WebKitTools/ChangeLog

   1 2010-11-23  Dirk Pranke  <dpranke@chromium.org>
   2
   3         Reviewed by Tony Chang.
   4
   5         This patch cleans up the logic used to shard tests into groups a
   6         bit and adds the --worker-model flag to NRWT. The flag is only
   7         used at the moment to control whether to run single-threaded or
   8         not, but eventually will also allow toggling between threads and
   9         processes.
  10
  11         Also add a minor cleanup with _test_is_slow(), which just
  12         eliminates some repetition and gives slightly better encapsulation.
  13
  14         https://bugs.webkit.org/show_bug.cgi?id=49773
  15
  16         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
  17         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
  18
  19 2010-11-23  Mihai Parparita  <mihaip@chromium.org>
  20
  21         Reviewed by Tony Chang.
  22
  23         Rebaseline server: list current baselines and platforms
  24         https://bugs.webkit.org/show_bug.cgi?id=49991
  25
  26         List existing baselines (with Trac links) for tests.
  27
  28         Add dropdowns for choosing with platform baselines to target (and what
  29         to do with existing ones).
  30
  31         Also fix MockFileSystem.join to behave more like os.path.join (unit
  32         test was ending up with two consecutive slashes in a layout test
  33         path).
  34
  35         * Scripts/webkitpy/common/system/filesystem_mock.py:
  36         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
  37         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
  38         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
  39         * Scripts/webkitpy/tool/commands/data/rebaselineserver/queue.js:
  40         * Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js:
  41         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
  42         * Scripts/webkitpy/tool/commands/rebaselineserver_unittest.py: Added.
  43
  44 2010-11-23  Ojan Vafai  <ojan@chromium.org>
  45
  46         Reviewed by Tony Chang.
  47
  48         remove timeout argument to urlopen
  49         https://bugs.webkit.org/show_bug.cgi?id=49995
  50
  51         Apparently the version of python we have on the bots doesn't accept a timeout argument.
  52
  53         * Scripts/webkitpy/layout_tests/layout_package/test_results_uploader.py:
  54
  55 2010-11-23  Sheriff Bot  <webkit.review.bot@gmail.com>
  56
  57         Unreviewed, rolling out r72628.
  58         http://trac.webkit.org/changeset/72628
  59         https://bugs.webkit.org/show_bug.cgi?id=49994
  60
  61         This patch is causing layout-test failtures on GTK Linux
  62         64-bit Debug (Requested by ctguil on #webkit).
  63
  64         * DumpRenderTree/chromium/AccessibilityUIElement.cpp:
  65         (AccessibilityUIElement::AccessibilityUIElement):
  66         (AccessibilityUIElement::parentElementCallback):
  67         * DumpRenderTree/chromium/AccessibilityUIElement.h:
  68         * DumpRenderTree/chromium/CppBoundClass.cpp:
  69         * DumpRenderTree/chromium/CppBoundClass.h:
  70
  71 2010-11-22  Ojan Vafai  <ojan@chromium.org>
  72
  73         Reviewed by Tony Chang.
  74
  75         speculative fix for upload errors: stop using mechanize to upload to test-results.appspot.com
  76         https://bugs.webkit.org/show_bug.cgi?id=49944
  77
  78         * Scripts/webkitpy/layout_tests/layout_package/test_results_uploader.py:
  79
  80 2010-11-23  Chris Guillory  <chris.guillory@google.com>
  81
  82         Reviewed by Dimitri Glazkov.
  83
  84         Include the FrameView widget of a RenderWidget in the accessibility tree.
  85         https://bugs.webkit.org/show_bug.cgi?id=49106
  86
  87         * DumpRenderTree/chromium/AccessibilityUIElement.cpp:
  88         (AccessibilityUIElement::AccessibilityUIElement):
  89         (AccessibilityUIElement::parentElementCallback):
  90         (AccessibilityUIElement::isEqualCallback):
  91         * DumpRenderTree/chromium/AccessibilityUIElement.h:
  92         * DumpRenderTree/chromium/CppBoundClass.cpp:
  93         (CppBoundClass::getFromCppVariant):
  94         * DumpRenderTree/chromium/CppBoundClass.h:
  95
  96 2010-11-23  Alex Grilo  <abgrilo@profusion.mobi>
  97
  98         Reviewed by Kenneth Rohde Christiansen.
  99
 100         [EFL] Add tiled backing store to EWebLauncher
 101         https://bugs.webkit.org/show_bug.cgi?id=45777
 102
 103         Allow EWebLauncher to choose between single and tiled backing store
 104
 105         * EWebLauncher/main.c:
 106         (on_key_down):
 107         (browserCreate):
 108         (main):
 109
 110 2010-11-23  Simon Hausmann  <simon.hausmann@nokia.com>
 111
 112         Reviewed by Tor Arne Vestbø.
 113
 114         [Qt] Review the setUserPermission & friends API
 115         https://bugs.webkit.org/show_bug.cgi?id=46810
 116
 117         Renamed requestPermissionFromUser to featurePermissionRequested
 118         and cancelRequestsForPermission to featurePermissionRequestCanceled.
 119
 120         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
 121         (WebCore::WebPage::WebPage):
 122         * QtTestBrowser/webpage.cpp:
 123         (WebPage::WebPage):
 124         (WebPage::featurePermissionRequestCanceled):
 125         * QtTestBrowser/webpage.h:
 126
 127 2010-11-23  Simon Hausmann  <simon.hausmann@nokia.com>
 128
 129         Reviewed by Tor Arne Vestbø.
 130
 131         [Qt] Review the setUserPermission & friends API
 132         https://bugs.webkit.org/show_bug.cgi?id=46810
 133
 134         Rename setUserPermission to setFeaturePermission
 135
 136         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
 137         (WebCore::WebPage::requestPermission):
 138         (WebCore::WebPage::permissionSet):
 139         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
 140         (LayoutTestController::grantDesktopNotificationPermission):
 141         * QtTestBrowser/webpage.cpp:
 142         (WebPage::requestPermission):
 143
 144 2010-11-23  Simon Hausmann  <simon.hausmann@nokia.com>
 145
 146         Reviewed by Tor Arne Vestbø.
 147
 148         [Qt] Review the setUserPermission & friends API
 149         https://bugs.webkit.org/show_bug.cgi?id=46810
 150
 151         Rename PermissionDomain to Feature, NotificationsPermissionDomain to Notifications
 152         and GeolocationPermissionDomain to Geolocation.
 153
 154         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
 155         (WebCore::WebPage::WebPage):
 156         (WebCore::WebPage::requestPermission):
 157         (WebCore::WebPage::cancelPermission):
 158         (WebCore::WebPage::permissionSet):
 159         (WebCore::DumpRenderTree::geolocationPermissionSet):
 160         * DumpRenderTree/qt/DumpRenderTreeQt.h:
 161         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
 162         (LayoutTestController::grantDesktopNotificationPermission):
 163         * QtTestBrowser/webpage.cpp:
 164         (WebPage::WebPage):
 165         (WebPage::requestPermission):
 166         (WebPage::cancelRequestsForPermission):
 167         * QtTestBrowser/webpage.h:
 168
 169 2010-11-23  Simon Hausmann  <simon.hausmann@nokia.com>
 170
 171         Reviewed by Tor Arne Vestbø.
 172
 173         [Qt] Review the setUserPermission & friends API
 174         https://bugs.webkit.org/show_bug.cgi?id=46810
 175
 176         Add a ByUser suffix to PermissionGranted/Denied. In the future
 177         we can add PermissionGrantedByDefault.
 178
 179
 180         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
 181         (WebCore::WebPage::requestPermission):
 182         (WebCore::WebPage::permissionSet):
 183         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
 184         (LayoutTestController::grantDesktopNotificationPermission):
 185         * QtTestBrowser/webpage.cpp:
 186         (WebPage::requestPermission):
 187
 188 2010-11-18  Satish Sampath  <satish@chromium.org>
 189
 190         Reviewed by Jeremy Orlow.
 191
 192         For speech input event, send an event object containing all the recognition results and metadata.
 193         https://bugs.webkit.org/show_bug.cgi?id=49736
 194
 195         Updated DRT by renaming the mock result method on all platforms to the new signature.
 196
 197         * DumpRenderTree/LayoutTestController.cpp:
 198         (addMockSpeechInputResultCallback):
 199         (LayoutTestController::staticFunctions):
 200         * DumpRenderTree/LayoutTestController.h:
 201         * DumpRenderTree/chromium/LayoutTestController.cpp:
 202         (LayoutTestController::LayoutTestController):
 203         (LayoutTestController::addMockSpeechInputResult):
 204         * DumpRenderTree/chromium/LayoutTestController.h:
 205         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
 206         (LayoutTestController::addMockSpeechInputResult):
 207         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
 208         (LayoutTestController::addMockSpeechInputResult):
 209         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
 210         (LayoutTestController::addMockSpeechInputResult):
 211         * DumpRenderTree/qt/LayoutTestControllerQt.h:
 212         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
 213         (LayoutTestController::addMockSpeechInputResult):
 214         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
 215         (LayoutTestController::addMockSpeechInputResult):
 216
 217 2010-11-22  Dirk Pranke  <dpranke@chromium.org>
 218
 219         Reviewed by Tony Chang.
 220
 221         Fix some spacing issues with log messages with
 222         new-run-webkit-tests --verbose.
 223
 224         https://bugs.webkit.org/show_bug.cgi?id=49936
 225
 226         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
 227
 228 2010-11-22  Ojan Vafai  <ojan@chromium.org>
 229
 230         Reviewed by Adam Barth.
 231
 232         update scm to work with different remote branches
 233         https://bugs.webkit.org/show_bug.cgi?id=49949
 234
 235         This works if there are multiple svn tracking branches
 236         and/or if the tracking branch is refs/remots/origin/master
 237         instead of refs/remotes/trunk which is the direction we're trying
 238         to head since that makes fetches faster.
 239
 240         * Scripts/webkitpy/common/checkout/scm.py:
 241         * Scripts/webkitpy/common/checkout/scm_unittest.py:
 242
 243 2010-11-22  Hayato Ito  <hayato@chromium.org>
 244
 245         Reviewed by Shinichiro Hamaji.
 246
 247         Ignore reference files which will be used by reftests when collecting
 248         test cases.
 249         https://bugs.webkit.org/show_bug.cgi?id=49835
 250
 251         * Scripts/webkitpy/layout_tests/port/test_files.py:
 252         * Scripts/webkitpy/layout_tests/port/test_files_unittest.py:
 253
 254 2010-11-22  Adam Roben  <aroben@apple.com>
 255
 256         Use paths relative to $WebKitVSPropsRedirectionDir to access shared .vsprops files
 257
 258         Apple's Windows build allows placing header files and import libraries for WebKit's
 259         dependencies (CoreGraphics, CFNetwork, SQLite, etc.) outside the source tree via the
 260         $WebKitLibrariesDir environment variable. This is both required for production builds and
 261         convenient for Apple-internal developer builds. Apple's production builds also require that
 262         WebKit's shared .vsprops files be accessed relative to $WebKitLibrariesDir. In production
 263         builds, the files are copied into that directory tree by the
 264         WebKitLibraries/win/tools/WinTools.make file. In Apple-internal developer builds, the
 265         copying is done by
 266         JavaScriptCore/JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.make.
 267
 268         This .vsprops copying is problematic in one very important case: when a developer updates
 269         their source tree and then tries to build. Visual Studio only reads .vsprops files when a
 270         project is first loaded. So, when Visual Studio is first opened after the .vsprops files are
 271         updated, it reads in the old files that were already residing in $WebKitLibrariesDir. When a
 272         build is started, JavaScriptCoreGenerated.make copies the new .vsprops files into
 273         $WebKitLibrariesDir, but Visual Studio will not pick up the changes.  The rest of the build
 274         will proceed with out-of-date .vsprops files, which will likely result in a build failure.
 275
 276         To fix this, we now use normal relative paths to access the .vsprops files in the source
 277         tree rather than in $WebKitLibrariesDir, but prefix those paths with a new environment
 278         variable, $WebKitVSPropsRedirectionDir. In developer builds, this environment variable is
 279         unset, so the normal relative paths are used to read the .vsprops files out of the source
 280         tree directly. In production builds, this environment variable is set to a fake directory
 281         that will cause the .vsprops files in $WebKitLibrariesDir to be found when the relative path
 282         is resolved.
 283
 284         For example, JavaScriptCore.vcproj uses this path for FeatureDefines.vsprops:
 285
 286         $(WebKitVSPropsRedirectionDir)..\..\..\WebKitLibraries\win\tools\vsprops\FeatureDefines.vsprops
 287
 288         In developer builds, where $WebKitVSPropsRedirectionDir is unset, this will point to the
 289         files in WebKitLibraries\win\tools\vsprops in the source tree. In production builds,
 290         JavaScriptCore.make sets $WebKitVSPropsRedirectionDir to
 291         "$(SRCROOT)\AppleInternal\tools\vsprops\OpenSource\1\2\3\", so the full path for
 292         FeatureDefines.vsprops becomes:
 293
 294         $(SRCROOT)\AppleInternal\tools\vsprops\OpenSource\1\2\3\..\..\..\WebKitLibraries\win\tools\vsprops\FeatureDefines.vsprops
 295
 296         which resolves to:
 297
 298         $(SRCROOT)\AppleInternal\tools\vsprops\OpenSource\WebKitLibraries\win\tools\vsprops\FeatureDefines.vsprops
 299
 300         (We rely on the fact that Windows doesn't care whether the directories "1", "2", and "3"
 301         actually exist since they are matched by an equal number of ".." path components.)
 302
 303         Note that Visual Studio still won't pick up changes made to .vsprops files while Visual
 304         Studio is open, but that problem hasn't seemed to cause developers many headaches so far.
 305
 306         Fixes <http://webkit.org/b/49181> Windows build fails mysteriously when .vsprops files are
 307         updated
 308
 309         Reviewed by Dave Hyatt.
 310
 311         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
 312         * DumpRenderTree/win/DumpRenderTree.vcproj:
 313         * DumpRenderTree/win/ImageDiff.vcproj:
 314         * FindSafari/FindSafari.vcproj:
 315         * MiniBrowser/MiniBrowser.vcproj:
 316         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
 317         * TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj:
 318         * WebKitAPITest/WebKitAPITest.vcproj:
 319         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
 320         * WebKitTestRunner/win/InjectedBundle.vcproj:
 321         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
 322         * WinLauncher/WinLauncher.vcproj:
 323         * record-memory-win/record-memory-win.vcproj:
 324         Changed to use paths relative to $WebKitVSPropsRedirectionDir to access shared .vsprops
 325         files.
 326
 327 2010-11-22  Adam Roben  <aroben@apple.com>
 328
 329         Make it possible to run tests on Windows without Visual Studio or VC++
 330         Express installed
 331
 332         webkitdir.pm::setupCygwinEnv dies when Visual Studio and VC++ Express
 333         are not installed. But this function doesn't need to be called when we
 334         already have a build available and are just trying to run the tests.
 335
 336         Fixes <http://webkit.org/b/49932> New Windows 7 bot can't run tests
 337         because Visual Studio/VC++ Express aren't installed
 338
 339         Reviewed by Dave Hyatt.
 340
 341         * Scripts/webkitdirs.pm:
 342         (determineConfigurationForVisualStudio): Don't call setupCygwinEnv, as
 343         it is not needed by this function. Also added a FIXME.
 344
 345         (usingVisualStudioExpress): Call setupCygwinEnv directly rather than
 346         relying on determineConfigurationForVisualStudio doing it.
 347
 348 2010-11-19  Hayato Ito  <hayato@chromium.org>
 349
 350         Reviewed by Shinichiro Hamaji.
 351
 352         Simplify Driver.run_test() so that it takes only one parameter, TestInput.
 353
 354         https://bugs.webkit.org/show_bug.cgi?id=49812
 355
 356         * Scripts/webkitpy/layout_tests/driver_test.py:
 357         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
 358         * Scripts/webkitpy/layout_tests/port/base.py:
 359         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
 360         * Scripts/webkitpy/layout_tests/port/chromium.py:
 361         * Scripts/webkitpy/layout_tests/port/dryrun.py:
 362         * Scripts/webkitpy/layout_tests/port/test.py:
 363         * Scripts/webkitpy/layout_tests/port/webkit.py:
 364         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
 365         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
 366
 367 2010-11-22  Hayato Ito  <hayato@chromium.org>
 368
 369         Reviewed by Tony Chang.
 370
 371         [NRWT] Retry a few times in reading a png image to avoid a race condition.
 372
 373         https://bugs.webkit.org/show_bug.cgi?id=49924
 374
 375         * Scripts/webkitpy/layout_tests/port/chromium.py:
 376
 377 2010-11-22  João Paulo Rechi Vita  <jprvita@profusion.mobi>
 378
 379         Reviewed by Kenneth Rohde Christiansen.
 380
 381         Fix EWebLauncher zoom discretization
 382         https://bugs.webkit.org/show_bug.cgi?id=49810
 383
 384         * EWebLauncher/main.c:
 385         (nearest_zoom_level_get):
 386         (zoom_level_set):
 387         (on_load_finished):
 388         (on_key_down):
 389
 390 2010-11-21  Sam Weinig  <sam@webkit.org>
 391
 392         Rubber-stamped by Dan Bernstein.
 393
 394         Fix ~300 WebKit2 tests by enabling developer extras preference.
 395
 396         * WebKitTestRunner/TestController.cpp:
 397         (WTR::TestController::resetStateToConsistentValues):
 398
 399 2010-11-21  Sam Weinig  <sam@webkit.org>
 400
 401         Reviewed by Dan Bernstein.
 402
 403         Textareas should be resizable by default
 404         https://bugs.webkit.org/show_bug.cgi?id=49892
 405
 406         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
 407         * TestWebKitAPI/Tests/WebKit2/WKPreferences.cpp: Added.
 408         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
 409         Add preferences test.
 410
 411 2010-11-19  Mihai Parparita  <mihaip@chromium.org>
 412
 413         Reviewed by Tony Chang.
 414
 415         Rebaseline server: checksum-only differences and prefetching
 416         https://bugs.webkit.org/show_bug.cgi?id=49841
 417
 418         Add support for image test results where the only difference is in the
 419         checksum.
