2010-11-24 Mihai Parparita <mihaip@chromium.org>
[WebKit-https.git] / WebKitTools / ChangeLog
1 2010-11-24  Mihai Parparita  <mihaip@chromium.org>
2
3         Reviewed by David Levin.
4
5         DRT should not generate pixel output for text/plain tests
6         https://bugs.webkit.org/show_bug.cgi?id=50002
7         
8         Modify Mac, Windows and GTK DRT implementations to disable pixel output
9         when encountering a text/plain response (the Chromium one already does
10         this, and none of the other ports special-case text/plain output in
11         the first place).
12
13         * DumpRenderTree/gtk/DumpRenderTree.cpp:
14         (dump):
15         * DumpRenderTree/mac/DumpRenderTree.mm:
16         (dump):
17         * DumpRenderTree/win/DumpRenderTree.cpp:
18         (dump):
19
20 2010-11-24  Andras Becsi  <abecsi@inf.u-szeged.hu>
21
22         Unreviewed.
23
24         Add my other email address to committers.py.
25
26         * Scripts/webkitpy/common/config/committers.py:
27
28 2010-11-24  Andras Becsi  <abecsi@inf.u-szeged.hu>
29
30         Reviewed by Csaba Osztrogonác.
31
32         Make build-webkit --minimal build with ENABLE_INSPECTOR=0
33         https://bugs.webkit.org/show_bug.cgi?id=49975
34
35         * Scripts/build-webkit: Add ENABLE_INSPECTOR define to features,
36         to enable the script to define it 0 in case of a minimal build.
37
38 2010-11-24  Martin Robinson  <mrobinson@igalia.com>
39
40         Reviewed by Xan Lopez.
41
42         [GTK] The WebKitWebView should expose a set of signals encapsulating the behavior of the EditorClient
43         https://bugs.webkit.org/show_bug.cgi?id=49143
44
45         * DumpRenderTree/gtk/DumpRenderTree.cpp:
46         (createWebView): Call the new connectEditingCallbacks function.
47         * DumpRenderTree/gtk/EditingCallbacks.cpp: Added.
48         (dumpNodePath):
49         (dumpRange):
50         (insertActionString):
51         (selectionAffinityString):
52         (shouldBeginEditing):
53         (shouldEndEditing):
54         (shouldInsertNode):
55         (shouldInsertText):
56         (shouldDeleteRange):
57         (shouldShowDeleteInterfaceForElement):
58         (shouldChangeSelectedRange):
59         (shouldApplyStyle):
60         (editingBegan):
61         (userChangedContents):
62         (editingEnded):
63         (selectionChanged):
64         (connectEditingCallbacks):
65         * DumpRenderTree/gtk/EditingCallbacks.h: Added.
66         * GNUmakefile.am: Added EditingCallbacks.{h,cpp} to the source list.
67
68 2010-11-24  Yi Shen  <yi.4.shen@nokia.com>
69
70         Reviewed by Andreas Kling.
71
72         [Qt] Make QtTestBrowser remember the selected user agent
73         https://bugs.webkit.org/show_bug.cgi?id=50021
74
75         Set user agent for a new page by using the old page's
76
77         * QtTestBrowser/mainwindow.cpp:
78         (MainWindow::setPage):
79
80 2010-11-24 Patrick Gansterer <paroga@webkit.org>
81
82         Reviewed by Csaba Osztrogonác.
83
84         Remove Bakefile build system files
85         https://bugs.webkit.org/show_bug.cgi?id=49983
86
87         r53757 only removed the content, but not the files.
88         This patch removes that empty files.
89
90         * DumpRenderTree/wx/DumpRenderTree.bkl: Removed.
91         * wx/browser/browser.bkl: Removed.
92         * wx/build-wxwebkit: Removed.
93
94 2010-11-23  Dirk Pranke  <dpranke@chromium.org>
95
96         Reviewed by Tony Chang.
97
98         This patch cleans up the logic used to shard tests into groups a
99         bit and adds the --worker-model flag to NRWT. The flag is only
100         used at the moment to control whether to run single-threaded or
101         not, but eventually will also allow toggling between threads and
102         processes.
103
104         Also add a minor cleanup with _test_is_slow(), which just
105         eliminates some repetition and gives slightly better encapsulation.
106
107         https://bugs.webkit.org/show_bug.cgi?id=49773
108
109         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
110         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
111
112 2010-11-23  Mihai Parparita  <mihaip@chromium.org>
113
114         Reviewed by Tony Chang.
115
116         Rebaseline server: list current baselines and platforms
117         https://bugs.webkit.org/show_bug.cgi?id=49991
118         
119         List existing baselines (with Trac links) for tests.
120         
121         Add dropdowns for choosing with platform baselines to target (and what
122         to do with existing ones).
123
124         Also fix MockFileSystem.join to behave more like os.path.join (unit
125         test was ending up with two consecutive slashes in a layout test
126         path).        
127
128         * Scripts/webkitpy/common/system/filesystem_mock.py:
129         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
130         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
131         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
132         * Scripts/webkitpy/tool/commands/data/rebaselineserver/queue.js:
133         * Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js:
134         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
135         * Scripts/webkitpy/tool/commands/rebaselineserver_unittest.py: Added.
136
137 2010-11-23  Ojan Vafai  <ojan@chromium.org>
138
139         Reviewed by Tony Chang.
140
141         remove timeout argument to urlopen
142         https://bugs.webkit.org/show_bug.cgi?id=49995
143
144         Apparently the version of python we have on the bots doesn't accept a timeout argument.
145
146         * Scripts/webkitpy/layout_tests/layout_package/test_results_uploader.py:
147
148 2010-11-23  Sheriff Bot  <webkit.review.bot@gmail.com>
149
150         Unreviewed, rolling out r72628.
151         http://trac.webkit.org/changeset/72628
152         https://bugs.webkit.org/show_bug.cgi?id=49994
153
154         This patch is causing layout-test failtures on GTK Linux
155         64-bit Debug (Requested by ctguil on #webkit).
156
157         * DumpRenderTree/chromium/AccessibilityUIElement.cpp:
158         (AccessibilityUIElement::AccessibilityUIElement):
159         (AccessibilityUIElement::parentElementCallback):
160         * DumpRenderTree/chromium/AccessibilityUIElement.h:
161         * DumpRenderTree/chromium/CppBoundClass.cpp:
162         * DumpRenderTree/chromium/CppBoundClass.h:
163
164 2010-11-22  Ojan Vafai  <ojan@chromium.org>
165
166         Reviewed by Tony Chang.
167
168         speculative fix for upload errors: stop using mechanize to upload to test-results.appspot.com
169         https://bugs.webkit.org/show_bug.cgi?id=49944
170
171         * Scripts/webkitpy/layout_tests/layout_package/test_results_uploader.py:
172
173 2010-11-23  Chris Guillory  <chris.guillory@google.com>
174
175         Reviewed by Dimitri Glazkov.
176
177         Include the FrameView widget of a RenderWidget in the accessibility tree.
178         https://bugs.webkit.org/show_bug.cgi?id=49106
179
180         * DumpRenderTree/chromium/AccessibilityUIElement.cpp:
181         (AccessibilityUIElement::AccessibilityUIElement):
182         (AccessibilityUIElement::parentElementCallback):
183         (AccessibilityUIElement::isEqualCallback):
184         * DumpRenderTree/chromium/AccessibilityUIElement.h:
185         * DumpRenderTree/chromium/CppBoundClass.cpp:
186         (CppBoundClass::getFromCppVariant):
187         * DumpRenderTree/chromium/CppBoundClass.h:
188
189 2010-11-23  Alex Grilo  <abgrilo@profusion.mobi>
190
191         Reviewed by Kenneth Rohde Christiansen.
192
193         [EFL] Add tiled backing store to EWebLauncher
194         https://bugs.webkit.org/show_bug.cgi?id=45777
195
196         Allow EWebLauncher to choose between single and tiled backing store
197
198         * EWebLauncher/main.c:
199         (on_key_down):
200         (browserCreate):
201         (main):
202
203 2010-11-23  Simon Hausmann  <simon.hausmann@nokia.com>
204
205         Reviewed by Tor Arne Vestbø.
206
207         [Qt] Review the setUserPermission & friends API
208         https://bugs.webkit.org/show_bug.cgi?id=46810
209
210         Renamed requestPermissionFromUser to featurePermissionRequested
211         and cancelRequestsForPermission to featurePermissionRequestCanceled.
212
213         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
214         (WebCore::WebPage::WebPage):
215         * QtTestBrowser/webpage.cpp:
216         (WebPage::WebPage):
217         (WebPage::featurePermissionRequestCanceled):
218         * QtTestBrowser/webpage.h:
219
220 2010-11-23  Simon Hausmann  <simon.hausmann@nokia.com>
221
222         Reviewed by Tor Arne Vestbø.
223
224         [Qt] Review the setUserPermission & friends API
225         https://bugs.webkit.org/show_bug.cgi?id=46810
226
227         Rename setUserPermission to setFeaturePermission
228
229         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
230         (WebCore::WebPage::requestPermission):
231         (WebCore::WebPage::permissionSet):
232         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
233         (LayoutTestController::grantDesktopNotificationPermission):
234         * QtTestBrowser/webpage.cpp:
235         (WebPage::requestPermission):
236
237 2010-11-23  Simon Hausmann  <simon.hausmann@nokia.com>
238
239         Reviewed by Tor Arne Vestbø.
240
241         [Qt] Review the setUserPermission & friends API
242         https://bugs.webkit.org/show_bug.cgi?id=46810
243
244         Rename PermissionDomain to Feature, NotificationsPermissionDomain to Notifications
245         and GeolocationPermissionDomain to Geolocation.
246
247         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
248         (WebCore::WebPage::WebPage):
249         (WebCore::WebPage::requestPermission):
250         (WebCore::WebPage::cancelPermission):
251         (WebCore::WebPage::permissionSet):
252         (WebCore::DumpRenderTree::geolocationPermissionSet):
253         * DumpRenderTree/qt/DumpRenderTreeQt.h:
254         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
255         (LayoutTestController::grantDesktopNotificationPermission):
256         * QtTestBrowser/webpage.cpp:
257         (WebPage::WebPage):
258         (WebPage::requestPermission):
259         (WebPage::cancelRequestsForPermission):
260         * QtTestBrowser/webpage.h:
261
262 2010-11-23  Simon Hausmann  <simon.hausmann@nokia.com>
263
264         Reviewed by Tor Arne Vestbø.
265
266         [Qt] Review the setUserPermission & friends API
267         https://bugs.webkit.org/show_bug.cgi?id=46810
268
269         Add a ByUser suffix to PermissionGranted/Denied. In the future
270         we can add PermissionGrantedByDefault.
271
272
273         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
274         (WebCore::WebPage::requestPermission):
275         (WebCore::WebPage::permissionSet):
276         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
277         (LayoutTestController::grantDesktopNotificationPermission):
278         * QtTestBrowser/webpage.cpp:
279         (WebPage::requestPermission):
280
281 2010-11-18  Satish Sampath  <satish@chromium.org>
282
283         Reviewed by Jeremy Orlow.
284
285         For speech input event, send an event object containing all the recognition results and metadata.
286         https://bugs.webkit.org/show_bug.cgi?id=49736
287
288         Updated DRT by renaming the mock result method on all platforms to the new signature.
289
290         * DumpRenderTree/LayoutTestController.cpp:
291         (addMockSpeechInputResultCallback):
292         (LayoutTestController::staticFunctions):
293         * DumpRenderTree/LayoutTestController.h:
294         * DumpRenderTree/chromium/LayoutTestController.cpp:
295         (LayoutTestController::LayoutTestController):
296         (LayoutTestController::addMockSpeechInputResult):
297         * DumpRenderTree/chromium/LayoutTestController.h:
298         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
299         (LayoutTestController::addMockSpeechInputResult):
300         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
301         (LayoutTestController::addMockSpeechInputResult):
302         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
303         (LayoutTestController::addMockSpeechInputResult):
304         * DumpRenderTree/qt/LayoutTestControllerQt.h:
305         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
306         (LayoutTestController::addMockSpeechInputResult):
307         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
308         (LayoutTestController::addMockSpeechInputResult):
309
310 2010-11-22  Dirk Pranke  <dpranke@chromium.org>
311
312         Reviewed by Tony Chang.
313
314         Fix some spacing issues with log messages with
315         new-run-webkit-tests --verbose.
316
317         https://bugs.webkit.org/show_bug.cgi?id=49936
318
319         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
320
321 2010-11-22  Ojan Vafai  <ojan@chromium.org>
322
323         Reviewed by Adam Barth.
324
325         update scm to work with different remote branches
326         https://bugs.webkit.org/show_bug.cgi?id=49949
327
328         This works if there are multiple svn tracking branches
329         and/or if the tracking branch is refs/remots/origin/master
330         instead of refs/remotes/trunk which is the direction we're trying
331         to head since that makes fetches faster.
332
333         * Scripts/webkitpy/common/checkout/scm.py:
334         * Scripts/webkitpy/common/checkout/scm_unittest.py:
335
336 2010-11-22  Hayato Ito  <hayato@chromium.org>
337
338         Reviewed by Shinichiro Hamaji.
339
340         Ignore reference files which will be used by reftests when collecting
341         test cases.
342         https://bugs.webkit.org/show_bug.cgi?id=49835
343
344         * Scripts/webkitpy/layout_tests/port/test_files.py:
345         * Scripts/webkitpy/layout_tests/port/test_files_unittest.py:
346
347 2010-11-22  Adam Roben  <aroben@apple.com>
348
349         Use paths relative to $WebKitVSPropsRedirectionDir to access shared .vsprops files
350
351         Apple's Windows build allows placing header files and import libraries for WebKit's
352         dependencies (CoreGraphics, CFNetwork, SQLite, etc.) outside the source tree via the
353         $WebKitLibrariesDir environment variable. This is both required for production builds and
354         convenient for Apple-internal developer builds. Apple's production builds also require that
355         WebKit's shared .vsprops files be accessed relative to $WebKitLibrariesDir. In production
356         builds, the files are copied into that directory tree by the
357         WebKitLibraries/win/tools/WinTools.make file. In Apple-internal developer builds, the
358         copying is done by
359         JavaScriptCore/JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.make.
360
361         This .vsprops copying is problematic in one very important case: when a developer updates
362         their source tree and then tries to build. Visual Studio only reads .vsprops files when a
363         project is first loaded. So, when Visual Studio is first opened after the .vsprops files are
364         updated, it reads in the old files that were already residing in $WebKitLibrariesDir. When a
365         build is started, JavaScriptCoreGenerated.make copies the new .vsprops files into
366         $WebKitLibrariesDir, but Visual Studio will not pick up the changes.  The rest of the build
367         will proceed with out-of-date .vsprops files, which will likely result in a build failure.
368
369         To fix this, we now use normal relative paths to access the .vsprops files in the source
370         tree rather than in $WebKitLibrariesDir, but prefix those paths with a new environment
371         variable, $WebKitVSPropsRedirectionDir. In developer builds, this environment variable is
372         unset, so the normal relative paths are used to read the .vsprops files out of the source
373         tree directly. In production builds, this environment variable is set to a fake directory
374         that will cause the .vsprops files in $WebKitLibrariesDir to be found when the relative path
375         is resolved.
376         
377         For example, JavaScriptCore.vcproj uses this path for FeatureDefines.vsprops:
378
379         $(WebKitVSPropsRedirectionDir)..\..\..\WebKitLibraries\win\tools\vsprops\FeatureDefines.vsprops
380
381         In developer builds, where $WebKitVSPropsRedirectionDir is unset, this will point to the
382         files in WebKitLibraries\win\tools\vsprops in the source tree. In production builds,
383         JavaScriptCore.make sets $WebKitVSPropsRedirectionDir to
384         "$(SRCROOT)\AppleInternal\tools\vsprops\OpenSource\1\2\3\", so the full path for
385         FeatureDefines.vsprops becomes:
386
387         $(SRCROOT)\AppleInternal\tools\vsprops\OpenSource\1\2\3\..\..\..\WebKitLibraries\win\tools\vsprops\FeatureDefines.vsprops
388
389         which resolves to:
390
391         $(SRCROOT)\AppleInternal\tools\vsprops\OpenSource\WebKitLibraries\win\tools\vsprops\FeatureDefines.vsprops
392
393         (We rely on the fact that Windows doesn't care whether the directories "1", "2", and "3"
394         actually exist since they are matched by an equal number of ".." path components.)
