Tools/ChangeLog

   1 2011-03-09  Adam Roben  <aroben@apple.com>
   2
   3         Add a link to the Leaks Viewer tool on the build.webkit.org homepage
   4
   5         Rubber-stamped by David Kilzer.
   6
   7         * BuildSlaveSupport/build.webkit.org-config/templates/root.html:
   8
   9 2011-03-09  Adam Roben  <aroben@apple.com>
  10
  11         Add a Leaks Viewer tool to build.webkit.org
  12
  13         This tool parses leaks from the files generated by the leaks bot and displays them in a
  14         view similar to the Web Inspector's Profiles pane (in fact, it uses the Web Inspector's
  15         code).
  16
  17         The URL of the leaks file can be entered in a prompt when the tools is first shown, or can
  18         be passed to the tool via the "url" query parameter.
  19
  20         Only one leaks file can be viewed at a time for now. Only the number of leaks, as opposed to
  21         the number of leaked bytes, is shown. And this tool has made apparent some bugs in the Web
  22         Inspector's Heavy profile view. But it's still pretty useful as-is.
  23
  24         Fixes <http://webkit.org/b/56028> Would like a Shark-like tool for viewing leaks from leak bots
  25
  26         Reviewed by Darin Adler.
  27
  28         * BuildSlaveSupport/build.webkit.org-config/public_html/LeaksViewer/LeaksViewer.css: Added.
  29         (#url-prompt-container): This is the dark overlay that is shown behind the URL prompt.
  30         (#url-prompt): This is the prompt itself.
  31
  32         (#loading-indicator):
  33         (#loading-indicator > img):
  34         These show a spinner in the bottom-right when we're loading a leaks file.
  35
  36         (.percent-time-status-bar-item): Hide the Percent button since we never want to show leaks
  37         as percentages.
  38
  39         * BuildSlaveSupport/build.webkit.org-config/public_html/LeaksViewer/LeaksViewer.js: Added.
  40         (LeaksViewer.loaded): Initialize ourselves and start loading the leaks file (or prompt for a
  41         URL if one wasn't given to us as a query parameter).
  42
  43         (LeaksViewer.get loading):
  44         (LeaksViewer.set loading):
  45         (LeaksViewer.get url):
  46         (LeaksViewer.set url):
  47         Pretty simple accessors.
  48
  49         (LeaksViewer.urlPromptButtonClicked): Start loading the URL from the prompt and hide the
  50         prompt.
  51         (LeaksViewer._displayURLPrompt): Show the prompt and focus the URL input.
  52         (LeaksViewer._loadLeaksFromURL): Start fetching the leaks file. When we've retrieved the
  53         file, hand it off to a Web Worker that will parse it into a fake "profile", then hand off
  54         the profile to the Web Inspector code once it's been created.
  55         (LeaksViewer._loadingStatusChanged): Update our loading indicator and title.
  56         (LeaksViewer._setLoadingIndicatorHidden): Simply hide or show the indicator.
  57         (LeaksViewer._updateTitle): Set our title based on our URL and loading status.
  58         (getResource): Helper function to fetch a resource via XHR and pass it to a callback.
  59
  60         * BuildSlaveSupport/build.webkit.org-config/public_html/LeaksViewer/WebInspectorShims.js: Added.
  61         (WebInspector.UIString): Just a simple implementation since this tool isn't localized.
  62         (ProfilerAgent.getProfile): Store the callback.
  63         (ProfilerAgent.profileReady): Call the callback.
  64         (monkeyPatchInspectorObjects): Fix up WebInspector.ProfileDataGridNode to show the self leak
  65         count and total leak counts as real values, not percentages or milliseconds.
  66
  67         * BuildSlaveSupport/build.webkit.org-config/public_html/LeaksViewer/Worker.js: Added.
  68         (onmessage): Create a profile based on the leaks file and pass it back to our caller.
  69         (parseLeaks): Extract call stacks out of a leaks file.
  70         (createNode): Helper function to create a new empty profile node.
  71         (createProfile): Take the leak stacks and turn them into a fake "profile" for consumption by
  72         WebInspector.ProfileView.
  73
  74         * BuildSlaveSupport/build.webkit.org-config/public_html/LeaksViewer/index.html: Added.
  75
  76 2011-03-09  Dan Bernstein  <mitz@apple.com>
  77
  78         Reviewed by Adam Roben.
  79
  80         Combine multiple --ignore-tests values
  81         https://bugs.webkit.org/show_bug.cgi?id=56016
  82
  83         * Scripts/old-run-webkit-tests:
  84
  85 2011-03-08  Kenji Imasaki  <imasaki@chromium.org>
  86
  87         Reviewed by David Levin.
  88
  89         Adding myself to the committer file.
  90
  91         * Scripts/webkitpy/common/config/committers.py:
  92
  93 2011-03-08  Jessie Berlin  <jberlin@apple.com>
  94
  95         Rubber-stamped by Adam Roben.
  96
  97         Fix two leaks I introduced into the API tests.
  98
  99         * TestWebKitAPI/Tests/WebKit2/win/DoNotCopyANullCFURLResponse.cpp:
 100         (TestWebKitAPI::TEST):
 101
 102 2011-03-08  Jessie Berlin  <jberlin@apple.com>
 103
 104         Reviewed by Anders Carlsson.
 105
 106         Crash in CFNetwork visiting google.com
 107         https://bugs.webkit.org/show_bug.cgi?id=55958
 108
 109         Add a test for not trying to copy a null CFURLResponse.
 110
 111         * TestWebKitAPI/Tests/WebKit2/win/DoNotCopyANullCFURLResponse.cpp: Added.
 112         (TestWebKitAPI::TEST):
 113         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
 114
 115 2011-03-08  Hayato Ito  <hayato@chromium.org>
 116
 117         Reviewed by Ojan Vafai.
 118
 119         [NRWT] Renamed a FailureWithType class to a ComparisonTestFailure class.
 120
 121         FailureWithType is a confusing naming because we have removed test_type/* classes in r79840.
 122
 123         https://bugs.webkit.org/show_bug.cgi?id=55450
 124
 125         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
 126
 127 2011-03-08  Carlos Garcia Campos  <cgarcia@igalia.com>
 128
 129         Reviewed by Martin Robinson.
 130
 131         [GTK] Add layoutTestController.setValueForUser() after r80412
 132         https://bugs.webkit.org/show_bug.cgi?id=55862
 133
 134         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
 135         (LayoutTestController::setValueForUser): Implement setValueForUser()
 136
 137 2011-03-08  Patrick Gansterer  <paroga@webkit.org>
 138
 139         Reviewed by Adam Barth.
 140
 141         Add script_shell_command to WebKitPort
 142         https://bugs.webkit.org/show_bug.cgi?id=55925
 143
 144         Add a central place where we generate the shell command from the script name.
 145         This function will be used later to detect the correct script interpreter on Win32.
 146
 147         * Scripts/webkitpy/common/config/ports.py:
 148         * Scripts/webkitpy/layout_tests/port/base.py:
 149         * Scripts/webkitpy/tool/steps/abstractstep.py:
 150         * Scripts/webkitpy/tool/steps/preparechangelog.py:
 151
 152 2011-03-07  Ojan Vafai  <ojan@chromium.org>
 153
 154         Reviewed by Adam Barth.
 155
 156         add webkit-patch attach-to-bug for posting to bugzilla from the commandline
 157         https://bugs.webkit.org/show_bug.cgi?id=55749
 158
 159         * Scripts/webkitpy/tool/commands/upload.py:
 160         * Scripts/webkitpy/tool/commands/upload_unittest.py:
 161         * Scripts/webkitpy/tool/steps/__init__.py:
 162         * Scripts/webkitpy/tool/steps/attachtobug.py: Added.
 163         * Scripts/webkitpy/tool/steps/options.py:
 164
 165 2011-03-07  Daniel Cheng  <dcheng@chromium.org>
 166
 167         Reviewed by David Levin.
 168
 169         Add support for DataTransferItems
 170         https://bugs.webkit.org/show_bug.cgi?id=55115
 171
 172         Enable data transfer items in DRT.
 173
 174         * DumpRenderTree/chromium/TestShell.cpp:
 175         (TestShell::TestShell):
 176
 177 2011-03-06  Ojan Vafai  <ojan@chromium.org>
 178
 179         Reviewed by Tony Chang.
 180
 181         [test dashboard] pull treemap files
 182         https://bugs.webkit.org/show_bug.cgi?id=55852
 183
 184         * TestResultServer/handlers/dashboardhandler.py:
 185
 186 2011-03-03  Hayato Ito  <hayato@chromium.org>
 187
 188         Reviewed by Mihai Parparita.
 189
 190         [NRWT] Clear output image file used by DumpRenderTree to make sure the
 191         previous image should not be used in the ChromiumDriver.
 192
 193         Also make sure that output image from the ChromiumDriver should be ''
 194         (empty string) if a test crashes for consistency with the WebKitDriver
 195         implementation.
 196
 197         https://bugs.webkit.org/show_bug.cgi?id=55746
 198
 199         * Scripts/webkitpy/layout_tests/port/chromium.py:
 200
 201 2011-03-07  Zan Dobersek  <zandobersek@gmail.com>
 202
 203         Reviewed by Martin Robinson.
 204
 205         [Gtk] fast/js/navigator-language.html fails locally
 206         https://bugs.webkit.org/show_bug.cgi?id=55880
 207
 208         Add LANG variable to the clean environment when running
 209         old-run-webkit-tests script.
 210
 211         * Scripts/old-run-webkit-tests:
 212
 213 2011-03-07  Tony Chang  <tony@chromium.org>
 214
 215         Reviewed by Mihai Parparita.
 216
 217         [Chromium] Find .checksum files without .pngs and vice-versa
 218         https://bugs.webkit.org/show_bug.cgi?id=55236
 219
 220         * Scripts/find-mismatched-layout-test-results: Added. A script for finding
 221             mismatched results.
 222
 223 2011-03-07  Chris Fleizach  <cfleizach@apple.com>
 224
 225         Reviewed by Beth Dakin.
 226
 227         AX: WK1 needs to use the ScrollView attachment for AXScrollArea, WK2 should not use this element.
 228         https://bugs.webkit.org/show_bug.cgi?id=55706
 229
 230         The rootElement in DRT should point to the scroll view. In WK1, that will be the parent of the rootObject.
 231
 232         * DumpRenderTree/mac/AccessibilityControllerMac.mm:
 233         (AccessibilityController::focusedElement):
 234            Remove FIXME comment about caching the focusedElement. In fact, we shouldn't cache this
 235            because it could change at any time.
 236         (AccessibilityController::rootElement):
 237
 238 2011-03-07  Sam Weinig  <sam@webkit.org>
 239
 240         Reviewed by Anders Carlsson.
 241
 242         Add unableToImplementPolicy callback for WebKit2 policy client
 243         <rdar://problem/9071902>
 244         https://bugs.webkit.org/show_bug.cgi?id=55884
 245
 246         * MiniBrowser/mac/BrowserWindowController.m:
 247         (-[BrowserWindowController awakeFromNib]):
 248         Update policy client initialization struct for new member.
 249
 250 2011-03-07  Sam Weinig  <sam@webkit.org>
 251
 252         Reviewed by Anders Carlsson.
 253
 254         Replace WebKit2's decidePolicyForMIMEType with decidePolicyForResponse
 255         https://bugs.webkit.org/show_bug.cgi?id=55827
 256
 257         * MiniBrowser/mac/BrowserWindowController.m:
 258         (decidePolicyForResponse):
 259         (-[BrowserWindowController awakeFromNib]):
 260         * TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp:
 261         (TestWebKitAPI::decidePolicyForResponse):
 262         (TestWebKitAPI::TEST):
 263         Update tools for new name.
 264
 265 2011-03-06  Daniel Bates  <dbates@rim.com>
 266
 267         Reviewed by Eric Seidel and David Kilzer.
 268
 269         Fix misspelled word in build-webkit's checkForJavaSDK() and sort forward declarations
 270         https://bugs.webkit.org/show_bug.cgi?id=55503
 271
 272         Fix misspelled word "Dowloads" [sic] in the message printed to standard output
 273         in checkForJavaSDK().
 274
 275         Also, remove extraneous space in function prototype for unlinkZeroFiles.
 276
 277         * Scripts/build-webkit:
 278
 279 2011-03-05  Brian Weinstein  <bweinstein@apple.com>
 280
 281         Reviewed by Sam Weinig.
 282
 283         Fix Web Process crashes caused by https://bugs.webkit.org/show_bug.cgi?id=53919.
 284
 285         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
 286         (WTR::InjectedBundlePage::willSendRequestForFrame): Retain the WKURLRequestRef returned
 287             by willSendRequestForFrame. The API expects a retained reference to the URL request.
 288
 289 2011-03-05  Ilya Sherman  <isherman@chromium.org>
 290
 291         Reviewed by Darin Adler.
 292
 293         HTMLInputElement::setValue() should schedule change event when the element is focused
 294         In service of https://code.google.com/p/chromium/issues/detail?id=42716
 295         https://bugs.webkit.org/show_bug.cgi?id=53160
 296
 297         * DumpRenderTree/LayoutTestController.cpp:
 298         (setValueForUserCallback):
 299         (LayoutTestController::staticFunctions):
 300         * DumpRenderTree/LayoutTestController.h:
 301         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
 302         (LayoutTestController::setValueForUser): Stubbed out; need real implementation.
 303         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
 304         (LayoutTestController::setValueForUser): Expose HTMLInputElement::setValueForUser() to tests
 305         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
 306         (LayoutTestController::setValueForUser): Stubbed out; need real implementation.
 307         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
 308         (LayoutTestController::setValueForUser): Stubbed out; need real implementation.
 309         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
 310         (LayoutTestController::setValueForUser): Stubbed out; need real implementation.
 311
 312 2011-03-04  Mark Rowe  <mrowe@apple.com>
 313
 314         Reviewed by Alexey Proskuryakov.
 315
 316         <rdar://problem/9055252> DumpRenderTree is crashing with GuardMalloc
 317
 318         GuardMalloc does not expect the memory containing its malloc_zone_t to ever be read-only.
 319
 320         * DumpRenderTree/mac/CheckedMalloc.cpp:
 321         (protectionOfRegion): Retrieve the protection flags for the VM region containing the given address.