 420
 421         Add prefetching of the image results from the next test.
 422
 423         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
 424         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
 425         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
 426
 427 2010-11-19  Dirk Pranke  <dpranke@chromium.org>
 428
 429         Reviewed by Tony Chang.
 430
 431         nrwt multiprocessing - add 'worker number' concept, move stuff to worker thread
 432
 433         Add the 'worker number' and 'worker name' concepts to the
 434         TestShellThread objects, and move test_types and test_args from
 435         the TestRunner to the TestShellThread.
 436
 437         https://bugs.webkit.org/show_bug.cgi?id=49768
 438
 439         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
 440         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
 441
 442 2010-11-19  Dirk Pranke  <dpranke@chromium.org>
 443
 444         Reviewed by Ojan Vafai.
 445
 446         Do some minor cleanup and bug fixing.
 447
 448         https://bugs.webkit.org/show_bug.cgi?id=49777
 449
 450         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
 451         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
 452
 453 2010-11-19  Dirk Pranke  <dpranke@chromium.org>
 454
 455         Reviewed by Ojan Vafai.
 456
 457         nrwt - config.build_directory() is busted
 458
 459         Fixes a typo that was causing us to usually return the top level
 460         directory WebKitBuild instead of WebKitBuild/{Debug,Release}. The
 461         bug was hidden by test stubs that were too simplistic :(.
 462
 463         https://bugs.webkit.org/show_bug.cgi?id=49815
 464
 465         * Scripts/webkitpy/common/system/executive_mock.py:
 466         * Scripts/webkitpy/layout_tests/port/config.py:
 467         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
 468
 469 2010-11-19  Steve Falkenburg  <sfalken@apple.com>
 470
 471         Reviewed by Adam Roben.
 472
 473         Add Debug_Cairo_CFLite and Release_Cairo_CFLite configurations for all vcproj files
 474         https://bugs.webkit.org/show_bug.cgi?id=49819
 475
 476         * DumpRenderTree/DumpRenderTree.sln:
 477         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
 478         * DumpRenderTree/win/DumpRenderTree.vcproj:
 479         * DumpRenderTree/win/ImageDiff.vcproj:
 480         * FindSafari/FindSafari.vcproj:
 481         * MiniBrowser/MiniBrowser.vcproj:
 482         * TestWebKitAPI/win/TestWebKitAPI.sln:
 483         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
 484         * TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj:
 485         * WebKitAPITest/WebKitAPITest.vcproj:
 486         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
 487         * WebKitTestRunner/win/InjectedBundle.vcproj:
 488         * WebKitTestRunner/win/InjectedBundleGenerated.vcproj:
 489         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
 490         * WinLauncher/WinLauncher.vcproj:
 491         * record-memory-win/record-memory-win.vcproj:
 492
 493 2010-11-19  Adam Roben  <aroben@apple.com>
 494
 495         Add some more Windows slaves
 496
 497         This is probably more than we will be using in the immediate future,
 498         but having some extra slave names makes experimenting with different
 499         configurations easier.
 500
 501         Reviewed by Jon Honeycutt.
 502
 503         * BuildSlaveSupport/build.webkit.org-config/config.json: Added 5 more
 504         Windows slaves and distributed them among the three different Test
 505         builders.
 506
 507 2010-11-19  Kinuko Yasuda  <kinuko@chromium.org>
 508
 509         Reviewed by Ojan Vafai.
 510
 511         [Chromium] REGRESSION: Record actual test results in the JSON result summary for non-layout tests
 512         https://bugs.webkit.org/show_bug.cgi?id=49702
 513
 514         Re-landing r72357 with a test fix.
 515
 516         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
 517         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
 518         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
 519
 520 2010-11-19  Avi Drissman  <avi@google.com>
 521
 522         Reviewed by Antonio Gomes.
 523
 524         Complete support for Unix editing mode
 525         https://bugs.webkit.org/show_bug.cgi?id=49757
 526
 527         * DumpRenderTree/LayoutTestController.cpp:
 528         (setEditingBehaviorCallback):
 529         * DumpRenderTree/chromium/LayoutTestController.cpp:
 530         (LayoutTestController::setEditingBehavior):
 531
 532 2010-11-19  Adam Roben  <aroben@apple.com>
 533
 534         Fix run-javascriptcore-tests when there's a space in the path to
 535         testapi
 536
 537         Reviewed by Sam Weinig.
 538
 539         * Scripts/run-javascriptcore-tests: Use an indirect object when calling
 540         system() to force it to interpret spaces in the path correctly.
 541
 542 2010-11-19  Adam Roben  <aroben@apple.com>
 543
 544         Update scripts for .exe name changes after r72327
 545
 546         As of r72327, the "_debug" suffix is only used in Debug_All builds.
 547
 548         Reviewed by Sam Weinig.
 549
 550         * Scripts/run-javascriptcore-tests:
 551         (testapiPath):
 552
 553         * Scripts/sunspider-compare-results:
 554         (pathToBuiltJSC):
 555
 556         * Scripts/webkitdirs.pm:
 557         (jscPath):
 558         (safariPath):
 559
 560 2010-11-19  Steve Falkenburg  <sfalken@apple.com>
 561
 562         Reviewed by Darin Adler.
 563
 564         Normalize Cairo/CFLite project/solution configuration names
 565         https://bugs.webkit.org/show_bug.cgi?id=49818
 566
 567         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
 568         * DumpRenderTree/win/DumpRenderTree.vcproj:
 569         * DumpRenderTree/win/ImageDiff.vcproj:
 570         * MiniBrowser/MiniBrowser.vcproj:
 571         * TestWebKitAPI/win/TestWebKitAPI.sln:
 572         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
 573         * WebKitAPITest/WebKitAPITest.vcproj:
 574         * WebKitTestRunner/win/InjectedBundle.vcproj:
 575         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
 576         * WinLauncher/WinLauncher.vcproj:
 577         * Scripts/webkitdirs.pm: Updated cairo build configs.
 578
 579 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
 580
 581         Reviewed by Adam Roben.
 582
 583         Windows vcproj configuration names should be normalized across projects
 584         https://bugs.webkit.org/show_bug.cgi?id=49776
 585
 586         * DumpRenderTree/DumpRenderTree.sln:
 587         * FindSafari/FindSafari.vcproj:
 588         * TestWebKitAPI/win/TestWebKitAPI.sln:
 589         * TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj:
 590         * TestWebKitAPI/win/TestWebKitAPIGeneratedCommon.vsprops: Added.
 591         * WebKitAPITest/WebKitAPITest.vcproj:
 592         * WebKitAPITest/WebKitAPITestCommon.vsprops:
 593         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
 594         * WebKitTestRunner/WebKitTestRunner.sln:
 595         * WebKitTestRunner/win/InjectedBundleGenerated.vcproj:
 596         * WebKitTestRunner/win/InjectedBundleGeneratedCommon.vsprops: Added.
 597         * record-memory-win/record-memory-win.vcproj:
 598
 599 2010-11-18  Gavin Barraclough  <barraclough@apple.com>
 600
 601         Build fix - as per DRT-mac, always enabled developer extras,
 602         in win/qt/gtk DRT.
 603
 604         This is necessary to enable rich exception messages to be generated
 605         following https://bugs.webkit.org/show_bug.cgi?id=49708.
 606
 607         * DumpRenderTree/gtk/DumpRenderTree.cpp:
 608         (shouldEnableDeveloperExtras):
 609         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
 610         (WebCore::shouldEnableDeveloperExtras):
 611         * DumpRenderTree/win/DumpRenderTree.cpp:
 612         (shouldEnableDeveloperExtras):
 613
 614 2010-11-18  Sheriff Bot  <webkit.review.bot@gmail.com>
 615
 616         Unreviewed, rolling out r72357.
 617         http://trac.webkit.org/changeset/72357
 618         https://bugs.webkit.org/show_bug.cgi?id=49784
 619
 620         It broke test-webkitpy tests since the patch didn't have the
 621         corresponding test update (Requested by kinuko on #webkit).
 622
 623         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
 624         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
 625
 626 2010-11-18  Mihai Parparita  <mihaip@chromium.org>
 627
 628         Reviewed by Tony Chang.
 629
 630         Rebaseline server: add rebaseline queue
 631         https://bugs.webkit.org/show_bug.cgi?id=49763
 632
 633         Add a UI queue (a multi-item <select>) where tests to be baselined
 634         can be enqueued (also supports basic removal). A queue is used instead
 635         of immediately doing rebaselines since SCM operations can take a while.
 636         It's better to go through lots of tests quickly to mark them as
 637         needing rebaselining and then batch the SCM operations.
 638
 639         Also adds a simple log where results can be displayed.
 640
 641         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
 642         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
 643         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
 644         * Scripts/webkitpy/tool/commands/data/rebaselineserver/queue.js: Added.
 645         * Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js:
 646         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
 647
 648 2010-11-18  Gavin Barraclough  <barraclough@apple.com>
 649
 650         Reviewed by Oliver Hunt.
 651
 652         Bug 49708 - Stop recompiling functions to regenerate exception info.
 653
 654         Instead only hold info as necessary – keep divot info is the inspector
 655         is enabled, line number info is debugging or profiling, and handler
 656         info for functions with try/catch.
 657
 658         * DumpRenderTree/mac/DumpRenderTree.mm:
 659         (shouldEnableDeveloperExtras):
 660             - always enable the developer tools from DRT, to ensure we
 661               produce rich error messages on JavaScript exceptions.
 662
 663 2010-11-18  Kinuko Yasuda  <kinuko@chromium.org>
 664
 665         Reviewed by Ojan Vafai.
 666
 667         [Chromium] REGRESSION: Record actual test results in the JSON result summary for non-layout tests
 668         https://bugs.webkit.org/show_bug.cgi?id=49702
 669
 670         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
 671         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
 672
 673 2010-11-18  Dirk Pranke  <dpranke@chromium.org>
 674
 675         Reviewed by Tony Chang.
 676
 677         new-run-webkit-tests: create first part of 'message_broker' class for multiprocessing fixes
 678
 679         Create the first version of the 'message_broker' package. This
 680         class will encapsulate all of the threading/multiprocessing and
 681         message-sending details for the communication between the
 682         'manager' object and the 'worker' objects. For the moment, it
 683         just holds some routines and tests for logging thread stacks.
 684
 685         There should be no functional changes in this patch, just moving stuff
 686         around.
 687
 688         https://bugs.webkit.org/show_bug.cgi?id=49707
 689
 690         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
 691         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread_unittest.py: Removed.
 692         * Scripts/webkitpy/layout_tests/layout_package/message_broker.py: Added.
 693         * Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py: Added.
 694         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
 695         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
 696
 697 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
 698
 699         Reviewed by Adam Roben.
 700
 701         Remove leftover Windows Debug_Internal configurations
 702         https://bugs.webkit.org/show_bug.cgi?id=49758
 703
 704         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
 705         * DumpRenderTree/win/DumpRenderTree.vcproj:
 706         * DumpRenderTree/win/ImageDiff.vcproj:
 707         * MiniBrowser/MiniBrowser.vcproj:
 708         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
 709         * WebKitAPITest/WebKitAPITest.vcproj:
 710         * WebKitTestRunner/win/InjectedBundle.vcproj:
 711         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
 712         * WinLauncher/WinLauncher.vcproj:
 713
 714 2010-11-18  Adam Roben  <aroben@apple.com>
 715
 716         Add three new Windows XP build slaves
 717
 718         Reviewed by Steve Falkenburg.
 719
 720         * BuildSlaveSupport/build.webkit.org-config/config.json: Added three
 721         new slaves, and added them to the Windows XP Debug (Tests) builder.
 722
 723 2010-11-18  Daniel Bates  <dbates@rim.com>
 724
 725         Reviewed by Adam Roben.
 726
 727         update-webkit-support-libs should fall back to existing
 728         WebKitSupportLibrary version if there is no internet connectivity
 729         https://bugs.webkit.org/show_bug.cgi?id=49503
 730
 731         Fall back to existing support libraries (if present) when there is
 732         no internet connection.
 733
 734         Currently, update-webkit-support-libs dies with an "out-of-date"
 735         error when there is no internet connection because it cannot
 736         retrieve versioning information from developer.apple.com. Because
 737         update-webkit-support-libs fails, build-webkit fails. Instead,
 738         if there is no internet connection and the support libraries are
 739         present then we should warn the user and exit() with success so
 740         that build-webkit can work without an internet connection.
 741
 742         * Scripts/update-webkit-support-libs:
 743
 744 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
 745
 746         Reviewed by Adam Roben.
 747
 748         Debug_Internal Windows configuration is unnecessary, should be removed
 749         https://bugs.webkit.org/show_bug.cgi?id=49753
 750
 751         * DumpRenderTree/win/DumpRenderTree.cpp:
 752         * MiniBrowser/win/stdafx.h:
 753         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp:
 754         * WebKitTestRunner/win/TestControllerWin.cpp:
 755
 756 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
 757
 758         Reviewed by Adam Roben.
 759
 760         Clean up vcproj errors
 761         https://bugs.webkit.org/show_bug.cgi?id=49741
 762
 763         * MiniBrowser/MiniBrowser.vcproj:
 764         * WebKitAPITest/WebKitAPITest.vcproj:
 765         * WebKitTestRunner/win/InjectedBundle.vcproj:
 766         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
 767
 768 2010-11-18  Hayato Ito  <hayato@chromium.org>
 769
 770         Reviewed by Shinichiro Hamaji.
 771
 772         [NRWT] Fix a break of '--new-baseline' feature in pixel_test
 773         which is caused by r72249.
 774
 775         https://bugs.webkit.org/show_bug.cgi?id=49751
 776
 777         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
 778
 779 2010-11-18  Adam Roben  <aroben@apple.com>
 780
 781         Implement layoutTestController.waitForPolicyDelegate on Windows
 782
 783         Fixes <http://webkit.org/b/25038> <rdar://problem/6790213>.
 784
 785         Reviewed by Simon Fraser.
 786
 787         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
 788         (LayoutTestController::waitForPolicyDelegate): Implemented by porting
 789         code from LayoutTestControllerMac.mm.
 790
 791 2010-11-18  Hayato Ito  <hayato@chromium.org>
 792
 793         Reviewed by Shinichiro Hamaji.
 794
 795         [NRWT] Fix a break of '--new-baseline' feature which is caused by r72249.
 796
 797         https://bugs.webkit.org/show_bug.cgi?id=49740
 798
 799         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
 800
 801 2010-11-18  Tony Chang  <tony@chromium.org>
 802
 803         Reviewed by Adam Barth.
 804
 805         run platform/chromium/plugins/refcount-leaks.html on all platforms
 806         https://bugs.webkit.org/show_bug.cgi?id=49485
 807
 808         Add PluginObject.testObjectCount which returns the number of allocated
 809         TestObjects. Add PluginObject.testCreateTestObject which allocates
 810         and returns a TestObject.
 811
 812         Add TestObject.refCount which returns the number of refs on the
 813         TestObject.
 814
 815         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
 816         (pluginGetProperty):
 817         (pluginInvoke):
 818         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
 819         (getTestObjectCount):
 820         (testAllocate):
 821         (testDeallocate):
 822         (testGetProperty):
 823         (testScriptObjectInvoke): Release a ref to a plugin object that was
 824             previously leaked.
 825         * DumpRenderTree/TestNetscapePlugIn/TestObject.h:
 826
 827 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
 828
 829         Reviewed by Adam Roben.
 830
 831         WebKitTools projects (WebKitTestRunner, record-memory-win) should use vsprops for common build settings
 832         https://bugs.webkit.org/show_bug.cgi?id=49711
 833
 834         * WebKitTestRunner/Configurations/WebKitTestRunnerCFLite.vsprops: Added.
 835         * WebKitTestRunner/Configurations/WebKitTestRunnerCommon.vsprops: Added.
 836         * WebKitTestRunner/Configurations/WebKitTestRunnerCoreFoundation.vsprops: Added.
 837         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
 838         * record-memory-win/record-memory-win-common.vsprops: Added.
 839         * record-memory-win/record-memory-win.vcproj:
 840
 841 2010-11-18  Andras Becsi  <abecsi@inf.u-szeged.hu>
 842
 843         Reviewed by Csaba Osztrogonác.
 844
 845         [Qt][WK2] Only add user agent strings to the list which aren't listed yet.
 846
 847         * MiniBrowser/qt/BrowserWindow.cpp:
 848         (BrowserWindow::updateUserAgentList):
 849
 850 2010-11-17  Adam Roben  <aroben@apple.com>
 851
 852         Don't trigger Windows builds when chromium-win test results change
 853
 854         Reviewed by Mark Rowe.
 855
 856         * Scripts/webkitpy/common/config/build.py:
 857         (_should_file_trigger_build): Changed the regular expression used to
 858         search for directory names to only consider full directory names,
 859         rather than matching directory names where the string we care about is
 860         a suffix of the name.
 861
 862         * Scripts/webkitpy/common/config/build_unittest.py:
 863         (ShouldBuildTest): Added a test.
 864
 865 2010-11-17  Hayato Ito  <hayato@chromium.org>
 866
 867         Reviewed by Shinichiro Hamaji.
 868
 869         Remove a uri member from TestInput class.
 870
 871         https://bugs.webkit.org/show_bug.cgi?id=49691
 872
 873         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
 874         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
 875
 876 2010-11-17  Adam Roben  <aroben@apple.com>
 877
 878         Move FindSafari's settings into a vsprops file
 879
 880         Fixes <http://webkit.org/b/49699> FindSafari should use a vsprops file
 881
 882         Reviewed by Steve Falkenburg.
 883
 884         * FindSafari/FindSafari.vcproj:
 885         * FindSafari/FindSafariCommon.vsprops: Added.
 886
 887 2010-11-17  Adam Roben  <aroben@apple.com>
 888
 889         Use a minimum font size of 0 in DRT on Windows
 890
 891         This is the Windows equivalent of r72164.
 892
 893         Rubber-stamped by Dave Hyatt.
 894
 895         * DumpRenderTree/win/DumpRenderTree.cpp:
 896         (resetDefaultsToConsistentValues):
 897
 898 2010-11-17  Dirk Pranke  <dpranke@chromium.org>
 899
 900         Reviewed by Andreas Kling.
 901
 902         webkit-patch pretty-diff shouldn't need to be run from a checkout root in git
 903         https://bugs.webkit.org/show_bug.cgi?id=49639
 904
 905         * Scripts/webkitpy/common/checkout/scm.py:
 906
 907 2010-11-17  Mihai Parparita  <mihaip@chromium.org>
 908
 909         Reviewed by Tony Chang.
 910
 911         Rebaseline server: add loupe for image diffs
 912         https://bugs.webkit.org/show_bug.cgi?id=49692
 913
 914         Add a loupe (magnifiying glass) for inspecting image diffs. Shows an
 915         enlarged area of the expected, actual and diff images side by side.