395
396         Note that Visual Studio still won't pick up changes made to .vsprops files while Visual
397         Studio is open, but that problem hasn't seemed to cause developers many headaches so far.
398
399         Fixes <http://webkit.org/b/49181> Windows build fails mysteriously when .vsprops files are
400         updated
401
402         Reviewed by Dave Hyatt.
403
404         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
405         * DumpRenderTree/win/DumpRenderTree.vcproj:
406         * DumpRenderTree/win/ImageDiff.vcproj:
407         * FindSafari/FindSafari.vcproj:
408         * MiniBrowser/MiniBrowser.vcproj:
409         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
410         * TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj:
411         * WebKitAPITest/WebKitAPITest.vcproj:
412         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
413         * WebKitTestRunner/win/InjectedBundle.vcproj:
414         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
415         * WinLauncher/WinLauncher.vcproj:
416         * record-memory-win/record-memory-win.vcproj:
417         Changed to use paths relative to $WebKitVSPropsRedirectionDir to access shared .vsprops
418         files.
419
420 2010-11-22  Adam Roben  <aroben@apple.com>
421
422         Make it possible to run tests on Windows without Visual Studio or VC++
423         Express installed
424
425         webkitdir.pm::setupCygwinEnv dies when Visual Studio and VC++ Express
426         are not installed. But this function doesn't need to be called when we
427         already have a build available and are just trying to run the tests.
428
429         Fixes <http://webkit.org/b/49932> New Windows 7 bot can't run tests
430         because Visual Studio/VC++ Express aren't installed
431
432         Reviewed by Dave Hyatt.
433
434         * Scripts/webkitdirs.pm:
435         (determineConfigurationForVisualStudio): Don't call setupCygwinEnv, as
436         it is not needed by this function. Also added a FIXME.
437
438         (usingVisualStudioExpress): Call setupCygwinEnv directly rather than
439         relying on determineConfigurationForVisualStudio doing it.
440
441 2010-11-19  Hayato Ito  <hayato@chromium.org>
442
443         Reviewed by Shinichiro Hamaji.
444
445         Simplify Driver.run_test() so that it takes only one parameter, TestInput.
446
447         https://bugs.webkit.org/show_bug.cgi?id=49812
448
449         * Scripts/webkitpy/layout_tests/driver_test.py:
450         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
451         * Scripts/webkitpy/layout_tests/port/base.py:
452         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
453         * Scripts/webkitpy/layout_tests/port/chromium.py:
454         * Scripts/webkitpy/layout_tests/port/dryrun.py:
455         * Scripts/webkitpy/layout_tests/port/test.py:
456         * Scripts/webkitpy/layout_tests/port/webkit.py:
457         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
458         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
459
460 2010-11-22  Hayato Ito  <hayato@chromium.org>
461
462         Reviewed by Tony Chang.
463
464         [NRWT] Retry a few times in reading a png image to avoid a race condition.
465
466         https://bugs.webkit.org/show_bug.cgi?id=49924
467
468         * Scripts/webkitpy/layout_tests/port/chromium.py:
469
470 2010-11-22  João Paulo Rechi Vita  <jprvita@profusion.mobi>
471
472         Reviewed by Kenneth Rohde Christiansen.
473
474         Fix EWebLauncher zoom discretization
475         https://bugs.webkit.org/show_bug.cgi?id=49810
476
477         * EWebLauncher/main.c:
478         (nearest_zoom_level_get):
479         (zoom_level_set):
480         (on_load_finished):
481         (on_key_down):
482
483 2010-11-21  Sam Weinig  <sam@webkit.org>
484
485         Rubber-stamped by Dan Bernstein.
486
487         Fix ~300 WebKit2 tests by enabling developer extras preference.
488
489         * WebKitTestRunner/TestController.cpp:
490         (WTR::TestController::resetStateToConsistentValues):
491
492 2010-11-21  Sam Weinig  <sam@webkit.org>
493
494         Reviewed by Dan Bernstein.
495
496         Textareas should be resizable by default
497         https://bugs.webkit.org/show_bug.cgi?id=49892
498
499         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
500         * TestWebKitAPI/Tests/WebKit2/WKPreferences.cpp: Added.
501         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
502         Add preferences test.
503
504 2010-11-19  Mihai Parparita  <mihaip@chromium.org>
505
506         Reviewed by Tony Chang.
507
508         Rebaseline server: checksum-only differences and prefetching
509         https://bugs.webkit.org/show_bug.cgi?id=49841
510         
511         Add support for image test results where the only difference is in the
512         checksum.
513         
514         Add prefetching of the image results from the next test.
515         
516         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
517         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
518         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
519
520 2010-11-19  Dirk Pranke  <dpranke@chromium.org>
521
522         Reviewed by Tony Chang.
523
524         nrwt multiprocessing - add 'worker number' concept, move stuff to worker thread
525
526         Add the 'worker number' and 'worker name' concepts to the 
527         TestShellThread objects, and move test_types and test_args from
528         the TestRunner to the TestShellThread.
529
530         https://bugs.webkit.org/show_bug.cgi?id=49768
531
532         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
533         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
534
535 2010-11-19  Dirk Pranke  <dpranke@chromium.org>
536
537         Reviewed by Ojan Vafai.
538
539         Do some minor cleanup and bug fixing.
540
541         https://bugs.webkit.org/show_bug.cgi?id=49777
542
543         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
544         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
545
546 2010-11-19  Dirk Pranke  <dpranke@chromium.org>
547
548         Reviewed by Ojan Vafai.
549
550         nrwt - config.build_directory() is busted
551
552         Fixes a typo that was causing us to usually return the top level
553         directory WebKitBuild instead of WebKitBuild/{Debug,Release}. The
554         bug was hidden by test stubs that were too simplistic :(.
555
556         https://bugs.webkit.org/show_bug.cgi?id=49815
557
558         * Scripts/webkitpy/common/system/executive_mock.py:
559         * Scripts/webkitpy/layout_tests/port/config.py:
560         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
561
562 2010-11-19  Steve Falkenburg  <sfalken@apple.com>
563
564         Reviewed by Adam Roben.
565
566         Add Debug_Cairo_CFLite and Release_Cairo_CFLite configurations for all vcproj files
567         https://bugs.webkit.org/show_bug.cgi?id=49819
568
569         * DumpRenderTree/DumpRenderTree.sln:
570         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
571         * DumpRenderTree/win/DumpRenderTree.vcproj:
572         * DumpRenderTree/win/ImageDiff.vcproj:
573         * FindSafari/FindSafari.vcproj:
574         * MiniBrowser/MiniBrowser.vcproj:
575         * TestWebKitAPI/win/TestWebKitAPI.sln:
576         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
577         * TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj:
578         * WebKitAPITest/WebKitAPITest.vcproj:
579         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
580         * WebKitTestRunner/win/InjectedBundle.vcproj:
581         * WebKitTestRunner/win/InjectedBundleGenerated.vcproj:
582         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
583         * WinLauncher/WinLauncher.vcproj:
584         * record-memory-win/record-memory-win.vcproj:
585
586 2010-11-19  Adam Roben  <aroben@apple.com>
587
588         Add some more Windows slaves
589
590         This is probably more than we will be using in the immediate future,
591         but having some extra slave names makes experimenting with different
592         configurations easier.
593
594         Reviewed by Jon Honeycutt.
595
596         * BuildSlaveSupport/build.webkit.org-config/config.json: Added 5 more
597         Windows slaves and distributed them among the three different Test
598         builders.
599
600 2010-11-19  Kinuko Yasuda  <kinuko@chromium.org>
601
602         Reviewed by Ojan Vafai.
603
604         [Chromium] REGRESSION: Record actual test results in the JSON result summary for non-layout tests
605         https://bugs.webkit.org/show_bug.cgi?id=49702
606
607         Re-landing r72357 with a test fix.
608
609         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
610         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
611         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
612
613 2010-11-19  Avi Drissman  <avi@google.com>
614
615         Reviewed by Antonio Gomes.
616
617         Complete support for Unix editing mode
618         https://bugs.webkit.org/show_bug.cgi?id=49757
619
620         * DumpRenderTree/LayoutTestController.cpp:
621         (setEditingBehaviorCallback):
622         * DumpRenderTree/chromium/LayoutTestController.cpp:
623         (LayoutTestController::setEditingBehavior):
624
625 2010-11-19  Adam Roben  <aroben@apple.com>
626
627         Fix run-javascriptcore-tests when there's a space in the path to
628         testapi
629
630         Reviewed by Sam Weinig.
631
632         * Scripts/run-javascriptcore-tests: Use an indirect object when calling
633         system() to force it to interpret spaces in the path correctly.
634
635 2010-11-19  Adam Roben  <aroben@apple.com>
636
637         Update scripts for .exe name changes after r72327
638
639         As of r72327, the "_debug" suffix is only used in Debug_All builds.
640
641         Reviewed by Sam Weinig.
642
643         * Scripts/run-javascriptcore-tests:
644         (testapiPath):
645
646         * Scripts/sunspider-compare-results:
647         (pathToBuiltJSC):
648
649         * Scripts/webkitdirs.pm:
650         (jscPath):
651         (safariPath):
652
653 2010-11-19  Steve Falkenburg  <sfalken@apple.com>
654
655         Reviewed by Darin Adler.
656
657         Normalize Cairo/CFLite project/solution configuration names
658         https://bugs.webkit.org/show_bug.cgi?id=49818
659
660         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
661         * DumpRenderTree/win/DumpRenderTree.vcproj:
662         * DumpRenderTree/win/ImageDiff.vcproj:
663         * MiniBrowser/MiniBrowser.vcproj:
664         * TestWebKitAPI/win/TestWebKitAPI.sln:
665         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
666         * WebKitAPITest/WebKitAPITest.vcproj:
667         * WebKitTestRunner/win/InjectedBundle.vcproj:
668         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
669         * WinLauncher/WinLauncher.vcproj:
670         * Scripts/webkitdirs.pm: Updated cairo build configs.
671
672 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
673
674         Reviewed by Adam Roben.
675
676         Windows vcproj configuration names should be normalized across projects
677         https://bugs.webkit.org/show_bug.cgi?id=49776
678
679         * DumpRenderTree/DumpRenderTree.sln:
680         * FindSafari/FindSafari.vcproj:
681         * TestWebKitAPI/win/TestWebKitAPI.sln:
682         * TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj:
683         * TestWebKitAPI/win/TestWebKitAPIGeneratedCommon.vsprops: Added.
684         * WebKitAPITest/WebKitAPITest.vcproj:
685         * WebKitAPITest/WebKitAPITestCommon.vsprops:
686         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
687         * WebKitTestRunner/WebKitTestRunner.sln:
688         * WebKitTestRunner/win/InjectedBundleGenerated.vcproj:
689         * WebKitTestRunner/win/InjectedBundleGeneratedCommon.vsprops: Added.
690         * record-memory-win/record-memory-win.vcproj:
691
692 2010-11-18  Gavin Barraclough  <barraclough@apple.com>
693
694         Build fix - as per DRT-mac, always enabled developer extras,
695         in win/qt/gtk DRT.
696
697         This is necessary to enable rich exception messages to be generated
698         following https://bugs.webkit.org/show_bug.cgi?id=49708.
699
700         * DumpRenderTree/gtk/DumpRenderTree.cpp:
701         (shouldEnableDeveloperExtras):
702         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
703         (WebCore::shouldEnableDeveloperExtras):
704         * DumpRenderTree/win/DumpRenderTree.cpp:
705         (shouldEnableDeveloperExtras):
706
707 2010-11-18  Sheriff Bot  <webkit.review.bot@gmail.com>
708
709         Unreviewed, rolling out r72357.
710         http://trac.webkit.org/changeset/72357
711         https://bugs.webkit.org/show_bug.cgi?id=49784
712
713         It broke test-webkitpy tests since the patch didn't have the
714         corresponding test update (Requested by kinuko on #webkit).
715
716         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
717         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
718
719 2010-11-18  Mihai Parparita  <mihaip@chromium.org>
720
721         Reviewed by Tony Chang.
722
723         Rebaseline server: add rebaseline queue
724         https://bugs.webkit.org/show_bug.cgi?id=49763
725         
726         Add a UI queue (a multi-item <select>) where tests to be baselined
727         can be enqueued (also supports basic removal). A queue is used instead
728         of immediately doing rebaselines since SCM operations can take a while.
729         It's better to go through lots of tests quickly to mark them as
730         needing rebaselining and then batch the SCM operations.
731         
732         Also adds a simple log where results can be displayed.
733
734         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
735         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
736         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
737         * Scripts/webkitpy/tool/commands/data/rebaselineserver/queue.js: Added.
738         * Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js:
739         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
740
741 2010-11-18  Gavin Barraclough  <barraclough@apple.com>
742
743         Reviewed by Oliver Hunt.
744
745         Bug 49708 - Stop recompiling functions to regenerate exception info.
746
747         Instead only hold info as necessary – keep divot info is the inspector
748         is enabled, line number info is debugging or profiling, and handler
749         info for functions with try/catch.
750
751         * DumpRenderTree/mac/DumpRenderTree.mm:
752         (shouldEnableDeveloperExtras):
753             - always enable the developer tools from DRT, to ensure we
754               produce rich error messages on JavaScript exceptions.
755
756 2010-11-18  Kinuko Yasuda  <kinuko@chromium.org>
757
758         Reviewed by Ojan Vafai.
759
760         [Chromium] REGRESSION: Record actual test results in the JSON result summary for non-layout tests
761         https://bugs.webkit.org/show_bug.cgi?id=49702
762
763         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
764         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
765
766 2010-11-18  Dirk Pranke  <dpranke@chromium.org>
767
768         Reviewed by Tony Chang.
769
770         new-run-webkit-tests: create first part of 'message_broker' class for multiprocessing fixes
771
772         Create the first version of the 'message_broker' package. This
773         class will encapsulate all of the threading/multiprocessing and
774         message-sending details for the communication between the
775         'manager' object and the 'worker' objects. For the moment, it
776         just holds some routines and tests for logging thread stacks.
777
778         There should be no functional changes in this patch, just moving stuff
779         around.
780
781         https://bugs.webkit.org/show_bug.cgi?id=49707
782
783         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
784         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread_unittest.py: Removed.
785         * Scripts/webkitpy/layout_tests/layout_package/message_broker.py: Added.
786         * Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py: Added.
787         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
788         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
789
790 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
791
792         Reviewed by Adam Roben.
793
794         Remove leftover Windows Debug_Internal configurations
795         https://bugs.webkit.org/show_bug.cgi?id=49758
796
797         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
798         * DumpRenderTree/win/DumpRenderTree.vcproj:
799         * DumpRenderTree/win/ImageDiff.vcproj:
800         * MiniBrowser/MiniBrowser.vcproj:
801         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
802         * WebKitAPITest/WebKitAPITest.vcproj:
803         * WebKitTestRunner/win/InjectedBundle.vcproj:
804         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
805         * WinLauncher/WinLauncher.vcproj:
806
807 2010-11-18  Adam Roben  <aroben@apple.com>
808
809         Add three new Windows XP build slaves
810
811         Reviewed by Steve Falkenburg.
812
813         * BuildSlaveSupport/build.webkit.org-config/config.json: Added three
814         new slaves, and added them to the Windows XP Debug (Tests) builder.
815
816 2010-11-18  Daniel Bates  <dbates@rim.com>
817
818         Reviewed by Adam Roben.
819
820         update-webkit-support-libs should fall back to existing
821         WebKitSupportLibrary version if there is no internet connectivity
822         https://bugs.webkit.org/show_bug.cgi?id=49503
823
824         Fall back to existing support libraries (if present) when there is
825         no internet connection.
826
827         Currently, update-webkit-support-libs dies with an "out-of-date"
828         error when there is no internet connection because it cannot
829         retrieve versioning information from developer.apple.com. Because
830         update-webkit-support-libs fails, build-webkit fails. Instead,
831         if there is no internet connection and the support libraries are
832         present then we should warn the user and exit() with success so
833         that build-webkit can work without an internet connection.
