplicates the builtin Object.values. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui.js: Accidentally deleted this line before committing r111793. 2012-03-22 Ojan Vafai Move garden-o-matic progress feedback from non-modal dialogs to a status console anchored to the bottom of the page https://bugs.webkit.org/show_bug.cgi?id=81983 Reviewed by Adam Barth. The dialogs get in the way if you are doing multiple rebaselines. Also, a number of people didn't realize that you could do multiple rebaselines in parallel. A non-modal dialog is just confusing. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/base.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/controllers.js: If we're only rebaselining 1 test, show the test name in the initial message. Otherwise, show the number of tests being rebaselined. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui.js: Turn MessageBox into StatusArea. StatusArea is a singleton and messages in the StatusArea are grouped by ID. That way, the results of each UI action (e.g. clicking rebaseline) are grouped together into a single area. Once addFinalMessage has been called for all IDs, we show the close button. Clicking close also serves to clear all teh content in the StatusArea. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui_unittests.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/styles/results.css: 2012-03-22 Tony Chang Unreviewed, fix chromium build after wtf move. Only use newwtf, remove references to wtf. * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp: 2012-03-22 Csaba Osztrogonác Actually move WTF files to their new home https://bugs.webkit.org/show_bug.cgi?id=81844 [Qt] Unreviewed buildfix after r111778. * DumpRenderTree/qt/DumpRenderTree.pro: 2012-03-22 Dan Bernstein Font fallback in WebKitTestRunner depends on user settings https://bugs.webkit.org/show_bug.cgi?id=81978 Reviewed by Jessie Berlin. * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm: (WTR::InjectedBundle::platformInitialize): Added an AppleLanguages key with a value consisting of the array ("en") to the override defaults dictionary. This is the same value used in DumpRenderTree, and it ensures that font fallback follows a consistent path that is not affected by the language preferences of the user running the test. 2012-03-22 Andy Estes check-for-inappropriate-objc-class-names should ignore the class defined by libarclite. https://bugs.webkit.org/show_bug.cgi?id=81971 Reviewed by Dan Bernstein. Xcode will automatically link ObjC binaries against libarclite in some cases, which defines a class called __ARCLite__. check-for-inappropriate-objc-class-names shouldn't complain about this class name. * Scripts/check-for-inappropriate-objc-class-names: 2012-03-22 Dirk Pranke new-run-webkit-tests blames the wrong test for crashing https://bugs.webkit.org/show_bug.cgi?id=81951 Reviewed by Adam Barth. NRWT wasn't extracting the WebProcess pid correct; should've used re.search instead of re.match :(. Add more tests! * Scripts/webkitpy/layout_tests/port/mac.py: (MacPort._get_crash_log): * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitDriver.has_crashed): (WebKitDriver._check_for_driver_crash): * Scripts/webkitpy/layout_tests/port/webkit_unittest.py: (WebKitDriverTest.test_no_timeout): (WebKitDriverTest): (WebKitDriverTest.test_check_for_driver_crash): (WebKitDriverTest.test_check_for_driver_crash.FakeServerProcess): (WebKitDriverTest.test_check_for_driver_crash.FakeServerProcess.__init__): (WebKitDriverTest.test_check_for_driver_crash.FakeServerProcess.pid): (WebKitDriverTest.test_check_for_driver_crash.FakeServerProcess.name): (WebKitDriverTest.test_check_for_driver_crash.FakeServerProcess.has_crashed): (WebKitDriverTest.test_check_for_driver_crash.assert_crash): 2012-03-22 Sudarsana Nagineni [EFL] [DRT] Implement LayoutTestController::markerTextForListItem() https://bugs.webkit.org/show_bug.cgi?id=81512 Adding missing implementation markerTextForListItem to EFL's LayoutTestController so that we can unskip related tests from the skip list. Reviewed by Antonio Gomes. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::markerTextForListItem): Implemented. 2012-03-22 Sudarsana Nagineni [EFL] Enable view mode media feature layout tests https://bugs.webkit.org/show_bug.cgi?id=81205 Adding missing implementation setViewModeMediaFeature to EFL's LayoutTestController so that we can unskip view mode media feature layout tests. Reviewed by Antonio Gomes. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::setViewModeMediaFeature): Implemented. 2012-03-22 Dirk Pranke nrwt: crash while stopping layout test helper on apple mac lion https://bugs.webkit.org/show_bug.cgi?id=81949 Reviewed by Eric Seidel. stop_helper() didn't work right if the helper process wasn't still running or didn't respond properly to being shut down. Fix this and add some tests. Also add code to MockExecutive() to be able to mock executive.popen() properly. * Scripts/webkitpy/common/system/executive_mock.py: (MockProcess.__init__): (MockProcess.wait): (MockExecutive.__init__): (MockExecutive.popen): * Scripts/webkitpy/common/system/systemhost_mock.py: (MockSystemHost.__init__): * Scripts/webkitpy/layout_tests/port/mac.py: (MacPort.start_helper): (MacPort.stop_helper): * Scripts/webkitpy/layout_tests/port/mac_unittest.py: (test_get_crash_log): (test_helper_starts): (test_helper_fails_to_start): (test_helper_fails_to_stop): (test_helper_fails_to_stop.bad_waiter): 2012-03-22 Gustavo Noronha Silva [GTK] Add make to the jhbuild moduleset https://bugs.webkit.org/show_bug.cgi?id=81769 Reviewed by Martin Robinson. Mostly based on a patch by Carlos Garcia Campos . Add GNU make with custom patches to fix several issues: - make-3.82-arg-list-length.patch: Patch by Ralf Wildenhues to fix the argument list limit of GNU make when running make distcheck. - make-3.82-parallel-build.patch: Patch by Kamil Mierzejewski and Jeremy Devenport to fix parallel build issues present in GNU make 3.82 * gtk/jhbuild.modules: Added make with patches. * gtk/patches/make-3.82-arg-list-length.patch: Added. * gtk/patches/make-3.82-parallel-build.patch: Added. 2012-03-22 Kevin Ollivier [wx] Unreviewed. Adding Source/WTF to the build and updating waf port list. * DumpRenderTree/wscript: * waf/build/settings.py: 2012-03-22 Dirk Pranke Re-land fix for r81603 yet again (!) with latest build fix. https://bugs.webkit.org/show_bug.cgi?id=81603 Unreviewed, build fix. The latest problem was a bad sprintf() in BasePort._get_crash_log(). I've reworked that routine and added unit tests for coverage. * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py: (SingleTestRunner._handle_error): * Scripts/webkitpy/layout_tests/controllers/test_result_writer.py: (write_test_result): (TestResultWriter.write_crash_log): * Scripts/webkitpy/layout_tests/port/base.py: (Port.diff_text): (Port._get_crash_log): * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumDriver.run_test): * Scripts/webkitpy/layout_tests/port/chromium_unittest.py: (ChromiumDriverTest.test_crash_log): * Scripts/webkitpy/layout_tests/port/driver.py: (DriverOutput.__init__): * Scripts/webkitpy/layout_tests/port/mac.py: (MacPort.is_lion): (MacPort._get_crash_log): * Scripts/webkitpy/layout_tests/port/mac_unittest.py: (test_default_child_processes): (test_get_crash_log): * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py: (MockDRTPortTest.test_get_crash_log): * Scripts/webkitpy/layout_tests/port/port_testcase.py: (PortTestCase.test_get_crash_log): * Scripts/webkitpy/layout_tests/port/test.py: (TestDriver.run_test): * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitDriver.__init__): (WebKitDriver._start): (WebKitDriver.has_crashed): (WebKitDriver._check_for_driver_crash): (WebKitDriver.run_test): 2012-03-22 Alexander Færøy [Qt] Enable support for handling database quota's in the MiniBrowser https://bugs.webkit.org/show_bug.cgi?id=81928 Reviewed by Simon Hausmann. * MiniBrowser/qt/qml/BrowserWindow.qml: 2012-03-22 Sudarsana Nagineni [EFL] Implement LayoutTestController::setEditingBehavior https://bugs.webkit.org/show_bug.cgi?id=81124 Adding missing implementation setEditingBehavior to EFL's LayoutTestController. This implementation allows us to unskip some tests from the skip list. Reviewed by Martin Robinson. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::setEditingBehavior): Implemented. 2012-03-22 Carlos Garcia Campos [GTK] Use the angle-bracket form to include wtf headers https://bugs.webkit.org/show_bug.cgi?id=81884 Reviewed by Eric Seidel. Use #include instead of #include "foo". * DumpRenderTree/gtk/AccessibilityCallbacks.cpp: * DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp: * DumpRenderTree/gtk/DumpRenderTree.cpp: * DumpRenderTree/gtk/EditingCallbacks.cpp: * DumpRenderTree/gtk/WorkQueueItemGtk.cpp: * WebKitTestRunner/gtk/EventSenderProxyGtk.cpp: 2012-03-22 Alexander Færøy Unreviewed, add Michael Brüning and Allan Jensen to committers.py. * Scripts/webkitpy/common/config/committers.py: 2012-03-22 Allan Sandfeld Jensen [Qt] Enable FAST_MOBILE_SCROLLING when scrolling is delegated. https://bugs.webkit.org/show_bug.cgi?id=81889 Reviewed by Kenneth Rohde Christiansen. Enable FAST_MOBILE_SCROLLING feature. * qmake/mkspecs/features/features.prf: 2012-03-22 Raphael Kubo da Costa [EFL] DRT: Do not pass a relative path to DOWNLOADED_FONTS_DIR. https://bugs.webkit.org/show_bug.cgi?id=81900 Reviewed by Gustavo Noronha Silva. r111598 added the DOWNLOADED_FONTS_DIR define, but it used a relative path which assumed a certain directory structure for the build directory which may noy be true. Instead, hardcode the path to the downloaded fonts starting from the top-level source directory. While this solution is also far from perfect, it makes it more consistent and reliable, and easier to tackle at once with the rest of bug 81475. * DumpRenderTree/efl/CMakeLists.txt: 2012-03-22 Carlos Garcia Campos [GTK] Allow to run any jhbuild command with jhbuild-wrapper script https://bugs.webkit.org/show_bug.cgi?id=81888 Reviewed by Philippe Normand. * Scripts/update-webkit-libs-jhbuild: (runJhbuild): Helper function to run jhbuild with a given command. Instead of running jhbuild directly, use the jhbuild-wrapper, so that we don't need to run it first to check whether jhbuild is installed or not. * efl/run-with-jhbuild: Call jhbuild-wrapper with run command. * gtk/run-with-jhbuild: Ditto. * jhbuild/jhbuild-wrapper: Allow to execute any jhbuild command, not only run. 2012-03-22 Philippe Normand Unreviewed, test-webkitpy fix after r111661. * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py: (MainTest.test_crash_log): (MainTest.test_web_process_crash_log): 2012-03-22 Philippe Normand Unreviewed, test-webkitpy fix after r111642. * Scripts/webkitpy/layout_tests/port/chromium_unittest.py: (ChromiumDriverTest.setUp): (ChromiumDriverTest.test_two_drivers): 2012-03-22 Csaba Osztrogonác webkitpy: clean up actually getting crash logs from DRT/WTR crashes https://bugs.webkit.org/show_bug.cgi?id=81603 Unreviewed rolling out r111609 and part of r111615, because it broke NRWT on Qt-WK2 platform. * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py: (SingleTestRunner._handle_error): * Scripts/webkitpy/layout_tests/controllers/test_result_writer.py: (write_test_result): (TestResultWriter.write_crash_report): * Scripts/webkitpy/layout_tests/port/base.py: (Port.is_crash_reporter): (Port._driver_class): * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumDriver.run_test): * Scripts/webkitpy/layout_tests/port/chromium_unittest.py: (ChromiumDriverTest.test_crashed_process_name): * Scripts/webkitpy/layout_tests/port/driver.py: (DriverOutput.__init__): * Scripts/webkitpy/layout_tests/port/mac.py: (MacPort): (MacPort.is_crash_reporter): (MacPort.release_http_lock): * Scripts/webkitpy/layout_tests/port/test.py: (TestDriver.run_test): * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitDriver.__init__): (WebKitDriver._start): (WebKitDriver.has_crashed): (WebKitDriver._check_for_driver_crash): (WebKitDriver.run_test): 2012-03-22 Csaba Osztrogonác [Qt] Add full platforms to Qt buildslaves https://bugs.webkit.org/show_bug.cgi?id=81877 Reviewed by Ryosuke Niwa. * BuildSlaveSupport/build.webkit.org-config/config.json: 2012-03-21 Johnny Ding Customize layout test timeout value for different ports. https://bugs.webkit.org/show_bug.cgi?id=79859 Reviewed by Ojan Vafai. * Scripts/webkitpy/layout_tests/controllers/manager.py: (Manager): * Scripts/webkitpy/layout_tests/port/base.py: (Port.default_test_timeout_ms): * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumDriver.stop): * Scripts/webkitpy/layout_tests/port/chromium_android.py: (ChromiumAndroidPort.default_test_timeout_ms): * Scripts/webkitpy/layout_tests/run_webkit_tests.py: (_set_up_derived_options): 2012-03-21 Mark Rowe Fix the build. Ensure that our test tools don't try and use RTTI unnecessarily. * DumpRenderTree/mac/Configurations/Base.xcconfig: * TestWebKitAPI/Configurations/Base.xcconfig: * WebKitTestRunner/Configurations/Base.xcconfig: 2012-03-21 Adam Barth WebKit should have some scripts to smooth interactions with GitHub https://bugs.webkit.org/show_bug.cgi?id=81842 Reviewed by Eric Seidel. These scripts support the workflow in https://trac.webkit.org/wiki/UsingGitHub by automating the configuration and syncing operations needed for a GitHub fork of WebKit to track the main WebKit repository. * Scripts/configure-github-as-upstream: Added. * Scripts/sync-master-with-upstream: Added. 2012-03-21 Dirk Pranke WTR - log the pid of a crashing WebProcess https://bugs.webkit.org/show_bug.cgi?id=81575 Reviewed by Alexey Proskuryakov. This allows run-webkit-tests to find the right crash log for a given crashed WebProcess in WebKit2. * WebKitTestRunner/TestController.cpp: (WTR::exceededDatabaseQuota): (WTR::TestController::createOtherPage): (WTR::TestController::initialize): (WTR::TestController::resetStateToConsistentValues): (WTR::TestController::runTest): (WTR::TestController::processDidCrash): 2012-03-21 Enrica Casucci WebKitURLWithTitles pasteboard format should support URLs containing Emoji characters. https://bugs.webkit.org/show_bug.cgi?id=81835 Reviewed by Brady Eidson. * TestWebKitAPI/Tests/mac/ContextMenuCanCopyURL.html: * TestWebKitAPI/Tests/mac/ContextMenuCanCopyURL.mm: (TestWebKitAPI::contextMenuCopyLink): 2012-03-21 Alexandru Chiculita [CSS Shaders] Make CSS Shaders compile on Chromium https://bugs.webkit.org/show_bug.cgi?id=81435 Reviewed by Stephen White. I've added WebKitCSSCustomFilterEnabled support for the layoutTestController.overridePreference function. * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::overridePreference): * DumpRenderTree/chromium/WebPreferences.cpp: (WebPreferences::reset): (WebPreferences::applyTo): * DumpRenderTree/chromium/WebPreferences.h: (WebPreferences): 2012-03-21 Dirk Pranke fix regression on gtk port introduced by r111609 https://bugs.webkit.org/show_bug.cgi?id=81839 Reviewed by Philippe Normand. r111609 should've initialized a couple of missing fields in WebKitDriver, and also should've checked for method overrides in the other ports :(. * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkDriver._start): * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitDriver.__init__): * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py: (MainTest.test_crash_log): (MainTest.test_web_process_crash_log): 2012-03-21 Dirk Pranke webkitpy: clean up actually getting crash logs from DRT/WTR crashes https://bugs.webkit.org/show_bug.cgi?id=81603 Reviewed by Adam Barth. Add a new _get_crash_log() overridable method on the Port object to customize how we fetch crash logs per port. Mac will now slow-spin with a timeout waiting for the crash log, rather than waiting for the ReportCrash process to exit (it appears that the mac will manage multiple crashes with a single ReportCrash process, the process waits around longer than necessary, presumably to avoid thrashing if processes are repeatedly crashing). Also, add the DriverOutput should contain the crash log and other info, which is created in a port-specific manner but can then be treated generically. Previously single_test_runner would get told that something crashed and attempt to do something to get the crash log, but it didn't have the information it needed to od the right thing; better to make the driver hand back the right info. * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py: (SingleTestRunner._handle_error): * Scripts/webkitpy/layout_tests/controllers/test_result_writer.py: (write_test_result): (TestResultWriter.write_crash_log): * Scripts/webkitpy/layout_tests/port/base.py: (Port.diff_text): (Port._get_crash_log): * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumDriver.run_test): * Scripts/webkitpy/layout_tests/port/chromium_unittest.py: (ChromiumDriverTest.test_crash_log): * Scripts/webkitpy/layout_tests/port/driver.py: (DriverOutput.__init__): * Scripts/webkitpy/layout_tests/port/mac.py: (MacPort.is_lion): (MacPort._get_crash_log): * Scripts/webkitpy/layout_tests/port/test.py: (TestDriver.run_test): * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitDriver.__init__): (WebKitDriver._start): (WebKitDriver.has_crashed): (WebKitDriver._check_for_driver_crash): (WebKitDriver.run_test): 2012-03-21 Dominik Röttsches [EFL] Use jhbuild downloaded fonts instead of hardcoded system font paths https://bugs.webkit.org/show_bug.cgi?id=81507 Now that we're only loading from one directory, no need to iterate over a list of directories anymore. Reviewed by Martin Robinson. * DumpRenderTree/efl/CMakeLists.txt: * DumpRenderTree/efl/FontManagement.cpp: (addFontDirectory): (addFontsToEnvironment): 2012-03-21 Dirk Pranke webkitpy: get ServerProcess out of the reportcrash business https://bugs.webkit.org/show_bug.cgi?id=81600 Unreviewed, build fix. Re-land r111307 and r111293 with another fix for a crash in NRWT; we need to check if the driver has crashed before attempting to read from it. * Scripts/webkitpy/layout_tests/port/server_process.py: (ServerProcess._reset): (ServerProcess._handle_possible_interrupt): (ServerProcess.write): (ServerProcess.read_stdout): (ServerProcess.has_crashed): (ServerProcess._read): (ServerProcess.stop): * Scripts/webkitpy/layout_tests/port/server_process_unittest.py: (TrivialMockPort.check_for_leaks): (TestServerProcess.test_broken_pipe): * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitPort._read_image_diff): (WebKitDriver.has_crashed): (WebKitDriver._check_for_driver_crash): (WebKitDriver.run_test): (WebKitDriver._read_block): * Scripts/webkitpy/layout_tests/port/webkit_unittest.py: (MockServerProcess.__init__): (MockServerProcess): (MockServerProcess.has_crashed): 2012-03-21 Zeno Albisser [Qt][Mac] ranlib segfaults when creating symbol tables for libWebCore.a. https://bugs.webkit.org/show_bug.cgi?id=81750 Building WebCore on mac currently creates an archive that is bigger than 4GB. But ranlib fails to create a symbol table for such a big archive, even on 64bit machines. Therefore we start using SVGAllInOne.cpp when building debug on mac. This reduces the size of the archive by about 300MB. Reviewed by Tor Arne Vestbø. * qmake/mkspecs/features/mac/default_pre.prf: 2012-03-21 Szilard Ledan Updated the address list of committers.py. Reviewed by Csaba Osztrogonác. * Scripts/webkitpy/common/config/committers.py: 2012-03-21 W. James MacLean Updating my nick in committers file. Unreviewed. * Scripts/webkitpy/common/config/committers.py: 2012-03-21 Peter Beverloo [Chromium] Remove Android build-fix when the proper fix rolled into WebKit https://bugs.webkit.org/show_bug.cgi?id=80861 Reviewed by Tony Chang. Remove most Android-specific logic from webkitdirs.pm in favor of handling this during project generation time. * Scripts/webkitdirs.pm: (buildChromiumMakefile): 2012-03-21 Carlos Garcia Campos [GTK] Allow running javascript from location bar in MiniBrowser https://bugs.webkit.org/show_bug.cgi?id=81331 Reviewed by Martin Robinson. * MiniBrowser/gtk/BrowserWindow.c: (activateUriEntryCallback): Use browser_window_load_uri(). (browser_window_load_uri): Check whether uri starts with javascript: and run the given script, or just call webkit_web_view_load_uri() to load the given uri. * MiniBrowser/gtk/BrowserWindow.h: * MiniBrowser/gtk/main.c: (createBrowserWindow): Use browser_window_load_uri(). 2012-03-21 Carlos Garcia Campos [GTK] Add webkit_web_view_run_javascript() to WebKit2 GTK+ https://bugs.webkit.org/show_bug.cgi?id=75543 Reviewed by Martin Robinson. * MiniBrowser/gtk/GNUmakefile.am: Add javascriptcore_cppflags to MiniBrowser CPP flags. 2012-03-20 Eric Seidel Move wtf/Platform.h from JavaScriptCore to Source/WTF/wtf https://bugs.webkit.org/show_bug.cgi?id=80911 Reviewed by Adam Barth. Update include paths for Source/WTF. * DumpRenderTree/efl/CMakeLists.txt: * DumpRenderTree/qt/DumpRenderTree.pro: * QtTestBrowser/QtTestBrowser.pro: * Scripts/build-webkit: Build Source/WTF on AppleWin as well. 2012-03-20 Adam Barth svn-apply should move inflight patches from WebCore/webaudio to WebCore/Modules/webaudio https://bugs.webkit.org/show_bug.cgi?id=81567 Reviewed by Eric Seidel. This temporary measure should help us avoid disrupting the folks working on WebAudio when we move the webaudio directory to Modules. * Scripts/VCSUtils.pm: (adjustPathForRecentRenamings): (parseGitDiffHeader): (parseSvnDiffHeader): 2012-03-20 Dirk Pranke Unreviewed, rolling out r111453. http://trac.webkit.org/changeset/111453 https://bugs.webkit.org/show_bug.cgi?id=81600 still broken * Scripts/webkitpy/layout_tests/port/server_process.py: (ServerProcess._reset): (ServerProcess.handle_interrupt): (ServerProcess.write): (ServerProcess._check_for_crash): (ServerProcess._handle_timeout): (ServerProcess._check_for_abort): (ServerProcess._read): (ServerProcess.stop): (ServerProcess): (ServerProcess.set_crashed): * Scripts/webkitpy/layout_tests/port/server_process_unittest.py: (TrivialMockPort.is_crash_reporter): (TestServerProcess.test_broken_pipe): * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitPort._read_image_diff): (WebKitDriver.has_crashed): (WebKitDriver._check_for_driver_crash): (WebKitDriver): (WebKitDriver._detected_crash): (WebKitDriver._subprocess_crashed): (WebKitDriver._crashed_process_name): (WebKitDriver.run_test): (WebKitDriver._read_block): * Scripts/webkitpy/layout_tests/port/webkit_unittest.py: (MockServerProcess.