 916
 917         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
 918         * Scripts/webkitpy/tool/commands/data/rebaselineserver/loupe.js: Added.
 919         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
 920         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
 921
 922 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
 923
 924         Reviewed by Adam Roben.
 925
 926         WebKitLauncherWin should use vsprops for shared build settings
 927         https://bugs.webkit.org/show_bug.cgi?id=49696
 928
 929         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
 930         * WebKitLauncherWin/WebKitLauncherWinCommon.vsprops: Added.
 931
 932 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
 933
 934         Reviewed by Adam Roben.
 935
 936         WinLauncher should use vsprops for shared build settings
 937         https://bugs.webkit.org/show_bug.cgi?id=49695
 938
 939         * WinLauncher/WinLauncher.cpp:
 940         (_tWinMain):
 941         (PrintView):
 942         (WndProc):
 943         (MyEditProc):
 944         * WinLauncher/WinLauncher.vcproj:
 945         * WinLauncher/WinLauncherCommon.vsprops: Added.
 946
 947 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
 948
 949         Delete unused file unintentionally added in last change.
 950
 951         * DumpRenderTree/win/DumpRenderTreeCURL.vsprops: Removed.
 952
 953 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
 954
 955         Reviewed by Adam Roben.
 956
 957         DumpRenderTree on Windows should use a vsprops file for shared project settings
 958         https://bugs.webkit.org/show_bug.cgi?id=49690
 959
 960         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
 961         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePluginCommon.vsprops: Added.
 962         * DumpRenderTree/win/DumpRenderTree.vcproj:
 963         * DumpRenderTree/win/DumpRenderTreeApple.vsprops: Added.
 964         * DumpRenderTree/win/DumpRenderTreeCFLite.vsprops: Added.
 965         * DumpRenderTree/win/DumpRenderTreeCURL.vsprops: Added.
 966         * DumpRenderTree/win/DumpRenderTreeCairo.vsprops: Added.
 967         * DumpRenderTree/win/DumpRenderTreeCommon.vsprops: Added.
 968         * DumpRenderTree/win/ImageDiff.vcproj:
 969         * DumpRenderTree/win/ImageDiffCommon.vsprops: Added.
 970
 971 2010-11-17  Hayato Ito  <hayato@chromium.org>
 972
 973         Refactor TestTypeBase.compare_output().
 974
 975         Introduce a TestOutput class and update compare_output() of each test
 976         types so that they can take both actual and expected TestOutput objects.
 977
 978         https://bugs.webkit.org/show_bug.cgi?id=49431
 979
 980         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
 981         * Scripts/webkitpy/layout_tests/layout_package/test_output.py: Added.
 982         * Scripts/webkitpy/layout_tests/port/base.py:
 983         * Scripts/webkitpy/layout_tests/port/chromium.py:
 984         * Scripts/webkitpy/layout_tests/port/dryrun.py:
 985         * Scripts/webkitpy/layout_tests/port/test.py:
 986         * Scripts/webkitpy/layout_tests/port/webkit.py:
 987         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
 988         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
 989         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
 990         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
 991
 992 2010-11-17  Adam Roben  <aroben@apple.com>
 993
 994         Make each Windows Test builder use the same OS for all its slaves
 995
 996         Currently each WebKit1 Windows Test builder has one Windows XP slave
 997         and one Windows 7 slave. Having different OSes in a single builder
 998         meant that the builders would alternate between red and green when
 999         there were OS-specific failures.
1000
1001         Fixes <http://webkit.org/b/49688> Windows Test builders alternate
1002         between red and green because of multiple OS versions
1003
1004         Reviewed by Mark Rowe.
1005
1006         * BuildSlaveSupport/build.webkit.org-config/config.json: Renamed
1007         "Windows Release (Tests)" to "Windows 7 Release (Tests)", "Windows
1008         Debug (Tests)" to "Windows XP Debug (Tests)", and "Windows Release
1009         (WebKit2 Tests)" to "Windows 7 Release (WebKit2 Tests)". Updated the
1010         builders for the two WebKit1 builders so that they would all use the
1011         appropriate OS.
1012
1013 2010-11-17  James Robinson  <jamesr@chromium.org>
1014
1015         Reviewed by Dan Bernstein.
1016
1017         [chromium] Set minimum font size to 0 for DRT
1018         https://bugs.webkit.org/show_bug.cgi?id=49677
1019
1020         Followup for http://trac.webkit.org/changeset/72141.
1021
1022         Test: http/tests/misc/acid3.html
1023
1024         * DumpRenderTree/chromium/WebPreferences.cpp:
1025         (WebPreferences::reset):
1026
1027 2010-11-17  Andras Becsi  <abecsi@webkit.org>
1028
1029         Unreviewed build fix.
1030
1031         [Qt] Add missing file I forgot in r72220.
1032
1033         * MiniBrowser/MiniBrowser.qrc: Added.
1034
1035 2010-11-17  Andras Becsi  <abecsi@webkit.org>
1036
1037         Reviewed by Andreas Kling.
1038
1039         [Qt][WK2] Make it possible to set custom user agent strings in MiniBrowser.
1040         https://bugs.webkit.org/show_bug.cgi?id=49627
1041
1042         Add a user agent dialog and a resource file for useragentlist.txt to MiniBrowser.
1043
1044         * MiniBrowser/MiniBrowser.qrc: Added.
1045         Add it here to prevent qmake from detecting it
1046         since it needs to be copied to the build directory.
1047         * MiniBrowser/qt/BrowserWindow.cpp:
1048         (BrowserWindow::BrowserWindow):
1049         (BrowserWindow::updateUserAgentList):
1050         (BrowserWindow::showUserAgentDialog):
1051         * MiniBrowser/qt/BrowserWindow.h:
1052         * MiniBrowser/qt/MiniBrowser.pro:
1053
1054 2010-11-17  Adam Roben  <aroben@apple.com>
1055
1056         Make changes to the Mac WebKit2 Skipped file trigger Windows builds
1057
1058         We use that file on Windows!
1059
1060         Fixes <http://webkit.org/b/49643> Changes to
1061         LayoutTests/platform/mac-wk2 should trigger a Windows build, but don't
1062
1063         Reviewed by Steve Falkenburg.
1064
1065         * Scripts/webkitpy/common/config/build.py:
1066         (_should_file_trigger_build): Added the mac-wk2 directory and make it
1067         trigger builds on SnowLeopard and Windows.
1068
1069         * Scripts/webkitpy/common/config/build_unittest.py:
1070         (ShouldBuildTest.test_should_build): Added a test.
1071
1072 2010-11-17  Satish Sampath  <satish@chromium.org>
1073
1074         Reviewed by Jeremy Orlow.
1075
1076         Clear the speech input mock explicitly before each test.
1077         https://bugs.webkit.org/show_bug.cgi?id=49660
1078
1079         * DumpRenderTree/chromium/LayoutTestController.cpp:
1080         (LayoutTestController::setMockSpeechInputResult):
1081         * DumpRenderTree/chromium/LayoutTestController.h:
1082         * DumpRenderTree/chromium/WebViewHost.cpp:
1083         (WebViewHost::speechInputController):
1084         (WebViewHost::reset): Invoke speech input mock's clearResults method.
1085         * DumpRenderTree/chromium/WebViewHost.h:
1086         (WebViewHost::speechInputControllerMock): Moved mock from LayoutTestController.
1087
1088 2010-11-17  Alexander Pavlov  <apavlov@chromium.org>
1089
1090         Unreviewed, specify IRC nickname in committers.py.
1091
1092         * Scripts/webkitpy/common/config/committers.py:
1093
1094 2010-11-17  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1095
1096         Reviewed by Tony Chang.
1097
1098         [NRWT] Make http locking similar to perl implementation
1099         https://bugs.webkit.org/show_bug.cgi?id=49187
1100
1101         * Scripts/webkitpy/common/system/file_lock.py: Added.
1102         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
1103         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1104
1105 2010-11-17  MORITA Hajime  <morrita@google.com>
1106
1107         Reviewed by Kent Tamura.
1108
1109         [Chromium][DRT] EventSender.contextClick() should aware spellchecking
1110         https://bugs.webkit.org/show_bug.cgi?id=49366
1111
1112         - EvenSender: Checked WebContextMenuData.misspelledWord and added extra context menu entries
1113           according to the spellchecker suggestion
1114         - MockSpellCheck: add fillSuggestionList to provide fake suggestions.
1115
1116         test_expectations.txt will be changed after this change is ported to test_shell.
1117
1118         * DumpRenderTree/chromium/EventSender.cpp:
1119         (makeMenuItemStringsFor):
1120         (EventSender::contextClick):
1121         * DumpRenderTree/chromium/MockSpellCheck.cpp:
1122         (MockSpellCheck::fillSuggestionList):
1123         (MockSpellCheck::initializeIfNeeded):
1124         * DumpRenderTree/chromium/MockSpellCheck.h:
1125         * DumpRenderTree/chromium/WebViewHost.cpp:
1126         (WebViewHost::mockSpellCheck):
1127         * DumpRenderTree/chromium/WebViewHost.h:
1128
1129 2010-11-16  Dave Hyatt  <hyatt@apple.com>
1130
1131         Reviewed by Sam Weinig.
1132
1133         Make sure the pref for minimum font size is just explicitly set to 0 every time,
1134         since the bots have 1 stuck in their plist otherwise.
1135
1136         * DumpRenderTree/mac/DumpRenderTree.mm:
1137         (resetDefaultsToConsistentValues):
1138
1139 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
1140
1141         Reviewed by Tony Chang.
1142
1143         Rebaseline server: compute diffs client-side
1144         https://bugs.webkit.org/show_bug.cgi?id=49640
1145
1146         The image diff output from the DRT is pretty bad on some ports (at
1147         least the Mac one), so it's better to compute diffs on the client by
1148         using <canvas>.
1149
1150         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
1151         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
1152         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
1153
1154 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
1155
1156         Reviewed by Tony Chang.
1157
1158         Rebaseline server: display test results
1159         https://bugs.webkit.org/show_bug.cgi?id=49626
1160
1161         Adds basic result display to the rebaseline server. On the Python side
1162         this involves:
1163         - Parsing the unexpected_results.json into a dictionary.
1164         - Serving it as JSON under /results.json.
1165         (the JSON -> dict -> JSON transform isn't strictly necessary right now,
1166         but I'll need to have access to the parsed results on the Python side
1167         for follow-up changes).
1168
1169         On the web UI side this adds:
1170         - Markup for display image and text results (expected, actual, diff),
1171           and JS for populating it.
1172         - Markup for breaking down test results by failure type and directory,
1173           and JS for populating it.
1174
1175         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
1176         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
1177         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
1178         * Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js: Added.
1179         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
1180
1181 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
1182
1183         Reviewed by Ojan Vafai.
1184
1185         See Chromium issue http://codereview.chromium.org/5133001/ - we
1186         are modifying the build bots to pass in the builder name with
1187         the "GPU" string appended instead of appending it in the code.
1188
1189         https://bugs.webkit.org/show_bug.cgi?id=49636
1190
1191         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
1192
1193 2010-11-16  Steve Falkenburg  <sfalken@apple.com>
1194
1195         Reviewed by Adam Roben.
1196
1197         Disable LTCG for Windows Release builds. Add new Release_LTCG configuration.
1198         https://bugs.webkit.org/show_bug.cgi?id=49632
1199
1200         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1201         * DumpRenderTree/win/DumpRenderTree.vcproj:
1202         * DumpRenderTree/win/ImageDiff.vcproj:
1203         * MiniBrowser/MiniBrowser.vcproj:
1204         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
1205         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
1206         * WebKitTestRunner/win/InjectedBundle.vcproj:
1207         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
1208         * WinLauncher/WinLauncher.vcproj:
1209         * record-memory-win/record-memory-win.vcproj:
1210
1211 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
1212
1213         Reviewed by Ojan Vafai.
1214
1215         new-run-webkit-tests: rename TestInfo to TestInput, move image hash to work thread
1216
1217         Rename the TestInfo class to TestInput to be clearer about its
1218         function, and move the checksum-reading code into dump_render_tree_thread
1219         to avoid cross-thread access.
1220
1221         https://bugs.webkit.org/show_bug.cgi?id=49573
1222
1223         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1224         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1225         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1226
1227 2010-11-16  Dave Hyatt  <hyatt@apple.com>
1228
1229         Fix bustage. Remove the minimum font size pref setting in DumpRenderTree.
1230
1231         * DumpRenderTree/mac/DumpRenderTree.mm:
1232         (resetDefaultsToConsistentValues):
1233
1234 2010-11-16  Stephanie Lewis  <slewis@apple.com>
1235
1236         Reviewed by Geoff Garen.
1237
1238         <rdar://problem/8624267> Leak creating offscreen webview running fast/dom tests
1239
1240         Use shared DeviceOrientationProviderMock. The old allocation method created a new WebDeviceOrientationProviderMock
1241         for every WebView, and the WebKit API doesn't support that ownership model.
1242
1243         I also fixed a leak in setMockDeviceOrientation by adding a missing release.
1244
1245         * DumpRenderTree/mac/DumpRenderTree.mm:
1246         (createWebViewAndOffscreenWindow):
1247         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1248         (LayoutTestController::setMockDeviceOrientation):
1249
1250 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
1251
1252         Reviewed by Ojan Vafai.
1253
1254         new-run-webkit-tests: log the process id in --verbose mode
1255
1256         As part of the switch to multi-process mode, we should log the
1257         process id when running w/ --verbose.
1258
1259         https://bugs.webkit.org/show_bug.cgi?id=49571
1260
1261         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
1262
1263 2010-11-16  Adam Roben  <aroben@apple.com>
1264
1265         Only trigger Windows builds when files that we actually use on Windows
1266         are changed
1267
1268         A new Scheduler subclass, PlatformSpecificScheduler, has been added.
1269         It uses the new webkitpy.common.config.build module to determine
1270         whether a particular change should trigger a build on a particular
1271         platform. The Windows builders have been switched to use a
1272         PlatformSpecificScheduler.
1273
1274         The logic to determine whether or not a particular change should
1275         trigger a build on a given platform has only been implemented/tested
1276         for Windows. I tried to make it easy to add more platforms in the
1277         future, but I don't have enough familiarity with all platforms to be
1278         able to implement it for them.
1279
1280         Fixes <http://webkit.org/b/49407> Windows builders kick off builds for
1281         lots irrelevant changes (e.g., rebaselining Chromium test results)
1282
1283         Reviewed by Eric Seidel.
1284
1285         * BuildSlaveSupport/build.webkit.org-config/config.json: Use a
1286         PlatformSpecificScheduler for the Windows builders.
1287
1288         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1289         (PlatformSpecificScheduler.__init__): Added. Stores our platform, sets
1290         up a ChangeFilter that filters to our branch and will call through to
1291         our filter method, and calls up to our base class.
1292         (PlatformSpecificScheduler.filter): Calls through to
1293         build.should_build to find out whether we should trigger a build for
1294         this change.
1295
1296         * Scripts/webkitpy/common/net/build.py: Added.
1297         (_should_file_trigger_build): Uses a set of directories and regexp
1298         patterns to determine whether the given file should trigger a build on
1299         the given platform. As mentioned earlier, this has only been
1300         implemented for Windows, though I did try to make some guesses about
1301         other platforms.
1302         (should_build): Returns true if any of the files should trigger a
1303         build on the given platform.
1304
1305         * Scripts/webkitpy/common/net/build_unittest.py: Added.
1306         (ShouldBuildTest.test_should_build): Does some basic testing to make
1307         sure we're triggering builds for the right files. It only tests
1308         Windows for now, though I tried to make some guesses about other
1309         platforms.
1310
1311 2010-11-16  Simon Hausmann  <simon.hausmann@nokia.com>
1312
1313         Reviewed by Andreas Kling.
1314
1315         [Qt] Remove synchronous QWebPage::checkPermissions signal
1316         https://bugs.webkit.org/show_bug.cgi?id=46810
1317
1318         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1319         (WebCore::WebPage::WebPage):
1320         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1321         (LayoutTestController::grantDesktopNotificationPermission): When granting
1322         permission, grant it directly on the QWebPage/Frame, that will remember it.
1323         * QtTestBrowser/webpage.cpp:
1324         (WebPage::WebPage):
1325         * QtTestBrowser/webpage.h:
1326
1327
1328 2010-11-16  Andras Becsi  <abecsi@inf.u-szeged.hu>
1329
1330         Reviewed by Csaba Osztrogonác.
1331
1332         [Qt] Add isPrinting and setPrinting methods to DRT's LayoutTestController.
1333
1334         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1335         (LayoutTestController::reset):
1336         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1337         (LayoutTestController::isPrinting):
1338         (LayoutTestController::setPrinting):
1339
1340 2010-11-16  Balazs Kelemen  <kbalazs@webkit.org>
1341
1342         Reviewed by Andreas Kling.
1343
1344         [Qt][WK2] Add a way to turn on tiled backing store in MiniBrowser
1345         https://bugs.webkit.org/show_bug.cgi?id=49587
1346
1347         * MiniBrowser/qt/BrowserView.cpp:
1348         (BrowserView::BrowserView): Added BackingStoreType
1349         argument to ctor. Propagating it to the QGraphicsWKView.
1350         * MiniBrowser/qt/BrowserView.h:
1351         * MiniBrowser/qt/BrowserWindow.cpp:
1352         (BrowserWindow::BrowserWindow):
1353         (BrowserWindow::newWindow): Create the new BrowserWindow
1354         with the same BackingStoreType that this instance has.
1355         * MiniBrowser/qt/BrowserWindow.h: Added BackingStoreType member.
1356         * MiniBrowser/qt/main.cpp:
1357         (main): Use tiled backing store if got -tiled command line argument.