834
835         * Scripts/update-webkit-support-libs:
836
837 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
838
839         Reviewed by Adam Roben.
840
841         Debug_Internal Windows configuration is unnecessary, should be removed
842         https://bugs.webkit.org/show_bug.cgi?id=49753
843
844         * DumpRenderTree/win/DumpRenderTree.cpp:
845         * MiniBrowser/win/stdafx.h:
846         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp:
847         * WebKitTestRunner/win/TestControllerWin.cpp:
848
849 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
850
851         Reviewed by Adam Roben.
852
853         Clean up vcproj errors
854         https://bugs.webkit.org/show_bug.cgi?id=49741
855
856         * MiniBrowser/MiniBrowser.vcproj:
857         * WebKitAPITest/WebKitAPITest.vcproj:
858         * WebKitTestRunner/win/InjectedBundle.vcproj:
859         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
860
861 2010-11-18  Hayato Ito  <hayato@chromium.org>
862
863         Reviewed by Shinichiro Hamaji.
864
865         [NRWT] Fix a break of '--new-baseline' feature in pixel_test
866         which is caused by r72249.
867
868         https://bugs.webkit.org/show_bug.cgi?id=49751
869
870         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
871
872 2010-11-18  Adam Roben  <aroben@apple.com>
873
874         Implement layoutTestController.waitForPolicyDelegate on Windows
875
876         Fixes <http://webkit.org/b/25038> <rdar://problem/6790213>.
877
878         Reviewed by Simon Fraser.
879
880         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
881         (LayoutTestController::waitForPolicyDelegate): Implemented by porting
882         code from LayoutTestControllerMac.mm.
883
884 2010-11-18  Hayato Ito  <hayato@chromium.org>
885
886         Reviewed by Shinichiro Hamaji.
887
888         [NRWT] Fix a break of '--new-baseline' feature which is caused by r72249.
889
890         https://bugs.webkit.org/show_bug.cgi?id=49740
891
892         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
893
894 2010-11-18  Tony Chang  <tony@chromium.org>
895
896         Reviewed by Adam Barth.
897
898         run platform/chromium/plugins/refcount-leaks.html on all platforms
899         https://bugs.webkit.org/show_bug.cgi?id=49485
900
901         Add PluginObject.testObjectCount which returns the number of allocated
902         TestObjects. Add PluginObject.testCreateTestObject which allocates
903         and returns a TestObject.
904
905         Add TestObject.refCount which returns the number of refs on the
906         TestObject.
907
908         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
909         (pluginGetProperty):
910         (pluginInvoke):
911         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
912         (getTestObjectCount):
913         (testAllocate):
914         (testDeallocate):
915         (testGetProperty):
916         (testScriptObjectInvoke): Release a ref to a plugin object that was
917             previously leaked.
918         * DumpRenderTree/TestNetscapePlugIn/TestObject.h:
919
920 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
921
922         Reviewed by Adam Roben.
923
924         WebKitTools projects (WebKitTestRunner, record-memory-win) should use vsprops for common build settings
925         https://bugs.webkit.org/show_bug.cgi?id=49711
926
927         * WebKitTestRunner/Configurations/WebKitTestRunnerCFLite.vsprops: Added.
928         * WebKitTestRunner/Configurations/WebKitTestRunnerCommon.vsprops: Added.
929         * WebKitTestRunner/Configurations/WebKitTestRunnerCoreFoundation.vsprops: Added.
930         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
931         * record-memory-win/record-memory-win-common.vsprops: Added.
932         * record-memory-win/record-memory-win.vcproj:
933
934 2010-11-18  Andras Becsi  <abecsi@inf.u-szeged.hu>
935
936         Reviewed by Csaba Osztrogonác.
937
938         [Qt][WK2] Only add user agent strings to the list which aren't listed yet.
939
940         * MiniBrowser/qt/BrowserWindow.cpp:
941         (BrowserWindow::updateUserAgentList):
942
943 2010-11-17  Adam Roben  <aroben@apple.com>
944
945         Don't trigger Windows builds when chromium-win test results change
946
947         Reviewed by Mark Rowe.
948
949         * Scripts/webkitpy/common/config/build.py:
950         (_should_file_trigger_build): Changed the regular expression used to
951         search for directory names to only consider full directory names,
952         rather than matching directory names where the string we care about is
953         a suffix of the name.
954
955         * Scripts/webkitpy/common/config/build_unittest.py:
956         (ShouldBuildTest): Added a test.
957
958 2010-11-17  Hayato Ito  <hayato@chromium.org>
959
960         Reviewed by Shinichiro Hamaji.
961
962         Remove a uri member from TestInput class.
963
964         https://bugs.webkit.org/show_bug.cgi?id=49691
965
966         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
967         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
968
969 2010-11-17  Adam Roben  <aroben@apple.com>
970
971         Move FindSafari's settings into a vsprops file
972
973         Fixes <http://webkit.org/b/49699> FindSafari should use a vsprops file
974
975         Reviewed by Steve Falkenburg.
976
977         * FindSafari/FindSafari.vcproj:
978         * FindSafari/FindSafariCommon.vsprops: Added.
979
980 2010-11-17  Adam Roben  <aroben@apple.com>
981
982         Use a minimum font size of 0 in DRT on Windows
983
984         This is the Windows equivalent of r72164.
985
986         Rubber-stamped by Dave Hyatt.
987
988         * DumpRenderTree/win/DumpRenderTree.cpp:
989         (resetDefaultsToConsistentValues):
990
991 2010-11-17  Dirk Pranke  <dpranke@chromium.org>
992
993         Reviewed by Andreas Kling.
994
995         webkit-patch pretty-diff shouldn't need to be run from a checkout root in git
996         https://bugs.webkit.org/show_bug.cgi?id=49639
997
998         * Scripts/webkitpy/common/checkout/scm.py:
999
1000 2010-11-17  Mihai Parparita  <mihaip@chromium.org>
1001
1002         Reviewed by Tony Chang.
1003
1004         Rebaseline server: add loupe for image diffs
1005         https://bugs.webkit.org/show_bug.cgi?id=49692
1006         
1007         Add a loupe (magnifiying glass) for inspecting image diffs. Shows an
1008         enlarged area of the expected, actual and diff images side by side.
1009
1010         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
1011         * Scripts/webkitpy/tool/commands/data/rebaselineserver/loupe.js: Added.
1012         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
1013         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
1014
1015 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
1016
1017         Reviewed by Adam Roben.
1018
1019         WebKitLauncherWin should use vsprops for shared build settings
1020         https://bugs.webkit.org/show_bug.cgi?id=49696
1021
1022         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
1023         * WebKitLauncherWin/WebKitLauncherWinCommon.vsprops: Added.
1024
1025 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
1026
1027         Reviewed by Adam Roben.
1028
1029         WinLauncher should use vsprops for shared build settings
1030         https://bugs.webkit.org/show_bug.cgi?id=49695
1031
1032         * WinLauncher/WinLauncher.cpp:
1033         (_tWinMain):
1034         (PrintView):
1035         (WndProc):
1036         (MyEditProc):
1037         * WinLauncher/WinLauncher.vcproj:
1038         * WinLauncher/WinLauncherCommon.vsprops: Added.
1039
1040 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
1041
1042         Delete unused file unintentionally added in last change.
1043
1044         * DumpRenderTree/win/DumpRenderTreeCURL.vsprops: Removed.
1045
1046 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
1047
1048         Reviewed by Adam Roben.
1049
1050         DumpRenderTree on Windows should use a vsprops file for shared project settings
1051         https://bugs.webkit.org/show_bug.cgi?id=49690
1052
1053         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1054         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePluginCommon.vsprops: Added.
1055         * DumpRenderTree/win/DumpRenderTree.vcproj:
1056         * DumpRenderTree/win/DumpRenderTreeApple.vsprops: Added.
1057         * DumpRenderTree/win/DumpRenderTreeCFLite.vsprops: Added.
1058         * DumpRenderTree/win/DumpRenderTreeCURL.vsprops: Added.
1059         * DumpRenderTree/win/DumpRenderTreeCairo.vsprops: Added.
1060         * DumpRenderTree/win/DumpRenderTreeCommon.vsprops: Added.
1061         * DumpRenderTree/win/ImageDiff.vcproj:
1062         * DumpRenderTree/win/ImageDiffCommon.vsprops: Added.
1063
1064 2010-11-17  Hayato Ito  <hayato@chromium.org>
1065
1066         Refactor TestTypeBase.compare_output().
1067
1068         Introduce a TestOutput class and update compare_output() of each test
1069         types so that they can take both actual and expected TestOutput objects.
1070
1071         https://bugs.webkit.org/show_bug.cgi?id=49431
1072
1073         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1074         * Scripts/webkitpy/layout_tests/layout_package/test_output.py: Added.
1075         * Scripts/webkitpy/layout_tests/port/base.py:
1076         * Scripts/webkitpy/layout_tests/port/chromium.py:
1077         * Scripts/webkitpy/layout_tests/port/dryrun.py:
1078         * Scripts/webkitpy/layout_tests/port/test.py:
1079         * Scripts/webkitpy/layout_tests/port/webkit.py:
1080         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1081         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
1082         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
1083         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
1084
1085 2010-11-17  Adam Roben  <aroben@apple.com>
1086
1087         Make each Windows Test builder use the same OS for all its slaves
1088
1089         Currently each WebKit1 Windows Test builder has one Windows XP slave
1090         and one Windows 7 slave. Having different OSes in a single builder
1091         meant that the builders would alternate between red and green when
1092         there were OS-specific failures.
1093
1094         Fixes <http://webkit.org/b/49688> Windows Test builders alternate
1095         between red and green because of multiple OS versions
1096
1097         Reviewed by Mark Rowe.
1098
1099         * BuildSlaveSupport/build.webkit.org-config/config.json: Renamed
1100         "Windows Release (Tests)" to "Windows 7 Release (Tests)", "Windows
1101         Debug (Tests)" to "Windows XP Debug (Tests)", and "Windows Release
1102         (WebKit2 Tests)" to "Windows 7 Release (WebKit2 Tests)". Updated the
1103         builders for the two WebKit1 builders so that they would all use the
1104         appropriate OS.
1105
1106 2010-11-17  James Robinson  <jamesr@chromium.org>
1107
1108         Reviewed by Dan Bernstein.
1109
1110         [chromium] Set minimum font size to 0 for DRT
1111         https://bugs.webkit.org/show_bug.cgi?id=49677
1112
1113         Followup for http://trac.webkit.org/changeset/72141.
1114
1115         Test: http/tests/misc/acid3.html
1116
1117         * DumpRenderTree/chromium/WebPreferences.cpp:
1118         (WebPreferences::reset):
1119
1120 2010-11-17  Andras Becsi  <abecsi@webkit.org>
1121
1122         Unreviewed build fix.
1123
1124         [Qt] Add missing file I forgot in r72220.
1125
1126         * MiniBrowser/MiniBrowser.qrc: Added.
1127
1128 2010-11-17  Andras Becsi  <abecsi@webkit.org>
1129
1130         Reviewed by Andreas Kling.
1131
1132         [Qt][WK2] Make it possible to set custom user agent strings in MiniBrowser.
1133         https://bugs.webkit.org/show_bug.cgi?id=49627
1134
1135         Add a user agent dialog and a resource file for useragentlist.txt to MiniBrowser.
1136
1137         * MiniBrowser/MiniBrowser.qrc: Added.
1138         Add it here to prevent qmake from detecting it
1139         since it needs to be copied to the build directory.
1140         * MiniBrowser/qt/BrowserWindow.cpp:
1141         (BrowserWindow::BrowserWindow):
1142         (BrowserWindow::updateUserAgentList):
1143         (BrowserWindow::showUserAgentDialog):
1144         * MiniBrowser/qt/BrowserWindow.h:
1145         * MiniBrowser/qt/MiniBrowser.pro:
1146
1147 2010-11-17  Adam Roben  <aroben@apple.com>
1148
1149         Make changes to the Mac WebKit2 Skipped file trigger Windows builds
1150
1151         We use that file on Windows!
1152
1153         Fixes <http://webkit.org/b/49643> Changes to
1154         LayoutTests/platform/mac-wk2 should trigger a Windows build, but don't
1155
1156         Reviewed by Steve Falkenburg.
1157
1158         * Scripts/webkitpy/common/config/build.py:
1159         (_should_file_trigger_build): Added the mac-wk2 directory and make it
1160         trigger builds on SnowLeopard and Windows.
1161
1162         * Scripts/webkitpy/common/config/build_unittest.py:
1163         (ShouldBuildTest.test_should_build): Added a test.
1164
1165 2010-11-17  Satish Sampath  <satish@chromium.org>
1166
1167         Reviewed by Jeremy Orlow.
1168
1169         Clear the speech input mock explicitly before each test.
1170         https://bugs.webkit.org/show_bug.cgi?id=49660
1171
1172         * DumpRenderTree/chromium/LayoutTestController.cpp:
1173         (LayoutTestController::setMockSpeechInputResult):
1174         * DumpRenderTree/chromium/LayoutTestController.h:
1175         * DumpRenderTree/chromium/WebViewHost.cpp:
1176         (WebViewHost::speechInputController):
1177         (WebViewHost::reset): Invoke speech input mock's clearResults method.
1178         * DumpRenderTree/chromium/WebViewHost.h:
1179         (WebViewHost::speechInputControllerMock): Moved mock from LayoutTestController.
1180
1181 2010-11-17  Alexander Pavlov  <apavlov@chromium.org>
1182
1183         Unreviewed, specify IRC nickname in committers.py.
1184
1185         * Scripts/webkitpy/common/config/committers.py:
1186
1187 2010-11-17  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1188
1189         Reviewed by Tony Chang.
1190
1191         [NRWT] Make http locking similar to perl implementation
1192         https://bugs.webkit.org/show_bug.cgi?id=49187
1193
1194         * Scripts/webkitpy/common/system/file_lock.py: Added.
1195         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
1196         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1197
1198 2010-11-17  MORITA Hajime  <morrita@google.com>
1199
1200         Reviewed by Kent Tamura.
1201
1202         [Chromium][DRT] EventSender.contextClick() should aware spellchecking
1203         https://bugs.webkit.org/show_bug.cgi?id=49366
1204
1205         - EvenSender: Checked WebContextMenuData.misspelledWord and added extra context menu entries
1206           according to the spellchecker suggestion
1207         - MockSpellCheck: add fillSuggestionList to provide fake suggestions.
1208
1209         test_expectations.txt will be changed after this change is ported to test_shell.
1210         
1211         * DumpRenderTree/chromium/EventSender.cpp:
1212         (makeMenuItemStringsFor):
1213         (EventSender::contextClick):
1214         * DumpRenderTree/chromium/MockSpellCheck.cpp:
1215         (MockSpellCheck::fillSuggestionList):
1216         (MockSpellCheck::initializeIfNeeded):
1217         * DumpRenderTree/chromium/MockSpellCheck.h:
1218         * DumpRenderTree/chromium/WebViewHost.cpp:
1219         (WebViewHost::mockSpellCheck):
1220         * DumpRenderTree/chromium/WebViewHost.h:
1221
1222 2010-11-16  Dave Hyatt  <hyatt@apple.com>
1223
1224         Reviewed by Sam Weinig.
1225
1226         Make sure the pref for minimum font size is just explicitly set to 0 every time,
1227         since the bots have 1 stuck in their plist otherwise.
1228
1229         * DumpRenderTree/mac/DumpRenderTree.mm:
1230         (resetDefaultsToConsistentValues):
1231
1232 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
1233
1234         Reviewed by Tony Chang.
1235
1236         Rebaseline server: compute diffs client-side
1237         https://bugs.webkit.org/show_bug.cgi?id=49640
1238         
1239         The image diff output from the DRT is pretty bad on some ports (at
1240         least the Mac one), so it's better to compute diffs on the client by
1241         using <canvas>.
1242
1243         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
1244         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
1245         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
1246
1247 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
1248
1249         Reviewed by Tony Chang.
1250
1251         Rebaseline server: display test results
1252         https://bugs.webkit.org/show_bug.cgi?id=49626
1253         
1254         Adds basic result display to the rebaseline server. On the Python side
1255         this involves:
1256         - Parsing the unexpected_results.json into a dictionary.