 322         (makeLargeMallocFailSilently): Save the initial protection flags, temporarily make the region writeable,
 323         and then restore the initial protection flags when we're done.
 324
 325 2011-03-04  Evan Martin  <evan@chromium.org>
 326
 327         Reviewed by Tony Chang.
 328
 329         [gdb] pretty-print KURL in gdb
 330         https://bugs.webkit.org/show_bug.cgi?id=55788
 331
 332         * gdb/webkit.py: extend the pretty-printers to print WTF::CString
 333         and KURLGooglePrivate.  As always, "p/r" can be used to sidestep
 334         pretty-printing.
 335
 336 2011-03-03  Timothy Hatcher  <timothy@apple.com>
 337
 338         Update the tests that use WKPageRunJavaScriptInMainFrame use the WKSerializedScriptValueRef.
 339
 340         Reviewed by Darin Adler.
 341
 342         * TestWebKitAPI/JavaScriptTest.cpp:
 343         (TestWebKitAPI::javaScriptCallback):
 344         * TestWebKitAPI/Tests/WebKit2/EvaluateJavaScript.cpp:
 345         (TestWebKitAPI::didRunJavaScript):
 346         * TestWebKitAPI/Tests/WebKit2/PageLoadDidChangeLocationWithinPageForFrame.cpp:
 347         (TestWebKitAPI::nullJavaScriptCallback):
 348         * TestWebKitAPI/Tests/WebKit2/PreventEmptyUserAgent.cpp:
 349         (TestWebKitAPI::didRunJavaScript):
 350
 351 2011-03-03  Timothy Hatcher  <timothy@apple.com>
 352
 353         Make run-api-tests show failed if a test fails.
 354
 355         https://webkit.org/b/55725
 356
 357         Reviewed by Sam Weinig.
 358
 359         * Scripts/run-api-tests:
 360         (runTest): Remove "my" from the nested $result assignment so the
 361         check outside the condition sees the real result.
 362
 363 2011-03-02  MORITA Hajime  <morrita@google.com>
 364
 365         Reviewed by Kent Tamura.
 366
 367         [Chromium][DRT] WebTask should use webkit_support::TaskAdaptor
 368         https://bugs.webkit.org/show_bug.cgi?id=55434
 369
 370         Changed WebTask to be a subclass of webkit_support::TaskAdaptor
 371         and passed the object webkit_support::PostDelayedTask directly.
 372
 373         Note that the change keeps postTask() as is because it has a
 374         semantics taht webkit_support::PostDelayedTask() has; It jumps to
 375         the main thread.
 376
 377         * DumpRenderTree/chromium/Task.cpp:
 378         (postDelayedTask):
 379         * DumpRenderTree/chromium/Task.h:
 380         (WebTask::Run):
 381
 382 2011-03-03  Mahesh Kulkarni  <mahesh.kulkarni@nokia.com>
 383
 384         Reviewed by Kenneth Rohde Christiansen.
 385
 386         [QT] Implement mock client-based geolocation for layout testing
 387         https://bugs.webkit.org/show_bug.cgi?id=54334
 388
 389         Implement testing for Client-Based geolocation for QtWebkit.
 390
 391         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
 392         (WebCore::DumpRenderTree::DumpRenderTree):
 393         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
 394         (LayoutTestController::reset):
 395         (LayoutTestController::setGeolocationPermission):
 396         (LayoutTestController::setMockGeolocationError):
 397         (LayoutTestController::setMockGeolocationPosition):
 398
 399 2011-03-03  Sheriff Bot  <webkit.review.bot@gmail.com>
 400
 401         Unreviewed, rolling out r80237.
 402         http://trac.webkit.org/changeset/80237
 403         https://bugs.webkit.org/show_bug.cgi?id=55714
 404
 405         Windows bots aren't ready to be core (Requested by aroben on
 406         #webkit).
 407
 408         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
 409         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
 410
 411 2011-03-03  Mihai Parparita  <mihaip@chromium.org>
 412
 413         Reviewed by Tony Chang.
 414
 415         NRWT: AttributeError: TestRunner2 instance has no attribute '_cancel_workers'
 416         https://bugs.webkit.org/show_bug.cgi?id=55694
 417
 418         Fixes for interruptions in NRWT:
 419         - Make TestRunInterruptedException be pickleable correctly (the base
 420           Exception class defines a __reduce__ that does not include the reason)
 421         - Fix ordering of returned arguments from TestRunner2._run_tests
 422           (interrupted and keyboard_interrupted were reversed)
 423         - Fix cancel_workers callsites (was using old name).
 424         - In handle_exception re-raise actual exception instance that was thrown
 425
 426         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
 427         * Scripts/webkitpy/layout_tests/layout_package/test_runner2.py:
 428         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
 429         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
 430
 431 2011-03-03  Qi Zhang  <qi.2.zhang@nokia.com>
 432
 433         Unreviewed.
 434
 435         Adding myself as a committer.
 436
 437         * Scripts/webkitpy/common/config/committers.py:
 438
 439 2011-03-03  Tony Gentilcore  <tonyg@chromium.org>
 440
 441         Reviewed by Mihai Parparita.
 442
 443         Teach sheriffbot to answer whois commands
 444         https://bugs.webkit.org/show_bug.cgi?id=55687
 445
 446         * Scripts/webkitpy/tool/bot/irc_command.py:
 447         * Scripts/webkitpy/tool/bot/irc_command_unittest.py:
 448         * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:
 449
 450 2011-03-03  Victoria Kirst  <vrk@google.com>
 451
 452         Reviewed by Eric Carlson.
 453
 454         Enable media tests in chromium-gpu/test_expectations
 455         https://bugs.webkit.org/show_bug.cgi?id=53608
 456
 457         Add media folder to the list of GPU test paths.
 458
 459         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
 460
 461 2011-03-03  Adam Roben  <aroben@apple.com>
 462
 463         Work around a bug in KURL's parsing of Windows-style absolute file: URLs
 464
 465         Fixes <http://webkit.org/b/55674> Many tests fail in WebKit2 mode on Windows due to
 466         incorrect parsing of absolute Windows-style file: URLs
 467
 468         Reviewed by Tim Hatcher.
 469
 470         * WebKitTestRunner/TestInvocation.cpp:
 471         (WTR::createWKURL): Add a "localhost" host to file: URLs we create to work around
 472         <http://webkit.org/b/55683>.
 473
 474 2011-03-03  Tony Chang  <tony@chromium.org>
 475
 476         Reviewed by Dimitri Glazkov.
 477
 478         [chromium] Run kill-old-processes on the chromium-win bot too
 479         https://bugs.webkit.org/show_bug.cgi?id=55630
 480
 481         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
 482         * BuildSlaveSupport/win/kill-old-processes: Add wdiff.exe to the kill list.
 483
 484 2011-03-03  Carlos Garcia Campos  <cgarcia@igalia.com>
 485
 486         Reviewed by Martin Robinson.
 487
 488         [GTK] DRT needs implementation of EventSender.scheduleAsynchronousClick
 489         https://bugs.webkit.org/show_bug.cgi?id=53960
 490
 491         * DumpRenderTree/gtk/EventSender.cpp:
 492         (sendClick):
 493         (scheduleAsynchronousClickCallback):
 494
 495 2011-03-03  Adam Roben  <aroben@apple.com>
 496
 497         Look for Windows-style paths when checking whether a test is an Inspector test
 498
 499         WTR part of <http://webkit.org/b/55672> <rdar://problem/9080867> All inspector tests time
 500         out or crash on Windows 7 Release (WebKit2 Tests)
 501
 502         Reviewed by Darin Adler.
 503
 504         * WebKitTestRunner/TestInvocation.cpp:
 505         (WTR::shouldOpenWebInspector): Also look for backslashes, since that's what we'll get on
 506         Windows.
 507
 508 2011-03-03  Adam Roben  <aroben@apple.com>
 509
 510         Make Windows 7 Release (Tests) a core builder
 511
 512         We can probably make Windows XP Debug (Tests) a core builder, too, but we should probably
 513         give it a few more days of greenness first.
 514
 515         Fixes <http://webkit.org/b/55665> Windows 7 Release (Tests) should be a core builder
 516
 517         Reviewed by Andreas Kling.
 518
 519         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
 520         (BuildBot): Added a regexp to match Windows 7 Release (Tests).
 521
 522         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
 523         (BuildBotTest): Updated the list of builders to match the current list on build.webkit.org.
 524         Updated the core regexps to match the buildbot module. Updated the expected results to match
 525         the current set of core builders.
 526
 527 2011-03-03  Benjamin Poulain  <benjamin.poulain@nokia.com>
 528
 529         Reviewed by Andreas Kling.
 530
 531         Adding myself as a reviewer.
 532
 533         * Scripts/webkitpy/common/config/committers.py:
 534
 535 2011-03-03  Lukasz Slachciak  <l.slachciak@samsung.com>
 536
 537         Reviewed by Xan Lopez.
 538
 539         [GTK] Extended application cache database API and added unit tests file.
 540         https://bugs.webkit.org/show_bug.cgi?id=55335
 541
 542         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: removed external declaration
 543         of webkit_application_cache_set_maximum_size due to API exposal.
 544
 545 2011-03-02  Ojan Vafai  <ojan@chromium.org>
 546
 547         Reviewed by Adam Barth.
 548
 549         make webkit-patch upload work for security bugs
 550         https://bugs.webkit.org/show_bug.cgi?id=55655
 551
 552         The problem is that we shell out to prepare-ChangeLog,
 553         which does not authenticate with bugzilla. All it needs from
 554         bugzilla is the bug title though. We just pass it through from
 555         webkit-patch instead.
 556
 557         * Scripts/prepare-ChangeLog:
 558         * Scripts/webkitpy/common/net/bugzilla/bug.py:
 559         * Scripts/webkitpy/tool/steps/preparechangelog.py:
 560
 561 2011-03-01  Ojan Vafai  <ojan@chromium.org>
 562
 563         Reviewed by Tony Chang.
 564
 565         [test-dashboard] make gtest JSON match layout-tests JSON
 566         https://bugs.webkit.org/show_bug.cgi?id=55556
 567
 568         http://trac.webkit.org/changeset/80090 stopped converting time_ms to t.
 569         Make this code match.
 570
 571         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
 572
 573 2011-03-02  Dirk Pranke  <dpranke@chromium.org>
 574
 575         Reviewed by Tony Chang.
 576
 577         NRWT - implement Linux Hardy 64-bit port support. This involves
 578         adding support for non-'x86' architectures and updating the
 579         "deduplicate_tests" script to correctly understand that
 580         port.name() is not always equal to basename(port.baseline_path()).
 581
 582         This change introduces two new port names: 'chromium-linux-x86'
 583         and 'chromium-linux-x86_64'. Using the prior 'chromium-linux'
 584         flag will cause the code to determine which one to use at
 585         runtime, just like we currently do with 'chromium-win'.
 586
 587         https://bugs.webkit.org/show_bug.cgi?id=55535
 588
 589         * Scripts/webkitpy/common/system/executive_mock.py:
 590           Handle 'return_stderr' keyword arg to run_command()
 591         * Scripts/webkitpy/layout_tests/deduplicate_tests.py:
 592         * Scripts/webkitpy/layout_tests/deduplicate_tests_unittest.py:
 593         * Scripts/webkitpy/layout_tests/port/base.py:
 594         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
 595         * Scripts/webkitpy/layout_tests/port/chromium_linux_unittest.py:
 596         * Scripts/webkitpy/layout_tests/port/factory.py:
 597         * Scripts/webkitpy/layout_tests/port/google_chrome.py:
 598
 599 2011-03-02  Kevin Ollivier  <kevino@theolliviers.com>
 600
 601         [wx] Build fix, make sure we put the wxWebKit dylib in the right directory and
 602         update symlinks accordingly.
 603
 604         * wx/packaging/build-mac-installer.py:
 605
 606 2011-03-02  Sheriff Bot  <webkit.review.bot@gmail.com>
 607
 608         Unreviewed, rolling out r80139.
 609         http://trac.webkit.org/changeset/80139
 610         https://bugs.webkit.org/show_bug.cgi?id=55620
 611
 612         failing and crashing tests on Snow Leopard bot (Requested by
 613         mihaip1 on #webkit).
 614
 615         * DumpRenderTree/LayoutTestController.cpp:
 616         (LayoutTestController::staticFunctions):
 617         * DumpRenderTree/LayoutTestController.h:
 618         * DumpRenderTree/chromium/LayoutTestController.cpp:
 619         (LayoutTestController::LayoutTestController):
 620         * DumpRenderTree/chromium/LayoutTestController.h:
 621         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
 622         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
 623         * DumpRenderTree/mac/UIDelegate.h:
 624         * DumpRenderTree/mac/UIDelegate.mm:
 625         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
 626         * DumpRenderTree/qt/LayoutTestControllerQt.h:
 627         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
 628         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
 629
 630 2011-03-02  Adam Roben  <aroben@apple.com>
 631
 632         Wait for force a paint in WebKitTestRunner until we're actually ready to dump the output
 633
 634         This matches DumpRenderTree on Windows.
 635
 636         Fixes <http://webkit.org/b/55469> <rdar://problem/9068539> REGRESSION (r79863): Lots of
 637         dom/html/level2/html/HTMLFrameElement*.html tests crashing in FrameView::paintOverhangAreas
 638         in WebKit2
 639
 640         Reviewed by Beth Dakin.
 641
 642         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
 643         (WTR::InjectedBundlePage::dump): Moved code to force a paint here...
 644         (WTR::InjectedBundlePage::didFinishLoadForFrame): ...from here.
 645
 646 2011-03-02  Daniel Cheng  <dcheng@chromium.org>
 647
 648         Reviewed by David Levin.
 649
 650         Add feature define for data transfer items
 651         https://bugs.webkit.org/show_bug.cgi?id=55510
 652
 653         * Scripts/build-webkit:
 654
 655 2011-03-02  Adam Roben  <aroben@apple.com>
 656
 657         Look in the new-to-SnowLeopard location for crash logs on SnowLeopard and newer
 658
 659         ReportCrash saves logs to ~/Library/Logs/CrashReporter on Leopard, but to
 660         ~/Library/Logs/DiagnosticReports on SnowLeopard. old-run-webkit-tests was only looking in
 661         the former location, but was getting lucky on some SnowLeopard machines because symlinks
 662         were being created there pointing into DiagnosticReports. For machines without these
 663         symlinks, crash logs were not getting captured.