__init__): 2012-03-20 Dirk Pranke webkitpy: get ServerProcess out of the reportcrash business https://bugs.webkit.org/show_bug.cgi?id=81600 Unreviewed, build fix. Re-land r111307 and r111293 with fix for crash in NRWT; I was failing to set the crashing process name. * Scripts/webkitpy/layout_tests/port/server_process.py: (ServerProcess._reset): (ServerProcess._handle_possible_interrupt): (ServerProcess.write): (ServerProcess.read_stdout): (ServerProcess.has_crashed): (ServerProcess._read): (ServerProcess.stop): * Scripts/webkitpy/layout_tests/port/server_process_unittest.py: (TrivialMockPort.check_for_leaks): (TestServerProcess.test_broken_pipe): * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitPort._read_image_diff): (WebKitDriver.has_crashed): (WebKitDriver._check_for_driver_crash): (WebKitDriver.run_test): (WebKitDriver._read_block): * Scripts/webkitpy/layout_tests/port/webkit_unittest.py: (MockServerProcess.__init__): (MockServerProcess): (MockServerProcess.has_crashed): 2012-03-20 Tim Horton [mac] Restore color space switching code to run-webkit-tests https://bugs.webkit.org/show_bug.cgi?id=80571 Reviewed by Simon Fraser. http://trac.webkit.org/changeset/111429 broke tools build on Snow Leopard. Use the old CM* API on Snow Leopard, since CGDisplayCreateUUIDFromDisplayID didn't exist until Lion. * DumpRenderTree/mac/LayoutTestHelper.m: (installLayoutTestColorProfile): (restoreUserColorProfile): 2012-03-20 Adele Peterson Update the last test to use EXPECT_WK_STREQ. Reviewed by Dan Bernstein. * TestWebKitAPI/Tests/mac/AttributedString.mm: (TestWebKitAPI::TEST): 2012-03-20 Adele Peterson "Attempt to insert nil value " exception when calling attributed string APIs on content with a custom font https://bugs.webkit.org/show_bug.cgi?id=81630 Reviewed by Alexey Proskuryakov. * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: * TestWebKitAPI/Tests/mac/Ahem.ttf: Added. * TestWebKitAPI/Tests/mac/AttributedString.mm: Added. (-[AttributedStringTest webView:didFinishLoadForFrame:]): (TestWebKitAPI): (TestWebKitAPI::didFinishLoadForFrame): (TestWebKitAPI::TEST): * TestWebKitAPI/Tests/mac/attributedStringCustomFont.html: Added. 2012-03-20 Eric Seidel Add link on queues.webkit.org to kov's fancy queue-depth graph https://bugs.webkit.org/show_bug.cgi?id=81690 Reviewed by Adam Barth. * QueueStatusServer/templates/recentstatus.html: 2012-03-20 Tim Horton [mac] Restore color space switching code to run-webkit-tests https://bugs.webkit.org/show_bug.cgi?id=80571 Reviewed by Simon Fraser. Make use of the layout test helper hooks to restore the Mac system-wide color profile switching code. This makes WkTR and DRT results match, and causes the main display's color profile to not affect results on Lion. * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: * DumpRenderTree/mac/LayoutTestHelper.m: Added. (installLayoutTestColorProfile): (restoreUserColorProfile): (simpleSignalHandler): (main): * Scripts/webkitpy/layout_tests/port/mac.py: (MacPort.release_http_lock): (MacPort): (MacPort._path_to_helper): (MacPort.start_helper): (MacPort.stop_helper): * WebKitTestRunner/mac/PlatformWebViewMac.mm: (WTR::PlatformWebView::PlatformWebView): (WTR::PlatformWebView::windowSnapshotImage): 2012-02-24 Darin Fisher Add more Chromium WebKit API reviewers https://bugs.webkit.org/show_bug.cgi?id=81576 Reviewed by Adam Barth and David Levin. Simplify watchlist to lump all WebKit API into one. * Scripts/webkitpy/common/config/watchlist: 2012-03-20 Gyuyoung Kim Convert hasSpellingMarker to use Internals interface. https://bugs.webkit.org/show_bug.cgi?id=81300 Reviewed by Ryosuke Niwa. Remove hasSpellingMarker functions, because it is able to work in the cross-port way through the Internals interface. * DumpRenderTree/LayoutTestController.cpp: (LayoutTestController::staticFunctions): * DumpRenderTree/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp: * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::LayoutTestController): * DumpRenderTree/chromium/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: * DumpRenderTree/mac/LayoutTestControllerMac.mm: * DumpRenderTree/qt/LayoutTestControllerQt.cpp: * DumpRenderTree/qt/LayoutTestControllerQt.h: * DumpRenderTree/win/LayoutTestControllerWin.cpp: * DumpRenderTree/wx/LayoutTestControllerWx.cpp: 2012-03-20 Csaba Osztrogonác Add new Qt-WK2 performance tester bot to build.webkit.org https://bugs.webkit.org/show_bug.cgi?id=80842 Reviewed by Ryosuke Niwa. * BuildSlaveSupport/build.webkit.org-config/config.json: * BuildSlaveSupport/build.webkit.org-config/master.cfg: (RunAndUploadPerfTestsWebKit2): (RunAndUploadPerfTestsWebKit2.start): (BuildAndPerfTestWebKit2Factory): (BuildAndPerfTestWebKit2Factory.__init__): (DownloadAndPerfTestWebKit2Factory): (DownloadAndPerfTestWebKit2Factory.__init__): 2012-03-20 Peter Beverloo [Chromium] Update to Android NDK r7b https://bugs.webkit.org/show_bug.cgi?id=81005 Reviewed by Adam Barth. Update to version r7b of the Android NDK, release notes of which are available on the following page: http://developer.android.com/sdk/ndk/index.html * Scripts/update-webkit-chromium: 2012-03-20 Csaba Osztrogonác x86 GCC bug triggers a compilation error in generated copy constructor of CollapsedBorderValue https://bugs.webkit.org/show_bug.cgi?id=81502 Unreviewed. * qmake/mkspecs/features/unix/default_post.prf: Reenable Werror after r111171. 2012-03-19 Jessie Berlin Unreviewed; roll out http://trac.webkit.org/changeset/111307 and http://trac.webkit.org/changeset/111293 because they broke running the layout tests on mac. * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitDriver._check_for_driver_crash): 2012-03-16 Martin Robinson [GTK] Allow running run-gtk-tests during 'make distcheck' https://bugs.webkit.org/show_bug.cgi?id=81415 Reviewed by Philippe Normand. Modify the build and run-gtk-tests so that it can run during 'make distcheck' * GNUmakefile.am: Distribute run-gtk-tests during 'make dist' and make it the main command run during 'make check.' * Scripts/run-gtk-tests: Refactored run-gtk-tests so that jhbuild is no longer a requirement to run it. If jhbuild isn't around it will simply not start the SPI daemons and skip the WebKit2 accessibility test. Remove test runs from the main loop and simply use the main loop in a synchronous method to start SPI daemons. Did a bunch of general cleanup around these tasks. 2012-03-19 Raphael Kubo da Costa jhbuild: Use the multiprocessing module to get the number of CPUs. https://bugs.webkit.org/show_bug.cgi?id=81585 Reviewed by Martin Robinson. Now that support for Python 2.5 has been dropped, we can safely use multiprocessing.cpu_count() to obtain the current number of CPUs in the system instead of running a webkitperl script for that. * efl/common.py: (top_level_path): * efl/jhbuildrc: * gtk/common.py: (build_path): * gtk/jhbuildrc: 2012-03-19 Ojan Vafai We should only add a live event listener once, not once per TestSelector. Adding one per TestSelector would cause accordion('option', 'active') to incorrectly be "false" if you examine multiple sets of tests. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results.js: 2012-03-19 Dirk Pranke Fix typo introduced in r111293. Unreviewed, build fix. * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitDriver._check_for_driver_crash): 2012-03-19 Dirk Pranke webkitpy: get ServerProcess out of the reportcrash business https://bugs.webkit.org/show_bug.cgi?id=81600 Reviewed by Adam Barth. ServerProcess was half-aware that ReportCrash might run sometimes, and that the process ServerProcess was talking to might have its own crashing subprocesses; neither of these things really worked right and it made the logic convoluted, so this change makes handling crashes completely separate from the server_process code, so that it can focus on just I/O to the subprocess. There should be no functional changes resulting from this patch. * Scripts/webkitpy/layout_tests/port/server_process.py: (ServerProcess._reset): (ServerProcess._handle_possible_interrupt): (ServerProcess.write): (ServerProcess.read_stdout): (ServerProcess.has_crashed): (ServerProcess._read): (ServerProcess.stop): * Scripts/webkitpy/layout_tests/port/server_process_unittest.py: (TrivialMockPort.check_for_leaks): (TestServerProcess.test_broken_pipe): * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitPort._read_image_diff): (WebKitDriver.has_crashed): (WebKitDriver._check_for_driver_crash): (WebKitDriver.run_test): (WebKitDriver._read_block): * Scripts/webkitpy/layout_tests/port/webkit_unittest.py: (MockServerProcess.__init__): (MockServerProcess): (MockServerProcess.has_crashed): 2012-03-19 Dirk Pranke webkitpy: crashlog parsing is broken https://bugs.webkit.org/show_bug.cgi?id=81586 Reviewed by Adam Barth. Seeking on a file opened through codecs() appears to not work correctly (at least on SL). The code was more complicated than it needed to be, so I have revamped it to just read a full crashlog at once and then look for matches. I have also added the ability to optionally propagate errors back (which can be helpful to debug races when ReportCrash is still running and you need to see that you couldn't open some files, rather than just ignoring them). * Scripts/webkitpy/common/system/crashlogs.py: (CrashLogs.find_newest_log): (CrashLogs._find_newest_log_darwin): * Scripts/webkitpy/common/system/crashlogs_unittest.py: (CrashLogsTest.test_find_log_darwin): 2012-03-19 Dirk Pranke webkitpy: clean up some port._filesystem references to not reference private members https://bugs.webkit.org/show_bug.cgi?id=81595 Reviewed by Eric Seidel. There should be no functional changes in this patch, just some refactoring. Also, fixes a minor nit in write_image_diff_files(() * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py: (SingleTestRunner.__init__): (SingleTestRunner._run_compare_test): (SingleTestRunner._run_rebaseline): (SingleTestRunner._save_baseline_data): (SingleTestRunner._handle_error): (SingleTestRunner._run_reftest): * Scripts/webkitpy/layout_tests/controllers/test_result_writer.py: (write_test_result): (TestResultWriter.__init__): (TestResultWriter._make_output_directory): (TestResultWriter.output_filename): (TestResultWriter._output_testname): (TestResultWriter.write_output_files): (TestResultWriter.write_stderr): (TestResultWriter.write_crash_report): (TestResultWriter.create_text_diff_and_write_result): (TestResultWriter.write_image_diff_files): (copy_file): * Scripts/webkitpy/layout_tests/controllers/test_result_writer_unittest.py: (TestResultWriterTest.test_reftest_diff_image): 2012-03-19 Dirk Pranke webkitpy: fix nit - incorrect reference to DRT in test_expectations.py https://bugs.webkit.org/show_bug.cgi?id=81598 Reviewed by Eric Seidel. "DumpRenderTree" shouldn't be hard-coded here, since some other process may actually be crashing (WebKitTestRunner, WebProcess, etc.) * Scripts/webkitpy/layout_tests/models/test_expectations.py: (TestExpectations): 2012-03-19 Dirk Pranke webkitpy: allow instance data in TestFailure objects https://bugs.webkit.org/show_bug.cgi?id=81580 Reviewed by Tony Chang. Current TestFailures are basically enums, but it would be nice, especially for crashes, if they contained the name and pid of the process that crashed. * Scripts/webkitpy/layout_tests/models/test_failures.py: (TestFailure.message): (FailureTimeout.message): (FailureCrash): (FailureCrash.__init__): (FailureCrash.message): (FailureMissingResult.message): (FailureTextMismatch.message): (FailureMissingImageHash.message): (FailureMissingImage.message): (FailureImageHashMismatch.message): (FailureImageHashIncorrect.message): (FailureReftestMismatch.message): (FailureReftestMismatchDidNotOccur.message): (FailureReftestNoImagesGenerated.message): (FailureMissingAudio.message): (FailureAudioMismatch.message): * Scripts/webkitpy/layout_tests/models/test_failures_unittest.py: (TestFailuresTest.test_equals): (TestFailuresTest): (TestFailuresTest.test_crashes): 2012-03-19 Brady Eidson and https://bugs.webkit.org/show_bug.cgi?id=81066 Crash in 3rd party WebKit apps under XHR/Cache code Reviewed by Antti Koivisto. * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: * TestWebKitAPI/Tests/mac/MemoryCachePruneWithinResourceLoadDelegate.html: Added. * TestWebKitAPI/Tests/mac/MemoryCachePruneWithinResourceLoadDelegate.mm: Added. (-[MemoryCachePruneTestResourceLoadDelegate webView:identifierForInitialRequest:fromDataSource:]): (-[MemoryCachePruneTestResourceLoadDelegate webView:resource:willSendRequest:redirectResponse:fromDataSource:]): (-[MemoryCachePruneTestResourceLoadDelegate webView:resource:didFinishLoadingFromDataSource:]): (-[MemoryCachePruneTestResourceLoadDelegate webView:resource:didFailLoadingWithError:fromDataSource:]): (TestWebKitAPI): (TestWebKitAPI::TEST): 2012-03-19 Brady Eidson and https://bugs.webkit.org/show_bug.cgi?id=81516 REGRESSION (r107435) URLs copied from WebKit apps aren't in the right pasteboard format Reviewed by Enrica Casucci. Bring up a context menu on a link, copy the link, then verify the pasteboard contents are in the correct format. * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: * TestWebKitAPI/Tests/mac/ContextMenuCanCopyURL.html: Added. * TestWebKitAPI/Tests/mac/ContextMenuCanCopyURL.mm: Added. (-[ContextMenuCanCopyURLDelegate webView:didFinishLoadForFrame:]): (TestWebKitAPI::contextMenuCopyLink): (TestWebKitAPI::TEST): 2012-03-19 Ojan Vafai Make the selected item in the accordion a link to the flakiness dashboard https://bugs.webkit.org/show_bug.cgi?id=81571 Reviewed by Dimitri Glazkov. Also, make it so that you can select the text of the link, instead of the current behavior of acting as if you had clicked on the accordion header. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/styles/results.css: 2012-03-19 Dirk Pranke NRWT runs some tests that are skipped with -i command line option https://bugs.webkit.org/show_bug.cgi?id=81535 Reviewed by Ojan Vafai. This change modifies the interaction of Skipped files and test_expectations files so that entries in Skipped files (and the -i command line) override *everything* in the expectations file. Specifically, a directory in a Skipped file will cause all of the tests in the dir to be skipped even if individual tests in the dir are listed in the test_expectations.txt. Skipped files also override anything in an overrides files. This seems to make more intuitive sense, since if you list something in the Skips file (and even more specify it on the command line) you probably want it to be universally applied. Theoretically we could add more precedence levels and have full paths in an expectations file override dirs in a Skipped file (but not the command line), but I don't know yet that that level of complexity is justified. * Scripts/webkitpy/layout_tests/models/test_expectations.py: (TestExpectations.__init__): (TestExpectations._add_skipped_tests): (TestExpectations._add_test): (TestExpectations._add_expectations): (TestExpectations._add_expectation_line): (TestExpectations._already_seen_better_match): * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: (test_more_specific_override_resets_skip): (SkippedTests): (SkippedTests.get_exp): (SkippedTests.assert_exp): (SkippedTests.test_skipped_tests_work): (SkippedTests.test_duplicate_skipped_test_fails_lint): (SkippedTests.test_skipped_file_overrides_expectations): (SkippedTests.test_skipped_dir_overrides_expectations): (SkippedTests.test_skipped_file_overrides_overrides): (SkippedTests.test_skipped_dir_overrides_overrides): 2012-03-19 Eric Seidel Fix WTF header include discipline in Chromium WebKit https://bugs.webkit.org/show_bug.cgi?id=81281 Reviewed by James Robinson. * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp: * Scripts/update-webkit-chromium: Made it possible to run this from a directory other than the root. 2012-03-19 Andy Chen [BlackBerry] Add new layout test for text match marker https://bugs.webkit.org/show_bug.cgi?id=81531 Make the LayoutTestControllerBlackBerry::findString to test the BlackBerry port implementation. Reviewed by Rob Buis. * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp: (LayoutTestController::findString): 2012-03-19 Ojan Vafai Don't show WONTFIX tests in garden-o-matic https://bugs.webkit.org/show_bug.cgi?id=81554 Reviewed by Dimitri Glazkov. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/results.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/results_unittests.js: 2012-03-19 Chris Fleizach platform/mac/accessibility/selected-rows-table.html failing on Lion Intel Debug WebKit2 testers https://bugs.webkit.org/show_bug.cgi?id=81530 Reviewed by Jessie Berlin. Add rowAtIndex to WebKitTestRunner. * WebKitTestRunner/InjectedBundle/AccessibilityUIElement.cpp: (WTR::AccessibilityUIElement::rowAtIndex): * WebKitTestRunner/InjectedBundle/AccessibilityUIElement.h: (AccessibilityUIElement): * WebKitTestRunner/InjectedBundle/Bindings/AccessibilityUIElement.idl: * WebKitTestRunner/InjectedBundle/mac/AccessibilityUIElementMac.mm: (WTR): (WTR::AccessibilityUIElement::rowAtIndex): 2012-03-19 Ojan Vafai Change garden-o-matic All Failures tab to Expected Failures https://bugs.webkit.org/show_bug.cgi?id=81540 Reviewed by Dimitri Glazkov. Also, fix some broken unittests that had just gotten stale. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/controllers.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/controllers/rebaseline.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/garden-o-matic.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/model.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/results.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/results_unittests.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/failures_unittests.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/notifications_unittests.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui_unittests.js: 2012-03-19 Hao Zheng [chromium] DRT crashes at shutdown. https://bugs.webkit.org/show_bug.cgi?id=81504 Reviewed by James Robinson. WebCompositor must be destroyed after WebView is properly destroyed, or some compositor objects like CCLayerTreeHost may not be destroyed. However, it is really hard to make the destruction order correct explicitly in ~TestShell. So move it out to DumpRenderTree. * DumpRenderTree/chromium/DumpRenderTree.cpp: (main): * DumpRenderTree/chromium/TestShell.cpp: (TestShell::~TestShell): 2012-03-19 Dirk Pranke add dpranke to webkitpy watchlist https://bugs.webkit.org/show_bug.cgi?id=81433 Reviewed by Benjamin Poulain. * Scripts/webkitpy/common/config/watchlist: 2012-03-19 Dirk Pranke test-webkitpy is failing on SL/Python 2.6.1 https://bugs.webkit.org/show_bug.cgi?id=81533 Reviewed by Ojan Vafai. It appears that itertools.combinations() is buggy in 2.6.1; this backports the implementation from 2.7. * Scripts/webkitpy/layout_tests/models/test_configuration.py: (TestConfigurationConverter.collapse_macros): (TestConfigurationConverter): (TestConfigurationConverter.combinations): 2012-03-19 David Michael Barr Optimize the cleanup sequence in the start-queue.sh script https://github.com/abarth/webkit/pull/1 Reviewed by Adam Barth. Updating the working copy is an expensive operation for a repository the size of WebKit. Attempt to reset only to the target revision. * EWSTools/start-queue.sh: 2012-03-19 Raphael Kubo da Costa Adjust architecture type of the EFL build bot. https://bugs.webkit.org/show_bug.cgi?id=81452 Reviewed by Tony Chang. * BuildSlaveSupport/build.webkit.org-config/config.json: The buildslave is a 64-bit machine, therefore describe it as being x86_64, not i386. 2012-03-19 Alexander Færøy [Qt] printLoadedUrls prints "Object" instead of the actual URL. https://bugs.webkit.org/show_bug.cgi?id=81514 Reviewed by Simon Hausmann. * MiniBrowser/qt/qml/BrowserWindow.qml: 2012-03-19 Raphael Kubo da Costa [EFL] jhbuild: Set CMAKE_PREFIX_PATH in jhbuildrc. https://bugs.webkit.org/show_bug.cgi?id=81491 Reviewed by Philippe Normand. For the find_path(), find_library() etc CMake calls to find the dependencies built with jhbuild, CMAKE_PREFIX_PATH must be properly set, otherwise CMake cannot know the dependencies exist and will only look in the standard directories. * efl/jhbuildrc: Set the CMAKE_PREFIX_PATH environment variable with the jhbuild Root directory. 2012-03-19 Dominik Röttsches [EFL] Add XDG env vars to jhbuildrc https://bugs.webkit.org/show_bug.cgi?id=81510 Needed to ensure correct glib behaviour in jhbuild environment. Reviewed by Philippe Normand. * efl/jhbuildrc: 2012-03-19 Allan Sandfeld Jensen Select best target for tap gesture. https://bugs.webkit.org/show_bug.cgi?id=78801 Reviewed by Kenneth Rohde Christiansen. Add TOUCH_ADJUSTMENT to enabled features. * qmake/mkspecs/features/features.prf: 2012-03-19 Robert Kroeger [chromium] synthesize wheel events for fling on main thread https://bugs.webkit.org/show_bug.cgi?id=81458 Added the ability to dispatch GestureFlingStart and GestureFlingCancel from EventSender. Reviewed by James Robinson. * DumpRenderTree/chromium/EventSender.cpp: (EventSender::EventSender): (EventSender::gestureFlingCancel): (EventSender::gestureFlingStart): * DumpRenderTree/chromium/EventSender.h: (EventSender): 2012-03-19 Alexander Færøy [Qt] Add command line option to set the user agent in the MiniBrowser https://bugs.webkit.org/show_bug.cgi?id=81508 Reviewed by Simon Hausmann. * MiniBrowser/qt/BrowserWindow.cpp: (BrowserWindow::BrowserWindow): (BrowserWindow::webViewExperimental): * MiniBrowser/qt/BrowserWindow.h: (BrowserWindow): * MiniBrowser/qt/MiniBrowserApplication.cpp: (printHelp): (MiniBrowserApplication::handleUserOptions): * MiniBrowser/qt/MiniBrowserApplication.h: (WindowOptions::WindowOptions): (WindowOptions::setUserAgent): (WindowOptions::userAgent): (WindowOptions): 2012-03-19 Alexander Færøy [Qt] Remove dead user agent code from the MiniBrowser. Reviewed by Simon Hausmann. * MiniBrowser/qt/BrowserWindow.cpp: * MiniBrowser/qt/BrowserWindow.h: (BrowserWindow): * MiniBrowser/qt/MiniBrowser.qrc: * MiniBrowser/qt/useragentlist.txt: Removed. 