1257         - Serving it as JSON under /results.json.
1258         (the JSON -> dict -> JSON transform isn't strictly necessary right now,
1259         but I'll need to have access to the parsed results on the Python side
1260         for follow-up changes).
1261         
1262         On the web UI side this adds:
1263         - Markup for display image and text results (expected, actual, diff),
1264           and JS for populating it.
1265         - Markup for breaking down test results by failure type and directory,
1266           and JS for populating it.
1267
1268         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
1269         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
1270         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
1271         * Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js: Added.
1272         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
1273
1274 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
1275
1276         Reviewed by Ojan Vafai.
1277
1278         See Chromium issue http://codereview.chromium.org/5133001/ - we
1279         are modifying the build bots to pass in the builder name with
1280         the "GPU" string appended instead of appending it in the code.
1281
1282         https://bugs.webkit.org/show_bug.cgi?id=49636
1283
1284         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
1285
1286 2010-11-16  Steve Falkenburg  <sfalken@apple.com>
1287
1288         Reviewed by Adam Roben.
1289
1290         Disable LTCG for Windows Release builds. Add new Release_LTCG configuration.
1291         https://bugs.webkit.org/show_bug.cgi?id=49632
1292
1293         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1294         * DumpRenderTree/win/DumpRenderTree.vcproj:
1295         * DumpRenderTree/win/ImageDiff.vcproj:
1296         * MiniBrowser/MiniBrowser.vcproj:
1297         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
1298         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
1299         * WebKitTestRunner/win/InjectedBundle.vcproj:
1300         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
1301         * WinLauncher/WinLauncher.vcproj:
1302         * record-memory-win/record-memory-win.vcproj:
1303
1304 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
1305
1306         Reviewed by Ojan Vafai.
1307
1308         new-run-webkit-tests: rename TestInfo to TestInput, move image hash to work thread
1309
1310         Rename the TestInfo class to TestInput to be clearer about its
1311         function, and move the checksum-reading code into dump_render_tree_thread
1312         to avoid cross-thread access.
1313
1314         https://bugs.webkit.org/show_bug.cgi?id=49573
1315
1316         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1317         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1318         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1319
1320 2010-11-16  Dave Hyatt  <hyatt@apple.com>
1321
1322         Fix bustage. Remove the minimum font size pref setting in DumpRenderTree.
1323
1324         * DumpRenderTree/mac/DumpRenderTree.mm:
1325         (resetDefaultsToConsistentValues):
1326
1327 2010-11-16  Stephanie Lewis  <slewis@apple.com>
1328
1329         Reviewed by Geoff Garen.
1330
1331         <rdar://problem/8624267> Leak creating offscreen webview running fast/dom tests
1332
1333         Use shared DeviceOrientationProviderMock. The old allocation method created a new WebDeviceOrientationProviderMock 
1334         for every WebView, and the WebKit API doesn't support that ownership model.
1335
1336         I also fixed a leak in setMockDeviceOrientation by adding a missing release.
1337
1338         * DumpRenderTree/mac/DumpRenderTree.mm:
1339         (createWebViewAndOffscreenWindow):
1340         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1341         (LayoutTestController::setMockDeviceOrientation):
1342
1343 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
1344
1345         Reviewed by Ojan Vafai.
1346
1347         new-run-webkit-tests: log the process id in --verbose mode
1348
1349         As part of the switch to multi-process mode, we should log the
1350         process id when running w/ --verbose.
1351
1352         https://bugs.webkit.org/show_bug.cgi?id=49571
1353
1354         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
1355
1356 2010-11-16  Adam Roben  <aroben@apple.com>
1357
1358         Only trigger Windows builds when files that we actually use on Windows
1359         are changed
1360
1361         A new Scheduler subclass, PlatformSpecificScheduler, has been added.
1362         It uses the new webkitpy.common.config.build module to determine
1363         whether a particular change should trigger a build on a particular
1364         platform. The Windows builders have been switched to use a
1365         PlatformSpecificScheduler.
1366
1367         The logic to determine whether or not a particular change should
1368         trigger a build on a given platform has only been implemented/tested
1369         for Windows. I tried to make it easy to add more platforms in the
1370         future, but I don't have enough familiarity with all platforms to be
1371         able to implement it for them.
1372
1373         Fixes <http://webkit.org/b/49407> Windows builders kick off builds for
1374         lots irrelevant changes (e.g., rebaselining Chromium test results)
1375
1376         Reviewed by Eric Seidel.
1377
1378         * BuildSlaveSupport/build.webkit.org-config/config.json: Use a
1379         PlatformSpecificScheduler for the Windows builders.
1380
1381         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1382         (PlatformSpecificScheduler.__init__): Added. Stores our platform, sets
1383         up a ChangeFilter that filters to our branch and will call through to
1384         our filter method, and calls up to our base class.
1385         (PlatformSpecificScheduler.filter): Calls through to
1386         build.should_build to find out whether we should trigger a build for
1387         this change.
1388
1389         * Scripts/webkitpy/common/net/build.py: Added.
1390         (_should_file_trigger_build): Uses a set of directories and regexp
1391         patterns to determine whether the given file should trigger a build on
1392         the given platform. As mentioned earlier, this has only been
1393         implemented for Windows, though I did try to make some guesses about
1394         other platforms.
1395         (should_build): Returns true if any of the files should trigger a
1396         build on the given platform.
1397
1398         * Scripts/webkitpy/common/net/build_unittest.py: Added.
1399         (ShouldBuildTest.test_should_build): Does some basic testing to make
1400         sure we're triggering builds for the right files. It only tests
1401         Windows for now, though I tried to make some guesses about other
1402         platforms.
1403
1404 2010-11-16  Simon Hausmann  <simon.hausmann@nokia.com>
1405
1406         Reviewed by Andreas Kling.
1407
1408         [Qt] Remove synchronous QWebPage::checkPermissions signal
1409         https://bugs.webkit.org/show_bug.cgi?id=46810
1410
1411         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1412         (WebCore::WebPage::WebPage):
1413         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1414         (LayoutTestController::grantDesktopNotificationPermission): When granting
1415         permission, grant it directly on the QWebPage/Frame, that will remember it.
1416         * QtTestBrowser/webpage.cpp:
1417         (WebPage::WebPage):
1418         * QtTestBrowser/webpage.h:
1419
1420
1421 2010-11-16  Andras Becsi  <abecsi@inf.u-szeged.hu>
1422
1423         Reviewed by Csaba Osztrogonác.
1424
1425         [Qt] Add isPrinting and setPrinting methods to DRT's LayoutTestController.
1426
1427         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1428         (LayoutTestController::reset):
1429         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1430         (LayoutTestController::isPrinting):
1431         (LayoutTestController::setPrinting):
1432
1433 2010-11-16  Balazs Kelemen  <kbalazs@webkit.org>
1434
1435         Reviewed by Andreas Kling.
1436
1437         [Qt][WK2] Add a way to turn on tiled backing store in MiniBrowser
1438         https://bugs.webkit.org/show_bug.cgi?id=49587
1439
1440         * MiniBrowser/qt/BrowserView.cpp:
1441         (BrowserView::BrowserView): Added BackingStoreType
1442         argument to ctor. Propagating it to the QGraphicsWKView.
1443         * MiniBrowser/qt/BrowserView.h:
1444         * MiniBrowser/qt/BrowserWindow.cpp:
1445         (BrowserWindow::BrowserWindow):
1446         (BrowserWindow::newWindow): Create the new BrowserWindow
1447         with the same BackingStoreType that this instance has.
1448         * MiniBrowser/qt/BrowserWindow.h: Added BackingStoreType member.
1449         * MiniBrowser/qt/main.cpp:
1450         (main): Use tiled backing store if got -tiled command line argument.
1451         Avoid redundant copy of command line args.
1452
1453 2010-11-16  John Knottenbelt  <jknotten@chromium.org>
1454
1455         Reviewed by Jeremy Orlow.
1456
1457         Move DeviceOrientationClientMock from LayoutTestController to WebViewHost.
1458         https://bugs.webkit.org/show_bug.cgi?id=48506
1459
1460         This ensures that there is one mock per page WebView / WebViewHost. This is a
1461         design constraint of the DeviceOrientationClientMock because the m_controller
1462         field can only store a pointer to a single instance of the corresponding
1463         DeviceOrientationController.
1464
1465         Test: fast/dom/DeviceOrientation/no-page-cache.html
1466
1467         * DumpRenderTree/chromium/LayoutTestController.cpp:
1468         (LayoutTestController::setMockDeviceOrientation):
1469         (LayoutTestController::deviceOrientationClientMock):
1470         * DumpRenderTree/chromium/LayoutTestController.h:
1471         * DumpRenderTree/chromium/WebViewHost.cpp:
1472         (WebViewHost::deviceOrientationClientMock):
1473         (WebViewHost::deviceOrientationClient):
1474         (WebViewHost::reset):
1475         * DumpRenderTree/chromium/WebViewHost.h:
1476
1477 2010-11-15  Andras Becsi  <abecsi@webkit.org>
1478
1479         Reviewed by Andreas Kling.
1480
1481         [Qt][WK2] Avoid polling in WebKitTestRunner.
1482         https://bugs.webkit.org/show_bug.cgi?id=49542
1483
1484         Make the WK2 testing session about 10-15% faster and decrease random flakiness
1485         resulting from timing skew by removing the millisecond polling from TestControllerQt.
1486
1487         * WebKitTestRunner/TestController.cpp:
1488         (WTR::TestController::didFinishLoadForFrame):
1489         * WebKitTestRunner/TestController.h:
1490         * WebKitTestRunner/TestInvocation.cpp:
1491         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
1492         * WebKitTestRunner/mac/TestControllerMac.mm:
1493         (WTR::TestController::notifyDone): Add method stub.
1494         * WebKitTestRunner/qt/TestControllerQt.cpp:
1495         (WTR::TestControllerRunLoop::instance):
1496         (WTR::TestControllerRunLoop::start):
1497         (WTR::TestControllerRunLoop::stop):
1498         (WTR::TestControllerRunLoop::TestControllerRunLoop):
1499         (WTR::TestControllerRunLoop::timerEvent):
1500         (WTR::TestController::notifyDone):
1501         (WTR::TestController::platformRunUntil):
1502         * WebKitTestRunner/win/TestControllerWin.cpp:
1503         (WTR::TestController::notifyDone): Add method stub.
1504
1505 2010-11-15  Darin Adler  <darin@apple.com>
1506
1507         * Scripts/webkitpy/common/net/bugzilla: Added property svn:ignore.
1508
1509 2010-11-15  Benjamin Poulain  <benjamin.poulain@nokia.com>
1510
1511         Reviewed by Andreas Kling.
1512
1513         [Qt] [WK2] Add Ctrl+L shortcut to Qt Minibrowser
1514         https://bugs.webkit.org/show_bug.cgi?id=49544
1515
1516         Add the shortcut to BrowserWindow.
1517
1518         * MiniBrowser/qt/BrowserWindow.cpp:
1519         (BrowserWindow::BrowserWindow):
1520         (BrowserWindow::openLocation):
1521         * MiniBrowser/qt/BrowserWindow.h:
1522
1523 2010-11-14  David Levin  <levin@chromium.org>
1524
1525         Reviewed by Daniel Bates.
1526
1527         check-webkit-style should detect PassRefPtr usage in functions.
1528         https://bugs.webkit.org/show_bug.cgi?id=49513
1529
1530         * Scripts/webkitpy/style/checkers/cpp.py:
1531         (check_for_function_lengths): Revert a comment change that I
1532          accidentally made in r71986.
1533         (check_pass_ptr_usage): Added the code to do the check.
1534         (process_line): Added the call to check_pass_ptr_usage.
1535         (CppChecker): Added the new error category.
1536         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
1537         (CppStyleTestBase::perform_pass_ptr_check): Runs the new check for
1538          testing purposes.
1539         (PassPtrTest::*): The class/functions to unit test the new
1540          functionality.
1541
1542 2010-11-14  David Levin  <levin@chromium.org>
1543
1544         Reviewed by Shinichiro Hamaji.
1545
1546         check-webkit-style function detection and the line count style checks should be separate.
1547         https://bugs.webkit.org/show_bug.cgi?id=49512
1548
1549         * Scripts/webkitpy/style/checkers/cpp.py: Do the separation.
1550         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Adjust the test to
1551           call the detection function and fix line counts in two places now that
1552           the code really only counts the lines in the body of the function.
1553
1554 2010-11-14  Andreas Kling  <kling@webkit.org>
1555
1556         Reviewed by Antonio Gomes.
1557
1558         http/tests/plugins tests print "Unhandled variable" to stderr under Qt
1559         https://bugs.webkit.org/show_bug.cgi?id=33438
1560
1561         Stifle "Unhandled variable" warning to match what the other
1562         TestNetscapePlugin does (NPP_GetValue in TestNetscapePlugIn/main.cpp
1563         simply returns NPERR_GENERIC_ERROR for unhandled variables.)
1564
1565         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
1566         (webkit_test_plugin_get_value):
1567
1568 2010-11-13  David Levin  <levin@chromium.org>
1569
1570         Reviewed by Eric Seidel.
1571
1572         check-webkit-style function detection crashes on functions in templates.
1573         https://bugs.webkit.org/show_bug.cgi?id=49504
1574
1575         For example "template <bool shouldClose(const Element*)>".
1576
1577         * Scripts/webkitpy/style/checkers/cpp.py: If not function is detected, bail out.
1578         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Modified the complex
1579           function detection test to expose this issue.
1580
1581 2010-11-12  David Levin  <levin@chromium.org>
1582
1583         Reviewed by Shinichiro Hamaji.
1584
1585         check-webkit-style function detection doesn't detect indented functions declaractions.
1586         https://bugs.webkit.org/show_bug.cgi?id=49446
1587
1588         Indented function declarations occur inside class definitions, so
1589         they are a pretty common (and worth detecting).
1590
1591         * Scripts/webkitpy/style/checkers/cpp.py:
1592           Changed regex to allow indentation.
1593           Changed the function start detection to only happen when not in a
1594           function.
1595           Changed function end detection to work based on matching braces
1596           instead of finding a close brace at the beginning of the line.
1597           Fixed close_expression to do what it says when it doesn't find
1598           the close.
1599         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Indented function test.
1600
1601 2010-11-12  Daniel Bates  <dbates@rim.com>
1602
1603         Rubber-stamped by Nikolas Zimmermann.
1604
1605         Make do-file-rename rename files in the directory WebKit2.
1606
1607         This also makes the list of searched directories in do-file-rename
1608         consistent with the list of searched directories in the script do-webcore-rename.
1609
1610         * Scripts/do-file-rename:
1611
1612 2010-11-12  Dirk Pranke  <dpranke@chromium.org>
1613
1614         Reviewed by James Robinson.
1615
1616         Attempt yet again to land the fix for bug 49360 (respecting
1617         set-webkit-configuration). We need to handle the cases where
1618         trying to run webkit-build-directory to find out where the
1619         default configuration might be fails (that shows up on some
1620         Chromium bots that apparently don't have perl installed).
1621
1622         https://bugs.webkit.org/show_bug.cgi?id=49360
1623
1624         * Scripts/webkitpy/layout_tests/port/config.py:
1625         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Added.
1626         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1627
1628 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
1629
1630         Reviewed by Adam Barth.
1631
1632         webkit-patch land logs commit message twice
1633         https://bugs.webkit.org/show_bug.cgi?id=49482
1634         
1635         Don't log the comment that's used by bugzilla.close_bug_as_fixed, since
1636         that's just the commit message, which we now output following the
1637         actual commit.
