Tools/ChangeLog

   1 2012-05-08  Kangil Han  <kangil.han@samsung.com>
   2
   3         [EFL][DRT] Implement TextInputController.
   4         https://bugs.webkit.org/show_bug.cgi?id=78559
   5
   6         Reviewed by Martin Robinson.
   7
   8         TextInputController has been created for an input emulation purpose in DRT.
   9         However, not like other ports, EFL doesn't have it.
  10         Therefore, this patch will enable EFL port to have an input emulation in DRT.
  11
  12         * DumpRenderTree/efl/CMakeLists.txt:
  13         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
  14         (DumpRenderTreeChrome::onWindowObjectCleared):
  15         * DumpRenderTree/efl/TextInputController.cpp: Added.
  16         (setMarkedTextCallback):
  17         (hasMarkedTextCallback):
  18         (markedRangeCallback):
  19         (insertTextCallback):
  20         (unmarkTextCallback):
  21         (firstRectForCharacterRangeCallback):
  22         (selectedRangeCallback):
  23         (makeTextInputController):
  24         * DumpRenderTree/efl/TextInputController.h: Added.
  25
  26 2012-04-23  Raphael Kubo da Costa  <rakuco@webkit.org>
  27
  28         [EFL] Pass a WTF::String directly from DRT to ewebkit when possible
  29         https://bugs.webkit.org/show_bug.cgi?id=84634
  30
  31         Reviewed by Gustavo Noronha Silva.
  32
  33         Drop unnecessary const char* -> WTF::String conversions by
  34         creating String objects directly in LayoutTestController out of
  35         JSStringRef's and passing them to DumpRenderTreeSupportEfl.
  36
  37         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp:
  38         (LayoutTestController::queueLoad):
  39         (LayoutTestController::setValueForUser):
  40         (LayoutTestController::findString):
  41         (LayoutTestController::addUserStyleSheet):
  42
  43 2012-05-08  Alexander Shalamov  <alexander.shalamov@intel.com>
  44
  45         [EFL][DRT] EventSender needs implementation of leapForward function
  46         https://bugs.webkit.org/show_bug.cgi?id=83486
  47
  48         Reviewed by Eric Seidel.
  49
  50         Implemented leapForward function in Efl's DRT. This function is required
  51         to delay and queue event execution and simulate real user behavior.
  52
  53         * DumpRenderTree/efl/EventSender.cpp:
  54         (DelayedEvent):
  55         (MouseEventInfo::MouseEventInfo):
  56         (MouseEventInfo):
  57         (sendClick):
  58         (mouseDownCallback):
  59         (mouseUpCallback):
  60         (mouseMoveToCallback):
  61         (leapForwardCallback):
  62         (mouseScrollByCallback):
  63         (feedOrQueueMouseEvent):
  64         (feedMouseEvent):
  65         (feedQueuedMouseEvents):
  66
  67 2012-05-08  Ryuan Choi  <ryuan.choi@samsung.com>
  68
  69         [EFL][DRT] WebKit/DRT should not paint webview for the pixel tests.
  70         https://bugs.webkit.org/show_bug.cgi?id=79853
  71
  72         Reviewed by Eric Seidel.
  73
  74         Screen buffer can be different from rendered results.
  75         For example, WebKit/Efl just moves tiles for ewk_view_tiled when scrolling
  76         contents.
  77         It means that pixel tests can't check regression of ewk_view_tiled.
  78
  79         This patch changes a way to provide bitmap from painting webview to capturing
  80         evas.
  81
  82         * DumpRenderTree/efl/PixelDumpSupportEfl.cpp:
  83         (createBitmapContextFromWebView):
  84
  85 2012-05-08  Sudarsana Nagineni  <sudarsana.nagineni@linux.intel.com>
  86
  87         [GTK] Memory leak in DRT EventSender.cpp
  88         https://bugs.webkit.org/show_bug.cgi?id=85886
  89
  90         Reviewed by Martin Robinson.
  91
  92         Fix a memory leak in gdkModifersFromJSValue() by using JSRetainPtr for
  93         JSString created with JSStringCreateWithUTF8CString.
  94
  95         * DumpRenderTree/gtk/EventSender.cpp:
  96         (gdkModifersFromJSValue):
  97
  98 2012-05-08  Christophe Dumez  <christophe.dumez@intel.com>
  99
 100         [EFL] glib dependency needs to be updated in jhbuild
 101         https://bugs.webkit.org/show_bug.cgi?id=85878
 102
 103         Reviewed by Martin Robinson.
 104
 105         Bump glib version from 2.31.8 to 2.32.2 in jhbuild.modules. This
 106         addresses flakiness in media tests due to possible deadlock with the
 107         following output on stderr:
 108         GStreamer-WARNING **: wrong STREAM_LOCK count 0
 109
 110         The bug was fixed upstream by:
 111         http://git.gnome.org/browse/glib/commit/?id=265f265c677bc5399dbb1cfa449cb4b915a6459f
 112
 113         * efl/jhbuild.modules:
 114
 115 2012-05-07  Alexis Menard  <alexis.menard@openbossa.org>
 116
 117         [Qt] Unbreak debugging of WebKit2.
 118         https://bugs.webkit.org/show_bug.cgi?id=85839
 119
 120         Reviewed by Simon Hausmann.
 121
 122         When you attach GDB to a running process, it stops it.
 123         http://trac.webkit.org/changeset/115958 introduced a pause()
 124         call to wait the debugger to be attached to then continue
 125         the execution of the WebProcess. Unfortunately the pause()
 126         function does not return unless a signal handler is called.
 127         This patch introduce an event handler to exit from the paused
 128         state when the debugger send the signal SIGCONT. The old code
 129         works with older version of GDB (<7.0) but not with newer
 130         versions where the behavior of pause() is correct.
 131
 132         * WebKitTestRunner/qt/main.cpp:
 133         (sigcontHandler):
 134         (main):
 135
 136 2012-05-08  Christophe Dumez  <christophe.dumez@intel.com>
 137
 138         [EFL][DRT] EFL's DRT does not fully support page visibility
 139         https://bugs.webkit.org/show_bug.cgi?id=85032
 140
 141         Reviewed by Antonio Gomes.
 142
 143         * DumpRenderTree/LayoutTestController.cpp:
 144         (resetPageVisibilityCallback): Fix incorrect argumentCount check.
 145         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
 146         (DumpRenderTreeChrome::resetDefaultsToConsistentValues): Properly
 147         reset page visibility setting between tests.
 148         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp:
 149         (LayoutTestController::setPageVisibility): Add support for "preview"
 150         visibility state.
 151
 152 2012-05-08  Gavin Peters  <gavinp@chromium.org>
 153
 154         [webkit-patch] Return meaningful results from single-result queries.
 155         https://bugs.webkit.org/show_bug.cgi?id=85367
 156
 157         Reviewed by Dirk Pranke.
 158
 159         Sometimes a quicksearch returns a single result by just redirecting to the page for that single
 160         bug. This patch hacks around that by noticing the redirect target is a single page, and constructing
 161         a result list consisting of that one bug.
 162
 163         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
 164         (BugzillaQueries._fetch_bugs_from_advanced_query):
 165
 166 2012-05-08  Gavin Peters  <gavinp@chromium.org>
 167
 168         [webkit-patch] Add new chrome-channels command to track down current chrome release channels for a committed bug.
 169         https://bugs.webkit.org/show_bug.cgi?id=85368
 170
 171         Reviewed by Eric Seidel.
 172
 173         * Scripts/webkitpy/common/config/urls.py:
 174         * Scripts/webkitpy/common/net/bugzilla/bugzilla_mock.py:
 175         (MockBugzillaQueries.fetch_bugs_matching_quicksearch):
 176         * Scripts/webkitpy/common/net/omahaproxy.py: Added.
 177         (OmahaProxy):
 178         (OmahaProxy.__init__):
 179         (OmahaProxy.set_url):
 180         (OmahaProxy._json_url):
 181         (OmahaProxy._get_json):
 182         (OmahaProxy.get_revisions):
 183         * Scripts/webkitpy/common/net/omahaproxy_unittest.py: Added.
 184         (MockOmahaProxy):
 185         (MockOmahaProxy.__init__):
 186         (OmahaProxyTest):
 187         (OmahaProxyTest.test_get_revisions):
 188         * Scripts/webkitpy/tool/commands/__init__.py:
 189         * Scripts/webkitpy/tool/commands/bugsearch.py:
 190         (BugSearch):
 191         * Scripts/webkitpy/tool/commands/chromechannels.py: Added.
 192         (ChromeChannels):
 193         (__init__):
 194         (execute):
 195         * Scripts/webkitpy/tool/commands/chromechannels_unittest.py: Added.
 196         (MockOmahaProxy):
 197         (MockOmahaProxy.get_revisions):
 198         (TestableChromeChannels):
 199         (TestableChromeChannels.__init__):
 200         (ChromeChannelsTest):
 201         (ChromeChannelsTest.test_single_bug):
 202         (ChromeChannelsTest.test_with_query):
 203
 204 2012-05-08  Carlos Garcia Campos  <cgarcia@igalia.com>
 205
 206         Unreviewed. Fix GTK+ unit tests.
 207
 208         * gtk/common.py:
 209         (pkg_config_file_variable):
 210
 211 2012-05-08  Carlos Garcia Campos  <cgarcia@igalia.com>
 212
 213         [GTK] Google tests that time out are leaked
 214         https://bugs.webkit.org/show_bug.cgi?id=84973
 215
 216         Reviewed by Philippe Normand.
 217
 218         Move Tools/Scripts/run-gtk-tests to Tools/gtk/run-api-tests
 219         leaving run-gtk-tests as simple wrapper that runs run-api-tests in
 220         a jhbuild environment when jhbuild is available.
 221
 222         * Scripts/run-gtk-tests: Run Tools/gtk/run-api-tests directly or
 223         with jhbuild if present.
 224         * gtk/common.py:
 225         (pkg_config_file_variable): Generic function to get a variable
 226         from a pkg-config file.
 227         (prefix_of_pkg_config_file): Use pkg_config_file_variable with 'prefix'.
 228         * gtk/run-api-tests: Copied from Tools/Scripts/run-gtk-tests.
 229         (TestRunner.__init__): Use common.build_path() to get the programs path.
 230         (TestRunner._lookup_atspi2_binary): Use
 231         common.pkg_config_file_variable() to get the 'exec_prefix'.
 232         (TestRunner._start_accessibility_daemons): Use subprocess.Popen()
 233         directly instead of helper _create_process that has been removed.
 234         (TestRunner._setup_testing_environment): Use subprocess.Popen()
 235         directly instead of helper _create_process that has been
 236         removed. Use common.top_level_path() and common.build_path()
 237         instead of custom methods that have been removed too.
 238         (TestRunner._run_test_command): Use subprocess.Popen() directly
 239         instead of helper _create_process that has been removed.
 240
 241 2012-05-08  Thiago Marcos P. Santos  <thiago.santos@intel.com>
 242
 243         [EFL] [DRT] Clear extra views before setting a blank page
 244         https://bugs.webkit.org/show_bug.cgi?id=85608
 245
 246         Reviewed by Kenneth Rohde Christiansen.
 247
 248         We need to clear the extra views before setting a blank page,
 249         otherwise it will cause more events interactions and sometimes
 250         trigger a unload(). This was causing frame-crash-with-page-cache.html
 251         to leak events to the next test causing any subsequent test to
 252         fail.
 253
 254         * DumpRenderTree/efl/DumpRenderTree.cpp:
 255         (runTest):
 256         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
 257         (DumpRenderTreeChrome::clearExtraViews):
 258         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
 259         * DumpRenderTree/efl/DumpRenderTreeChrome.h:
 260         (DumpRenderTreeChrome):
 261
 262 2012-05-08  Balazs Kelemen  <kbalazs@webkit.org>
 263
 264         [Qt] X11 plugins need to be reworked for Qt5+WK1
 265         https://bugs.webkit.org/show_bug.cgi?id=80691
 266
 267         Reviewed by Simon Hausmann.
 268
 269         Implement basic windowless plugin support with Qt5.
 270
 271         * qmake/mkspecs/features/features.prf:
 272         Enable NPAPI plugins if X11 libraries are available.
 273         * qmake/mkspecs/features/functions.prf:
 274         Added a convenience function to determine availability
 275         of X11 libraries.
 276
 277 2012-05-07  Raphael Kubo da Costa  <rakuco@webkit.org>
 278
 279         [webkitpy] Remove Python 2.5-specific workaround from http_server_base.py.
 280         https://bugs.webkit.org/show_bug.cgi?id=85852
 281
 282         Reviewed by Dirk Pranke.
 283
 284         * Scripts/webkitpy/layout_tests/servers/http_server_base.py:
 285         (HttpServerBase._is_server_running_on_all_ports): Revert the part
 286         of r89412 which was a workaround for an issue with Python 2.5 on
 287         Mac 10.5.
 288
 289 2012-05-07  Raphael Kubo da Costa  <rakuco@FreeBSD.org>
 290
 291         webkitdirs.pm: Make determineNumberOfCPUs work on FreeBSD.
 292         https://bugs.webkit.org/show_bug.cgi?id=81481
 293
 294         Reviewed by Dirk Pranke.
 295
 296         FreeBSD has the same sysctl interface present in Darwin to query
 297         the number of CPUs available, so use that too.
 298
 299         * Scripts/webkitdirs.pm:
 300         (determineNumberOfCPUs): Share the sysctl call with Darwin.
 301         (isFreeBSD): Added.
 302
 303 2012-05-07  Alexis Menard  <alexis.menard@openbossa.org>
 304
 305         Unreviewed Qt build fix in Debug.
 306
 307         We need some system includes here for getpid() to be recognized.
 308
 309         * WebKitTestRunner/qt/main.cpp:
 310
 311 2012-05-07  Ojan Vafai  <ojan@chromium.org>
 312
 313         When embedding the flakiness dashboard, hide popups when the frame is blurred.
 314         https://bugs.webkit.org/show_bug.cgi?id=85684
 315
 316         Reviewed by Dirk Pranke.
 317
 318         If you show a popup, then click elsewhere in garden-o-matic, it will now
 319         properly hide it.
 320
 321         * TestResultServer/static-dashboards/dashboard_base.js:
 322         (showPopup):
 323         * TestResultServer/static-dashboards/flakiness_dashboard.js:
 324         (showPopupForBuild):
 325         (generatePageForIndividualTests.if):
 326         (generatePageForIndividualTests):
 327         * TestResultServer/static-dashboards/flakiness_dashboard_embedded_unittests.js: Added.
 328         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
 329         * TestResultServer/static-dashboards/run-embedded-unittests.html: Added.
 330
 331 2012-05-07  Ian Vollick  <vollick@chromium.org>
 332
 333         [chromium] WebViewHost should setCompositorSurfaceReady
 334         https://bugs.webkit.org/show_bug.cgi?id=85418
 335
 336         Reviewed by James Robinson.
 337
 338         * DumpRenderTree/chromium/WebViewHost.cpp:
 339         (WebViewHost::setWebWidget):
 340
 341 2012-05-07  Jon Lee  <jonlee@apple.com>
 342
 343         WebKitTestRunner needs layoutTestController.dumpFrameLoadCallbacks
 344         https://bugs.webkit.org/show_bug.cgi?id=42331
 345         <rdar://problem/8193641>
 346
 347         Reviewed by Darin Adler.
 348
 349         This patch sets up the framework for supporting dumpFrameLoadCallbacks, as well as
 350         outputting some of the required strings. As tests become unskipped the other callbacks
 351         can be supported.
 352
 353         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl: Add dumpFrameLoadCallbacks()
 354         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
 355         (WTR::InjectedBundle::didReceiveMessage): pass in the dictionary containing keys needed to initialize
 356         dump callbacks for the layoutTestController.
 357         (WTR::InjectedBundle::booleanForKey): Helper function to extract boolean value from key in WKDictionaryRef.
 358         Output some warning if we are trying to extract a value that is not a boolean.
 359         (WTR::InjectedBundle::beginTesting): Update to take in the dictionary passed in by the TestController.
 360         We cannot set the bits on the layoutTestController until it has been created in beginTesting().
 361         * WebKitTestRunner/InjectedBundle/InjectedBundle.h:
 362         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
 363         (WTR::InjectedBundle::didReceiveMessage):
 364         (WTR):
 365         (WTR::InjectedBundle::beginTesting):
 366         * WebKitTestRunner/InjectedBundle/InjectedBundle.h:
 367         (InjectedBundle):
 368         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
 369         (WTR):
 370         (WTR::dumpFrameDescriptionSuitableForTestResult): Static function similar to
 371         -[WebFrame _drt_descriptionSuitableForTestResult] in DumpRenderTree.
 372         (WTR::InjectedBundlePage::didStartProvisionalLoadForFrame): Output string.
 373         (WTR::InjectedBundlePage::didCommitLoadForFrame): Output string.
 374         (WTR::InjectedBundlePage::didFinishLoadForFrame): Output string.
 375         (WTR::InjectedBundlePage::didReceiveTitleForFrame): Output string.
 376         (WTR::InjectedBundlePage::didCancelClientRedirectForFrame): Output string.
 377         (WTR::InjectedBundlePage::willPerformClientRedirectForFrame): Output string. In WK1 it's
 378         "willPerformClientRedirectToURL" so we will have to use that value here.
 379         (WTR::InjectedBundlePage::didFinishDocumentLoadForFrame): Output string.
 380         (WTR::InjectedBundlePage::didHandleOnloadEventsForFrame): Output string.
 381         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
 382         (WTR::LayoutTestController::LayoutTestController): Initialize to not dump frame load callbacks.
 383         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
 384         (WTR::LayoutTestController::dumpFrameLoadCallbacks): Change to use setShouldDumpFrameLoadCallbacks.
 385         (WTR::LayoutTestController::setShouldDumpFrameLoadCallbacks): Set bit.
 386         (WTR::LayoutTestController::shouldDumpFrameLoadCallbacks): Return bit.
 387         (LayoutTestController): Added m_dumpFrameLoadCallbacks to track whether to dump the output.
 388         * WebKitTestRunner/TestInvocation.cpp:
 389         (WTR::shouldLogFrameLoadDelegates): Similar to DRT.
 390         (WTR::TestInvocation::invoke): Based on the URL loaded, set the bit in the dictionary that gets passed
 391         to the injected bundle.
 392
 393 2012-05-07  Tommy Widenflycht  <tommyw@google.com>
 394
 395         [chromium] MediaStream API: Enhance WebUserMediaClientMock and add a test for it
 396         https://bugs.webkit.org/show_bug.cgi?id=85618
 397
 398         Reviewed by Adam Barth.
 399
 400         Adding the capability to control which MediaStreamTracks actually get created,
 401         and adding a test that exercises the JS api using that capability.
 402
 403         * DumpRenderTree/chromium/WebUserMediaClientMock.cpp:
 404         (WebKit::WebUserMediaClientMock::requestUserMedia):
 405         * DumpRenderTree/chromium/WebUserMediaClientMock.h:
 406         * DumpRenderTree/chromium/WebViewHost.cpp:
 407         (WebViewHost::testMediaStreamClient):
 408         * DumpRenderTree/chromium/WebViewHost.h:
 409         (WebViewHost):
 410
 411 2012-05-07  Zan Dobersek  <zandobersek@gmail.com>
 412
 413         [Gtk] WebGL feature is not built anymore through build-webkit after r116251
 414         https://bugs.webkit.org/show_bug.cgi?id=85788
 415
 416         Reviewed by Martin Robinson.
 417
 418         Enable WebGL by default for the Gtk port. The feature is already enabled by
 419         default in configure.ac, so there's no reason it shouldn't be enabled when
 420         building through build-webkit.