2012-03-19 Raphael Kubo da Costa jhbuild: Call jhbuild's autogen.sh via bash. https://bugs.webkit.org/show_bug.cgi?id=81485 Reviewed by Philippe Normand. jhbuild's autogen.sh uses "/bin/bash" in the shebang line, however this is usually true only on Linux systems, and even then bash for some reason might not be available in that location. It is safer to call bash directly and pass autogen.sh to it instead. * jhbuild/jhbuild-wrapper: (install_jhbuild): 2012-03-19 Philippe Normand [GTK] Prefix run-launcher with run-with-jhbuild https://bugs.webkit.org/show_bug.cgi?id=81503 Reviewed by Hajime Morita. * Scripts/run-launcher: 2012-03-19 Dominik Röttsches [EFL] Prefix run-launcher with run-with-jhbuild https://bugs.webkit.org/show_bug.cgi?id=80026 Reviewed by Hajime Morita. * Scripts/run-launcher: 2012-03-18 Csaba Osztrogonác [Qt] Unreviewed. Disable -Werror on x86 until proper fix. https://bugs.webkit.org/show_bug.cgi?id=81498 * qmake/mkspecs/features/unix/default_post.prf: 2012-03-17 Raphael Kubo da Costa Unreviewed; switch to using my FreeBSD.org email address, the other one is going to disappear soon. * Scripts/webkitpy/common/config/committers.py: * Scripts/webkitpy/common/config/watchlist: 2012-03-17 Raphael Kubo da Costa Unreviewed; fix webkitpy unit test failures after r111116. It turns out that cmd_line in the patch in bug 80025 had the right signature when it was submitted, but r109242 added a few more required parameters, and since the method (and EflDriver) did not exist at the time we ended up committing a broken method. * Scripts/webkitpy/layout_tests/port/efl.py: (EflDriver.cmd_line): Add the `pixel_tests' and `per_test_args' arguments. 2012-03-17 Charles Wei [BlackBerry] Enable Web Timing for performance profiling and improvement. https://bugs.webkit.org/show_bug.cgi?id=81085 Reviewed by Rob Buis. * Scripts/build-webkit: 2012-03-17 Dominik Röttsches Add a dependencies installation step for EFL to buildbot master.cfg https://bugs.webkit.org/show_bug.cgi?id=81337 Reviewed by Gustavo Noronha Silva. * BuildSlaveSupport/build.webkit.org-config/master.cfg: (InstallEflDependencies): (Factory.__init__): 2012-03-17 Dominik Röttsches [EFL] Prefixing run-webkit-tests with run-with-jhbuild https://bugs.webkit.org/show_bug.cgi?id=80025 Reviewed by Gustavo Noronha Silva. * Scripts/webkitpy/layout_tests/port/efl.py: (EflDriver): (EflDriver.cmd_line): (EflPort._driver_class): 2012-03-17 Dominik Röttsches [EFL] Add and use run-with-jhbuild and update-webkitefl-libs scripts for EFL https://bugs.webkit.org/show_bug.cgi?id=79904 build-webkit --efl will use jhbuild for fetching and building dependencies. Reviewed by Gustavo Noronha Silva. * Scripts/update-webkitefl-libs: Added. * Scripts/webkitdirs.pm: (jhbuildWrapperPrefixIfNeeded): (generateBuildSystemFromCMakeProject): (buildCMakeGeneratedProject): * efl/common.py: Added. (script_path): (top_level_path): (number_of_cpus): * efl/jhbuild.modules: Added. * efl/jhbuildrc: Added. * efl/run-with-jhbuild: Added. 2012-03-16 Stephanie Lewis https://bugs.webkit.org/show_bug.cgi?id=81065 CrashTracer: [USER] 111 crashes in WebProcess at WebKitTestRunnerInjectedBundle: WTR::InjectedBundle::done + 142 Fix crash preventing WebKitTestRunner from working in Release. WebKitTestRunner used ostringstream to collect output from the test. ostringstream is incompatible with our FastMalloc overrides. When copying strings out ostringstream a new string is allocated with FastMalloc because the allocation is inlined. When that string is freed is uses the system malloc because that call is not inlined. I removed ostringstream in favor of StringBuilder. Properly set NDEBUG on Release builds so we don't crash when when we free the strings StringBuilder created. Reviewed by Geoff Garen. Add DEBUG_DEFINES to set NDEBUG on release builds. * WebKitTestRunner/Configurations/Base.xcconfig: * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj: Remove ostringstream member in favor of StringBuilder. * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp: (WTR::InjectedBundle::initialize): (WTR::InjectedBundle::beginTesting): (WTR::InjectedBundle::done): * WebKitTestRunner/InjectedBundle/InjectedBundle.h: (WTF): (WTR::InjectedBundle::stringBuilder): (InjectedBundle): Append output to StringBuilder. * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp: (WTR::hasPrefix): (WTR::dumpPath): (WTR::rangeToStr): (WTR::styleDecToStr): (WTR::frameToStr): (WTR::dumpFrameScrollPosition): (WTR::dumpFrameText): (WTR::dumpDescendantFramesText): (WTR::InjectedBundlePage::dump): (WTR::InjectedBundlePage::didReceiveTitleForFrame): (WTR::InjectedBundlePage::didFinishDocumentLoadForFrame): (WTR::InjectedBundlePage::willSendRequestForFrame): (WTR::InjectedBundlePage::decidePolicyForNavigationAction): (WTR::lastFileURLPathComponent): (WTR::InjectedBundlePage::willAddMessageToConsole): (WTR::InjectedBundlePage::willSetStatusbarText): (WTR::InjectedBundlePage::willRunJavaScriptAlert): (WTR::InjectedBundlePage::willRunJavaScriptConfirm): (WTR::InjectedBundlePage::willRunJavaScriptPrompt): (WTR::InjectedBundlePage::shouldBeginEditing): (WTR::InjectedBundlePage::shouldEndEditing): (WTR::InjectedBundlePage::shouldInsertNode): (WTR::InjectedBundlePage::shouldInsertText): (WTR::InjectedBundlePage::shouldDeleteRange): (WTR::InjectedBundlePage::shouldChangeSelectedRange): (WTR::InjectedBundlePage::shouldApplyStyle): (WTR::InjectedBundlePage::didBeginEditing): (WTR::InjectedBundlePage::didEndEditing): (WTR::InjectedBundlePage::didChange): (WTR::InjectedBundlePage::didChangeSelection): (WTR::InjectedBundlePage::supportsFullScreen): (WTR::InjectedBundlePage::enterFullScreenForElement): (WTR::InjectedBundlePage::exitFullScreenForElement): (WTR::InjectedBundlePage::beganEnterFullScreen): (WTR::InjectedBundlePage::beganExitFullScreen): (WTR::dumpBackForwardListItem): (WTR::InjectedBundlePage::dumpBackForwardList): * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: (WTR::LayoutTestController::waitToDumpWatchdogTimerFired): (WTR::LayoutTestController::dumpConfigurationForViewport): Convert between WTF::String and WKStringRef. * WebKitTestRunner/StringFunctions.h: (WTR::toWTFString): (WTR): * WebKitTestRunner/TestInvocation.cpp: (WTR::TestInvocation::didReceiveMessageFromInjectedBundle): 2012-03-16 Eric Seidel Add a land-from-url command https://bugs.webkit.org/show_bug.cgi?id=81411 Reviewed by Adam Barth. Very basic so far. The goal is to make it landing smarter so that users don't have to bother to parse out the ids out of the urls. * Scripts/webkitpy/common/config/urls.py: (parse_bug_id): (parse_attachment_id): * Scripts/webkitpy/common/config/urls_unittest.py: (URLsTest.test_parse_bug_id): (URLsTest): (URLsTest.test_parse_attachment_id): * Scripts/webkitpy/tool/commands/download.py: (ProcessURLsMixin): (ProcessURLsMixin._fetch_list_of_patches_to_process): (LandFromURL): * Scripts/webkitpy/tool/commands/download_unittest.py: (test_land_from_bug): (test_land_from_url): 2012-03-16 Dave Tharp build-webkit launcher instructions use wrong relative path for Qt https://bugs.webkit.org/show_bug.cgi?id=81357 Reviewed by Eric Seidel. Ensure we are at the proper directory before generating run-launcher instructions. * Scripts/build-webkit: 2012-03-16 Brady Eidson and https://bugs.webkit.org/show_bug.cgi?id=81412 REGRESSION (r107435) Copy a link and paste to Mail: Nothing is pasted Reviewed by Geoff Garen. * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: * TestWebKitAPI/Tests/mac/WebViewCanPasteURL.mm: Added. 2012-03-16 Ojan Vafai Mark wontfix tests in full_results.json so we can ignore them in garden-o-matic https://bugs.webkit.org/show_bug.cgi?id=81407 Reviewed by Adam Barth. * Scripts/webkitpy/layout_tests/controllers/manager.py: (summarize_results): * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py: (ResultSummaryTest.summarized_results): (ResultSummaryTest.test_no_svn_revision): (ResultSummaryTest.test_svn_revision): (ResultSummaryTest): (ResultSummaryTest.test_summarized_results_wontfix): 2012-03-16 Ojan Vafai Have webkit-patch rebaseline-test update test_expectations.txt https://bugs.webkit.org/show_bug.cgi?id=81054 Recommit http://trac.webkit.org/changeset/110783 now that http://trac.webkit.org/changeset/111034 has landed. * Scripts/webkitpy/layout_tests/models/test_expectations.py: (TestExpectations.remove_configuration_from_test): * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: (RemoveConfigurationsTest): (RemoveConfigurationsTest.test_remove): (test_remove_line): * Scripts/webkitpy/tool/commands/rebaseline.py: (RebaselineTest._update_expectations_file): (RebaselineTest._rebaseline_test_and_update_expectations): (RebaselineTest): (RebaselineTest.execute): * Scripts/webkitpy/tool/commands/rebaseline_unittest.py: (TestRebaseline.test_rebaseline_updates_expectations_file_noop): (test_rebaseline_updates_expectations_file): 2012-03-15 Ojan Vafai Specifier collapsing when writing test expectations lines gets a number of cases wrong https://bugs.webkit.org/show_bug.cgi?id=81309 Reviewed by Dimitri Glazkov. I've run this over all the lines in the current Chromium test_expectations.txt file, so I'm relatively confident we now cover all the cases. * Scripts/webkitpy/layout_tests/models/test_configuration.py: (TestConfigurationConverter.__init__): (TestConfigurationConverter.collapse_macros): (TestConfigurationConverter.collapse_macros.collapse_individual_specifier_set): (TestConfigurationConverter): (TestConfigurationConverter.intersect_combination): (TestConfigurationConverter.symmetric_difference): (TestConfigurationConverter.to_specifiers_list): (TestConfigurationConverter.to_specifiers_list.try_collapsing): (TestConfigurationConverter.to_specifiers_list.try_abbreviating): * Scripts/webkitpy/layout_tests/models/test_configuration_unittest.py: (make_mock_all_test_configurations_set): (TestConfigurationConverterTest.test_symmetric_difference): (TestConfigurationConverterTest.test_to_config_set): (TestConfigurationConverterTest.test_macro_expansion): (TestConfigurationConverterTest.test_to_specifier_lists): (TestConfigurationConverterTest.test_converter_macro_collapsing): * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumPort): * Scripts/webkitpy/layout_tests/port/chromium_android.py: (ChromiumAndroidPort.__init__): The android port uses "arm" as it's architecture, which is technically correct, but considerably complicates making collapsing work. We probably should kill the concept of architecture entirely. The benefits are not worth the code complexity. * Scripts/webkitpy/layout_tests/port/chromium_unittest.py: (ChromiumPortTest.test_all_test_configurations): * Scripts/webkitpy/tool/commands/rebaseline_unittest.py: (TestRebaseline.test_rebaseline_updates_expectations_file_noop): (test_rebaseline_updates_expectations_file): 2012-03-16 Dinu Jacob [Qt][Wk2] Assertion Failure and crash on file upload https://bugs.webkit.org/show_bug.cgi?id=80854 Reviewed by Simon Hausmann. Added filePicker to WebView using experimental API. * MiniBrowser/qt/MiniBrowser.qrc: * MiniBrowser/qt/icons/folder.png: Added. * MiniBrowser/qt/icons/titlebar.png: Added. * MiniBrowser/qt/icons/up.png: Added. * MiniBrowser/qt/qml/BrowserWindow.qml: * MiniBrowser/qt/qml/FilePicker.qml: Added. 2012-03-16 Mahesh Kulkarni Updating email for committer.py script. * Scripts/webkitpy/common/config/committers.py: 2012-03-16 Robert Kroeger [chromium] DRT: Add support for sending gesture scrolls without a begin https://bugs.webkit.org/show_bug.cgi?id=81292 Reviewed by Kent Tamura. * DumpRenderTree/chromium/EventSender.cpp: (EventSender::EventSender): (EventSender::gestureScrollFirstPoint): * DumpRenderTree/chromium/EventSender.h: (EventSender): 2012-03-02 Carlos Garcia Campos and Sergio Villar Senin [GTK] WebKitIconDatabase doesn't keep icons cached https://bugs.webkit.org/show_bug.cgi?id=56200 Reviewed by Martin Robinson. Add support for favicons in GtkLauncher. * GtkLauncher/main.c: (activateUriEntryCb): (iconLoadedCb): (createBrowser): 2012-03-16 Adam Barth Remove sheriff-bot's last-green-revision command https://bugs.webkit.org/show_bug.cgi?id=81314 Reviewed by Eric Seidel. This command is a cruel joke. WebKit doesn't really ever have a last-green-revision. * Scripts/webkitpy/common/net/buildbot/buildbot.py: (BuildBot._find_green_revision): * Scripts/webkitpy/common/net/buildbot/buildbot_mock.py: (MockBuildBot.builder_statuses): * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py: (test_find_green_revision): * Scripts/webkitpy/tool/bot/irc_command.py: (IRCCommand.execute): * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py: (SheriffIRCBotTest.test_help): * Scripts/webkitpy/tool/commands/queries.py: (PatchesToReview.execute): 2012-03-16 Tor Arne Vestbø [Qt] Make gccdepends test aware of broken icecream deps generation Icecream seems to build the files without passing -o to gcc with the original destination path, which means the target name in the dependency file is wrong. Also, echo something to the final file instead of just touching it, so it's not deleted on every build. Reviewed by Simon Hausmann. * qmake/config.tests/gccdepends/gccdepends.pro: 2012-03-16 Kihong Kwon Support for Battery Status API https://bugs.webkit.org/show_bug.cgi?id=62698 Add include directory and file. http://www.w3.org/TR/battery-status/ Reviewed by Adam Barth. * Scripts/build-webkit: 2012-03-16 Kishore Bolisetty check-webkit-style fails on LayoutTests/platform/efl/test_expectations.txt https://bugs.webkit.org/show_bug.cgi?id=81007 Reviewed by Adam Barth. A dedidcated build slave is available for efl on build.webkit.org. This change add's an entry for the elf port in list of available port builders. This is required because test_expectations.py when triggered, looks for a valid port object for the specified port. * Scripts/webkitpy/layout_tests/port/builders.py: 2012-03-16 Adam Barth [webkitpy] StatusServer.next_work_item has an incorrectly named local variable https://bugs.webkit.org/show_bug.cgi?id=81302 Reviewed by Eric Seidel. This is a trivial patch, but the variable name turned up in a grep of webkitpy and looks like copy pasta. * Scripts/webkitpy/common/net/statusserver.py: (StatusServer.next_work_item): 2012-03-16 Andrew Lo [BlackBerry] Enable PAGE_VISIBILITY_API https://bugs.webkit.org/show_bug.cgi?id=81257 Enable Page Visibility API for BlackBerry. Reviewed by Rob Buis. * Scripts/build-webkit: 2012-03-16 Zan Dobersek [Gtk] Add computedStyleIncludingVisitedInfo to DumpRenderTreeSupport https://bugs.webkit.org/show_bug.cgi?id=81232 Reviewed by Philippe Normand. Use the newly-added helper method in DumpRenderTreeSupportGtk. * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: (LayoutTestController::computedStyleIncludingVisitedInfo): 2012-03-15 Eric Seidel Move parse_bug_id into config/urls in preparation for re-use elsewhere. https://bugs.webkit.org/show_bug.cgi?id=81313 Reviewed by Adam Barth. * Scripts/webkitpy/common/checkout/changelog.py: (parse_bug_id_from_changelog): * Scripts/webkitpy/common/config/urls.py: (parse_bug_id): * Scripts/webkitpy/common/config/urls_unittest.py: Copied from Tools/Scripts/webkitpy/common/config/urls.py. (URLsTest): (URLsTest.test_parse_bug_id): * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py: (test_parse_bug_id): * Scripts/webkitpy/tool/bot/irc_command.py: (_post_error_and_check_for_bug_url): * Scripts/webkitpy/tool/bot/sheriff.py: (Sheriff.post_rollout_patch): (Sheriff.post_chromium_deps_roll): 2012-03-15 Adam Barth The commit-queue should fast-track patches that have already passed the testing EWS bots https://bugs.webkit.org/show_bug.cgi?id=81305 Reviewed by Eric Seidel. If a patch has already passed the testing EWS bots (currently only the chromium-ews bot), we should be able to land the patch without running the tests again. This patch is somewhat risky in that the tests might have been broken between when the EWS bots ran and when the commit-queue is processing the patch. If that turns out to be a problem in practice, we'll tighten up this fast-track in some way, likely by making sure that the pass status is recent. * Scripts/webkitpy/tool/bot/commitqueuetask.py: (CommitQueueTaskDelegate.did_pass_testing_ews): (CommitQueueTask._did_pass_tests_recently): (CommitQueueTask.run): * Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py: (MockCommitQueue.did_pass_testing_ews): (test_fast_success_case): * Scripts/webkitpy/tool/commands/queues.py: (CommitQueue.did_pass_testing_ews): * Scripts/webkitpy/tool/commands/queues_unittest.py: (test_did_pass_testing_ews): 2012-03-15 Kentaro Hara [PerformanceTests] run-perf-tests should output correct units https://bugs.webkit.org/show_bug.cgi?id=78303 Reviewed by Ryosuke Niwa. Previously run-perf-tests just supported "ms" for units. Consequently, Dromaeo perf tests had been reporting runs/s as "ms". This patch fixes run-perf-tests so that they can support custom units. The default unit is "ms". The test result looks like this: $ ./Tools/Scripts/run-perf-tests Dromaeo/dom-attr.html RESULT Dromaeo: dom-attr= 6465.9525483 runs/s median= 0.0 runs/s, stdev= 24.2983433436 runs/s, min= 6435.87649402 runs/s, max= 6515.63693392 runs/s * Scripts/webkitpy/performance_tests/perftestsrunner.py: (PerfTestsRunner._process_parser_test_result): 2012-03-15 Brent Fulgham [WinCairo] Unreviewed build change after wtf path changes. * DumpRenderTree/win/ImageDiffWinCairoCommon.vsprops: Updated include directories to look at the private header directory to find 'wtf'. 2012-03-15 Kent Tamura REGRESSION(r110595): webkit-patch suggest-reviewers doesn't work https://bugs.webkit.org/show_bug.cgi?id=81293 Reviewed by Adam Barth. r110595 introduced new format of the author field of ChangeLog entry: "Adam Barth && Benjamin Poulain " Updte chanelog.py to support it. * Scripts/webkitpy/common/checkout/changelog.py: (ChangeLogEntry._split_contributor_names): * Scripts/webkitpy/common/checkout/changelog_unittest.py: (test_parse_authors): 2012-03-15 Lynn Neir Implement LayoutTestController::setValueForUser on Windows https://bugs.webkit.org/show_bug.cgi?id=55834 Implemented setValueForUser in both DumpRenderTree and WebKitTestRunner. Reviewed by Jessie Berlin. * DumpRenderTree/win/LayoutTestControllerWin.cpp: (LayoutTestController::setValueForUser): * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl: * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: (WTR::LayoutTestController::setValueForUser): (WTR): * WebKitTestRunner/InjectedBundle/LayoutTestController.h: (LayoutTestController): 2012-03-15 Adam Barth Update the start-queue.sh script to match what we're using on the bots https://bugs.webkit.org/show_bug.cgi?id=81277 Reviewed by Eric Seidel. Over time we've added more magic git commands to clean up the local repo after it has been borked in various ways. The idea behind this sequence is to clear out any changes in the working copy and then to rebuild the master branch from origin/master. This patch also removes the watchdog timer that we've stopped using on the bots. This patch was reviewed on . Note: Eric and I are experimenting with GitHub to see if it might be useful for WebKit. If you have thoughts on this topic, please feel free to grab Eric or me on #webkit to chat. I assume that he and I are the only ones who care about this file so using this alternative process for this patch shouldn't cause any trouble. * EWSTools/start-queue.sh: 2012-03-15 Adam Barth The commit-queue can hang when the test_expectations.txt style check fails during commit https://bugs.webkit.org/show_bug.cgi?id=81251 Reviewed by Eric Seidel. We forgot to check the --non-interactive flag, so we end up prompting the "user" who doesn't exist on the bot. We might consider teaching the User class about non-interactive if we have this sort of bug again. * Scripts/webkitpy/tool/steps/commit.py: (Commit.options): (Commit._check_test_expectations): (Commit.run): * Scripts/webkitpy/tool/steps/commit_unittest.py: (CommitTest.test_check_test_expectations): 2012-03-15 Jessie Berlin fast/dom/Window/window-properties.html and fast/dom/prototype-inheritance-2.html failing on Mac WK2 because WebAudio is not enabled https://bugs.webkit.org/show_bug.cgi?id=81259 Reviewed by Dan Bernstein. Enable WebAudio so that it appears in the window object in the tests. The webaudio tests are all still skipped in the Mac Skipped list. * WebKitTestRunner/TestController.cpp: (WTR::TestController::resetStateToConsistentValues): 2012-03-15 David Levin Various autoresize fixes. https://bugs.webkit.org/show_bug.cgi?id=81199 Reviewed by Darin Fisher. * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::reset): Stop using the deprecated api that was removed. 2012-03-15 Carlos Garcia Campos [GTK] Actually use --release and --debug command line options in run-gtk-test https://bugs.webkit.org/show_bug.cgi?id=81234 Reviewed by Philippe Normand. Actually handle --release and --debug command line options to create the build directory, falling back to current hack to detect whether it's a Release or Debug build when both options are missed. * Scripts/run-gtk-tests: (TestRunner.__init__): Receive options too. (TestRunner._get_top_level_directory): Helper function to get the top level directory. (TestRunner._get_build_directory): Helper function to get the build directory depending on --release/--debug command line options or gessing it if both options are missing. (TestRunner._ensure_accessibility_daemon_is_running): Use self-options now. (TestRunner.run): Ditto. (TestRunner.run.run_tests): Pass options to the constructor instead of run(). 2012-03-15 Carlos Garcia Campos [GTK] Run unit tests with --verbose in the bots https://bugs.webkit.org/show_bug.cgi?id=81230 Reviewed by Martin Robinson. * BuildSlaveSupport/build.webkit.org-config/master.cfg: (RunGtkAPITests): Add --verbose to run-gtk-test command. 