1638
1639         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
1640
1641 2010-11-12  Tony Chang  <tony@chromium.org>
1642
1643         Reviewed by Adam Barth.
1644
1645         run platform/chromium/plugins/script-object-invoke.html on all platforms
1646         https://bugs.webkit.org/show_bug.cgi?id=49280
1647
1648         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1649         (testScriptObjectInvoke):
1650         (pluginInvoke):
1651
1652 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
1653
1654         Reviewed by Adam Barth.
1655
1656         webkit-patch land-cowboy (and other commands) should display the committed SVN revision after landing a patch
1657         https://bugs.webkit.org/show_bug.cgi?id=49471
1658
1659         Right now there's no indication of a successful commit.
1660         
1661         * Scripts/webkitpy/tool/commands/download_unittest.py:
1662         * Scripts/webkitpy/tool/steps/commit.py:
1663
1664 2010-11-12  Kinuko Yasuda  <kinuko@chromium.org>
1665
1666         Reviewed by Ojan Vafai.
1667
1668         Include detailed test modifiers in results.json and enable incremental uploading for non-layout tests
1669         https://bugs.webkit.org/show_bug.cgi?id=49354
1670
1671         Also moved/integrated the upload method from run_webkit_tests.py to json_results_generator.py.
1672
1673         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
1674         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1675         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
1676         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1677
1678 2010-11-12  Tony Chang  <tony@chromium.org>
1679
1680         Reviewed by Kent Tamura.
1681
1682         [chromium] Load a blank page when closing DRT WebViewHosts
1683         https://bugs.webkit.org/show_bug.cgi?id=49417
1684
1685         In test_shell, when we close a window, we first load about:blank
1686         and run GC to fire the destruction logic of the page.  In test_shell,
1687         this happens in ~TestShell.  In DRT, we manage virtual windows as
1688         WebViewHosts, so we need to replicate this logic in ~WebViewHost.
1689
1690         Managing focus is a bit trickier, since normally you would get window
1691         events to reset the focus to the right window.  We do it manually
1692         instead.
1693
1694         This fixes 2 plugin tests on DRT.
1695
1696         * DumpRenderTree/chromium/TestShell.cpp:
1697         (TestShell::~TestShell): Move window close logic to ~WebViewHost.
1698         (TestShell::closeWindow): Move widget close() to WVH destructor and
1699             manage focus manually.
1700         (TestShell::closeRemainingWindows): Comment typo.
1701         * DumpRenderTree/chromium/WebViewHost.cpp:
1702         (WebViewHost::closeWidget):
1703         (invokeCloseWidget):
1704         (WebViewHost::closeWidgetSoon): Close the window asynchronously (matches
1705             test_shell).
1706         (WebViewHost::~WebViewHost):
1707         * DumpRenderTree/chromium/WebViewHost.h:
1708
1709 2010-11-12  David Levin  <levin@chromium.org>
1710
1711         Reviewed by Adam Barth.
1712
1713         cpp.py needs some tidying.
1714         https://bugs.webkit.org/show_bug.cgi?id=49443
1715
1716         * Scripts/webkitpy/style/checkers/cpp.py:
1717           Fix parameter ordering for iteratively_replace_matches_with_char
1718           to mirror other similar functions (and improved comments, etc.)
1719           Also, use a python-ism (for/else) to get rid of a variable that
1720           was there to determine if something had been found in the loop.
1721
1722 2010-11-12  Robert Kroeger  <rjkroege@chromium.org>
1723
1724         Reviewed by James Robinson.
1725
1726         [Chromium/DRT] Make EventSendingController send complete touch events.
1727         https://bugs.webkit.org/show_bug.cgi?id=49285
1728
1729         * DumpRenderTree/chromium/EventSender.cpp:
1730         (EventSender::addTouchPoint):
1731         (EventSender::updateTouchPoint):
1732         (EventSender::sendCurrentTouchEvent):
1733
1734 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
1735
1736         Unreviewed rollout of r71858.
1737
1738         Rollout out r71858 since it breaks new-run-webkit-httpd as used by the
1739         NaCl tests.
1740
1741         * Scripts/webkitpy/layout_tests/port/config.py:
1742         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Removed.
1743         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1744
1745 2010-11-12  Benjamin Poulain  <benjamin.poulain@nokia.com>
1746
1747         Reviewed by Kenneth Rohde Christiansen.
1748
1749         [Qt] TestController::runUntil doesn't honor the timeout parameter
1750         https://bugs.webkit.org/show_bug.cgi?id=48941
1751
1752         Implement the missing timeout of WebKitTestRunner for Qt.
1753
1754         A QElapsedTimer has been added to RunUntilConditionLoop in order
1755         to measure how long the loop has been running. When the timer
1756         is bigger or equal than the timeout, the loop ends even if the
1757         condition is not met.
1758
1759         * WebKitTestRunner/qt/TestControllerQt.cpp:
1760         (WTR::RunUntilConditionLoop::start):
1761         (WTR::RunUntilConditionLoop::run):
1762         (WTR::RunUntilConditionLoop::timerEvent):
1763         (WTR::TestController::platformRunUntil):
1764
1765 2010-11-11  David Levin  <levin@chromium.org>
1766
1767         Reviewed by Shinichiro Hamaji.
1768
1769         check-webkit-style function detection doesn't handle templates with spaces.
1770         https://bugs.webkit.org/show_bug.cgi?id=49427
1771
1772         * Scripts/webkitpy/style/checkers/cpp.py: Stripped out templates when
1773           finding the function name.
1774         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test have a
1775           template with a space.
1776
1777 2010-11-11  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
1778
1779         Reviewed by Shinichiro Hamaji.
1780
1781         Enable check-webkit-style on Qt files
1782         https://bugs.webkit.org/show_bug.cgi?id=48322
1783
1784         Adding exception for _q_... private slots and
1785         ..._data() methods in tests
1786
1787         * Scripts/webkitpy/style/checker.py: enable Qt dirs, add exceptions
1788         * Scripts/webkitpy/style/checker_unittest.py: improve unittests
1789
1790 2010-11-11  David Levin  <levin@chromium.org>
1791
1792         Reviewed by Shinichiro Hamaji.
1793
1794         check-webkit-style function detection doesn't work for templates and destructors.
1795         https://bugs.webkit.org/show_bug.cgi?id=49425
1796
1797         * Scripts/webkitpy/style/checkers/cpp.py: Changed regex for templates and destructors.
1798         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test hit this issue.
1799
1800 2010-11-11  Joone Hur  <joone@kldp.org>
1801
1802         Reviewed by Antonio Gomes.
1803
1804         [GTK][DRT] Implement LayoutTestController::nodesFromRect
1805         https://bugs.webkit.org/show_bug.cgi?id=46598
1806
1807         Support nodesFromRect in DRT
1808
1809         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1810         (LayoutTestController::nodesFromRect):
1811
1812 2010-11-11  MORITA Hajime  <morrita@google.com>
1813
1814         Reviewed by Kent Tamura.
1815
1816         editing/selection/context-menu-on-text.html fails on chromium
1817         https://bugs.webkit.org/show_bug.cgi?id=45898
1818
1819         Enhance Chromium's EventSender::contextClick() to return
1820         array of menu strings. The implementation incomplete because
1821         Although actual context menu is implemented by the browser,
1822         it reflects editability of underlying context thus makes
1823         context-menu-on-text.html passed.
1824         
1825         * DumpRenderTree/chromium/EventSender.cpp:
1826         (makeMenuItemStringsFor): Added.
1827         (EventSender::contextClick): Now returns array of menu strings.
1828         * DumpRenderTree/chromium/WebViewHost.cpp:
1829         (WebViewHost::showContextMenu): Now storing given WebContextMenuData object.
1830         (WebViewHost::clearContextMenuData): Added.
1831         (WebViewHost::lastContextMenuData): Added.
1832         * DumpRenderTree/chromium/WebViewHost.h:
1833
1834 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1835
1836         Unreviewed, build fix.
1837
1838         fix build breakage caused by fix for bug 49380 (r71858).
1839
1840         config_standalone needs to ensure that the packages it needs are
1841         in sys.path.
1842
1843         https://bugs.webkit.org/show_bug.cgi?id=49419
1844
1845         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
1846
1847 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1848
1849         Reviewed by Adam Roben.
1850
1851         Fix NRWT to respect set-webkit-configuration again :(
1852
1853         This change fixes a typo in config.py that was causing the wrong
1854         value to be read initially and us never actually looking into
1855         the filesystem to get the default configuration.
1856         
1857         * Scripts/webkitpy/layout_tests/port/config.py:
1858         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
1859         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1860
1861 2010-11-11  Eric Seidel  <eric@webkit.org>
1862
1863         Reviewed by Adam Barth.
1864
1865         Split out Bug, Attachment and CommitterValidator from bugzilla.py
1866         https://bugs.webkit.org/show_bug.cgi?id=49403
1867
1868         CommitterValidator really had nothing to do with bugzilla.py.
1869         I've put it next to committers.py for now since its rather tied to that file.
1870
1871         * Scripts/webkitpy/common/config/committervalidator.py: Added.
1872         * Scripts/webkitpy/common/config/committervalidator_unittest.py: Added.
1873         * Scripts/webkitpy/common/net/bugzilla/__init__.py:
1874         * Scripts/webkitpy/common/net/bugzilla/attachment.py: Added.
1875         * Scripts/webkitpy/common/net/bugzilla/bug.py: Added.
1876         * Scripts/webkitpy/common/net/bugzilla/bug_unittest.py: Added.
1877         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
1878         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py:
1879         * Scripts/webkitpy/tool/bot/feeders.py:
1880         * Scripts/webkitpy/tool/commands/queues.py:
1881
1882 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1883
1884         Reviewed by Eric Seidel.
1885
1886         fix missing 'pid' argument for http lockfile checking on win32
1887         https://bugs.webkit.org/show_bug.cgi?id=49363
1888
1889         * Scripts/webkitpy/common/system/executive.py:
1890
1891 2010-11-11  Eric Seidel  <eric@webkit.org>
1892
1893         Rubber-stamped by Adam Barth.
1894
1895         Move bugzilla.py into its own module in preparation for splitting one-file-per-class
1896         https://bugs.webkit.org/show_bug.cgi?id=49402
1897
1898         * Scripts/webkitpy/common/net/bugzilla/__init__.py: Added.
1899         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla.py.
1900         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla_unittest.py.
1901
1902 2010-11-10  Chris Guillory  <chris.guillory@google.com>
1903
1904         Reviewed by Chris Fleizach.
1905
1906         Use IAccessibleComparable to compare accessibility objects.
1907         https://bugs.webkit.org/show_bug.cgi?id=49118
1908
1909         * DumpRenderTree/AccessibilityUIElement.cpp:
1910         (AccessibilityUIElement::isEqual):
1911         * DumpRenderTree/AccessibilityUIElement.h:
1912         * DumpRenderTree/win/AccessibilityUIElementWin.cpp:
1913         (comparableObject):
1914         (AccessibilityUIElement::isEqual):
1915
1916 2010-11-10  Csaba Osztrogonác  <ossy@webkit.org>
1917
1918         Reviewed by David Hyatt.
1919
1920         HTML5 Ruby support should be mandatory feature
1921         https://bugs.webkit.org/show_bug.cgi?id=49272
1922
1923         Remove Ruby as optional feature.
1924
1925         * Scripts/build-webkit:
1926
1927 2010-11-10  Adam Roben  <aroben@apple.com>
1928
1929         Print VC++ Express build logs in the same order that the projects are
1930         built
1931
1932         Fixes <http://webkit.org/b/49326> print-vse-failure-logs prints logs
1933         in an unhelpful order
1934
1935         Reviewed by Eric Seidel.
1936
1937         * Scripts/print-vse-failure-logs:
1938         (PrintVisualStudioExpressLogs._build_order): Added. Uses
1939         print-msvc-project-dependencies to figure out the order in which
1940         projects are built.
1941         (PrintVisualStudioExpressLogs._sort_buildlogs): Added. Sorts the logs
1942         based on their build order and project name.
1943         (PrintVisualStudioExpressLogs._obj_directory): Moved code to find the
1944         scripts directory from here...
1945         (PrintVisualStudioExpressLogs._scripts_directory): ...to here.
1946         (PrintVisualStudioExpressLogs.main): Sort the logs before printing
1947         them.
1948
1949 2010-11-10  Adam Roben  <aroben@apple.com>
1950
1951         Only print the interesting text from VC++ Express build logs
1952
1953         Fixes <http://webkit.org/b/49325> It's hard to find the interesting
1954         output from print-vse-failure-logs
1955
1956         Reviewed by Eric Seidel.
1957
1958         * Scripts/print-vse-failure-logs: Removed unnecessary os.path import,
1959         added newly-required imports. Removed a comment that was essentially a
1960         FIXME, and is now fixed.
1961         (PrintVisualStudioExpressLogs._relevant_text): Added. Uses
1962         BeautifulSoup to extract the relevant text from the build log.
1963         (PrintVisualStudioExpressLogs.main): Only print the relevant text.
1964
1965 2010-11-09  Mihai Parparita  <mihaip@chromium.org>
1966
1967         Reviewed by Tony Chang.
1968
1969         [Chromium] http/tests/security/XFrameOptions fail with Chromium DRT
1970         https://bugs.webkit.org/show_bug.cgi?id=49286
1971         
1972         The tests in http/tests/security/XFrameOptions started to fail after
1973         r71297, since it changed the WebViewHost::assignIdentifierToRequest
1974         logic to always assign identifiers to requests, even if we're not going
1975         to be dumping resource load callbacks at that point. These tests
1976         only call LayoutTestController.dumpResourceLoadCallbacks after the
1977         request is started, thus their expectations have "<unknown>" as the
1978         identifier. Change WebViewHost to the previous behavior, which also
1979         match the other ports'.
1980
1981         * DumpRenderTree/chromium/WebViewHost.cpp:
1982         (WebViewHost::assignIdentifierToRequest):
1983         (WebViewHost::removeIdentifierForRequest):
1984
1985 2010-11-08  Tony Chang  <tony@chromium.org>
1986
1987         Reviewed by Adam Barth.
1988
1989         run platform/chromium/plugins/return-npobject.html on all platforms
1990         https://bugs.webkit.org/show_bug.cgi?id=49103
1991
1992         This tests that we can get an NPObject returned through a method on
1993         an NPAPI Object.
1994
1995         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1996         (testCallbackReturn):
1997         (pluginInvoke):
1998
1999 2010-11-09  Sheriff Bot  <webkit.review.bot@gmail.com>
2000
2001         Unreviewed, rolling out r71672.
2002         http://trac.webkit.org/changeset/71672
2003         https://bugs.webkit.org/show_bug.cgi?id=49276
2004
2005         It doesn't work on Chromium Windows bot (Requested by Ossy on
2006         #webkit).
2007
2008         * Scripts/webkitpy/common/system/file_lock.py: Removed.
2009         * Scripts/webkitpy/common/system/file_lock_unittest.py: Removed.
2010         * Scripts/webkitpy/layout_tests/port/http_lock.py:
2011
2012 2010-11-09  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2013
2014         Reviewed by Tony Chang.
2015
2016         [NRWT] Make http locking similar to perl implementation
2017         https://bugs.webkit.org/show_bug.cgi?id=49187
2018
2019         * Scripts/webkitpy/common/system/file_lock.py: Added.
2020         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
2021         * Scripts/webkitpy/layout_tests/port/http_lock.py:
2022
2023 2010-11-09  James Kozianski  <koz@chromium.org>
2024
2025         Reviewed by Eric Seidel.
2026
2027         Clean up imports for webkit-patch.
2028         https://bugs.webkit.org/show_bug.cgi?id=49083
2029
2030         Move some imports from commands into main.py where they are used.
2031         Move command imports into the __init__ of the commands module.
2032
2033         * Scripts/webkitpy/tool/commands/__init__.py:
2034         * Scripts/webkitpy/tool/commands/download.py:
2035         * Scripts/webkitpy/tool/commands/queues.py:
2036         * Scripts/webkitpy/tool/main.py:
2037
2038 2010-11-09  Yi Shen  <yi.4.shen@nokia.com>
2039
2040         Reviewed by Andreas Kling.
2041
2042         [Qt][QtTestBrowser] Don't erase incorrect url in the Url Bar
2043         https://bugs.webkit.org/show_bug.cgi?id=49047
2044
2045         * QtTestBrowser/mainwindow.cpp:
2046         (MainWindow::setAddressUrl):
2047
2048 2010-11-09  Andras Becsi  <abecsi@webkit.org>
2049
2050         Reviewed by Csaba Osztrogonác.
2051
2052         Improve old-run-webkit-tests --verbose to show DumpTool opening/closing
2053         https://bugs.webkit.org/show_bug.cgi?id=49252
2054
2055         * Scripts/old-run-webkit-tests:
2056
2057 2010-11-09  Csaba Osztrogonác  <ossy@webkit.org>
2058
2059         Reviewed by Tor Arne Vestbø.
2060
2061         Output of old-run-webkit-tests is buggy in verbose mode
2062         https://bugs.webkit.org/show_bug.cgi?id=49249
2063
2064         * Scripts/old-run-webkit-tests: Redundant outputs removed.