1358         Avoid redundant copy of command line args.
1359
1360 2010-11-16  John Knottenbelt  <jknotten@chromium.org>
1361
1362         Reviewed by Jeremy Orlow.
1363
1364         Move DeviceOrientationClientMock from LayoutTestController to WebViewHost.
1365         https://bugs.webkit.org/show_bug.cgi?id=48506
1366
1367         This ensures that there is one mock per page WebView / WebViewHost. This is a
1368         design constraint of the DeviceOrientationClientMock because the m_controller
1369         field can only store a pointer to a single instance of the corresponding
1370         DeviceOrientationController.
1371
1372         Test: fast/dom/DeviceOrientation/no-page-cache.html
1373
1374         * DumpRenderTree/chromium/LayoutTestController.cpp:
1375         (LayoutTestController::setMockDeviceOrientation):
1376         (LayoutTestController::deviceOrientationClientMock):
1377         * DumpRenderTree/chromium/LayoutTestController.h:
1378         * DumpRenderTree/chromium/WebViewHost.cpp:
1379         (WebViewHost::deviceOrientationClientMock):
1380         (WebViewHost::deviceOrientationClient):
1381         (WebViewHost::reset):
1382         * DumpRenderTree/chromium/WebViewHost.h:
1383
1384 2010-11-15  Andras Becsi  <abecsi@webkit.org>
1385
1386         Reviewed by Andreas Kling.
1387
1388         [Qt][WK2] Avoid polling in WebKitTestRunner.
1389         https://bugs.webkit.org/show_bug.cgi?id=49542
1390
1391         Make the WK2 testing session about 10-15% faster and decrease random flakiness
1392         resulting from timing skew by removing the millisecond polling from TestControllerQt.
1393
1394         * WebKitTestRunner/TestController.cpp:
1395         (WTR::TestController::didFinishLoadForFrame):
1396         * WebKitTestRunner/TestController.h:
1397         * WebKitTestRunner/TestInvocation.cpp:
1398         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
1399         * WebKitTestRunner/mac/TestControllerMac.mm:
1400         (WTR::TestController::notifyDone): Add method stub.
1401         * WebKitTestRunner/qt/TestControllerQt.cpp:
1402         (WTR::TestControllerRunLoop::instance):
1403         (WTR::TestControllerRunLoop::start):
1404         (WTR::TestControllerRunLoop::stop):
1405         (WTR::TestControllerRunLoop::TestControllerRunLoop):
1406         (WTR::TestControllerRunLoop::timerEvent):
1407         (WTR::TestController::notifyDone):
1408         (WTR::TestController::platformRunUntil):
1409         * WebKitTestRunner/win/TestControllerWin.cpp:
1410         (WTR::TestController::notifyDone): Add method stub.
1411
1412 2010-11-15  Darin Adler  <darin@apple.com>
1413
1414         * Scripts/webkitpy/common/net/bugzilla: Added property svn:ignore.
1415
1416 2010-11-15  Benjamin Poulain  <benjamin.poulain@nokia.com>
1417
1418         Reviewed by Andreas Kling.
1419
1420         [Qt] [WK2] Add Ctrl+L shortcut to Qt Minibrowser
1421         https://bugs.webkit.org/show_bug.cgi?id=49544
1422
1423         Add the shortcut to BrowserWindow.
1424
1425         * MiniBrowser/qt/BrowserWindow.cpp:
1426         (BrowserWindow::BrowserWindow):
1427         (BrowserWindow::openLocation):
1428         * MiniBrowser/qt/BrowserWindow.h:
1429
1430 2010-11-14  David Levin  <levin@chromium.org>
1431
1432         Reviewed by Daniel Bates.
1433
1434         check-webkit-style should detect PassRefPtr usage in functions.
1435         https://bugs.webkit.org/show_bug.cgi?id=49513
1436
1437         * Scripts/webkitpy/style/checkers/cpp.py:
1438         (check_for_function_lengths): Revert a comment change that I
1439          accidentally made in r71986.
1440         (check_pass_ptr_usage): Added the code to do the check.
1441         (process_line): Added the call to check_pass_ptr_usage.
1442         (CppChecker): Added the new error category.
1443         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
1444         (CppStyleTestBase::perform_pass_ptr_check): Runs the new check for
1445          testing purposes.
1446         (PassPtrTest::*): The class/functions to unit test the new
1447          functionality.
1448
1449 2010-11-14  David Levin  <levin@chromium.org>
1450
1451         Reviewed by Shinichiro Hamaji.
1452
1453         check-webkit-style function detection and the line count style checks should be separate.
1454         https://bugs.webkit.org/show_bug.cgi?id=49512
1455
1456         * Scripts/webkitpy/style/checkers/cpp.py: Do the separation.
1457         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Adjust the test to
1458           call the detection function and fix line counts in two places now that
1459           the code really only counts the lines in the body of the function.
1460
1461 2010-11-14  Andreas Kling  <kling@webkit.org>
1462
1463         Reviewed by Antonio Gomes.
1464
1465         http/tests/plugins tests print "Unhandled variable" to stderr under Qt
1466         https://bugs.webkit.org/show_bug.cgi?id=33438
1467
1468         Stifle "Unhandled variable" warning to match what the other
1469         TestNetscapePlugin does (NPP_GetValue in TestNetscapePlugIn/main.cpp
1470         simply returns NPERR_GENERIC_ERROR for unhandled variables.)
1471
1472         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
1473         (webkit_test_plugin_get_value):
1474
1475 2010-11-13  David Levin  <levin@chromium.org>
1476
1477         Reviewed by Eric Seidel.
1478
1479         check-webkit-style function detection crashes on functions in templates.
1480         https://bugs.webkit.org/show_bug.cgi?id=49504
1481
1482         For example "template <bool shouldClose(const Element*)>".
1483
1484         * Scripts/webkitpy/style/checkers/cpp.py: If not function is detected, bail out.
1485         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Modified the complex
1486           function detection test to expose this issue.
1487
1488 2010-11-12  David Levin  <levin@chromium.org>
1489
1490         Reviewed by Shinichiro Hamaji.
1491
1492         check-webkit-style function detection doesn't detect indented functions declaractions.
1493         https://bugs.webkit.org/show_bug.cgi?id=49446
1494
1495         Indented function declarations occur inside class definitions, so
1496         they are a pretty common (and worth detecting).
1497
1498         * Scripts/webkitpy/style/checkers/cpp.py:
1499           Changed regex to allow indentation.
1500           Changed the function start detection to only happen when not in a
1501           function.
1502           Changed function end detection to work based on matching braces
1503           instead of finding a close brace at the beginning of the line.
1504           Fixed close_expression to do what it says when it doesn't find
1505           the close.
1506         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Indented function test.
1507
1508 2010-11-12  Daniel Bates  <dbates@rim.com>
1509
1510         Rubber-stamped by Nikolas Zimmermann.
1511
1512         Make do-file-rename rename files in the directory WebKit2.
1513
1514         This also makes the list of searched directories in do-file-rename
1515         consistent with the list of searched directories in the script do-webcore-rename.
1516
1517         * Scripts/do-file-rename:
1518
1519 2010-11-12  Dirk Pranke  <dpranke@chromium.org>
1520
1521         Reviewed by James Robinson.
1522
1523         Attempt yet again to land the fix for bug 49360 (respecting
1524         set-webkit-configuration). We need to handle the cases where
1525         trying to run webkit-build-directory to find out where the
1526         default configuration might be fails (that shows up on some
1527         Chromium bots that apparently don't have perl installed).
1528
1529         https://bugs.webkit.org/show_bug.cgi?id=49360
1530
1531         * Scripts/webkitpy/layout_tests/port/config.py:
1532         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Added.
1533         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1534
1535 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
1536
1537         Reviewed by Adam Barth.
1538
1539         webkit-patch land logs commit message twice
1540         https://bugs.webkit.org/show_bug.cgi?id=49482
1541
1542         Don't log the comment that's used by bugzilla.close_bug_as_fixed, since
1543         that's just the commit message, which we now output following the
1544         actual commit.
1545
1546         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
1547
1548 2010-11-12  Tony Chang  <tony@chromium.org>
1549
1550         Reviewed by Adam Barth.
1551
1552         run platform/chromium/plugins/script-object-invoke.html on all platforms
1553         https://bugs.webkit.org/show_bug.cgi?id=49280
1554
1555         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1556         (testScriptObjectInvoke):
1557         (pluginInvoke):
1558
1559 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
1560
1561         Reviewed by Adam Barth.
1562
1563         webkit-patch land-cowboy (and other commands) should display the committed SVN revision after landing a patch
1564         https://bugs.webkit.org/show_bug.cgi?id=49471
1565
1566         Right now there's no indication of a successful commit.
1567
1568         * Scripts/webkitpy/tool/commands/download_unittest.py:
1569         * Scripts/webkitpy/tool/steps/commit.py:
1570
1571 2010-11-12  Kinuko Yasuda  <kinuko@chromium.org>
1572
1573         Reviewed by Ojan Vafai.
1574
1575         Include detailed test modifiers in results.json and enable incremental uploading for non-layout tests
1576         https://bugs.webkit.org/show_bug.cgi?id=49354
1577
1578         Also moved/integrated the upload method from run_webkit_tests.py to json_results_generator.py.
1579
1580         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
1581         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1582         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
1583         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1584
1585 2010-11-12  Tony Chang  <tony@chromium.org>
1586
1587         Reviewed by Kent Tamura.
1588
1589         [chromium] Load a blank page when closing DRT WebViewHosts
1590         https://bugs.webkit.org/show_bug.cgi?id=49417
1591
1592         In test_shell, when we close a window, we first load about:blank
1593         and run GC to fire the destruction logic of the page.  In test_shell,
1594         this happens in ~TestShell.  In DRT, we manage virtual windows as
1595         WebViewHosts, so we need to replicate this logic in ~WebViewHost.
1596
1597         Managing focus is a bit trickier, since normally you would get window
1598         events to reset the focus to the right window.  We do it manually
1599         instead.
1600
1601         This fixes 2 plugin tests on DRT.
1602
1603         * DumpRenderTree/chromium/TestShell.cpp:
1604         (TestShell::~TestShell): Move window close logic to ~WebViewHost.
1605         (TestShell::closeWindow): Move widget close() to WVH destructor and
1606             manage focus manually.
1607         (TestShell::closeRemainingWindows): Comment typo.
1608         * DumpRenderTree/chromium/WebViewHost.cpp:
1609         (WebViewHost::closeWidget):
1610         (invokeCloseWidget):
1611         (WebViewHost::closeWidgetSoon): Close the window asynchronously (matches
1612             test_shell).
1613         (WebViewHost::~WebViewHost):
1614         * DumpRenderTree/chromium/WebViewHost.h:
1615
1616 2010-11-12  David Levin  <levin@chromium.org>
1617
1618         Reviewed by Adam Barth.
1619
1620         cpp.py needs some tidying.
1621         https://bugs.webkit.org/show_bug.cgi?id=49443
1622
1623         * Scripts/webkitpy/style/checkers/cpp.py:
1624           Fix parameter ordering for iteratively_replace_matches_with_char
1625           to mirror other similar functions (and improved comments, etc.)
1626           Also, use a python-ism (for/else) to get rid of a variable that
1627           was there to determine if something had been found in the loop.
1628
1629 2010-11-12  Robert Kroeger  <rjkroege@chromium.org>
1630
1631         Reviewed by James Robinson.
1632
1633         [Chromium/DRT] Make EventSendingController send complete touch events.
1634         https://bugs.webkit.org/show_bug.cgi?id=49285
1635
1636         * DumpRenderTree/chromium/EventSender.cpp:
1637         (EventSender::addTouchPoint):
1638         (EventSender::updateTouchPoint):
1639         (EventSender::sendCurrentTouchEvent):
1640
1641 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
1642
1643         Unreviewed rollout of r71858.
1644
1645         Rollout out r71858 since it breaks new-run-webkit-httpd as used by the
1646         NaCl tests.
1647
1648         * Scripts/webkitpy/layout_tests/port/config.py:
1649         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Removed.
1650         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1651
1652 2010-11-12  Benjamin Poulain  <benjamin.poulain@nokia.com>
1653
1654         Reviewed by Kenneth Rohde Christiansen.
1655
1656         [Qt] TestController::runUntil doesn't honor the timeout parameter
1657         https://bugs.webkit.org/show_bug.cgi?id=48941
1658
1659         Implement the missing timeout of WebKitTestRunner for Qt.
1660
1661         A QElapsedTimer has been added to RunUntilConditionLoop in order
1662         to measure how long the loop has been running. When the timer
1663         is bigger or equal than the timeout, the loop ends even if the
1664         condition is not met.
1665
1666         * WebKitTestRunner/qt/TestControllerQt.cpp:
1667         (WTR::RunUntilConditionLoop::start):
1668         (WTR::RunUntilConditionLoop::run):
1669         (WTR::RunUntilConditionLoop::timerEvent):
1670         (WTR::TestController::platformRunUntil):
1671
1672 2010-11-11  David Levin  <levin@chromium.org>
1673
1674         Reviewed by Shinichiro Hamaji.
1675
1676         check-webkit-style function detection doesn't handle templates with spaces.
1677         https://bugs.webkit.org/show_bug.cgi?id=49427
1678
1679         * Scripts/webkitpy/style/checkers/cpp.py: Stripped out templates when
1680           finding the function name.
1681         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test have a
1682           template with a space.
1683
1684 2010-11-11  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
1685
1686         Reviewed by Shinichiro Hamaji.
1687
1688         Enable check-webkit-style on Qt files
1689         https://bugs.webkit.org/show_bug.cgi?id=48322
1690
1691         Adding exception for _q_... private slots and
1692         ..._data() methods in tests
1693
1694         * Scripts/webkitpy/style/checker.py: enable Qt dirs, add exceptions
1695         * Scripts/webkitpy/style/checker_unittest.py: improve unittests
1696
1697 2010-11-11  David Levin  <levin@chromium.org>
1698
1699         Reviewed by Shinichiro Hamaji.
1700
1701         check-webkit-style function detection doesn't work for templates and destructors.
1702         https://bugs.webkit.org/show_bug.cgi?id=49425
1703
1704         * Scripts/webkitpy/style/checkers/cpp.py: Changed regex for templates and destructors.
1705         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test hit this issue.
1706
1707 2010-11-11  Joone Hur  <joone@kldp.org>
1708
1709         Reviewed by Antonio Gomes.
1710
1711         [GTK][DRT] Implement LayoutTestController::nodesFromRect
1712         https://bugs.webkit.org/show_bug.cgi?id=46598
1713
1714         Support nodesFromRect in DRT
1715
1716         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1717         (LayoutTestController::nodesFromRect):
1718
1719 2010-11-11  MORITA Hajime  <morrita@google.com>
1720
1721         Reviewed by Kent Tamura.
1722
1723         editing/selection/context-menu-on-text.html fails on chromium
1724         https://bugs.webkit.org/show_bug.cgi?id=45898
1725
1726         Enhance Chromium's EventSender::contextClick() to return
1727         array of menu strings. The implementation incomplete because
1728         Although actual context menu is implemented by the browser,
1729         it reflects editability of underlying context thus makes
1730         context-menu-on-text.html passed.
1731
1732         * DumpRenderTree/chromium/EventSender.cpp:
1733         (makeMenuItemStringsFor): Added.
1734         (EventSender::contextClick): Now returns array of menu strings.
1735         * DumpRenderTree/chromium/WebViewHost.cpp:
1736         (WebViewHost::showContextMenu): Now storing given WebContextMenuData object.
1737         (WebViewHost::clearContextMenuData): Added.
1738         (WebViewHost::lastContextMenuData): Added.
1739         * DumpRenderTree/chromium/WebViewHost.h:
1740
1741 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1742
1743         Unreviewed, build fix.
1744
1745         fix build breakage caused by fix for bug 49380 (r71858).
1746
1747         config_standalone needs to ensure that the packages it needs are
1748         in sys.path.
1749
1750         https://bugs.webkit.org/show_bug.cgi?id=49419
1751
1752         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
1753
1754 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1755
1756         Reviewed by Adam Roben.
1757
1758         Fix NRWT to respect set-webkit-configuration again :(
1759
1760         This change fixes a typo in config.py that was causing the wrong
1761         value to be read initially and us never actually looking into
1762         the filesystem to get the default configuration.
1763
1764         * Scripts/webkitpy/layout_tests/port/config.py:
1765         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
1766         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1767
1768 2010-11-11  Eric Seidel  <eric@webkit.org>
1769
1770         Reviewed by Adam Barth.
1771
1772         Split out Bug, Attachment and CommitterValidator from bugzilla.py
1773         https://bugs.webkit.org/show_bug.cgi?id=49403
1774
1775         CommitterValidator really had nothing to do with bugzilla.py.
1776         I've put it next to committers.py for now since its rather tied to that file.
1777
1778         * Scripts/webkitpy/common/config/committervalidator.py: Added.
1779         * Scripts/webkitpy/common/config/committervalidator_unittest.py: Added.
1780         * Scripts/webkitpy/common/net/bugzilla/__init__.py:
1781         * Scripts/webkitpy/common/net/bugzilla/attachment.py: Added.
1782         * Scripts/webkitpy/common/net/bugzilla/bug.py: Added.
1783         * Scripts/webkitpy/common/net/bugzilla/bug_unittest.py: Added.
1784         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
1785         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py:
1786         * Scripts/webkitpy/tool/bot/feeders.py:
1787         * Scripts/webkitpy/tool/commands/queues.py:
1788
1789 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1790
1791         Reviewed by Eric Seidel.
1792
1793         fix missing 'pid' argument for http lockfile checking on win32
1794         https://bugs.webkit.org/show_bug.cgi?id=49363
1795
1796         * Scripts/webkitpy/common/system/executive.py:
1797
1798 2010-11-11  Eric Seidel  <eric@webkit.org>
1799
1800         Rubber-stamped by Adam Barth.
1801
1802         Move bugzilla.py into its own module in preparation for splitting one-file-per-class
1803         https://bugs.webkit.org/show_bug.cgi?id=49402
1804
1805         * Scripts/webkitpy/common/net/bugzilla/__init__.py: Added.
1806         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla.py.
1807         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla_unittest.py.
1808
1809 2010-11-10  Chris Guillory  <chris.guillory@google.com>
1810
1811         Reviewed by Chris Fleizach.
1812
1813         Use IAccessibleComparable to compare accessibility objects.
1814         https://bugs.webkit.org/show_bug.cgi?id=49118
1815
1816         * DumpRenderTree/AccessibilityUIElement.cpp:
1817         (AccessibilityUIElement::isEqual):
1818         * DumpRenderTree/AccessibilityUIElement.h:
1819         * DumpRenderTree/win/AccessibilityUIElementWin.cpp:
1820         (comparableObject):
1821         (AccessibilityUIElement::isEqual):
1822
1823 2010-11-10  Csaba Osztrogonác  <ossy@webkit.org>
1824
1825         Reviewed by David Hyatt.
1826
1827         HTML5 Ruby support should be mandatory feature
1828         https://bugs.webkit.org/show_bug.cgi?id=49272
1829
1830         Remove Ruby as optional feature.
1831
1832         * Scripts/build-webkit:
1833
1834 2010-11-10  Adam Roben  <aroben@apple.com>
1835
1836         Print VC++ Express build logs in the same order that the projects are
1837         built
1838
1839         Fixes <http://webkit.org/b/49326> print-vse-failure-logs prints logs
1840         in an unhelpful order
1841
1842         Reviewed by Eric Seidel.
1843
1844         * Scripts/print-vse-failure-logs:
1845         (PrintVisualStudioExpressLogs._build_order): Added. Uses
1846         print-msvc-project-dependencies to figure out the order in which
1847         projects are built.