2012-03-15 Kevin Ollivier Move wx port to using export macros https://bugs.webkit.org/show_bug.cgi?id=77279 Reviewed by Hajime Morita. * waf/build/settings.py: (common_configure): 2012-03-15 Philippe Normand Unreviewed, GTK build fix attempt after r110845. * Scripts/run-gtk-tests: Added --release and --debug options. 2012-03-15 Dominik Röttsches Parametrize run-with-jhbuild and update-webkitgtk-libs with platform --gtk/--efl https://bugs.webkit.org/show_bug.cgi?id=79673 Preparatory work for bringing jhbuild enabled build to EFL port. Reviewed by Philippe Normand. * Scripts/update-webkit-libs-jhbuild: Copied from Tools/Scripts/update-webkitgtk-libs. * Scripts/update-webkitgtk-libs: * gtk/run-with-jhbuild: * jhbuild/jhbuild-wrapper: Copied from Tools/gtk/run-with-jhbuild. (top_level_path): (jhbuild_installed): (jhbuild_cloned): (jhbuild_at_expected_revision): (update_jhbuild): (clone_jhbuild): (install_jhbuild): (update_webkit_libs_jhbuild): (determine_platform): (ensure_jhbuild): 2012-03-15 Carlos Garcia Campos [GTK] Add command line options to run-gtk-tests script https://bugs.webkit.org/show_bug.cgi?id=81218 Reviewed by Philippe Normand. Following command line options have been added: -v, --verbose: to run gtester with --verbose. That would be used by the bots to make it easier to know which test case it failing. --display: to use a different display for Xvfb It's also possible to pass one or more tests as arguments. * Scripts/run-gtk-tests: (TestRunner.__init__): Receive command line arguments. (TestRunner._get_tests): Get all the tests or the ones passed by command line arguments. (TestRunner._ensure_accessibility_daemon_is_running): Use display command line argument. (TestRunner.run): Pas command line options. (TestRunner.run.run_tests): Pass command line arguments. 2012-03-15 Carlos Garcia Campos [GTK] WebKit2APITests/TestPrinting timeout in the bots https://bugs.webkit.org/show_bug.cgi?id=81211 Reviewed by Philippe Normand. Add gtk+ patch to fix timeout in the bots. A similar patch has already landed in updatream gtk+, see http://git.gnome.org/browse/gtk+/commit/?id=5ebfad5d091f4c9d9c6e60364fad4c60842f027f * gtk/jhbuild.modules: * gtk/patches/gtk+-3.0.12-gtkprinter.patch: Added. 2012-03-14 Gustavo Noronha Silva sheriffbot should also be addressable with a comma in addition to colon https://bugs.webkit.org/show_bug.cgi?id=81101 Reviewed by Adam Barth. * Scripts/webkitpy/common/net/irc/ircbot.py: (IRCBot.on_pubmsg): 2012-03-15 Alexander Færøy [Qt] MiniBrowser scroll indicators should not scroll out of view https://bugs.webkit.org/show_bug.cgi?id=81131 Reviewed by Kenneth Rohde Christiansen. * MiniBrowser/qt/qml/ScrollIndicator.qml: 2012-03-14 Chris Fleizach AX: ARIA tables need to support selected rows attribute/select rows changed notification https://bugs.webkit.org/show_bug.cgi?id=79821 Reviewed by Beth Dakin. * DumpRenderTree/AccessibilityUIElement.cpp: (rowAtIndexCallback): (AccessibilityUIElement::getJSClass): * DumpRenderTree/AccessibilityUIElement.h: (AccessibilityUIElement): * DumpRenderTree/blackberry/AccessibilityUIElementBlackBerry.cpp: (AccessibilityUIElement::rowAtIndex): * DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp: (AccessibilityUIElement::rowAtIndex): * DumpRenderTree/mac/AccessibilityUIElementMac.mm: (AccessibilityUIElement::rowAtIndex): * DumpRenderTree/win/AccessibilityUIElementWin.cpp: (AccessibilityUIElement::rowAtIndex): 2012-03-14 Tim Horton NRWT should respect --additional-platform-directory's Skipped list https://bugs.webkit.org/show_bug.cgi?id=81172 Reviewed by Timothy Hatcher. Respect the Skipped file inside all paths included via the --additional-platform-directory option. Test results from said path are already used, but the Skipped file is ignored. * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitPort._skipped_file_search_paths): * Scripts/webkitpy/layout_tests/port/webkit_unittest.py: (test_skipped_file_search_paths): 2012-03-14 Ojan Vafai Unreviewed, rolling out r110783. http://trac.webkit.org/changeset/110783 https://bugs.webkit.org/show_bug.cgi?id=81054 Caused existing test expectation lines to expand to one for each test configuration. * Scripts/webkitpy/layout_tests/models/test_expectations.py: (TestExpectations.has_warnings): * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: (test_macro_overrides): * Scripts/webkitpy/tool/commands/rebaseline.py: (RebaselineTest._save_baseline): (RebaselineTest._rebaseline_test): (RebaselineTest.execute): * Scripts/webkitpy/tool/commands/rebaseline_unittest.py: (TestRebaseline.test_tests_to_update): 2012-03-13 Jon Lee Separate NOTIFICATIONS and LEGACY_NOTIFICATIONS https://bugs.webkit.org/show_bug.cgi?id=80922 Reviewed by Jian Li. You can include either NOTIFICATIONS or LEGACY_NOTIFICATIONS and have a complete API. LEGACY_NOTIFICATIONS should cover all of the previous functionality, and NOTIFICATIONS will cover the new API. Therefore, APIs that are common between the two will have: #if ENABLE(NOTIFICATIONS) || ENABLE(LEGACY_NOTIFICATIONS) This patch initially sets both to be exactly the same. As other bugs with patches begin to migrate to the new API, the defines will begin to split. This allows ports to decide which set of APIs to include. Update everything to be #if ENABLE(NOTIFICATIONS) || ENABLE(LEGACY_NOTIFICATIONS) * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::grantDesktopNotificationPermission): (LayoutTestController::simulateDesktopNotificationClick): * DumpRenderTree/chromium/NotificationPresenter.cpp: * DumpRenderTree/chromium/TestShell.cpp: (TestShell::TestShell): (TestShell::resetTestController): * DumpRenderTree/win/DRTDesktopNotificationPresenter.cpp: (DRTDesktopNotificationPresenter::checkNotificationPermission): 2012-03-13 Ojan Vafai Have webkit-patch rebaseline-test update test_expectations.txt https://bugs.webkit.org/show_bug.cgi?id=81054 Reviewed by Dirk Pranke. This will allow us to use the All Failures tab in garden-o-matic to rebaseline tests without making test_expectations.txt stale. It's conservative and only removes lines if the test is not flaky and fails in a way the rebaseline would fix (e.g. CRASH/TIMEOUT will not be touched). * Scripts/webkitpy/layout_tests/models/test_expectations.py: (TestExpectations.remove_configuration_from_test): * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: (RemoveConfigurationsTest): (RemoveConfigurationsTest.test_remove): (test_remove_line): * Scripts/webkitpy/tool/commands/rebaseline.py: (RebaselineTest._is_supported_port): (RebaselineTest): (RebaselineTest._update_expectations_file): (RebaselineTest._rebaseline_test_and_update_expectations): (RebaselineTest.execute): * Scripts/webkitpy/tool/commands/rebaseline_unittest.py: (TestRebaseline.test_rebaseline_updates_expectations_file): 2012-03-14 Jessie Berlin Crash in WKTR under addChromeInputField due to using 0 as a key in a HashMap https://bugs.webkit.org/show_bug.cgi?id=81167 Reviewed by Sam Weinig. Do not use 0 as a key in a HashMap. * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: cacheLayoutTestControllerCallback uses the enum as keys for the callback map. 2012-03-14 Ojan Vafai Add the chromium lion bot to garden-o-matic and remove the platform_move_to in builders.py since the Chromium Lion port is now a normal port. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/config.js: * Scripts/webkitpy/layout_tests/port/builders.py: 2012-03-14 Dirk Pranke Fix regression/crash introduced in r110724 - bad type passed from skipped_tests() by default. Unreviewed, build fix. * Scripts/webkitpy/layout_tests/port/base.py: (Port.skipped_tests): 2012-03-14 Dirk Pranke Please add a way to manually skip some tests in NRWT https://bugs.webkit.org/show_bug.cgi?id=81019 Reviewed by Ojan Vafai. This patch implements the -i / --ignore-tests flag from ORWT and refactors the test_expectations.py code slightly to handle it. * Scripts/webkitpy/layout_tests/controllers/manager.py: (Manager.parse_expectations): * Scripts/webkitpy/layout_tests/models/test_expectations.py: (TestExpectations.__init__): * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: (test_add_skipped_tests): (test_add_skipped_tests_duplicate): * Scripts/webkitpy/layout_tests/run_webkit_tests.py: (parse_args): * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py: (MainTest.test_ignore_tests): (MainTest.test_ignore_tests.assert_ignored): 2012-03-14 Dirk Pranke nrwt: run with no args on lion, is putting baselines in platform/mac-lion, not platform/mac https://bugs.webkit.org/show_bug.cgi?id=81028 Reviewed by Ojan Vafai. The mac port had incorrect lookups for 'mac-lion' and 'mac-future' in some places. This patch fixes that and adds a bunch of tests. * Scripts/webkitpy/layout_tests/port/mac.py: (MacPort.baseline_path): * Scripts/webkitpy/layout_tests/port/mac_unittest.py: (MacTest.assert_skipped_file_search_paths): (MacTest.test_skipped_file_search_paths): (_assert_search_path): (test_baseline_search_path): * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitPort._skipped_file_search_paths): 2012-03-13 Jer Noble Lion Intel Debug WebKit2 Tests crashing under [WKFullScreenWindowController _startEnterFullScreenAnimationWithDuration:] https://bugs.webkit.org/show_bug.cgi?id=81056 Reviewed by Jessie Berlin. Intercept beganEnterFullScreen and beganExitFullScreen and turn them into no-ops (plus logging). * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp: (WTR::InjectedBundlePage::InjectedBundlePage): (WTR::InjectedBundlePage::beganEnterFullScreen): (WTR::InjectedBundlePage::beganExitFullScreen): 2012-03-14 Alexander Færøy The width and height of the scroll indicators in the MiniBrowser does not consider the offset margins. https://bugs.webkit.org/show_bug.cgi?id=81112 Reviewed by Kenneth Rohde Christiansen. This patch fixes the issue by withdrawing two times the anchor margins. * MiniBrowser/qt/qml/ScrollIndicator.qml: 2012-03-14 Sudarsana Nagineni [EFL] Implement LayoutTestController::clearBackForwardList https://bugs.webkit.org/show_bug.cgi?id=80142 Adding missing implementation clearBackForwardList to EFL's LayoutTestController. This implementation allows us to unskip tests connected with history. Reviewed by Antonio Gomes. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::clearBackForwardList): Implemented. 2012-03-13 Thiago Marcos P. Santos [EFL][DRT] Expose window.internals object https://bugs.webkit.org/show_bug.cgi?id=79900 Exposes window.internals for EFL's DRT. Makes possible to remove some tests from the skip list. Reviewed by Antonio Gomes. * DumpRenderTree/efl/CMakeLists.txt: * DumpRenderTree/efl/DumpRenderTreeChrome.cpp: (DumpRenderTreeChrome::onWindowObjectCleared): 2012-03-13 Adam Barth && Benjamin Poulain Always enable ENABLE(CLIENT_BASED_GEOLOCATION) https://bugs.webkit.org/show_bug.cgi?id=78853 Reviewed by Adam Barth. * Scripts/build-webkit: * qmake/mkspecs/features/features.prf: 2012-03-12 Antonio Gomes Convert nodesFromRect tests to use Internals interface https://bugs.webkit.org/show_bug.cgi?id=80886 Reviewed by Ryosuke Niwa. This clean up allows to remove port specific bits of each DRT implementation that supports testing Document::nodesFromRect, in favor of making use of the 'Internals' interface to test it. This makes its code conveniently more expansible and cleaner. Also removed stubs from Win and EFL DRTs. * DumpRenderTree/LayoutTestController.cpp: (LayoutTestController::staticFunctions): * DumpRenderTree/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: * DumpRenderTree/mac/LayoutTestControllerMac.mm: * DumpRenderTree/qt/LayoutTestControllerQt.cpp: * DumpRenderTree/qt/LayoutTestControllerQt.h: (LayoutTestController): * DumpRenderTree/win/LayoutTestControllerWin.cpp: 2012-03-13 Ojan Vafai Rebaselining for a new port doesn't work right with multiple fallback ports https://bugs.webkit.org/show_bug.cgi?id=80932 Reviewed by Adam Barth. Since chromium-leopard also falls back to mac-leopard, we need to copy the existing result for both chromium-leopard and chromium-snowleopard before doing lion rebaselines. * Scripts/webkitpy/layout_tests/port/builders.py: (fallback_port_names_for_new_port): * Scripts/webkitpy/tool/commands/rebaseline.py: (RebaselineTest): (RebaselineTest._copy_existing_baseline): (RebaselineTest._rebaseline_test): (RebaselineTest.execute): * Scripts/webkitpy/tool/commands/rebaseline_unittest.py: * Scripts/webkitpy/tool/servers/gardeningserver.py: (GardeningHTTPRequestHandler.rebaseline): * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py: (GardeningServerTest.test_rebaseline_new_port): 2012-03-13 Tor Arne Vestbø [Qt] Tweak minibrowser UI Add some margins to the scroll indicators, and set a dark background color so you see the page dimensions. Reviewed by Antonio Gomes. * MiniBrowser/qt/qml/BrowserWindow.qml: * MiniBrowser/qt/qml/ScrollIndicator.qml: 2012-03-13 ChangSeok Oh [EFL] [DRT] Implement scheduleAsynchronousKeyDown. https://bugs.webkit.org/show_bug.cgi?id=79601 Reviewed by Gustavo Noronha Silva. Extracted shareable code from keyDownCallback and then added createKeyEventInfo. createKeyEventInfo creates a new structure KeyEventInfo that has keyName and modifiers and is taken over to sendKeyDown. Added sendKeyDown and sendAsynchronousKeyDown to evaluate keyDown event directly and asynchronously. Eventually, sendAsynchronousKeyDown calls sendKeyDown. * DumpRenderTree/efl/EventSender.cpp: (KeyEventInfo::KeyEventInfo): (KeyEventInfo): (createKeyEventInfo): (sendKeyDown): (keyDownCallback): (sendAsynchronousKeyDown): (scheduleAsynchronousKeyDownCallback): 2012-03-12 Ojan Vafai webkit-patch optimize-baselines sometimes creates a mac-future result https://bugs.webkit.org/show_bug.cgi?id=80717 Reviewed by Adam Barth. * Scripts/webkitpy/common/checkout/baselineoptimizer.py: (_baseline_search_hypergraph): (BaselineOptimizer._find_optimal_result_placement): (BaselineOptimizer._filter_virtual_ports): * Scripts/webkitpy/common/checkout/baselineoptimizer_unittest.py: (BaselineOptimizerTest.test_no_add_mac_future): 2012-03-12 Jessie Berlin More Windows build fixing. * DumpRenderTree/win/WorkQueueItemWin.cpp: 2012-03-12 Jessie Berlin More Windows build fixing. * DumpRenderTree/win/AccessibilityControllerWin.cpp: * DumpRenderTree/win/EditingDelegate.cpp: * DumpRenderTree/win/EventSender.cpp: * DumpRenderTree/win/FrameLoadDelegate.cpp: * DumpRenderTree/win/LayoutTestControllerWin.cpp: * DumpRenderTree/win/UIDelegate.cpp: 2012-03-12 Dirk Pranke remove the 'gpu' group from the flakiness dashboard https://bugs.webkit.org/show_bug.cgi?id=80353 Reviewed by Adam Barth. Since those bots are obsolete, we don't need this any more. * TestResultServer/static-dashboards/builders.js: * TestResultServer/static-dashboards/flakiness_dashboard_tests.js: (testPlatformAndBuildType): (testHtmlForTestTypeSwitcherGroup): 2012-03-12 Dirk Pranke remove graphics_type from webkitpy.layout_tests.models.test_configuration https://bugs.webkit.org/show_bug.cgi?id=80643 Reviewed by Ojan Vafai. Now that the chromium-gpu ports are gone, there's no need for the graphics_type CPU/GPU distinction. Most of this change is just updating tests accordingly. * Scripts/webkitpy/layout_tests/controllers/test_expectations_editor_unittest.py: (TestExpectationEditorTests): * Scripts/webkitpy/layout_tests/models/test_configuration.py: (TestConfiguration.__init__): (TestConfiguration.category_order): (TestConfiguration.__str__): (TestConfiguration.__repr__): (TestConfiguration.__hash__): (TestConfigurationConverter.to_specifiers_list): * Scripts/webkitpy/layout_tests/models/test_configuration_unittest.py: (make_mock_all_test_configurations_set): (TestConfigurationTest.test_items): (TestConfigurationTest.test_keys): (TestConfigurationTest.test_str): (TestConfigurationTest.test_repr): (TestConfigurationTest.test_hash): (TestConfigurationTest.test_hash.query_unknown_key): (TestConfigurationTest.test_eq): (TestConfigurationTest.test_values): (SpecifierSorterTest.test_specifier_priority): (SpecifierSorterTest.test_sort_specifiers): (TestConfigurationConverterTest.test_to_config_set): (TestConfigurationConverterTest.test_macro_expansion): (TestConfigurationConverterTest.test_to_specifier_lists): (TestConfigurationConverterTest.test_converter_macro_collapsing): * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: (test_error_on_different_build_type): (TestExpectationSerializerTests.test_parsed_to_string): (TestExpectationSerializerTests.test_reconstitute_only_these): * Scripts/webkitpy/layout_tests/port/apple.py: (ApplePort._generate_all_test_configurations): * Scripts/webkitpy/layout_tests/port/base.py: (Port.__init__): (Port.name): (Port.version): (Port.test_configuration): * Scripts/webkitpy/layout_tests/port/builders.py: * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumPort): (ChromiumPort._generate_all_test_configurations): * Scripts/webkitpy/layout_tests/port/chromium_linux.py: (ChromiumLinuxPort.__init__): * Scripts/webkitpy/layout_tests/port/chromium_linux_unittest.py: (ChromiumLinuxPortTest.test_check_illegal_port_names): * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py: (ChromiumMacPortTest.test_baseline_path): * Scripts/webkitpy/layout_tests/port/chromium_unittest.py: (ChromiumPortTest.test_all_test_configurations): * Scripts/webkitpy/layout_tests/port/efl.py: (EflPort._generate_all_test_configurations): * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkPort._generate_all_test_configurations): * Scripts/webkitpy/layout_tests/port/qt.py: (QtPort._generate_all_test_configurations): * Scripts/webkitpy/layout_tests/port/test.py: (TestPort.all_test_configurations): (TestPort._all_build_types): * Scripts/webkitpy/tool/commands/rebaseline_unittest.py: * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py: (BuildCoverageExtrapolatorTest.test_extrapolate): (GardeningExpectationsUpdaterTest.test_unknown_failure_type): (GardeningExpectationsUpdaterTest.test_add_new_expectation): (GardeningExpectationsUpdaterTest.test_replace_old_expectation): (GardeningExpectationsUpdaterTest.test_pass_expectation): (GardeningExpectationsUpdaterTest.test_supplement_old_expectation): (GardeningExpectationsUpdaterTest.test_spurious_updates): 2012-03-12 Dirk Pranke _runtime_feature_list isn't supported on Apple Mac DRT https://bugs.webkit.org/show_bug.cgi?id=80906 Reviewed by Eric Seidel. It looks like this feature is actually only supported on the Win DRT port, so this patch reworks the code to make that clearer and reduce unnecessary overrides. * Scripts/webkitpy/layout_tests/port/efl.py: (EflPort._path_to_webcore_library): * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkPort._path_to_webcore_library): * Scripts/webkitpy/layout_tests/port/qt.py: (QtPort._skipped_file_search_paths): * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitPort._runtime_feature_list): * Scripts/webkitpy/layout_tests/port/webkit_unittest.py: (TestWebKitPort.__init__): (TestWebKitPort.all_test_configurations): (test_skipped_directories_for_features): * Scripts/webkitpy/layout_tests/port/win.py: (WinPort): (WinPort._runtime_feature_list): * Scripts/webkitpy/layout_tests/port/win_unittest.py: (WinPortTest.test_operating_system): (WinPortTest): (WinPortTest.test_runtime_feature_list): 2012-03-12 Peter Beverloo [Chromium] Temporary build-fix for the Android bot https://bugs.webkit.org/show_bug.cgi?id=80862 Reviewed by Adam Barth. Manually set the environmental variables required to compile source files through the Android NDK instead of relying on the envsetup.sh script. A proper fix will be landed to Chromium ASAP, and following a Chromium roll into WebKit this code will be removed. * Scripts/webkitdirs.pm: (buildChromiumMakefile): 2012-03-12 Tor Arne Vestbø [Qt] Use right construct to detect config_test_gccdepends Reviewed by Andreas Kling. * qmake/mkspecs/features/default_pre.prf: 2012-03-12 Tor Arne Vestbø [Qt] Expose location of WebKit's bin and imports dir through module file Reviewed by Andreas Kling. * qmake/qt_webkit.pri: 2012-03-12 Andras Becsi [Qt][WK2] Add support for rudimentary scroll indicators in MiniBrowser https://bugs.webkit.org/show_bug.cgi?id=80832 Reviewed by Tor Arne Vestbø. Add a basic scroll indicator component to MiniBrowser so that it behaves similar to the ScrollDecorator QML component. * MiniBrowser/qt/MiniBrowser.qrc: * MiniBrowser/qt/qml/BrowserWindow.qml: * MiniBrowser/qt/qml/ScrollIndicator.qml: Added. 