2065
2066 2010-11-08  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
2067
2068         Reviewed by Martin Robinson.
2069
2070         [GTK] Link with target name set does not work
2071         https://bugs.webkit.org/show_bug.cgi?id=48865
2072
2073         When a new page is created with a name (target=myFrame), the new
2074         mainFrame could not be found because they where not stored in the
2075         same PageGroup. As PageGroup are not exposed externally so the
2076         simpliest solution is to use a global page group name. This also fixes
2077         issue with visited link coloration across pages. After this change the
2078         private function webkit_web_view_set_group_name() was no longer used
2079         so it was removed completly.
2080
2081         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2082         (createWebView):
2083
2084 2010-11-08  John Knottenbelt  <jknotten@chromium.org>
2085
2086         Reviewed by Kent Tamura.
2087
2088         WebViewHost::reset() uses placement new.
2089         https://bugs.webkit.org/show_bug.cgi?id=49069
2090
2091         WebViewHost is using placement destruction / new to simulate a fresh
2092         WebViewHost object at the same address. This is because the WebView remains
2093         open across tests and maintains a pointer to the WebViewHost.
2094         This change resets member variables explictly instead of the placement new dance.
2095
2096         * DumpRenderTree/chromium/WebViewHost.cpp:
2097         (WebViewHost::WebViewHost):
2098         (WebViewHost::reset):
2099
2100 2010-11-08  Dirk Pranke <dpranke@chromium.org>
2101
2102         Reviewed by Eric Seidel.
2103
2104         Fix the webkit_base_dir logic in webkitpy/layout_tests/port/config.py. 
2105         It turns out that NRWT can't use the code in scm.find_checkout_root()
2106         because the Chromium bots don't do full checkouts of the WebKit
2107         tree; they only check out subdirectories like WebKitTools/Scripts.
2108         Until we can figure out a better approach for this, I've
2109         restored the base_dir-detecting code from NRWT, which works in
2110         any directory tree, scm or no.
2111
2112         This also restores the files modified in r71475 and r71474.
2113
2114         https://bugs.webkit.org/show_bug.cgi?id=49151
2115
2116         * WebKitTools/Scripts/webkitpy/layout_tests/port/config.py
2117         * WebKitTools/Scripts/webkitpy/layout_tests/port/config_unittest.py
2118
2119 2010-11-08  Eric Seidel  <eric@webkit.org>
2120
2121         Reviewed by Adam Barth.
2122
2123         webkit-patch failure-reason explodes if a build is missing
2124         https://bugs.webkit.org/show_bug.cgi?id=49195
2125
2126         This is likely a recent regression.  Adding a unit test for this case.
2127
2128         * Scripts/webkitpy/common/net/buildbot.py:
2129         * Scripts/webkitpy/common/net/buildbot_unittest.py:
2130
2131 2010-11-08  Dan Bernstein  <mitz@apple.com>
2132
2133         Reviewed by Darin Adler.
2134
2135         Ignore the system scrollbar setting.
2136
2137         * DumpRenderTree/mac/DumpRenderTree.mm:
2138         (resetDefaultsToConsistentValues): Set the scrollbar default.
2139
2140 2010-11-05  Adam Roben  <aroben@apple.com>
2141
2142         Make webkitpy.common.system.executive_unittest pass when running under
2143         Win32 Python
2144
2145         Fixes <http://webkit.org/b/49033>.
2146
2147         Reviewed by Dave Levin and Eric Seidel.
2148
2149         * Scripts/webkitpy/common/system/executive.py:
2150         (Executive._run_command_with_teed_output): Pass the arguments through
2151         encode_argument_if_needed rather than using Cygwin-specific code here.
2152         (Executive.run_and_throw_if_fail): Use child_process_encoding to decode
2153         the output.
2154         (Executive.run_command): Use encode_argument_if_needed to encode the
2155         arguments and child_process_encoding to decode the output.
2156         (Executive._child_process_encoding): Returns the encoding that should be
2157         used when communicating with child processes. On Windows we use mbcs,
2158         which maps to the current code page. On all other platforms we use
2159         UTF-8.
2160         (Executive._should_encode_child_process_arguments): Returns True if
2161         arguments to child processes need to be encoded. This is currently
2162         only needed on Cygwin and Win32 Python 2.x.
2163         (Executive._encode_argument_if_needed): Encode the argument using
2164         child_process_encoding if we need to encode arguments to child
2165         processes on this platform.
2166
2167         * Scripts/webkitpy/common/system/executive_unittest.py:
2168         (never_ending_command): Added. Returns arguments to run a command that
2169         will not quit until we kill it. On Windows we use wmic, on other
2170         platforms we use yes.
2171         (ExecutiveTest.test_run_command_with_unicode): Changed to expect the
2172         mbcs encoding to be used and for output from the child processes to
2173         have been roundtripped through encode/decode on Win32 Python. When
2174         UTF-8 is the encoding the roundtripping is undetectable, but with mbcs
2175         it's possible that some characters will not be able to be converted
2176         and will be replaced by question marks; the round-tripping allows us
2177         to expect this result.
2178
2179         (ExecutiveTest.test_kill_process):
2180         (ExecutiveTest.test_kill_all):
2181         Use never_ending_command instead of invoking "yes" directly. Expect an
2182         exit code of 1 when using Win32 Python, as that's what seems to happen.
2183
2184 2010-11-08  Adam Roben  <aroben@apple.com>
2185
2186         Roll out r71532
2187
2188         It broke the build for Cygwin 1.7 installs. Cygwin 1.7's default
2189         .bashrc unsets %TEMP%, which broke copy-tools.cmd.
2190
2191         * Scripts/webkitdirs.pm:
2192
2193 2010-11-08  Tony Chang  <tony@chromium.org>
2194
2195         Reviewed by Adam Barth.
2196
2197         run platform/chromium/plugins/nested-plugin-objects.html on all platforms
2198         https://bugs.webkit.org/show_bug.cgi?id=49094
2199
2200         This tests that objects created by plugins are proplery cleaned up.
2201
2202         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
2203         (testPassTestObject):
2204         (pluginInvoke):
2205         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
2206         (testAllocate):
2207         (testDeallocate):
2208         (testGetProperty):
2209         (testConstruct):
2210
2211 2010-11-08  Adam Roben  <aroben@apple.com>
2212
2213         Mark Windows builds triggered from Perl as being non-interactive
2214
2215         This affects whether some of our scripts will show alerts vs. printing
2216         to the build log.
2217
2218         Fixes <http://webkit.org/b/49181> Windows build fail mysteriously when
2219         .vsprops files are updated
2220
2221         Reviewed by Steve Falkenburg.
2222
2223         * Scripts/webkitdirs.pm:
2224         (buildVisualStudioProject): Set WEBKIT_NONINTERACTIVE_BUILD to 1.
2225
2226 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2227
2228         Reviewed by Ojan Vafai.
2229
2230         Make http locking default in NRWT.
2231         https://bugs.webkit.org/show_bug.cgi?id=48053
2232
2233         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2234         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2235
2236 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2237
2238         Reviewed by Csaba Osztrogonác.
2239
2240         [NRWT] If the http lock fails we shouldn't do any locking
2241         https://bugs.webkit.org/show_bug.cgi?id=49164
2242
2243         If something goes wrong with the locking, the test should keep going.
2244
2245         * Scripts/webkitpy/layout_tests/port/http_lock.py:
2246
2247 2010-11-08  Adam Roben  <aroben@apple.com>
2248
2249         Switch back to using kCGImageAlphaPremultipliedFirst when generating
2250         pixel dumps on Windows
2251
2252         I changed this behavior in r71418 thinking that it was required for
2253         getting plugins to show up in pixel dumps. But it doesn't seem to be
2254         necessary, and was making it impossible to compare new Windows pixel
2255         dumps with existing Windows or Mac pixel dumps (because ImageDiff won't
2256         compare an image with alpha to an image without alpha).
2257
2258         Fixes <http://webkit.org/b/49172> REGRESION (r71418): Can't compare
2259         new Windows pixel results to existing Windows or Mac results
2260
2261         Reviewed by Antti Koivisto.
2262
2263         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
2264         (createBitmapContextFromWebView): Replaced kCGImageAlphaNoneSkipFirst
2265         with kCGImageAlphaPremultipliedFirst.
2266
2267 2010-11-08  Csaba Osztrogonac  <ossy@webkit.org>
2268
2269         Unreviewed, rolling out r71466.
2270         http://trac.webkit.org/changeset/71466
2271         https://bugs.webkit.org/show_bug.cgi?id=48865
2272
2273         It broke layout tests on GTK bots.
2274
2275         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2276         (createWebView):
2277
2278 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2279
2280         Reviewed by Csaba Osztrogonác.
2281
2282         Enable running of Qt API tests on BuildBot
2283         https://bugs.webkit.org/show_bug.cgi?id=49004
2284
2285         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
2286
2287 2010-11-08  Eric Seidel  <eric@webkit.org>
2288
2289         Reviewed by Adam Barth.
2290
2291         Add clean-review-queue command to remove closed bugs from the webkit.org/pending-review
2292         https://bugs.webkit.org/show_bug.cgi?id=49160
2293
2294         Bugzilla doesn't automatically remove r? when a bug gets closed.
2295         This script takes care of that for webkit.org.
2296
2297         * Scripts/webkitpy/common/net/bugzilla.py:
2298         * Scripts/webkitpy/tool/commands/upload.py:
2299
2300 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
2301
2302         Unreviewed, rolling out r71474.
2303         http://trac.webkit.org/changeset/71474
2304         https://bugs.webkit.org/show_bug.cgi?id=48280
2305
2306         breaks chromium webkit tests
2307         https://bugs.webkit.org/show_bug.cgi?id=49151
2308
2309         * Scripts/webkitpy/layout_tests/port/base.py:
2310         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2311         * Scripts/webkitpy/layout_tests/port/config.py:
2312         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
2313         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
2314         * Scripts/webkitpy/layout_tests/port/webkit.py:
2315         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2316         * Scripts/webkitpy/tool/mocktool.py:
2317
2318 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
2319
2320         Unreviewed, rolling out r71475.
2321         http://trac.webkit.org/changeset/71475
2322  
2323         breaks chromium webkit tests
2324         https://bugs.webkit.org/show_bug.cgi?id=49151
2325
2326         * Scripts/webkitpy/common/newstringio.py: Removed.
2327         * Scripts/webkitpy/common/newstringio_unittest.py: Removed.
2328         * Scripts/webkitpy/common/system/executive_mock.py: Removed.
2329         * Scripts/webkitpy/common/system/filesystem_mock.py: Removed.
2330         * Scripts/webkitpy/layout_tests/port/config_mock.py: Removed.
2331
2332 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2333
2334         Unreviewed, build fix.
2335
2336         Add files inexplicably not committed in r71474 as part of the
2337         fix for bug 48280.
2338
2339         * Scripts/webkitpy/common/newstringio.py: Added.
2340         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
2341         * Scripts/webkitpy/common/system/executive_mock.py: Added.
2342         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
2343         * Scripts/webkitpy/layout_tests/port/config_mock.py: Added.
2344
2345 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2346
2347         Reviewed by Eric Siedel.
2348
2349         new-run-webkit-tests: update port/base and port/webkit to use the
2350         new FileSystem and Config abstractions, pulling more logic out of
2351         the base Port classes into separate, mockable objects.
2352
2353         Also create a MockFileSystem object, a MockConfig object, move
2354         MockExecutive into common/system to be next to executive, and
2355         update the config object to use a FileSystem.
2356
2357         https://bugs.webkit.org/show_bug.cgi?id=48280
2358
2359         * Scripts/webkitpy/common/newstringio.py: Added.
2360         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
2361         * Scripts/webkitpy/common/system/executive_mock.py: Added.
2362         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
2363         * Scripts/webkitpy/layout_tests/port/base.py:
2364         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2365         * Scripts/webkitpy/layout_tests/port/config.py:
2366         * Scripts/webkitpy/layout_tests/port/config_mock.py:
2367         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
2368         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
2369         * Scripts/webkitpy/layout_tests/port/webkit.py:
2370         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2371         * Scripts/webkitpy/tool/mocktool.py:
2372
2373 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2374
2375         Unreviewed, build breakage.
2376
2377         Apparently I uploaded the wrong version of the file to fix 49122
2378         and neither Eric or I noticed - it was missing a dirname() call.
2379         Fixing ...
2380
2381         https://bugs.webkit.org/show_bug.cgi?id=49122
2382
2383         * Scripts/webkitpy/common/checkout/scm.py:
2384
2385 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2386
2387         Reviewed by Eric Seidel.
2388
2389         webkitpy/tool/* unittests change cwd and don't clean up properly
2390
2391         https://bugs.webkit.org/show_bug.cgi?id=49122
2392
2393         * Scripts/webkitpy/common/checkout/scm.py:
2394
2395 2010-11-05  Chris Marrin  <cmarrin@apple.com>
2396
2397         Reviewed by Simon Fraser.
2398
2399         suspendAnimations/resumeAnimations not present in WebKit2
2400         https://bugs.webkit.org/show_bug.cgi?id=49109
2401
2402         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
2403         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
2404         (WTR::LayoutTestController::suspendAnimations):
2405         (WTR::LayoutTestController::resumeAnimations):
2406         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
2407
2408 2010-11-05  Tony Chang  <tony@chromium.org>
2409
2410         Reviewed by David Levin.
2411
2412         cleanup style in TestNetscapePlugIn/PluginObject.cpp
2413         https://bugs.webkit.org/show_bug.cgi?id=49044
2414
2415         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
2416         (getPluginClass):
2417         (pluginGetProperty):
2418         (pluginSetProperty):
2419         (variantToIdentifier):
2420         (testCallback):
2421         (getURL):
2422         (getURLNotify):
2423         (testInvokeDefault):
2424         (testGetProperty):
2425         (testDocumentOpen):
2426         (testWindowOpen):
2427         (pluginInvoke):
2428         (handleCallback):
2429         (notifyStream):
2430         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
2431
2432 2010-11-05  Eric Seidel  <eric@webkit.org>
2433
2434         Reviewed by David Levin.
2435
2436         Add QueueStatusServer/__init__.py so others can run the QueueStatusServer tests
2437         https://bugs.webkit.org/show_bug.cgi?id=49032
2438
2439         I wrote this file as part of bug 47847, but I forgot to commit it.
2440         No one else noticed it missing because test-webkitpy knows how
2441         to recover in the case where it can't import QueueStatusServer
2442         (which generally occurs due to not having installed the AppEngine SDK).
2443
2444         * QueueStatusServer/__init__.py: Added.
2445         * QueueStatusServer/model/workitems_unittest.py:
2446          - Remove a test which fails.  This was probably landed (by me)
2447            from my other machine, which since this __init__.py was missing
2448            I never noticed the failure and landed this invalid test.
2449            Sadly we can't really test remove_work_item as it depends
2450            on .key() working.  .key() will throw unless the object
2451            has already been saved it seems.
2452            This may be a mis-design in our remove_work_item implementation,
2453            but for now, just removing the test.
2454
2455 2010-11-04  Adam Roben  <aroben@apple.com>
2456
2457         Teach check-webkit-style about TestNetscapePlugIn
2458
2459         Fixes <http://webkit.org/b/49030> check-webkit-style is ignorant of
2460         TestNetscapePlugIn's coding conventions
2461
2462         Reviewed by John Sullivan.
2463
2464         * Scripts/webkitpy/style/checker.py: Lump TestNetscapePlugIn in with
2465         WebKitAPITest and TestWebKitAPI in having different include order and
2466         naming conventions than WebCore. Added some comments about why the
2467         exceptions exist.