1848         (PrintVisualStudioExpressLogs._sort_buildlogs): Added. Sorts the logs
1849         based on their build order and project name.
1850         (PrintVisualStudioExpressLogs._obj_directory): Moved code to find the
1851         scripts directory from here...
1852         (PrintVisualStudioExpressLogs._scripts_directory): ...to here.
1853         (PrintVisualStudioExpressLogs.main): Sort the logs before printing
1854         them.
1855
1856 2010-11-10  Adam Roben  <aroben@apple.com>
1857
1858         Only print the interesting text from VC++ Express build logs
1859
1860         Fixes <http://webkit.org/b/49325> It's hard to find the interesting
1861         output from print-vse-failure-logs
1862
1863         Reviewed by Eric Seidel.
1864
1865         * Scripts/print-vse-failure-logs: Removed unnecessary os.path import,
1866         added newly-required imports. Removed a comment that was essentially a
1867         FIXME, and is now fixed.
1868         (PrintVisualStudioExpressLogs._relevant_text): Added. Uses
1869         BeautifulSoup to extract the relevant text from the build log.
1870         (PrintVisualStudioExpressLogs.main): Only print the relevant text.
1871
1872 2010-11-09  Mihai Parparita  <mihaip@chromium.org>
1873
1874         Reviewed by Tony Chang.
1875
1876         [Chromium] http/tests/security/XFrameOptions fail with Chromium DRT
1877         https://bugs.webkit.org/show_bug.cgi?id=49286
1878
1879         The tests in http/tests/security/XFrameOptions started to fail after
1880         r71297, since it changed the WebViewHost::assignIdentifierToRequest
1881         logic to always assign identifiers to requests, even if we're not going
1882         to be dumping resource load callbacks at that point. These tests
1883         only call LayoutTestController.dumpResourceLoadCallbacks after the
1884         request is started, thus their expectations have "<unknown>" as the
1885         identifier. Change WebViewHost to the previous behavior, which also
1886         match the other ports'.
1887
1888         * DumpRenderTree/chromium/WebViewHost.cpp:
1889         (WebViewHost::assignIdentifierToRequest):
1890         (WebViewHost::removeIdentifierForRequest):
1891
1892 2010-11-08  Tony Chang  <tony@chromium.org>
1893
1894         Reviewed by Adam Barth.
1895
1896         run platform/chromium/plugins/return-npobject.html on all platforms
1897         https://bugs.webkit.org/show_bug.cgi?id=49103
1898
1899         This tests that we can get an NPObject returned through a method on
1900         an NPAPI Object.
1901
1902         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1903         (testCallbackReturn):
1904         (pluginInvoke):
1905
1906 2010-11-09  Sheriff Bot  <webkit.review.bot@gmail.com>
1907
1908         Unreviewed, rolling out r71672.
1909         http://trac.webkit.org/changeset/71672
1910         https://bugs.webkit.org/show_bug.cgi?id=49276
1911
1912         It doesn't work on Chromium Windows bot (Requested by Ossy on
1913         #webkit).
1914
1915         * Scripts/webkitpy/common/system/file_lock.py: Removed.
1916         * Scripts/webkitpy/common/system/file_lock_unittest.py: Removed.
1917         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1918
1919 2010-11-09  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1920
1921         Reviewed by Tony Chang.
1922
1923         [NRWT] Make http locking similar to perl implementation
1924         https://bugs.webkit.org/show_bug.cgi?id=49187
1925
1926         * Scripts/webkitpy/common/system/file_lock.py: Added.
1927         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
1928         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1929
1930 2010-11-09  James Kozianski  <koz@chromium.org>
1931
1932         Reviewed by Eric Seidel.
1933
1934         Clean up imports for webkit-patch.
1935         https://bugs.webkit.org/show_bug.cgi?id=49083
1936
1937         Move some imports from commands into main.py where they are used.
1938         Move command imports into the __init__ of the commands module.
1939
1940         * Scripts/webkitpy/tool/commands/__init__.py:
1941         * Scripts/webkitpy/tool/commands/download.py:
1942         * Scripts/webkitpy/tool/commands/queues.py:
1943         * Scripts/webkitpy/tool/main.py:
1944
1945 2010-11-09  Yi Shen  <yi.4.shen@nokia.com>
1946
1947         Reviewed by Andreas Kling.
1948
1949         [Qt][QtTestBrowser] Don't erase incorrect url in the Url Bar
1950         https://bugs.webkit.org/show_bug.cgi?id=49047
1951
1952         * QtTestBrowser/mainwindow.cpp:
1953         (MainWindow::setAddressUrl):
1954
1955 2010-11-09  Andras Becsi  <abecsi@webkit.org>
1956
1957         Reviewed by Csaba Osztrogonác.
1958
1959         Improve old-run-webkit-tests --verbose to show DumpTool opening/closing
1960         https://bugs.webkit.org/show_bug.cgi?id=49252
1961
1962         * Scripts/old-run-webkit-tests:
1963
1964 2010-11-09  Csaba Osztrogonác  <ossy@webkit.org>
1965
1966         Reviewed by Tor Arne Vestbø.
1967
1968         Output of old-run-webkit-tests is buggy in verbose mode
1969         https://bugs.webkit.org/show_bug.cgi?id=49249
1970
1971         * Scripts/old-run-webkit-tests: Redundant outputs removed.
1972
1973 2010-11-08  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
1974
1975         Reviewed by Martin Robinson.
1976
1977         [GTK] Link with target name set does not work
1978         https://bugs.webkit.org/show_bug.cgi?id=48865
1979
1980         When a new page is created with a name (target=myFrame), the new
1981         mainFrame could not be found because they where not stored in the
1982         same PageGroup. As PageGroup are not exposed externally so the
1983         simpliest solution is to use a global page group name. This also fixes
1984         issue with visited link coloration across pages. After this change the
1985         private function webkit_web_view_set_group_name() was no longer used
1986         so it was removed completly.
1987
1988         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1989         (createWebView):
1990
1991 2010-11-08  John Knottenbelt  <jknotten@chromium.org>
1992
1993         Reviewed by Kent Tamura.
1994
1995         WebViewHost::reset() uses placement new.
1996         https://bugs.webkit.org/show_bug.cgi?id=49069
1997
1998         WebViewHost is using placement destruction / new to simulate a fresh
1999         WebViewHost object at the same address. This is because the WebView remains
2000         open across tests and maintains a pointer to the WebViewHost.
2001         This change resets member variables explictly instead of the placement new dance.
2002
2003         * DumpRenderTree/chromium/WebViewHost.cpp:
2004         (WebViewHost::WebViewHost):
2005         (WebViewHost::reset):
2006
2007 2010-11-08  Dirk Pranke <dpranke@chromium.org>
2008
2009         Reviewed by Eric Seidel.
2010
2011         Fix the webkit_base_dir logic in webkitpy/layout_tests/port/config.py.
2012         It turns out that NRWT can't use the code in scm.find_checkout_root()
2013         because the Chromium bots don't do full checkouts of the WebKit
2014         tree; they only check out subdirectories like WebKitTools/Scripts.
2015         Until we can figure out a better approach for this, I've
2016         restored the base_dir-detecting code from NRWT, which works in
2017         any directory tree, scm or no.
2018
2019         This also restores the files modified in r71475 and r71474.
2020
2021         https://bugs.webkit.org/show_bug.cgi?id=49151
2022
2023         * WebKitTools/Scripts/webkitpy/layout_tests/port/config.py
2024         * WebKitTools/Scripts/webkitpy/layout_tests/port/config_unittest.py
2025
2026 2010-11-08  Eric Seidel  <eric@webkit.org>
2027
2028         Reviewed by Adam Barth.
2029
2030         webkit-patch failure-reason explodes if a build is missing
2031         https://bugs.webkit.org/show_bug.cgi?id=49195
2032
2033         This is likely a recent regression.  Adding a unit test for this case.
2034
2035         * Scripts/webkitpy/common/net/buildbot.py:
2036         * Scripts/webkitpy/common/net/buildbot_unittest.py:
2037
2038 2010-11-08  Dan Bernstein  <mitz@apple.com>
2039
2040         Reviewed by Darin Adler.
2041
2042         Ignore the system scrollbar setting.
2043
2044         * DumpRenderTree/mac/DumpRenderTree.mm:
2045         (resetDefaultsToConsistentValues): Set the scrollbar default.
2046
2047 2010-11-05  Adam Roben  <aroben@apple.com>
2048
2049         Make webkitpy.common.system.executive_unittest pass when running under
2050         Win32 Python
2051
2052         Fixes <http://webkit.org/b/49033>.
2053
2054         Reviewed by Dave Levin and Eric Seidel.
2055
2056         * Scripts/webkitpy/common/system/executive.py:
2057         (Executive._run_command_with_teed_output): Pass the arguments through
2058         encode_argument_if_needed rather than using Cygwin-specific code here.
2059         (Executive.run_and_throw_if_fail): Use child_process_encoding to decode
2060         the output.
2061         (Executive.run_command): Use encode_argument_if_needed to encode the
2062         arguments and child_process_encoding to decode the output.
2063         (Executive._child_process_encoding): Returns the encoding that should be
2064         used when communicating with child processes. On Windows we use mbcs,
2065         which maps to the current code page. On all other platforms we use
2066         UTF-8.
2067         (Executive._should_encode_child_process_arguments): Returns True if
2068         arguments to child processes need to be encoded. This is currently
2069         only needed on Cygwin and Win32 Python 2.x.
2070         (Executive._encode_argument_if_needed): Encode the argument using
2071         child_process_encoding if we need to encode arguments to child
2072         processes on this platform.
2073
2074         * Scripts/webkitpy/common/system/executive_unittest.py:
2075         (never_ending_command): Added. Returns arguments to run a command that
2076         will not quit until we kill it. On Windows we use wmic, on other
2077         platforms we use yes.
2078         (ExecutiveTest.test_run_command_with_unicode): Changed to expect the
2079         mbcs encoding to be used and for output from the child processes to
2080         have been roundtripped through encode/decode on Win32 Python. When
2081         UTF-8 is the encoding the roundtripping is undetectable, but with mbcs
2082         it's possible that some characters will not be able to be converted
2083         and will be replaced by question marks; the round-tripping allows us
2084         to expect this result.
2085
2086         (ExecutiveTest.test_kill_process):
2087         (ExecutiveTest.test_kill_all):
2088         Use never_ending_command instead of invoking "yes" directly. Expect an
2089         exit code of 1 when using Win32 Python, as that's what seems to happen.
2090
2091 2010-11-08  Adam Roben  <aroben@apple.com>
2092
2093         Roll out r71532
2094
2095         It broke the build for Cygwin 1.7 installs. Cygwin 1.7's default
2096         .bashrc unsets %TEMP%, which broke copy-tools.cmd.
2097
2098         * Scripts/webkitdirs.pm:
2099
2100 2010-11-08  Tony Chang  <tony@chromium.org>
2101
2102         Reviewed by Adam Barth.
2103
2104         run platform/chromium/plugins/nested-plugin-objects.html on all platforms
2105         https://bugs.webkit.org/show_bug.cgi?id=49094
2106
2107         This tests that objects created by plugins are proplery cleaned up.
2108
2109         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
2110         (testPassTestObject):
2111         (pluginInvoke):
2112         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
2113         (testAllocate):
2114         (testDeallocate):
2115         (testGetProperty):
2116         (testConstruct):
2117
2118 2010-11-08  Adam Roben  <aroben@apple.com>
2119
2120         Mark Windows builds triggered from Perl as being non-interactive
2121
2122         This affects whether some of our scripts will show alerts vs. printing
2123         to the build log.
2124
2125         Fixes <http://webkit.org/b/49181> Windows build fail mysteriously when
2126         .vsprops files are updated
2127
2128         Reviewed by Steve Falkenburg.
2129
2130         * Scripts/webkitdirs.pm:
2131         (buildVisualStudioProject): Set WEBKIT_NONINTERACTIVE_BUILD to 1.
2132
2133 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2134
2135         Reviewed by Ojan Vafai.
2136
2137         Make http locking default in NRWT.
2138         https://bugs.webkit.org/show_bug.cgi?id=48053
2139
2140         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2141         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2142
2143 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2144
2145         Reviewed by Csaba Osztrogonác.
2146
2147         [NRWT] If the http lock fails we shouldn't do any locking
2148         https://bugs.webkit.org/show_bug.cgi?id=49164
2149
2150         If something goes wrong with the locking, the test should keep going.
2151
2152         * Scripts/webkitpy/layout_tests/port/http_lock.py:
2153
2154 2010-11-08  Adam Roben  <aroben@apple.com>
2155
2156         Switch back to using kCGImageAlphaPremultipliedFirst when generating
2157         pixel dumps on Windows
2158
2159         I changed this behavior in r71418 thinking that it was required for
2160         getting plugins to show up in pixel dumps. But it doesn't seem to be
2161         necessary, and was making it impossible to compare new Windows pixel
2162         dumps with existing Windows or Mac pixel dumps (because ImageDiff won't
2163         compare an image with alpha to an image without alpha).
2164
2165         Fixes <http://webkit.org/b/49172> REGRESION (r71418): Can't compare
2166         new Windows pixel results to existing Windows or Mac results
2167
2168         Reviewed by Antti Koivisto.
2169
2170         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
2171         (createBitmapContextFromWebView): Replaced kCGImageAlphaNoneSkipFirst
2172         with kCGImageAlphaPremultipliedFirst.
2173
2174 2010-11-08  Csaba Osztrogonac  <ossy@webkit.org>
2175
2176         Unreviewed, rolling out r71466.
2177         http://trac.webkit.org/changeset/71466
2178         https://bugs.webkit.org/show_bug.cgi?id=48865
2179
2180         It broke layout tests on GTK bots.
2181
2182         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2183         (createWebView):
2184
2185 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2186
2187         Reviewed by Csaba Osztrogonác.
2188
2189         Enable running of Qt API tests on BuildBot
2190         https://bugs.webkit.org/show_bug.cgi?id=49004
2191
2192         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
2193
2194 2010-11-08  Eric Seidel  <eric@webkit.org>
2195
2196         Reviewed by Adam Barth.
2197
2198         Add clean-review-queue command to remove closed bugs from the webkit.org/pending-review
2199         https://bugs.webkit.org/show_bug.cgi?id=49160
2200
2201         Bugzilla doesn't automatically remove r? when a bug gets closed.
2202         This script takes care of that for webkit.org.
2203
2204         * Scripts/webkitpy/common/net/bugzilla.py:
2205         * Scripts/webkitpy/tool/commands/upload.py:
2206
2207 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
2208
2209         Unreviewed, rolling out r71474.
2210         http://trac.webkit.org/changeset/71474
2211         https://bugs.webkit.org/show_bug.cgi?id=48280
2212
2213         breaks chromium webkit tests
2214         https://bugs.webkit.org/show_bug.cgi?id=49151
2215
2216         * Scripts/webkitpy/layout_tests/port/base.py:
2217         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2218         * Scripts/webkitpy/layout_tests/port/config.py:
2219         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
2220         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
2221         * Scripts/webkitpy/layout_tests/port/webkit.py:
2222         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2223         * Scripts/webkitpy/tool/mocktool.py:
2224
2225 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
2226
2227         Unreviewed, rolling out r71475.
2228         http://trac.webkit.org/changeset/71475
2229
2230         breaks chromium webkit tests
2231         https://bugs.webkit.org/show_bug.cgi?id=49151
2232
2233         * Scripts/webkitpy/common/newstringio.py: Removed.
2234         * Scripts/webkitpy/common/newstringio_unittest.py: Removed.
2235         * Scripts/webkitpy/common/system/executive_mock.py: Removed.
2236         * Scripts/webkitpy/common/system/filesystem_mock.py: Removed.
2237         * Scripts/webkitpy/layout_tests/port/config_mock.py: Removed.
2238
2239 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2240
2241         Unreviewed, build fix.
2242
2243         Add files inexplicably not committed in r71474 as part of the
2244         fix for bug 48280.
2245
2246         * Scripts/webkitpy/common/newstringio.py: Added.
2247         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
2248         * Scripts/webkitpy/common/system/executive_mock.py: Added.
2249         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
2250         * Scripts/webkitpy/layout_tests/port/config_mock.py: Added.
2251
2252 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2253
2254         Reviewed by Eric Siedel.
2255
2256         new-run-webkit-tests: update port/base and port/webkit to use the
2257         new FileSystem and Config abstractions, pulling more logic out of
2258         the base Port classes into separate, mockable objects.
2259
2260         Also create a MockFileSystem object, a MockConfig object, move
2261         MockExecutive into common/system to be next to executive, and
2262         update the config object to use a FileSystem.
2263
2264         https://bugs.webkit.org/show_bug.cgi?id=48280
2265
2266         * Scripts/webkitpy/common/newstringio.py: Added.
2267         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
2268         * Scripts/webkitpy/common/system/executive_mock.py: Added.
2269         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
2270         * Scripts/webkitpy/layout_tests/port/base.py:
2271         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2272         * Scripts/webkitpy/layout_tests/port/config.py:
2273         * Scripts/webkitpy/layout_tests/port/config_mock.py:
2274         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
2275         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
2276         * Scripts/webkitpy/layout_tests/port/webkit.py:
2277         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2278         * Scripts/webkitpy/tool/mocktool.py:
2279
2280 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2281
2282         Unreviewed, build breakage.
2283
2284         Apparently I uploaded the wrong version of the file to fix 49122
2285         and neither Eric or I noticed - it was missing a dirname() call.