 421
 422         * Scripts/webkitperl/FeatureList.pm:
 423
 424 2012-05-07  Csaba Osztrogonác  <ossy@webkit.org>
 425
 426         Unittest fix after r116288, because now there are more than 5 hit for the 'Eric' word.
 427         We have only 2 'Gavin', so it is a better choice than 'Eric' for unit testing.
 428
 429         Reviewed by Tor Arne Vestbø.
 430
 431         * Scripts/webkitpy/tool/bot/irc_command_unittest.py:
 432         (IRCCommandTest.test_whois):
 433
 434 2012-05-07  Sheriff Bot  <webkit.review.bot@gmail.com>
 435
 436         Unreviewed, rolling out r116299, r116301, and r116303.
 437         http://trac.webkit.org/changeset/116299
 438         http://trac.webkit.org/changeset/116301
 439         http://trac.webkit.org/changeset/116303
 440         https://bugs.webkit.org/show_bug.cgi?id=85795
 441
 442         Build is still broken (Requested by Ossy on #webkit).
 443
 444         * DumpRenderTree/qt/DumpRenderTree.pro:
 445         * DumpRenderTree/qt/main.cpp:
 446         (main):
 447         * QtTestPlatformPlugin/QtTestPlatformPlugin.pro: Removed.
 448         * QtTestPlatformPlugin/TestIntegration.cpp: Removed.
 449         * QtTestPlatformPlugin/TestIntegration.h: Removed.
 450         * QtTestPlatformPlugin/mac/TestFontDatabase.h: Removed.
 451         * QtTestPlatformPlugin/mac/TestFontDatabase.mm: Removed.
 452         * QtTestPlatformPlugin/mac/TestIntegrationMac.mm: Removed.
 453         * QtTestPlatformPlugin/main.cpp: Removed.
 454         * QtTestPlatformPlugin/testplatform.json: Removed.
 455         * Tools.pro:
 456         * WebKitTestRunner/Target.pri:
 457         * WebKitTestRunner/qt/main.cpp:
 458         (main):
 459
 460 2012-05-07  Balazs Kelemen  <kbalazs@webkit.org>
 461
 462         Unreviewed Qt build fix.
 463
 464         One more try to learn how to use
 465         const with a pointer array.
 466
 467         * DumpRenderTree/qt/main.cpp:
 468         (initializeTestPlatformPlugin):
 469
 470 2012-05-07  Balazs Kelemen  <kbalazs@webkit.org>
 471
 472         Unreviewed build fix after r116299.
 473
 474         * DumpRenderTree/qt/main.cpp:
 475         (initializeTestPlatformPlugin):
 476
 477 2012-05-07  Balazs Kelemen  <kbalazs@webkit.org>
 478
 479         [Qt] Add test specific platform plugin to achieve unified layout test results
 480         https://bugs.webkit.org/show_bug.cgi?id=80996
 481
 482         Reviewed by Simon Hausmann.
 483
 484         Added QtTestPlatformPlugin as a new project under Tools.
 485         This is a Qt5-ish platform plugin that can be used to tweak the
 486         platform support interfaces in order to unify layout test results.
 487         For now it only overrides the font database on Mac and redirects
 488         everything else to the real platform plugin. The font database it
 489         provides mimics the way how we set up test fonts with fontconfig on Linux.
 490         Make DumpRenderTree and WebKitTestRunner use this platform plugin.
 491
 492         * DumpRenderTree/qt/DumpRenderTree.pro:
 493         * DumpRenderTree/qt/main.cpp:
 494         (initializeTestPlatformPlugin):
 495         (main):
 496         * QtTestPlatformPlugin/QtTestPlatformPlugin.pro: Added.
 497         * QtTestPlatformPlugin/TestIntegration.cpp: Added.
 498         (TestIntegration::TestIntegration):
 499         (TestIntegration::fontDatabase):
 500         * QtTestPlatformPlugin/TestIntegration.h: Added.
 501         (TestIntegration):
 502         (TestIntegration::hasCapability):
 503         (TestIntegration::createPlatformPixmap):
 504         (TestIntegration::createPlatformWindow):
 505         (TestIntegration::createPlatformBackingStore):
 506         (TestIntegration::createPlatformOpenGLContext):
 507         (TestIntegration::createPlatformSharedGraphicsCache):
 508         (TestIntegration::guiThreadEventDispatcher):
 509         (TestIntegration::clipboard):
 510         (TestIntegration::drag):
 511         (TestIntegration::inputContext):
 512         (TestIntegration::accessibility):
 513         (TestIntegration::nativeInterface):
 514         (TestIntegration::services):
 515         (TestIntegration::styleHint):
 516         (TestIntegration::platformTheme):
 517         * QtTestPlatformPlugin/mac/TestFontDatabase.h: Added.
 518         (TestFontDatabase):
 519         * QtTestPlatformPlugin/mac/TestFontDatabase.mm: Added.
 520         (TestFontDatabase::populateFontDatabase):
 521         * QtTestPlatformPlugin/mac/TestIntegrationMac.mm: Added.
 522         (TestIntegration::fontDatabase):
 523         * QtTestPlatformPlugin/main.cpp: Added.
 524         (TestIntegrationPlugin::keys):
 525         (TestIntegrationPlugin::create):
 526         (TestIntegrationPlugin::initialize):
 527         * QtTestPlatformPlugin/testplatform.json: Added.
 528         * Tools.pro:
 529         * WebKitTestRunner/Target.pri:
 530         * WebKitTestRunner/qt/main.cpp:
 531         (main):
 532
 533 2012-05-07  Christophe Dumez  <christophe.dumez@intel.com>
 534
 535         [EFL] media/video-poster-blocked-by-willsendrequest.html is flaky
 536         https://bugs.webkit.org/show_bug.cgi?id=85609
 537
 538         Reviewed by Simon Hausmann.
 539
 540         Flakiness was caused by the fact that
 541         media/crash-closing-page-with-media-as-plugin-fallback.html and
 542         media/video-poster-blocked-by-willsendrequest.html use the same PNG
 543         file as poster attribute in the video tag. The PNG file gets cached
 544         when the first test is executed and this somehow causes the second
 545         test to fail.
 546
 547         The flakiness was addressed by clearing the memory and page caches
 548         between tests.
 549
 550         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
 551         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
 552
 553 2012-05-07  Eric Seidel  <eric@webkit.org>
 554
 555         Add values for all features to Qt's features.pri
 556         https://bugs.webkit.org/show_bug.cgi?id=85746
 557
 558         Reviewed by Tor Arne Vestbø.
 559
 560         When generate-feature-files generates features.pri it provides a value
 561         for every possible ENABLE_.  In prepration for landing generate-feature-files
 562         it seems appropriate to add a value for all possible ENABLEs to the existing
 563         features.pri.  This lets me make sure that I'm getting them all right!
 564
 565         * qmake/mkspecs/features/features.pri:
 566
 567 2012-05-07  Adam Bergkvist  <adam.bergkvist@ericsson.com>
 568
 569         Added myself as a committer.
 570
 571         Not reviewed.
 572
 573         * Scripts/webkitpy/common/config/committers.py:
 574
 575 2012-05-07  Sudarsana Nagineni  <sudarsana.nagineni@linux.intel.com>
 576
 577         [EFL][DRT] Implement LayoutTestController::execCommand
 578         https://bugs.webkit.org/show_bug.cgi?id=82286
 579
 580         Reviewed by Antonio Gomes.
 581
 582         Add missing implementation execCommand and isCommandEnabled to EFL's
 583         LayoutTestController.
 584
 585         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp:
 586         (LayoutTestController::execCommand):
 587         (LayoutTestController::isCommandEnabled):
 588
 589 2012-05-06  Hayato Ito  <hayato@chromium.org>
 590
 591         [prepare-ChangeLog] Excludes a reference file used by reftests from test files.
 592         https://bugs.webkit.org/show_bug.cgi?id=74268
 593
 594         Reviewed by Ryosuke Niwa.
 595
 596         * Scripts/prepare-ChangeLog:
 597         (generateFileList):
 598
 599 2012-05-06  Mary Wu  <mary.wu@torchmobile.com.cn>
 600
 601         Add a contributor to committers.py
 602         https://bugs.webkit.org/show_bug.cgi?id=85761
 603
 604         Reviewed by Unreviewed.
 605
 606         Add myself to committers.py contributor.
 607
 608         * Scripts/webkitpy/common/config/committers.py:
 609
 610 2012-05-06  Dan Bernstein  <mitz@apple.com>
 611
 612         Tools part of: Building and debugging WebKit in the Xcode IDE requires a lot of setup
 613         https://bugs.webkit.org/show_bug.cgi?id=85739
 614
 615         Reviewed by Daniel Bates.
 616
 617         * Scripts/build-webkit: Moved code that copies from WebKitLibraries to the product directory
 618         from here...
 619         * Scripts/copy-webkitlibraries-to-product-directory: ...to this new script.
 620
 621 2012-05-06  Andy Estes  <aestes@apple.com>
 622
 623         Fix the build after r116246.
 624
 625         * DumpRenderTree/LayoutTestController.cpp: Remove an unused function.
 626
 627 2012-05-06  Eric Seidel  <eric@webkit.org>
 628
 629         Update several build-webkit options to match the #defines they toggle
 630         https://bugs.webkit.org/show_bug.cgi?id=85744
 631
 632         Reviewed by Adam Barth.
 633
 634         Most notable here is fixing --3d-canvas to be --webgl these days --
 635         the define was renamed 15 months ago!
 636
 637         * Scripts/webkitperl/FeatureList.pm:
 638
 639 2012-05-06  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
 640
 641         [EFL] EFL's LayoutTestController setJavaScriptCanAccessClipboard implementation
 642         https://bugs.webkit.org/show_bug.cgi?id=83687
 643
 644         Reviewed by Antonio Gomes.
 645
 646         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp:
 647         (LayoutTestController::setJavaScriptCanAccessClipboard):
 648
 649 2012-05-06  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
 650
 651         Convert isPageBoxVisible to use Internals interface.
 652         https://bugs.webkit.org/show_bug.cgi?id=85692
 653
 654         Reviewed by Darin Adler.
 655
 656         Remove isPageBoxVisible functions, because it is able to work in the
 657         cross-port way through the Internals interface.
 658
 659         * DumpRenderTree/LayoutTestController.cpp:
 660         (LayoutTestController::staticFunctions):
 661         * DumpRenderTree/LayoutTestController.h:
 662         (LayoutTestController):
 663         * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp:
 664         * DumpRenderTree/chromium/LayoutTestController.cpp:
 665         (LayoutTestController::LayoutTestController):
 666         * DumpRenderTree/chromium/LayoutTestController.h:
 667         (LayoutTestController):
 668         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp:
 669         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
 670         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
 671         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
 672         * DumpRenderTree/qt/LayoutTestControllerQt.h:
 673         (LayoutTestController):
 674         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
 675         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
 676
 677 2012-05-06  Tom Zakrajsek  <tomz@codeaurora.org>
 678
 679         sheriffbot isn't reopening patches after it lands rollouts
 680         https://bugs.webkit.org/show_bug.cgi?id=64418
 681
 682         Reviewed by Adam Barth.
 683
 684         * Scripts/webkitpy/tool/commands/download_unittest.py:
 685         * Scripts/webkitpy/tool/steps/createbug.py:
 686         (CreateBug.run):
 687
 688 2012-05-04  Raphael Kubo da Costa  <rakuco@webkit.org>
 689
 690         webkitpy: Use os.pathsep instead of manually finding out the path separator in NRWT.
 691         https://bugs.webkit.org/show_bug.cgi?id=85697
 692
 693         Reviewed by Eric Seidel.
 694
 695         * Scripts/new-run-webkit-tests: Instead of manually detecting whether
 696         the path separator should be ':' or ';' by checking sys.platform, rely
 697         on os.pathsep which has exactly the same information (os.pathsep is
 698         ':' for Cygwin, from what I could check in Python's sources).
 699
 700 2012-05-05  Sheriff Bot  <webkit.review.bot@gmail.com>
 701
 702         Unreviewed, rolling out r116223.
 703         http://trac.webkit.org/changeset/116223
 704         https://bugs.webkit.org/show_bug.cgi?id=85718
 705
 706         The bots need to set XDG_DATA_DIRS (Requested by rakuco on
 707         #webkit).
 708
 709         * efl/jhbuildrc:
 710         * gtk/jhbuildrc:
 711
 712 2012-05-05  Raphael Kubo da Costa  <rakuco@webkit.org>
 713
 714         [jhbuild] Do not set XDG_DATA_DIRS in jhbuildrc.
 715         https://bugs.webkit.org/show_bug.cgi?id=85717
 716
 717         Reviewed by Martin Robinson.
 718
 719         After r116209, the XDG_DATA_DIRS environment variable is now
 720         passed by webkitpy's layout_tests code to run-with-jhbuild, so
 721         instead of setting a few defaults in jhbuildrc we can now rely on
 722         the values set and really used by each system.
 723
 724         * efl/jhbuildrc:
 725         * gtk/jhbuildrc:
 726
 727 2012-05-05  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
 728
 729         [Qt] Buildfix for --minimal build after r116107.
 730
 731         Reviewed by Csaba Osztrogonác.
 732
 733         * qmake/mkspecs/features/features.prf:
 734
 735 2012-05-05  Zan Dobersek  <zandobersek@gmail.com>
 736
 737         [Gtk] GtkDriver is not required anymore after r116134
 738         https://bugs.webkit.org/show_bug.cgi?id=85699
 739
 740         Reviewed by Csaba Osztrogonác.
 741
 742         Remove GtkDriver as it was replaced by the XvfbDriver and is
 743         no longer needed.
 744
 745         * Scripts/webkitpy/layout_tests/port/gtk.py:
 746
 747 2012-05-05  Zan Dobersek  <zandobersek@gmail.com>
 748
 749         Increase the display ID for the Xvfb process to use if pixel_tests argument
 750         is true rather than if pixel_tests option is present. This fixes the
 751         XvfbDriver in the same way r115825 fixed the now-redundant GtkDriver.
 752
 753         Rubber-stamped by Csaba Osztrogonác.
 754
 755         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py:
 756         (XvfbDriver._start):
 757
 758 2012-05-04  Raphael Kubo da Costa  <rakuco@webkit.org>
 759
 760         webkitpy: Preserve XDG_DATA_DIRS in the environment passed to ServerProcess.
 761         https://bugs.webkit.org/show_bug.cgi?id=85694
 762
 763         Reviewed by Martin Robinson.
 764
 765         This is sort of a follow-up to r100674: $XDG_DATA_DIRS should be
 766         preserved, as it is particularly useful for Linux/Unix
 767         environments.  As a real use case, if shared-mime-info is
 768         installed into a non-trivial prefix, Glib and other libraries will
 769         have trouble to properly identify mimetypes (in ResourceResponses,
 770         for example) as the base freedesktop.org mimetype file will not be
 771         found.
 772
 773         * Scripts/webkitpy/layout_tests/port/base.py:
 774         (Port.to.setup_environ_for_server):
 775
 776 2012-05-04  Gustavo Noronha Silva  <gns@gnome.org>
 777
 778         [GTK] Simplify how libWebCoreModules is linked in, and fix WebKit2 build
 779         https://bugs.webkit.org/show_bug.cgi?id=85691
 780
 781         * GNUmakefile.am: no longer link libWebCoreModules to DumpRenderTree.
 782
 783 2012-05-04  Eric Seidel  <eric@webkit.org>
 784
 785         Make capitalization and descriptions in FeatureList.pm more consistent to make easier to autogenerate
 786         https://bugs.webkit.org/show_bug.cgi?id=85583
 787
 788         Reviewed by Daniel Bates.
 789
 790         * Scripts/webkitperl/FeatureList.pm:
 791
 792 2012-05-04  Eric Seidel  <eric@webkit.org>
 793
 794         Sort entries in FeaturesList.pm to make them easier to autogenerate
 795         https://bugs.webkit.org/show_bug.cgi?id=85584
 796
 797         Reviewed by Daniel Bates.
 798
 799         No functional changes, only sorting.
 800
 801         * Scripts/webkitperl/FeatureList.pm:
 802
 803 2012-05-04  Andy Estes  <aestes@apple.com>
 804
 805         REGRESSION (r115752): WebKit2.NewFirstVisuallyNonEmptyLayout API test times out
 806         https://bugs.webkit.org/show_bug.cgi?id=85675
 807
 808         * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayout.cpp:
 809         (TestWebKitAPI::TEST): Disable the test.
 810
 811 2012-04-30  Jer Noble  <jer.noble@apple.com>
 812
 813         Full screen will exit during a provisional load of a non-ancestor iframe.
 814         https://bugs.webkit.org/show_bug.cgi?id=85230
 815
 816         Reviewed by Maciej Stachowiak .
 817
 818         Add support for the new UIDelegate webView:closeFullScreenWithListener: method:
 819         * DumpRenderTree/mac/UIDelegate.mm:
 820         (-[UIDelegate webView:closeFullScreenWithListener:]):
 821
 822         Add support for the new closeFullScreen injected bundle method:
 823         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
 824         (WTR::InjectedBundlePage::InjectedBundlePage):
 825         (WTR::InjectedBundlePage::closeFullScreen):
 826         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
 827
 828 2012-05-04  Ojan Vafai  <ojan@chromium.org>
 829
 830         Fix test naming to be camel-case like the methods they are testing.
 831
 832         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
 833         (test):
 834
 835 2012-05-04  Ojan Vafai  <ojan@chromium.org>
 836
 837         Show the results links in the embedded flakiness dashboard
 838         https://bugs.webkit.org/show_bug.cgi?id=85660
 839
 840         Reviewed by Adam Barth.
 841
 842         Turns out these links don't add much visual noise and it's annoying
 843         to have to go to the dashboard from garden-o-matic just to view the
 844         results for a test.
 845
 846         This is useful for cases where garden-o-matic doesn't properly show the
 847         results (e.g. missing tests). If we address all those cases, then
 848         we could easily hide the results links again if we wanted to.
 849
 850         * TestResultServer/static-dashboards/flakiness_dashboard.js:
 851         (showUpdateInfoForTest):
 852         (htmlForIndividualTestOnAllBuildersWithResultsLinks):
 853         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
 854         (test):
 855
 856 2012-05-04  Dirk Pranke  <dpranke@chromium.org>
 857
 858         Revert to using chromium --test-shell mode on SL in NRWT
 859
 860         https://bugs.webkit.org/show_bug.cgi?id=83076
 861
 862         Unreviewed, build fix.
 863
 864         We seem to be seeing a lot more flakiness on the bot since
 865         I switched NRWT to "DRT" mode. We'll try reverting back and
 866         see if this helps things.
 867
 868         * Scripts/webkitpy/layout_tests/port/chromium.py:
 869         (ChromiumDriver.__init__):
 870
 871 2012-05-04  Christophe Dumez  <christophe.dumez@intel.com>
 872
 873         [EFL] Implement layoutTestController.dumpResourceResponseMIMETypes
 874         https://bugs.webkit.org/show_bug.cgi?id=84941
 875
 876         Reviewed by Martin Robinson.
 877
 878         Print MIME type detection messages in EFL's DumpRenderTree if
 879         LayoutTestController's dumpResourceResponseMIMETypes() returns true.
 880
 881         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
 882         (DumpRenderTreeChrome::onResponseReceived):
 883
 884 2012-05-04  Sudarsana Nagineni  <sudarsana.nagineni@linux.intel.com>
 885
 886         [GTK] DRT needs an implementation of LayoutTestController::setDefersLoading and ::goBack
 887         https://bugs.webkit.org/show_bug.cgi?id=85134
 888
 889         Reviewed by Martin Robinson.
 890
 891         Add missing implementation setDefersLoading and goBack to GTK's
 892         LayoutTestController so that the test case rely on these methods
 893         can be enabled.
 894
 895         * DumpRenderTree/gtk/DumpRenderTree.cpp:
 896         (resetDefaultsToConsistentValues):
 897         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
 898         (LayoutTestController::goBack):
 899         (LayoutTestController::setDefersLoading):
 900
 901 2012-05-04  Kristóf Kosztyó  <kkristof@inf.u-szeged.hu>
 902
 903         [Qt][NRWT] Run each DRT in it's own xvfb
 904         https://bugs.webkit.org/show_bug.cgi?id=77335
 905
 906         Reviewed by Dirk Pranke.
 907
 908         * Scripts/webkitpy/layout_tests/port/gtk.py:
 909         (GtkPort._driver_class):
 910         * Scripts/webkitpy/layout_tests/port/qt.py:
 911         (QtPort._driver_class):
 912         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py: Added.