 664
 665         Fixes <http://webkit.org/b/55607> old-run-webkit-tests doesn't capture crash logs on some
 666         SnowLeopard machines
 667
 668         Reviewed by Eric Seidel.
 669
 670         * Scripts/old-run-webkit-tests:
 671         (captureSavedCrashLog): Look in ~/Library/Logs/DiagnosticReports on SnowLeopard and newer.
 672
 673 2011-03-02  Adam Roben  <aroben@apple.com>
 674
 675         Don't try to capture crash logs on platforms that haven't implemented that feature
 676
 677         Fixes <http://webkit.org/b/55504> Use of uninitialized value in numeric lt (<) at
 678         Tools/Scripts/old-run-webkit-tests line 1778 seen on Qt bots after a crash
 679
 680         Reviewed by Joseph Pecoraro.
 681
 682         * Scripts/old-run-webkit-tests:
 683         (captureSavedCrashLog): Bail if we don't have a crash log glob.
 684         (findNewestFileMatchingGlob): Fix a buggy test for when no paths match the glob. This isn't
 685         strictly required to fix bug 55504, but seems worthwhile.
 686
 687 2011-01-26  John Knottenbelt  <jknotten@chromium.org>
 688
 689         Reviewed by Dmitry Titov.
 690
 691         Detach Geolocation from Frame when Page destroyed.
 692         https://bugs.webkit.org/show_bug.cgi?id=52877
 693
 694         Extend the layout test controller to expose the number of pending
 695         geolocation requests, so that we can test that the requests have
 696         been cancelled on page close.
 697
 698         * DumpRenderTree/LayoutTestController.cpp:
 699         (numberOfPendingGeolocationPermissionRequestsCallback):
 700         (LayoutTestController::staticFunctions):
 701         * DumpRenderTree/LayoutTestController.h:
 702         * DumpRenderTree/chromium/LayoutTestController.cpp:
 703         (LayoutTestController::LayoutTestController):
 704         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
 705         * DumpRenderTree/chromium/LayoutTestController.h:
 706         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
 707         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
 708         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
 709         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
 710         * DumpRenderTree/mac/UIDelegate.h:
 711         * DumpRenderTree/mac/UIDelegate.mm:
 712         (-[UIDelegate numberOfPendingGeolocationPermissionRequests]):
 713         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
 714         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
 715         * DumpRenderTree/qt/LayoutTestControllerQt.h:
 716         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
 717         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
 718         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
 719         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
 720
 721
 722 2011-03-01  Kent Tamura  <tkent@chromium.org>
 723
 724         Reviewed by Dimitri Glazkov.
 725
 726         Assertion fails when a form validation bubble appears
 727         https://bugs.webkit.org/show_bug.cgi?id=55550
 728
 729         Test: fast/forms/interactive-validation-attach-assertion.html
 730
 731         * DumpRenderTree/chromium/WebPreferences.cpp:
 732         (WebPreferences::applyTo): Enable the interactive validation feature by
 733           default. This matches to Chromium browser.
 734
 735 2011-03-01  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
 736
 737         Reviewed by Kent Tamura.
 738
 739         Remove efl port style exception of Webcore/JavaScriptCore
 740         https://bugs.webkit.org/show_bug.cgi?id=55549
 741
 742         WebCore and JavaScriptCore need to adhere WebKit coding style.
 743
 744         * Scripts/webkitpy/style/checker.py:
 745
 746 2011-03-01  Tony Chang  <tony@chromium.org>
 747
 748         Reviewed by Ojan Vafai.
 749
 750         [chromium] clean up temp files on chromium-mac builder too
 751         https://bugs.webkit.org/show_bug.cgi?id=55525
 752
 753         * BuildSlaveSupport/chromium/remove-crash-logs:
 754
 755 2011-02-28  Ojan Vafai  <ojan@chromium.org>
 756
 757         Reviewed by Tony Chang.
 758
 759         wrap json in a function call to afford cross-domain loading
 760         https://bugs.webkit.org/show_bug.cgi?id=55353
 761
 762         Also delete the code that compacts the JSON. It turns out this
 763         breaks the rebaseline tool and is probably a premature optimization anyways.
 764
 765         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
 766         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
 767         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
 768         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
 769
 770 2011-03-01  Sheriff Bot  <webkit.review.bot@gmail.com>
 771
 772         Unreviewed, rolling out r80079.
 773         http://trac.webkit.org/changeset/80079
 774         https://bugs.webkit.org/show_bug.cgi?id=55547
 775
 776         "Broke the Win debug build?" (Requested by dcheng on #webkit).
 777
 778         * Scripts/build-webkit:
 779
 780 2011-03-01  Mark Rowe  <mrowe@apple.com>
 781
 782         Reviewed by Sam Weinig.
 783
 784         Update verification scripts to handle WebKit2.
 785
 786         * Scripts/check-for-global-initializers: Skip object files that use RefCountedLeakChecker.
 787         * Scripts/check-for-inappropriate-files-in-framework: Add some inappropriate file types for
 788         WebKit2.
 789
 790 2011-03-01  Daniel Cheng  <dcheng@chromium.org>
 791
 792         Reviewed by David Levin.
 793
 794         Add feature define for data transfer items
 795         https://bugs.webkit.org/show_bug.cgi?id=55510
 796
 797         * Scripts/build-webkit:
 798
 799 2011-02-23  Joseph Pecoraro  <joepeck@webkit.org>
 800
 801         Reviewed by Timothy Hatcher.
 802
 803         All Console Messages should be passed to ChromeClients.
 804         https://bugs.webkit.org/show_bug.cgi?id=54926
 805
 806         Update DumpRenderTree's UIDelegate to use the new console message delegate.
 807
 808         * DumpRenderTree/mac/UIDelegate.mm:
 809         (-[UIDelegate webView:addMessageToConsole:withSource:]):
 810
 811 2011-03-01  Dirk Pranke  <dpranke@chromium.org>
 812
 813         Reviewed by Adam Barth.
 814
 815         Tragically, we stopped running the python unit tests
 816         on Chromium Win just as I was getting them to be green.
 817         This patch turns them back on.
 818
 819         https://bugs.webkit.org/show_bug.cgi?id=55521
 820
 821         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
 822
 823 2011-03-01  Dirk Pranke  <dpranke@chromium.org>
 824
 825         Reviewed by Adam Roben.
 826
 827         Disable
 828         webkitpy.common.system.executive_unittest.ExecutiveTest.\
 829            test_kill_all fails on win until it can be properly triaged.
 830
 831         https://bugs.webkit.org/show_bug.cgi?id=54790
 832
 833         * Scripts/webkitpy/common/system/executive_unittest.py:
 834
 835     2011-03-01  Helder Correia  <helder@sencha.com>
 836
 837             Unreviewed. Add myself to committers.py.
 838
 839             * Scripts/webkitpy/common/config/committers.py:
 840
 841 2011-03-01  Adam Roben  <aroben@apple.com>
 842
 843         Give up if a crash log for the web process is taking too long to be saved on Windows
 844
 845         This should keep the bots from getting stuck while waiting for a crash log, if writing a
 846         crash log ever hangs or the UI process for some reason doesn't notice it's finished.
 847
 848         Fixes <http://webkit.org/b/55499> WebKitTestRunner can hang forever waiting for a crash log
 849         to be saved for the web process
 850
 851         Reviewed by Steve Falkenburg.
 852
 853         * WebKitTestRunner/win/TestControllerWin.cpp:
 854         (WTR::runRunLoopUntil): Moved code to run the run loop here here from platformRunUntil.
 855         Generalized the code slightly to handle the optional object to wait on.
 856         (WTR::TestController::platformRunUntil): Use the new runRunLoopUntil function, and also use
 857         it (with a timeout) when waiting for a crash log for the web process to be saved. This will
 858         prevent us from waiting forever.
 859
 860 2011-03-01  Adam Roben  <aroben@apple.com>
 861
 862         Give WebKitLauncherWin's PDB file a name different from WebKit's
 863
 864         WebKitLauncherWin (which builds WebKit.exe) and WebKit (which builds WebKit.dll) were
 865         fighting over the same PDB file, meaning that one of them was going to lose and not have
 866         symbols. WebKitLauncherWin's PDB file is now named WebKitLauncherWin.pdb.
 867
 868         Fixes <http://webkit.org/b/55495> Crashes in WebKit.dll have bad backtraces for some test
 869         runs (due to missing symbols)
 870
 871         Reviewed by David Kilzer.
 872
 873         * WebKitLauncherWin/WebKitLauncherWinCommon.vsprops: Use the project name,
 874         "WebKitLauncherWin", instead of the target name, "WebKit", when naming the PDB file.
 875
 876 2011-03-01  Adam Roben  <aroben@apple.com>
 877
 878         Count web process crashes toward the --exit-after-n-crashes-or-timeouts limit
 879
 880         Fixes <http://webkit.org/b/55492> Web process crashes aren't counted toward the
 881         --exit-after-n-crashes-or-timeouts limit
 882
 883         Reviewed by David Kilzer.
 884
 885         * Scripts/old-run-webkit-tests:
 886         (stopRunningTestsEarlyIfNeeded): Include web process crashes in the computation, and print
 887         the number of tests that caused the web process to crash when exiting early.
 888
 889 2011-03-01  Adam Roben  <aroben@apple.com>
 890
 891         Save a crash log when the web process crashes
 892
 893         On Windows, WebKitTestRunner now detects when the web process is crashing and waits to exit
 894         until it has finished crashing, which guarantees that the crash log will have had time to be
 895         saved, too. On Mac, we always wait until ReportCrash has exited before capturing the crash
 896         log, so all we have to do is choose the right crash log out of the CrashReporter directory.
 897
 898         Fixes <http://webkit.org/b/44121> <rdar://problem/8320759> When the web process crashes and
 899         a crash log is being saved, WebKitTestRunner thinks the web process has become unresponsive
 900
 901         Reviewed by Sam Weinig.
 902
 903         * Scripts/old-run-webkit-tests:
 904         (testCrashedOrTimedOut): Don't kill WebKitTestRunner when the web process crashes. It will
 905         kill itself. On Windows, this will cause us to wait until the crash log has been saved. On
 906         Mac, it should have no effect. Capture saved crash logs for web process crashes, too.
 907         (captureSavedCrashLog): Added $webProcessCrashed argument. On Mac, look for
 908         WebProces_*.crash files when the web process crashes.
 909
 910         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
 911         (WTR::InjectedBundle::initialize): Added an initializationUserData argument. Updated for
 912         initializePlatformDefaults -> platformInitialize rename. Pass the initializationUserData
 913         along to platformInitialize.
 914
 915         * WebKitTestRunner/InjectedBundle/InjectedBundle.h: See above.
 916
 917         * WebKitTestRunner/InjectedBundle/InjectedBundleMain.cpp:
 918         (WKBundleInitialize): Pass along the initializationUserData to the InjectedBundle.
 919
 920         * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm:
 921         (WTR::InjectedBundle::platformInitialize):
 922         * WebKitTestRunner/InjectedBundle/qt/InjectedBundleQt.cpp:
 923         (WTR::InjectedBundle::platformInitialize):
 924         Updated function signature.
 925
 926         * WebKitTestRunner/InjectedBundle/win/InjectedBundleWin.cpp:
 927         (WTR::exceptionFilter): Added. Tells the UI process we're crashing by signaling the
 928         webProcessCrashingEvent, then lets the crash continue as normal.
 929
 930         (WTR::InjectedBundle::platformInitialize): Hook up exceptionFilter. Retrieve the name of the
 931         event we should use to tell the UI process we're crashing from the initializationUserData,
 932         and get a handle to that event.
 933
 934         * WebKitTestRunner/TestController.cpp:
 935         (WTR::TestController::TestController): Initialize new members.
 936         (WTR::TestController::processDidCrash): Removed unnecessary WKPageRef argument. Changed to
 937         only print the "#CRASHED - WebProcess" message once, since this can be called more than once
 938         when a crash log is being saved on Windows. Exit right away if specified. (This is the
 939         default.)
 940
 941         * WebKitTestRunner/TestController.h: Added new members.
 942
 943         * WebKitTestRunner/win/TestControllerWin.cpp:
 944         (WTR::TestController::platformInitialize): Set up the event the web process will use to tell
 945         us it's crashing.
 946         (WTR::TestController::platformRunUntil): Pass MWMO_INPUTAVAILABLE to
 947         ::MsgWaitForMultipleObjectsEx so we'll process messages that have already been seen by
 948         ::PeekMessage. (This is unrelated to the bug fix.) Notice when the webProcessCrashingEvent
 949         has been signaled. When this happens, print the "#CRASHED - WebProcess" message right away
 950         so the test harness will know the web process has crashed and not try to kill us, then wait
 951         for the web process to finish crashing so a crash log will have time to be saved.
 952         (WTR::toWK): Simple hepler function.
 953         (WTR::TestController::platformInitializeContext): Pass along the name of the event the web
 954         process should use to tell us it is crashing in the context's initialization user data.
 955
 956 2011-03-01  Dimitri Glazkov  <dglazkov@chromium.org>
 957
 958         Reviewed by Tony Gentilcore.
 959
 960         Move the checks to exclude Python/Perl tests for Chromium Win build to the right place.
 961         https://bugs.webkit.org/show_bug.cgi?id=55476
 962
 963         Chromium Win uses BuildAndTestFactory, not TestFactory.
 964
 965         * BuildSlaveSupport/build.webkit.org-config/master.cfg: Moved to the right place.
 966
 967 2011-03-01  Adam Roben  <aroben@apple.com>
 968
 969         Test that the WebKit2 responsiveness timer doesn't fire too early
 970
 971         Test for <http://webkit.org/b/55417> <rdar://problem/9065287> RunLoop::Timer fires 1000x too
 972         early on Windows
 973
 974         Reviewed by Anders Carlsson.
 975
 976         * TestWebKitAPI/PlatformUtilities.h: Added sleep().
 977
 978         * TestWebKitAPI/Tests/WebKit2/ResponsivenessTimerDoesntFireEarly.cpp: Added.