2286         Fixing ...
2287
2288         https://bugs.webkit.org/show_bug.cgi?id=49122
2289
2290         * Scripts/webkitpy/common/checkout/scm.py:
2291
2292 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2293
2294         Reviewed by Eric Seidel.
2295
2296         webkitpy/tool/* unittests change cwd and don't clean up properly
2297
2298         https://bugs.webkit.org/show_bug.cgi?id=49122
2299
2300         * Scripts/webkitpy/common/checkout/scm.py:
2301
2302 2010-11-05  Chris Marrin  <cmarrin@apple.com>
2303
2304         Reviewed by Simon Fraser.
2305
2306         suspendAnimations/resumeAnimations not present in WebKit2
2307         https://bugs.webkit.org/show_bug.cgi?id=49109
2308
2309         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
2310         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
2311         (WTR::LayoutTestController::suspendAnimations):
2312         (WTR::LayoutTestController::resumeAnimations):
2313         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
2314
2315 2010-11-05  Tony Chang  <tony@chromium.org>
2316
2317         Reviewed by David Levin.
2318
2319         cleanup style in TestNetscapePlugIn/PluginObject.cpp
2320         https://bugs.webkit.org/show_bug.cgi?id=49044
2321
2322         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
2323         (getPluginClass):
2324         (pluginGetProperty):
2325         (pluginSetProperty):
2326         (variantToIdentifier):
2327         (testCallback):
2328         (getURL):
2329         (getURLNotify):
2330         (testInvokeDefault):
2331         (testGetProperty):
2332         (testDocumentOpen):
2333         (testWindowOpen):
2334         (pluginInvoke):
2335         (handleCallback):
2336         (notifyStream):
2337         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
2338
2339 2010-11-05  Eric Seidel  <eric@webkit.org>
2340
2341         Reviewed by David Levin.
2342
2343         Add QueueStatusServer/__init__.py so others can run the QueueStatusServer tests
2344         https://bugs.webkit.org/show_bug.cgi?id=49032
2345
2346         I wrote this file as part of bug 47847, but I forgot to commit it.
2347         No one else noticed it missing because test-webkitpy knows how
2348         to recover in the case where it can't import QueueStatusServer
2349         (which generally occurs due to not having installed the AppEngine SDK).
2350
2351         * QueueStatusServer/__init__.py: Added.
2352         * QueueStatusServer/model/workitems_unittest.py:
2353          - Remove a test which fails.  This was probably landed (by me)
2354            from my other machine, which since this __init__.py was missing
2355            I never noticed the failure and landed this invalid test.
2356            Sadly we can't really test remove_work_item as it depends
2357            on .key() working.  .key() will throw unless the object
2358            has already been saved it seems.
2359            This may be a mis-design in our remove_work_item implementation,
2360            but for now, just removing the test.
2361
2362 2010-11-04  Adam Roben  <aroben@apple.com>
2363
2364         Teach check-webkit-style about TestNetscapePlugIn
2365
2366         Fixes <http://webkit.org/b/49030> check-webkit-style is ignorant of
2367         TestNetscapePlugIn's coding conventions
2368
2369         Reviewed by John Sullivan.
2370
2371         * Scripts/webkitpy/style/checker.py: Lump TestNetscapePlugIn in with
2372         WebKitAPITest and TestWebKitAPI in having different include order and
2373         naming conventions than WebCore. Added some comments about why the
2374         exceptions exist.
2375
2376 2010-11-04  Adam Roben  <aroben@apple.com>
2377
2378         Add a test that shows that windowed plugins are able to paint
2379
2380         Somehow we never had a test for this before.
2381
2382         Fixes <http://webkit.org/b/49029> Should add a test that shows
2383         windowed plugins can paint
2384
2385         Reviewed by Jon Honeycutt.
2386
2387         * DumpRenderTree/TestNetscapePlugIn/Tests/win/DrawsGradient.cpp: Added.
2388         (DrawsGradient::DrawsGradient):
2389         (DrawsGradient::wndProc): We handle the WM_PAINT and WM_PRINTCLIENT messages.
2390
2391         (DrawsGradient::onPaint):
2392         (DrawsGradient::onPrintClient):
2393         These both just call through to paint.
2394
2395         (DrawsGradient::paint): Fills our client area with some gradients.
2396
2397         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Link
2398         against Msimg32.lib for ::GradientFill and added DrawsGradient.
2399
2400         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
2401         (createBitmapContextFromWebView): Use WM_PRINT instead of
2402         WM_PRINTCLIENT so that ::DefWindowProc will send
2403         WM_PRINT/WM_PRINTCLIENT messages to the WebView's child windows.
2404         Replaced kCGImageAlphaPremultipledFirst with kCGImageAlphaNoneSkipFirst
2405         because GDI doesn't support alpha and kCGImageBitmapOrder32Little with
2406         kCGImageBitmapOrder32Host because, who knows, maybe someday Windows
2407         will run on a big-endian platform.
2408
2409 2010-11-04  Adam Roben  <aroben@apple.com>
2410
2411         Extract much of NPNInvalidateRectInvalidatesWindow's code into a
2412         WindowedPluginTest base class
2413
2414         The base class takes care of subclassing the plugin's window so that a
2415         custom WNDPROC is called. This will make it easier to write tests that
2416         need to handle window messages.
2417
2418         Fixes <http://webkit.org/b/49028> It's hard to write a PluginTest with
2419         a custom WNDPROC
2420
2421         Reviewed by Jon Honeycutt.
2422
2423         * DumpRenderTree/TestNetscapePlugIn/Tests/win/NPNInvalidateRectInvalidatesWindow.cpp:
2424         (NPNInvalidateRectInvalidatesWindow::NPNInvalidateRectInvalidatesWindow):
2425         (NPNInvalidateRectInvalidatesWindow::NPP_SetWindow):
2426         (NPNInvalidateRectInvalidatesWindow::wndProc):
2427         (NPNInvalidateRectInvalidatesWindow::testInvalidateRect):
2428         Moved code from here to WindowedPluginTest. Changed to use window()
2429         instead of m_window.
2430
2431         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2432         Added WindowedPluginTest and added TestNetscapePlugIn/win to the
2433         include path. Also reordered the include path to make a little more
2434         sense and simplified the entry that added TestNetscapePlugIn itself to
2435         the include path.
2436
2437         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.cpp: Added.
2438         (WindowedPluginTest::WindowedPluginTest):
2439         (WindowedPluginTest::NPP_SetWindow):
2440         (WindowedPluginTest::staticWndProc):
2441         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.h: Added.
2442         (WindowedPluginTest::window):
2443         Code came from NPNInvalidateRectInvalidatesWindow.
2444
2445 2010-11-04  Adam Roben  <aroben@apple.com>
2446
2447         Add a plugin test to show that windowed plugins are clipped correctly
2448
2449         Fixes <http://webkit.org/b/49024> <rdar://problem/8487847> Windowed
2450         plugins aren't clipped in WebKit2 on Windows
2451
2452         Reviewed by Jon Honeycutt.
2453
2454         * DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowRegionIsSetToClipRect.cpp: Added.
2455         (WindowRegionIsSetToClipRect::WindowRegionIsSetToClipRect): Initialize members.
2456         (WindowRegionIsSetToClipRect::NPP_SetWindow): Check that our window
2457         region matches the clip rect we know we should have based on
2458         window-region-is-set-to-clip-rect.html, and check that our window class
2459         doesn't have the CS_PARENTDC style.
2460
2461         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2462         Added WindowRegionIsSetToClipRect.
2463
2464 2010-11-05  Alejandro G. Castro  <alex@igalia.com>
2465
2466         Reviewed by Martin Robinson.
2467
2468         [GTK] Avoid font hinting in the DRT
2469         https://bugs.webkit.org/show_bug.cgi?id=48548
2470
2471         Change the settings to avoid font hinting, it was causing
2472         different results depending on the distribution.
2473
2474         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2475         (initializeGtkFontSettings):
2476         * DumpRenderTree/gtk/fonts/fonts.conf:
2477
2478 2010-11-05  Eric Seidel  <eric@webkit.org>
2479
2480         Reviewed by Adam Barth.
2481
2482         build-webkit spams system.log by repeatedly checking for PBXProductDirectory in com.apple.Xcode
2483         https://bugs.webkit.org/show_bug.cgi?id=49051
2484
2485         This is a speculative fix.  The unit tests cover these methods, however
2486         I don't know if this will fully stop the system.log spam.
2487
2488         * Scripts/webkitpy/layout_tests/port/base.py:
2489         * Scripts/webkitpy/layout_tests/port/webkit.py:
2490
2491 2010-11-05  Eric Seidel  <eric@webkit.org>
2492
2493         Reviewed by Adam Barth.
2494
2495         Add basic support for showing bot id on /queue-status/ pages
2496         https://bugs.webkit.org/show_bug.cgi?id=49037
2497
2498         This support is really simple.  Eventually we'll want to
2499         show the bot id in the lock table too, but we don't have
2500         that information stored in the server yet.
2501
2502         * QueueStatusServer/handlers/queuestatus.py:
2503         * QueueStatusServer/handlers/queuestatus_unittest.py: Added.
2504         * QueueStatusServer/templates/includes/singlequeuestatus.html:
2505
2506 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
2507
2508         Reviewed by Adam Barth.
2509
2510         Create a filesystem wrapper that we can use to enforce
2511         particular conventions and use for mocking and dependency
2512         injection down the line.
2513
2514         https://bugs.webkit.org/show_bug.cgi?id=48144
2515
2516         * Scripts/webkitpy/common/system/filesystem.py: Added.
2517         * Scripts/webkitpy/common/system/filesystem_unittest.py: Added.
2518
2519 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2520
2521         Reviewed by Tony Chang.
2522
2523         run_webkit_tests_unittest fails under Python 2.5
2524         https://bugs.webkit.org/show_bug.cgi?id=49043
2525
2526         Switch from itertools.chain.from_iterable (which was added in 2.6)
2527         to using itertools.chain directly.
2528
2529         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2530
2531 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2532
2533         Unreviewed fix to webkit-patch rebaseline-server so that it runs under
2534         Python 2.5 (it needs an import to use the with statement).
2535
2536         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
2537
2538 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
2539
2540         Reviewed by Adam Barth.
2541
2542         new-run-webkit-tests: split out webkit-specific configuration stuff into a new module
2543
2544         The current NRWT code has webkit-specific configuration code (like
2545         _script_path, default configuration, etc.) mixed in with
2546         layout-test-specific stuff in port/base. The configuration code
2547         should be split out into a separate module for easier mocking,
2548         testing, and isolation.
2549
2550         https://bugs.webkit.org/show_bug.cgi?id=48264
2551
2552         * Scripts/webkitpy/layout_tests/port/config.py: Added.
2553         * Scripts/webkitpy/layout_tests/port/config_unittest.py: Added.
2554
2555 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2556
2557         Reviewed by Tony Chang.
2558
2559         Rebaseline server: initial framework
2560         https://bugs.webkit.org/show_bug.cgi?id=48892
2561
2562         Adds the basic framework for the rebaseline server (details at
2563         http://webkit.org/b/47761). Includes the rebaseline-server webkit-patch
2564         command, which starts an HTTP server that can serve static files or
2565         invoke handler methods on a class.
2566
2567         * Scripts/webkitpy/tool/commands/__init__.py:
2568         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html: Added.
2569         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css: Added.
2570         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js: Added.
2571         * Scripts/webkitpy/tool/commands/rebaselineserver.py: Added.
2572
2573 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
2574
2575         Reviewed by James Robinson.
2576
2577         new-run-webkit-tests wasn't using DRT by default for
2578         --platform chromium-gpu
2579
2580         The default value was set to False instead of None, which meant
2581         that the platform specific logic wasn't firing to change the
2582         value to True (b/c we were afraid we'd be overriding the user
2583         preference).
2584
2585         https://bugs.webkit.org/show_bug.cgi?id=49038
2586
2587         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2588
2589 2010-11-04  Tony Chang  <tony@chromium.org>
2590
2591         Reviewed by Adam Barth.
2592
2593         make platform/chromium/plugins/multiple-plugins.html pass on all platforms
2594         https://bugs.webkit.org/show_bug.cgi?id=49026
2595
2596         Have the testObject.bar property return the string "bar".  This lets
2597         us run and pass platform/chromium/plugins/multiple-plugins.html on all
2598         platforms.
2599
2600         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
2601         (testGetProperty):
2602
2603 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2604
2605         Reviewed by Tony Chang.
2606
2607         NRWT doesn't have good test coverage for --run-chunk, --batch-size, --run-part, etc.
2608         https://bugs.webkit.org/show_bug.cgi?id=48878
2609
2610         Add get_tests_run so that it's easy to see which tests get run (and with
2611         what batching) for a given flag combination. Flesh out the various
2612         test cases that have FIXMEs.
2613
2614         Also fixes an off-by-one error (batch sizes were one larger than
2615         expected) and makes --run-part also have wraparound behavior, like
2616         --run-chunk.
2617
2618         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2619         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2620         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2621
2622 2010-11-04  Erik Arvidsson  <arv@chromium.org>
2623
2624         Reviewed by Dimitri Glazkov.
2625
2626         Support box-sizing without the vendor prefix
2627         https://bugs.webkit.org/show_bug.cgi?id=36713
2628
2629         Based on patch by Peter Beverloo <peter@lvp-media.com>
2630
2631         * iExploder/htdocs/cssproperties.in: Change -webkit-box-sizing to box-sizing.
2632
2633 2010-11-04  Csaba Osztrogonác  <ossy@webkit.org>
2634
2635         Unreviewed rollout r71340, because it broke Chromium Windows bot.
2636
2637         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2638         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2639
2640 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2641
2642         Reviewed by Ojan Vafai.
2643
2644         Make http locking default in NRWT.
2645         https://bugs.webkit.org/show_bug.cgi?id=48053
2646
2647         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2648         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2649
2650 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2651
2652         Reviewed by Eric Seidel.
2653
2654         [NRWT] Clear invalid http locks on Windows platform as well
2655         https://bugs.webkit.org/show_bug.cgi?id=48515
2656
2657         * Scripts/webkitpy/common/system/executive.py:
2658         * Scripts/webkitpy/common/system/executive_unittest.py:
2659         * Scripts/webkitpy/layout_tests/port/http_lock.py:
2660
2661 2010-11-02  Adam Roben  <aroben@apple.com>
2662
2663         Reduce our dependence on coreutils when running Python tests
2664
2665         This patch introduces versions of the standard echo and cat utilities
2666         implemented in Python. They are probably missing features of their
2667         coreutils equivalents, but they can do what's necessary for our Python
2668         tests. This is useful on Windows, where these utilities typically
2669         aren't available.
2670
2671         Fixes <http://webkit.org/b/48883> executive_unittest relies on echo
2672         and cat utilities from coreutils, which are not present on Windows
2673
2674         Reviewed by Eric Seidel.
2675
2676         * Scripts/webkitpy/common/system/executive_unittest.py: Changed to use
2677         our Python-based echo and cat.
2678
2679         * Scripts/webkitpy/common/system/fileutils.py: Added.
2680         (make_stdout_binary): On Windows, puts sys.stdout into binary mode so
2681         that \n won't be translated into \r\n. I couldn't think of a good way
2682         to test this directly without touching the filesystem, but it is tested
2683         indirectly by echo_unittest.
2684
2685         * Scripts/webkitpy/test/cat.py: Added.
2686         (command_arguments): Returns a list for invoking cat with the given arguments.
2687         (main): Acts like a simplified version of the coreutils cat utility.
2688
2689         * Scripts/webkitpy/test/cat_unittest.py: Added.
2690         (CatTest.assert_cat): Runs cat with the given input and ensures the
2691         output matches the input.
2692         (CatTest.test_basic): Performs a simple test of cat.
2693         (CatTest.test_no_newline): Tests what happens when the input string
2694         doesn't have a trailing newline.
2695         (CatTest.test_unicode): Tests passing a unicode string to cat.
2696         (CatTest.test_as_command): Tests running cat as a separate command.
2697
2698         * Scripts/webkitpy/test/echo.py: Added.
2699         (command_arguments): Returns a list for invoking echo with the given arguments.
2700         (main): Acts like a simplified version of the coreutils echo utility.
2701
2702         * Scripts/webkitpy/test/echo_unittest.py: Added.
2703         (EchoTest.test_basic): Performs a simple test of echo.
2704         (EchoTest.test_no_newline): Tests passing -n to echo to suppress the
2705         trailing newline.
2706         (EchoTest.test_unicode): Tests passing unicode and non-unicode strings
2707         to echo.
2708         (EchoTest.test_argument_order): Tests what happens when -n is not the
2709         first argument.
2710         (EchoTest.test_empty_arguments): Tests what happens when you pass [] to
2711         echo.main.
2712         (EchoTest.test_no_arguments): Tests what happens when you call
2713         echo.main with no arguments.
2714         (EchoTest.test_as_command): Tests running echo as a separate command.
2715
2716 2010-11-04  Renata Hodovan  <reni@inf.u-szeged.hu>
2717
2718         Unreviewed: Add myself to the list of Committers.
2719
2720         * Scripts/webkitpy/common/config/committers.py:
2721
2722 2010-11-04  Andreas Kling  <kling@webkit.org>
2723
2724         Reviewed by Kenneth Rohde Christiansen.
2725
2726         [Qt] Hook into QWebPage::scrollRequested for resizes-to-contents mode
2727
2728         In RTC mode the QGraphicsWebView item is the size of the contents,
2729         scrolling works a bit differently  (we need to react to scrollRequested.)
2730
2731         Normally QGraphicsView will replay the last mouse event when scrolling,
2732         so to prevent WebKit from getting confused by this we temporarily make
2733         the QGraphicsView non-interactive.
2734
2735         * QtTestBrowser/webview.cpp:
2736         (WebViewGraphicsBased::setPage):
2737         (WebViewGraphicsBased::scrollRequested):
2738         * QtTestBrowser/webview.h:
2739
2740 2010-11-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2741
2742         Reviewed by Andreas Kling.
2743
2744         Use OS(MAC_OS_X) rather than PLATFORM(MAC) in TestControllerQt
2745
2746         PLATFORM(MAC) is not defined for the Qt port, as it refers to
2747         the Mac-port, not the Mac OS X operating system.
2748
2749         * WebKitTestRunner/qt/TestControllerQt.cpp:
2750         (WTR::isExistingLibrary):
2751
2752 2010-11-04  Leonid Ebril  <leonid.ebril@nokia.com>
2753
2754         Reviewed by Andreas Kling.
2755
2756         [Qt] Adding iphone user agent string the user agent list for QtTestBrowser
2757         https://bugs.webkit.org/show_bug.cgi?id=48869
2758
2759         * QtTestBrowser/useragentlist.txt:
2760
2761 2010-11-03  Adam Roben  <aroben@apple.com>
2762
2763         Always use uppercase drive names in strings returned by abspath_to_uri
2764
2765         Some versions of cygpath use lowercase drive letters while others use
2766         uppercase, which makes it hard to test the output of code that uses
2767         cygpath.