 913         (XvfbDriver):
 914         (XvfbDriver._start):
 915         (XvfbDriver._start.x_filter):
 916         (XvfbDriver.stop):
 917
 918 2012-05-04  Zan Dobersek  <zandobersek@gmail.com>
 919
 920         Use destructuring assingment for the return values of the
 921         port.diff_image() method call.
 922
 923         Rubber-stamped by Dirk Pranke.
 924
 925         * Scripts/webkitpy/layout_tests/controllers/test_result_writer.py:
 926         (write_test_result):
 927
 928 2012-05-03  Ojan Vafai  <ojan@chromium.org>
 929
 930         Store fewer runs by default in the flakiness dashboard
 931         https://bugs.webkit.org/show_bug.cgi?id=85568
 932
 933         Reviewed by Dirk Pranke.
 934
 935         This will greatly shrink the size of the json files we load,
 936         resulting in a much faster initial load. People rarely care
 937         about more than 100 runs. For the cases they do, they can still click
 938         the show all runs checkbox to show 500 runs.
 939
 940         * TestResultServer/model/jsonresults.py:
 941
 942 2012-05-04  Christophe Dumez  <christophe.dumez@intel.com>
 943
 944         [EFL] Frame load callbacks output is missing for redirections
 945         https://bugs.webkit.org/show_bug.cgi?id=85173
 946
 947         Reviewed by Antonio Gomes.
 948
 949         Catch the new ewk_frame signals to print the expected output in case
 950         of redirections, when LayoutTestController's dumpFrameLoadCallbacks()
 951         returns true.
 952
 953         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
 954         (DumpRenderTreeChrome::createView):
 955         (DumpRenderTreeChrome::onWillSendRequest):
 956         (DumpRenderTreeChrome::onFrameCreated):
 957         (DumpRenderTreeChrome::onFrameRedirectForProvisionalLoad):
 958         (DumpRenderTreeChrome::onFrameRedirectRequested):
 959         (DumpRenderTreeChrome::onResponseReceived):
 960         * DumpRenderTree/efl/DumpRenderTreeChrome.h:
 961         (DumpRenderTreeChrome):
 962
 963 2012-05-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
 964
 965         [Qt] Fix up warning about missing GLib/Gio/GStreamer
 966
 967         We don't actually fall back.
 968
 969         Rubber-stamped by Csaba Osztrogonác.
 970
 971         * qmake/mkspecs/features/features.prf:
 972
 973 2012-05-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
 974
 975         [Qt] Clarify warning about missing GLib/Gio/GStreamer for media support
 976
 977         And only print it once.
 978
 979         Reviewed by Csaba Osztrogonác.
 980
 981         * qmake/mkspecs/features/features.prf:
 982
 983 2012-05-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
 984
 985         [Qt] Clean up and split features.prf into a static list of defaults
 986
 987         The static list of feature defaults is used as a fallback for any
 988         feature that's not dynamically detected or overriden on the command
 989         line (though build-webkit or passing DEFINES+= to qmake).
 990
 991         The static list is complete, which allows for auto-generation based
 992         on Features.py (see bug https://bugs.webkit.org/show_bug.cgi?id=85456)
 993
 994         https://bugs.webkit.org/show_bug.cgi?id=85611
 995
 996         Reviewed by Simon Hausmann.
 997
 998         * Scripts/build-webkit:
 999         * Tools.pro:
1000         * qmake/mkspecs/features/features.prf:
1001         * qmake/mkspecs/features/features.pri: Added.
1002
1003 2012-05-04  Sheriff Bot  <webkit.review.bot@gmail.com>
1004
1005         Unreviewed, rolling out r116085, r116091, and r116095.
1006         http://trac.webkit.org/changeset/116085
1007         http://trac.webkit.org/changeset/116091
1008         http://trac.webkit.org/changeset/116095
1009         https://bugs.webkit.org/show_bug.cgi?id=85628
1010
1011         We are not ready with dependencies on all platform yet (mac) +
1012         problems with debug builds. (Requested by Zoltan on #webkit).
1013
1014         * qmake/mkspecs/features/features.prf:
1015
1016 2012-05-04  Christophe Dumez  <christophe.dumez@intel.com>
1017
1018         [EFL] LayoutTestController does not support overriding WebKitLoadSiteIconsKey
1019         https://bugs.webkit.org/show_bug.cgi?id=85171
1020
1021         Reviewed by Andreas Kling.
1022
1023         Implement support for overriding "WebKitLoadSiteIconsKey" setting in
1024         EFL's LayoutTestController.
1025
1026         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
1027         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
1028         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp:
1029         (LayoutTestController::overridePreference):
1030
1031 2012-05-03  Alexis Menard  <alexis.menard@openbossa.org>
1032
1033         [Qt] Enable fullscreen API for WebKit2.
1034         https://bugs.webkit.org/show_bug.cgi?id=85498
1035
1036         Reviewed by Simon Hausmann.
1037
1038         Enable the fullscreen API on the MiniBrowser and turn
1039         it default on WebKit2, disable it on WK1.
1040
1041         * MiniBrowser/qt/BrowserWindow.cpp:
1042         (BrowserWindow::BrowserWindow):
1043         * MiniBrowser/qt/qml/BrowserWindow.qml:
1044         * Scripts/webkitperl/FeatureList.pm:
1045         * qmake/mkspecs/features/features.prf:
1046
1047 2012-05-04  Allan Sandfeld Jensen  <allan.jensen@nokia.com>
1048
1049         REGRESSION(r116008) old-run-webkit-tests broken
1050         https://bugs.webkit.org/show_bug.cgi?id=85597
1051
1052         Reviewed by Simon Hausmann.
1053
1054         Reinstate libraryContainsSymbol.
1055
1056         * Scripts/webkitperl/features.pm:
1057         (libraryContainsSymbol):
1058
1059 2012-05-04  Zoltan Horvath  <zoltan@webkit.org>
1060
1061         [Qt] Set WebCore imagedecoders as default and add fallback to QImageDecoder
1062         https://bugs.webkit.org/show_bug.cgi?id=80400
1063
1064         Remove WTF_USE_QT_IMAGE_DECODER macro.
1065
1066         Reviewed by Simon Hausmann.
1067
1068         * qmake/mkspecs/features/features.prf:
1069
1070 2012-05-03  Christophe Dumez  <christophe.dumez@intel.com>
1071
1072         [EFL] DRT should obey layoutTestController's addURLToRedirect()
1073         https://bugs.webkit.org/show_bug.cgi?id=82722
1074
1075         Reviewed by Maciej Stachowiak.
1076
1077         Implement support for layoutTestController's addURLToRedirect()
1078         in EFL's DRT so that it redirects properly when indicated.
1079
1080         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
1081         (DumpRenderTreeChrome::onWillSendRequest):
1082
1083 2012-05-03  Christophe Dumez  <christophe.dumez@intel.com>
1084
1085         [EFL] DRT should support LayoutTestController's willSendRequestReturnsNullOnRedirect()
1086         https://bugs.webkit.org/show_bug.cgi?id=82704
1087
1088         Reviewed by Antonio Gomes.
1089
1090         EFL's DRT needs to obey LayoutTestController's
1091         willSendRequestReturnsNullOnRedirect().
1092
1093         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
1094         (DumpRenderTreeChrome::onWillSendRequest):
1095
1096 2012-05-03  Raphael Kubo da Costa  <rakuco@webkit.org>
1097
1098         git-add-reviewer: Do not put the bug title and its URL in the same line in the commit message.
1099         https://bugs.webkit.org/show_bug.cgi?id=85570
1100
1101         Reviewed by Adam Roben.
1102
1103         The standard format of commit messages and ChangeLog entries has a
1104         short summary in the first line, followed by the bug URL in
1105         Bugzilla. Passing "%s" in the pretty format line to git-rev-list
1106         merges these two lines into a single one separated by a space
1107         character, which is not what we want.
1108
1109         Since "%s%n%n%b" is in practice "%B" (raw message), use it
1110         instead, as the first two lines do not end up being merged
1111         together.
1112
1113         * Scripts/git-add-reviewer:
1114         (writeCommitMessageToFile):
1115
1116 2012-05-03  Ojan Vafai  <ojan@chromium.org>
1117
1118         Improve UI for garden-o-matic examine view when there are fewer than 5 tests.
1119         https://bugs.webkit.org/show_bug.cgi?id=85566
1120
1121         Reviewed by Adam Barth.
1122
1123         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results.js:
1124         (.):
1125         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results_unittests.js:
1126         (.):
1127         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/styles/results.css:
1128
1129 2012-05-03  Raphael Kubo da Costa  <rakuco@webkit.org>
1130
1131         webkitpy: Recognize FreeBSD as a valid platform.
1132         https://bugs.webkit.org/show_bug.cgi?id=81467
1133
1134         Reviewed by Adam Barth.
1135
1136         Add PlatformInfo.is_freebsd() so that trying to use webkit-patch
1137         on FreeBSD does not raise an "unrecognized platform" exception.
1138
1139         Most of this patch contains some unit tests for the mentioned
1140         method, as currently platform-specific code is needed only on
1141         Windows.
1142
1143         * Scripts/webkitpy/common/system/platforminfo.py:
1144         (PlatformInfo.__init__):
1145         (PlatformInfo.is_freebsd):
1146         (PlatformInfo._determine_os_name):
1147         * Scripts/webkitpy/common/system/platforminfo_mock.py:
1148         (MockPlatformInfo.is_freebsd):
1149         * Scripts/webkitpy/common/system/platforminfo_unittest.py:
1150         (fake_platform):
1151         (fake_platform.FakePlatformModule.release):
1152         (TestPlatformInfo.test_real_code):
1153         (TestPlatformInfo.test_os_name_and_wrappers):
1154         (TestPlatformInfo.test_os_version):
1155         (TestPlatformInfo.test_display_name):
1156         (TestPlatformInfo.test_total_bytes_memory):
1157         (TestPlatformInfo.test_free_bytes_memory):
1158         * Scripts/webkitpy/layout_tests/port/factory.py:
1159         (PortFactory._default_port):
1160
1161 2012-05-03  Ojan Vafai  <ojan@chromium.org>
1162
1163         Make the table headers white-space:nowrap. This decreases the vertical size,
1164         which makes for more minimal embedding in garden-o-matic.
1165
1166         * TestResultServer/static-dashboards/flakiness_dashboard_tests.css:
1167         (.table-header-content *):
1168
1169 2012-05-03  Eric Seidel  <eric@webkit.org>
1170
1171         Split build-webkit's feature option list off into its own module in preparation for autogeneration
1172         https://bugs.webkit.org/show_bug.cgi?id=85548
1173
1174         Reviewed by Daniel Bates.
1175
1176         My Perl-fu is very weak.  This moves the feature option list off into a separate module
1177         file so that I can easily autogenerate just that file with the fancy new generate-feature-files command.
1178         It's a bit odd to have an array which we return out to build-webkit to modify, but it turns out
1179         not to matter.  Also, build-webkit doesn't actually use any of the variables we're using to
1180         store the results from the option parse, so those end up inaccessible from build-webkit (except by reference)
1181         which is actually cleaner.  Callers can't assume that they would get a new array every time, but someone
1182         with more Perl-fu than I could create something fancier than this if we need in the future.
1183
1184         * Scripts/build-webkit:
1185         * Scripts/webkitperl/FeatureList.pm: Added.
1186         (getFeatureOptionList):
1187
1188 2012-05-03  Dirk Pranke  <dpranke@chromium.org>
1189
1190         REGRESSION: run-webkit-tests failing on Chromium Mac
1191         https://bugs.webkit.org/show_bug.cgi?id=85459
1192
1193         Unreviewed, build fix.
1194
1195         Turns out select doesn't throw IOErrors, it has it's own kind of
1196         error :(. Trap that instead.
1197
1198         * Scripts/webkitpy/layout_tests/port/server_process.py:
1199         (ServerProcess._wait_for_data_and_update_buffers_using_select):
1200
1201 2012-05-03  Raphael Kubo da Costa  <rakuco@webkit.org>
1202
1203         [CMake] Rewrite FindCairo.cmake.
1204         https://bugs.webkit.org/show_bug.cgi?id=84895
1205
1206         Reviewed by Daniel Bates.
1207
1208         The old approach relied on pkg-config for finding Cairo (which
1209         introduced a dependency on pkg-config that could be avoided), used
1210         the LibFindMacros code that we should probably remove in the
1211         future and did not use the FindPackageHandleStandardArguments
1212         module.
1213
1214         Change all that by rewriting the module.
1215         - Use the pkg-config output optionally instead of requiring it
1216         like LibFindMacros did.
1217         - Remove the implicit dependency on FreeType which often found it
1218         the wrong way via pkg-config and without considering
1219         CMAKE_PREFIX_PATH.
1220         - Retrieve the Cairo version by looking at cairo-version.h instead
1221         of relying on pkg-config. It requires some additional code for
1222         checking if the desired version has been found, but that will not
1223         be needed once we start depending on CMake 2.8.3 or later.
1224
1225         The only downside is that FPHSA sets <UPPERCASED_NAME>_FOUND
1226         instead of <Name>_FOUND, and to keep things consistent
1227         Cairo_LIBRARIES and Cairo_INCLUDE_DIRS have become CAIRO_LIBRARIES
1228         and CAIRO_INCLUDE_DIRS.
1229
1230         * DumpRenderTree/efl/CMakeLists.txt: Use CAIRO_FOO instead of
1231         Cairo_FOO.
1232         * EWebLauncher/CMakeLists.txt: Ditto.
1233
1234 2012-05-03  Ojan Vafai  <ojan@chromium.org>
1235
1236         Fix modifier and bug sorting on the flakiness dashboard
1237         https://bugs.webkit.org/show_bug.cgi?id=85554
1238
1239         Reviewed by Dirk Pranke.
1240
1241         The sorting assumes that the column header name matches a field
1242         in the testResults object and uses that field for sorting.
1243         Filter the bugs out of the modifiers list and put them in a bugs field
1244         and only put the non-bug modifiers in the modifiers list.
1245
1246         * TestResultServer/static-dashboards/flakiness_dashboard.js:
1247         (populateExpectationsData):
1248         (htmlForSingleTestRow):
1249         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
1250
1251 2012-05-03  Ojan Vafai  <ojan@chromium.org>
1252
1253         Fix some state resetting when running unittests. HtmlForSingleTestRow
1254         was flaky.
1255
1256         * TestResultServer/static-dashboards/flakiness_dashboard.js:
1257         (htmlForSingleTestRow):
1258         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
1259         (resetGlobals):
1260
1261 2012-05-03  Raphael Kubo da Costa  <rakuco@webkit.org>
1262
1263         webkitpy: Use PlatformInfo wherever possible in the User class.
1264         https://bugs.webkit.org/show_bug.cgi?id=85549
1265
1266         Reviewed by Dirk Pranke.
1267
1268         Instead of relying on sys.platform to detect whether the current
1269         platform is a Mac, use the PlatformInfo class instead.
1270
1271         * Scripts/webkitpy/common/system/user.py:
1272         (User.__init__):
1273         (User.edit_changelog):
1274
1275 2012-05-03  Ojan Vafai  <ojan@chromium.org>
1276
1277         Remove the percent failed column from the flakiness dashboard
1278         https://bugs.webkit.org/show_bug.cgi?id=85552
1279
1280         Reviewed by Dirk Pranke.
1281
1282         This column isn't useful enough to justify the visual noise.
1283
1284         * TestResultServer/static-dashboards/flakiness_dashboard.js:
1285         (processMissingAndExtraExpectations):
1286         (tableHeaders):
1287         (htmlForSingleTestRow):
1288         (sortTests):
1289         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
1290         (test):
1291
1292 2012-05-03  Eric Seidel  <eric@webkit.org>
1293
1294         --coverage should not not use the "feature" options system in build-webkit, similarly remove special-casing for WTF_URL
1295         https://bugs.webkit.org/show_bug.cgi?id=85542
1296
1297         Reviewed by Daniel Bates.
1298
1299         Remove these last two odd-men-out, so that we can generate the default feature options
1300         for build-webkit using the new generate-feature-files command.
1301         The Mac build system has some special casing for ENABLE_ defines which gets them
1302         defined as -D on the build command, even though initially they start out as
1303         environment variables ENABLE_FOO = ENABLE_FOO.  The Feature options in build-webkit
1304         depend on this support, but it doesn't work for WTF_USE_ defines like WTF_USE_WTF_URL.
1305         I'm told that WTF_URL is not actively being worked on, so just removing it for now.
1306         It should probably come back to life as an ENABLE_, or build-webkit or the Xcode projects
1307         should be made generically smarter about these USE_WTF defines.
1308         (There is still WTF_USE_TILED_BACKING_STORE in the feature list, but it only seems
1309         used by Qt.  If someone tried build-webkit --tiled-backing-store on Mac today it
1310         would not work!)
1311
1312         * Scripts/build-webkit:
1313
1314 2012-05-03  Eric Seidel  <eric@webkit.org>
1315
1316         Remove the last bits of Leopard support from our perl scripts
1317         https://bugs.webkit.org/show_bug.cgi?id=85545
1318
1319         Reviewed by Daniel Bates.
1320
1321         Chromium still supports Leopard (for a little while yet), but AppleMacWebKit dropped
1322         Leopard support several months ago.  Remove the last bits of isLeopard() branching
1323         in our Perl scripts, since no one is using this (Chromium has never used this Perl code).
1324
1325         * Scripts/build-webkit:
1326         * Scripts/old-run-webkit-tests:
1327         (countAndPrintLeaks):
1328         (captureSavedCrashLog):
1329         * Scripts/webkitdirs.pm:
1330         (determineArchitecture):
1331         (argumentsForRunAndDebugMacWebKitApp):
1332
1333 2012-05-03  Ojan Vafai  <ojan@chromium.org>
1334
1335         [Chromium] Don't show the WebKit Linux ASAN bot on the flakiness dashboard.
1336         https://bugs.webkit.org/show_bug.cgi?id=85551
1337
1338         Reviewed by Dirk Pranke.
1339
1340         It's not uploading results and in the short term we don't plan to have it start.
1341
1342         * TestResultServer/static-dashboards/builders.js:
1343         (isChromiumWebkitTipOfTreeTestRunner):
1344         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
1345
1346 2012-05-03  Ojan Vafai  <ojan@chromium.org>
1347
1348         Show the test name in the dashboard if showChrome==false and we're showing multiple tests
1349         https://bugs.webkit.org/show_bug.cgi?id=85243
1350
1351         Reviewed by Dirk Pranke.
1352
1353         This happens for cases where a test name matches multiple tests.
1354         In practice, this only happens for virtual test suites (e.g. the canvas tests).
1355
1356         * TestResultServer/static-dashboards/flakiness_dashboard.js:
1357         (htmlForIndividualTestOnAllBuildersWithChrome):
1358         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
1359         (testHtmlForIndividualTestOnAllBuildersWithChromeNonexistant):
1360         (testHtmlForIndividualTestOnAllBuildersWithChrome):
1361         (testHtmlForIndividualTestOnAllBuildersWithChromeWebkitMaster):
1362         (testHtmlForIndividualTests):
1363
1364 2012-05-03  Ojan Vafai  <ojan@chromium.org>
1365
1366         Port the flakiness dashboard unittests to qunit
1367         https://bugs.webkit.org/show_bug.cgi?id=85544
1368
1369         Reviewed by Adam Barth.
1370
1371         * TestResultServer/static-dashboards/dashboard_base.js:
1372         (appendJSONScriptElements):
1373         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
1374         (resetGlobals):
1375         (runExpectationsTest):
1376         (test):
1377         * TestResultServer/static-dashboards/run-unittests.html: Added.
1378
1379 2012-05-03  Raphael Kubo da Costa  <rakuco@webkit.org>
1380
1381         webkitpy: Remove check for readline on Mac.
1382         https://bugs.webkit.org/show_bug.cgi?id=85547
1383
1384         Reviewed by Eric Seidel.
1385
1386         Python 2.5 is not supported by the WebKit project anymore, and
1387         Python's readline documentation does not mention the module not
1388         being available in the supported versions.
1389
1390         * Scripts/webkitpy/common/system/user.py:
1391
1392 2012-05-03  Thiago Marcos P. Santos  <thiago.santos@intel.com>
1393
1394         [NRWT] Run performance tests with lock
1395         https://bugs.webkit.org/show_bug.cgi?id=78628
1396
1397         Reviewed by Tony Chang.
1398
1399         Locking performance tests (like we do for http tests) will force them
1400         to run in serial. This reduces the load of the machine when running perf
1401         tests and minimizes the chances of the tests to fail, specially
1402         timeouts.