 979         (TestWebKitAPI::didReceiveMessageFromInjectedBundle):
 980         (TestWebKitAPI::didFinishLoadForFrame):
 981         (TestWebKitAPI::processDidBecomeUnresponsive):
 982         (TestWebKitAPI::setInjectedBundleClient):
 983         (TestWebKitAPI::setPageLoaderClient):
 984         Simple helper functions.
 985
 986         (TestWebKitAPI::TEST): Load an HTML file to make sure the web process is initialized. Then
 987         tell the web process to pause and press the spacebar key. The spacebar keypress should cause
 988         the responsiveness timer to start, but the web process should unpause before it has a chance
 989         to fire. Run until the web process has unpaused, and assert that the timer didn't fire.
 990
 991         * TestWebKitAPI/Tests/WebKit2/ResponsivenessTimerDoesntFireEarly_Bundle.cpp: Added.
 992         (TestWebKitAPI::ResponsivenessTimerDoesntFireEarlyTest::ResponsivenessTimerDoesntFireEarlyTest):
 993         Call up to the base class.
 994         (TestWebKitAPI::ResponsivenessTimerDoesntFireEarlyTest::didReceiveMessage): When asked to
 995         pause, sleep for 0.5 seconds, then send back a message saying we paused.
 996
 997         * TestWebKitAPI/mac/PlatformUtilitiesMac.mm:
 998         (TestWebKitAPI::Util::sleep): Added. Calls through to usleep.
 999         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp:
1000         (TestWebKitAPI::Util::sleep): Added. Calls through to ::Sleep.
1001
1002         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
1003         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
1004         * TestWebKitAPI/win/TestWebKitAPIInjectedBundle.vcproj:
1005         Added new files.
1006
1007 2011-03-01  Andras Becsi  <abecsi@webkit.org>
1008
1009         Reviewed by Csaba Osztrogonác.
1010
1011         [Qt] Clean up the project files and move common options to WebKit.pri.
1012
1013         * WebKitTestRunner/qt/WebKitTestRunner.pro: Deduplicate.
1014
1015
1016 2011-03-01  Adam Roben  <aroben@apple.com>
1017
1018         Skip another multiprocessing test on Windows
1019
1020         It is sometimes failing (and may be leaving child Python processes around).
1021         <http://webkit.org/b/55087> tracks the failure.
1022
1023         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
1024         (FuncitonTests.test_get__processes): Skip this test on Windows.
1025
1026 2011-03-01  Ilya Tikhonovsky  <loislo@chromium.org>
1027
1028         Reviewed by Pavel Feldman.
1029
1030         Web Inspector: reduce inspector tests flakiness.
1031         https://bugs.webkit.org/show_bug.cgi?id=55401
1032
1033         * DumpRenderTree/chromium/TestShell.cpp:
1034         (TestShell::runFileTest):
1035         * DumpRenderTree/chromium/WebPreferences.cpp:
1036         (WebPreferences::reset):
1037
1038 2011-02-28  Mihai Parparita  <mihaip@chromium.org>
1039
1040         Reviewed by James Robinson.
1041
1042         NRWT should still upload results when it exits early due to too many failures
1043         https://bugs.webkit.org/show_bug.cgi?id=55432
1044
1045         Only keyboard interrupts should prevent results from being uploaded.
1046
1047         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
1048         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1049
1050 2011-02-28  Alice Liu  <alice.liu@apple.com>
1051
1052         Reviewed by Adam Roben and Gavin Barraclough.
1053
1054         Test for https://bugs.webkit.org/show_bug.cgi?id=54898
1055         InjectedBundleNodeHandle dies too early in WKBundleHitTestResultGetNodeHandle.
1056
1057         * TestWebKitAPI/InjectedBundleController.h:
1058         (TestWebKitAPI::InjectedBundleController::bundle):
1059         * TestWebKitAPI/PlatformWebView.h:
1060         * TestWebKitAPI/Tests/WebKit2/HitTestResultNodeHandle.cpp: Added.
1061         (TestWebKitAPI::didFinishLoadForFrame):
1062         (TestWebKitAPI::didReceiveMessageFromInjectedBundle):
1063         (TestWebKitAPI::setPageLoaderClient):
1064         (TestWebKitAPI::TEST):
1065         * TestWebKitAPI/Tests/WebKit2/HitTestResultNodeHandle_Bundle.cpp: Added.
1066         (TestWebKitAPI::HitTestResultNodeHandleTest::HitTestResultNodeHandleTest):
1067         (TestWebKitAPI::HitTestResultNodeHandleTest::getContextMenuFromDefaultMenu):
1068         (TestWebKitAPI::HitTestResultNodeHandleTest::didCreatePage):
1069         * TestWebKitAPI/mac/PlatformWebViewMac.mm:
1070         (TestWebKitAPI::PlatformWebView::simulateRightClick):
1071         * TestWebKitAPI/win/PlatformWebViewWin.cpp:
1072         (TestWebKitAPI::PlatformWebView::simulateRightClick):
1073
1074         Added files to:
1075         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
1076         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
1077         * TestWebKitAPI/win/TestWebKitAPIInjectedBundle.vcproj:
1078
1079 2011-02-28  Adam Roben  <aroben@apple.com>
1080
1081         Start running InjectedBundle API tests on Windows
1082
1083         A new project, TestWebKitAPIInjectedBundle, builds the injected bundle DLL.
1084
1085         Fixes <http://webkit.org/b/55420> run-api-tests doesn't run injected bundle tests on Windows
1086
1087         Reviewed by Darin Adler.
1088
1089         * TestWebKitAPI/Configurations/TestWebKitAPIDebug.vsprops: Added.
1090         * TestWebKitAPI/Configurations/TestWebKitAPIDebugAll.vsprops: Added.
1091         * TestWebKitAPI/Configurations/TestWebKitAPIDebugCairoCFLite.vsprops: Added.
1092         * TestWebKitAPI/Configurations/TestWebKitAPIRelease.vsprops: Added.
1093         * TestWebKitAPI/Configurations/TestWebKitAPIReleaseCairoCFLite.vsprops: Added.
1094         These are just extracted from TestWebKitAPI.vcproj.
1095
1096         * TestWebKitAPI/Configurations/TestWebKitAPIInjectedBundleCommon.vsprops: Added. Just sets
1097         up the DLL's filename.
1098
1099         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp: Fixed DLL name.
1100         (TestWebKitAPI::Util::createInjectedBundlePath): Changed to use kCFAllocatorNull instead of
1101         0 (== kCFAllocatorDefault) so that CF won't try to deallocate our C string.
1102
1103         * TestWebKitAPI/win/TestWebKitAPI.sln: Added TestWebKitAPIInjectedBundle. Removed the bogus
1104         "all" configuration.
1105
1106         * TestWebKitAPI/win/TestWebKitAPI.vcproj: Changed to use the new .vsprops files. Added
1107         injected bundle tests. Let VS resort things.
1108
1109         * TestWebKitAPI/win/TestWebKitAPIInjectedBundle.vcproj: Copied from Tools/TestWebKitAPI/win/TestWebKitAPI.vcproj.
1110
1111 2011-02-28  Adam Roben  <aroben@apple.com>
1112
1113         Add a test for the paint rect passed via WM_PAINT to windowless plugins
1114
1115         Test for <http://webkit.org/b/55365> <rdar://problem/9031089> REGRESSION (r79040): WebKit2:
1116         Large portions of pages with plugins paint black
1117
1118         Reviewed by Anders Carlsson.
1119
1120         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
1121         (PluginTest::NPP_HandleEvent):
1122         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
1123         Added NPP_HandleEvent. Just returns 0 at this level.
1124
1125         * DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowlessPaintRectCoordinates.cpp: Added.
1126         (WindowlessPaintRectCoordinates::WindowlessPaintRectCoordinates): Call up to the base class.
1127         (WindowlessPaintRectCoordinates::NPP_New): Mark ourselves as windowless.
1128         (WindowlessPaintRectCoordinates::NPP_HandleEvent): Check that the paint rect passed via
1129         WM_PAINT has the expected coordinates.
1130
1131         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
1132         (NPP_HandleEvent): Give the PluginTest a chance to handle the event before doing anything
1133         else.
1134
1135         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Added new file, let VS
1136         resort things.
1137
1138         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1139         (WTR::InjectedBundlePage::didFinishLoadForFrame): Added a call to WKBundlePageForceRepaint.
1140         This is roughly equivalent to the call to -[WebView displayIfNeeded] in -[FrameLoadDelegate
1141         webView:didFinishLoadForFrame:] in DumpRenderTree.
1142
1143         * WebKitTestRunner/win/PlatformWebViewWin.cpp:
1144         (WTR::PlatformWebView::PlatformWebView): Tell the WKView it's in a window so that plugins
1145         can start running.
1146
1147 2011-02-28  Adam Roben  <aroben@apple.com>
1148
1149         Tell MiniBrowser's WKViews that they're in a window
1150
1151         Fixes <http://webkit.org/b/55364> Plugins don't work in MiniBrowser on Windows
1152
1153         Reviewed by Anders Carlsson.
1154
1155         * MiniBrowser/win/BrowserView.cpp:
1156         (BrowserView::create): Call WKViewSetIsInWindow after creating the view.
1157
1158 2011-02-28  Alexis Menard  <alexis.menard@openbossa.org>
1159
1160         Reviewed by Csaba Osztrogonác.
1161
1162         The new QML tests require QML_IMPORT_PATH variable to be set in order to
1163         load the proper plugin.
1164
1165         * Scripts/run-qtwebkit-tests:
1166
1167 2011-02-28  Joone Hur  <joone.hur@collabora.co.uk>
1168
1169         Unreviewed, fix spelling mistake.
1170
1171         hardware-accelated => hardware-accelerated.
1172
1173         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1174
1175 2011-02-28  Kristian Amlie  <kristian.amlie@nokia.com>
1176
1177         Reviewed by Andreas Kling.
1178
1179         Made sure that the build-webkit --qmake argument is always respected.
1180
1181         Previously, it would be respected during the build, but not during
1182         the early feature detection. The build would also fail if qmake was
1183         not in your path, even if you specified --qmake with a valid qmake
1184         binary.
1185
1186         [Qt] WebKit patches required to work with a modularized version of Qt
1187         https://bugs.webkit.org/show_bug.cgi?id=53916
1188
1189         * Scripts/build-webkit:
1190         * Scripts/webkitdirs.pm:
1191
1192 2011-02-27  Ojan Vafai  <ojan@chromium.org>
1193
1194         Fix failing tests from http://trac.webkit.org/changeset/79837.
1195
1196         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
1197         This was just a mistake that would break if someone tried printing with this configuration.
1198         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
1199         Method signature changed. Didn't notice that this method was being calling in a unittest.
1200
1201 2011-02-24  Hayato Ito  <hayato@chromium.org>
1202
1203         Reviewed by Eric Seidel.
1204
1205         [NRWT] Remove test_types/* classes.
1206
1207         This is the last patch in the series of efforts which tried to remove test_types classes.
1208         A single_test_runner module will take over responsibilities of test_type classes.
1209
1210         From this patch, we won't display statistics per test_type even if a '--verbose' option is given.
1211
1212         https://bugs.webkit.org/show_bug.cgi?id=55123
1213
1214         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
1215         * Scripts/webkitpy/layout_tests/layout_package/test_results.py:
1216         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:
1217         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
1218         * Scripts/webkitpy/layout_tests/layout_package/worker_mixin.py:
1219         * Scripts/webkitpy/layout_tests/test_types/__init__.py: Removed.
1220         * Scripts/webkitpy/layout_tests/test_types/image_diff.py: Removed.
1221         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py: Removed.
1222         * Scripts/webkitpy/layout_tests/test_types/test_type_base_unittest.py: Removed.
1223         * Scripts/webkitpy/layout_tests/test_types/text_diff.py: Removed.
1224
1225 2011-02-25  Ojan Vafai  <ojan@chromium.org>
1226
1227         Reviewed by Tony Chang.
1228
1229         Change results.json format to the one used by unexpected_results.json
1230         https://bugs.webkit.org/show_bug.cgi?id=52267
1231
1232         Also add runtimes in milliseconds to the JSON and make the output format more compact.
1233         Named the file full_results.json to avoid conflicting with the results.json
1234         file the test-results server currently serves up.
1235
1236         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
1237         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1238         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
1239         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
1240         * Scripts/webkitpy/layout_tests/layout_package/result_summary.py:
1241         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
1242
1243 2011-02-27  Adam Roben  <aroben@apple.com>
1244
1245         Followup fix to r79827
1246
1247         * DumpRenderTree/win/UIDelegate.cpp:
1248         (UIDelegate::webViewAddMessageToConsole): Replaced call to urlSuitableForTestResult with
1249         lastPathComponent. This matches what we did here before r79827, and matches Mac.
1250
1251 2011-02-27  Dan Bernstein  <mitz@apple.com>
1252
1253         Reviewed by Adam Roben.
1254
1255         Make bisect-builds print a trac link for the regression range
1256         https://bugs.webkit.org/show_bug.cgi?id=55331
1257
1258         * Scripts/bisect-builds:
1259
1260 2011-02-27  Adam Roben  <aroben@apple.com>
1261
1262         Use iswalpha instead of isalpha when dealing with wchar_ts in EditingDelegate
1263
1264         Fixes <http://webkit.org/b/55062> <rdar://problem/9059907> Crash beneath
1265         EditingDelegate::checkSpellingOfString when running fast/forms/input-text-maxlength.html or
1266         fast/forms/input-text-paste-maxlength.html on Windows with full page heap enabled
1267
1268         Reviewed by Anders Carlsson.
1269
1270         * DumpRenderTree/win/EditingDelegate.cpp:
1271         (indexOfFirstWordCharacter):
1272         (wordLength):
1273         Use iswalpha instead of isalpha, since TCHAR is really just wchar_t.
1274
1275 2011-02-27  Adam Roben  <aroben@apple.com>
1276
1277         Make ResourceLoadDelegate print URLs relative to the main resource on Windows
1278
1279         This matches Mac. (In fact, the code was ported from Mac!)
1280
1281         Fixes <http://webkit.org/b/55328> URLs printed by ResourceLoadDelegate on Windows don't
1282         match Mac
1283
1284         Reviewed by Anders Carlsson.
1285
1286         * DumpRenderTree/win/DumpRenderTree.cpp:
1287         (substringFromIndex): Added. Emulates -[NSString substringFromIndex:].