2768
2769         Fixes <http://webkit.org/b/48914> webkitpy.common.system.path_unittest
2770         fails with Cygwin 1.5
2771
2772         Reviewed by Eric Seidel.
2773
2774         * Scripts/webkitpy/common/system/path.py:
2775         (cygpath): Updated the docstring to indicate that only absolute paths
2776         should be passed for now (though relative paths will work fine).
2777         (_Cygpath.convert): Upper-case the first letter of the converted Windows path.
2778
2779 2010-11-03  George Guo  <George.Guo@Nokia.com>
2780
2781         Reviewed by Andreas Kling.
2782
2783         [Qt] QtTestBrowser : set mmp rule pageddata in Symbian
2784         https://bugs.webkit.org/show_bug.cgi?id=48767
2785
2786         Paging is needd on Symbian devices to support benchmarks tests like
2787         dromaeo.com and Celtic Kane that need a lot of memory to run
2788
2789         * QtTestBrowser/QtTestBrowser.pro:
2790
2791 2010-11-03  Jenn Braithwaite  <jennb@chromium.org>
2792
2793         Reviewed by Dmitry Titov.
2794
2795         Chromium: Update resource tracking when moving a frame between documents
2796         https://bugs.webkit.org/show_bug.cgi?id=48363
2797
2798         * DumpRenderTree/chromium/WebViewHost.cpp:
2799         (WebViewHost::assignIdentifierToRequest):
2800         Always put resource id in map so we can make assumptions about its
2801         presence.
2802         (WebViewHost::removeIdentifierForRequest):
2803         Added.
2804         (WebViewHost::didFinishResourceLoad):
2805         (WebViewHost::didFailResourceLoad):
2806         Check existence of resource id before removing from map.
2807         * DumpRenderTree/chromium/WebViewHost.h:
2808
2809 2010-11-03  Victor Wang  <victorw@chromium.org>
2810
2811         Reviewed by Adam Barth.
2812
2813         [Chromium] update buildbot names in chromium rebaseline tool.
2814
2815         https://bugs.webkit.org/show_bug.cgi?id=48881
2816
2817         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2818
2819 2010-11-03  Darin Adler  <darin@apple.com>
2820
2821         Updated Xcode projects by opening them with Xcode 3.2.4.
2822         Updated svn:ignore for Xcode projects.
2823
2824         * MiniBrowser/MiniBrowser.xcodeproj: Added property svn:ignore.
2825         * TestWebKitAPI/TestWebKitAPI.xcodeproj: Added property svn:ignore.
2826         * WebKitLauncher/WebKitLauncher.xcodeproj: Modified property svn:ignore.
2827         * WebKitLauncher/WebKitLauncher.xcodeproj/project.pbxproj: Updated with Xcode 3.2.4.
2828
2829 2010-11-03  Adam Roben  <aroben@apple.com>
2830
2831         Stop waiting for messages from the web process after a timeout elapses
2832
2833         TestController::runUntil now takes a timeout parameter that specifies
2834         how long to wait for the condition to become true. WebKitTestRunner
2835         takes a --timeout flag that specifies how long the timeout should be.
2836         run-webkit-tests passes this flag to WebKitTestRunner so its timeout
2837         will be similar to run-webkit-tests.
2838
2839         Fixes <http://webkit.org/b/43047> <rdar://problem/8365833>
2840         WebKitTestRunner waits forever if the web process crashes
2841
2842         Reviewed by Darin Adler and Anders Carlsson.
2843
2844         * Scripts/old-run-webkit-tests:
2845         (top level): Moved the GuardMalloc timeout adjustment here from
2846         readFromDumpToolWithTimer.
2847         (openDumpTool): Make WTR use a timeout similar to but slightly shorter
2848         than the one that was specified on the command line.
2849
2850         * WebKitTestRunner/TestController.cpp:
2851         (WTR::TestController::TestController): Initialize our timeout values.
2852         (WTR::TestController::initialize): Parse the --timeout flag and use it
2853         to modify our timeout values.
2854         (WTR::TestController::resetStateToConsistentValues): Changed to use a
2855         short timeout while waiting for the web process to reset and to return
2856         a boolean indicating whether we were able to reset the web process.
2857         Uses a 5-second timeout while waiting for the process to be reset.
2858         (WTR::TestController::runTest): Changed to return a boolean indicating
2859         whether we were able to reset the web process (and thus run the test).
2860         (WTR::TestController::runUntil): Call through to platformRunUntil.
2861
2862         (WTR::TestController::runTestingServerLoop):
2863         (WTR::TestController::run):
2864         Changed to bail if any test can't be run. This will cause the process
2865         to exit. (Unfortunately this will make run-webkit-tests think we
2866         crashed; see <http://webkit.org/b/48943>.)
2867
2868         * WebKitTestRunner/TestController.h: Added platformRunUntil,
2869         m_longTimeout, and m_shortTimeout.
2870
2871         * WebKitTestRunner/TestInvocation.cpp:
2872         (WTR::TestInvocation::invoke): Use a short timeout when waiting for the
2873         initial response and a long timeout when waiting for the test to
2874         complete. Check whether runUntil timed out and print an error message
2875         if so.
2876
2877         * WebKitTestRunner/mac/TestControllerMac.mm:
2878         (WTR::TestController::platformRunUntil): Renamed from runUntil. Pass
2879         [NSDate distantPast] to -[NSRunLoop runMode:beforeDate:] so that we
2880         won't block waiting for the run loop. Only loop until the timeout
2881         elapses.
2882
2883         * WebKitTestRunner/qt/TestControllerQt.cpp:
2884         (WTR::TestController::platformRunUntil): Renamed from runUntil. Added a
2885         FIXME about honoring the timeout.
2886
2887         * WebKitTestRunner/win/TestControllerWin.cpp:
2888         (WTR::TestController::platformRunUntil): Renamed from runUntil. Use
2889         ::MsgWaitForMultipleObjectsEx to implement the timeout. Changed to use
2890         ::PeekMessageW so that we don't block waiting for messages to become
2891         available.
2892
2893 2010-11-03  Adam Roben  <aroben@apple.com>
2894
2895         Add a plugin test that evaluates JS after removing the plugin element
2896         from the document
2897
2898         This test replaces platform/win/plugins/plugin-delayed-destroy.html.
2899         That test was made to prevent a crash very similar to this one, but
2900         unfortunately tested only the mechanism that prevented the crash and
2901         not whether the crash itself was prevented. Since WebKit2 uses a
2902         different mechanism to prevent the crash, the test was failing even
2903         though WebKit2 was not vulnerable to the crash. This new test crashes
2904         if there is no mechanism in place to prevent it and passes in both
2905         WebKit1 and WebKit2.
2906
2907         Fixes <http://webkit.org/b/46711> <rdar://problem/8485903>
2908         platform/win/plugins/plugin-delayed-destroy.html fails in WebKit2
2909
2910         Reviewed by Anders Carlsson.
2911
2912         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
2913         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2914         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
2915         * GNUmakefile.am:
2916         Added new file.
2917
2918         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
2919         (pluginDeallocate): Make sure we delete the PluginTest object. This
2920         prevents a leak and also allows us to test the crash.
2921
2922         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
2923         (PluginTest::executeScript): Made this into a non-static member
2924         function.
2925
2926         (PluginTest::waitUntilDone):
2927         (PluginTest::notifyDone):
2928         Updated for changes to executeScript.
2929
2930         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Added executeScript.
2931
2932         * DumpRenderTree/TestNetscapePlugIn/Tests/EvaluateJSAfterRemovingPluginElement.cpp: Added.
2933         (EvaluateJSAfterRemovingPluginElement::EvaluateJSAfterRemovingPluginElement):
2934         Initialize ourselves and tell the test harness to wait.
2935         (EvaluateJSAfterRemovingPluginElement::NPP_DestroyStream): Remove our
2936         plugin element from the document, then execute some JavaScript. If
2937         WebKit does not have appropriate mechanisms in place, we'll be
2938         destroyed inside the first call to executeScript and crash on the
2939         second call.
2940
2941 2010-11-02  Stephen White  <senorblanco@chromium.org>
2942
2943         Reviewed by Tony Chang.
2944
2945         [chromium] Fix LayoutTestController UMRs.
2946         https://bugs.webkit.org/show_bug.cgi?id=48872
2947
2948         * DumpRenderTree/chromium/LayoutTestController.cpp:
2949         (LayoutTestController::LayoutTestController):
2950
2951 2010-11-03  Kent Tamura  <tkent@chromium.org>
2952
2953         Reviewed by Dimitri Glazkov.
2954
2955         REGRESSION: rebaseline-chromium-webkit-tests uses non-zero tolerance for
2956         image dup detection
2957         https://bugs.webkit.org/show_bug.cgi?id=48744
2958
2959         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2960          - Make a function for option parsing for ease of test
2961          - Set 0 to options.tolerance
2962         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
2963          - Add a test for this change
2964
2965 2010-11-02  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
2966
2967         Unreviewed: Add myself to the list of Committers.
2968
2969         * Scripts/webkitpy/common/config/committers.py:
2970
2971 2010-11-02  Anders Carlsson  <andersca@apple.com>
2972
2973         Fix build.
2974
2975         * TestWebKitAPI/PlatformUtilities.cpp:
2976         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
2977
2978 2010-11-02  Sam Weinig  <sam@webkit.org>
2979
2980         Reviewed by Anders Carlsson.
2981
2982         Add a way to send startup messages on the context which can be posted when a process launches
2983         <rdar://problem/8617928>
2984         https://bugs.webkit.org/show_bug.cgi?id=48838
2985
2986         * MiniBrowser/mac/WebBundle/WebBundleMain.m:
2987         (WKBundleInitialize):
2988         * TestWebKitAPI/InjectedBundleController.cpp:
2989         (TestWebKitAPI::InjectedBundleController::initialize):
2990         (TestWebKitAPI::InjectedBundleController::didReceiveMessage):
2991         * TestWebKitAPI/InjectedBundleController.h:
2992         * TestWebKitAPI/InjectedBundleMain.cpp:
2993         (WKBundleInitialize):
2994         * TestWebKitAPI/PlatformUtilities.cpp:
2995         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
2996         * WebKitTestRunner/InjectedBundle/InjectedBundleMain.cpp:
2997         (WKBundleInitialize):
2998         Update implementations of WKBundleInitialize to take an initial userData
2999         argument.  Change TestWebKitAPI to use the new initial userData to initialize
3000         each test's bundle.
3001
3002 2010-11-02  Benjamin Kalman  <kalman@google.com>
3003
3004         Reviewed by Ojan Vafai.
3005
3006         new-run-webkit-tests doesn't strip "LayoutTests/" from prefix, unlike old-run-webkit-tests
3007         https://bugs.webkit.org/show_bug.cgi?id=48794
3008
3009         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: Strip the "LayoutTests/" prefix from test argument paths.
3010
3011 2010-11-02  Adam Roben  <aroben@apple.com>
3012
3013         Skip webkitpy.layout_tests.run_webkit_tests_unittest.MainTest on Cygwin
3014         Python 2.5.x
3015
3016         It is known to hang on that version of Python. See
3017         <http://webkit.org/b/48614>.
3018
3019         Reviewed by Adam Barth.
3020
3021         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py: Use
3022         skip_if to skip MainTest on Cygwin Python 2.5.x.
3023
3024         * Scripts/webkitpy/test/skip.py: Added.
3025         (skip_if): If the passed-in condition is false, find all the test_*
3026         methods of the given class and replace them with a function that just
3027         logs that we're skipping these tests. This is loosely based on the
3028         unittest.skip_if decorator added in Python 3.1.
3029         (_skipped_method): Returns a function that just logs that the tests are
3030         being skipped. This is used in place of the actual test_* functions.
3031
3032         * Scripts/webkitpy/test/skip_unittest.py: Added.
3033         (SkipTest.setUp): Create a logger for ourselves and give it a handler
3034         that logs to a private stream.
3035         (SkipTest.tearDown): Reset the logger.
3036         (SkipTest.create_fixture_class): Returns a class that we use to test
3037         skip_if. It has a single test_* method, test_foo, that just calls
3038         through to a callback.
3039         (SkipTest.foo_callback): Record that test_foo was called.
3040         (SkipTest.test_skip_if_false): Pass skip_if a False condition and test
3041         that test_foo does get called.
3042         (SkipTest.test_skip_if_true): Pass skip_if a True condition and test
3043         that test_foo does not get called and the appropriate message gets
3044         logged.
3045
3046 2010-11-02  Adam Barth  <abarth@webkit.org>
3047
3048         Reviewed by Eric Seidel.
3049
3050         webkit-patch should tell check-webkit-style which files were changed so
3051         check-webkit-style doesn't have to stat the whole working copy again
3052         https://bugs.webkit.org/show_bug.cgi?id=48792
3053
3054         * Scripts/webkitpy/tool/mocktool.py:
3055         * Scripts/webkitpy/tool/steps/checkstyle.py:
3056
3057 2010-11-02  Robert Kroeger  <rjkroege@chromium.org>
3058
3059         Reviewed by James Robinson.
3060
3061         [Chromium/DRT] Make EventSendingController honour leapForward for touch events.
3062         https://bugs.webkit.org/show_bug.cgi?id=48777
3063
3064         * DumpRenderTree/chromium/EventSender.cpp:
3065         (EventSender::sendCurrentTouchEvent):
3066
3067 2010-11-02  Adam Roben  <aroben@apple.com>
3068
3069         Only track resource identifiers in DRT when dumpResourceLoadCallbacks
3070         is on
3071
3072         This reverts Windows to our pre-r71097 behavior. That patch made us
3073         track all resource identifiers, including the main resource, so the
3074         main resource's URL started appearing in test output instead of
3075         "<unknown>". Arguably having the main resource's URL is better, but all
3076         other platforms print "<unknown>" and we want to match.
3077
3078         Fixes <http://webkit.org/b/48837> <rdar://problem/8620351> REGRESSION
3079         (r71097): Many http tests failing on Windows
3080
3081         Reviewed by Anders Carlsson.
3082
3083         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
3084         (ResourceLoadDelegate::identifierForInitialRequest): Don't add the
3085         identifier to the URL map if we're not supposed to dump resource load
3086         callbacks.
3087         (ResourceLoadDelegate::removeIdentifierForRequest): Always remove the
3088         identifier from the URL map even if we're already "done". There's no
3089         point in keeping out-of-date identifiers around.
3090
3091 2010-11-01  Jenn Braithwaite  <jennb@chromium.org>
3092
3093         Reviewed by Adam Roben.
3094
3095         Windows: Update resource tracking when moving a frame between documents
3096         https://bugs.webkit.org/show_bug.cgi?id=48364
3097
3098         * DumpRenderTree/win/DumpRenderTree.cpp:
3099         (createWebViewAndOffscreenWindow):
3100         (main):
3101         Give each WebView its own ResourceLoadDelegate instance in order to
3102         make assertions about resource ids on a particular WebView.
3103         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
3104         (ResourceLoadDelegate::identifierForInitialRequest):
3105         Always add id to the map.
3106         (ResourceLoadDelegate::removeIdentifierForRequest):
3107         Added.
3108         (ResourceLoadDelegate::willSendRequest):
3109         (ResourceLoadDelegate::didReceiveAuthenticationChallenge):
3110         (ResourceLoadDelegate::didReceiveResponse):
3111         (ResourceLoadDelegate::didFinishLoadingFromDataSource):
3112         (ResourceLoadDelegate::didFailLoadingWithError):
3113         (ResourceLoadDelegate::descriptionSuitableForTestResult):
3114         Replace static descriptionSuitableForTestResult with static member function to access identifier map.
3115         * DumpRenderTree/win/ResourceLoadDelegate.h:
3116         (ResourceLoadDelegate::urlMap):
3117         Moved within class so that each WebView has its own id map.
3118
3119 2010-11-01  Ojan Vafai  <ojan@chromium.org>
3120
3121         Reviewed by Dimitri Glazkov.
3122
3123         remove debug code from run_webkit_tests.py
3124         https://bugs.webkit.org/show_bug.cgi?id=48800
3125
3126         Remove temporary debug code and make --master-name required
3127         if --test-results-server is set now that all clients set
3128         --master-name.
3129
3130         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3131
3132 2010-11-01  Tony Chang  <tony@chromium.org>
3133
3134         Reviewed by Kent Tamura.
3135
3136         add plugin event logging for linux
3137         https://bugs.webkit.org/show_bug.cgi?id=48779
3138
3139         This is taken from Chromium's fork of the layout test plugin:
3140         http://git.chromium.org/gitweb/?p=chromium.git/.git;a=blob;f=webkit/tools/npapi_layout_test_plugin/main.cpp;h=3ebdada2f049b3624756438cff852364f86a2ede;hb=HEAD#l348
3141
3142         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
3143         (handleEventX11):
3144         (NPP_HandleEvent):
3145
3146 2010-11-01  Dirk Pranke  <dpranke@chromium.org>
3147
3148         Reviewed by James Robinson.
3149
3150         new-run-webkit-tests: use DRT, child-processes=1 for GPU tests by default
3151         https://bugs.webkit.org/show_bug.cgi?id=48790
3152
3153         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
3154         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
3155
3156 2010-11-01  Adam Barth  <abarth@webkit.org>
3157
3158         Reviewed by Eric Seidel.
3159
3160         Teach check-webkit-style how to accept a list of files to diff on the
3161         command line
3162         https://bugs.webkit.org/show_bug.cgi?id=48784
3163
3164         In a future patch, webkit-patch will use this option to improve
3165         performance.  I'm landing this in two pieces to avoid causing a version
3166         skew problem for the style-bot.
3167
3168         * Scripts/check-webkit-style:
3169         * Scripts/webkitpy/style/optparser.py:
3170         * Scripts/webkitpy/style/optparser_unittest.py:
3171         * Scripts/webkitpy/style_references.py:
3172
3173 2010-11-01  Anders Carlsson  <andersca@apple.com>
3174
3175         Reviewed by John Sullivan.
3176
3177         Tear down the related WebProcessProxy when a WebContext is deallocated
3178         https://bugs.webkit.org/show_bug.cgi?id=48769
3179
3180         * TestWebKitAPI/Tests/WebKit2/FailedLoad.cpp:
3181         (TestWebKitAPI::didFailProvisionalLoadWithErrorForFrame):
3182         We don't support empty URLs anymore, update test to expect a null URL instead.
3183
3184 2010-11-01  Søren Gjesse  <sgjesse@chromium.org>
3185
3186         Reviewed by Andreas Kling.
3187
3188         Fix warning when compiling the chromium port of DumpRenderShell
3189         with clang.