1403
1404         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1405         (Manager.__init__):
1406         (Manager._is_perf_test):
1407         (Manager._test_requires_lock):
1408         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1409
1410 2012-05-03  Dirk Pranke  <dpranke@chromium.org>
1411
1412         Attempt to fix a crash on SL ...
1413
1414         Unreviewed, build fix.
1415
1416         It looks like we may be getting an IOError raised during
1417         select() on chromium mac, and that error isn't being caught.
1418         Pushing it into the try/catch block will hopefully catch this
1419         and allow us to see why the subprocess is crashing.
1420
1421         * Scripts/webkitpy/layout_tests/port/server_process.py:
1422         (ServerProcess._wait_for_data_and_update_buffers_using_select):
1423
1424 2012-05-02  Jon Honeycutt  <jhoneycutt@apple.com>
1425
1426         WebFrameLoaderClient::dispatchWillSendSubmitEvent() needs to be
1427         implemented for WebKit2
1428         https://bugs.webkit.org/show_bug.cgi?id=84304
1429
1430         Reviewed by Jessie Berlin.
1431
1432         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
1433         Added new files to project.
1434
1435         * TestWebKitAPI/Tests/WebKit2/WillSendSubmitEvent.cpp: Added.
1436         (TestWebKitAPI::didReceiveMessageFromInjectedBundle):
1437         Look for the "DidReceiveWillSendSubmitEvent" message. The message body
1438         will be the dictionary of form control names and values passed to the
1439         injected bundle's willSendSubmitEvent handler; look for the pairs
1440         "textField" -> "text field" and "passwordField" -> "password field".
1441         Check that "hiddenField" is not included.
1442         (TestWebKitAPI::setInjectedBundleClient):
1443         (TestWebKitAPI::TEST):
1444
1445         * TestWebKitAPI/Tests/WebKit2/WillSendSubmitEvent_Bundle.cpp: Added.
1446         (TestWebKitAPI::willSendSubmitEvent):
1447         Post a message to the TestWebKitAPI process with the dictionary of form
1448         control names and values that we received.
1449         (TestWebKitAPI::WillSendSubmitEventTest::WillSendSubmitEventTest):
1450         (TestWebKitAPI::WillSendSubmitEventTest::didCreatePage):
1451         Set up the page's form client.
1452
1453         * TestWebKitAPI/Tests/WebKit2/auto-submitting-form.html: Added.
1454
1455 2012-05-03  Eric Seidel  <eric@webkit.org>
1456
1457         Remove dead code from build-webkit for working around old XCode dependency bug
1458         https://bugs.webkit.org/show_bug.cgi?id=85535
1459
1460         Reviewed by Adam Barth.
1461
1462         Xcode used to have a bug whereby it didn't notice if you changed the #defines
1463         passed on the command line to xcodebuild.  I wrote code to work around this
1464         while bringing up SVG support 5-6 years ago.  I think we can finally remove this
1465         code from build-webkit, since SVG is the only feature flag to ever use it
1466         (modern in-progress features like MathML don't seem to be using the support)
1467         an I suspect the XCode bug has long been fixed.
1468
1469         * Scripts/build-webkit:
1470         * Scripts/webkitperl/features.pm:
1471
1472 2012-05-03  Andy Estes  <aestes@apple.com>
1473
1474         InjectedBundleControllerMac.mm needs to include config.h
1475         https://bugs.webkit.org/show_bug.cgi?id=85531
1476
1477         Reviewed by Andreas Kling.
1478
1479         * TestWebKitAPI/mac/InjectedBundleControllerMac.mm:
1480
1481 2012-05-03  Raphael Kubo da Costa  <rakuco@webkit.org>
1482
1483         webkitpy: Use PlatformInfo whenever possible in server_process.py.
1484         https://bugs.webkit.org/show_bug.cgi?id=85516
1485
1486         Reviewed by Eric Seidel.
1487
1488         Eliminate direct access to sys.platform wherever possible by using
1489         PlatformInfo.is_win() instead.
1490
1491         * Scripts/webkitpy/layout_tests/port/server_process.py:
1492         (ServerProcess.__init__): Remove the executive parameter as it was
1493         not passed by any caller, obtain a SystemHost from port_obj
1494         instead.
1495         (ServerProcess._log): Style, remove extra empty line.
1496         (ServerProcess._start):
1497         (ServerProcess.stop):
1498         (ServerProcess.kill):
1499         * Scripts/webkitpy/layout_tests/port/server_process_unittest.py:
1500         (TrivialMockPort.__init__): Create a MockSystemHost used by
1501         ServerProcess.
1502         (TestServerProcess.test_broken_pipe): Test different platforms.
1503
1504 2012-05-03  Thiago Marcos P. Santos  <thiago.santos@intel.com>
1505
1506         webkitpy: Style checker broken when committing a PNG
1507         https://bugs.webkit.org/show_bug.cgi?id=85504
1508
1509         Reviewed by Dirk Pranke.
1510
1511         * Scripts/webkitpy/style/checkers/png.py:
1512         (PNGChecker.__init__):
1513
1514 2012-05-03  Philippe Normand  <pnormand@igalia.com>
1515
1516         Unreviewed, fix another call to join(), similar to the ones landed
1517         in r115975.
1518
1519         * Scripts/webkitdirs.pm:
1520         (jhbuildConfigurationChanged):
1521
1522 2012-05-03  Raphael Kubo da Costa  <rakuco@webkit.org>
1523
1524         webkitpy: Remove sys.platform accesses in CrashLogs.
1525         https://bugs.webkit.org/show_bug.cgi?id=85436
1526
1527         Reviewed by Dirk Pranke.
1528
1529         Instead of directly checking for sys.platform == 'darwin', use a
1530         SystemHost and call PlatformInfo.is_mac().
1531
1532         * Scripts/webkitpy/common/system/crashlogs.py:
1533         (CrashLogs.__init__):
1534         (CrashLogs.find_newest_log):
1535         (CrashLogs._log_directory_darwin):
1536         (CrashLogs._find_newest_log_darwin):
1537         * Scripts/webkitpy/common/system/crashlogs_unittest.py:
1538         (CrashLogsTest.test_find_log_darwin):
1539         * Scripts/webkitpy/layout_tests/port/mac.py:
1540         (MacPort._get_crash_log):
1541         * Scripts/webkitpy/layout_tests/port/test.py:
1542         (TestDriver.run_test):
1543         * Scripts/webkitpy/tool/commands/queries.py:
1544         (execute):
1545
1546 2012-05-03  Philippe Normand  <pnormand@igalia.com>
1547
1548         [GTK] join() is wrongly used in webkitdirs.pm since r115532
1549         https://bugs.webkit.org/show_bug.cgi?id=85501
1550
1551         Reviewed by Martin Robinson.
1552
1553         * Scripts/webkitdirs.pm:
1554         (getJhbuildPath):
1555         (buildAutotoolsProject):
1556
1557 2012-05-03  Zan Dobersek  <zandobersek@gmail.com>
1558
1559         [Gtk] ImageDiff crashes when it can't open a display
1560         https://bugs.webkit.org/show_bug.cgi?id=85476
1561
1562         Reviewed by Martin Robinson.
1563
1564         Replace gdk_init() initialization call with the g_type_init()
1565         call. GdkPixbufLoader doesn't actually need the current call to be
1566         made and this change will also remove crashes when ImageDiff
1567         is started in an environment without a DISPLAY env.
1568
1569         * DumpRenderTree/gtk/ImageDiff.cpp:
1570         (main):
1571
1572 2012-05-03  Zan Dobersek  <zandobersek@gmail.com>
1573
1574         ImageDiff should be run inside a properly established environment
1575         https://bugs.webkit.org/show_bug.cgi?id=85292
1576
1577         Reviewed by Martin Robinson.
1578
1579         When creating the ImageDiff server process, pass along a freshly-set-up environment
1580         in which the process should be executed in.
1581
1582         * Scripts/webkitpy/layout_tests/port/webkit.py:
1583         (WebKitPort._start_image_diff_process):
1584
1585 2012-05-03  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
1586
1587         [Qt] Allow the web process and WTR to be paused on startup
1588
1589         Makes it easier to debug the web process or run-webkit-tests -2, as you
1590         have ample time to attach gdb to the process.
1591
1592         Reviewed by Simon Hausmann.
1593
1594         * Scripts/webkitpy/layout_tests/port/qt.py:
1595         (QtPort.setup_environ_for_server):
1596         * WebKitTestRunner/qt/main.cpp:
1597         (main):
1598
1599 2012-05-03  Csaba Osztrogonác  <ossy@webkit.org>
1600
1601         Password protect "Stop build" button
1602         https://bugs.webkit.org/show_bug.cgi?id=81982
1603
1604         Reviewed by Ryosuke Niwa.
1605
1606         Only authenticated users should be able to use stopBuild, stopAllBuilds, cancelPendingBuild
1607         features to prevent killing builds on build.webkit.org by SPAM bots.
1608
1609         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1610
1611 2012-05-03  Dongwoo Im  <dw.im@samsung.com>
1612
1613         [DRT] Remove the name of parameter, visibility, from the setPageVisibility stub implementations.
1614         https://bugs.webkit.org/show_bug.cgi?id=85468
1615
1616         Reviewed by Nikolas Zimmermann.
1617
1618         * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp:
1619         (LayoutTestController::setPageVisibility): Remove the name of parameter, visibility.
1620         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1621         (LayoutTestController::setPageVisibility): Remove the name of parameter, visibility.
1622         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1623         (LayoutTestController::setPageVisibility): Remove the name of parameter, visibility.
1624         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1625         (LayoutTestController::setPageVisibility): Remove the name of parameter, visibility.
1626         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
1627         (LayoutTestController::setPageVisibility): Remove the name of parameter, visibility.
1628
1629 2012-05-03  Stephanie Lewis  <slewis@apple.com>
1630
1631         https://bugs.webkit.org/show_bug.cgi?id=85450 unbounded growth of JSDOMWindowShells loading pages in the same window
1632         <rdar://problem/11320059> REGRESSION (r115083): PLT3 shows linear memory growth and gets slower with each run
1633
1634         Reviewed by Brady Eidson.
1635
1636         The API added for DOMWindowExtension, didCreateGlobalObjectForFrame, would create a global object
1637         for every world, even those that did not need the callback.  This had the side effect of creating a
1638         JSDOMWindowShell that the associated world didn't necessarily know to clean up.  Instead of creating
1639         unnecessary objects change the API to globalObjectIsAvailableForFrame and do not pass the global object
1640         in the API.  The object can be accessed later by those worlds which require it.
1641
1642         * TestWebKitAPI/Tests/WebKit2/DOMWindowExtensionBasic.cpp:
1643         (TestWebKitAPI):
1644         (TestWebKitAPI::didReceiveMessageFromInjectedBundle):
1645         * TestWebKitAPI/Tests/WebKit2/DOMWindowExtensionBasic_Bundle.cpp:
1646         (TestWebKitAPI):
1647         (DOMWindowExtensionBasic):
1648         (TestWebKitAPI::DOMWindowExtensionBasic::didCreatePage):
1649         (TestWebKitAPI::DOMWindowExtensionBasic::globalObjectIsAvailableForFrame):
1650         (TestWebKitAPI::globalObjectIsAvailableForFrameCallback):
1651
1652 2012-05-03  Nikolas Zimmermann  <nzimmermann@rim.com>
1653
1654         Not reviewed. Fix mac build, it was missing setPageVisibility/resetPageVisibility stub implementations.
1655
1656         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1657         (LayoutTestController::resetPageVisibility):
1658         (LayoutTestController::setPageVisibility):
1659
1660 2012-05-02  Alexander Færøy  <ahf@0x90.dk>
1661
1662         Rename deviceDPI to devicePixelRatio
1663         https://bugs.webkit.org/show_bug.cgi?id=85049
1664
1665         Reviewed by Kenneth Rohde Christiansen.
1666
1667         Use 1.5 as device pixel ratio for the Qt MiniBrowser.
1668
1669         * MiniBrowser/qt/qml/BrowserWindow.qml:
1670
1671 2012-05-03  Dongwoo Im  <dw.im@samsung.com>
1672
1673         [EFL][DRT] Implement the LayoutTestController's methods related to the Page Visibility API.
1674         https://bugs.webkit.org/show_bug.cgi?id=85347
1675
1676         Reviewed by Nikolas Zimmermann.
1677
1678         * DumpRenderTree/LayoutTestController.h:
1679         (LayoutTestController):
1680         * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp:
1681         (LayoutTestController::resetPageVisibility): Not implemented.
1682         (LayoutTestController::setPageVisibility): Not implemented.
1683         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp:
1684         (LayoutTestController::resetPageVisibility): Reset the page visibility.
1685         (LayoutTestController::setPageVisibility): Set the page visibility.
1686         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1687         (LayoutTestController::resetPageVisibility): Not implemented.
1688         (LayoutTestController::setPageVisibility): Not implemented.
1689         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1690         (LayoutTestController::resetPageVisibility): Not implemented.
1691         (LayoutTestController::setPageVisibility): Not implemented.
1692         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1693         (LayoutTestController::resetPageVisibility): Not implemented.
1694         (LayoutTestController::setPageVisibility): Not implemented.
1695         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
1696         (LayoutTestController::resetPageVisibility): Not implemented.
1697         (LayoutTestController::setPageVisibility): Not implemented.
1698
1699 2012-05-02  Sheriff Bot  <webkit.review.bot@gmail.com>
1700
1701         Unreviewed, rolling out r115918.
1702         http://trac.webkit.org/changeset/115918
1703         https://bugs.webkit.org/show_bug.cgi?id=85452
1704
1705         "Broke test-webkitpy on Apple Lion bots" (Requested by estes
1706         on #webkit).
1707
1708         * Scripts/webkitpy/common/system/crashlogs.py:
1709         (CrashLogs.__init__):
1710         (CrashLogs.find_newest_log):
1711         (CrashLogs._log_directory_darwin):
1712         (CrashLogs._find_newest_log_darwin):
1713         * Scripts/webkitpy/common/system/crashlogs_unittest.py:
1714         (CrashLogsTest.test_find_log_darwin):
1715         * Scripts/webkitpy/layout_tests/port/mac.py:
1716         (MacPort._get_crash_log):
1717         * Scripts/webkitpy/layout_tests/port/test.py:
1718         (TestDriver.run_test):
1719         * Scripts/webkitpy/tool/commands/queries.py:
1720         (execute):
1721
1722 2012-05-02  Raphael Kubo da Costa  <rakuco@webkit.org>
1723
1724         webkitpy: Remove sys.platform accesses in CrashLogs.
1725         https://bugs.webkit.org/show_bug.cgi?id=85436
1726
1727         Reviewed by Dirk Pranke.
1728
1729         Instead of directly checking for sys.platform == 'darwin', use a
1730         SystemHost and call PlatformInfo.is_mac().
1731
1732         * Scripts/webkitpy/common/system/crashlogs.py:
1733         (CrashLogs.__init__):
1734         (CrashLogs.find_newest_log):
1735         (CrashLogs._log_directory_darwin):
1736         (CrashLogs._find_newest_log_darwin):
1737         * Scripts/webkitpy/common/system/crashlogs_unittest.py:
1738         (CrashLogsTest.test_find_log_darwin):
1739         * Scripts/webkitpy/layout_tests/port/mac.py:
1740         (MacPort._get_crash_log):
1741         * Scripts/webkitpy/layout_tests/port/test.py:
1742         (TestDriver.run_test):
1743         * Scripts/webkitpy/tool/commands/queries.py:
1744         (execute):
1745
1746 2012-05-02  Raphael Kubo da Costa  <rakuco@webkit.org>
1747
1748         webkitpy: Special-case win32 calls in file_lock and executive.
1749         https://bugs.webkit.org/show_bug.cgi?id=85431
1750
1751         Reviewed by Dirk Pranke.
1752
1753         Simplify accesses to sys.platform by only checking if sys.platform
1754         is "win32" and removing explicit checks for the else case.
1755
1756         * Scripts/webkitpy/common/system/executive.py:
1757         (Executive.check_running_pid):
1758         * Scripts/webkitpy/common/system/file_lock.py:
1759         (FileLock._create_lock):
1760         (FileLock._remove_lock):
1761
1762 2012-05-02  Ojan Vafai  <ojan@chromium.org>
1763
1764         Remove pink from flakiness dashboard
1765         https://bugs.webkit.org/show_bug.cgi?id=85422
1766
1767         Reviewed by Dirk Pranke.
1768
1769         The pink indicates that the test doesn't match it's listing in test_expectations.txt.
1770         In practice, this just added noise and confused people.
1771
1772         * TestResultServer/static-dashboards/flakiness_dashboard.js:
1773         (htmlForSingleTestRow):
1774         (showLegend):
1775         * TestResultServer/static-dashboards/flakiness_dashboard_tests.css:
1776         (#legend-contents *):
1777         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
1778         (testHtmlForSingleTestRow):
1779
1780 2012-05-02  Gustavo Noronha Silva  <gns@gnome.org>
1781
1782         Unreviewed. Fix mistake I committed when applying review comments.
1783
1784         * jhbuild/jhbuild-wrapper: dependencies_path was getting 'Root'
1785         appended to its path, so jhbuild itself was being cloned and
1786         installed in the wrong place when using WEBKITOUTPUTDIR.
1787
1788 2012-05-02  Dirk Pranke  <dpranke@chromium.org>
1789
1790         nrwt: make ServerProcess work on chromium win
1791         https://bugs.webkit.org/show_bug.cgi?id=85333
1792
1793         Reviewed by Ojan Vafai.
1794
1795         This change implements a poor man's select() that will
1796         slow-spin doing non-blocking reads on the stdout and stderr
1797         named pipes connecting the worker to the driver. Seems to work
1798         and I have yet to see much overhead or ill effects but it
1799         probably needs more testing.
1800
1801         * Scripts/webkitpy/layout_tests/port/chromium.py:
1802         (ChromiumDriver.__init__):
1803         * Scripts/webkitpy/layout_tests/port/server_process.py:
1804         (ServerProcess._start):
1805         (ServerProcess._wait_for_data_and_update_buffers_using_select):
1806         (ServerProcess._wait_for_data_and_update_buffers_using_win32_apis):
1807         (ServerProcess):
1808         (ServerProcess._non_blocking_read_win32):
1809         (ServerProcess._read):
1810         * Scripts/webkitpy/layout_tests/port/server_process_unittest.py:
1811         (TestServerProcess.test_basic): Added.
1812
1813 2012-05-02  Raphael Kubo da Costa  <rakuco@webkit.org>
1814
1815         webkitpy: Use a SystemHost instead of PlatformInfo in the PNG checker.
1816         https://bugs.webkit.org/show_bug.cgi?id=85426
1817
1818         Reviewed by Dirk Pranke.
1819
1820         r115880 replaced the direct use of sys.platform with a
1821         PlatformInfo object. Dirk Pranke later told me SystemHost objects
1822         should be preferred, as nothing out of webkitpy.common.system
1823         should create its own PlatformInfo instances.
1824
1825         * Scripts/webkitpy/common/system/systemhost_mock.py:
1826         (MockSystemHost.__init__): Make it possible to specify a custom
1827         filesystem object.
1828         * Scripts/webkitpy/style/checkers/png.py:
1829         (PNGChecker.__init__): Accept a host object instead of a
1830         filesystem one, which is now obtained from host.
1831         (PNGChecker._config_file_path): Retrieve PlatformInfo from
1832         SystemHost.
1833         * Scripts/webkitpy/style/checkers/png_unittest.py:
1834         (PNGCheckerTest.test_init): Adjust calls to PNGChecker to its new
1835         constructor signature.
1836         (PNGCheckerTest.test_check): Ditto.
1837
1838 2012-05-02  Raphael Kubo da Costa  <rakuco@webkit.org>
1839
1840         webkitpy: Remove direct checks for sys.platform == 'darwin' in the layout test classes.