1288         (urlSuitableForTestResult): Ported code from -[NSURL(DRTExtras)
1289         _drt_descriptionSuitableForTestResult].
1290         (cfStringRefToWString): Moved here from LayoutTestControllerWin.cpp.
1291
1292         * DumpRenderTree/win/DumpRenderTreeWin.h: Added declaration of cfStringRefToWString.
1293
1294         * DumpRenderTree/win/LayoutTestControllerWin.cpp: Moved cfStringRefToWString to
1295         DumpRenderTree.cpp.
1296
1297 2011-02-27  Adam Roben  <aroben@apple.com>
1298
1299         Change the order of Chang Shu's email addresses so the first one is their Bugzilla username
1300
1301         This makes Bugzilla autocompletion work correctly.
1302
1303         * Scripts/webkitpy/common/config/committers.py:
1304
1305 2011-02-27  Patrick Gansterer  <paroga@webkit.org>
1306
1307         Reviewed by Andreas Kling.
1308
1309         [Qt] fast/loader/user-stylesheet-fast-path.html fails
1310         https://bugs.webkit.org/show_bug.cgi?id=50911
1311
1312         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1313         (LayoutTestController::setUserStyleSheetLocation):
1314         Use QUrl::fromEncoded since the passed string is already encoded.
1315
1316 2011-02-27  Adam Roben  <aroben@apple.com>
1317
1318         Ensure $testResultsDirectory is an absolute path before setting up Windows crash log saving
1319
1320         Windows 7 (and maybe Vista?) launches the post-mortem debugger with a working directory that
1321         is different from old-run-webkit-test's. Because of this, the path to the crash log file
1322         (which is relative to $testResultsDirectory) must be an absolute path. See the bug for more
1323         details.
1324
1325         Fixes <http://webkit.org/b/55318> Crash logs aren't saved on Windows 7 when a relative
1326         --results-directory path is used (like on the test slaves)
1327
1328         Reviewed by Dan Bernstein.
1329
1330         * Scripts/old-run-webkit-tests: Moved the call to setUpWindowsCrashLogSaving after we've
1331         made $testResultsDirectory an absolute path.
1332
1333 2011-02-26  Dominic Mazzoni  <dmazzoni@google.com>
1334
1335         Reviewed by Adam Barth.
1336
1337         m_dumpAccessibilityNotifications should be initialized.
1338         https://bugs.webkit.org/show_bug.cgi?id=54874
1339
1340         * DumpRenderTree/chromium/AccessibilityController.cpp:
1341         (AccessibilityController::AccessibilityController):
1342
1343 2011-02-26  Vsevolod Vlasov  <vsevik@chromium.org>
1344
1345         Reviewed by Pavel Feldman.
1346
1347         DumpRenderTree should reset frame opener between tests.
1348         https://bugs.webkit.org/show_bug.cgi?id=54874
1349
1350         Added clearOpener() calls between tests.
1351         Some tests are using window.open() to change current window location.
1352         This changes window.opener property whose nullity is checked before
1353         showing xml viewer.
1354
1355         * DumpRenderTree/chromium/TestShell.cpp:
1356         (TestShell::resetTestController):
1357         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1358         (resetDefaultsToConsistentValues):
1359         * DumpRenderTree/mac/DumpRenderTree.mm:
1360         (resetWebViewToConsistentStateBeforeTesting):
1361         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1362         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
1363         * DumpRenderTree/win/DumpRenderTree.cpp:
1364         (resetWebViewToConsistentStateBeforeTesting):
1365         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1366         (WTR::InjectedBundlePage::reset):
1367
1368 2011-02-26  Yael Aharon  <yael.aharon@nokia.com>
1369
1370         Reviewed by Andreas Kling.
1371
1372         [Qt] Fix the focus in MiniBrowser.
1373         https://bugs.webkit.org/show_bug.cgi?id=55288
1374
1375         Give focus to the web view when a load starts.
1376
1377         * MiniBrowser/qt/BrowserWindow.cpp:
1378         (BrowserWindow::urlChanged):
1379
1380 2011-02-26  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
1381
1382         Reviewed by Adam Barth.
1383
1384         [EFL] Add coding style exceptions for EFL port
1385         https://bugs.webkit.org/show_bug.cgi?id=54733
1386
1387         The EFL APIs use EFL naming style, which includes both lower-cased and camel-cased,
1388         underscore-sparated values. But, style checking rule have found parameter_name errors
1389         in efl directories so far.
1390
1391         * Scripts/webkitpy/style/checker.py: Add prameter_name rule's exception and efl directories.
1392
1393 2011-02-26  David Levin  <levin@chromium.org>
1394
1395         Reviewed by Darin Adler.
1396
1397         check-webkit-style should check for spaces after periods in comments
1398         https://bugs.webkit.org/show_bug.cgi?id=55253
1399
1400         * Scripts/webkitpy/style/checkers/cpp.py: Added the check.
1401         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added tests
1402           and fixed the name of one related test.
1403
1404 2011-02-26  Kenneth Rohde Christiansen  <kenneth@webkit.org>
1405
1406         Reviewed by Andreas Kling.
1407
1408         Make it possible to test the targetdensity-dpi support
1409         https://bugs.webkit.org/show_bug.cgi?id=55142
1410
1411         Test the viewport meta tag feature targetdensity-dpi by
1412         adding extra arguments to dumpConfigurationForViewport
1413
1414         Add the new arguments to the shared header and all
1415         implementation.
1416
1417         * DumpRenderTree/LayoutTestController.cpp:
1418         (dumpConfigurationForViewportCallback):
1419         * DumpRenderTree/LayoutTestController.h:
1420         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1421         (LayoutTestController::dumpConfigurationForViewport):
1422         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1423         (LayoutTestController::dumpConfigurationForViewport):
1424         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1425         (LayoutTestController::dumpConfigurationForViewport):
1426         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
1427         (LayoutTestController::dumpConfigurationForViewport):
1428         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1429         (LayoutTestController::dumpConfigurationForViewport):
1430         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1431
1432 2011-02-26  Sheriff Bot  <webkit.review.bot@gmail.com>
1433
1434         Unreviewed, rolling out r79764.
1435         http://trac.webkit.org/changeset/79764
1436         https://bugs.webkit.org/show_bug.cgi?id=55295
1437
1438         "broke Chromium builds" (Requested by rniwa on #webkit).
1439
1440         * DumpRenderTree/chromium/TestShell.cpp:
1441         (TestShell::resetTestController):
1442         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1443         (resetDefaultsToConsistentValues):
1444         * DumpRenderTree/mac/DumpRenderTree.mm:
1445         (resetWebViewToConsistentStateBeforeTesting):
1446         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1447         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
1448         * DumpRenderTree/win/DumpRenderTree.cpp:
1449         (resetWebViewToConsistentStateBeforeTesting):
1450         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1451         (WTR::InjectedBundlePage::reset):
1452
1453 2011-02-26  Yongjun Zhang  <yongjun_zhang@apple.com>
1454
1455         Reviewed by David Kilzer.
1456
1457         https://bugs.webkit.org/show_bug.cgi?id=48781
1458
1459         Add a resource load delegate method to query if WebCore should paint the default broken image for failed images.
1460
1461         Add a new resource load client method (shouldPaintBrokenImage).  WebKit client can decide if WebCore
1462         should paint the default broken image when an image fails to load or decode.
1463
1464         * DumpRenderTree/LayoutTestController.cpp:
1465         (LayoutTestController::LayoutTestController):
1466         (setShouldPaintBrokenImageCallback):
1467         (LayoutTestController::staticFunctions):
1468         (LayoutTestController::setShouldPaintBrokenImage):
1469         * DumpRenderTree/LayoutTestController.h:
1470         (LayoutTestController::shouldPaintBrokenImage):
1471         * DumpRenderTree/mac/ResourceLoadDelegate.mm:
1472         (-[ResourceLoadDelegate webView:shouldPaintBrokenImageForURL:]):
1473
1474 2011-02-26  Vsevolod Vlasov  <vsevik@chromium.org>
1475
1476         Reviewed by Pavel Feldman.
1477
1478         DumpRenderTree should reset frame opener between tests.
1479         https://bugs.webkit.org/show_bug.cgi?id=54874
1480
1481         Added clearOpener() calls between tests.
1482         Some tests are using window.open() to change current window location.
1483         This changes window.opener property whose nullity is checked before
1484         showing xml viewer.
1485
1486         * DumpRenderTree/chromium/TestShell.cpp:
1487         (TestShell::resetTestController):
1488         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1489         (resetDefaultsToConsistentValues):
1490         * DumpRenderTree/mac/DumpRenderTree.mm:
1491         (resetWebViewToConsistentStateBeforeTesting):
1492         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1493         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
1494         * DumpRenderTree/win/DumpRenderTree.cpp:
1495         (resetWebViewToConsistentStateBeforeTesting):
1496         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1497         (WTR::InjectedBundlePage::reset):
1498
1499 2011-02-25  Lukasz Slachciak  <l.slachciak@samsung.com>
1500
1501         Reviewed by Kenneth Rohde Christiansen.
1502
1503         [EFL] Added EWebLauncher command line option to turn on/off frame flattening
1504         https://bugs.webkit.org/show_bug.cgi?id=54809
1505
1506         * EWebLauncher/main.c: Handling -f option added.
1507         (on_key_down):
1508         (browserCreate):
1509         (main):
1510
1511 2011-02-25  Sheriff Bot  <webkit.review.bot@gmail.com>
1512
1513         Unreviewed, rolling out r79324.
1514         http://trac.webkit.org/changeset/79324
1515         https://bugs.webkit.org/show_bug.cgi?id=55260
1516
1517         Leopard Debug is once again running out of address space while
1518         linking (Requested by aroben on #webkit).
1519
1520         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
1521         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1522
1523 2011-02-25  Tony Chang  <tony@chromium.org>
1524
1525         Reviewed by James Robinson.
1526
1527         Clean up temp files left by crashing DRTs on chromium-win
1528         https://bugs.webkit.org/show_bug.cgi?id=55182
1529
1530         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1531         * BuildSlaveSupport/chromium/remove-crash-logs: Added.
1532
1533 2011-02-25  Adam Roben  <aroben@apple.com>
1534
1535         Double-check that modifying the registry worked
1536
1537         On Windows Vista/7 with UAC enabled, regtool will fail to modify the registry, but it will
1538         still return a successful exit code. So we double-check here that the value we tried to
1539         write to the registry was really written.
1540
1541         Fixes <http://webkit.org/b/55227> old-run-webkit-tests falsely claims to be saving crash
1542         logs on Windows Vista/7
1543
1544         Reviewed by Darin Adler.
1545
1546         * Scripts/old-run-webkit-tests:
1547         (writeRegistryString): After trying to write the string value, check that it was really
1548         written.
1549
1550 2011-02-25  Adam Roben  <aroben@apple.com>
1551
1552         Set the 32-bit post-mortem debugger on 64-bit OSes
1553
1554         The post-mortem debugger is controlled by some registry values in
1555         HKLM\SOFTWARE\Microsoft\Windows NT\CurrentVersion\AeDebug. On 64-bit OSes, you can either
1556         set the values in that key, which will be used for 64-bit apps, or in the equivalent key
1557         under Wow6432Node, which will be used for 32-bit apps. Since DumpRenderTree and
1558         WebKitTestRunner are 32-bit, we want to use the latter location.
1559
1560         Fixes <http://webkit.org/b/55225> old-run-webkit-tests sets the post-mortem debugger for
1561         64-bit apps, but should set it for 32-bit apps instead
1562
1563         Reviewed by Darin Adler.
1564
1565         * Scripts/old-run-webkit-tests:
1566         (readRegistryString):
1567         (writeRegistryString):
1568         Pass --wow32 to regtool so it will get/set values beneath Wow6432Node, which is where
1569         Windows looks for the post-mortem debugger values when a 32-bit app crashes on a 64-bit OS.
1570         Passing --wow32 on a 32-bit OS will not have any effect.
1571
1572 2011-02-25  Adam Roben  <aroben@apple.com>
1573
1574         Clean up code to set up crash log saving on Windows
1575
1576         Fixes <http://webkit.org/b/55224> old-run-webkit-test's Windows registry code is ugly
1577
1578         Reviewed by David Kilzer.
1579
1580         * Scripts/old-run-webkit-tests:
1581         (readRegistryString):
1582         (writeRegistryString):
1583         Moved code to read strings from and write strings to the registry here...
1584
1585         (setUpWindowsCrashLogSaving): ...from here.
1586
1587 2011-02-25  Adam Roben  <aroben@apple.com>
1588
1589         Work around Cygwin's crash-suppression behavior
1590
1591         Cygwin calls ::SetErrorMode(SEM_FAILCRITICALERRORS), which any processes it launches will
1592         inherit. This is bad for testing/debugging, as it causes the post-mortem debugger not to be
1593         invoked. (Cygwin does this because it makes crashes more UNIX-y.) We reset the error mode
1594         when our test apps launch to work around Cygwin's behavior.
1595
1596         Fixes <http://webkit.org/b/55222> Test apps crash silently (without invoking post-mortem
1597         debugger) when launched from Cygwin 1.7
1598
1599         Reviewed by Darin Adler.
1600
1601         * DumpRenderTree/win/DumpRenderTree.cpp:
1602         (main):
1603         * TestWebKitAPI/win/main.cpp:
1604         (main):
1605         * WebKitAPITest/main.cpp:
1606         (main):
1607         * WebKitTestRunner/win/TestControllerWin.cpp:
1608         (WTR::TestController::platformInitialize):
1609         Call ::SetErrorMode(0) to undo Cygwin's folly.
1610
1611 2011-02-24  Jocelyn Turcotte  <jocelyn.turcotte@nokia.com>
1612
1613         Reviewed by Andreas Kling.
1614
1615         [Qt] Revert the support for QNAM affined to a different thread.
1616         https://bugs.webkit.org/show_bug.cgi?id=55149
1617
1618         Qt 4.8 will have QNAM use its own thread internally by default,
1619         no need to keep this complexity in WebKit.