2012-03-12 C Anthony Risinger [GTK] r110296 included an extra $ in @$(AM_V_GEN) https://bugs.webkit.org/show_bug.cgi?id=80809 Reviewed by Philippe Normand. * GNUmakefile.am: Add `$` to @(AM_V_GEN), else /bin/sh syntax error 2012-03-12 Ming Xie [BlackBerry] build-webkit doesn't recognize "--cmakearg" for the BlackBerry port https://bugs.webkit.org/show_bug.cgi?id=80812 Reviewed by Rob Buis. * Scripts/build-webkit: Add $cmakeArgs to buildCMakeProjectOrExit() * Scripts/webkitdirs.pm: (blackberryCMakeArguments): Remove "PUBLIC_BUILD=0" define 2012-03-12 Simon Hausmann [Qt] Fix build with newer Qt 5. Reviewed by Tor Arne Vestbø. Newer Qt 5 renames QDeclarativeFoo to QQmlFoo and provides backwards compatible renames, which however do trigger a #warning. Until we _require_ a newer Qt 5 we should disable -Werror. Also load(moc) for the availability of QMAKE_MOC in various derived sources. The implicit load was removed recently. * qmake/mkspecs/features/default_pre.prf: * qmake/mkspecs/features/unix/default_post.prf: 2012-03-11 Maciej Stachowiak Add --instruments option to sunspider to profile with Instruments https://bugs.webkit.org/show_bug.cgi?id=80783 Reviewed by Filip Pizlo. * Scripts/run-sunspider: 2012-03-11 Viatcheslav Ostapenko [Qt] [WK2] Support threaded renderer in WK2 https://bugs.webkit.org/show_bug.cgi?id=76661 Remove disabling of threaded rendering. Reviewed by Noam Rosenthal. * MiniBrowser/qt/main.cpp: (main): 2012-03-09 Zan Dobersek [WK2] run-perf-tests should be able to run with WTR https://bugs.webkit.org/show_bug.cgi?id=80042 Reviewed by Martin Robinson. Add an option to PerfTestRunner to use WebKitTestRunner instead of DumpRenderTree and add the '--no-timeout' flag to WebKitTestRunner to avoid timing out when running performance tests. * Scripts/webkitpy/performance_tests/perftestsrunner.py: (PerfTestsRunner._parse_args): * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp: (WTR::InjectedBundle::InjectedBundle): Initialize m_useWaitToDumpWatchdogTimer to true. (WTR::InjectedBundle::didReceiveMessage): Unpack the 'BeginTest' message as a dictionary containing DumpPixels and UseWaitToDumpWatchdogTimer keys. * WebKitTestRunner/InjectedBundle/InjectedBundle.h: (WTR::InjectedBundle::useWaitToDumpWatchdogTimer): Make m_useWaitToDumpWatchdogTimer publicly accessible. * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: (WTR::LayoutTestController::waitUntilDone): Only initialize the watchdog if it is required. * WebKitTestRunner/TestController.cpp: (WTR::TestController::TestController): Initialize m_useWaitToDumpWatchdogTimer to true. (WTR::TestController::initialize): Set m_useWaitToDumpWatchdogTimer to false if '--no-timeout' option is present. (WTR::TestController::runUntil): Translate TimeoutDuration to a proper timeout value, or -1.0 if timeout should not occur. * WebKitTestRunner/TestController.h: (WTR::TestController::useWaitToDumpWatchdogTimer): Make m_useWaitToDumpWatchdogTimer publicly accessible. * WebKitTestRunner/TestInvocation.cpp: (WTR::TestInvocation::invoke): Pack the 'BeginTest' message as a dictionary, containing DumpPixels and UseWaitToDumpWatchdogTimer keys and corresponding values. Waiting for final message should not time out if not necessary. * WebKitTestRunner/gtk/TestControllerGtk.cpp: (WTR::TestController::platformRunUntil): Only set up a timeout if necessary. * WebKitTestRunner/qt/TestControllerQt.cpp: (WTR::TestController::platformRunUntil): Ditto. * WebKitTestRunner/mac/TestControllerMac.mm: (WTR::TestController::platformRunUntil): Add a FIXME, indicating functionality is missing. * WebKitTestRunner/win/TestControllerWin.cpp: (WTR::TestController::platformRunUntil): Ditto. 2012-03-09 Jon Lee Rename NotificationPresenter to NotificationClient https://bugs.webkit.org/show_bug.cgi?id=80488 Reviewed by Kentaro Hara. Refactor to use renamed WebCore::NotificationClient. * DumpRenderTree/win/DRTDesktopNotificationPresenter.cpp: (DRTDesktopNotificationPresenter::checkNotificationPermission): 2012-03-07 James Robinson [chromium] Deprecate renderDirectlyToWebView parameter of WebViewClient::createGraphicsContext3D() https://bugs.webkit.org/show_bug.cgi?id=80523 Reviewed by Darin Fisher. * DumpRenderTree/chromium/WebViewHost.cpp: (WebViewHost::createGraphicsContext3D): * DumpRenderTree/chromium/WebViewHost.h: (WebViewHost): 2012-03-09 Eric Seidel build-webkit should build Source/WTF on Mac https://bugs.webkit.org/show_bug.cgi?id=80728 Reviewed by Adam Barth. We'll add other ports to this list in future commits. * Scripts/build-webkit: 2012-03-09 Jon Lee Add support for ENABLE(LEGACY_NOTIFICATIONS) https://bugs.webkit.org/show_bug.cgi?id=80497 Reviewed by Adam Barth. Prep for b80472: Update API for Web Notifications * Scripts/build-webkit: * qmake/mkspecs/features/features.prf: 2012-03-09 Ashod Nakashian Bash scripts should support LF endings only https://bugs.webkit.org/show_bug.cgi?id=79509 Reviewed by David Kilzer. * EWSTools/boot.sh: Added properties svn:executable and svn:eol-style. * EWSTools/start-queue.sh: Added property svn:eol-style. * Scripts/webkit-tools-completion.sh: Added properties svn:executable and svn:eol-style. * WebKitTestRunner/win/build-generated-files.sh: Added properties svn:executable and svn:eol-style. * iExploder/iexploder-1.7.2/tools/release_src.sh: Added property svn:eol-style. * iExploder/iexploder-1.7.2/tools/update_html_tags_from_sources.sh: Added property svn:eol-style. 2012-03-09 Ionut Biru [BUILD] use detected python to run generate-gtkdoc https://bugs.webkit.org/show_bug.cgi?id=80695 Reviewed by Gustavo Noronha Silva. * GNUmakefile.am: Use $(PYTHON) instead of plain `python` 2012-03-09 Nat Duca Make check-webkit-style complain about "no new tests" lines in changelogs https://bugs.webkit.org/show_bug.cgi?id=80582 Reviewed by David Levin. * Scripts/webkitpy/style/checkers/changelog.py: (ChangeLogChecker.check_entry): * Scripts/webkitpy/style/checkers/changelog_unittest.py: (ChangeLogCheckerTest.test_no_new_tests): 2012-03-09 Sheriff Bot Unreviewed, rolling out r110250. http://trac.webkit.org/changeset/110250 https://bugs.webkit.org/show_bug.cgi?id=80682 It made NRWT ignore test_expectations.txt on Qt-WK2 and ignore qt-linux at all. (Requested by Ossy on #webkit). * Scripts/old-run-webkit-tests: * Scripts/webkitpy/layout_tests/port/qt.py: (QtPort): (QtPort.determine_full_port_name): (QtPort.baseline_search_path): * Scripts/webkitpy/layout_tests/port/qt_unittest.py: (QtPortTest): (QtPortTest._assert_search_path): (QtPortTest.test_baseline_search_path): 2012-03-08 Tor Arne Vestbø [Qt] Use Qt's module system for install rules and depending on QtWebKit Instead of rolling our own install rules we now use the same approach as every other Qt module, by loading qt_module.prf and qt_module_config.prf. This ensures that we follow the same semantics as the rest of Qt on what sort of config options are enabled by default (create_cmake eg.). It also allows us to use QT += webkit instead of the workaround we had with CONFIG += qtwebkit. We do however force Qt to always treat our build as a non-developer build, so the libraries will end up in the WebKit lib directory instead of the qtbase directory (as with a normal developer-build). This allows us to keep the webkit-build self-contained. If Qt is a developer build we still copy the module file manually to Qt, so that you don't have to install WebKit to make it available. For non-developer builds of Qt, it is still possible to use the built WebKit libraries without having to install them, by having the variable QMAKE_EXTRA_MODULE_FORWARDS set in the project's .qmake.cache file, pointing to $WEBKITOUTUTDIR/$CONFIGURATION/modules. https://bugs.webkit.org/show_bug.cgi?id=80590 Reviewed by Simon Hausmann. * DumpRenderTree/qt/DumpRenderTree.pro: * MiniBrowser/qt/BrowserWindow.cpp: * MiniBrowser/qt/MiniBrowser.pro: * MiniBrowser/qt/MiniBrowserApplication.cpp: * MiniBrowser/qt/UrlLoader.cpp: * QtTestBrowser/QtTestBrowser.pro: * Tools.pro: * WebKitTestRunner/InjectedBundle/Target.pri: * WebKitTestRunner/Target.pri: * qmake/mkspecs/features/default_post.prf: * qmake/mkspecs/features/default_pre.prf: * qmake/mkspecs/features/functions.prf: * qmake/mkspecs/features/qtwebkit-private.prf: Removed. * qmake/mkspecs/features/qtwebkit.prf: * qmake/mkspecs/features/rpath.prf: * qmake/mkspecs/modules/qt_webkit.pri: Removed. * qmake/qt_webkit.pri: Added. 2012-03-08 Igor Oliveira Unreviewed. Update my email. * Scripts/webkitpy/common/config/committers.py: 2012-03-08 Xianzhu Wang Recent changes to chromium.py etc. broke chromium_android.py https://bugs.webkit.org/show_bug.cgi?id=80623 Reviewed by Dirk Pranke. * Scripts/webkitpy/layout_tests/port/chromium_android.py: (ChromiumAndroidDriver.__init__): (ChromiumAndroidDriver._start): 2012-03-08 Xianzhu Wang [Chromium] Exception running reftest with --no-pixel-tests https://bugs.webkit.org/show_bug.cgi?id=80642 Reviewed by Dirk Pranke. * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumDriver.__init__): (ChromiumDriver._wrapper_options): 2012-03-08 Rafael Brandao [NRWT] Fix --platform=qt-5.0 --new-baseline combo https://bugs.webkit.org/show_bug.cgi?id=72489 Reviewed by Dirk Pranke. NRWT: The default platform name for Qt combined with --webkit-test-runner is now "qt-5.0-wk2" instead of "qt-linux" and that name is now prefered for tests. --platform will override the platform name and then we'll use it. ORWT: We check for qt version and use platform name "qt-4.8" if the version is lower than 5.0.0, and then we choose between "qt-5.0-wk1" and "qt-5.0-wk2". --platform will override it in the end. * Scripts/old-run-webkit-tests: * Scripts/webkitpy/layout_tests/port/qt.py: (QtPort): (QtPort.determine_full_port_name): (QtPort.baseline_search_path): * Scripts/webkitpy/layout_tests/port/qt_unittest.py: (QtPortTest): (QtPortTest._assert_search_path): (QtPortTest.test_baseline_search_path): 2012-03-08 Beth Dakin https://bugs.webkit.org/show_bug.cgi?id=80463 RenderImage is using the wrong origin when calling addRelevantRepaintedObject -and corresponding- Reviewed by Dan Bernstein. New test! * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayoutForImages.cpp: Added. (TestWebKitAPI): (TestWebKitAPI::didNewFirstVisuallyNonEmptyLayout): (TestWebKitAPI::setPageLoaderClient): (TestWebKitAPI::TEST): * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayoutForImages_Bundle.cpp: Added. (TestWebKitAPI): (NewFirstVisuallyNonEmptyLayoutForImagesTest): (TestWebKitAPI::NewFirstVisuallyNonEmptyLayoutForImagesTest::NewFirstVisuallyNonEmptyLayoutForImagesTest): (TestWebKitAPI::NewFirstVisuallyNonEmptyLayoutForImagesTest::didCreatePage): * TestWebKitAPI/Tests/WebKit2/lots-of-images.html: Added. 2012-03-08 Tor Arne Vestbø [Qt] Prospective Windows build fix Don't assume that windows-builds will always run inside a cmd.exe shell. * qmake/mkspecs/features/default_pre.prf: 2012-03-08 Dan Bernstein Dashboard regions should not be in device space Reviewed by John Sullivan. * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: * TestWebKitAPI/Tests/mac/DeviceScaleFactorInDashboardRegions.mm: Added. (-[DeviceScaleFactorInDashboardRegionsUIDelegate webView:dashboardRegionsChanged:]): (TestWebKitAPI): (TestWebKitAPI::TEST): 2012-03-08 Dirk Pranke webkit-patch works oddly on local Git branches https://bugs.webkit.org/show_bug.cgi?id=76958 Reviewed by Ojan Vafai. This change makes the '-g' flag to webkit-patch behave a little more consistently with git. The 'HEAD..' syntax is changed to 'HEAD....' (four dots), so as to not collide with the underlying git syntax. We also generalize the syntax to support any refname, so 'foo....' means diff against foo and include working copy changes. This change also adds the 'UPSTREAM' shortcut to identify whatever the current upstream/tracking branch is; this is useful if you have multiple local branches stacked up. * Scripts/webkitpy/common/checkout/scm/git.py: (Git._branch_from_ref): (Git): (Git._current_branch): (Git._upstream_branch): (Git.merge_base): (Git._commit_on_branch): * Scripts/webkitpy/common/checkout/scm/scm_unittest.py: (GitSVNTest._second_local_commit): (GitSVNTest._two_local_commits): (GitSVNTest.test_commit_with_message_git_commit_range): (GitSVNTest.test_upstream_branch): (GitSVNTest.test_create_patch_working_copy_only): (GitSVNTest.test_changed_files_local_plus_working_copy): (GitSVNTest.test_changed_files_working_copy_only): (GitSVNTest.test_changed_files_upstream): * Scripts/webkitpy/tool/steps/options.py: (Options): 2012-03-08 John Sullivan Made extract-localizable-strings ignore NSLocalizedFileSizeDescription Reviewed by Eric Carlson. * Scripts/extract-localizable-strings: Added NSLocalizedFileSizeDescription to allowed list of tokens that start with NSLocalized. Also improved an error message to include the troublesome token. 2012-03-08 Sheriff Bot Unreviewed, rolling out r110115. http://trac.webkit.org/changeset/110115 https://bugs.webkit.org/show_bug.cgi?id=80581 broke test-webkitpy (Requested by philn-tp on #webkit). * Scripts/old-run-webkit-tests: * Scripts/webkitpy/layout_tests/port/qt.py: (QtPort): (QtPort.determine_full_port_name): (QtPort.baseline_search_path): 2012-03-08 Kentaro Hara Unreviewed. Added myself to watchlist. * Scripts/webkitpy/common/config/watchlist: 2012-03-07 Jessie Berlin Clean Windows build fails after r110033 https://bugs.webkit.org/show_bug.cgi?id=80553 Rubber-stamped by Jon Honeycutt and Eric Seidel. Add the new JavaScriptCore include directory to some more vsprops files. * DumpRenderTree/win/ImageDiffCommon.vsprops: * WebKitTestRunner/Configurations/InjectedBundleCommon.vsprops: * WebKitTestRunner/Configurations/WebKitTestRunnerCommon.vsprops: 2012-03-07 Rafael Brandao [NRWT] Fix --platform=qt-5.0 --new-baseline combo https://bugs.webkit.org/show_bug.cgi?id=72489 Reviewed by Dirk Pranke. NRWT: The default platform name for Qt combined with --webkit-test-runner is now "qt-5.0-wk2" instead of "qt-linux" and that name is now prefered for tests. --platform will override the platform name and then we'll use it. ORWT: We check for qt version and use platform name "qt-4.8" if the version is lower than 5.0.0, and then we choose between "qt-5.0-wk1" and "qt-5.0-wk2". --platform will override it in the end. * Scripts/old-run-webkit-tests: * Scripts/webkitpy/layout_tests/port/qt.py: (QtPort): (QtPort.determine_full_port_name): (QtPort.baseline_search_path): 2012-03-07 Hugo Parente Lima [Qt][WK2] MiniBrowser on desktop does not react to click events https://bugs.webkit.org/show_bug.cgi?id=80314 Reviewed by Kenneth Rohde Christiansen. Do not enable touch mocking on desktop mode. * MiniBrowser/qt/MiniBrowserApplication.cpp: (MiniBrowserApplication::handleUserOptions): 2012-03-07 Ojan Vafai Automatically move to the next test after rebaseline in the results view https://bugs.webkit.org/show_bug.cgi?id=80527 Reviewed by Dimitri Glazkov. * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/controllers.js: * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results.js: 2012-03-07 Ojan Vafai garden-o-matic rebaseline is broken for ports that don't have a fallback port https://bugs.webkit.org/show_bug.cgi?id=80526 Reviewed by Adam Barth. Passing "None" as the fallback port throws an error in "webkit-patch rebaseline-test" since there is no "None" port. * Scripts/webkitpy/tool/servers/gardeningserver.py: (GardeningHTTPRequestHandler.rebaseline): * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py: (GardeningServerTest.test_rebaseline): 2012-03-07 James Robinson [chromium] Remove the TestWebPlugin's use of implicit FBO for offscreen contexts https://bugs.webkit.org/show_bug.cgi?id=80521 Reviewed by Kenneth Russell. Switches TestWebPlugin from using a context created by WebViewClient::createGraphicsContext3D(..., false) over to an explicitly offscreen context with explicitly managed texture / FBO, so we can deprecate the renderDirectlyToHostWindow flag from WebViewClient::createGraphicsContext3D. Covered by platform/chromium/compositing/plugins/* * DumpRenderTree/chromium/TestWebPlugin.cpp: (TestWebPlugin::TestWebPlugin): (TestWebPlugin::initialize): (TestWebPlugin::updateGeometry): (TestWebPlugin::initScene): (TestWebPlugin::destroyScene): * DumpRenderTree/chromium/TestWebPlugin.h: (TestWebPlugin): * DumpRenderTree/chromium/WebViewHost.cpp: (WebViewHost::createPlugin): 2012-03-07 Philippe Normand [GTK] remove webkitpy dependency in run-gtk-tests https://bugs.webkit.org/show_bug.cgi?id=80500 Reviewed by Martin Robinson. Replace Executive calls with bare subprocess calls. * Scripts/run-gtk-tests: (TestRunner): (TestRunner._lookup_atspi2_binary): (TestRunner.run): (TestRunner.run.run_for_real): 2012-03-07 Philippe Normand Unreviewed, GTK build fix after r110059. * Scripts/run-gtk-tests: (TestRunner._check_if_tests_have_timed_out): (TestRunner.run.run_tests): (TestRunner): 2012-03-07 Philippe Normand [GTK] race condition in run-gtk-tests https://bugs.webkit.org/show_bug.cgi?id=80495 Reviewed by Martin Robinson. Refactored the script to wait the a11y dbus service becomes available before starting the tests. * Scripts/run-gtk-tests: (TestRunner._lookup_atspi2_binary): (TestRunner): (TestRunner._wait_dbus_service_and_run): (TestRunner._wait_dbus_service_and_run.on_name_appeared): (TestRunner._wait_dbus_service_and_run.on_name_vanished): (TestRunner.run): (TestRunner.run.bailout): (TestRunner.run.run_for_real): 2012-03-07 Dinu Jacob [Qt] REGRESSION(r109575) Broke tap-to-zoom gesture https://bugs.webkit.org/show_bug.cgi?id=80425 Reviewed by Simon Hausmann. Set startScreenPos in mock touch point * MiniBrowser/qt/MiniBrowserApplication.cpp: (MiniBrowserApplication::notify): * MiniBrowser/qt/MiniBrowserApplication.h: (MiniBrowserApplication): 2012-03-07 Simon Hausmann [Qt] Windows build fix. Reviewed by Tor Arne Vestbø. Pass the -j parameter only to "make", not "nmake". * Scripts/webkitdirs.pm: (buildQMakeProjects): 2012-03-07 Simon Hausmann [Qt] Fix compilation without QtQuick1 https://bugs.webkit.org/show_bug.cgi?id=80503 Reviewed by Tor Arne Vestbø. Introduce HAVE_QQUICK1 as feature and added haveQtModule as convenience function. * qmake/mkspecs/features/features.prf: * qmake/mkspecs/features/functions.prf: 2012-03-07 Zoltan Horvath [Qt] Add qmake config tests for JPEG and PNG library https://bugs.webkit.org/show_bug.cgi?id=80398 Add JPEG and PNG library tests to config.tests and add error() abort to WebCore.pri if tests don't work. Reviewed by Tor Arne Vestbø. * qmake/config.tests/libjpeg/libjpeg.cpp: Added. (main): * qmake/config.tests/libjpeg/libjpeg.pro: Added. * qmake/config.tests/libpng/libpng.cpp: Added. (main): * qmake/config.tests/libpng/libpng.pro: Added. * qmake/sync.profile: 2012-03-07 Byungwoo Lee [EFL] Build warning: Fix warn_unused_result warnings. https://bugs.webkit.org/show_bug.cgi?id=79194 Reviewed by Filip Pizlo. Fixed build warning which is getting generated when not using return value of a fuction, especially declared with warn_unused_result attribute. * DumpRenderTree/efl/ImageDiff.cpp: (printImage): 2012-03-07 Kangil Han [DRT] Remove PlainTextController implementations. https://bugs.webkit.org/show_bug.cgi?id=79959 Reviewed by Hajime Morita. PlainTextController usages in existing tests have been replaced by internals API by bug 78570. So this patch will remove PlainTextController implementations to avoid further usage in new tests. * DumpRenderTree/DumpRenderTree.gypi: * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: * DumpRenderTree/chromium/PlainTextController.cpp: Removed. * DumpRenderTree/chromium/PlainTextController.h: Removed. * DumpRenderTree/chromium/TestShell.cpp: (TestShell::TestShell): (TestShell::bindJSObjectsToWindow): * DumpRenderTree/chromium/TestShell.h: (TestShell): * DumpRenderTree/gtk/DumpRenderTree.cpp: (webViewWindowObjectCleared): * DumpRenderTree/gtk/PlainTextController.cpp: Removed. * DumpRenderTree/gtk/PlainTextController.h: Removed. * DumpRenderTree/mac/FrameLoadDelegate.mm: (-[FrameLoadDelegate didClearWindowObjectInStandardWorldForFrame:]): * DumpRenderTree/mac/PlainTextController.h: Removed. * DumpRenderTree/mac/PlainTextController.mm: Removed. * DumpRenderTree/qt/DumpRenderTree.pro: * DumpRenderTree/qt/DumpRenderTreeQt.cpp: (WebCore::DumpRenderTree::DumpRenderTree): (WebCore::DumpRenderTree::initJSObjects): * DumpRenderTree/qt/DumpRenderTreeQt.h: (DumpRenderTree): * DumpRenderTree/qt/PlainTextControllerQt.cpp: Removed. * DumpRenderTree/qt/PlainTextControllerQt.h: Removed. * GNUmakefile.am: 2012-03-06 Grzegorz Czajkowski [EFL][DRT] Implements LayoutTestController's setPopupBlockingEnabled. https://bugs.webkit.org/show_bug.cgi?id=80137 Reviewed by Eric Seidel. Adds implementation LayoutTestController's setPopupBlockingEnabled. To pass popup-blocking-click-in-iframe.html EventSender should not repeat click events in idler's function. EventSender allows to asynchronously send click event through ecore_idler_add. This event will be called repeatedly as long as it return true. Click event should be called only once by returning ECORE_CALLBACK_CANCEL. * DumpRenderTree/efl/EventSender.cpp: (sendMouseEvent): Changed return value from bool to void because it always returns true. (sendClick): Idler is deleted by returning ECORE_CALLBACK_CANCEL. (mouseDownCallback): Remove unnecessary checking of return value. * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::setPopupBlockingEnabled): 2012-03-06 Dirk Pranke rebaseline_chromium_webkit_tests is obsolete and should be removed https://bugs.webkit.org/show_bug.cgi?id=80347 Reviewed by Tony Chang. * Scripts/rebaseline-chromium-webkit-tests: Removed. * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests.py: Removed. * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests_unittest.py: Removed. 2012-03-06 Dirk Pranke test_start_cmd (webkitpy.layout_tests.servers.http_server_unittest.TestHttpServer) is broken https://bugs.webkit.org/show_bug.cgi?id=80468 Unreviewed, build fix. Fixing a bug revealed by r80443 when we fixed filesystem.copyfile(). * Scripts/webkitpy/layout_tests/servers/http_server.py: (Lighttpd._prepare_config): 2012-03-06 Dirk Pranke remove chromium_gpu ports https://bugs.webkit.org/show_bug.cgi?id=80345 Reviewed by Ojan Vafai. This change leaves in the 'graphics_type' distinction in the test_expectations; removing that will require more updates and a pass over test_expectations.txt, so I'm separating that out. * Scripts/webkitpy/layout_tests/port/builders.py: * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumPort): * Scripts/webkitpy/layout_tests/port/chromium_gpu.py: Removed. * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py: Removed. * Scripts/webkitpy/layout_tests/port/factory.py: (PortFactory): (PortFactory.get): * Scripts/webkitpy/layout_tests/port/factory_unittest.py: (FactoryTest.test_qt): * Scripts/webkitpy/tool/commands/rebaseline_unittest.py: (TestRebaseline.test_rebaseline_expectations): 2012-03-06 Raphael Kubo da Costa undefined reference to JSC::IdentifierTable::~IdentifierTable() on EFL port https://bugs.webkit.org/show_bug.cgi?id=80282 Reviewed by Antonio Gomes. Remove transitive library dependencies; they are especially dangerous when one ends up linking against WTF, JSC and then WTF again, since some symbols will not be defined. Passing --no-copy-dt-needed-entries and --as-needed to the linker (which some recent Linux distros do by default) makes the issue even more evident. * DumpRenderTree/efl/CMakeLists.txt: Do not link directly to WTF, as the other libraries have the needed symbols. Explicitly link against fontconfig. 