1841         https://bugs.webkit.org/show_bug.cgi?id=85424
1842
1843         Reviewed by Tony Chang.
1844
1845         We should reduce direct usage of sys.platform in webkitpy and use
1846         PlatformInfo instead.
1847
1848         * Scripts/webkitpy/layout_tests/servers/http_server.py:
1849         (Lighttpd._prepare_config): Call PlatformInfo.is_mac() instead of
1850         checking for 'darwin' directly.
1851         * Scripts/webkitpy/layout_tests/servers/http_server_base.py:
1852         (HttpServerBase.__init__): Ditto.
1853
1854 2012-05-02  Raphael Kubo da Costa  <rakuco@webkit.org>
1855
1856         webkitpy: Remove direct sys.platform usage in run_webkit_tests_integrationtest.
1857         https://bugs.webkit.org/show_bug.cgi?id=85419
1858
1859         Reviewed by Dirk Pranke.
1860
1861         As pointed out by Dirk Pranke, a PlatformInfo is not supposed to
1862         be created directly, we should should a SystemHost instead.
1863
1864         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1865         (MainTest.setUp): Create a SystemHost and obtain the PlatformInfo
1866         object from it instead of creating PlatformInfo directly.
1867
1868 2012-05-02  Raphael Kubo da Costa  <rakuco@webkit.org>
1869
1870         webkitpy: Remove direct sys.platform usage in run_webkit_tests_integrationtest.
1871         https://bugs.webkit.org/show_bug.cgi?id=85419
1872
1873         Rubber-stamped by Dirk Pranke.
1874
1875         Follow-up patch to r115887 which fixes some capitalization problems
1876         pointed out in comments 2 and 4.
1877
1878         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1879         (MainTest.setUp):
1880         (MainTest.test_child_processes_2):
1881         (MainTest.test_child_processes_min):
1882         (MainTest.test_exception_raised):
1883         (MainTest.test_keyboard_interrupt):
1884
1885 2012-05-02  Raphael Kubo da Costa  <rakuco@webkit.org>
1886
1887         webkitpy: Remove direct sys.platform usage in run_webkit_tests_integrationtest.
1888         https://bugs.webkit.org/show_bug.cgi?id=85419
1889
1890         Reviewed by Dirk Pranke.
1891
1892         Another part of the work towards replacing direct sys.platform
1893         usage with PlatformInfo.
1894
1895         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1896         (MainTest.setUp): Move the setting of SHOULD_TEST_PROCESSES
1897         here. Even though it is costlier, unittest.setUpClass was
1898         introduced in Python 2.7.
1899         (MainTest.test_child_processes_2): Use self.SHOULD_TEST_PROCESSES.
1900         (MainTest.test_child_processes_min): Ditto.
1901         (MainTest.test_exception_raised): Ditto.
1902         (MainTest.test_keyboard_interrupt): Ditto.
1903         (MainTest.test_crash_log): Check for PlatformInfo.is_mac().
1904         (MainTest.test_web_process_crash_log): Ditto, and remove duplicate
1905         check at the end.
1906         (MainTest.test_no_http_tests): Remove check for (cygwin && Python
1907         2.5.2), as we only support Python 2.6+.
1908
1909 2012-05-02  Raphael Kubo da Costa  <rakuco@webkit.org>
1910
1911         webkitpy: Remove direct sys.platform usage in the PNG checker.
1912         https://bugs.webkit.org/show_bug.cgi?id=85407
1913
1914         Reviewed by Tony Chang.
1915
1916         Work towards replacing direct sys.platform usage with
1917         PlatformInfo.
1918
1919         * Scripts/webkitpy/common/system/platforminfo_mock.py:
1920         (MockPlatformInfo.__init__): Accept os_name and os_version as
1921         arguments.
1922         * Scripts/webkitpy/style/checkers/png.py:
1923         (PNGChecker.__init__): Create a PlatformInfo object instead of
1924         using sys.platform when needed.
1925         (PNGChecker._config_file_path): Check for the current platform via
1926         PlatformInfo.
1927         * Scripts/webkitpy/style/checkers/png_unittest.py:
1928         (PNGCheckerTest.test_check): Update tests after the changes above.
1929
1930 2012-05-02  Raphael Kubo da Costa  <rakuco@webkit.org>
1931
1932         [watchlist] Update watchlist rules for EFL-related patches.
1933         http://bugs.webkit.org/show_bug.cgi?id=85160
1934
1935         Reviewed by Eric Seidel.
1936
1937         * Scripts/webkitpy/common/config/watchlist: Monitor
1938         Source/WTF/wtf/efl and LayoutTests/platform/efl as well.
1939
1940 2012-05-02  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
1941
1942         Revert r115191 - "Make the web view's url property follow the active url"
1943
1944         It uncovered/caused issues in the icon-implementation that can't be
1945         easily fixed, so rolling out instead.
1946
1947         https://bugs.webkit.org/show_bug.cgi?id=77554
1948
1949         Rubber-stamped by Simon Hausmann.
1950
1951         * MiniBrowser/qt/qml/BrowserWindow.qml:
1952         * TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp:
1953         (TestWebKitAPI::TEST):
1954
1955 2012-05-02  Zan Dobersek  <zandobersek@gmail.com>
1956
1957         Reftest mismatches are providing undefined diff percent in results.html
1958         https://bugs.webkit.org/show_bug.cgi?id=85283
1959
1960         Reviewed by Ojan Vafai.
1961
1962         When writing test results for a reftest mismatch, also take into account the
1963         difference percent that is produced when generating image diff. This value is
1964         then used when creating test results in the form of dictionaries, stored under
1965         the 'image_diff_percent' key. It ultimately ends up in JSON results data,
1966         providing additional information about the reftest difference.
1967
1968         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1969         (interpret_test_failures): Store the diff percent for reftest failure under the
1970         'image_diff_percent' key.
1971         * Scripts/webkitpy/layout_tests/controllers/test_result_writer.py:
1972         (write_test_result): Also set the new diff_percent attribute of the reftest
1973         mismatch failure to the value of the difference percent ImageDiff outputs.
1974         * Scripts/webkitpy/layout_tests/models/test_failures.py:
1975         (FailureReftestMismatch.__init__): Add the new diff_percent attribute with the
1976         default value of None.
1977         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1978         (EndToEndTest.test_reftest_with_two_notrefs): Update the test case expected results.
1979
1980 2012-05-02  Terry Anderson  <tdanderson@chromium.org>
1981
1982         Addition of a contributor to committers.py
1983         https://bugs.webkit.org/show_bug.cgi?id=85324
1984
1985         Reviewed by Unreviewed.
1986
1987         * Scripts/webkitpy/common/config/committers.py:
1988
1989 2012-05-02  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
1990
1991         [EFL] EFL's LayoutTestController overridePreference implementation
1992         https://bugs.webkit.org/show_bug.cgi?id=83007
1993
1994         Reviewed by Antonio Gomes.
1995
1996         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
1997         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
1998         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp:
1999         (toBool): aux function converting JSStringRef to bool
2000         (toInt): aux function converting JSStringRef to int
2001         (LayoutTestController::overridePreference): added implementation
2002
2003 2012-05-01  Raphael Kubo da Costa  <rakuco@webkit.org>
2004
2005         webkitpy: Fix sys.platform check in find_files_unittest.
2006         https://bugs.webkit.org/show_bug.cgi?id=85339
2007
2008         Reviewed by Ryosuke Niwa.
2009
2010         Follow-up to r101618 -- 'win' is not a valid sys.platform value,
2011         'win32' should be used instead. The check always returned false
2012         due to this, so assert_normalize() was never called.
2013
2014         * Scripts/webkitpy/common/find_files_unittest.py:
2015         (TestWinNormalize.test_win): Add the missing import for the
2016         FileSystem class and fix the sys.platform check.
2017
2018 2012-05-02  Zan Dobersek  <zandobersek@gmail.com>
2019
2020         [GTK] drag&drop related failures
2021         https://bugs.webkit.org/show_bug.cgi?id=84675
2022
2023         Reviewed by Philippe Normand.
2024
2025         Rather than checking options to see whether we're running pixel tests
2026         the pixel_tests argument passed to GtkDriver._start should be used as
2027         it also covers cases when pixel tests should be run because of testing
2028         reftests. This again creates different Xvfb processes with proper
2029         display IDs for a worker that's performing both ordinary and pixel
2030         tests, as was the case before changes made in r114788.
2031
2032         * Scripts/webkitpy/layout_tests/port/gtk.py:
2033         (GtkDriver._start):
2034
2035 2012-05-02  Simon Hausmann  <simon.hausmann@nokia.com>
2036
2037         [Qt] Enable device orientation/motion with Qt 5
2038
2039         Reviewed by Tor Arne Vestbø.
2040
2041         Enable device orientation/motion features with Qt 5 if Qt Sensors is available.
2042
2043         * qmake/mkspecs/features/features.prf:
2044
2045 2012-05-01  Raphael Kubo da Costa  <rakuco@webkit.org>
2046
2047         webkitpy: Remove unused platform imports.
2048         https://bugs.webkit.org/show_bug.cgi?id=85341
2049
2050         Reviewed by Eric Seidel.
2051
2052         Remove a few uses of `import platform' which were either rendered
2053         unneeded after r105931 or were never needed at all.
2054
2055         * Scripts/webkitpy/common/system/executive.py:
2056         * Scripts/webkitpy/tool/bot/flakytestreporter.py:
2057
2058 2012-05-01  Terry Anderson  <tdanderson@chromium.org>
2059
2060         [chromium] Accept four parameters when dispatching a WebInputEvent::GestureTap in chromium DRT eventSender
2061         https://bugs.webkit.org/show_bug.cgi?id=85289
2062
2063         Reviewed by Eric Seidel.
2064
2065         * DumpRenderTree/chromium/EventSender.cpp:
2066         (EventSender::gestureEvent):
2067             This change will allow for testing of WK85101. Currently
2068             no tests use the extra parameters deltaX and deltaY. Added
2069             a check for the size of |arguments|.
2070
2071 2012-05-01  Eric Seidel  <eric@webkit.org>
2072
2073         Add myself to the watchlist to watch all rendering changes, as well as sort the watchers in the list
2074         https://bugs.webkit.org/show_bug.cgi?id=85305
2075
2076         Reviewed by Adam Barth.
2077
2078         * Scripts/webkitpy/common/config/watchlist:
2079
2080 2012-05-01  Sheriff Bot  <webkit.review.bot@gmail.com>
2081
2082         Unreviewed, rolling out r115735.
2083         http://trac.webkit.org/changeset/115735
2084         https://bugs.webkit.org/show_bug.cgi?id=85314
2085
2086         cause two tests to crash due to an ASSERTION failure
2087         (Requested by zhenyao on #webkit).
2088
2089         * DumpRenderTree/chromium/EventSender.cpp:
2090         (EventSender::gestureEvent):
2091
2092 2012-05-01  Terry Anderson  <tdanderson@chromium.org>
2093
2094         [chromium] Accept four parameters when dispatching a WebInputEvent::GestureTap in chromium DRT eventSender
2095         https://bugs.webkit.org/show_bug.cgi?id=85289
2096
2097         Reviewed by Eric Seidel.
2098
2099         * DumpRenderTree/chromium/EventSender.cpp:
2100         (EventSender::gestureEvent):
2101             This change will allow for testing of WK85101. Currently no tests use the extra
2102             parameters deltaX and deltaY.
2103
2104 2012-05-01  Zan Dobersek  <zandobersek@gmail.com>
2105
2106         Print out stderr output of ImageDiff if it is present
2107         https://bugs.webkit.org/show_bug.cgi?id=85285
2108
2109         Reviewed by Philippe Normand.
2110
2111         Log the stderr output of ImageDiff as a warning if present. This is
2112         especially helpful when ImageDiff crashes and stderr data might contain
2113         relevant information about what caused the crash.
2114
2115         * Scripts/webkitpy/layout_tests/port/webkit.py:
2116         (WebKitPort._read_image_diff):
2117
2118 2012-05-01  Kenneth Rohde Christiansen  <kenneth@webkit.org>
2119
2120         [Qt] Add an experimental extension to set the min. contents width
2121         https://bugs.webkit.org/show_bug.cgi?id=85281
2122
2123         Reviewed by Antonio Gomes.
2124
2125         * MiniBrowser/qt/qml/BrowserWindow.qml:
2126
2127         Set the value of preferredMinimumContentsWidth to 980.
2128
2129 2012-04-30  Ojan Vafai  <ojan@chromium.org>
2130
2131         Fix tests to not depend on the actual list of builders at build.chromium.org.
2132
2133         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
2134         (testHtmlForIndividualTestOnAllBuildersWithChrome):
2135         (testHtmlForIndividualTestOnAllBuildersWithChromeWebkitMaster):
2136
2137 2012-04-29  Sam Weinig  <sam@webkit.org>
2138
2139         Remove BlobBuilder
2140         https://bugs.webkit.org/show_bug.cgi?id=84036
2141
2142         Reviewed by Anders Carlsson.
2143
2144         * Scripts/build-webkit:
2145         * qmake/mkspecs/features/features.prf:
2146         Make exposing the WebKitBlobBuilder JS object conditional on a new
2147         ENABLE_LEGACY_WEBKIT_BLOB_BUILDER flag. Don't enable this for the Mac,
2148         but do for everyone else.
2149
2150 2012-04-30  Amruth Raj Padmanabhuni  <amruthraj@motorola.com>
2151
2152         [GTK] Add glib as a dependency to build glib-networking
2153         https://bugs.webkit.org/show_bug.cgi?id=85181
2154
2155         Reviewed by Martin Robinson.
2156
2157         * gtk/jhbuild.modules:
2158
2159 2012-04-30  Alexis Menard  <alexis.menard@openbossa.org>
2160
2161         Move myself to reviewers.
2162
2163         Reviewed by Andreas Kling.
2164
2165         * Scripts/webkitpy/common/config/committers.py:
2166
2167 2012-04-30  Jocelyn Turcotte  <jocelyn.turcotte@nokia.com>
2168
2169         Unreviewed, move myself to reviewers
2170
2171         * Scripts/webkitpy/common/config/committers.py:
2172
2173 2012-04-30  Carlos Garcia Campos  <cgarcia@igalia.com>
2174
2175         Unreviewed. Fix make distcheck.
2176
2177         * TestWebKitAPI/GNUmakefile.am: Add missing headers to compilation
2178         and resources to EXTRA_DIST.
2179
2180 2012-04-30  Sudarsana Nagineni  <sudarsana.nagineni@linux.intel.com>
2181
2182         [GTK] DRT support for layoutTestController.removeOriginAccessWhitelistEntry
2183         https://bugs.webkit.org/show_bug.cgi?id=85156
2184
2185         Reviewed by Martin Robinson.
2186
2187         Add missing implementation removeOriginAccessWhitelistEntry to
2188         GTK's LayoutTestController.
2189
2190         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2191         (LayoutTestController::removeOriginAccessWhitelistEntry):
2192
2193 2012-04-29  Maciej Stachowiak  <mjs@apple.com>
2194
2195         REGRESSION: On Lion, run-webkit-tests changes the display color profile even when not running pixel tests
2196         https://bugs.webkit.org/show_bug.cgi?id=81729
2197
2198         Reviewed by Dan Bernstein.
2199
2200         * Scripts/webkitpy/layout_tests/controllers/manager.py:
2201         (Manager._set_up_run): Only launch LayoutTestHelper when running pixel tests.
2202
2203 2012-04-29  Sudarsana Nagineni  <sudarsana.nagineni@linux.intel.com>
2204
2205         [GTK] DRT needs an implementation of layoutTestController.setDomainRelaxationForbiddenForURLScheme
2206         https://bugs.webkit.org/show_bug.cgi?id=85131
2207
2208         Reviewed by Martin Robinson.
2209
2210         Add missing implementation setDomainRelaxationForbiddenForURLScheme to
2211         GTK's LayoutTestController.
2212
2213         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2214         (LayoutTestController::setDomainRelaxationForbiddenForURLScheme):
2215
2216 2012-04-28  Sam Weinig  <sam@webkit.org>
2217
2218         Smooth scrolling needs a new key
2219         <rdar://problem/11331632>
2220
2221         Reviewed by Dan Bernstein.
2222
2223         * DumpRenderTree/mac/DumpRenderTree.mm:
2224         (resetDefaultsToConsistentValues):
2225         * TestWebKitAPI/mac/InjectedBundleControllerMac.mm:
2226         (TestWebKitAPI::InjectedBundleController::platformInitialize):
2227         * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm:
2228         (WTR::InjectedBundle::platformInitialize):
2229         Update for new key.
2230
2231 2012-04-28  Zan Dobersek  <zandobersek@gmail.com>
2232
2233         [WK2] http/tests/navigation/anchor-frames-gbk.html fails
2234         https://bugs.webkit.org/show_bug.cgi?id=76896
2235
2236         Reviewed by Darin Adler.
2237
2238         Only dump as text if currently dumping render tree. This ensures that
2239         calling dumpAsText in a test with dumpChildFramesAsText already being
2240         called doesn't override the first decision, possibly making tests fail.
2241
2242         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
2243         (WTR::LayoutTestController::dumpAsText):
2244
2245 2012-04-28  Dan Winship  <danw@gnome.org>
2246
2247         Unreviewed; Add me to the SoupNetwork watchlist
2248
2249         * Scripts/webkitpy/common/config/watchlist:
2250
2251 2012-04-27  Gustavo Noronha Silva  <gns@gnome.org>
2252
2253         [GTK] jhbuild cloning is not following WEBKITOUTPUTDIR.
2254         https://bugs.webkit.org/show_bug.cgi?id=76161
2255
2256         Reviewed by Martin Robinson.
2257
2258         * Scripts/webkitdirs.pm:
2259         (getJhbuildPath): New method to obtain the jhbuild base directory,
2260         using the product base directory
2261         (jhbuildConfigurationChanged): Use the new method
2262         (buildAutotoolsProject): Ditto.
2263         * efl/jhbuildrc: Use WEBKITOUTPUTDIR when calculating the path.
2264         * gtk/jhbuildrc: Ditto.
2265         * jhbuild/jhbuild-wrapper: Ditto.
2266
2267 2012-04-27  Scott Graham  <scottmg@chromium.org>
2268
2269         [chromium] make case of npTestNetscapePlugIn.dll match case in copy rule
2270         https://bugs.webkit.org/show_bug.cgi?id=85083
2271
2272         Reviewed by Dirk Pranke.
2273
2274         Make output product_name for npTestNetscapePlugIn.dll match the case
2275         of the copy_TestNetscapePlugIn rule. This is required for ninja, which
2276         is more particular about case matching in rules than the VS build.
2277
2278         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
2279
2280 2012-04-27  Dirk Pranke  <dpranke@chromium.org>
2281
2282         [chromium] use "drt-style" output, not "test-shell-style" output, on mac and linux DRT
2283         https://bugs.webkit.org/show_bug.cgi?id=84917
2284
2285         Unreviewed, build fix.
2286
2287         Reland the change in r115453 with a fix for chromium win.
2288         The logic in ChromiumDriver was busted on windows, causing DRT
2289         to get launched without the --test-shell flag.
2290
2291         * Scripts/webkitpy/layout_tests/port/chromium.py:
2292         (ChromiumDriver):
2293         (ChromiumDriver.__init__):
2294         (ChromiumDriver._wrapper_options):
2295         (ChromiumDriver.cmd_line):
2296         (ChromiumDriver._start):
2297         (ChromiumDriver.has_crashed):
2298         (ChromiumDriver.run_test):
2299         (ChromiumDriver.stop):
2300         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2301         (ChromiumDriverTest.setUp):
2302         (ChromiumDriverTest.test_stop):
2303         (ChromiumDriverTest.test_two_drivers.MockDriver.__init__):
2304         (ChromiumDriverTest.test_two_drivers):
2305
2306 2012-04-27  Peter Beverloo  <peter@chromium.org>
2307
2308         Add John Grabowski as a non-committer to committers.py
2309         https://bugs.webkit.org/show_bug.cgi?id=85080
2310
2311         Reviewed by Dirk Pranke.
2312
2313         He's being CC'ed often enough because of his involvement in the Chromium
2314         side of Chrome for Android upstreaming.
2315
2316         * Scripts/webkitpy/common/config/committers.py:
2317           add jrg@chromium.org
2318
2319 2012-04-27  Allan Sandfeld Jensen  <allan.jensen@nokia.com>
2320
2321         [Qt] Fix minimal build.