1620
1621         This mainly reverts:
1622         http://trac.webkit.org/changeset/73710
1623         http://trac.webkit.org/changeset/73712
1624
1625         * QtTestBrowser/launcherwindow.cpp:
1626         (LauncherWindow::~LauncherWindow):
1627         (LauncherWindow::initializeView):
1628         (LauncherWindow::createChrome):
1629         * QtTestBrowser/launcherwindow.h:
1630         (WindowOptions::WindowOptions):
1631         * QtTestBrowser/webpage.cpp:
1632         (WebPage::WebPage):
1633         * QtTestBrowser/webpage.h:
1634
1635 2011-02-24  Adam Barth  <abarth@webkit.org>
1636
1637         Reviewed by Eric Seidel.
1638
1639         Add a script for creating gypi files
1640         https://bugs.webkit.org/show_bug.cgi?id=55113
1641
1642         This is a trivial script to create gypi files from the source code in
1643         the current directory.
1644
1645         * Scripts/make-gypi: Added.
1646
1647 2011-02-24  Sheriff Bot  <webkit.review.bot@gmail.com>
1648
1649         Unreviewed, rolling out r79628.
1650         http://trac.webkit.org/changeset/79628
1651         https://bugs.webkit.org/show_bug.cgi?id=55195
1652
1653         "broke chromium win tests" (Requested by dpranke on #webkit).
1654
1655         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1656         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1657
1658 2011-02-24  Dirk Pranke  <dpranke@chromium.org>
1659
1660         Reviewed by Mihai Parparita.
1661
1662         rebaseline-chromium-webkit-tests can't handle GPU variants
1663         https://bugs.webkit.org/show_bug.cgi?id=55002
1664
1665         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
1666         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
1667
1668 2011-02-24  Dirk Pranke  <dpranke@chromium.org>
1669
1670         Reviewed by Mihai Parparita.
1671
1672         nrwt: log command used to start web server to --verbose
1673
1674         https://bugs.webkit.org/show_bug.cgi?id=55188
1675
1676         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1677         * Scripts/webkitpy/layout_tests/port/http_server.py:
1678
1679 2011-02-24  Victoria Kirst  <vrk@google.com>
1680
1681         Reviewed by Mihai Parparita.
1682
1683         [chromium] Move media-file.js and video-test.js back to media/ for UILayoutTests
1684         https://bugs.webkit.org/show_bug.cgi?id=55089
1685
1686         This adds an alias into the media/ directory so that http/tests/media
1687         tests can access the media resources when running in an httpd process.
1688
1689         * Scripts/webkitperl/httpd.pm:
1690         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1691         * Scripts/webkitpy/layout_tests/port/http_server.py:
1692
1693 2011-02-24  Dirk Pranke  <dpranke@chromium.org>
1694
1695         Reviewed by Mihai Parparita.
1696
1697         nrwt: enable multiple processes by default on Chromium Win.
1698
1699         https://bugs.webkit.org/show_bug.cgi?id=55163
1700
1701         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1702         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1703
1704 2011-02-24  Sam Weinig  <sam@webkit.org>
1705
1706         Attempt to fix Qt.
1707
1708         * WebKitTestRunner/TestInvocation.cpp:
1709
1710 2011-02-24  James Robinson  <jamesr@chromium.org>
1711
1712         Unreviewed, rolling out r79598.
1713         http://trac.webkit.org/changeset/79598
1714         https://bugs.webkit.org/show_bug.cgi?id=55089
1715
1716         Appears to be causing trouble with httpd on windows
1717
1718         * Scripts/webkitperl/httpd.pm:
1719         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1720         * Scripts/webkitpy/layout_tests/port/http_server.py:
1721
1722 2011-02-24  Sam Weinig  <sam@webkit.org>
1723
1724         Reviewed by Darin Adler.
1725
1726         Add basic pixel testing support to WebKitTestRunner
1727         https://bugs.webkit.org/show_bug.cgi?id=55161
1728
1729         * Scripts/old-run-webkit-tests:
1730         Don't print color space information when using WebKitTestRunner since
1731         we don't need to change the screen.
1732
1733         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
1734         (WTR::InjectedBundle::initialize):
1735         Add call to initializePlatformDefaults.
1736
1737         (WTR::InjectedBundle::done):
1738         Return a dictionary on completion with both the text output and pixel dump override.
1739
1740         * WebKitTestRunner/InjectedBundle/InjectedBundle.h:
1741         Add declaration for initializePlatformDefaults.
1742
1743         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
1744         (WTR::LayoutTestController::LayoutTestController):
1745         (WTR::LayoutTestController::dumpAsText):
1746         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
1747         (WTR::LayoutTestController::shouldDumpPixels):
1748         Make setting dumpAsText override the dump pixels preference.
1749
1750         * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm: Added.
1751         (WTR::InjectedBundle::initializePlatformDefaults):
1752         Set platform defaults matching DumpRenderTree.
1753
1754         * WebKitTestRunner/InjectedBundle/qt/InjectedBundleQt.cpp: Added.
1755         (WTR::InjectedBundle::initializePlatformDefaults):
1756         * WebKitTestRunner/InjectedBundle/win/InjectedBundleWin.cpp: Added.
1757         (WTR::InjectedBundle::initializePlatformDefaults):
1758         Add stubs for initializePlatformDefaults.
1759
1760         * WebKitTestRunner/TestController.cpp:
1761         (WTR::TestController::runTest):
1762         (WTR::TestController::runTestingServerLoop):
1763         (WTR::TestController::run):
1764         Parse expected pixel results out of stdin and setup the invocation with it.
1765
1766         * WebKitTestRunner/TestInvocation.cpp:
1767         (WTR::TestInvocation::TestInvocation):
1768         Initialize new members.
1769
1770         (WTR::TestInvocation::~TestInvocation):
1771         (WTR::TestInvocation::setIsPixelTest):
1772         (WTR::sizeWebViewForCurrentTest):
1773         (WTR::TestInvocation::invoke):
1774         (WTR::TestInvocation::dump):
1775         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
1776         * WebKitTestRunner/TestInvocation.h:
1777         Dump pixels in addition to text by calling dumpPixelsAndCompareWithExpected.
1778
1779         * WebKitTestRunner/mac/PlatformWebViewMac.mm:
1780         (WTR::PlatformWebView::PlatformWebView):
1781         Use -[NSColorSpace genericRGBColorSpace] instead of the main screen color space.
1782
1783         * WebKitTestRunner/mac/TestInvocationMac.mm: Added.
1784         (WTR::createCGContextFromPlatformView):
1785         (WTR::computeMD5HashStringForContext):
1786         (WTR::dumpBitmap):
1787         (WTR::forceRepaintFunction):
1788         (WTR::TestInvocation::dumpPixelsAndCompareWithExpected):
1789         Add mac specific pixel dumping support. Right now, we always pull pixels
1790         from the window server.
1791
1792         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
1793         * WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro:
1794         * WebKitTestRunner/qt/WebKitTestRunner.pro:
1795         * WebKitTestRunner/win/InjectedBundle.vcproj:
1796         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
1797         Add new files.
1798
1799 2011-02-24  Chang Shu  <cshu@webkit.org>
1800
1801         Reviewed by Kenneth Rohde Christiansen.
1802
1803         [Qt] Launch MiniBrowser if the test run is WebKit2.
1804         https://bugs.webkit.org/show_bug.cgi?id=55145
1805
1806         * Scripts/old-run-webkit-tests:
1807         * Scripts/run-launcher:
1808         * Scripts/webkitdirs.pm:
1809
1810 2011-02-24  Victoria Kirst  <vrk@google.com>
1811
1812         Reviewed by Mihai Parparita.
1813
1814         [chromium] Move media-file.js and video-test.js back to media/ for UILayoutTests
1815         https://bugs.webkit.org/show_bug.cgi?id=55089
1816
1817         This adds an alias into the media/ directory so that http/tests/media
1818         tests can access the media resources when running in an httpd process.
1819
1820         * Scripts/webkitperl/httpd.pm:
1821         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1822         * Scripts/webkitpy/layout_tests/port/http_server.py:
1823
1824 2011-02-24  Andrew Wilson  <atwilson@chromium.org>
1825
1826         Unreviewed, rolling out r79570.
1827         http://trac.webkit.org/changeset/79570
1828         https://bugs.webkit.org/show_bug.cgi?id=54874
1829
1830         Breaks chromium build because glue/mocks/mock_web_frame.h/cc
1831         was not updated
1832
1833         * DumpRenderTree/chromium/TestShell.cpp:
1834         (TestShell::resetTestController):
1835         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1836         (resetDefaultsToConsistentValues):
1837         * DumpRenderTree/mac/DumpRenderTree.mm:
1838         (resetWebViewToConsistentStateBeforeTesting):
1839         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1840         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
1841         * DumpRenderTree/win/DumpRenderTree.cpp:
1842         (resetWebViewToConsistentStateBeforeTesting):
1843         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1844         (WTR::InjectedBundlePage::reset):
1845
1846 2011-02-24  Mahesh Kulkarni  <mahesh.kulkarni@nokia.com>
1847
1848         Reviewed by Laszlo Gombos.
1849
1850         [Qt] add android 2.2 user agent to QtTestBrowser
1851         https://bugs.webkit.org/show_bug.cgi?id=55085
1852
1853         Credit to Forrest Hodgkins to find right user-agent which works for youtube.
1854         * QtTestBrowser/useragentlist.txt:
1855
1856 2011-02-24  Adam Roben  <aroben@apple.com>
1857
1858         Change FrameLoadDelegate to support any number of delegates with delayed work to process
1859
1860         This makes our behavior match Mac more closely, and allows us to remove an incorrect
1861         assertion that was firing during some tests. (The assertion was claiming that there was
1862         never more than one delegate with delayed work to process, but that was not the case.)
1863
1864         Fixes <http://webkit.org/b/55146> Assertion failure in FrameLoadDelegate::locationChangeDone
1865         when running http/tests/navigation/back-twice-without-commit.html
1866
1867         Reviewed by Eric Carlson.
1868
1869         * DumpRenderTree/win/FrameLoadDelegate.cpp:
1870         (delegatesWithDelayedWork): Added. Returns all FrameLoadDelegates that have delayed work to
1871         process. A single delegate may appear in this Vector more than once (just as, on Mac, a
1872         single delegate may have multiple performSelector requests).
1873         (processWorkTimer): Pass the HWND to ::KillTimer, for pedantic brownie points. Added an
1874         assertion that the timer firing is the shared process work timer. Instead of using the
1875         single, global "delegate waiting for timer" delegate, give all delegates that have delayed
1876         work to process a chance to process their work.
1877         (FrameLoadDelegate::locationChangeDone): If we don't already have an active timer for
1878         processing delayed work, create one. Then add ourselves to the delegatesWithDelayedWork
1879         Vector so our processWork function will be called when the timer fires.
1880
1881 2011-02-24  Vsevolod Vlasov  <vsevik@chromium.org>
1882
1883         Reviewed by Alexey Proskuryakov.
1884
1885         DumpRenderTree should reset frame opener between tests.
1886         https://bugs.webkit.org/show_bug.cgi?id=54874
1887
1888         Added clearOpener() calls between tests.
1889         Some tests are using window.open() to change current window location.
1890         This changes window.opener property whose nullity is checked before
1891         showing xml viewer.
1892
1893         * DumpRenderTree/chromium/TestShell.cpp:
1894         (TestShell::resetTestController):
1895         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1896         (resetDefaultsToConsistentValues):
1897         * DumpRenderTree/mac/DumpRenderTree.mm:
1898         (resetWebViewToConsistentStateBeforeTesting):
1899         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1900         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
1901         * DumpRenderTree/win/DumpRenderTree.cpp:
1902         (resetWebViewToConsistentStateBeforeTesting):
1903         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1904         (WTR::InjectedBundlePage::reset):
1905
1906 2011-02-24  Adam Roben  <aroben@apple.com>
1907
1908         Don't trigger Windows builds for changes to WebCore.exp.in or the top-level GNUmakefile.am
1909
1910         Fixes <http://webkit.org/b/55144> Windows builders should not have built r79343 or r79440,
1911         but did
1912
1913         Reviewed by Anders Carlsson.
1914
1915         * Scripts/webkitpy/common/config/build.py:
1916         (_should_file_trigger_build): Loosened the pattern for GNUmakefile.am to include the
1917         top-level GNUmakefile.am, and loosened the pattern for .exp files to include WebCore.exp.in.
1918
1919         * Scripts/webkitpy/common/config/build_unittest.py:
1920         (ShouldBuildTest): Added some more tests.
1921
1922 2011-02-24  Andras Becsi  <abecsi@webkit.org>
1923
1924         Reviewed by Laszlo Gombos.
1925
1926         [Qt] MinGW build fails to link
1927         https://bugs.webkit.org/show_bug.cgi?id=55050
1928
1929         Prepend the libraries of subcomponents instead of appending them
1930         to fix the library order according to the dependency of the libraries
1931
1932         * WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro:
1933         * WebKitTestRunner/qt/WebKitTestRunner.pro:
1934
1935 2011-02-24  Gabor Loki  <loki@webkit.org>
1936
1937         Reviewed by Csaba Osztrogonác.
1938
1939         [Qt] Remove ARMv5 Release buildbot.
1940
1941         Remove ARMv5 Release buildbot, because there is very little difference
1942         between ARMv5 and ARMv7 Release bots, and there is also very little
1943         interest in ARMv5 build slave (compared to ARMv7).
1944
1945         * BuildSlaveSupport/build.webkit.org-config/config.json:
1946         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1947
1948 2011-02-22  Ojan Vafai  <ojan@chromium.org>
1949
1950         Reviewed by Tony Chang.
1951
1952         fix revision number finding when creating git patches
1953         https://bugs.webkit.org/show_bug.cgi?id=55024
1954
1955         HEAD~n does not walk merge commits correctly. git log however does.
1956
1957         * Scripts/webkitpy/common/checkout/scm.py:
1958         * Scripts/webkitpy/common/checkout/scm_unittest.py:
1959
1960 2011-02-23  Kenneth Russell  <kbr@google.com>
1961
1962         Unreviewed, rolling out r79387.
1963         https://bugs.webkit.org/show_bug.cgi?id=54885
1964
1965         Reapply the code changes from r79038 since they weren't the root
1966         cause of the test timeouts.