3190         https://bugs.webkit.org/show_bug.cgi?id=48414
3191
3192         * DumpRenderTree/chromium/TestShell.h:
3193         (TestShell::javaScriptFlagsForLoad):
3194
3195 2010-11-01  Adam Roben  <aroben@apple.com>
3196
3197         Fix typo from r71022
3198
3199         * BuildSlaveSupport/build.webkit.org-config/config.json:
3200
3201 2010-11-01  Adam Roben  <aroben@apple.com>
3202
3203         Trigger the Windows Release WebKit2 tests when a Release build
3204         finishes, not when a Debug build finishes
3205
3206         Fixes <http://webkit.org/b/48754> Windows Release WebKit2 tests are
3207         triggered at the wrong time
3208
3209         Reviewed by Sam Weinig.
3210
3211         * BuildSlaveSupport/build.webkit.org-config/config.json: Fixed
3212         triggerable name and trigger.
3213
3214 2010-11-01  Mario Sanchez Prada  <msanchez@igalia.com>
3215
3216         Unreviewed. Adding my IRC nickname to the list of committers.
3217
3218         * Scripts/webkitpy/common/config/committers.py:
3219
3220 2010-10-31  Robert Hogan  <robert@webkit.org>
3221
3222         Reviewed by Antonio Gomes.
3223
3224         [Qt] [Gtk] Plug-ins having upper case in mime type are failing to load
3225
3226         Qt and Gtk are case-sensitive when storing the declared mime-type
3227         of plugins. Since plugin mime-types are lowercased prior to searching
3228         for them in the plugin database, ensure they are loaded with the
3229         mime-type in lower case too.
3230
3231         Change the test netscape plugin to declare its mimetype in sentence
3232         case so that the correct behaviour is enforced.
3233
3234         https://bugs.webkit.org/show_bug.cgi?id=36815
3235
3236         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
3237         (NP_GetMIMEDescription):
3238
3239 2010-10-31  Robert Hogan  <robert@webkit.org>
3240
3241         Reviewed by Antonio Gomes.
3242
3243         [Qt] Support nodesFromRect in DRT
3244
3245         https://bugs.webkit.org/show_bug.cgi?id=48716
3246
3247         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3248         (LayoutTestController::nodesFromRect):
3249         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3250
3251 2010-10-30  Andreas Kling  <kling@webkit.org>
3252
3253         Reviewed by Antonio Gomes.
3254
3255         [Qt] QtTestBrowser: Switching view type moves the embedded inspector
3256         https://bugs.webkit.org/show_bug.cgi?id=48705
3257
3258         Reinsert the embedded inspector into the splitter after changing
3259         between QWebView/QGraphicsWebView.
3260
3261         * QtTestBrowser/launcherwindow.cpp:
3262         (LauncherWindow::init):
3263         (LauncherWindow::initializeView):
3264         * QtTestBrowser/webinspector.h:
3265         (WebInspector::WebInspector):
3266
3267 2010-10-28  Antonio Gomes  <agomes@rim.com>
3268
3269         Reviewed by Ojan Vafai.
3270
3271         Needs a "LinuxEditingBehavior", perhaps with a better name
3272         https://bugs.webkit.org/show_bug.cgi?id=36627
3273
3274         Adding support to Mac's, GTK+'s, Windows' and Chromium's LayoutTestController class to test the newly introduced Unix editing behavior.
3275
3276         * DumpRenderTree/chromium/LayoutTestController.cpp:
3277         (LayoutTestController::setEditingBehavior):
3278         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3279         (LayoutTestController::setEditingBehavior):
3280         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3281         (LayoutTestController::setEditingBehavior):
3282         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3283         (LayoutTestController::setEditingBehavior):
3284
3285 2010-10-29  Andreas Kling  <kling@webkit.org>
3286
3287         Reviewed by Kenneth Rohde Christiansen.
3288
3289         [Qt] QtTestBrowser: Add keyboard shortcut to toggle full screen (F11)
3290         https://bugs.webkit.org/show_bug.cgi?id=48695
3291
3292         * QtTestBrowser/launcherwindow.cpp:
3293         (LauncherWindow::createChrome):
3294
3295 2010-10-29  Andreas Kling  <kling@webkit.org>
3296
3297         Reviewed by Kenneth Rohde Christiansen.
3298
3299         [Qt] QtTestBrowser: Fix uninitialized read in FpsTimer
3300         https://bugs.webkit.org/show_bug.cgi?id=48675
3301
3302         FpsTimer::m_timer was never initialized and passed to QObject::killTimer()
3303         on startup with in -graphicsbased mode.
3304
3305         * QtTestBrowser/fpstimer.cpp:
3306         (FpsTimer::FpsTimer):
3307         (FpsTimer::numFrames):
3308         (FpsTimer::stop):
3309         (FpsTimer::timerEvent):
3310         * QtTestBrowser/fpstimer.h:
3311
3312 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
3313
3314         Reviewed by Tony Chang.
3315
3316         new-run-webkit-tests: change TestResults to be serializable
3317
3318         In preparation for changing new-run-webkit-tests from
3319         multithreaded to multiprocess, we need to make sure the data
3320         going between the threads is easily serialized over a socket.
3321
3322         This change adds serialization/pickling for the TestResults and
3323         TestFailure objects (using cPickle).
3324
3325         The TestFailure objects included a "has_wdiff" flag for Text
3326         results, but the flag wasn't being used, so I've removed it,
3327         simplifying the state to basically a set of enum objects with
3328         associated methods.
3329
3330         https://bugs.webkit.org/show_bug.cgi?id=48616
3331
3332         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3333         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
3334         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
3335         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
3336         * Scripts/webkitpy/layout_tests/layout_package/test_results.py: Added.
3337         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:Added.
3338         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3339         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
3340
3341 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
3342
3343         Reviewed by Ojan Vafai.
3344
3345         new-run-webkit-tests: need to provide separate dashboard results for GPU tests
3346         https://bugs.webkit.org/show_bug.cgi?id=48687
3347
3348         Modify the steps to upload the JSON files to the dashboards so
3349         that the GPU tests don't conflict with the regular tests on a
3350         port. We do this by modifying the --builder-name parameter to
3351         append " - GPU", which should cause the app to treat the results
3352         as a completely new builder. This is a little non-obvious, but
3353         keeps us from having to restructure the app.
3354
3355         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
3356         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
3357         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3358
3359 2010-10-29  Ojan Vafai  <ojan@chromium.org>
3360
3361         Reviewed by Tony Chang.
3362
3363         [chromium] add debug logging to help diagnose flakiness dashboard issues
3364         https://bugs.webkit.org/show_bug.cgi?id=48657
3365
3366         The appengine app thinks it's getting empty files uploaded. Add some logging
3367         to see if new-run-webkit-tests agrees.
3368         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3369
3370 2010-10-29  Dimitri Glazkov  <dglazkov@chromium.org>
3371
3372         Reviewed by Ojan Vafai.
3373
3374         [Chromium/DRT] Add master-name flag to new-run-webkit-tests.
3375         https://bugs.webkit.org/show_bug.cgi?id=48649
3376
3377         * BuildSlaveSupport/build.webkit.org-config/master.cfg: Added flag.
3378
3379 2010-10-29  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
3380
3381         Reviewed by Martin Robinson.
3382
3383         Enable popup window in GtkLauncher
3384         https://bugs.webkit.org/show_bug.cgi?id=48335
3385
3386         The GtkLauncher application does not support opening new window when
3387         a link with "target=_blank" is clicked or similar call to
3388         window.open(). Instead, GtkLauncher does nothing which breaks
3389         navigation of some websites.
3390
3391         * GtkLauncher/main.c:
3392         (activate_uri_entry_cb):
3393         (update_title):
3394         (link_hover_cb):
3395         (notify_title_cb):
3396         (notify_load_status_cb):
3397         (notify_progress_cb):
3398         (destroy_cb):
3399         (go_back_cb):
3400         (go_forward_cb):
3401         (create_web_view_cb):
3402         (web_view_ready_cb):
3403         (close_web_view_cb):
3404         (create_browser):
3405         (create_statusbar):
3406         (create_toolbar):
3407         (create_window):
3408         (main):
3409
3410 2010-10-29  Adam Roben  <aroben@apple.com>
3411
3412         Teach check-webkit-style about WebKit2's idiosyncracies
3413
3414         Fixes <http://webkit.org/b/48638> Style bot complains about a number
3415         of WebKit2 conventions
3416
3417         Reviewed by Anders Carlsson.
3418
3419         * Scripts/webkitpy/style/checker.py: Excluded some rules for various
3420         WebKit2-related files. Also updated the excluded rules for
3421         WebKitAPITest to match the current code.
3422
3423 2010-10-29  Csaba Osztrogonác  <ossy@webkit.org>
3424
3425         Reviewed by Adam Roben and David Kilzer.
3426
3427         Fix and cleanup of build systems
3428         https://bugs.webkit.org/show_bug.cgi?id=48342
3429
3430         * Scripts/build-webkit:
3431          - Remove unnecessary ENABLE_SANDBOX option.
3432          - Add ENABLE_FULLSCREEN_API option.
3433
3434 2010-10-28  Adam Roben  <aroben@apple.com>
3435
3436         Switch the Windows WebKit2 bot to the Release configuration
3437
3438         We only have one machine testing WebKit2 on Windows right now, and
3439         Debug is just too slow for it to keep up.
3440
3441         Fixes (hopefully!) <http://webkit.org/b/48615> Windows WebKit2 bot is
3442         always way behind
3443
3444         Reviewed by Jon Honeycutt.
3445
3446         * BuildSlaveSupport/build.webkit.org-config/config.json:
3447
3448 2010-10-29  Leandro Gracia Gil  <leandrogracia@google.com>
3449
3450         Reviewed by Jeremy Orlow.
3451
3452         Added a second parameter to setMockSpeechInputResult for
3453         the language used in speech input.
3454         https://bugs.webkit.org/show_bug.cgi?id=47089
3455
3456         * DumpRenderTree/LayoutTestController.cpp:
3457         (setMockSpeechInputResultCallback):
3458         * DumpRenderTree/LayoutTestController.h:
3459         * DumpRenderTree/chromium/LayoutTestController.cpp:
3460         (LayoutTestController::setMockSpeechInputResult):
3461         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3462         (LayoutTestController::setMockSpeechInputResult):
3463         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3464         (LayoutTestController::setMockSpeechInputResult):
3465         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3466         (LayoutTestController::setMockSpeechInputResult):
3467         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3468         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3469         (LayoutTestController::setMockSpeechInputResult):
3470         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3471         (LayoutTestController::setMockSpeechInputResult):
3472         (LayoutTestController::pageSizeAndMarginsInPixels):
3473
3474 2010-10-28  Ojan Vafai  <ojan@chromium.org>
3475
3476         Reviewed by Tony Chang.
3477
3478         [chromium] add a result-small.json file for the test dashboard
3479         https://bugs.webkit.org/show_bug.cgi?id=48547
3480
3481         Output both a results.json file and a results-small.json file.
3482         The dashboard will load results-small.json by default so it loads faster.
3483
3484         * TestResultServer/model/jsonresults.py:
3485         * TestResultServer/model/jsonresults_unittest.py:
3486         Added a bunch of sys.path hackery. Unforunately, this uses hardcoded
3487         paths. That obviously needs to be fixed, but at least this way it
3488         clearly documents what paths are necessary.
3489
3490 2010-10-28  Eric Seidel  <eric@webkit.org>
3491
3492         Reviewed by Adam Barth.
3493
3494         webkit-patch upload calls changed_files more often than it should
3495         https://bugs.webkit.org/show_bug.cgi?id=48567
3496
3497         Passing changed_files around everywhere isn't a very elegant solution
3498         but it's the one we have for the moment.  I think keeping an explicit
3499         cache on Checkout (or making StepState() a real class) is a better
3500         long-term option.
3501
3502         Previously bug_id_for_this_commit was calling changed_files and the
3503         result was never getting cached on the state.  Now we're explicitly
3504         caching the result on the state and passing that to the bug_id_for_this_commit call.
3505
3506         I looked into building unit tests for this.  Doing so would require
3507         using a real Checkout object with a MockSCM and overriding the appropriate
3508         calls on SCM to count how often we're stating the file system.
3509         That's a useful set of tests to build for a separate change.
3510
3511         * Scripts/webkitpy/common/checkout/api.py:
3512         * Scripts/webkitpy/tool/commands/download.py:
3513         * Scripts/webkitpy/tool/commands/upload.py:
3514         * Scripts/webkitpy/tool/mocktool.py:
3515
3516 2010-10-28  Eric Seidel  <eric@webkit.org>
3517
3518         Reviewed by Adam Barth.
3519
3520         Make suggest-reviewers slightly faster
3521         https://bugs.webkit.org/show_bug.cgi?id=48562
3522
3523         Add @memoized to one more common call.
3524
3525         * Scripts/webkitpy/common/checkout/api.py:
3526
3527 2010-10-28  Eric Seidel  <eric@webkit.org>
3528
3529         Reviewed by Adam Barth.
3530
3531         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is still very slow
3532         https://bugs.webkit.org/show_bug.cgi?id=48536
3533
3534         This patch makes it about 40 seconds faster, but it still
3535         takes 1:40.  This will require more refinement.
3536
3537         The suggested reviewers list appears to be the same.
3538
3539         I think the next step may be to have it stop the search after
3540         5 reviewers are found.  We never want to suggest 30 people.
3541
3542         * Scripts/webkitpy/common/checkout/scm.py:
3543          - Using --remove-empty to theoretically stop lookups past
3544            when a file is removed.  I'm not entirely clear that the option
3545            does what it says it does.  Example:
3546            git log --pretty=format:%H -5 --remove-empty -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
3547            returns only one commit
3548            vs.
3549            git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
3550            which returns 5.  I was not aware that wx files were ever removed from the repository?
3551
3552 2010-10-28  Kinuko Yasuda  <kinuko@chromium.org>
3553
3554         Reviewed by David Levin.
3555
3556         [Chromium] Support FileSystem in chromium DRT
3557         https://bugs.webkit.org/show_bug.cgi?id=47643
3558
3559         * DumpRenderTree/chromium/WebViewHost.cpp:
3560         (WebViewHost::openFileSystem): Added.
3561         * DumpRenderTree/chromium/WebViewHost.h:
3562         (WebViewHost::openFileSystem): Added.
3563
3564 2010-10-28  Sam Weinig  <sam@webkit.org>
3565
3566         Reviewed by Anders Carlsson.
3567
3568         WKURLRefs should be allowed to be null
3569         <rdar://problem/8575621>
3570         https://bugs.webkit.org/show_bug.cgi?id=48535
3571
3572         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
3573         * TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp:
3574         (TestWebKitAPI::didStartProvisionalLoadForFrame):
3575         (TestWebKitAPI::didCommitLoadForFrame):
3576         (TestWebKitAPI::didFinishLoadForFrame):
3577         Test that URLs are null pointers when unset.
3578
3579 2010-10-28  Chris Fleizach  <cfleizach@apple.com>
3580
3581         Reviewed by Adele Peterson.
3582
3583         AX: multi select group option does not handle setting of AXSelectedChildren correctly
3584         https://bugs.webkit.org/show_bug.cgi?id=48464
3585
3586         Add support for querying information about selected children to DRT, including:
3587            selectedChildAtIndex
3588            selectedChildrenCount
3589            setSelectedChild
3590
3591         * DumpRenderTree/AccessibilityUIElement.cpp:
3592         (selectedChildAtIndexCallback):
3593         (setSelectedChildCallback):
3594         (selectedChildrenCountCallback):
3595         (AccessibilityUIElement::setSelectedChild):
3596         (AccessibilityUIElement::selectedChildrenCount):
3597         (AccessibilityUIElement::selectedChildAtIndex):
3598         (AccessibilityUIElement::getJSClass):
3599         * DumpRenderTree/AccessibilityUIElement.h:
3600         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
3601         (AccessibilityUIElement::selectedChildAtIndex):
3602         (AccessibilityUIElement::selectedChildrenCount):
3603         (AccessibilityUIElement::setSelectedChild):
3604
3605 2010-10-27  Sam Weinig  <sam@webkit.org>
3606
3607         Reviewed by Anders Carlsson.
3608
3609         Add WebKit2 API for window feature getter/setters
3610         <rdar://problem/8590373>
3611         https://bugs.webkit.org/show_bug.cgi?id=48496
3612
3613         * MiniBrowser/mac/BrowserWindowController.m:
3614         (-[BrowserWindowController awakeFromNib]):
3615         * MiniBrowser/win/BrowserView.cpp:
3616         (BrowserView::create):
3617         * WebKitTestRunner/TestController.cpp:
3618         (WTR::createOtherPage):
3619         (WTR::TestController::initialize):
3620
3621 2010-10-28  Søren Gjesse  <sgjesse@chromium.org>
3622
3623         Reviewed by Tony Chang.
3624
3625         Added support for the DumpRenderTree flags --multiple-loads and --js-flags to the Python test runner.
3626         https://bugs.webkit.org/show_bug.cgi?id=48236
3627
3628         * Scripts/webkitpy/layout_tests/port/chromium.py:
3629         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3630
3631 2010-10-28  Sergio Villar Senin  <svillar@igalia.com>
3632
3633         Reviewed by Martin Robinson.
3634
3635         [GTK] WebKitWebFrame's load-status is not properly notified to the tests
3636         https://bugs.webkit.org/show_bug.cgi?id=48048
3637
3638         DRT now listens to WebKitWebFrame load-status signals for each
3639         created frame instead of the load-status signal of the
3640         WebKitWebView that only notifies about changes in the main frame.
3641
3642         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3643         (webViewLoadFinished):
3644         (webFrameLoadStatusNotified):
3645         (frameCreatedCallback):
3646         (createWebView):
3647         (main):
3648
3649 2010-10-28  Tony Chang  <tony@chromium.org>
3650
3651         Reviewed by Kent Tamura.
3652
3653         [chromium] fix textInputController.{selectedRange,markedRange}
3654         https://bugs.webkit.org/show_bug.cgi?id=48487
3655
3656         * DumpRenderTree/chromium/TextInputController.cpp:
3657         (TextInputController::markedRange): Return arrays of ints, rather than a string
3658         (TextInputController::selectedRange): Ditto.