2322         https://bugs.webkit.org/show_bug.cgi?id=85045
2323
2324         Reviewed by Tor Arne Vestbø.
2325
2326         Do not set conflicting values of ENABLE_XSLT.
2327
2328         * qmake/mkspecs/features/features.prf:
2329
2330 2012-04-27  Dominik Röttsches  <dominik.rottsches@linux.intel.com>
2331
2332         [EFL][DRT] EFL should sanitize pulseaudio module configuration too
2333         https://bugs.webkit.org/show_bug.cgi?id=85047
2334
2335         Reviewed by Dirk Pranke.
2336
2337         PulseAudio test setup sanitization shared between EFL and GTK.
2338
2339         * Scripts/webkitpy/layout_tests/port/efl.py:
2340         (EflPort):
2341         (EflPort.setup_test_run):
2342         (EflPort.clean_up_test_run):
2343         * Scripts/webkitpy/layout_tests/port/gtk.py:
2344         (GtkPort): Moving pulseaudio unload/restore out to shared class.
2345         * Scripts/webkitpy/layout_tests/port/pulseaudio_sanitizer.py: Added.
2346         (PulseAudioSanitizer):
2347         (PulseAudioSanitizer._unload_pulseaudio_module):
2348         (PulseAudioSanitizer._restore_pulseaudio_module):
2349
2350 2012-04-27  Dirk Pranke  <dpranke@chromium.org>
2351
2352         Unreviewed, revert r115453 - broke chromium win bots
2353
2354         * Scripts/webkitpy/layout_tests/port/chromium.py:
2355         (ChromiumPort._path_to_image_diff):
2356         (ChromiumDriver):
2357         (ChromiumDriver.__init__):
2358         (ChromiumDriver._wrapper_options):
2359         (ChromiumDriver.cmd_line):
2360         (ChromiumDriver.stop):
2361         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2362         (ChromiumDriverTest.setUp):
2363         (ChromiumDriverTest.test_stop):
2364         (ChromiumDriverTest.test_two_drivers):
2365         (ChromiumDriverTest.test_two_drivers.MockDriver.__init__):
2366
2367 2012-04-27  Ryosuke Niwa  <rniwa@webkit.org>
2368
2369         Chromium Windows Perf bots timeout due to not output
2370         https://bugs.webkit.org/show_bug.cgi?id=84940
2371
2372         Reviewed by Dirk Pranke.
2373
2374         Dirk and I investigated the issue on the bot but we couldn't figure out what is going wrong.
2375         Since we run-perf-tests don't need any of fancy feature printer provides, just use python's
2376         built-in logging module instead. Printing out to stdout and stderr seem to work so hopefully
2377         this will fix the issue on the bot.
2378
2379         * Scripts/run-perf-tests:
2380         * Scripts/webkitpy/performance_tests/perftest.py:
2381         (PerfTest.run):
2382         (PerfTest.run_failed):
2383         (PerfTest.parse_output):
2384         (PerfTest.output_statistics):
2385         (ChromiumStylePerfTest.parse_output):
2386         (PageLoadingPerfTest.run):
2387         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
2388         (MainTest.test_parse_output):
2389         (MainTest.test_parse_output_with_failing_line):
2390         (TestPageLoadingPerfTest):
2391         (TestPageLoadingPerfTest.test_run):
2392         (TestPageLoadingPerfTest.test_run_with_bad_output):
2393         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
2394         (PerfTestsRunner.__init__):
2395         (PerfTestsRunner._parse_args):
2396         (PerfTestsRunner.run):
2397         (PerfTestsRunner._upload_json):
2398         (PerfTestsRunner._print_status):
2399         (PerfTestsRunner._run_tests_set):
2400         (PerfTestsRunner._run_single_test):
2401         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
2402         (create_runner):
2403         (test_run_test_set):
2404         (test_run_test_set_kills_drt_per_run):
2405         (test_run_test_pause_before_testing):
2406         (test_run_test_set_for_parser_tests):
2407         (test_run_test_set_with_json_output):
2408         (test_run_test_set_with_json_source):
2409         (test_run_test_set_with_multiple_repositories):
2410         (test_upload_json):
2411         (test_parse_args):
2412
2413 2012-04-27  Dirk Pranke  <dpranke@chromium.org>
2414
2415         [chromium] use "drt-style" output, not "test-shell-style" output, on mac and linux DRT
2416         https://bugs.webkit.org/show_bug.cgi?id=84917
2417
2418         Reviewed by Ojan Vafai.
2419
2420         This change changes the chromium new-run-webkit-tests
2421         implementation to use the code in WebKitDriver and ServerProcess
2422         by default on Mac and Linux, instead of using the old-style
2423         Chromium/test_shell style of output.
2424
2425         We will still use the test_shell style on Windows for now
2426         and you can still get the old style of behavior with
2427         new-run-webkit-tests --additional-drt-flag --test-shell
2428
2429         * Scripts/webkitpy/layout_tests/port/chromium.py:
2430         (ChromiumDriver):
2431         (ChromiumDriver.__init__):
2432         (ChromiumDriver._wrapper_options):
2433         (ChromiumDriver.cmd_line):
2434         (ChromiumDriver._start):
2435         (ChromiumDriver.has_crashed):
2436         (ChromiumDriver.run_test):
2437         (ChromiumDriver.stop):
2438         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2439         (ChromiumDriverTest.setUp):
2440         (ChromiumDriverTest.test_stop):
2441         (ChromiumDriverTest.test_two_drivers.MockDriver.__init__):
2442         (ChromiumDriverTest.test_two_drivers):
2443
2444 2012-04-27  Dirk Pranke  <dpranke@chromium.org>
2445
2446         webkitpy: fix race in unit tests for manager_worker_broker
2447         https://bugs.webkit.org/show_bug.cgi?id=84925
2448
2449         Reviewed by Ojan Vafai.
2450
2451         I have been intermittently seeing hangs when running
2452         test-webkitpy on Mac SL; it appears that there's a bug (?) in
2453         the multiprocessing module where queue references are getting
2454         leaked or not cleaned up somehow. By explicitly closing the
2455         queues, things seem to be happy, so this patch adds a cleanup()
2456         method to the message broker and modifies the manager (and the
2457         unit tests) to call it. This may get rid of the intermittent
2458         error on shutdown on Linux I've seen as well.
2459
2460         * Scripts/webkitpy/layout_tests/controllers/manager.py:
2461         (Manager._run_tests):
2462         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
2463         (_Broker.__del__):
2464         (_Broker):
2465         (_Broker.cleanup):
2466         (_BrokerConnection.cleanup):
2467         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py:
2468         (_TestsMixin.test_name):
2469         (_TestsMixin.test_cancel):
2470         (_TestsMixin.test_done):
2471         (_TestsMixin.test_unknown_message):
2472
2473 2012-04-27  Gavin Peters  <gavinp@chromium.org>
2474
2475         Add new ENABLE_LINK_PRERENDER define to control the Prerendering API
2476         https://bugs.webkit.org/show_bug.cgi?id=84871
2477
2478         Reviewed by Adam Barth.
2479
2480         Prerendering is currently covered by the ENABLE_LINK_PREFETCH macro, but the new Prerendering
2481         API separates it from prefetching.  Having separate include guards lets ports enable prefetching,
2482         a relatively easy change, without needing to build the infrastructure for prerendering, which
2483         is considerably more complicated.
2484
2485 2012-04-27  Philippe Normand  <pnormand@igalia.com>
2486
2487         Unreviewed, GTK gardening.
2488
2489         * Scripts/run-gtk-tests:
2490         (TestRunner): Skip flaky WebKit2 API test.
2491
2492 2012-04-27  Philippe Normand  <pnormand@igalia.com>
2493
2494         Unreviewed, GTK gardening.
2495
2496         * Scripts/run-gtk-tests:
2497         (TestRunner): Fix another typo. So easy to get confused by a Test
2498         name and the actual name of the executable...
2499
2500 2012-04-27  Philippe Normand  <pnormand@igalia.com>
2501
2502         Unreviewed, GTK gardening.
2503
2504         * Scripts/run-gtk-tests:
2505         (TestRunner): Skip failing WebKit2 API test.
2506
2507 2012-04-27  Christophe Dumez  <christophe.dumez@intel.com>
2508
2509         [EFL] DRT should support LayoutTestController's dumpResourceLoadCallbacks()
2510         https://bugs.webkit.org/show_bug.cgi?id=81891
2511
2512         Reviewed by Philippe Normand.
2513
2514         Implement support for LayoutTestController's
2515         dumpResourceLoadCallbacks() in EFL's DumpRenderTree by catching new
2516         ewk_view signals about resource loading and print out the expected
2517         information. This allows for several test cases to be removed from the
2518         skip list.
2519
2520         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
2521         (DumpRenderTreeChrome::createView):
2522         (descriptionSuitableForTestResult):
2523         (DumpRenderTreeChrome::onResponseReceived):
2524         (DumpRenderTreeChrome::onResourceLoadFinished):
2525         (DumpRenderTreeChrome::onResourceLoadFailed):
2526         (DumpRenderTreeChrome::onNewResourceRequest):
2527         * DumpRenderTree/efl/DumpRenderTreeChrome.h:
2528         (DumpRenderTreeChrome):
2529
2530 2012-04-27  Philippe Normand  <pnormand@igalia.com>
2531
2532         Unreviewed, GTK typo fix.
2533
2534         * Scripts/run-gtk-tests:
2535         (TestRunner): Fix-up skipped WebKit2 API test name TestWKConnection.
2536
2537 2012-04-26  Carlos Garcia Campos  <cgarcia@igalia.com>
2538
2539         [SOUP] Add a way to register custom uri schemes in WebKit2
2540         https://bugs.webkit.org/show_bug.cgi?id=84130
2541
2542         Reviewed by Martin Robinson.
2543
2544         * Scripts/webkitpy/style/checkers/cpp.py:
2545         (check_identifier_name_in_declaration): Add webkit_soup prefix to
2546         the list of exceptions to the underscores in identifiers rule.
2547
2548 2012-04-26  Ryosuke Niwa  <rniwa@webkit.org>
2549
2550         Extract PerfTestFactory
2551         https://bugs.webkit.org/show_bug.cgi?id=83944
2552
2553         Reviewed by Dirk Pranke.
2554
2555         Extracted PerfTestFactory out of PerfTest. This class creates appropriate PerfTest object given
2556         a test name and a path. Also removed dirname from the constructors of PerfTest and its subclasses.
2557
2558         Also added simple test cases to sanity check PerfTestFactory.
2559
2560         * Scripts/webkitpy/performance_tests/perftest.py:
2561         (PerfTest.__init__):
2562         (PerfTest.test_name):
2563         (ChromiumStylePerfTest.__init__):
2564         (PageLoadingPerfTest.__init__):
2565         (PageLoadingPerfTest.run):
2566         (PerfTestFactory):
2567         (PerfTestFactory.create_perf_test):
2568         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
2569         (MainTest.test_parse_output):
2570         (MainTest.test_parse_output_with_failing_line):
2571         (TestPageLoadingPerfTest.test_run):
2572         (TestPageLoadingPerfTest.test_run_with_bad_output):
2573         (TestPerfTestFactory):
2574         (TestPerfTestFactory.test_regular_test):
2575         (TestPerfTestFactory.test_inspector_test):
2576         (TestPerfTestFactory.test_page_loading_test):
2577         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
2578         (PerfTestsRunner):
2579         (PerfTestsRunner._collect_tests):
2580         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
2581         (run_test):
2582         (_tests_for_runner):
2583
2584 2012-04-26  Dirk Pranke  <dpranke@chromium.org>
2585
2586         nrwt: clean up server process, webkit driver so chromium can use it
2587         https://bugs.webkit.org/show_bug.cgi?id=84910
2588
2589         Reviewed by Ojan Vafai.
2590
2591         This change moves the "sample a process" logic out of
2592         server_process.py and into a port-specific class (where really
2593         only the mac has an implementation), and also preemptively kills
2594         DRT when a test times out in WebKitDriver (rather than waiting through the
2595         additional delays caused by calling stop() when we would want to
2596         restart the driver generically in worker.py).
2597
2598         These changes will make it possible for the chromium port to
2599         switch over to the stock WebKitDriver implementation, at least
2600         on mac and linux.
2601
2602         * Scripts/webkitpy/layout_tests/port/base.py:
2603         (Port.sample_process):
2604         * Scripts/webkitpy/layout_tests/port/mac.py:
2605         (MacPort.sample_process):
2606         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
2607         (test_helper_fails_to_stop):
2608         (test_sample_process):
2609         (test_sample_process.logging_run_command):
2610         (test_sample_process_throws_exception):
2611         (test_sample_process_throws_exception.throwing_run_command):
2612         * Scripts/webkitpy/layout_tests/port/server_process.py:
2613         (ServerProcess._log):
2614         (ServerProcess._handle_timeout):
2615         (ServerProcess.stop):
2616         (ServerProcess):
2617         (ServerProcess.kill): Here we add a method to immediately stop
2618         the process rather than trying to shut it down cleanly.
2619         * Scripts/webkitpy/layout_tests/port/server_process_unittest.py:
2620         (TestServerProcess.test_broken_pipe):
2621         * Scripts/webkitpy/layout_tests/port/webkit.py:
2622         (WebKitDriver.run_test): Fix an issue where we weren't passing
2623         along any per-test args (only needed for Chromium, but still).
2624         Also, kill the driver immediately if we time out a test.
2625
2626 2012-04-26  Benjamin Poulain  <bpoulain@apple.com>
2627
2628         ObjcClass::methodsNamed() can leak if buffer is dynamically allocated
2629         https://bugs.webkit.org/show_bug.cgi?id=84668
2630
2631         Reviewed by Alexey Proskuryakov.
2632
2633         Extend ObjCPlugin to support the new layout tests of the Objective-C bridge.
2634
2635         * DumpRenderTree/mac/ObjCPlugin.m:
2636         (+[ObjCPlugin isSelectorExcludedFromWebScript:]):
2637         (+[ObjCPlugin webScriptNameForSelector:]):
2638         (-[ObjCPlugin methodMappedToLongName]):
2639         (-[ObjCPlugin testConversionColon:]):
2640         (-[ObjCPlugin _:]):
2641
2642 2012-04-26  Dimitri Glazkov  <dglazkov@chromium.org>
2643
2644         Unreviewed, rolling out r115340.
2645         http://trac.webkit.org/changeset/115340
2646         https://bugs.webkit.org/show_bug.cgi?id=83048
2647
2648         Does not work with Windows.
2649
2650         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py:
2651         (SingleTestRunner._run_reftest):
2652         * Scripts/webkitpy/layout_tests/controllers/test_result_writer.py:
2653         (copy_file):
2654         * Scripts/webkitpy/layout_tests/port/base.py:
2655         (Port._parse_reftest_list):
2656         (Port.reference_files):
2657         (Port._real_tests):
2658         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2659         (PortTest.test_reference_files):
2660         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2661         (ChromiumDriverTest.test_write_command_and_read_line):
2662
2663 2012-04-26  Christophe Dumez  <christophe.dumez@intel.com>
2664
2665         [EFL] Enable VIDEO_TRACK feature
2666         https://bugs.webkit.org/show_bug.cgi?id=84830
2667
2668         Reviewed by Gustavo Noronha Silva.
2669
2670         Enable VIDEO_TRACK by default on EFL port whenever build-webkit is
2671         used.
2672
2673         * Scripts/build-webkit:
2674
2675 2012-04-08  Robert Hogan  <robert@webkit.org>
2676
2677         new-webkit-run-tests: handle ref tests from the CSS test suite
2678         https://bugs.webkit.org/show_bug.cgi?id=83048
2679
2680         Reviewed by Ryosuke Niwa.
2681
2682         This allows new-run-webkit-tests to run reference tests created for the CSS test suite. It removes a big swathe
2683         of pixel and text results that are no longer required now that the reference results are used and adds in the support
2684         files from the CSS test suite that allow the ref tests to run.
2685
2686         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py:
2687         (SingleTestRunner._run_reftest): Assert that the reference result exists. Previously the test would
2688          run without it and report a pass!
2689         * Scripts/webkitpy/layout_tests/controllers/test_result_writer.py:
2690         (copy_file): ensure the output directory exists, it may not if the reference test is
2691           contained in a folder that contains nothing but reference tests.
2692         * Scripts/webkitpy/layout_tests/port/base.py:
2693         (Port.available_reference_files): Add as a helper. Handle embedded relative paths, e.g: some/../path
2694         (Port.reference_files): Only return reference results in the reftest list that actually exist on
2695           disk. Also, don't ignore potential reference results on disk that are not in the reftest list, and
2696           don't ignore them just because we have a reftest list for the test's directory.
2697         (Port._real_tests): ignore anything that looks like a test in a 'support' directory, this semantic
2698           is used by the CSS 2.1 test suite.
2699         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2700         (PortTest.test_reference_files_when_reftestlist_and_reference_file_present): Ensure that a test with a reference result
2701           on disk but no reference result listed in the directory's reftest list picks up its reference result.
2702         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2703         (ChromiumDriverTest.test_strip_uri): Test that some/test.html and some/other/../test.html are treated
2704          as the same file.
2705
2706 2012-04-26  Carlos Garcia Campos  <cgarcia@igalia.com>
2707
2708         Unreviewed. Fix GTK+ build after r115314.
2709
2710         For some reason it fails while bulding GtkVersioning for WTR, but
2711         it didn't fail neither in EWS nor in my laptop. We don't really
2712         need GtkVersioning in WTR since it already depends on GTK+3
2713         unconditionally.
2714
2715         * WebKitTestRunner/GNUmakefile.am: Remove GtkVersioning.c from
2716         compilation.
2717         * WebKitTestRunner/gtk/EventSenderProxyGtk.cpp:
2718         (WTR::EventSenderProxy::createMouseButtonEvent): Use GTK+ 3 API
2719         directly instead of getDefaultGDKPointerDevice() helper.
2720         (WTR::EventSenderProxy::keyDown): Ditto.
2721         (WTR::EventSenderProxy::mouseMoveTo): Ditto.
2722
2723 2012-04-24  Ojan Vafai  <ojan@chromium.org>
2724
2725         Show flakiness dashboard data in garden-o-matic
2726         https://bugs.webkit.org/show_bug.cgi?id=83716
2727
2728         Reviewed by Dimitri Glazkov.
2729
2730         Put an iframe below the expected/actual results in the Results view.
2731         Size the iframe to it's height.
2732
2733         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/garden-o-matic.html:
2734         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/run-unittests.html:
2735         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui.js:
2736         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results.js:
2737         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results_unittests.js:
2738         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui_unittests.js:
2739         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/styles/results.css:
2740
2741 2012-04-26  Sudarsana Nagineni  <sudarsana.nagineni@linux.intel.com>
2742
2743         [EFL] [DRT] LayoutTestController needs implementation of setJavaScriptProfilingEnabled
2744         https://bugs.webkit.org/show_bug.cgi?id=84576
2745
2746         Reviewed by Martin Robinson.
2747
2748         Add missing implementation setJavaScriptProfilingEnabled to EFL's
2749         LayoutTestController in order to unskip tests rely on this setting.
2750
2751         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
2752         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
2753         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp:
2754         (LayoutTestController::setJavaScriptProfilingEnabled):
2755
2756 2012-04-26  Xianzhu Wang  <wangxianzhu@chromium.org>
2757
2758         Add Jing Zhao, Min Qin, Tien-Ren Chen and Hao Zheng into contributors list.
2759
2760         Not reviewed.
2761
2762         * Scripts/webkitpy/common/config/committers.py:
2763
2764 2012-04-26  Carlos Garcia Campos  <cgarcia@igalia.com>
2765
2766         [GTK] Test TestWebKitAPI/WebKit2/TestNewFirstVisuallyNonEmptyLayout times out
2767         https://bugs.webkit.org/show_bug.cgi?id=84957
2768
2769         Reviewed by Martin Robinson.
2770
2771         * Scripts/run-gtk-tests:
2772         (TestRunner): Unskip TestNewFirstVisuallyNonEmptyLayout and
2773         TestNewFirstVisuallyNonEmptyLayoutForImages.
2774         * TestWebKitAPI/gtk/PlatformWebViewGtk.cpp:
2775         (TestWebKitAPI::PlatformWebView::PlatformWebView): Show the
2776         toplevel window containing the web view to make sure events are
2777         received.