2468
2469 2010-11-04  Adam Roben  <aroben@apple.com>
2470
2471         Add a test that shows that windowed plugins are able to paint
2472
2473         Somehow we never had a test for this before.
2474
2475         Fixes <http://webkit.org/b/49029> Should add a test that shows
2476         windowed plugins can paint
2477
2478         Reviewed by Jon Honeycutt.
2479
2480         * DumpRenderTree/TestNetscapePlugIn/Tests/win/DrawsGradient.cpp: Added.
2481         (DrawsGradient::DrawsGradient):
2482         (DrawsGradient::wndProc): We handle the WM_PAINT and WM_PRINTCLIENT messages.
2483
2484         (DrawsGradient::onPaint):
2485         (DrawsGradient::onPrintClient):
2486         These both just call through to paint.
2487
2488         (DrawsGradient::paint): Fills our client area with some gradients.
2489
2490         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Link
2491         against Msimg32.lib for ::GradientFill and added DrawsGradient.
2492
2493         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
2494         (createBitmapContextFromWebView): Use WM_PRINT instead of
2495         WM_PRINTCLIENT so that ::DefWindowProc will send
2496         WM_PRINT/WM_PRINTCLIENT messages to the WebView's child windows.
2497         Replaced kCGImageAlphaPremultipledFirst with kCGImageAlphaNoneSkipFirst
2498         because GDI doesn't support alpha and kCGImageBitmapOrder32Little with
2499         kCGImageBitmapOrder32Host because, who knows, maybe someday Windows
2500         will run on a big-endian platform.
2501
2502 2010-11-04  Adam Roben  <aroben@apple.com>
2503
2504         Extract much of NPNInvalidateRectInvalidatesWindow's code into a
2505         WindowedPluginTest base class
2506
2507         The base class takes care of subclassing the plugin's window so that a
2508         custom WNDPROC is called. This will make it easier to write tests that
2509         need to handle window messages.
2510
2511         Fixes <http://webkit.org/b/49028> It's hard to write a PluginTest with
2512         a custom WNDPROC
2513
2514         Reviewed by Jon Honeycutt.
2515
2516         * DumpRenderTree/TestNetscapePlugIn/Tests/win/NPNInvalidateRectInvalidatesWindow.cpp:
2517         (NPNInvalidateRectInvalidatesWindow::NPNInvalidateRectInvalidatesWindow):
2518         (NPNInvalidateRectInvalidatesWindow::NPP_SetWindow):
2519         (NPNInvalidateRectInvalidatesWindow::wndProc):
2520         (NPNInvalidateRectInvalidatesWindow::testInvalidateRect):
2521         Moved code from here to WindowedPluginTest. Changed to use window()
2522         instead of m_window.
2523
2524         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2525         Added WindowedPluginTest and added TestNetscapePlugIn/win to the
2526         include path. Also reordered the include path to make a little more
2527         sense and simplified the entry that added TestNetscapePlugIn itself to
2528         the include path.
2529
2530         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.cpp: Added.
2531         (WindowedPluginTest::WindowedPluginTest):
2532         (WindowedPluginTest::NPP_SetWindow):
2533         (WindowedPluginTest::staticWndProc):
2534         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.h: Added.
2535         (WindowedPluginTest::window):
2536         Code came from NPNInvalidateRectInvalidatesWindow.
2537
2538 2010-11-04  Adam Roben  <aroben@apple.com>
2539
2540         Add a plugin test to show that windowed plugins are clipped correctly
2541
2542         Fixes <http://webkit.org/b/49024> <rdar://problem/8487847> Windowed
2543         plugins aren't clipped in WebKit2 on Windows
2544
2545         Reviewed by Jon Honeycutt.
2546
2547         * DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowRegionIsSetToClipRect.cpp: Added.
2548         (WindowRegionIsSetToClipRect::WindowRegionIsSetToClipRect): Initialize members.
2549         (WindowRegionIsSetToClipRect::NPP_SetWindow): Check that our window
2550         region matches the clip rect we know we should have based on
2551         window-region-is-set-to-clip-rect.html, and check that our window class
2552         doesn't have the CS_PARENTDC style.
2553
2554         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2555         Added WindowRegionIsSetToClipRect.
2556
2557 2010-11-05  Alejandro G. Castro  <alex@igalia.com>
2558
2559         Reviewed by Martin Robinson.
2560
2561         [GTK] Avoid font hinting in the DRT
2562         https://bugs.webkit.org/show_bug.cgi?id=48548
2563
2564         Change the settings to avoid font hinting, it was causing
2565         different results depending on the distribution.
2566
2567         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2568         (initializeGtkFontSettings):
2569         * DumpRenderTree/gtk/fonts/fonts.conf:
2570
2571 2010-11-05  Eric Seidel  <eric@webkit.org>
2572
2573         Reviewed by Adam Barth.
2574
2575         build-webkit spams system.log by repeatedly checking for PBXProductDirectory in com.apple.Xcode
2576         https://bugs.webkit.org/show_bug.cgi?id=49051
2577
2578         This is a speculative fix.  The unit tests cover these methods, however
2579         I don't know if this will fully stop the system.log spam.
2580
2581         * Scripts/webkitpy/layout_tests/port/base.py:
2582         * Scripts/webkitpy/layout_tests/port/webkit.py:
2583
2584 2010-11-05  Eric Seidel  <eric@webkit.org>
2585
2586         Reviewed by Adam Barth.
2587
2588         Add basic support for showing bot id on /queue-status/ pages
2589         https://bugs.webkit.org/show_bug.cgi?id=49037
2590
2591         This support is really simple.  Eventually we'll want to
2592         show the bot id in the lock table too, but we don't have
2593         that information stored in the server yet.
2594
2595         * QueueStatusServer/handlers/queuestatus.py:
2596         * QueueStatusServer/handlers/queuestatus_unittest.py: Added.
2597         * QueueStatusServer/templates/includes/singlequeuestatus.html:
2598
2599 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
2600
2601         Reviewed by Adam Barth.
2602
2603         Create a filesystem wrapper that we can use to enforce
2604         particular conventions and use for mocking and dependency
2605         injection down the line.
2606
2607         https://bugs.webkit.org/show_bug.cgi?id=48144
2608
2609         * Scripts/webkitpy/common/system/filesystem.py: Added.
2610         * Scripts/webkitpy/common/system/filesystem_unittest.py: Added.
2611
2612 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2613
2614         Reviewed by Tony Chang.
2615
2616         run_webkit_tests_unittest fails under Python 2.5
2617         https://bugs.webkit.org/show_bug.cgi?id=49043
2618         
2619         Switch from itertools.chain.from_iterable (which was added in 2.6)
2620         to using itertools.chain directly.
2621
2622         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2623
2624 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2625
2626         Unreviewed fix to webkit-patch rebaseline-server so that it runs under
2627         Python 2.5 (it needs an import to use the with statement).
2628
2629         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
2630
2631 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
2632
2633         Reviewed by Adam Barth.
2634
2635         new-run-webkit-tests: split out webkit-specific configuration stuff into a new module
2636
2637         The current NRWT code has webkit-specific configuration code (like
2638         _script_path, default configuration, etc.) mixed in with
2639         layout-test-specific stuff in port/base. The configuration code
2640         should be split out into a separate module for easier mocking,
2641         testing, and isolation.
2642
2643         https://bugs.webkit.org/show_bug.cgi?id=48264
2644
2645         * Scripts/webkitpy/layout_tests/port/config.py: Added.
2646         * Scripts/webkitpy/layout_tests/port/config_unittest.py: Added.
2647
2648 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2649
2650         Reviewed by Tony Chang.
2651
2652         Rebaseline server: initial framework
2653         https://bugs.webkit.org/show_bug.cgi?id=48892
2654         
2655         Adds the basic framework for the rebaseline server (details at 
2656         http://webkit.org/b/47761). Includes the rebaseline-server webkit-patch
2657         command, which starts an HTTP server that can serve static files or
2658         invoke handler methods on a class.
2659
2660         * Scripts/webkitpy/tool/commands/__init__.py:
2661         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html: Added.
2662         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css: Added.
2663         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js: Added.
2664         * Scripts/webkitpy/tool/commands/rebaselineserver.py: Added.
2665
2666 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
2667
2668         Reviewed by James Robinson.
2669
2670         new-run-webkit-tests wasn't using DRT by default for
2671         --platform chromium-gpu
2672
2673         The default value was set to False instead of None, which meant
2674         that the platform specific logic wasn't firing to change the
2675         value to True (b/c we were afraid we'd be overriding the user
2676         preference).
2677
2678         https://bugs.webkit.org/show_bug.cgi?id=49038
2679
2680         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2681
2682 2010-11-04  Tony Chang  <tony@chromium.org>
2683
2684         Reviewed by Adam Barth.
2685
2686         make platform/chromium/plugins/multiple-plugins.html pass on all platforms
2687         https://bugs.webkit.org/show_bug.cgi?id=49026
2688
2689         Have the testObject.bar property return the string "bar".  This lets
2690         us run and pass platform/chromium/plugins/multiple-plugins.html on all
2691         platforms.
2692
2693         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
2694         (testGetProperty):
2695
2696 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2697
2698         Reviewed by Tony Chang.
2699
2700         NRWT doesn't have good test coverage for --run-chunk, --batch-size, --run-part, etc.
2701         https://bugs.webkit.org/show_bug.cgi?id=48878
2702         
2703         Add get_tests_run so that it's easy to see which tests get run (and with
2704         what batching) for a given flag combination. Flesh out the various
2705         test cases that have FIXMEs.
2706         
2707         Also fixes an off-by-one error (batch sizes were one larger than 
2708         expected) and makes --run-part also have wraparound behavior, like
2709         --run-chunk.
2710
2711         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2712         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2713         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2714
2715 2010-11-04  Erik Arvidsson  <arv@chromium.org>
2716
2717         Reviewed by Dimitri Glazkov.
2718
2719         Support box-sizing without the vendor prefix
2720         https://bugs.webkit.org/show_bug.cgi?id=36713
2721
2722         Based on patch by Peter Beverloo <peter@lvp-media.com>
2723
2724         * iExploder/htdocs/cssproperties.in: Change -webkit-box-sizing to box-sizing.
2725
2726 2010-11-04  Csaba Osztrogonác  <ossy@webkit.org>
2727
2728         Unreviewed rollout r71340, because it broke Chromium Windows bot.
2729
2730         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2731         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2732
2733 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2734
2735         Reviewed by Ojan Vafai.
2736
2737         Make http locking default in NRWT.
2738         https://bugs.webkit.org/show_bug.cgi?id=48053
2739
2740         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2741         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2742
2743 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2744
2745         Reviewed by Eric Seidel.
2746
2747         [NRWT] Clear invalid http locks on Windows platform as well
2748         https://bugs.webkit.org/show_bug.cgi?id=48515
2749
2750         * Scripts/webkitpy/common/system/executive.py:
2751         * Scripts/webkitpy/common/system/executive_unittest.py:
2752         * Scripts/webkitpy/layout_tests/port/http_lock.py:
2753
2754 2010-11-02  Adam Roben  <aroben@apple.com>
2755
2756         Reduce our dependence on coreutils when running Python tests
2757
2758         This patch introduces versions of the standard echo and cat utilities
2759         implemented in Python. They are probably missing features of their
2760         coreutils equivalents, but they can do what's necessary for our Python
2761         tests. This is useful on Windows, where these utilities typically
2762         aren't available.
2763
2764         Fixes <http://webkit.org/b/48883> executive_unittest relies on echo
2765         and cat utilities from coreutils, which are not present on Windows
2766
2767         Reviewed by Eric Seidel.
2768
2769         * Scripts/webkitpy/common/system/executive_unittest.py: Changed to use
2770         our Python-based echo and cat.
2771
2772         * Scripts/webkitpy/common/system/fileutils.py: Added.
2773         (make_stdout_binary): On Windows, puts sys.stdout into binary mode so
2774         that \n won't be translated into \r\n. I couldn't think of a good way
2775         to test this directly without touching the filesystem, but it is tested
2776         indirectly by echo_unittest.
2777
2778         * Scripts/webkitpy/test/cat.py: Added.
2779         (command_arguments): Returns a list for invoking cat with the given arguments.
2780         (main): Acts like a simplified version of the coreutils cat utility.
2781
2782         * Scripts/webkitpy/test/cat_unittest.py: Added.
2783         (CatTest.assert_cat): Runs cat with the given input and ensures the
2784         output matches the input.
2785         (CatTest.test_basic): Performs a simple test of cat.
2786         (CatTest.test_no_newline): Tests what happens when the input string
2787         doesn't have a trailing newline.
2788         (CatTest.test_unicode): Tests passing a unicode string to cat.
2789         (CatTest.test_as_command): Tests running cat as a separate command.
2790
2791         * Scripts/webkitpy/test/echo.py: Added.
2792         (command_arguments): Returns a list for invoking echo with the given arguments.
2793         (main): Acts like a simplified version of the coreutils echo utility.
2794
2795         * Scripts/webkitpy/test/echo_unittest.py: Added.
2796         (EchoTest.test_basic): Performs a simple test of echo.
2797         (EchoTest.test_no_newline): Tests passing -n to echo to suppress the
2798         trailing newline.
2799         (EchoTest.test_unicode): Tests passing unicode and non-unicode strings
2800         to echo.
2801         (EchoTest.test_argument_order): Tests what happens when -n is not the
2802         first argument.
2803         (EchoTest.test_empty_arguments): Tests what happens when you pass [] to
2804         echo.main.
2805         (EchoTest.test_no_arguments): Tests what happens when you call
2806         echo.main with no arguments.
2807         (EchoTest.test_as_command): Tests running echo as a separate command.
2808
2809 2010-11-04  Renata Hodovan  <reni@inf.u-szeged.hu>
2810
2811         Unreviewed: Add myself to the list of Committers.
2812
2813         * Scripts/webkitpy/common/config/committers.py:
2814
2815 2010-11-04  Andreas Kling  <kling@webkit.org>
2816
2817         Reviewed by Kenneth Rohde Christiansen.
2818
2819         [Qt] Hook into QWebPage::scrollRequested for resizes-to-contents mode
2820
2821         In RTC mode the QGraphicsWebView item is the size of the contents,
2822         scrolling works a bit differently  (we need to react to scrollRequested.)
2823
2824         Normally QGraphicsView will replay the last mouse event when scrolling,
2825         so to prevent WebKit from getting confused by this we temporarily make
2826         the QGraphicsView non-interactive.
2827
2828         * QtTestBrowser/webview.cpp:
2829         (WebViewGraphicsBased::setPage):
2830         (WebViewGraphicsBased::scrollRequested):
2831         * QtTestBrowser/webview.h:
2832
2833 2010-11-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2834
2835         Reviewed by Andreas Kling.
2836
2837         Use OS(MAC_OS_X) rather than PLATFORM(MAC) in TestControllerQt
2838
2839         PLATFORM(MAC) is not defined for the Qt port, as it refers to
2840         the Mac-port, not the Mac OS X operating system.
2841
2842         * WebKitTestRunner/qt/TestControllerQt.cpp:
2843         (WTR::isExistingLibrary):
2844
2845 2010-11-04  Leonid Ebril  <leonid.ebril@nokia.com>
2846
2847         Reviewed by Andreas Kling.
2848
2849         [Qt] Adding iphone user agent string the user agent list for QtTestBrowser
2850         https://bugs.webkit.org/show_bug.cgi?id=48869
2851
2852         * QtTestBrowser/useragentlist.txt:
2853
2854 2010-11-03  Adam Roben  <aroben@apple.com>
2855
2856         Always use uppercase drive names in strings returned by abspath_to_uri
2857
2858         Some versions of cygpath use lowercase drive letters while others use
2859         uppercase, which makes it hard to test the output of code that uses
2860         cygpath.
2861
2862         Fixes <http://webkit.org/b/48914> webkitpy.common.system.path_unittest
2863         fails with Cygwin 1.5
2864
2865         Reviewed by Eric Seidel.
2866
2867         * Scripts/webkitpy/common/system/path.py:
2868         (cygpath): Updated the docstring to indicate that only absolute paths
2869         should be passed for now (though relative paths will work fine).
2870         (_Cygpath.convert): Upper-case the first letter of the converted Windows path.