1967
1968         * DumpRenderTree/chromium/WebViewHost.cpp:
1969         (WebViewHost::reset):
1970
1971 2011-02-23  Siddharth Mathur  <siddharth.mathur@nokia.com>
1972
1973         Reviewed by Laszlo Gombos.
1974
1975         [Qt] Make sure Symbian binary UIDs are unique
1976
1977         * MiniBrowser/qt/MiniBrowser.pro: Resolve the UID collision with
1978         QtTestBrowser.pro by changing the UID.
1979
1980 2011-02-23  Patrick Gansterer  <paroga@webkit.org>
1981
1982         Reviewed by Darin Adler.
1983
1984         Rename PLATFORM(CF) to USE(CF)
1985         https://bugs.webkit.org/show_bug.cgi?id=53540
1986
1987         * DumpRenderTree/config.h:
1988
1989 2011-02-23  Dirk Pranke  <dpranke@chromium.org>
1990
1991         Reviewed by Ojan Vafai.
1992
1993         nrwt: turn on multiprocessing on chromium_linux, clean up unit tests for worker model defaults
1994         https://bugs.webkit.org/show_bug.cgi?id=55016
1995
1996         * Scripts/webkitpy/layout_tests/port/base.py:
1997         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1998         * Scripts/webkitpy/layout_tests/port/chromium_linux_unittest.py: Added.
1999         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
2000         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
2001         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2002         * Scripts/webkitpy/layout_tests/port/mac.py:
2003         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
2004         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
2005         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
2006
2007 2011-02-23  Tony Chang  <tony@chromium.org>
2008
2009         Reviewed by Kent Tamura.
2010
2011         [chromium] DumpRenderTree --check-layout-test-sys-deps leaks 4 temp dirs
2012         https://bugs.webkit.org/show_bug.cgi?id=55004
2013
2014         * DumpRenderTree/chromium/DumpRenderTree.cpp:
2015         (WebKitSupportTestEnvironment::WebKitSupportTestEnvironment):
2016         (WebKitSupportTestEnvironment::~WebKitSupportTestEnvironment):
2017         (main): Use a C++ object to ensure that TearDownTestEnvironment
2018             gets called when --check-layout-test-sys-deps is called
2019
2020 2011-02-23  Gopal Raghavan  <gopal.1.raghavan@nokia.com>
2021
2022         Reviewed by Anders Carlsson.
2023
2024         [Qt] [Qt] check webkit style fails in qt/tests
2025         https://bugs.webkit.org/show_bug.cgi?id=54998
2026
2027         check-webkit-style was complaining about missing config.h in WebKit/qt/test/qdeclarativewebview. This patch was originally part of 50222. Reviewer suggested to split and create a separate patch for checker.
2028
2029         * Scripts/webkitpy/style/checker.py:
2030
2031 2011-02-22  Vsevolod Vlasov  <vsevik@chromium.org>
2032
2033         Reviewed by Pavel Feldman.
2034
2035         DumpRenderTree should dump as text if test path contains "dumpAsText/"
2036         https://bugs.webkit.org/show_bug.cgi?id=54873
2037
2038         * DumpRenderTree/chromium/LayoutTestController.h:
2039         (LayoutTestController::setShouldDumpAsText):
2040         (LayoutTestController::setShouldGeneratePixelResults):
2041         * DumpRenderTree/chromium/TestShell.cpp:
2042         (TestShell::runFileTest):
2043         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2044         (shouldDumpAsText):
2045         (dump):
2046         (runTest):
2047         * DumpRenderTree/mac/DumpRenderTree.mm:
2048         (shouldDumpAsText):
2049         (runTest):
2050         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
2051         (WebCore::isDumpAsTextTest):
2052         (WebCore::DumpRenderTree::open):
2053         * DumpRenderTree/win/DumpRenderTree.cpp:
2054         (shouldDumpAsText):
2055         (runTest):
2056         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2057         (WTR::InjectedBundlePage::dump):
2058
2059 2011-02-22  Kenneth Russell  <kbr@google.com>
2060
2061         Reviewed by James Robinson.
2062
2063         Many tests started timing out on the chrome mac canaries around r79035
2064         https://bugs.webkit.org/show_bug.cgi?id=54885
2065
2066         Revert the code changes, but not the Chromium DEPS roll, from
2067         r79038 to see whether this clears up the timing-out tests. Was not
2068         able to reproduce the timeouts locally with a Release mode build
2069         of DRT (on Snow Leopard).
2070
2071         * DumpRenderTree/chromium/WebViewHost.cpp:
2072         (WebViewHost::reset):
2073
2074 2011-02-22  Dirk Pranke  <dpranke@chromium.org>
2075
2076         Reviewed by James Robinson.
2077
2078         nrwt: --platform chromium-gpu doesn't work with --worker-model=processes
2079         https://bugs.webkit.org/show_bug.cgi?id=55009
2080
2081         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2082
2083 2011-02-18  Mihai Parparita  <mihaip@chromium.org>
2084
2085         Reviewed by Ojan Vafai.
2086
2087         Default to 500 for --exit-after-n-failures
2088         https://bugs.webkit.org/show_bug.cgi?id=54773
2089
2090         Make 500 the default value for --exit-after-n-failures (ORWT bots
2091         already default to this as of r75726). Similarly, change the default for
2092         --exit-after-n-crashes-or-timeouts to 20.
2093
2094         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2095
2096 2011-02-02  Martin Robinson  <mrobinson@igalia.com>
2097
2098         Reviewed by Xan Lopez.
2099
2100         [GTK] DRT needs an implementation of LayoutTestController.setSelectTrailingWhitespace
2101         https://bugs.webkit.org/show_bug.cgi?id=53603
2102
2103         Add support for LayoutTestController.setSelectTrailingWhitespace by calling through
2104         to DumpRenderTreeSupportGtk for this functionality.
2105
2106         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2107         (resetDefaultsToConsistentValues): Always call LayoutTestController.setSelectTrailingWhitespace to
2108         false when reseting values between tests.
2109         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2110         (LayoutTestController::setSelectTrailingWhitespaceEnabled): Call LayoutTestController.setSelectTrailingWhitespace
2111         with the appropriate value.
2112
2113 2011-02-22  Adam Roben  <aroben@apple.com>
2114
2115         Don't trigger a Windows build when the top-level ChangeLog changes
2116
2117         r79320 is an example of a change that shouldn't have built on Windows, but did, because of
2118         this bug.
2119
2120         Fixes <http://webkit.org/b/54957> Windows bots build when top-level ChangeLog changes, but
2121         shouldn't
2122
2123         Reviewed by Anders Carlsson.
2124
2125         * Scripts/webkitpy/common/config/build.py:
2126         (_should_file_trigger_build): Modified the ChangeLog pattern to match the top-level
2127         ChangeLog, and re-sorted it in the list.
2128
2129         * Scripts/webkitpy/common/config/build_unittest.py:
2130         (ShouldBuildTest): Added a test that ChangeLogs don't trigger builds.
2131
2132 2011-02-22  Sheriff Bot  <webkit.review.bot@gmail.com>
2133
2134         Unreviewed, rolling out r78691.
2135         http://trac.webkit.org/changeset/78691
2136         https://bugs.webkit.org/show_bug.cgi?id=54956
2137
2138         Leopard Debug is successfully building again (Requested by
2139         aroben on #webkit).
2140
2141         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
2142         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2143
2144 2011-02-22  Ilya Tikhonovsky  <loislo@chromium.org>
2145
2146         Reviewed by Pavel Feldman.
2147
2148         Web Inspector: flakyness of inspector tests.
2149         https://bugs.webkit.org/show_bug.cgi?id=54729
2150
2151         As far as we have the protocol with sequence numbers we can simplify test harness support
2152         and drop out chomium specific methods.
2153
2154         * DumpRenderTree/DumpRenderTree.gypi:
2155         * DumpRenderTree/chromium/DRTDevToolsAgent.cpp:
2156         (DRTDevToolsAgent::sendMessageToInspectorFrontend):
2157         (DRTDevToolsAgent::runtimePropertyChanged):
2158         (DRTDevToolsAgent::asyncCall):
2159         (DRTDevToolsAgent::call):
2160         * DumpRenderTree/chromium/DRTDevToolsAgent.h:
2161         * DumpRenderTree/chromium/DRTDevToolsCallArgs.cpp: Removed.
2162         * DumpRenderTree/chromium/DRTDevToolsCallArgs.h: Removed.
2163         * DumpRenderTree/chromium/DRTDevToolsClient.cpp:
2164         (DRTDevToolsClient::sendMessageToBackend):
2165         (DRTDevToolsClient::asyncCall):
2166         (DRTDevToolsClient::call):
2167         * DumpRenderTree/chromium/DRTDevToolsClient.h:
2168
2169 2011-02-22  Andras Becsi  <abecsi@webkit.org>
2170
2171         Reviewed by Laszlo Gombos.
2172
2173         [Qt] Redesign the build system
2174         https://bugs.webkit.org/show_bug.cgi?id=51339
2175
2176         Part 2.
2177
2178         Build WebCore as a static library, compile the WebKit API and WebKit2 API
2179         in a final step and link to WebKit2, WebCore and JSC libraries to fix
2180         linking issues resulting from stripped away symbols.
2181
2182         * WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro: Remove superfluous includepaths.
2183
2184 2011-02-21  Lukasz Slachciak  <l.slachciak@samsung.com>
2185
2186         Reviewed by Antonio Gomes.
2187
2188         [EFL] Restored spatial navigation option in EWebLauncher since #18662 bug is closed
2189         https://bugs.webkit.org/show_bug.cgi?id=54806
2190
2191         * EWebLauncher/main.c:
2192         (on_key_down): calling ewk_view_setting_spatial_navigation_set when "F12" pressed added.
2193
2194 2011-02-21  Dirk Pranke  <dpranke@chromium.org>
2195
2196         Reviewed by Kent Tamura.
2197
2198         nrwt: revert behavior on 10.5 to "old-threads". It's possible
2199         that r79062 has introduced a bunch of flakiness, although I
2200         can't explain why right now. This change should restore the
2201         old behavior, and then we can look at the build logs to see
2202         if the tests stop timing out or being flaky.
2203
2204         https://bugs.webkit.org/show_bug.cgi?id=54925
2205
2206         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
2207         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
2208         * Scripts/webkitpy/layout_tests/port/mac.py:
2209         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
2210
2211 2011-02-21  Dirk Pranke  <dpranke@chromium.org>
2212
2213         Reviewed by Dimitri Glazkov.
2214
2215         nrwt: fix typo causing us to try and repeatedly shut down the servers
2216         https://bugs.webkit.org/show_bug.cgi?id=54904
2217
2218         * Scripts/webkitpy/layout_tests/layout_package/worker_mixin.py:
2219
2220 2011-02-21  Alejandro G. Castro  <alex@igalia.com>
2221
2222         Reviewed by Csaba Osztrogonác.
2223
2224         REGRESSION(r79157):
2225         plugins/get-url-with-javascript-destroying-plugin.html make the
2226         next test crash
2227         https://bugs.webkit.org/show_bug.cgi?id=54863
2228
2229         * GNUmakefile.am:
2230
2231 2011-02-20  Dirk Pranke  <dpranke@chromium.org>
2232
2233         Reviewed by Ojan Vafai.
2234
2235         test-webkitpy: stop skipping a bunch of tests on win32
2236         https://bugs.webkit.org/show_bug.cgi?id=54788
2237
2238         * Scripts/webkitpy/common/net/testoutputset_unittest.py:
2239         * Scripts/webkitpy/test/main.py:
2240
2241 2011-02-20  Dirk Pranke  <dpranke@chromium.org>
2242
2243         Reviewed by Ojan Vafai.
2244
2245         Fix various bugs related to handling of dir separators and real
2246         vs. fake filesystems that were causing test-webkitpy to fail
2247         tests on win32. These bugs have a variety of causes but are
2248         lumped together because they should be easy to review.
2249
2250         https://bugs.webkit.org/show_bug.cgi?id=54700
2251
2252         * Scripts/webkitpy/common/net/testoutput.py:
2253           Here we were using os.path.sep but not escaping it prior to
2254           using in a regexp, which  caused bad things on win32.
2255         * Scripts/webkitpy/common/net/testoutput_unittest.py:
2256           Same here.
2257         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2258           Here we were using the native filesystem instead of the mock
2259           filesystem, making some tests unnecessarily platform dependent.
2260           Also we weren't setting the default configuration for the
2261           TestChromiumWinPort.
2262         * Scripts/webkitpy/layout_tests/update_webgl_conformance_tests.py:
2263           Here we *shouldn't* use os.path.join(), just so the code becomes
2264           less platform dependent.
2265         * Scripts/webkitpy/style/checker.py:
2266           Here we should be using os.path.sep instead of "/"; the tests
2267           require the native filesystem.
2268         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
2269           Here the underlying code is using os.path.sep, so we probably
2270           should as well.
2271         * Scripts/webkitpy/style/test_expectations_unittest.py:
2272           Here we should be using the filesystem hanging of the port
2273           object (which is the mocked filesystem) rather than the native
2274           one.
2275
2276 2011-02-20  Anders Carlsson  <andersca@apple.com>
2277
2278         Reviewed by Maciej Stachowiak.
2279
2280         Crash when a plug-in requests a javascript: url that destroys the plug-in
2281         https://bugs.webkit.org/show_bug.cgi?id=54837
2282         <rdar://problem/9005475>
2283
2284         Add new plug-in test.
2285
2286         * DumpRenderTree/DumpRenderTree.gypi:
2287         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
2288         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
2289         (PluginTest::NPN_GetURL):
2290         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
2291         * DumpRenderTree/TestNetscapePlugIn/Tests/GetURLWithJavaScriptURLDestroyingPlugin.cpp: Added.
2292         (GetURLWithJavaScriptURLDestroyingPlugin::GetURLWithJavaScriptURLDestroyingPlugin):
2293         (GetURLWithJavaScriptURLDestroyingPlugin::NPP_New):
2294         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2295         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
2296
2297 2011-02-19  Patrick Gansterer  <paroga@webkit.org>
2298
2299         Reviewed by Andreas Kling.
2300
2301         Make WinCE a core builder
2302         https://bugs.webkit.org/show_bug.cgi?id=54801
2303
2304         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
2305         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2306
2307 2011-02-19  Lukasz Slachciak  <l.slachciak@samsung.com>
2308
2309         Reviewed by Andreas Kling.
2310
2311         Fixed problem with launching EWebLauncher with "run-launcher --efl". Incorrect path referring to non-existing .libs directory ws set.