2778
2779 2012-04-26  Carlos Garcia Campos  <cgarcia@igalia.com>
2780
2781         Unreviewed. Add bug numbers to skipped tests.
2782
2783         * Scripts/run-gtk-tests:
2784         (TestRunner):
2785
2786 2012-04-26  Carlos Garcia Campos  <cgarcia@igalia.com>
2787
2788         [GTK] Build and run TestWebKitAPI WebKit2 unit tests
2789         https://bugs.webkit.org/show_bug.cgi?id=84446
2790
2791         Reviewed by Philippe Normand.
2792
2793         * Scripts/run-gtk-tests:
2794         (TestRunner): Add TestWebKitAPI/WebKit2 to the list of directories
2795         containing tests. Unskip FunctionalTest.RefCountedStorage, since
2796         we are now using a custom main that initializes WTF threading.
2797         (TestRunner._setup_testing_environment): Set environment variables
2798         required to run WebKit2 tests.
2799         * TestWebKitAPI/GNUmakefile.am:
2800         * TestWebKitAPI/JavaScriptTest.cpp: Use
2801         JavaScriptCore/JSContextRef.h instead of
2802         JavaScriptCore/JavaScriptCore.h since it includes JSStringRefCF.h
2803         unconditionally.
2804         * TestWebKitAPI/PlatformWebView.h: Define PlatformWKView and
2805         PlatformWindow for GTK.
2806         * TestWebKitAPI/Tests/WebKit2/EvaluateJavaScript.cpp: Remove
2807         unneded header that breaks the build of GTK port.
2808         * TestWebKitAPI/Tests/WebKit2/PreventEmptyUserAgent.cpp: Use
2809         JavaScriptCore/JSContextRef.h instead of
2810         JavaScriptCore/JavaScriptCore.h since it includes JSStringRefCF.h
2811         unconditionally.
2812         * TestWebKitAPI/Tests/WebKit2/WKPreferences.cpp:
2813         (TestWebKitAPI::TEST): Add expected values for GTK+ port.
2814         * TestWebKitAPI/gtk/InjectedBundleControllerGtk.cpp:
2815         (TestWebKitAPI):
2816         (TestWebKitAPI::InjectedBundleController::platformInitialize):
2817         * TestWebKitAPI/gtk/PlatformUtilitiesGtk.cpp: Added.
2818         (TestWebKitAPI::Util::checkTestFinished):
2819         (TestWebKitAPI::Util::run):
2820         (TestWebKitAPI::Util::sleep):
2821         (TestWebKitAPI::Util::getFilenameFromEnvironmentVariableAsUTF8):
2822         (TestWebKitAPI::Util::createInjectedBundlePath):
2823         (TestWebKitAPI::Util::createURLForResource):
2824         (TestWebKitAPI::Util::URLForNonExistentResource):
2825         (TestWebKitAPI::Util::isKeyDown):
2826         * TestWebKitAPI/gtk/PlatformWebViewGtk.cpp: Added.
2827         (TestWebKitAPI::PlatformWebView::PlatformWebView):
2828         (TestWebKitAPI::PlatformWebView::~PlatformWebView):
2829         (TestWebKitAPI::PlatformWebView::page):
2830         (TestWebKitAPI::PlatformWebView::resizeTo):
2831         (TestWebKitAPI::doKeyStroke):
2832         (TestWebKitAPI::PlatformWebView::simulateSpacebarKeyPress):
2833         (TestWebKitAPI::PlatformWebView::simulateAltKeyPress):
2834         (TestWebKitAPI::doMouseButtonEvent):
2835         (TestWebKitAPI::PlatformWebView::simulateRightClick):
2836         (TestWebKitAPI::PlatformWebView::simulateMouseMove):
2837         * TestWebKitAPI/gtk/main.cpp:
2838         (main):
2839
2840 2012-04-26  Carlos Garcia Campos  <cgarcia@igalia.com>
2841
2842         [GTK] run-gtk-tests: Use a timeout per test instead of a global timeout
2843         https://bugs.webkit.org/show_bug.cgi?id=84695
2844
2845         Reviewed by Philippe Normand.
2846
2847         It also adds a command line option to be able to pass a custom
2848         timeout value, instead of hard-coding it. The default timeout is 10
2849         seconds if no other value is passed to the script.
2850
2851         * Scripts/run-gtk-tests:
2852         (TestTimeout): Exception raised when a test times out.
2853         (TestRunner._get_child_pid_from_test_output): Helper function to
2854         get the pid of the running test from gtester output.
2855         (TestRunner._kill_process): Helper funtion ot kill a process
2856         ignoring exceptions if the process is already died.
2857         (TestRunner._run_test_command): Run the test command raising
2858         TestTimeout exception if the test doesn't finish before the given
2859         timeout in seconds.
2860         (TestRunner._run_test_command.alarm_handler): Alarm handler that
2861         raises TestTimeout exception.
2862         (TestRunner._run_test_glib): Use _run_test_command() to run the test.
2863         (TestRunner._run_test_google): Ditto.
2864         (TestRunner.run_tests): Add current test to the list of timed out
2865         test if TestTimeout exception is raised. Show the list of timed
2866         out test at the end.
2867
2868 2012-04-26  Sudarsana Nagineni  <sudarsana.nagineni@linux.intel.com>
2869
2870         [EFL] [DRT] LayoutTestController needs implementation of addOriginAccessWhitelistEntry and removeOriginAccessWhitelistEntry
2871         https://bugs.webkit.org/show_bug.cgi?id=83880
2872
2873         Reviewed by Antonio Gomes.
2874
2875         Add missing implementation addOriginAccessWhitelistEntry and
2876         removeOriginAccessWhitelistEntry to EFL's LayoutTestController
2877         in order to unskip tests which rely on this.
2878
2879         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
2880         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
2881         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp:
2882         (LayoutTestController::addOriginAccessWhitelistEntry):
2883         (LayoutTestController::removeOriginAccessWhitelistEntry):
2884
2885 2012-04-26  Carlos Garcia Campos  <cgarcia@igalia.com>
2886
2887         [GTK] Add downloads support to MiniBrowser
2888         https://bugs.webkit.org/show_bug.cgi?id=73662
2889
2890         Reviewed by Philippe Normand.
2891
2892         * MiniBrowser/gtk/BrowserDownloadsBar.c: Added.
2893         (browserDownloadsBarChildRemoved): Close the downloads bar if
2894         there aren't active downloads.
2895         (browserDownloadsBarResponse): Destroy the downloads bar when
2896         closed.
2897         (browser_downloads_bar_init):
2898         (browser_downloads_bar_class_init):
2899         (browser_downloads_bar_new): Create a new downloads bar.
2900         (actionButtonClicked): Cancel the download if cancel button was
2901         clicked, or open the downloaded file if the download finished.
2902         (browserDownloadFinalize):
2903         (browser_download_init):
2904         (browser_download_class_init):
2905         (downloadReceivedResponse): Update download message using the
2906         response URI.
2907         (remainingTime): Helper function to get human readable remaining
2908         time.
2909         (downloadProgress): Update downloads progress.
2910         (downloadReceivedData):
2911         (downloadFinished): Update download message and turn download
2912         button into open.
2913         (downloadFailed): Show error message or destroy the download UI if
2914         it was cancelled.
2915         (browserDownloadNew): Create a new download widget for the given
2916         WebKitDownload.
2917         (browser_downloads_bar_add_download): Add download widget to the
2918         downloads bar.
2919         * MiniBrowser/gtk/BrowserDownloadsBar.h: Added.
2920         * MiniBrowser/gtk/BrowserWindow.c:
2921         (downloadStarted): Create a download widget and add it to the
2922         downloads bar when a new download is started.
2923         (browserWindowConstructed): Connect to
2924         WebKitWebContext::download-started signal.
2925         * MiniBrowser/gtk/GNUmakefile.am: Add new files to compilation.
2926
2927 2012-04-26  Carlos Garcia Campos  <cgarcia@igalia.com>
2928
2929         Unreviewed. Skip unit test failing in debug bot.
2930
2931         * Scripts/run-gtk-tests:
2932         (TestRunner):
2933
2934 2012-04-26  Christophe Dumez  <christophe.dumez@intel.com>
2935
2936         [EFL][DRT] "resource,request,willsend" signal needs to provide more information
2937         https://bugs.webkit.org/show_bug.cgi?id=84670
2938
2939         Reviewed by Andreas Kling.
2940
2941         Handle in EFL's DumpRenderTree the new Ewk_Frame_Resource_Messages now
2942         passed with the "resource,request,willsend" signal and dump
2943         information about the network request and the possible network
2944         response if LayoutTestController's dumpResourceLoadCallbacks() returns
2945         true.
2946
2947         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
2948         (pathSuitableForTestResult):
2949         (urlSuitableForTestResult):
2950         (descriptionSuitableForTestResult):
2951         (DumpRenderTreeChrome::onWillSendRequest):
2952
2953 2012-04-25  Csaba Osztrogonác  <ossy@webkit.org>
2954
2955         [Qt][WK2] Unreviewed, rolling out r113703 because of crashing tests.
2956         http://trac.webkit.org/changeset/113703
2957         https://bugs.webkit.org/show_bug.cgi?id=83024
2958
2959         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
2960         * Tools.pro:
2961         * qmake/mkspecs/features/features.prf:
2962         * qmake/mkspecs/features/functions.prf:
2963
2964 2012-04-25  Maciej Stachowiak  <mjs@apple.com>
2965
2966         Teach LayoutTestHelper to correctly restore the color profile when it's a factory profile
2967         https://bugs.webkit.org/show_bug.cgi?id=84931
2968
2969         Reviewed by Filip Pizlo.
2970
2971         * DumpRenderTree/mac/LayoutTestHelper.m:
2972         (installLayoutTestColorProfile): Account for the possibility of the default
2973         profile being a factory profile.
2974
2975 2012-04-25  Sheriff Bot  <webkit.review.bot@gmail.com>
2976
2977         Unreviewed, rolling out r115240.
2978         http://trac.webkit.org/changeset/115240
2979         https://bugs.webkit.org/show_bug.cgi?id=84928
2980
2981         It broke everything (Requested by Ossy_HOME on #webkit).
2982
2983         * Scripts/webkitpy/layout_tests/port/gtk.py:
2984         (GtkDriver):
2985         (GtkDriver._start):
2986         (GtkDriver._start.x_filter):
2987         (GtkDriver.stop):
2988         (GtkPort._driver_class):
2989         * Scripts/webkitpy/layout_tests/port/qt.py:
2990         (QtPort._build_driver):
2991         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py: Removed.
2992
2993 2012-04-25  Dirk Pranke  <dpranke@chromium.org>
2994
2995         [chromium] change DRT's "DRT" output to match the other ports
2996         https://bugs.webkit.org/show_bug.cgi?id=84904
2997
2998         Reviewed by Kent Tamura.
2999
3000         Since we run Chromium's DRT in --test-shell mode, the
3001         non-test-shell output has never been really beaten on to make
3002         sure it's right. This fixes a few issues and will be tested
3003         with upcoming NRWT changes that'll land in a separate patch.
3004         In particular we should never output framing text like "\n"
3005         or "#EOF" outside of something outside of TestEventPrinter,
3006         and we weren't handling audio output properly at all.
3007
3008         * DumpRenderTree/chromium/TestEventPrinter.cpp:
3009         (DRTPrinter):
3010         (TestShellPrinter):
3011         (DRTPrinter::handleTextFooter):
3012         (DRTPrinter::handleAudioFooter):
3013         (DRTPrinter::handleTestFooter):
3014         (TestShellPrinter::handleAudioFooter):
3015         * DumpRenderTree/chromium/TestEventPrinter.h:
3016         (TestEventPrinter):
3017         * DumpRenderTree/chromium/TestShell.cpp:
3018         (TestShell::TestShell):
3019         (TestShell::dump):
3020
3021 2012-04-25  Joshua Bell  <jsbell@chromium.org>
3022
3023         Unreviewed: Adding Alec Flett as a contributor (non-committer).
3024
3025         * Scripts/webkitpy/common/config/committers.py:
3026
3027 2012-04-25  Abhishek Arya  <inferno@chromium.org>
3028
3029         Mark myself as a reviewer.
3030         https://bugs.webkit.org/show_bug.cgi?id=84902
3031
3032         Reviewed by Dimitri Glazkov.
3033
3034         * Scripts/webkitpy/common/config/committers.py:
3035
3036 2012-04-25  Kristóf Kosztyó  <kkristof@inf.u-szeged.hu>
3037
3038         [Qt][NRWT] Run each DRT in it's own xvfb
3039         https://bugs.webkit.org/show_bug.cgi?id=77335
3040
3041         Reviewed by Dirk Pranke.
3042
3043         * Scripts/webkitpy/layout_tests/port/gtk.py:
3044         (GtkPort._driver_class):
3045         * Scripts/webkitpy/layout_tests/port/qt.py:
3046         (QtPort._driver_class):
3047         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py: Added.
3048         (XvfbDriver):
3049         (XvfbDriver._start):
3050         (XvfbDriver._start.x_filter):
3051         (XvfbDriver.stop):
3052
3053 2012-04-25  Sheriff Bot  <webkit.review.bot@gmail.com>
3054
3055         Unreviewed, rolling out r115222.
3056         http://trac.webkit.org/changeset/115222
3057         https://bugs.webkit.org/show_bug.cgi?id=84874
3058
3059         Caused test regressions on the EFL build (Requested by rakuco
3060         on #webkit).
3061
3062         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp:
3063         (LayoutTestController::addOriginAccessWhitelistEntry):
3064         (LayoutTestController::removeOriginAccessWhitelistEntry):
3065
3066 2012-04-25  Sudarsana Nagineni  <sudarsana.nagineni@linux.intel.com>
3067
3068         [EFL] [DRT] LayoutTestController needs implementation of addOriginAccessWhitelistEntry and removeOriginAccessWhitelistEntry
3069         https://bugs.webkit.org/show_bug.cgi?id=83880
3070
3071         Reviewed by Antonio Gomes.
3072
3073         Add missing implementation addOriginAccessWhitelistEntry and
3074         removeOriginAccessWhitelistEntry to EFL's LayoutTestController
3075         in order to unskip tests which rely on this.
3076
3077         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp:
3078         (LayoutTestController::addOriginAccessWhitelistEntry):
3079         (LayoutTestController::removeOriginAccessWhitelistEntry):
3080
3081 2012-04-25  Thiago Marcos P. Santos  <thiago.santos@intel.com>
3082
3083         [EFL] [DRT] Reset text zoom factor before running a test
3084         https://bugs.webkit.org/show_bug.cgi?id=84861
3085
3086         Reviewed by Antonio Gomes.
3087
3088         This fix eliminates most of the current flakiness on EFL build
3089         bots, although some CSS tests will now need rebasiling.
3090
3091         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
3092         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
3093
3094 2012-04-25  Milian Wolff  <milian.wolff@kdab.com>
3095
3096         [Qt] add LayoutTestController::setPrinting support to Qt unit tests
3097         https://bugs.webkit.org/show_bug.cgi?id=84246
3098
3099         Reviewed by Simon Hausmann.
3100
3101         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
3102         (WebCore::DumpRenderTree::dump):
3103
3104 2012-04-18  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3105
3106         [Qt] Make the web view's url property follow the active url
3107
3108         https://bugs.webkit.org/show_bug.cgi?id=77554
3109
3110         Update  the location bar in the minibrowser to behave
3111         a bit more like normal browsers in terms of when the url will change
3112         and how active focus is handled.
3113
3114         Reviewed by Simon Hausmann.
3115
3116         * MiniBrowser/qt/qml/BrowserWindow.qml:
3117
3118 2012-04-25  Philippe Normand  <pnormand@igalia.com>
3119
3120         Webkit build fails due to missing gstreamer include file on Kubuntu 8.04
3121         https://bugs.webkit.org/show_bug.cgi?id=81913
3122
3123         Reviewed by Tor Arne Vestbø.
3124
3125         * qmake/mkspecs/features/features.prf: Check for gstreamer-0.10
3126         and gstreamer-plugins-base >= 0.10.30 just like the GTK port
3127         does. 0.10.27 is unfortunately not recent enough to get the media
3128         player working reliably.
3129
3130
3131 2012-04-25  Xiaobo Wang  <xbwang@torchmobile.com.cn>
3132
3133         [BlackBerry] Run tests in '/http/tests/local' with local URI
3134         https://bugs.webkit.org/show_bug.cgi?id=84820
3135
3136         Reviewed by Nikolas Zimmermann.
3137
3138         Currently we run all tests under folder "/http/tests/" as HTTP
3139         tests (http://...). However in WebKit ORWT and NRWT scripts, tests in
3140         'http/tests/local' are run with local path (like '/developer/LayoutTests
3141         /http/tests/local/absolute-url-strip-whitespace.html').
3142         With this patch we can keep the same behaviour as other portings.
3143
3144         * DumpRenderTree/blackberry/DumpRenderTree.cpp:
3145         (BlackBerry::WebKit::DumpRenderTree::isHTTPTest):
3146
3147 2012-04-25  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
3148
3149         [EFL][DRT] Reset DRT settings to their default values after a testcase was executed.
3150         https://bugs.webkit.org/show_bug.cgi?id=84709
3151
3152         Reviewed by Eric Seidel.
3153
3154         Not all the settings that can be changed by layout testcases through a JS API were
3155         reset in DumpRenderTreeChrome::resetDefaultsToConsistentValues().
3156
3157         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
3158         (defaultEditingBehavior): returns editing behavior depending on current platform.
3159         (DumpRenderTreeChrome::resetDefaultsToConsistentValues): reset some missed DRT settings
3160
3161 2012-04-24  Dirk Pranke  <dpranke@chromium.org>
3162
3163         nrwt: add ASAN handling to chromium crash logs
3164         https://bugs.webkit.org/show_bug.cgi?id=84664
3165
3166         Reviewed by Ojan Vafai.
3167
3168         * Scripts/webkitpy/layout_tests/port/chromium.py:
3169         (ChromiumPort._get_crash_log):
3170
3171 2012-04-24  Dirk Pranke  <dpranke@chromium.org>
3172
3173         webkit-patch rebaseline-test is retrieving stale expectations
3174         https://bugs.webkit.org/show_bug.cgi?id=84762
3175
3176         Reviewed by Ojan Vafai.
3177
3178         webkit-patch rebaseline-expectations is fetching files directly
3179         from the layout-test-results directory on the buildbots, rather
3180         than fetching them from the archive. This is problematic because
3181         we don't clobber the directory after each run and so you can end
3182         up fetching stale failing results.
3183
3184         This change temporarily changes the code to fetch the zip file
3185         instead -- making things much slower -- until I can confer w/
3186         abarth and ojan over the best way to fix this for the long-term.
3187
3188         * Scripts/webkitpy/tool/commands/rebaseline.py:
3189         (RebaselineTest.__init__):
3190         (RebaselineTest._zip_file_set):
3191         (RebaselineTest):
3192         (RebaselineTest._fetch_baseline):
3193         (RebaselineTest._rebaseline_test):
3194         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
3195         (TestRebaseline.stub_rebaseline_test_command_and_tool):
3196         (TestRebaseline.stub_rebaseline_test_command_and_tool.FakeZipFileSet):
3197         (TestRebaseline.stub_rebaseline_test_command_and_tool.FakeZipFileSet.read):
3198         (TestRebaseline.test_rebaseline_updates_expectations_file_noop):
3199         (test_rebaseline_updates_expectations_file):
3200         (test_rebaseline_test):
3201         (test_rebaseline_and_copy_test):
3202         (test_rebaseline_and_copy_test_no_existing_result):
3203         (test_rebaseline_and_copy_test_with_lion_result):
3204         (test_rebaseline_and_copy_no_overwrite_test):
3205
3206 2012-04-24  Ojan Vafai  <ojan@chromium.org>
3207
3208         Improve code for embedding flakiness dashboard as an iframe
3209         https://bugs.webkit.org/show_bug.cgi?id=84782
3210
3211         Reviewed by Dirk Pranke.
3212
3213         -Include the horizontal scrollbar height in the measured height of the page.