2012-03-06 Dean Jackson filter-build-webkit needs more filters https://bugs.webkit.org/show_bug.cgi?id=80440 Reviewed by Joseph Pecoraro. Added as many missing filters as I could find for Apple builds. Plus a few minor style updates. * Scripts/filter-build-webkit: (printLine): 2012-03-06 Ojan Vafai Copying old baselines throws an error if the new directory doesn't exist https://bugs.webkit.org/show_bug.cgi?id=80443 Reviewed by Adam Barth. Existing tests cover the new code with the fix to filesystem_mock. * Scripts/webkitpy/common/system/filesystem_mock.py: (MockFileSystem.copyfile): Fix the mock to better match python's copyfile semantics. * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests.py: (Rebaseliner._create_html_baseline_files): * Scripts/webkitpy/tool/commands/rebaseline.py: (RebaselineTest._copy_existing_baseline): 2012-03-05 Ojan Vafai Add a mechanism to rebaseline new ports https://bugs.webkit.org/show_bug.cgi?id=80355 Reviewed by Adam Barth. For a test that's only failing on a new port, we want to first copy the existing result into the location for the port it's replacing in order to not break that port. For example, bringing up the chromium-lion port, if we just stick the results in the chromium-mac directory, the snow leopard will start failing. Instead, we first copy the existing result to the chromium-mac-snowleopard directory. * Scripts/webkitpy/layout_tests/port/builders.py: (builder_path_for_port_name): (fallback_port_name_for_new_port): * Scripts/webkitpy/tool/commands/rebaseline.py: (RebaselineTest): (RebaselineTest._copy_existing_baseline): (RebaselineTest._rebaseline_test): (RebaselineTest.execute): * Scripts/webkitpy/tool/commands/rebaseline_unittest.py: (TestRebaseline.test_rebaseline_test): (TestRebaseline): (TestRebaseline.test_rebaseline_and_copy_test): (test_rebaseline_and_copy_test_no_existing_result): (test_rebaseline_and_copy_test_with_lion_result): (test_rebaseline_and_copy_no_overwrite_test): * Scripts/webkitpy/tool/servers/gardeningserver.py: (GardeningHTTPRequestHandler.rebaseline): * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py: (GardeningServerTest.test_rebaseline): (GardeningServerTest): (GardeningServerTest.test_rebaseline_new_port): 2012-03-06 Beth Dakin https://bugs.webkit.org/show_bug.cgi?id=80351 InlineTextBox is using the wrong origin when calling addRelevantRepaintedObject -and corresponding- Reviewed by Sam Weinig. This test has actually been broken since revision 109273. This change makes the test pass again, and it makes it specifically exercise the fix for this bug. * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayout.cpp: (TestWebKitAPI::TEST): Adjust comments in the other tests to reflect the API changes from 109273. * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayoutFails.cpp: (TestWebKitAPI::TEST): * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayoutFails_Bundle.cpp: (TestWebKitAPI::NewFirstVisuallyNonEmptyLayoutFailsTest::didCreatePage): * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayoutFrames_Bundle.cpp: (TestWebKitAPI::NewFirstVisuallyNonEmptyLayoutFramesTest::didCreatePage): * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayout_Bundle.cpp: (TestWebKitAPI::NewFirstVisuallyNonEmptyLayoutTest::didCreatePage): * TestWebKitAPI/Tests/WebKit2/lots-of-text.html: Added. 2012-03-06 Hans Wennborg Speech JavaScript API: add SpeechRecognitionError https://bugs.webkit.org/show_bug.cgi?id=80410 Chromium: enable the runtime flag for scripted speech so we can run the layout tests for it. Reviewed by Adam Barth. * DumpRenderTree/chromium/TestShell.cpp: (TestShell::TestShell): 2012-03-06 Sami Kyostila Enable universal newlines in chromium android layout test runner https://bugs.webkit.org/show_bug.cgi?id=80418 Reviewed by Adam Barth. The Chromium Android layout test runner uses the 'adb' program to communicate with the Android device. Adb uses CRLF ('\r\n') to delimit output lines from the test runner, while the test expectation files are written with LF ('\n') delimiters. This causes the results to have spurious whitespace differences, making them more difficult to interpret. The fix is to enable universal newlines for the DumpRenderTree pipe, causing the output to only contain LF newlines like on other platforms. * Scripts/webkitpy/layout_tests/port/chromium_android.py: (ChromiumAndroidDriver._start): 2012-03-06 Tor Arne Vestbø [Qt] Revert r109900, it was completly busted HAVE_FOO does _not_ have a WTF_ prefix, go figure. We still leave in the HAVE(FOO) in QtInitializeTestFonts.cpp, instead of manual define checks. Reviewed by Ossy. * DumpRenderTree/qt/DumpRenderTree.pro: * QtTestBrowser/QtTestBrowser.pro: * WebKitTestRunner/InjectedBundle/Target.pri: * qmake/mkspecs/features/features.prf: 2012-03-05 Tor Arne Vestbø [Qt] Use config test to determine if the compiler supports -MD We used to assume that unix systems building against Qt5 would have a compiler that supported -MD. Although this was usually the case, we failed to take into account broken tools for distributed builds (e.g. teambuilder or icecream) which would end up producing .d files in the root of the build dir instead of along-side the object file. Now that we use a config-test to determine the support, we'll capture those cases, unless distributed builds are enabled after the initial build-webkit run. Reviewed by Csaba Osztrogonác. * Tools.pro: * qmake/config.tests/gccdepends/empty.cpp: Added. * qmake/config.tests/gccdepends/gccdepends.pro: Added. * qmake/mkspecs/features/default_pre.prf: * qmake/mkspecs/features/features.prf: * qmake/mkspecs/features/unix/default_pre.prf: * qmake/sync.profile: 2012-03-06 Csaba Osztrogonác REGRESSION(r109294): It broke the unittest of master.cfg https://bugs.webkit.org/show_bug.cgi?id=80253 Reviewed by Tor Arne Vestbø. * BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py: Use json instead of simplejson. (BuildBotConfigLoader._create_mock_passwords_dict): (BuildBotConfigLoader._mock_open): (BuildBotConfigLoader._add_dependant_modules_to_sys_modules): (BuildBotConfigLoader.load_config): 2012-03-06 Kristóf Kosztyó Qt API test buildstep throws exception after buildmaster update https://bugs.webkit.org/show_bug.cgi?id=80249 This patch based on Peter Gal's work. Reviewed by Csaba Osztrogonác. * BuildSlaveSupport/build.webkit.org-config/master.cfg: (RunQtAPITests): (BuildAndTestFactory.__init__): Enable API tests on the Qt platform again. 2012-03-06 Tor Arne Vestbø [Qt] Use correct define names for HAVE(FOO) style optional dependencies Reviewed by Simon Hausmann. Reviewed by Ossy. * DumpRenderTree/qt/DumpRenderTree.pro: * DumpRenderTree/qt/QtInitializeTestFonts.cpp: (WebKit::initializeTestFonts): * QtTestBrowser/QtTestBrowser.pro: * WebKitTestRunner/InjectedBundle/Target.pri: * qmake/mkspecs/features/features.prf: 2012-03-06 Tor Arne Vestbø [Qt] Don't add recursive qmake_all target if there are no subdirs to recurse Reviewed by Ossy. * qmake/mkspecs/features/default_post.prf: 2012-03-06 Tor Arne Vestbø [Qt] Make sure clean builds are followed by make qmake && make if needed On non-GNUmake systems we have to run make qmake so that the generated sources are generated before we run qmake on the target pro files. This is achieved by using the 'incremental' target. Reviewed by Ossy. * Scripts/webkitdirs.pm: (buildQMakeProjects): 2012-03-06 Filip Pizlo DumpRenderTree should support preciseTime() https://bugs.webkit.org/show_bug.cgi?id=80397 Reviewed by Gavin Barraclough. * DumpRenderTree/LayoutTestController.cpp: (preciseTimeCallback): (LayoutTestController::staticFunctions): 2012-03-06 Sheriff Bot Unreviewed, rolling out r109887. http://trac.webkit.org/changeset/109887 https://bugs.webkit.org/show_bug.cgi?id=80402 It broke the build (Requested by Ossy on #webkit). * DumpRenderTree/qt/DumpRenderTree.pro: * DumpRenderTree/qt/QtInitializeTestFonts.cpp: (WebKit::initializeTestFonts): * QtTestBrowser/QtTestBrowser.pro: * WebKitTestRunner/InjectedBundle/Target.pri: * qmake/mkspecs/features/features.prf: 2012-03-06 Sheriff Bot Unreviewed, rolling out r109837. http://trac.webkit.org/changeset/109837 https://bugs.webkit.org/show_bug.cgi?id=80399 breaks Mac Productions builds, too late to try and fix it tonight (Requested by eseidel on #webkit). * Scripts/build-webkit: 2012-03-06 Tor Arne Vestbø [Qt] Use correct define names for HAVE(FOO) style optional dependencies Reviewed by Simon Hausmann. * DumpRenderTree/qt/DumpRenderTree.pro: * DumpRenderTree/qt/QtInitializeTestFonts.cpp: (WebKit::initializeTestFonts): * QtTestBrowser/QtTestBrowser.pro: * WebKitTestRunner/InjectedBundle/Target.pri: * qmake/mkspecs/features/features.prf: 2012-03-06 Kristóf Kosztyó [Qt] Create EWS to build WebKit2 https://bugs.webkit.org/show_bug.cgi?id=80277 Reviewed by Csaba Osztrogonác. * QueueStatusServer/model/queues.py: (Queue): * Scripts/webkitpy/tool/commands/earlywarningsystem.py: (QtWK2EWS): * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py: (EarlyWarningSytemTest.test_builder_ewses): 2012-03-05 Sheriff Bot Unreviewed, rolling out r109733. http://trac.webkit.org/changeset/109733 https://bugs.webkit.org/show_bug.cgi?id=80378 It broke Qt5 build system (Requested by Ossy on #webkit). * Tools.pro: * qmake/config.tests/gccdepends/empty.cpp: Removed. * qmake/config.tests/gccdepends/gccdepends.pro: Removed. * qmake/mkspecs/features/default_pre.prf: * qmake/mkspecs/features/features.prf: * qmake/mkspecs/features/unix/default_pre.prf: * qmake/sync.profile: 2012-03-05 Eric Seidel Update JavaScriptCore files to use fully-qualified WTF include path https://bugs.webkit.org/show_bug.cgi?id=79960 Reviewed by Adam Barth. Build WTF/WTF.xcodeproj by default on Mac. * Scripts/build-webkit: 2012-03-05 Dirk Pranke There should be a way to disable optimizer in webkit-patch rebaseline-expectations https://bugs.webkit.org/show_bug.cgi?id=69590 Reviewed by Ryosuke Niwa. Adds a --no-optimize flag to webkit-patch rebaseline-expectations to skip the 'optimize-expectations' step. * Scripts/webkitpy/tool/commands/rebaseline.py: (RebaselineExpectations.__init__): (RebaselineExpectations.execute): * Scripts/webkitpy/tool/commands/rebaseline_unittest.py: 2012-03-05 Kent Tamura [Chromium] Add new popup type: PagePopup https://bugs.webkit.org/show_bug.cgi?id=80106 Reviewed by Darin Fisher. * DumpRenderTree/chromium/WebViewHost.cpp: (WebViewHost::createPopupMenu): 2012-03-05 Jessie Berlin Add myself as a reviewer. * Scripts/webkitpy/common/config/committers.py: Also, add my IRC nick so it is easier to find and yell at me. 2012-03-05 Robert Kroeger Correctly compute the positions of GestureScroll* events on Chromium. https://bugs.webkit.org/show_bug.cgi?id=80075 Reviewed by Adam Barth. * DumpRenderTree/chromium/EventSender.cpp: (EventSender::reset): (EventSender::gestureEvent): * DumpRenderTree/chromium/EventSender.h: (EventSender): 2012-03-05 Dirk Pranke flakiness dashboard doesn't display baselines for virtual tests correctly https://bugs.webkit.org/show_bug.cgi?id=80341 Reviewed by Ojan Vafai. If the test is part of a virtual test suite, we need to look for baselines in the underlying fallback path for the actual test being run, as well as baselines for the virtual version. This patch adds code to handle both branches and also pulls the mapping of the test suites up into a constant. * TestResultServer/static-dashboards/flakiness_dashboard.html: * TestResultServer/static-dashboards/flakiness_dashboard_tests.js: (testLookupVirtualTestSuite): (testBaseTest): 2012-03-05 Sheriff Bot Unreviewed, rolling out r109760. http://trac.webkit.org/changeset/109760 https://bugs.webkit.org/show_bug.cgi?id=80320 Caused many GTK+ tests to crash (Requested by mrobinson on #webkit). * gtk/jhbuild.modules: 2012-03-05 Charlie Reis [chromium] Pass WebNavigationPolicy to WebViewClient::createView https://bugs.webkit.org/show_bug.cgi?id=80057 Reviewed by Darin Fisher. * DumpRenderTree/chromium/WebViewHost.cpp: (WebViewHost::createView): * DumpRenderTree/chromium/WebViewHost.h: (WebViewHost): 2012-03-05 Martin Robinson [soup] Crash while loading http://www.jusco.cn https://bugs.webkit.org/show_bug.cgi?id=68238 Reviewed by Philippe Normand. * gtk/jhbuild.modules: Bumped the libsoup and glib dependencies in the jhbuild file. 2012-03-05 Sheriff Bot Unreviewed, rolling out r109748. http://trac.webkit.org/changeset/109748 https://bugs.webkit.org/show_bug.cgi?id=80296 Made some tests crash, will fix and recommit (Requested by noamr on #webkit). * MiniBrowser/qt/main.cpp: (main): 2012-03-05 Zan Dobersek [GTK] plugins/netscape-plugin-page-cache-works.html fails https://bugs.webkit.org/show_bug.cgi?id=74409 Reviewed by Martin Robinson. Add support for overriding the page-cache-for-plugins preference through DumpRenderTreeSupport. * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: (booleanFromValue): A helper function to convert a string to boolean. (LayoutTestController::overridePreference): 2012-03-05 Rob Buis [BlackBerry] Update Tools/DumpRenderTree https://bugs.webkit.org/show_bug.cgi?id=80287 Reviewed by Antonio Gomes. Update so we are sure we can build our implementation of DRT. * DumpRenderTree/DumpRenderTree.h: include DumpRenderTreeBlackBerry.h. * DumpRenderTree/PixelDumpSupport.cpp: include PixelDumpSupportBlackBerry.h. * DumpRenderTree/blackberry/AccessibilityControllerBlackBerry.cpp: (AccessibilityController::addNotificationListener): add missing hook. (AccessibilityController::removeNotificationListener): add missing hook. * DumpRenderTree/blackberry/AccessibilityUIElementBlackBerry.cpp: (AccessibilityUIElement::isSelectedOptionActive): add missing hook. * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp: (LayoutTestController::setMockSpeechInputDumpRect): add missing hook. (LayoutTestController::simulateDesktopNotificationClick): add missing hook. 2012-03-05 No'am Rosenthal [Qt] [WK2] Support threaded renderer in WK2 https://bugs.webkit.org/show_bug.cgi?id=76661 Remove the QML_NO_THREADED_RENDERER environment variable from MiniBrowser. Reviewed by Kenneth Rohde Christiansen. * MiniBrowser/qt/main.cpp: (main): 2012-03-05 Alexander Færøy Unreviewed. Add myself to committer list. * Scripts/webkitpy/common/config/committers.py: 2012-03-05 Tor Arne Vestbø [Qt] Use config test to determine if the compiler supports -MD We used to assume that unix systems building against Qt5 would have a compiler that supported -MD. Although this was usually the case, we failed to take into account broken tools for distributed builds (e.g. teambuilder or icecream) which would end up producing .d files in the root of the build dir instead of along-side the object file. Now that we use a config-test to determine the support, we'll capture those cases, unless distributed builds are enabled after the initial build-webkit run. Reviewed by Csaba Osztrogonác. * Tools.pro: * qmake/config.tests/gccdepends/empty.cpp: Added. * qmake/config.tests/gccdepends/gccdepends.pro: Added. * qmake/mkspecs/features/default_pre.prf: * qmake/mkspecs/features/features.prf: * qmake/mkspecs/features/unix/default_pre.prf: * qmake/sync.profile: 2012-03-05 Csaba Osztrogonác Qt API test buildstep throws exception after buildmaster update https://bugs.webkit.org/show_bug.cgi?id=80249 Reviewed by Zoltan Herczeg. * BuildSlaveSupport/build.webkit.org-config/master.cfg: Disable API tests on the Qt bot temporarily. (BuildAndTestFactory.__init__): 2012-03-04 Martin Robinson [GTK] Unit tests are noisy when run under jhbuild https://bugs.webkit.org/show_bug.cgi?id=80241 Reviewed by Philippe Normand. Force the 'memory' backend in GSettings so that each unit test doesn't produce a warning when it's forced to fall back to the 'memory' backend. * Scripts/run-gtk-tests: (TestRunner.run): Force the 'memory' backend. 2012-03-04 Martin Robinson [GTK] "Console: undefined @0: ReferenceError: Can't find variable: WebInspector" during testwebinspector https://bugs.webkit.org/show_bug.cgi?id=80235 Reviewed by Philippe Normand. * Scripts/run-gtk-tests: (TestRunner.run): Use the WEBKIT_INSPECTOR_PATH environment variable so that inspector resources path determination is shared by all tests. 2012-03-04 Martin Robinson [GTK] REGRESSION: r108278 broke some tests that expect focus https://bugs.webkit.org/show_bug.cgi?id=79084 Reviewed by Philippe Normand. Grab focus when dumping test results and also when sending events. During the course of testing, the WebView may lose focus and this seems to snap it back when tests demand it. * DumpRenderTree/gtk/DumpRenderTree.cpp: (dump): Grab focus. * DumpRenderTree/gtk/EventSender.cpp: (dispatchEvent): Ditto. 2012-03-04 Charles Wei Unreviewed, Add myself to committer list. https://bugs.webkit.org/show_bug.cgi?id=80232 * Scripts/webkitpy/common/config/committers.py: 2012-03-04 Anders Carlsson Newer versions of clang fail to compile WebKit. Fix build. * DumpRenderTree/mac/AccessibilityUIElementMac.mm: (descriptionOfValue): (AccessibilityUIElement::rowIndexRange): (AccessibilityUIElement::columnIndexRange): (AccessibilityUIElement::selectedTextRange): * DumpRenderTree/mac/EditingDelegate.mm: (-[DOMRange dump]): * DumpRenderTree/mac/PolicyDelegate.mm: (-[PolicyDelegate webView:unableToImplementPolicyWithError:frame:]): * DumpRenderTree/mac/ResourceLoadDelegate.mm: (-[NSError _drt_descriptionSuitableForTestResult]): * WebKitTestRunner/InjectedBundle/mac/AccessibilityUIElementMac.mm: (WTR::descriptionOfValue): (WTR::AccessibilityUIElement::rowIndexRange): (WTR::AccessibilityUIElement::columnIndexRange): (WTR::AccessibilityUIElement::selectedTextRange): 2012-03-03 Simon Hausmann [Qt] Fix static_libs_as_shared build https://bugs.webkit.org/show_bug.cgi?id=80214 Reviewed by Tor Arne Vestbø. * WebKitTestRunner/qt/main.cpp: (main): Set the correct environment variable for choosing the theme in the web process. 2012-03-03 Hans Wennborg Implement Speech JavaScript API https://bugs.webkit.org/show_bug.cgi?id=80019 Reviewed by Adam Barth. Add ENABLE_SCRIPTED_SPEECH. * Scripts/build-webkit: * qmake/mkspecs/features/features.prf: 2012-03-02 Alexey Proskuryakov Fix commit queue, remove duplicate entry. * Scripts/webkitpy/common/config/committers.py: 2012-03-02 Stephen Chenney Unreviewed, add myself as a committer * Scripts/webkitpy/common/config/committers.py: 2012-03-02 Ojan Vafai Add the Chromium Lion canary bot to the flakiness dashboard. * TestResultServer/static-dashboards/builders.js: 2012-03-02 Dirk Pranke links for virtual tests are broken in the flakiness dashboard https://bugs.webkit.org/show_bug.cgi?id=80180 Reviewed by Ojan Vafai. * TestResultServer/static-dashboards/flakiness_dashboard.html: (linkHTMLToOpenWindow): * TestResultServer/static-dashboards/flakiness_dashboard_tests.js: (testLinkHTMLToOpenWindow): added. 2012-03-02 Philippe Normand [GTK] security/block-test.html baselines differ from mac https://bugs.webkit.org/show_bug.cgi?id=78343 Reviewed by Martin Robinson. Improved the FrameLoader error heuristics to match mac behavior a bit more. * DumpRenderTree/gtk/DumpRenderTree.cpp: (pathFromSoupURI): Handle ftp urls the same as http urls. (descriptionSuitableForTestResult): Map WebKitPolicyError to WebKitErrorDomain. 2012-03-02 Andras Becsi [Qt][WK2] Fix typo in patch committed in r109575 Unreviewed typo fix. * MiniBrowser/qt/MiniBrowserApplication.cpp: (MiniBrowserApplication::notify): Cache the position in floating point coordinates. 2012-03-02 Andras Becsi [Qt][WK2] Infinite loop on history navigation, when panning https://bugs.webkit.org/show_bug.cgi?id=79119 Reviewed by Simon Hausmann. Make the touch mocking more robust by creating touch events directly and sending them synchronously to the browser window. If the sent touch event was not accepted we propagate the original mouse event to its target. This way we make sure we only send either a touch or a mouse event and do not end up in an infinite loop between our mocking code and the Qt5 automatic mouse event synthesizing. * MiniBrowser/qt/BrowserWindow.cpp: (BrowserWindow::updateVisualMockTouchPoints): * MiniBrowser/qt/BrowserWindow.h: (BrowserWindow): * MiniBrowser/qt/MiniBrowserApplication.cpp: (touchRectForPosition): (MiniBrowserApplication::notify): (MiniBrowserApplication::sendTouchEvent): * MiniBrowser/qt/MiniBrowserApplication.h: (MiniBrowserApplication): 2012-03-02 Tor Arne Vestbø [Qt] Use 'all' as default target when debug_and_release is in effect When debug_and_release is enabled, in combination with build_all, qmake will generate an aggregate makefile that calls make on both the debug and the release makefiles, but with the 'all' target. In this case we can't use 'first' as the default target, as the extra targets would then not be run, so we append extra targets to 'all' instead. Reviewed by Simon Hausmann. * qmake/mkspecs/features/default_post.prf: 2012-03-02 Tor Arne Vestbø [Qt] Make sure activeBuildConfig() returns the right config Using isActiveConfig fails, as that only checks if the given argument is in CONFIG, not the order. Reviewed by Simon Hausmann. * qmake/mkspecs/features/functions.prf: 2012-03-02 Simon Hausmann [Qt] Fix tests run with WTR not using QStyle theme https://bugs.webkit.org/show_bug.cgi?id=80147 Reviewed by Csaba Osztrogonác. * WebKitTestRunner/qt/main.cpp: (main): Set environment variable for web process to use QStyle theme. 2012-03-02 Philippe Normand Unreviewed, remove debug print call. * DumpRenderTree/gtk/DumpRenderTree.cpp: (descriptionSuitableForTestResult): 2012-03-02 Robert Kroeger Updated EventSender to send complete gesture events. https://bugs.webkit.org/show_bug.cgi?id=80075 Reviewed by Adam Barth. * DumpRenderTree/chromium/EventSender.cpp: (EventSender::gestureEvent): set additional parameters in sent events. 2012-03-01 Kalev Lember WebKitTestRunner: Explicitly include unistd.h for gcc 4.7 https://bugs.webkit.org/show_bug.cgi?id=80033 Reviewed by Gustavo Noronha Silva. GCC 4.7 no longer includes unistd.h from most of the standard C++ header files, so we need to explicitly include it for getcwd(). * WebKitTestRunner/TestInvocation.cpp: Make sure unistd.h gets included not only on Mac, but on all Unix platforms. 