2312         https://webkit.org/b/54778
2313
2314         * Scripts/webkitdirs.pm:
2315
2316 2011-02-18  David Levin  <levin@chromium.org>
2317
2318         Reviewed by Eric Seidel.
2319
2320         check-webkit-style: Misses brace style error when the line with the { has a } in it.
2321         https://bugs.webkit.org/show_bug.cgi?id=54769
2322
2323         * Scripts/webkitpy/style/checkers/cpp.py: Fixed the check for the close brace
2324           to only look after the last open brace, so that the open brace in this line
2325           "} else {" will still be able to trigger the error.
2326         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added related unit tests.
2327
2328 2011-02-18  Zan Dobersek  <zandobersek@gmail.com>
2329
2330         Reviewed by Martin Robinson.
2331
2332         [Gtk] Re-enable meter tag support
2333         https://bugs.webkit.org/show_bug.cgi?id=54762
2334
2335         Meter tag is supported on the Gtk port, enable it by default.
2336
2337         * Scripts/build-webkit:
2338
2339 2011-02-18  David Levin  <levin@chromium.org>
2340
2341         Reviewed by Eric Seidel.
2342
2343         check-webkit-style falsely complains about WebKitGTK+ public headers
2344         https://bugs.webkit.org/show_bug.cgi?id=54650
2345
2346         * Scripts/webkitpy/style/checker.py: Add some exceptions for the gtk files.
2347         * Scripts/webkitpy/style/checkers/cpp.py: Restrict the checks for WEBKIT_API
2348           to the chromium directory (and improved the checks slightly).
2349         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added corresponding tests.
2350
2351 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
2352
2353         Reviewed by James Robinson.
2354
2355         new-run-webkit-tests: enable multiple processes by default on
2356         mac. This change removes the artificial restrictions we placed
2357         on the # of child processes we used with the old threading
2358         model, and switches to the new message-based model and multiple
2359         processes, where available. If multiple processes are not
2360         available (Leopard / Python 2.5), then we use the 'inline' model
2361         instead of the 'old-threads' model on the 'mac' port or one
2362         process on the 'chromium-mac' port. We need additional testing
2363         to see if the new 'threads' model works reliably and is worth
2364         supporting, or if we should just live with things being slightly
2365         slow.
2366
2367         https://bugs.webkit.org/show_bug.cgi?id=54596
2368
2369         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
2370         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
2371         * Scripts/webkitpy/layout_tests/port/mac.py:
2372         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
2373         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2374
2375 2011-02-18  Andrew Wilson  <atwilson@chromium.org>
2376
2377         Unreviewed, rolling out r79047.
2378         http://trac.webkit.org/changeset/79047
2379         https://bugs.webkit.org/show_bug.cgi?id=54596
2380
2381         Broke canary bots - please watch the canaries the next time
2382         you land this
2383
2384         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
2385         * Scripts/webkitpy/layout_tests/port/mac.py:
2386
2387 2011-02-18  Emil A Eklund  <eae@chromium.org>
2388
2389         Reviewed by Darin Adler.
2390
2391         Crash in EventHandler::sendContextMenuEventForKey
2392         https://bugs.webkit.org/show_bug.cgi?id=54495
2393
2394         Add support for the context menu key (VK_APPS) to EventSender for
2395         platforms that support the key.
2396
2397         * DumpRenderTree/chromium/EventSender.cpp:
2398         (EventSender::keyDown):
2399         * DumpRenderTree/gtk/EventSender.cpp:
2400         (keyDownCallback):
2401         * DumpRenderTree/qt/EventSenderQt.cpp:
2402         (EventSender::keyDown):
2403         * DumpRenderTree/win/EventSender.cpp:
2404         (keyDownCallback):
2405
2406 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
2407
2408         Reviewed by Ojan Vafai.
2409
2410         new-run-webkit-tests: enable multiple processes by default on mac
2411         https://bugs.webkit.org/show_bug.cgi?id=54596
2412
2413         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
2414         * Scripts/webkitpy/layout_tests/port/mac.py:
2415
2416 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
2417
2418         Reviewed by Adam Roben.
2419
2420         Skip pretty patch unit tests if ruby isn't installed.
2421
2422         https://bugs.webkit.org/show_bug.cgi?id=54699
2423
2424         * Scripts/webkitpy/common/prettypatch_unittest.py:
2425
2426 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
2427
2428         Reviewed by Adam Roben.
2429
2430         Skip mac.test_skipped_file_paths() when running on win32; the
2431         test has hardcoded '/' directory separators and since it's a
2432         port-specific test and we have coverage on other ports it's not
2433         worth it to make the test more generic.
2434
2435         https://bugs.webkit.org/show_bug.cgi?id=54698
2436
2437         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
2438
2439 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
2440
2441         Reviewed by Adam Roben.
2442
2443         Fix webkitpy.layout_tests.port.config_unittest to work on Win32.
2444
2445         https://bugs.webkit.org/show_bug.cgi?id=54703
2446
2447         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
2448
2449 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
2450
2451         Reviewed by Adam Roben.
2452
2453         Change the default port we're mocking out on win32 when we
2454         run mock_drt_unittest. Normally we'd default to the
2455         port-specific default, but that would be "win", which doesn't
2456         currently work.
2457
2458         https://bugs.webkit.org/show_bug.cgi?id=54708
2459
2460         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
2461
2462 2011-02-18  Kenneth Russell  <kbr@google.com>
2463
2464         Reviewed by James Robinson.
2465
2466         [chromium] Clean up initialization of minimum timer interval
2467         https://bugs.webkit.org/show_bug.cgi?id=54772
2468
2469         Use new GetForegroundTabTimerInterval function in webkit_support
2470         to reset the minimum timer interval before each test.
2471
2472         * DumpRenderTree/chromium/WebViewHost.cpp:
2473         (WebViewHost::reset):
2474
2475 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
2476
2477         Reviewed by Adam Roben.
2478
2479         r78494 fixed the chromium_win unittest failing under win32,
2480         but not cygwin. This fixes cygwin as well.
2481
2482         https://bugs.webkit.org/show_bug.cgi?id=54652
2483
2484         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2485
2486 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
2487
2488         Reviewed by Ojan Vafai, Mihai Parparita.
2489
2490         new-run-webkit-tests: r78522 made it impossible to create a
2491         'chromium-win' or 'chromium-mac' port; they would always get the
2492         version tacked on to the end. It turns out that we actually need
2493         to be able to create these ports in order for
2494         rebaseline-chromium-webkit-tests to work correctly given the
2495         broken coverage of our layout bots. As soon as we get SL and
2496         Win 7 bots running reliably on the canaries, we can get rid of
2497         this change. Note that we only need to change the chromium_mac
2498         and win ports because linux doesn't yet use versions and that
2499         script doesn't work with non-chromium ports.
2500
2501         Also, this change fixes a bug where we were using
2502         "platform/chromium-win-win7" and
2503         "platform/chromium-mac-snowleopard" as the baseline dirs for
2504         Win 7 and SL; we should've been using the generic directories
2505         instead since they are the newest versions of the os's.
2506
2507         https://bugs.webkit.org/show_bug.cgi?id=54691
2508
2509         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
2510         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
2511         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2512         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2513         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2514
2515 2011-02-16  MORITA Hajime  <morrita@google.com>
2516
2517         Reviewed by Kent Tamura.
2518
2519         [Chromium][DRT] some PostDelayedTask() accesses possibly dangling this pointer.
2520         http://webkit.org/b/53899
2521
2522         Replaced static invocation wrappers with MethodTask objects,
2523         which are removed on WebViewHost destructor.
2524
2525         * DumpRenderTree/chromium/WebViewHost.cpp:
2526         (WebViewHost::requestCheckingOfText):
2527         (WebViewHost::scheduleAnimation):
2528         (WebViewHost::closeWidgetSoon):
2529         * DumpRenderTree/chromium/WebViewHost.h:
2530         (WebViewHost::HostMethodTask::HostMethodTask):
2531         (WebViewHost::HostMethodTask::runIfValid):
2532         (WebViewHost::taskList):
2533
2534 2011-02-07  Hayato Ito  <hayato@chromium.org>
2535
2536         Reviewed by Ojan Vafai.
2537
2538         Get rid of code which writes test results from test_type's
2539         compare_output() method.
2540         Writing test results now happens in single_test_runner, using
2541         newly introduced test_result_writer module.
2542
2543         Subsequent patches will eliminate test_type/* classes.
2544
2545         https://bugs.webkit.org/show_bug.cgi?id=53908
2546
2547         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
2548         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py: Added.
2549         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
2550         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
2551         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2552
2553 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
2554
2555         Reviewed by Adam Roben.
2556
2557         Add blacklist of remaining webkitpy/* modules that fail on
2558         Win32, so that test-webkitpy will at least run cleanly.
2559
2560         https://bugs.webkit.org/show_bug.cgi?id=54709
2561
2562         * Scripts/webkitpy/layout_tests/test/main.py:
2563
2564 2011-02-17  Simon Fraser  <simon.fraser@apple.com>
2565
2566         Reviewed by Sam Weinig.
2567
2568         Set a preference to ensure that scrollbars are in a predictable
2569         state for DRT.
2570
2571         * DumpRenderTree/mac/DumpRenderTree.mm:
2572         (resetDefaultsToConsistentValues):
2573
2574 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
2575
2576         Reviewed by Adam Roben.
2577
2578         test-webkitpy doesn't work with the multiprocessing module on win32
2579         or cygwin, so we skip it for now.
2580
2581         https://bugs.webkit.org/show_bug.cgi?id=54520
2582
2583         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
2584         * Scripts/webkitpy/layout_tests/layout_package/run_webkit_tests_unittest.py:
2585
2586 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
2587
2588         Reviewed by Adam Roben.
2589
2590         Fix version, baseline_search_path for chromium-win-win7. Also
2591         fix chromium_win_unittests, and fix port_testcase.make_port()
2592         to correctly use the options passed into it and only set
2593         results_directory() if necessary.
2594
2595         https://bugs.webkit.org/show_bug.cgi?id=54652
2596
2597         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2598         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2599         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
2600
2601 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
2602
2603         Reviewed by Ojan Vafai.
2604
2605         Fix a hang in new-run-webkit-tests when run with
2606         --dry-run --worker-model=processes
2607
2608         https://bugs.webkit.org/show_bug.cgi?id=54595
2609
2610         * Scripts/webkitpy/layout_tests/layout_package/test_runner2.py:
2611         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2612
2613 2011-02-17  Kevin Ollivier  <kevino@theolliviers.com>
2614
2615         [wx] Build fixes after recent changes.
2616
2617         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2618         (LayoutTestController::setMinimumTimerInterval):
2619         * wx/build/settings.py:
2620
2621 2011-02-17  Csaba Osztrogonác  <ossy@webkit.org>
2622
2623         Reviewed by Andreas Kling.
2624
2625         [Qt] ImageDiff sometimes hangs
2626         https://bugs.webkit.org/show_bug.cgi?id=54641
2627
2628         * DumpRenderTree/qt/ImageDiff.cpp: Clear actualImage and baselineImage after all comparison.
2629         (main):
2630
2631 2011-02-17  Peter Varga  <pvarga@webkit.org>
2632
2633         Reviewed by Csaba Osztrogonác.
2634
2635         The run-sunspider --v8-suite doesn't work
2636         https://bugs.webkit.org/show_bug.cgi?id=54660
2637
2638         * Scripts/run-sunspider:
2639
2640 2011-02-17  Gabor Rapcsanyi  <rgabor@webkit.org>
2641
2642         Reviewed by Csaba Osztrogonác.
2643
2644         [NRWT] Fix environment setup on Qt port.
2645         https://bugs.webkit.org/show_bug.cgi?id=54549
2646
2647         * Scripts/webkitpy/layout_tests/port/qt.py:
2648
2649 2011-02-16  Yael Aharon  <yael.aharon@nokia.com>
2650
2651         Reviewed by Andreas Kling.
2652
2653         [Qt] Enable tiled backing store by default in MiniBrowser.
2654         https://bugs.webkit.org/show_bug.cgi?id=54581
2655
2656         Instead of a flag to turn it on, a new flag is defined to turn
2657         tiled backing store off.
2658
2659         * MiniBrowser/qt/MiniBrowserApplication.cpp:
2660         (MiniBrowserApplication::handleUserOptions):
2661         * MiniBrowser/qt/MiniBrowserApplication.h:
2662         (WindowOptions::WindowOptions):
2663
2664 2011-02-16  Eric Seidel  <eric@webkit.org>
2665
2666         Reviewed by Mihai Parparita.
2667
2668         flaky test archive uploads are always 22 bytes long
2669         https://bugs.webkit.org/show_bug.cgi?id=54593
2670
2671         Speculative fix for the 22-bytes bug.
2672
2673         * Scripts/webkitpy/tool/bot/flakytestreporter.py:
2674
2675 2011-02-16  Anna Cavender  <annacc@chromium.org>
2676
2677         Reviewed by Mihai Parparita.
2678
2679         http/tests/media should be served over HTTP (not from local file)
2680         https://bugs.webkit.org/show_bug.cgi?id=54028
2681
2682         Removed special file:// treatment of http/tests/media
2683
2684         * Scripts/old-run-webkit-tests:
2685         * Scripts/webkitpy/layout_tests/port/base.py:
2686
2687 2011-02-16  Mihai Parparita  <mihaip@chromium.org>
2688
2689         Reviewed by Ojan Vafai.
2690
2691         ChangeLog.parse_latest_entry_from_file does not handle rolled over ChangeLogs
2692         https://bugs.webkit.org/show_bug.cgi?id=54609
2693
2694         r78737 rolled over ChangeLogs, and webkit-patch complains when landing
2695         patches such as this one that are the first to add an entry to the
2696         ChangeLog (since they don't see a second date line to know where the
2697         entry ends).
2698
2699         * Scripts/webkitpy/common/checkout/changelog.py:
2700         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
2701
2702 == Rolled over to ChangeLog-2011-02-16 ==