3214         -Post a message everytime the height changes instead of in response to a queryContentHeight message.
3215         This way, the embedder doesn't need to poll.
3216
3217         No new tests. This is hard to test because it involves postMessage and iframes.
3218         Once we switch the dashboard tests over to qunit, we should be able to write an
3219         asyncTest for this.
3220
3221         * TestResultServer/static-dashboards/flakiness_dashboard.js:
3222         (generatePage):
3223         (appendHTML):
3224         (appendExpectations):
3225         (postHeightChangedMessage):
3226         (generatePageForIndividualTests.if):
3227         (generatePageForIndividualTests):
3228         * TestResultServer/static-dashboards/flakiness_dashboard_embedded.html: Added.
3229
3230 2012-04-16  Ojan Vafai  <ojan@chromium.org>
3231
3232         Replace garden-o-matic accordion UI with a panel UI
3233         https://bugs.webkit.org/show_bug.cgi?id=84081
3234
3235         Reviewed by Dimitri Glazkov.
3236
3237         This is a nicer UI to work with and makes it easier to embed the flakiness dashboard,
3238         since we can reuse the same iframe without moving it in the DOM.
3239
3240         Also, excise unnecessary jquery-isms.
3241
3242         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results.js:
3243         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results_unittests.js:
3244         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/styles/results.css:
3245
3246 2012-04-24  Peter Beverloo  <peter@chromium.org>
3247
3248         [Chromium] DumpRenderTree should depend on base.gyp:test_support_base
3249         https://bugs.webkit.org/show_bug.cgi?id=84729
3250
3251         Reviewed by Dirk Pranke.
3252
3253         This dependency is necessary for Android since it'll soon be
3254         calling InitAndroidOSPathStub() in it's platform support. The
3255         TestWebKitAPI and webkit_unit_tests already have this dependency.
3256
3257         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
3258
3259 2012-04-24  Ojan Vafai  <ojan@chromium.org>
3260
3261         Replace vague informative statement with the list of builders where the test isn't listed
3262         https://bugs.webkit.org/show_bug.cgi?id=84779
3263
3264         Reviewed by Dirk Pranke.
3265
3266         This is more informative and more likely to be noticed. With the old UI, people never
3267         read the line and if they did, they didn't really understand what it meant. Hopefully
3268         have the list of builders will be more clear.
3269
3270         * TestResultServer/static-dashboards/flakiness_dashboard.js:
3271         * TestResultServer/static-dashboards/flakiness_dashboard_tests.css:
3272         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
3273
3274 2012-04-24  Tommy Widenflycht  <tommyw@google.com>
3275
3276         MediaStream API: Create a new flag for PeerConnection
3277         https://bugs.webkit.org/show_bug.cgi?id=84723
3278
3279         Reviewed by Dimitri Glazkov.
3280
3281         To allow Chrome to make the GetUserMedia functionality available by default we
3282         need another flag for the PeerConnection related parts. This new flag is in
3283         addition to the general media stream flag, and both need to be enabled for
3284         PeerConnection object creation.
3285
3286         * DumpRenderTree/chromium/TestShell.cpp:
3287         (TestShell::TestShell):
3288
3289 2012-04-24  Joe Mason  <jmason@rim.com>
3290
3291         [BlackBerry] Revert broken changes to authentication dialog
3292         https://bugs.webkit.org/show_bug.cgi?id=80135
3293
3294         Reviewed by Antonio Gomes.
3295
3296         The previous patches from this bug caused an infinite loop when using digest auth;
3297         apparently they were only tested with basic.
3298
3299         * DumpRenderTree/blackberry/DumpRenderTree.cpp:
3300         * DumpRenderTree/blackberry/DumpRenderTreeBlackBerry.h:
3301         (WebCore):
3302
3303 2012-04-24  Zan Dobersek  <zandobersek@gmail.com>
3304
3305         [TestResultsServer] Add a unit test for changes in r114868
3306         https://bugs.webkit.org/show_bug.cgi?id=84713
3307
3308         Reviewed by Ojan Vafai.
3309
3310         Add a new unit test to test the isToTWebKit attribute of a BuilderGroup object
3311         that is created for either DEPS or ToT builder groups. The unit test covers changes
3312         made in r114868.
3313
3314         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
3315         (testBuilderGroupIsToTWebKitAttribute.g_handleBuildersListLoaded):
3316         (testBuilderGroupIsToTWebKitAttribute.assertEquals):
3317
3318 2012-04-24  Brady Eidson  <beidson@apple.com>
3319
3320         <rdar://problem/10120155> and https://bugs.webkit.org/show_bug.cgi?id=82664
3321          Need DOMWindow mechanism to supplement UserScripts for page cache notifications
3322
3323         Reviewed by Sam Weinig.
3324
3325         Add a new bundle test that stresses the basics of the WKBundleDOMWindowExtension API.
3326
3327         It makes sure new extensions can be added at the right times, and that all of the expected
3328         callbacks are made.
3329
3330         * TestWebKitAPI/Tests/WebKit2/DOMWindowExtensionBasic.cpp: Added.
3331         (TestWebKitAPI::didReceiveMessageFromInjectedBundle): Record all messages sent from the bundle.
3332         (TestWebKitAPI::TEST): Run a page through some basic exercises, then compared messages sent from the
3333            bundle to a set of known expected messages.
3334
3335         Listen to all the new client callbacks, create some extensions, and monitor their lifetimes:
3336         * TestWebKitAPI/Tests/WebKit2/DOMWindowExtensionBasic_Bundle.cpp: Added.
3337         (TestWebKitAPI::DOMWindowExtensionBasic::DOMWindowExtensionBasic):
3338         (TestWebKitAPI::DOMWindowExtensionBasic::frameLoadFinished):
3339         (TestWebKitAPI::DOMWindowExtensionBasic::sendExtensionStateMessage):
3340         (TestWebKitAPI::DOMWindowExtensionBasic::initialize):
3341         (TestWebKitAPI::DOMWindowExtensionBasic::didCreatePage):
3342         (TestWebKitAPI::DOMWindowExtensionBasic::updateExtensionStateRecord):
3343         (TestWebKitAPI::DOMWindowExtensionBasic::sendBundleMessage):
3344         (TestWebKitAPI::DOMWindowExtensionBasic::didCreateGlobalObjectForFrame):
3345         (TestWebKitAPI::DOMWindowExtensionBasic::willDisconnectDOMWindowExtensionFromGlobalObject):
3346         (TestWebKitAPI::DOMWindowExtensionBasic::didReconnectDOMWindowExtensionToGlobalObject):
3347         (TestWebKitAPI::DOMWindowExtensionBasic::willDestroyGlobalObjectForDOMWindowExtension):
3348         (TestWebKitAPI::didFinishLoadForFrameCallback):
3349         (TestWebKitAPI::didCreateGlobalObjectForFrameCallback):
3350         (TestWebKitAPI::willDisconnectDOMWindowExtensionFromGlobalObjectCallback):
3351         (TestWebKitAPI::didReconnectDOMWindowExtensionToGlobalObjectCallback):
3352         (TestWebKitAPI::willDestroyGlobalObjectForDOMWindowExtensionCallback):
3353
3354         * TestWebKitAPI/Tests/WebKit2/InjectedBundleBasic.cpp:
3355         (TestWebKitAPI::didReceiveMessageFromInjectedBundle): Make this method static, as it always should have been.
3356
3357         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
3358
3359 2012-04-24  Carlos Garcia Campos  <cgarcia@igalia.com>
3360
3361         [GTK] Build and run TestWebKitAPI unit tests
3362         https://bugs.webkit.org/show_bug.cgi?id=84325
3363
3364         Reviewed by Philippe Normand.
3365
3366         We only build WTF unit tests for now. WebKit2 tests will be
3367         enabled in a follow up patch.
3368
3369         * Scripts/run-gtk-tests:
3370         (TestRunner): Add TestWebKitAPI/WTF to the list of directories
3371         containing unit tests.
3372         (TestRunner._run_test_glib): Run glib tests using gtester.
3373         (TestRunner._run_test_google): Run google tests.
3374         (TestRunner._run_test): Use _run_test_glib or _run_test_google
3375         depending on whether given test is glib or google test
3376         * TestWebKitAPI/GNUmakefile.am: Added.
3377         * TestWebKitAPI/config.h: Use #ifdef instead of # to fix a compile
3378         warning with -Wundef. Include WebKit2 headers only when building
3379         WebKit2.
3380
3381 2012-04-24  Allan Sandfeld Jensen  <allan.jensen@nokia.com>
3382
3383         [Qt][WK2] Convert touch-point area
3384         https://bugs.webkit.org/show_bug.cgi?id=84434
3385
3386         Reviewed by Kenneth Rohde Christiansen.
3387
3388         Fix failing test after r114917 by also adding the new EventSender API
3389         to WebKitTestRunner.
3390
3391         * WebKitTestRunner/EventSenderProxy.h:
3392         (EventSenderProxy):
3393         * WebKitTestRunner/InjectedBundle/Bindings/EventSendingController.idl:
3394         * WebKitTestRunner/InjectedBundle/EventSendingController.cpp:
3395         (WTR::EventSendingController::setTouchPointRadius):
3396         * WebKitTestRunner/InjectedBundle/EventSendingController.h:
3397         (EventSendingController):
3398         * WebKitTestRunner/TestController.cpp:
3399         (WTR::TestController::didReceiveSynchronousMessageFromInjectedBundle):
3400         * WebKitTestRunner/qt/EventSenderProxyQt.cpp:
3401         (WTR::EventSenderProxy::addTouchPoint):
3402         (WTR::EventSenderProxy::updateTouchPoint):
3403         (WTR::EventSenderProxy::setTouchPointRadius):
3404         (WTR::EventSenderProxy::clearTouchPoints):
3405
3406 2012-04-24  Mikhail Naganov  <mnaganov@chromium.org>
3407
3408         [Chromium] Unreviewed test expectations update.
3409         https://bugs.webkit.org/show_bug.cgi?id=84726
3410
3411         Mark webkitpy.layout_tests.servers.http_server_unittest.TestHttpServer.test_start_cmd as failing on Windows.
3412
3413         * Scripts/webkitpy/test/test_finder.py:
3414         (TestFinder._default_names):
3415
3416 2012-04-24  Christophe Dumez  <christophe.dumez@intel.com>
3417
3418         [EFL] Enable Web Timing
3419         https://bugs.webkit.org/show_bug.cgi?id=84705
3420
3421         Reviewed by Tony Gentilcore.
3422
3423         Enable web timing by default whenever build-webkit is used.
3424
3425         * Scripts/build-webkit:
3426
3427 2012-04-23  Eriq Augustine  <eaugusti@chromium.org>
3428
3429         Return value from executed script in Chromium.
3430         https://bugs.webkit.org/show_bug.cgi?id=79851
3431
3432         Reviewed by Adam Barth.
3433
3434         Adding a function to LayoutTestController that evaluates a script in
3435         an isolated wold an returns the result of the evaluation.
3436
3437         * DumpRenderTree/LayoutTestController.cpp:
3438         (evaluateScriptInIsolatedWorldAndReturnValue):
3439         (LayoutTestController::staticFunctions):
3440         * DumpRenderTree/LayoutTestController.h:
3441         (LayoutTestController):
3442         * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp:
3443         (LayoutTestController::evaluateScriptInIsolatedWorldAndReturnValue):
3444         * DumpRenderTree/chromium/LayoutTestController.cpp:
3445         (LayoutTestController::LayoutTestController):
3446         (LayoutTestController::setXSSAuditorEnabled):
3447         (LayoutTestController::evaluateScriptInIsolatedWorldAndReturnValue):
3448         * DumpRenderTree/chromium/LayoutTestController.h:
3449         (LayoutTestController):
3450         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp:
3451         (LayoutTestController::evaluateScriptInIsolatedWorldAndReturnValue):
3452         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3453         (LayoutTestController::evaluateScriptInIsolatedWorldAndReturnValue):
3454         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3455         (LayoutTestController::evaluateScriptInIsolatedWorldAndReturnValue):
3456         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3457         (LayoutTestController::evaluateScriptInIsolatedWorldAndReturnValue):
3458         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3459         (LayoutTestController):
3460         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3461         (LayoutTestController::evaluateScriptInIsolatedWorldAndReturnValue):
3462         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3463         (LayoutTestController::evaluateScriptInIsolatedWorldAndReturnValue):
3464
3465 2012-04-23  Dirk Pranke  <dpranke@chromium.org>
3466
3467         run-perf-tests fail intermittently with an exception
3468         https://bugs.webkit.org/show_bug.cgi?id=83532
3469
3470         Reviewed by Ryosuke Niwa.
3471
3472         This bulletproofs port.get_option('child_processes')
3473         against the case where the option wasn't set.
3474
3475         * Scripts/webkitpy/layout_tests/port/mac.py:
3476         (MacPort._get_crash_log):
3477         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
3478         (test_get_crash_log):
3479         (test_get_crash_log.fake_time_cb):
3480
3481 2012-04-23  Balazs Kelemen  <kbalazs@webkit.org>
3482
3483         [Qt] Add desktop zooming support for QQuickWebView
3484         https://bugs.webkit.org/show_bug.cgi?id=82337
3485
3486         Unreviewed typo fix after r114900.
3487
3488         * MiniBrowser/qt/BrowserWindow.cpp:
3489         (BrowserWindow::BrowserWindow): Fix the mistyped
3490         value of a zoom level.
3491
3492 2012-04-23  Hao Zheng  <zhenghao@chromium.org>
3493
3494         [chromium-android] Get rid of the trick that replaces system fallback font with DejaVuSans.
3495         https://bugs.webkit.org/show_bug.cgi?id=84580
3496
3497         Reviewed by Ojan Vafai.
3498
3499         Android's first fallback font is changed to DroidSansArabic.ttf in JB.
3500         Take the opportunity to get rid of the trick.
3501
3502         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
3503         (ChromiumAndroidPort.start_helper):
3504         (ChromiumAndroidPort._push_fonts):
3505
3506 2012-04-23  Dirk Pranke  <dpranke@chromium.org>
3507
3508         test-webkitpy is failing
3509         https://bugs.webkit.org/show_bug.cgi?id=84617
3510
3511         Reviewed by Eric Seidel.
3512
3513         * Scripts/webkitpy/layout_tests/port/http_lock_unittest.py:
3514         (HttpLockTestWithRealFileSystem.clean_all_lockfile):
3515
3516 2012-04-23  Ojan Vafai  <ojan@chromium.org>
3517
3518         Split up flakiness_dashboard.html into multiple files. This is a step in the direction
3519         of making the code more maintainable (e.g. being able to use qunit for testing).
3520
3521         * TestResultServer/static-dashboards/dashboard_base.js:
3522         * TestResultServer/static-dashboards/flakiness_dashboard.css: Added.
3523         * TestResultServer/static-dashboards/flakiness_dashboard.html:
3524         * TestResultServer/static-dashboards/flakiness_dashboard.js: Copied from Tools/TestResultServer/static-dashboards/flakiness_dashboard.html.
3525         * TestResultServer/static-dashboards/flakiness_dashboard_tests.css: Added.
3526         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js: Renamed from Tools/TestResultServer/static-dashboards/flakiness_dashboard_tests.js.
3527
3528 2012-04-23  Allan Sandfeld Jensen  <allan.jensen@nokia.com>
3529
3530         [Qt][WK2] Convert touch-point area.
3531         https://bugs.webkit.org/show_bug.cgi?id=84434
3532
3533         Reviewed by Kenneth Rohde Christiansen.
3534
3535         Add feature to set touch-point radius for emitted touch-events.
3536
3537         * DumpRenderTree/qt/EventSenderQt.cpp:
3538         (EventSender::addTouchPoint):
3539         (EventSender::updateTouchPoint):
3540         (EventSender::setTouchPointRadius):
3541         (EventSender::clearTouchPoints):
3542         * DumpRenderTree/qt/EventSenderQt.h:
3543         (EventSender):
3544
3545 2012-04-23  Christophe Dumez  <christophe.dumez@intel.com>
3546
3547         [EFL] EFL's LayoutTestController does not implement elementDoesAutoCompleteForElementWithId
3548         https://bugs.webkit.org/show_bug.cgi?id=84425
3549
3550         Reviewed by Antonio Gomes.
3551
3552         Use the new helper in DumpRenderTreeSupportEfl to properly
3553         test whether an element performs autocompletion.
3554
3555         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp:
3556         (LayoutTestController::elementDoesAutoCompleteForElementWithId):
3557
3558 2012-04-23  Csaba Osztrogonác  <ossy@webkit.org>
3559
3560         master.cfg cleanup, remove dead codes after r92792.
3561         https://bugs.webkit.org/show_bug.cgi?id=84433
3562
3563         Relanding r114881 after typo fix in UploadBuiltProduct.
3564
3565         Reviewed by Ryosuke Niwa.
3566
3567         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3568
3569 2012-04-23  Dominik Röttsches  <dominik.rottsches@linux.intel.com>
3570
3571         [EFL][DRT] Need to support requestAnimationFrame
3572         https://bugs.webkit.org/show_bug.cgi?id=84585
3573
3574         Reviewed by Martin Robinson.
3575
3576         Making requestAnimationFrame() support a default on the EFL build.
3577
3578         * Scripts/build-webkit:
3579
3580 2012-04-23  Balazs Kelemen  <kbalazs@webkit.org>
3581
3582         [Qt] Add desktop zooming support for QQuickWebView
3583         https://bugs.webkit.org/show_bug.cgi?id=82337
3584
3585         Reviewed by Simon Hausmann.
3586
3587         Add zooming support for MiniBrowser to
3588         test the new API.
3589
3590         * MiniBrowser/qt/BrowserWindow.cpp:
3591         (BrowserWindow::BrowserWindow):
3592         (BrowserWindow::zoomIn):
3593         (BrowserWindow::zoomOut):
3594         (BrowserWindow::keyPressEvent):
3595         (BrowserWindow::wheelEvent):
3596         * MiniBrowser/qt/BrowserWindow.h:
3597         (BrowserWindow):
3598
3599 2012-04-23  Csaba Osztrogonác  <ossy@webkit.org>
3600
3601         Unreviewed rolling out r114881, because it broke upload buildstep.
3602
3603         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3604         (ConfigureBuild.__init__):
3605         (ConfigureBuild.start):
3606         (CompileWebKit.start):
3607         (UploadBuiltProduct):
3608         (UploadBuiltProduct.determineExtraFeatures):
3609         (Factory.__init__):
3610         (BuildFactory.__init__):
3611         (TestFactory.__init__):
3612         (loadBuilderConfig):
3613
3614 2012-04-23  Csaba Osztrogonác  <ossy@webkit.org>
3615
3616         master.cfg cleanup, remove dead codes after r92792.
3617         https://bugs.webkit.org/show_bug.cgi?id=84433
3618
3619         Reviewed by Ryosuke Niwa.
3620
3621         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3622
3623 2012-04-23  Milian Wolff  <mail@milianw.de>
3624
3625         [Qt] Make DRT not dump pixel results if test is calling layoutTestController.dumpAsText()
3626         https://bugs.webkit.org/show_bug.cgi?id=84327
3627
3628         Reviewed by Csaba Osztrogonác.
3629
3630         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
3631         (WebCore::DumpRenderTree::dump):
3632
3633 2012-04-22  Sriram Neelakandan  <sriram.neelakandan@gmail.com>
3634
3635         [Gtk] Added MOZ_X11 build flag for TARGET_X11
3636         https://bugs.webkit.org/show_bug.cgi?id=40785
3637
3638         Reviewed by Anders Carlsson.
3639
3640         * GNUmakefile.am:
3641
3642 2012-04-22  Zan Dobersek  <zandobersek@gmail.com>
3643
3644         [TestResultsServer] Impossible to inspect builds on non-Chromium builders
3645         https://bugs.webkit.org/show_bug.cgi?id=84544
3646
3647         Reviewed by Ryosuke Niwa.
3648
3649         Use the groupEnum parameter that is passed through to the onBuilderListLoad
3650         function rather than an undefined attribute of the BuilderGroup object. This
3651         makes it again possible to inspect builds on non-Chromium builders in the
3652