2012-03-01 Dirk Pranke Actually look at the args in driver_input :(. Unreviewed, build fix. * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumDriver.run_test): 2012-03-01 Dirk Pranke Fix typo in flag used to run gpu virtual suites. Unreviewed, build fix. * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumPort.virtual_test_suites): 2012-03-01 Adam Barth Unreviewed, rolling out r109317. http://trac.webkit.org/changeset/109317 Let's try watchlists again now that we're using a pure git working copy * Scripts/webkitpy/tool/bot/stylequeuetask.py: (StyleQueueTask.run): * Scripts/webkitpy/tool/commands/queues_unittest.py: 2012-03-01 Dirk Pranke webkitpy: test_run_test_pause_before_testing is failing after r109482 https://bugs.webkit.org/show_bug.cgi?id=80084 Unreviewed, build fix. Fixes the test string compare to ignore the time spent running the tests. * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py: (test_run_test_pause_before_testing): 2012-03-01 Dirk Pranke nrwt: virtual tests aren't being run https://bugs.webkit.org/show_bug.cgi?id=80083 Reviewed by Eric Seidel. If we tried to run all of the tests by default (and didn't specify any paths), the virtual tests weren't picked up. * Scripts/webkitpy/layout_tests/port/base.py: (Port._virtual_tests): * Scripts/webkitpy/layout_tests/port/base_unittest.py: (PortTest.test_tests): * Scripts/webkitpy/layout_tests/port/test.py: (TestPort.virtual_test_suites): 2012-03-01 Ryosuke Niwa run-perf-tests should report the test runtime https://bugs.webkit.org/show_bug.cgi?id=80078 Reviewed by Hajime Morita. Record and output the aggreated test run time. * Scripts/webkitpy/performance_tests/perftestsrunner.py: (PerfTestsRunner._run_single_test): 2012-03-01 Michael Nordman [chromium] Use webkit_support to create session storage namespaces in DRT. https://bugs.webkit.org/show_bug.cgi?id=79933 Reviewed by Tony Chang. * DumpRenderTree/chromium/WebViewHost.cpp: (WebViewHost::createSessionStorageNamespace): 2012-03-01 Dirk Pranke nrwt: test_isfile, test_isdir, and test_exists don't work for virtual tests https://bugs.webkit.org/show_bug.cgi?id=80048 Reviewed by Adam Barth. * Scripts/webkitpy/layout_tests/port/base.py: (Port.test_isfile): (Port.test_isdir): (Port): (Port.test_exists): * Scripts/webkitpy/layout_tests/port/base_unittest.py: (PortTest.test_test_exists): (PortTest): (PortTest.test_test_isfile): (PortTest.test_test_isdir): * Scripts/webkitpy/layout_tests/port/test.py: (TestPort.virtual_test_suites): * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py: (MainTest.test_virtual): 2012-03-01 Kalev Lember [GTK] Avoid building any parts of WebKitTestRunner when WK2 is disabled https://bugs.webkit.org/show_bug.cgi?id=80031 Reviewed by Gustavo Noronha Silva. Only build the libTestRunnerInjectedBundle.la convenience library when ENABLE_WEBKIT2 is defined. * WebKitTestRunner/GNUmakefile.am: 2012-03-01 Philippe Normand Unreviewed, second attempt to fix GTK NRWT after r109242. * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkDriver._start): 2012-03-01 Philippe Normand Unreviewed, GTK NRWT build fix attempt after r109242. * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkDriver._start): 2012-03-01 Kentaro Hara REGRESSION(r109036): run-bindings-tests fails because resolve-supplemental.pl was renamed https://bugs.webkit.org/show_bug.cgi?id=80008 Reviewed by Adam Barth. In r109036, I forgot to rename "resolve-supplemental.pl" in run-bindings-tests to "preprocess-idls.pl". This patch renames it. Test: Confirm that run-bindings-tests runs. * Scripts/webkitpy/bindings/main.py: (BindingsTests.generate_supplemental_dependency): 2012-02-29 Simon Hausmann [Qt][WK2] QQuickWebView::event should lookup faster which events QQuickWebPage can handle https://bugs.webkit.org/show_bug.cgi?id=78047 Reviewed by Kenneth Rohde Christiansen. Send events for the webview through the QQuickCanvas instead of directly to the item. Direct sending to the item requires the QObject::event re-implementation to dispatch to the correct event handling function. However QQuickItem::event doesn't do that. Instead QQuickCanvas calls the event handlers directly, and with the removal of our own dispatching we need to go through that code path. * WebKitTestRunner/qt/PlatformWebViewQt.cpp: (WTR::PlatformWebView::sendEvent): (WTR::PlatformWebView::postEvent): 2012-02-29 Adam Barth Disable watchlist. It's been spamming CCs. * Scripts/webkitpy/tool/bot/stylequeuetask.py: (StyleQueueTask.run): * Scripts/webkitpy/tool/commands/queues_unittest.py: 2012-02-29 Dirk Pranke need to clone the gpu baselines and expectations into the new "virtual" dir https://bugs.webkit.org/show_bug.cgi?id=79956 Reviewed by Adam Barth. * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumPort.virtual_test_suites): 2012-02-29 Mark Rowe Switch to using json rather than simplejson now that the master is using Python 2.7. * BuildSlaveSupport/build.webkit.org-config/master.cfg: (loadBuilderConfig): 2012-02-29 David Levin [chromium] Add the ability to turn off autoresize. https://bugs.webkit.org/show_bug.cgi?id=77452 Reviewed by Darin Fisher. * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::LayoutTestController): Expose a way to turn off auto-resize. (LayoutTestController::enableAutoResizeMode): Ditto and adjust to new WebKit api. (LayoutTestController::disableAutoResizeMode): Ditto. * DumpRenderTree/chromium/LayoutTestController.h: Ditto. 2012-02-29 Dirk Pranke nrwt: make --print trace work with virtual tests https://bugs.webkit.org/show_bug.cgi?id=79952 Reviewed by Adam Barth. * Scripts/webkitpy/layout_tests/views/printing.py: (Printer._print_test_trace): (Printer._print_baseline): 2012-02-28 Brian Weinstein WebKit2: didNewFirstVisuallyNonEmptyLayout should be sent to injected bundle https://bugs.webkit.org/show_bug.cgi?id=79849 Reviewed by Beth Dakin. * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp: (WTR::InjectedBundlePage::InjectedBundlePage): Add an entry to the struct. 2012-02-29 Lucas Forschler Removed unused Apple buildslaves from the master configuration https://bugs.webkit.org/show_bug.cgi?id=79925 Reviewed by Stephanie Lewis. * BuildSlaveSupport/build.webkit.org-config/config.json: 2012-02-29 Dirk Pranke nrwt: implement simple 'virtual test suite' support https://bugs.webkit.org/show_bug.cgi?id=79737 Reviewed by Adam Barth. Add very basic support for 'virtual test suites' to NRWT: a virtual suite is a directory that may contain new baselines, contains a pointer to a "base" directory of tests, and an optional list of command line arguments to pass to DRT; it gives us a way to run all of the tests in a given directory multiple times with multiple (potentially differing) sets of results and baselines. This patch implements the support needed for this feature, and some basic tests, but no actual port will use the feature yet. This is probably the simplest implementation possible - the list of virtual suites will be hard-coded into the port's implementation. One can imagine a more data-driven approach where the suite instructions are stored in a manifest file either in LayoutTests (all suites in one file), or per-port, or per-directory. If this feature ends up being useful we should probably add something like that so people don't have to hunt around in the code to add suites. * Scripts/webkitpy/layout_tests/port/base.py: (Port.__init__): (Port.expected_filename): (Port.tests): (Port): (Port._real_tests): (Port._driver_class): (Port.virtual_test_suites): (Port.virtual_suite): (Port.virtual_suite.VirtualTestSuite): (Port.virtual_suite.VirtualTestSuite.__init__): (Port.virtual_suite.VirtualTestSuite.__repr__): (Port.populated_virtual_test_suites): (Port._virtual_tests): (Port.lookup_virtual_test_base): (Port.lookup_virtual_test_args): * Scripts/webkitpy/layout_tests/port/base_unittest.py: (PortTest.test_find_with_skipped_directories): * Scripts/webkitpy/layout_tests/port/driver.py: (DriverInput.__init__): (DriverProxy.run_test): * Scripts/webkitpy/layout_tests/port/test.py: (TestPort.virtual_test_suites): (TestDriver.run_test): * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py: (MainTest.test_all): (MainTest.test_virtual): 2012-02-29 Dirk Pranke nrwt: support more than two drivers in DriverProxy https://bugs.webkit.org/show_bug.cgi?id=79736 Reviewed by Adam Barth. Now that we can support per-test command lines for Drivers, modify DriverProxy to keep a map of running drivers for each needed command-line; this will allow us to transparently maintain a pool of appropriately configured DRTs without having to constantly start and stop them. Note that this potentially raises a garbage collection problem - the number of running DRTs will grow with the number of different sets of command line args. For now this is no worse than the current code - if you're running with pixel tests, you will only need one DRT per worker, and if you aren't, you'll need two (one for text-only tests, and one for reftests). An alternative would be to only ever have one running driver, and restart the driver as the command line changes, but this might (?) slow down execution in the text-only case - we should benchmark this because it would be simpler and possibly allow us to eliminate DriverProxy altogether. * Scripts/webkitpy/layout_tests/port/driver.py: (DriverProxy.__init__): (DriverProxy): (DriverProxy._make_driver): (DriverProxy.run_test): (DriverProxy.has_crashed): (DriverProxy.stop): (DriverProxy.cmd_line): (DriverProxy._cmd_line_as_key): 2012-02-29 Adrienne Walker Unreviewed, add myself as a reviewer * Scripts/webkitpy/common/config/committers.py: 2012-02-29 Greg Billock Add more tests for web intents https://bugs.webkit.org/show_bug.cgi?id=79527 Reviewed by Adam Barth. * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::LayoutTestController): (LayoutTestController::sendWebIntentResponse): * DumpRenderTree/chromium/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/chromium/WebViewHost.h: (WebViewHost): (WebViewHost::currentIntentRequest): 2012-02-29 Dirk Pranke nrwt: port/Driver needs to support per-test command line args https://bugs.webkit.org/show_bug.cgi?id=79733 Reviewed by Adam Barth. As part of removing the 'gpu' configurations and adding support for 'virtual test suites', the Driver classes need to support per-test command lines (since different tests will need to be run with different command line options). The per-test args are not yet used, so this change should have no visible effects and need no additional testing. * Scripts/webkitpy/layout_tests/port/base.py: (Port.driver_cmd_line): * Scripts/webkitpy/layout_tests/port/chromium.py: (ChromiumDriver.cmd_line): (ChromiumDriver._start): (ChromiumDriver.start): * Scripts/webkitpy/layout_tests/port/chromium_unittest.py: (ChromiumDriverTest.test_two_drivers.MockDriver.cmd_line): (ChromiumDriverTest): (ChromiumDriverTest.test_two_drivers): * Scripts/webkitpy/layout_tests/port/driver.py: (Driver.cmd_line): (DriverProxy.__init__): (DriverProxy.start): (DriverProxy.cmd_line): * Scripts/webkitpy/layout_tests/port/driver_unittest.py: (DriverTest.test_virtual_driver_methods): * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkDriver.cmd_line): * Scripts/webkitpy/layout_tests/port/mock_drt.py: * Scripts/webkitpy/layout_tests/port/test.py: (TestPort._path_to_driver): (TestDriver.cmd_line): (TestDriver.start): * Scripts/webkitpy/layout_tests/port/webkit.py: (WebKitDriver.cmd_line): (WebKitDriver._start): (WebKitDriver.run_test): (WebKitDriver.start): * Scripts/webkitpy/layout_tests/port/webkit_unittest.py: (WebKitDriverTest.test_no_timeout): * Scripts/webkitpy/performance_tests/perftestsrunner.py: (PerfTestsRunner._run_tests_set): * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py: (test_run_test_pause_before_testing): 2012-02-29 Dirk Pranke perftestsrunner can call printer.write() after printer.cleanup() https://bugs.webkit.org/show_bug.cgi?id=79872 Reviewed by Ryosuke Niwa. This patch fixes a unittest that was incorrectly calling PerfTestsRunner.run() twice (which led to a weird state in the printer object). * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py: (MainTest.test_run_with_upload_json): 2012-02-29 Balazs Kelemen [Qt][WK2] WebKitTestRunner should use 480x360 sized view for W3C SVG tests https://bugs.webkit.org/show_bug.cgi?id=76546 Reviewed by Simon Hausmann. Revert the workaround introduced in r107868 and reintroduce the qml binding so we can test the code path that we are more interested in. The reason that it did not work before is that we don't have an active platform window which we now workaround by propagating the resize through QWindowSystemInterface. * WebKitTestRunner/qt/PlatformWebViewQt.cpp: (WTR::WrapperWindow::handleStatusChanged): (WTR::PlatformWebView::resizeTo): 2012-02-29 Gabor Rapcsanyi [GTK] Small fix for cross compilation to not generate gtkdoc. Rubber-stamped by Csaba Osztrogonác. * Scripts/webkitdirs.pm: (isCrossCompilation): 2012-02-28 Lucas Forschler Fix a typo in config.json linx -> linux * BuildSlaveSupport/build.webkit.org-config/config.json: 2012-02-28 Mahesh Kulkarni [Qt] Allow read/write to the WebView.url property https://bugs.webkit.org/show_bug.cgi?id=77554 Reviewed by Tor Arne Vestbø. Fix MiniBrowser to use WebView.url property and remove WebView.load usage. * MiniBrowser/qt/qml/BrowserWindow.qml: 2012-02-28 Dan Winship [GTK] make tests explicitly request GSettingsBackendMemory https://bugs.webkit.org/show_bug.cgi?id=79788 Reviewed by Martin Robinson. Explicitly request the fallback memory-based GSettings backend; if other backends are available then we don't want to use them, because they would make tests depend on the user's preferences. And if other backends aren't available, we don't need glib to print a warning about that fact. * Scripts/webkitpy/layout_tests/port/gtk.py: (GtkPort.setup_environ_for_server): 2012-02-28 Sheriff Bot Unreviewed, rolling out r109110. http://trac.webkit.org/changeset/109110 https://bugs.webkit.org/show_bug.cgi?id=79806 webkitpy makes it too easy to add wkb.ug URLs to ChangeLogs (Requested by aroben on #webkit). * Scripts/webkitpy/common/checkout/changelog.py: (parse_bug_id): (parse_bug_id_from_changelog): * Scripts/webkitpy/common/checkout/changelog_unittest.py: * Scripts/webkitpy/common/config/urls.py: 2012-02-28 Adam Roben Add my new email address Fixes Adam Roben's email address in webkitpy is soon to be incorrect Reviewed by Anders Carlsson. * Scripts/webkitpy/common/config/committers.py: 2012-02-28 Adam Roben Teach webkitpy about wkb.ug URLs Fixes webkit-patch can't find bug numbers from ChangeLog entries that use wkb.ug URLs Reviewed by Anders Carlsson. * Scripts/webkitpy/common/checkout/changelog.py: (parse_bug_id): (parse_bug_id_from_changelog): Fall back to the tiny URL format if neither the short nor long URL formats work. * Scripts/webkitpy/common/checkout/changelog_unittest.py: (ChangeLogTest.test_parse_bug_id_from_changelog): Added tests for the above changes. * Scripts/webkitpy/common/config/urls.py: Added bug_url_tiny, which can be used to match wkb.ug URLs. 2012-02-28 János Badics [Qt] build-webkit does not detect the number of CPUs https://bugs.webkit.org/show_bug.cgi?id=76445 Based on Seo Sanghyeon's work (modification: Automatically determine the number of CPUs for make only if -j make argument haven't already been specified.) Reviewed by Csaba Osztrogonác. * Scripts/webkitdirs.pm: (buildQMakeProjects): 2012-02-28 Kent Tamura Update a pattern for forms files of watchlist https://bugs.webkit.org/show_bug.cgi?id=79767 Reviewed by Kentaro Hara. * Scripts/webkitpy/common/config/watchlist: Catches the following files with the "Forms" pattern. - HTMLDatListElement.* - HTMLKeygenElement.* - HTMLLegendElement.* - HTMLOutputElement.* - FormAssociatedElement.* - SliderThumbElement.* - TextControlInnerElements.* - RenderFileUploadControl.* 2012-02-27 Sheriff Bot Unreviewed, rolling out r109041. http://trac.webkit.org/changeset/109041 https://bugs.webkit.org/show_bug.cgi?id=79741 Broke Chromium Win build (Requested by bashi on #webkit). * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::LayoutTestController): * DumpRenderTree/chromium/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/chromium/WebViewHost.h: 2012-02-27 James Robinson [chromium] Remove media tests from ChromiumGpu configurations https://bugs.webkit.org/show_bug.cgi?id=79720 Reviewed by Adam Barth. We no longer need to run the media/ tests in both GPU and non-GPU configurations since we can exercise the GPU composited playback path via tests in the compositing/ directory. * Scripts/webkitpy/layout_tests/port/chromium_gpu.py: (_default_tests_paths): * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py: (ChromiumGpuTest.test_default_tests_paths): (ChromiumGpuTest.test_test_files): 2012-02-27 Dirk Pranke nrwt: log multiline stderr/crash output properly https://bugs.webkit.org/show_bug.cgi?id=79722 Reviewed by Adam Barth. This cleans up some nits in the log output to make things slightly easier to read (every line gets timestamped, etc.). Also tweaks the log message for pixel hash mismatches. * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py: (SingleTestRunner._handle_error): (SingleTestRunner._compare_image): 2012-02-27 Ryosuke Niwa Touch the master.cfg to force a reconfig. * BuildSlaveSupport/build.webkit.org-config/master.cfg: (BuildAndTestLeaksFactory): (NewBuildAndTestFactory): 2012-02-27 Jacky Jiang [BlackBerry] Upstream BlackBerry API web page related files https://bugs.webkit.org/show_bug.cgi?id=74380 Reviewed by Antonio Gomes. Adapt to the removing of WebPage::mainFrame(), use DumpRenderTreeSupport::corePage() and its mainFrame() instead. No new tests as this patch doesn't change behavior. * DumpRenderTree/blackberry/DumpRenderTree.cpp: (BlackBerry::WebKit::DumpRenderTree::resetToConsistentStateBeforeTesting): (BlackBerry::WebKit::DumpRenderTree::runTests): 2012-02-27 Ryosuke Niwa Add Windows and Linux Chromium perf bots https://bugs.webkit.org/show_bug.cgi?id=79418 Reviewed by Tony Chang. Add Chromium Win and Linux perf bots. Also replace Chromium Mac by the dedicated perf bot slave (uses the same hardware). * BuildSlaveSupport/build.webkit.org-config/config.json: 2012-02-27 Xianzhu Wang Undefined variable 'current' when running chromium_android.py https://bugs.webkit.org/show_bug.cgi?id=79696 Reviewed by Adam Barth. * Scripts/webkitpy/layout_tests/port/chromium_android.py: (ChromiumAndroidDriver._read_prompt): 2012-02-27 Greg Billock Add more tests for web intents https://bugs.webkit.org/show_bug.cgi?id=79527 Reviewed by Adam Barth. * DumpRenderTree/chromium/LayoutTestController.cpp: (LayoutTestController::LayoutTestController): (LayoutTestController::sendWebIntentResponse): * DumpRenderTree/chromium/LayoutTestController.h: (LayoutTestController): * DumpRenderTree/chromium/WebViewHost.h: (WebViewHost): (WebViewHost::currentIntentRequest): 2012-02-27 Xianzhu Wang Chromium-Android layout test script: the adb shell prompt is not always '# ' https://bugs.webkit.org/show_bug.cgi?id=79506 Reviewed by Adam Barth. * Scripts/webkitpy/layout_tests/port/chromium_android.py: (ChromiumAndroidDriver._start): (ChromiumAndroidDriver._read_prompt): 2012-02-27 Dan Bernstein Added a test for REGRESSION (WebKit2): Non-activating links sometimes don’t work https://bugs.webkit.org/show_bug.cgi?id=79607 Reviewed by Adele Peterson. * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Added new files. * TestWebKitAPI/Tests/mac/AcceptsFirstMouse.mm: Added. (TestWebKitAPI::AcceptsFirstMouse::url): (TestWebKitAPI::AcceptsFirstMouse::didLoadURL): (TestWebKitAPI::AcceptsFirstMouse::runTest): (TestWebKitAPI::TEST_F): * TestWebKitAPI/Tests/mac/acceptsFirstMouse.html: Added. 2012-02-27 Carlos Garcia Campos Unreviewed. Fix make distcheck. * GNUmakefile.am: Fix typo. 2012-02-27 Tor Arne Vestbø [Qt] Use USE() macro instead of ENABLE() for using the Qt image decoder Reviewed by Kenneth Rohde Christiansen.. * qmake/mkspecs/features/features.prf: 2012-02-27 Hao Zheng Add a ChromiumGpuAndroid port. https://bugs.webkit.org/show_bug.cgi?id=79628 Reviewed by Adam Barth. Android does have a gpu port, so revert r107697. * Scripts/webkitpy/layout_tests/port/chromium_android.py: (ChromiumAndroidPort.__init__): * Scripts/webkitpy/layout_tests/port/chromium_gpu.py: (ChromiumGpuAndroidPort): (ChromiumGpuAndroidPort.__init__): (ChromiumGpuAndroidPort.baseline_search_path): (ChromiumGpuAndroidPort.default_child_processes): (ChromiumGpuAndroidPort.tests): * Scripts/webkitpy/layout_tests/port/factory.py: (PortFactory): 2012-02-25 Gyuyoung Kim Change Build bot with BuildAndTest bot for EFL port. https://bugs.webkit.org/show_bug.cgi?id=77518 Reviewed by Ryosuke Niwa. EFL build bot starts to support layout test. * BuildSlaveSupport/build.webkit.org-config/config.json: * BuildSlaveSupport/built-product-archive: (archiveBuiltProduct): (extractBuiltProduct): * BuildSlaveSupport/test-result-archive: (archiveTestResults): * Scripts/webkitdirs.pm: Enable SHARED_CORE option when ENABLE_DRT is enable in EFL port. (generateBuildSystemFromCMakeProject): 2012-02-24 Darin Fisher Add watchlist for ChromiumPlatformApi matching Source/Platform/chromium/public/ Reviewed by James Robinson. * Scripts/webkitpy/common/config/watchlist: 2012-02-24 Dirk Pranke webkitpy: changelog_unittest test_fuzzy_reviewer_match is slow https://bugs.webkit.org/show_bug.cgi?id=79258 Reviewed by Ryosuke Niwa. Split the fuzzy matching into separate functions so that they can be sharded and run in parallel. Also, fix a bug in the svn_blame_regexp that was causing us to require ChangeLog lines to have at least one space after the author, making the code require trailing whitespace in the changelog in some places (this was making my life harder because I have my text editor set to strip trailing whitespace in python files, and that was causing test_parse_log_entries_from_changelog() to fail when the whitespace on line 275 was removed). * Scripts/webkitpy/common/checkout/changelog.py: (ChangeLog): * Scripts/webkitpy/common/checkout/changelog_unittest.py: (test_fuzzy_reviewer_match__none): (test_fuzzy_reviewer_match_adam_barth): (test_fuzzy_reviewer_match_darin_adler_et_al): (test_fuzzy_reviewer_match_dimitri_glazkov): (test_fuzzy_reviewer_match_george_staikos): (test_fuzzy_reviewer_match_mark_rowe): 2012-02-24 Dirk Pranke webkitpy: committers_unittest fuzzy matching is really slow https://bugs.webkit.org/show_bug.cgi?id=79179 Reviewed by Ryosuke Niwa. Break the fuzzy matching tests into individual routines for each contributor so that the intent is a little clearer, so that it's easier to test individual names (and identify duplicate tests), and so that we can eventually run them in parallel. Also, mark most of them as 'integration' tests so that they can be skipped if we want to go fast (we leave a few as regular tests for functional coverage). * Scripts/webkitpy/common/config/committers_unittest.py: (CommittersTest): (CommittersTest.test_contributors_by_fuzzy_match): (CommittersTest.integration_test_contributors__none): (CommittersTest.integration_test_contributors__none_2): (CommittersTest.integration_test_contributors__none_3): (CommittersTest.integration_test_contributors_ada_chan): (CommittersTest.integration_test_contributors_adele_peterson): (CommittersTest.integration_test_contributors_alexey_proskuryakov): (CommittersTest.integration_test_contributors_alice_liu): (CommittersTest.integration_test_contributors_alp_toker): (CommittersTest.integration_test_contributors_anders_carlsson): (CommittersTest.integration_test_contributors_antti_koivisto): (CommittersTest.integration_test_contributors_beth_dakin): (CommittersTest.integration_test_contributors_brady_eidson): (CommittersTest.integration_test_contributors_cameron_zwarich): (CommittersTest.integration_test_contributors_chris_blumenberg): (CommittersTest.integration_test_contributors_dan_bernstein): (CommittersTest.integration_test_contributors_dan_bernstein_2): (CommittersTest.integration_test_contributors_darin_adler): (CommittersTest.integration_test_contributors_david_harrison): (CommittersTest.integration_test_contributors_david_harrison_2): (CommittersTest.integration_test_contributors_david_hyatt): (CommittersTest.integration_test_contributors_david_kilzer): (CommittersTest.integration_test_contributors_don_melton): (CommittersTest.integration_test_contributors_eric_seidel): (CommittersTest.integration_test_contributors_geoffrey_garen): (CommittersTest.integration_test_contributors_greg_bolsinga): (CommittersTest.integration_test_contributors_holger_freyther): (CommittersTest.integration_test_contributors_jon_sullivan): (CommittersTest.integration_test_contributors_jon_honeycutt): (CommittersTest.integration_test_contributors_joseph_pecoraro): (CommittersTest.integration_test_contributors_ken_kocienda): (CommittersTest.integration_test_contributors_kenneth_russell): (CommittersTest.integration_test_contributors_kevin_decker): (CommittersTest.integration_test_contributors_kevin_mccullough): (CommittersTest.integration_test_contributors_lars_knoll): (CommittersTest.integration_test_contributors_lars_weintraub): (CommittersTest.integration_test_contributors_maciej_stachowiak): (CommittersTest.integration_test_contributors_mark_rowe): (CommittersTest.integration_test_contributors_nikolas_zimmermann): (CommittersTest.integration_test_contributors_oliver_hunt): (CommittersTest.integration_test_contributors_rniwa): (CommittersTest.disabled_integration_test_contributors_simon_fraser): (CommittersTest.integration_test_contributors_steve_falkenburg): (CommittersTest.integration_test_contributors_sam_weinig): (CommittersTest.integration_test_contributors_tim_omernick): (CommittersTest.integration_test_contributors_timothy_hatcher): (CommittersTest.integration_test_contributors_tor_arne_vestbo): (CommittersTest.integration_test_contributors_vicki_murley): (CommittersTest.integration_test_contributors_zack_rusin): 2012-02-24 Dirk Pranke webkitpy: should autoinstall coverage https://bugs.webkit.org/show_bug.cgi?id=79535 Reviewed by Eric Seidel. We should automatically print the coverage report after the run, too. * Scripts/webkitpy/test/main.py: (Tester._run_tests): * Scripts/webkitpy/thirdparty/__init__.py: (AutoinstallImportHook.find_module): (AutoinstallImportHook._install_coverage): 2012-02-24 Pablo Flouret Fix copypasta in build-webkit's --vibration option description https://bugs.webkit.org/show_bug.cgi?id=79528 Reviewed by Andreas Kling. * Scripts/build-webkit: 2012-02-24 Antonio Gomes Adding myself to ScrollingCoordinator watchlist https://bugs.webkit.org/show_bug.cgi?id=79525 Reviewed by Rob Buis. * Scripts/webkitpy/common/config/watchlist: 2012-02-24 Darin Fisher Update watchlist for ChromiumPublicApi to include Platform/chromium/public/ Reviewed by Adam Barth. * Scripts/webkitpy/common/config/watchlist: 2012-02-24 Dirk Pranke webkitpy: config_unittest is sensitive to other tests that might've already run https://bugs.webkit.org/show_bug.cgi?id=79382 Reviewed by Adam Barth. If another test runs prior to config_unittest and pokes the cached configuration, the first test in config_unittest might fail; the tearDown() will then allow subsequent tests to pass. Clearing the configuration prior to each test makes things slightly slower but safer. * Scripts/webkitpy/layout_tests/port/config_unittest.py: (ConfigTest.setUp): 2012-02-24 David Kilzer Use xcrun to find compiler paths for Generate Derived Sources build phase script Reviewed by Dan Bernstein. * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj: (Generate Derived Sources): Use xcrun to find the path to the compiler since that works on both iOS and OS X. 2012-02-24 James Robinson Create a watchlist for ScrollingCoordinator changes https://bugs.webkit.org/show_bug.cgi?id=79511 Reviewed by Anders Carlsson. * Scripts/webkitpy/common/config/watchlist: 2012-02-24 Grzegorz Czajkowski [EFL][DRT] Implement setValueForUser and setAutofilled. https://bugs.webkit.org/show_bug.cgi?id=79237 Reviewed by Andreas Kling. Adds missing implementation setValueForUser and setAutofilled to EFL's LayoutTestController. Those implementations are related with input fields and allow to pass following tests: fast/forms/onchange-setvalueforuser.html fast/forms/input-autofilled.html fast/forms/reset-autofilled.html * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: (LayoutTestController::setValueForUser): (LayoutTestController::setAutofilled): 2012-02-24 Zoltan Horvath [Qt] Allow to use WebCore imagedecoders https://bugs.webkit.org/show_bug.cgi?id=32410 Add ENABLE_QT_IMAGE_DECODER macro, it's enabled by default. Reviewed by Simon Hausmann. * qmake/mkspecs/features/features.prf: 2012-02-24 Balazs Kelemen [Qt] Add Qt5 way to force 96 DPI for tests https://bugs.webkit.org/show_bug.cgi?id=79364 Reviewed by Simon Hausmann. * DumpRenderTree/qt/main.cpp: (main): * WebKitTestRunner/InjectedBundle/qt/ActivateFontsQt.cpp: (WTR::activateFonts): 2012-02-24 Carlos Garcia Campos [GTK] MiniBrowser should reload the view then the reload button is clicked https://bugs.webkit.org/show_bug.cgi?id=79472 Reviewed by Philippe Normand. * MiniBrowser/gtk/BrowserWindow.c: (reloadCallback): Call webkit_web_view_reload(). (browser_window_init): Connect reload item clicked even to reloadCallback instead of activateUriEntryCallback. 2012-02-24 Kentaro Hara Add performance tests for DOM attribute getters and setters https://bugs.webkit.org/show_bug.cgi?id=79208 Reviewed by Ryosuke Niwa. We want to allow perf tests to output info messages that do not affect the test results. To accomplish this, this patch modifies perftestsrunner.py so that it ignores a line that starts from "Info:". * Scripts/webkitpy/performance_tests/perftestsrunner.py: (PerfTestsRunner): 2012-02-24 Andrew Lo [BlackBerry] Enable requestAnimationFrame https://bugs.webkit.org/show_bug.cgi?id=79408 Enable requestAnimationFrame for BlackBerry. Reviewed by Rob Buis. * Scripts/build-webkit: 2012-02-24 Adam Barth style-queue tries to apply the patch twice https://bugs.webkit.org/show_bug.cgi?id=79459 Reviewed by Eric Seidel. Previously, we applied the patch using both apply-patch and check-style. This patch introduces a check-style-local to mirror the apply-watchlist-local, which operates on the current working diff. This patch also cleans up some other bugs I found by running the queue locally for a while. The queue still prints out a slightly less-than-ideal message on the bugs when it find an error, but it's ok for now. We'll need to iterate a bit. * Scripts/webkitpy/tool/bot/stylequeuetask.py: (StyleQueueTask.validate): (StyleQueueTask._check_style): * Scripts/webkitpy/tool/commands/download.py: (CheckStyleLocal): * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py: (EarlyWarningSytemTest._default_expected_stderr): (EarlyWarningSytemTest._test_testing_ews): * Scripts/webkitpy/tool/commands/queues.py: (AbstractReviewQueue.handle_script_error): (StyleQueue.review_patch): * Scripts/webkitpy/tool/commands/queues_unittest.py: * Scripts/webkitpy/tool/commands/queuestest.py: (QueuesTest.assert_queue_outputs): 2012-02-23 Adam Barth should_proceed_with_work_item is unused and can be removed https://bugs.webkit.org/show_bug.cgi?id=79416 Reviewed by Eric Seidel. We used to use this function to check whether the tree is red. Now, we don't use external measures of whether to proceed with work items. Instead, we analyze them with the idea in mind that the tree might be red. * Scripts/webkitpy/tool/bot/queueengine.py: (QueueEngineDelegate.next_work_item): (QueueEngine.run): * Scripts/webkitpy/tool/bot/queueengine_unittest.py: (LoggingDelegate): (LoggingDelegate.next_work_item): (RaisingDelegate.process_work_item): (QueueEngineTest.test_terminating_error): * Scripts/webkitpy/tool/commands/earlywarningsystem.py: (AbstractEarlyWarningSystem.__init__): * Scripts/webkitpy/tool/commands/queues.py: (AbstractQueue.next_work_item): (FeederQueue.next_work_item): (CommitQueue.next_work_item): (AbstractReviewQueue.next_work_item): (StyleQueue.__init__): * Scripts/webkitpy/tool/commands/queues_unittest.py: (FeederQueueTest.test_feeder_queue): (CommitQueueTest.test_commit_queue): (test_commit_queue_failure): (test_commit_queue_failure_with_failing_tests): (test_rollout): (test_rollout_lands): (StyleQueueTest.test_style_queue_with_style_exception): (test_style_queue_with_watch_list_exception): * Scripts/webkitpy/tool/commands/queuestest.py: (QueuesTest.assert_queue_outputs): * Scripts/webkitpy/tool/commands/sheriffbot.py: (SheriffBot.next_work_item): 2012-02-23 Adrienne Walker Unreviewed, add Stephen Chenney to committers.py as a contributor. * Scripts/webkitpy/common/config/committers.py: 2012-02-23 Adam Barth style-queue logs bogus messages to the QueueStatusServer https://bugs.webkit.org/show_bug.cgi?id=79424 Reviewed by Eric Seidel. Copy/paste error. * Scripts/webkitpy/tool/bot/stylequeuetask.py: (StyleQueueTask._check_style): (StyleQueueTask._apply_watch_list): * Scripts/webkitpy/tool/commands/queues_unittest.py: 2012-02-23 Eric Seidel Use self._filesystem in more places in the SCM classes https://bugs.webkit.org/show_bug.cgi?id=79415 Reviewed by Adam Barth. * Scripts/webkitpy/common/checkout/scm/detection_unittest.py: (SCMDetectorTest.test_detect_scm_system): * Scripts/webkitpy/common/checkout/scm/git.py: (Git.find_checkout_root): (Git.to_object_name): * Scripts/webkitpy/common/checkout/scm/scm.py: (SCM.__init__): (SCM.in_working_directory): * Scripts/webkitpy/common/checkout/scm/svn.py: (SVN.find_uuid): (SVN.find_checkout_root): 2012-02-23 Sheriff Bot Unreviewed, rolling out r108685. http://trac.webkit.org/changeset/108685 https://bugs.webkit.org/show_bug.cgi?id=79414 Broke Chromium builds (Requested by enne on #webkit). * qmake/mkspecs/features/features.prf: 2012-02-23 Adam Barth style-queue shouldn't spam bugs if it fails to update its working copy https://bugs.webkit.org/show_bug.cgi?id=79412 Reviewed by Eric Seidel. This patch moves style-queue over to the new PatchAnalysisTask infrastructure, which is smarter about retrying patches when update fails. * Scripts/webkitpy/common/system/executive_mock.py: (MockExecutive.run_and_throw_if_fail): * Scripts/webkitpy/tool/bot/earlywarningsystemtask.py: * Scripts/webkitpy/tool/bot/patchanalysistask.py: (UnableToApplyPatch): (UnableToApplyPatch.__init__): (PatchAnalysisTask.__init__): * Scripts/webkitpy/tool/bot/stylequeuetask.py: Added. (StyleQueueTaskDelegate): (StyleQueueTaskDelegate.parent_command): (StyleQueueTask): (StyleQueueTask.validate): (StyleQueueTask._check_style): (StyleQueueTask._apply_watch_list): (StyleQueueTask.run): * Scripts/webkitpy/tool/commands/earlywarningsystem.py: * Scripts/webkitpy/tool/commands/queues.py: (StyleQueue): (StyleQueue.should_proceed_with_work_item): (StyleQueue.review_patch): (StyleQueue.run_command): (StyleQueue.command_passed): (StyleQueue.command_failed): (StyleQueue.expected_failures): (StyleQueue.refetch_patch): * Scripts/webkitpy/tool/commands/queues_unittest.py: (StyleQueueTest.test_style_queue_with_style_exception): (test_style_queue_with_watch_list_exception): 2012-02-23 Zoltan Horvath [Qt] Allow to use WebCore imagedecoders https://bugs.webkit.org/show_bug.cgi?id=32410 Add ENABLE_QT_IMAGE_DECODER macro, it's enabled by default. Reviewed by Simon Hausmann. * qmake/mkspecs/features/features.prf: 2012-02-23 Eric Seidel Remove old broken find_checkout_root tests now that find_checkout_root is gone https://bugs.webkit.org/show_bug.cgi?id=79405 Reviewed by Adam Barth. * Scripts/webkitpy/common/checkout/scm/detection_unittest.py: Added. (SCMDetectorTest): (SCMDetectorTest.test_find_checkout_root): * Scripts/webkitpy/common/checkout/scm/scm_unittest.py: (SVNTestRepository.tear_down): 2012-02-23 Michael Brüning [Qt][WK2] Implement proxy authentication dialog. https://bugs.webkit.org/show_bug.cgi?id=79350 Reviewed by Simon Hausmann. * MiniBrowser/qt/MiniBrowser.qrc: * MiniBrowser/qt/qml/BrowserWindow.qml: * MiniBrowser/qt/qml/ProxyAuthenticationDialog.qml: Added. 2012-02-23 Kevin Ollivier [wx] Build fix, add recently added source dirs. * waf/build/settings.py: 2012-02-23 Simon Hausmann [Qt] Add support for touch cancellation https://bugs.webkit.org/show_bug.cgi?id=79348 Reviewed by Kenneth Rohde Christiansen. Added straight-forward touch cancellation event sending to Qt WK1 DRT and WK2 WTR. * DumpRenderTree/qt/EventSenderQt.cpp: (EventSender::touchCancel): (EventSender::cancelTouchPoint): * DumpRenderTree/qt/EventSenderQt.h: (EventSender): * WebKitTestRunner/EventSenderProxy.h: (EventSenderProxy): * WebKitTestRunner/InjectedBundle/Bindings/EventSendingController.idl: * WebKitTestRunner/InjectedBundle/EventSendingController.cpp: (WTR::EventSendingController::touchCancel): (WTR): (WTR::EventSendingController::cancelTouchPoint): * WebKitTestRunner/InjectedBundle/EventSendingController.h: (EventSendingController): * WebKitTestRunner/TestController.cpp: (WTR::TestController::didReceiveSynchronousMessageFromInjectedBundle): * WebKitTestRunner/qt/EventSenderProxyQt.cpp: (WTR::EventSenderProxy::touchCancel): (WTR): (WTR::EventSenderProxy::cancelTouchPoint): 2012-02-23 Sheriff Bot Unreviewed, rolling out r108626. http://trac.webkit.org/changeset/108626 https://bugs.webkit.org/show_bug.cgi?id=79370 broke qt-wk2 bot, somehow the workaround does not work on the bot (Requested by Ossy on #webkit). * WebKitTestRunner/qt/PlatformWebViewQt.cpp: (WTR::WrapperWindow::handleStatusChanged): (WTR::PlatformWebView::resizeTo): 2012-02-23 ChangSeok Oh [GTK] DRT doesn't support scheduleAsynchronousKeyDown. https://bugs.webkit.org/show_bug.cgi?id=78481 Reviewed by Gustavo Noronha Silva. Extracted shareable code from keyDownCallback and then added createKeyPressEvent to create gdk press key event. Added sendKeyDown and sendAsynchronousKeyDown to evaluate keyDown event directly and asynchronously. Eventually, sendAsynchronousKeyDown calls sendKeyDown. * DumpRenderTree/gtk/EventSender.cpp: (scheduleAsynchronousClickCallback): (createKeyPressEvent): (sendKeyDown): (keyDownCallback): (sendAsynchronousKeyDown): (scheduleAsynchronousKeyDownCallback): 2012-02-23 Balazs Kelemen [Qt][WK2] WebKitTestRunner should use 480x360 sized view for W3C SVG tests https://bugs.webkit.org/show_bug.cgi?id=76546 Reviewed by Simon Hausmann. Co-authored with Simon Hausmann. Revert the workaround introduced in r107868 and reintroduce the qml binding so we can test the code path that we are more interested in. The reason that it did not work before is that we don't have an active platform window which we now workaround by propagating the resize through QWindowSystemInterface. * WebKitTestRunner/qt/PlatformWebViewQt.cpp: (WTR::WrapperWindow::handleStatusChanged): (WTR::PlatformWebView::resizeTo): 2012-02-23 Sheriff Bot Unreviewed, rolling out r108577. http://trac.webkit.org/changeset/108577 https://bugs.webkit.org/show_bug.cgi?id=79359 It doesn't work as expected (Requested by Ossy on #webkit). * DumpRenderTree/qt/EventSenderQt.cpp: (EventSender::postEvent): * DumpRenderTree/qt/EventSenderQt.h: (EventSender): 2012-02-23 Philippe Normand [GTK] FullScreen signals https://bugs.webkit.org/show_bug.cgi?id=76181 Reviewed by Martin Robinson. Use the two new entering/leaving fullscreen signals to hide/show the tool bar, the status bar and ask the user's permission before entering fullscreen. * GtkLauncher/main.c: (webViewFullscreenMessageWindowClose): (webViewWindowStateEvent): (hide_widget): (show_widget): (webViewEnteringFullScreen): (webViewLeavingFullScreen): (createBrowser): (createWindow): 2012-02-22 Dongwoo Im [EFL] Fix to launch the EWebLauncher using the script, run-launcher. https://bugs.webkit.org/show_bug.cgi?id=79076 Reviewed by Andreas Kling. * Scripts/webkitdirs.pm: Fix the path of libewebkit.so. (builtDylibPathForName): 2012-02-22 Allan Sandfeld Jensen [Qt] Add eventSender.gestureTap https://bugs.webkit.org/show_bug.cgi?id=66173 Reviewed by Kenneth Rohde Christiansen. * DumpRenderTree/qt/EventSenderQt.cpp: (EventSender::gestureTap): * DumpRenderTree/qt/EventSenderQt.h: 2012-02-22 Brady Eidson and https://bugs.webkit.org/show_bug.cgi?id=79279 Synchronous ShouldGoToBackForwardListItem causes lots of WebProcess hangs Reviewed by Anders Carlsson. Update our tools to the new form of the PageLoadClient. * MiniBrowser/mac/BrowserWindowController.m: (-[BrowserWindowController awakeFromNib]): * WebKitTestRunner/TestController.cpp: (WTR::TestController::initialize): 2012-02-22 Daniel Bates new-run-webkit-tests shouldn't restore Safari state on Lion when showing results https://bugs.webkit.org/show_bug.cgi?id=79277 Reviewed by Adam Barth. Pass command line argument no-saved-state to run-safari to disable Safari state restoration for the session on OS 10.7. * Scripts/webkitpy/layout_tests/port/mac.py: (MacPort.show_results_html_file): * Scripts/webkitpy/layout_tests/port/mac_unittest.py: (test_show_results_html_file): 2012-02-22 Kalev Lember Explicitly include unistd.h in TestNetscapePlugIn for GCC 4.7 https://bugs.webkit.org/show_bug.cgi?id=77759 Reviewed by Tony Chang. Fixes build with gcc 4.7. r100432 added the include only for ANDROID; this changes the ifdef to also cover other unix platforms. On Fedora 17 the gcc 4.7 compiler no longer implicitly includes unistd.h from standard headers and other unix platforms are likely to get gcc 4.7 as well in the future. * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp: 2012-02-22 Philippe Normand Unreviewed, rolling out r108522. http://trac.webkit.org/changeset/108522 https://bugs.webkit.org/show_bug.cgi?id=76181 Broke 4 fullscreen tests on GTK. * GtkLauncher/main.c: (createBrowser): (createWindow): 2012-02-20 Philippe Normand [GTK] FullScreen signals https://bugs.webkit.org/show_bug.cgi?id=76181 Reviewed by Martin Robinson. Use the two new entering/leaving fullscreen signals to hide/show the tool bar, the status bar and ask the user's permission before entering fullscreen. * GtkLauncher/main.c: (webViewFullscreenMessageWindowClose): (webViewWindowStateEvent): (hide_widget): (show_widget): (webViewEnteringFullScreen): (webViewLeavingFullScreen): (createBrowser): (createWindow): 2012-02-22 Rob Flack Update Linux ChromiumOS builders in flakiness dashboard. https://bugs.webkit.org/show_bug.cgi?id=79238 Reviewed by Ojan Vafai. * TestResultServer/static-dashboards/builders.js: 2012-02-22 Dirk Pranke webkitpy: test_function_length_check_definition_huge_lines is slow https://bugs.webkit.org/show_bug.cgi?id=79185 Reviewed by Adam Barth. This test takes ~3 seconds to run, apparently because we're running a test on a 10,000 line function :). Dropping the function length to 640 still exercises the code path and provides a 10x speedup. * Scripts/webkitpy/style/checkers/cpp_unittest.py: (CppStyleTest): (CheckForFunctionLengthsTest.test_function_length_check_definition_huge_lines): 2012-02-22 Dirk Pranke webkitpy: speed up FileLockTest.test_lock_lifecycle https://bugs.webkit.org/show_bug.cgi?id=79189 Reviewed by Adam Barth. The unit test had a second lock trying to acquire a lock already held, and a hard-coded timeout of one second. I have changed the timeouts to a tenth of a second (which is probably still fine) and changed the hard-spin on the lock to a slow-spin every hundredth of a second. * Scripts/webkitpy/common/system/file_lock.py: (FileLock.acquire_lock): * Scripts/webkitpy/common/system/file_lock_integrationtest.py: (FileLockTest.setUp): (FileLockTest.test_stuck_lock): 2012-02-22 Gabor Rapcsanyi build-webkit should use environment variable at number of CPU detection on all platform https://bugs.webkit.org/show_bug.cgi?id=79227 Reviewed by Martin Robinson. * Scripts/webkitdirs.pm: (determineNumberOfCPUs): 2012-02-22 Mihnea Ovidenie [Qt] REGRESSION(r108108): It made 3 fast/repaint tests fail https://bugs.webkit.org/show_bug.cgi?id=78960 Reviewed by Csaba Osztrogonác. * DumpRenderTree/qt/LayoutTestControllerQt.cpp: (LayoutTestController::overridePreference): 2012-02-22 Hao Zheng [chromium] Build WebKit with MEDIA_STREAM disabled. https://bugs.webkit.org/show_bug.cgi?id=79214 Reviewed by Kent Tamura. * DumpRenderTree/chromium/WebUserMediaClientMock.cpp: * DumpRenderTree/chromium/WebUserMediaClientMock.h: * DumpRenderTree/chromium/WebViewHost.cpp: (WebViewHost::createMediaPlayer): * DumpRenderTree/chromium/WebViewHost.h: (WebViewHost): 2012-02-22 Zeno Albisser [Qt][WK2][Mac] WebKitTestRunner does not build with clang. https://bugs.webkit.org/show_bug.cgi?id=79124 WebKitTestRunnerPrefix.h should also be included when building with clang. Reviewed by Kenneth Rohde Christiansen. * WebKitTestRunner/Target.pri: 2012-02-20 Roland Steiner