2871
2872 2010-11-03  George Guo  <George.Guo@Nokia.com>
2873
2874         Reviewed by Andreas Kling.
2875
2876         [Qt] QtTestBrowser : set mmp rule pageddata in Symbian
2877         https://bugs.webkit.org/show_bug.cgi?id=48767
2878     
2879         Paging is needd on Symbian devices to support benchmarks tests like
2880         dromaeo.com and Celtic Kane that need a lot of memory to run  
2881
2882         * QtTestBrowser/QtTestBrowser.pro:
2883
2884 2010-11-03  Jenn Braithwaite  <jennb@chromium.org>
2885
2886         Reviewed by Dmitry Titov.
2887
2888         Chromium: Update resource tracking when moving a frame between documents
2889         https://bugs.webkit.org/show_bug.cgi?id=48363
2890
2891         * DumpRenderTree/chromium/WebViewHost.cpp:
2892         (WebViewHost::assignIdentifierToRequest):
2893         Always put resource id in map so we can make assumptions about its
2894         presence.
2895         (WebViewHost::removeIdentifierForRequest):
2896         Added.
2897         (WebViewHost::didFinishResourceLoad):
2898         (WebViewHost::didFailResourceLoad):
2899         Check existence of resource id before removing from map.
2900         * DumpRenderTree/chromium/WebViewHost.h:
2901
2902 2010-11-03  Victor Wang  <victorw@chromium.org>
2903
2904         Reviewed by Adam Barth.
2905
2906         [Chromium] update buildbot names in chromium rebaseline tool.
2907
2908         https://bugs.webkit.org/show_bug.cgi?id=48881
2909
2910         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2911
2912 2010-11-03  Darin Adler  <darin@apple.com>
2913
2914         Updated Xcode projects by opening them with Xcode 3.2.4.
2915         Updated svn:ignore for Xcode projects.
2916
2917         * MiniBrowser/MiniBrowser.xcodeproj: Added property svn:ignore.
2918         * TestWebKitAPI/TestWebKitAPI.xcodeproj: Added property svn:ignore.
2919         * WebKitLauncher/WebKitLauncher.xcodeproj: Modified property svn:ignore.
2920         * WebKitLauncher/WebKitLauncher.xcodeproj/project.pbxproj: Updated with Xcode 3.2.4.
2921
2922 2010-11-03  Adam Roben  <aroben@apple.com>
2923
2924         Stop waiting for messages from the web process after a timeout elapses
2925
2926         TestController::runUntil now takes a timeout parameter that specifies
2927         how long to wait for the condition to become true. WebKitTestRunner
2928         takes a --timeout flag that specifies how long the timeout should be.
2929         run-webkit-tests passes this flag to WebKitTestRunner so its timeout
2930         will be similar to run-webkit-tests.
2931
2932         Fixes <http://webkit.org/b/43047> <rdar://problem/8365833>
2933         WebKitTestRunner waits forever if the web process crashes
2934
2935         Reviewed by Darin Adler and Anders Carlsson.
2936
2937         * Scripts/old-run-webkit-tests:
2938         (top level): Moved the GuardMalloc timeout adjustment here from
2939         readFromDumpToolWithTimer.
2940         (openDumpTool): Make WTR use a timeout similar to but slightly shorter
2941         than the one that was specified on the command line.
2942
2943         * WebKitTestRunner/TestController.cpp:
2944         (WTR::TestController::TestController): Initialize our timeout values.
2945         (WTR::TestController::initialize): Parse the --timeout flag and use it
2946         to modify our timeout values.
2947         (WTR::TestController::resetStateToConsistentValues): Changed to use a
2948         short timeout while waiting for the web process to reset and to return
2949         a boolean indicating whether we were able to reset the web process.
2950         Uses a 5-second timeout while waiting for the process to be reset.
2951         (WTR::TestController::runTest): Changed to return a boolean indicating
2952         whether we were able to reset the web process (and thus run the test).
2953         (WTR::TestController::runUntil): Call through to platformRunUntil.
2954
2955         (WTR::TestController::runTestingServerLoop):
2956         (WTR::TestController::run):
2957         Changed to bail if any test can't be run. This will cause the process
2958         to exit. (Unfortunately this will make run-webkit-tests think we
2959         crashed; see <http://webkit.org/b/48943>.)
2960
2961         * WebKitTestRunner/TestController.h: Added platformRunUntil,
2962         m_longTimeout, and m_shortTimeout.
2963
2964         * WebKitTestRunner/TestInvocation.cpp:
2965         (WTR::TestInvocation::invoke): Use a short timeout when waiting for the
2966         initial response and a long timeout when waiting for the test to
2967         complete. Check whether runUntil timed out and print an error message
2968         if so.
2969
2970         * WebKitTestRunner/mac/TestControllerMac.mm:
2971         (WTR::TestController::platformRunUntil): Renamed from runUntil. Pass
2972         [NSDate distantPast] to -[NSRunLoop runMode:beforeDate:] so that we
2973         won't block waiting for the run loop. Only loop until the timeout
2974         elapses.
2975
2976         * WebKitTestRunner/qt/TestControllerQt.cpp:
2977         (WTR::TestController::platformRunUntil): Renamed from runUntil. Added a
2978         FIXME about honoring the timeout.
2979
2980         * WebKitTestRunner/win/TestControllerWin.cpp:
2981         (WTR::TestController::platformRunUntil): Renamed from runUntil. Use
2982         ::MsgWaitForMultipleObjectsEx to implement the timeout. Changed to use
2983         ::PeekMessageW so that we don't block waiting for messages to become
2984         available.
2985
2986 2010-11-03  Adam Roben  <aroben@apple.com>
2987
2988         Add a plugin test that evaluates JS after removing the plugin element
2989         from the document
2990
2991         This test replaces platform/win/plugins/plugin-delayed-destroy.html.
2992         That test was made to prevent a crash very similar to this one, but
2993         unfortunately tested only the mechanism that prevented the crash and
2994         not whether the crash itself was prevented. Since WebKit2 uses a
2995         different mechanism to prevent the crash, the test was failing even
2996         though WebKit2 was not vulnerable to the crash. This new test crashes
2997         if there is no mechanism in place to prevent it and passes in both
2998         WebKit1 and WebKit2.
2999
3000         Fixes <http://webkit.org/b/46711> <rdar://problem/8485903>
3001         platform/win/plugins/plugin-delayed-destroy.html fails in WebKit2
3002
3003         Reviewed by Anders Carlsson.
3004
3005         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3006         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
3007         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
3008         * GNUmakefile.am:
3009         Added new file.
3010
3011         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
3012         (pluginDeallocate): Make sure we delete the PluginTest object. This
3013         prevents a leak and also allows us to test the crash.
3014
3015         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
3016         (PluginTest::executeScript): Made this into a non-static member
3017         function.
3018
3019         (PluginTest::waitUntilDone):
3020         (PluginTest::notifyDone):
3021         Updated for changes to executeScript.
3022
3023         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Added executeScript.
3024
3025         * DumpRenderTree/TestNetscapePlugIn/Tests/EvaluateJSAfterRemovingPluginElement.cpp: Added.
3026         (EvaluateJSAfterRemovingPluginElement::EvaluateJSAfterRemovingPluginElement):
3027         Initialize ourselves and tell the test harness to wait.
3028         (EvaluateJSAfterRemovingPluginElement::NPP_DestroyStream): Remove our
3029         plugin element from the document, then execute some JavaScript. If
3030         WebKit does not have appropriate mechanisms in place, we'll be
3031         destroyed inside the first call to executeScript and crash on the
3032         second call.
3033
3034 2010-11-02  Stephen White  <senorblanco@chromium.org>
3035
3036         Reviewed by Tony Chang.
3037
3038         [chromium] Fix LayoutTestController UMRs.
3039         https://bugs.webkit.org/show_bug.cgi?id=48872
3040
3041         * DumpRenderTree/chromium/LayoutTestController.cpp:
3042         (LayoutTestController::LayoutTestController):
3043
3044 2010-11-03  Kent Tamura  <tkent@chromium.org>
3045
3046         Reviewed by Dimitri Glazkov.
3047
3048         REGRESSION: rebaseline-chromium-webkit-tests uses non-zero tolerance for
3049         image dup detection
3050         https://bugs.webkit.org/show_bug.cgi?id=48744
3051
3052         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
3053          - Make a function for option parsing for ease of test
3054          - Set 0 to options.tolerance
3055         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
3056          - Add a test for this change
3057
3058 2010-11-02  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
3059
3060         Unreviewed: Add myself to the list of Committers.
3061
3062         * Scripts/webkitpy/common/config/committers.py:
3063
3064 2010-11-02  Anders Carlsson  <andersca@apple.com>
3065
3066         Fix build.
3067
3068         * TestWebKitAPI/PlatformUtilities.cpp:
3069         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
3070
3071 2010-11-02  Sam Weinig  <sam@webkit.org>
3072
3073         Reviewed by Anders Carlsson.
3074
3075         Add a way to send startup messages on the context which can be posted when a process launches
3076         <rdar://problem/8617928>
3077         https://bugs.webkit.org/show_bug.cgi?id=48838
3078
3079         * MiniBrowser/mac/WebBundle/WebBundleMain.m:
3080         (WKBundleInitialize):
3081         * TestWebKitAPI/InjectedBundleController.cpp:
3082         (TestWebKitAPI::InjectedBundleController::initialize):
3083         (TestWebKitAPI::InjectedBundleController::didReceiveMessage):
3084         * TestWebKitAPI/InjectedBundleController.h:
3085         * TestWebKitAPI/InjectedBundleMain.cpp:
3086         (WKBundleInitialize):
3087         * TestWebKitAPI/PlatformUtilities.cpp:
3088         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
3089         * WebKitTestRunner/InjectedBundle/InjectedBundleMain.cpp:
3090         (WKBundleInitialize):
3091         Update implementations of WKBundleInitialize to take an initial userData
3092         argument.  Change TestWebKitAPI to use the new initial userData to initialize
3093         each test's bundle.
3094
3095 2010-11-02  Benjamin Kalman  <kalman@google.com>
3096
3097         Reviewed by Ojan Vafai.
3098
3099         new-run-webkit-tests doesn't strip "LayoutTests/" from prefix, unlike old-run-webkit-tests
3100         https://bugs.webkit.org/show_bug.cgi?id=48794
3101
3102         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: Strip the "LayoutTests/" prefix from test argument paths.
3103
3104 2010-11-02  Adam Roben  <aroben@apple.com>
3105
3106         Skip webkitpy.layout_tests.run_webkit_tests_unittest.MainTest on Cygwin
3107         Python 2.5.x
3108
3109         It is known to hang on that version of Python. See
3110         <http://webkit.org/b/48614>.
3111
3112         Reviewed by Adam Barth.
3113
3114         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py: Use
3115         skip_if to skip MainTest on Cygwin Python 2.5.x.
3116
3117         * Scripts/webkitpy/test/skip.py: Added.
3118         (skip_if): If the passed-in condition is false, find all the test_*
3119         methods of the given class and replace them with a function that just
3120         logs that we're skipping these tests. This is loosely based on the
3121         unittest.skip_if decorator added in Python 3.1.
3122         (_skipped_method): Returns a function that just logs that the tests are
3123         being skipped. This is used in place of the actual test_* functions.
3124
3125         * Scripts/webkitpy/test/skip_unittest.py: Added.
3126         (SkipTest.setUp): Create a logger for ourselves and give it a handler
3127         that logs to a private stream.
3128         (SkipTest.tearDown): Reset the logger.
3129         (SkipTest.create_fixture_class): Returns a class that we use to test
3130         skip_if. It has a single test_* method, test_foo, that just calls
3131         through to a callback.
3132         (SkipTest.foo_callback): Record that test_foo was called.
3133         (SkipTest.test_skip_if_false): Pass skip_if a False condition and test
3134         that test_foo does get called.
3135         (SkipTest.test_skip_if_true): Pass skip_if a True condition and test
3136         that test_foo does not get called and the appropriate message gets
3137         logged.
3138
3139 2010-11-02  Adam Barth  <abarth@webkit.org>
3140
3141         Reviewed by Eric Seidel.
3142
3143         webkit-patch should tell check-webkit-style which files were changed so
3144         check-webkit-style doesn't have to stat the whole working copy again
3145         https://bugs.webkit.org/show_bug.cgi?id=48792
3146
3147         * Scripts/webkitpy/tool/mocktool.py:
3148         * Scripts/webkitpy/tool/steps/checkstyle.py:
3149
3150 2010-11-02  Robert Kroeger  <rjkroege@chromium.org>
3151
3152         Reviewed by James Robinson.
3153
3154         [Chromium/DRT] Make EventSendingController honour leapForward for touch events.
3155         https://bugs.webkit.org/show_bug.cgi?id=48777
3156
3157         * DumpRenderTree/chromium/EventSender.cpp:
3158         (EventSender::sendCurrentTouchEvent):
3159
3160 2010-11-02  Adam Roben  <aroben@apple.com>
3161
3162         Only track resource identifiers in DRT when dumpResourceLoadCallbacks
3163         is on
3164
3165         This reverts Windows to our pre-r71097 behavior. That patch made us
3166         track all resource identifiers, including the main resource, so the
3167         main resource's URL started appearing in test output instead of
3168         "<unknown>". Arguably having the main resource's URL is better, but all
3169         other platforms print "<unknown>" and we want to match.
3170
3171         Fixes <http://webkit.org/b/48837> <rdar://problem/8620351> REGRESSION
3172         (r71097): Many http tests failing on Windows
3173
3174         Reviewed by Anders Carlsson.
3175
3176         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
3177         (ResourceLoadDelegate::identifierForInitialRequest): Don't add the
3178         identifier to the URL map if we're not supposed to dump resource load
3179         callbacks.
3180         (ResourceLoadDelegate::removeIdentifierForRequest): Always remove the
3181         identifier from the URL map even if we're already "done". There's no
3182         point in keeping out-of-date identifiers around.
3183
3184 2010-11-01  Jenn Braithwaite  <jennb@chromium.org>
3185
3186         Reviewed by Adam Roben.
3187
3188         Windows: Update resource tracking when moving a frame between documents
3189         https://bugs.webkit.org/show_bug.cgi?id=48364
3190
3191         * DumpRenderTree/win/DumpRenderTree.cpp:
3192         (createWebViewAndOffscreenWindow):
3193         (main):
3194         Give each WebView its own ResourceLoadDelegate instance in order to
3195         make assertions about resource ids on a particular WebView.
3196         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
3197         (ResourceLoadDelegate::identifierForInitialRequest):
3198         Always add id to the map.
3199         (ResourceLoadDelegate::removeIdentifierForRequest):
3200         Added.
3201         (ResourceLoadDelegate::willSendRequest):
3202         (ResourceLoadDelegate::didReceiveAuthenticationChallenge):
3203         (ResourceLoadDelegate::didReceiveResponse):
3204         (ResourceLoadDelegate::didFinishLoadingFromDataSource):
3205         (ResourceLoadDelegate::didFailLoadingWithError):
3206         (ResourceLoadDelegate::descriptionSuitableForTestResult):
3207         Replace static descriptionSuitableForTestResult with static member function to access identifier map.
3208         * DumpRenderTree/win/ResourceLoadDelegate.h:
3209         (ResourceLoadDelegate::urlMap):
3210         Moved within class so that each WebView has its own id map.
3211
3212 2010-11-01  Ojan Vafai  <ojan@chromium.org>
3213
3214         Reviewed by Dimitri Glazkov.
3215
3216         remove debug code from run_webkit_tests.py
3217         https://bugs.webkit.org/show_bug.cgi?id=48800
3218
3219         Remove temporary debug code and make --master-name required
3220         if --test-results-server is set now that all clients set
3221         --master-name.
3222
3223         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3224
3225 2010-11-01  Tony Chang  <tony@chromium.org>
3226
3227         Reviewed by Kent Tamura.
3228
3229         add plugin event logging for linux
3230         https://bugs.webkit.org/show_bug.cgi?id=48779
3231
3232         This is taken from Chromium's fork of the layout test plugin:
3233         http://git.chromium.org/gitweb/?p=chromium.git/.git;a=blob;f=webkit/tools/npapi_layout_test_plugin/main.cpp;h=3ebdada2f049b3624756438cff852364f86a2ede;hb=HEAD#l348
3234
3235         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
3236         (handleEventX11):
3237         (NPP_HandleEvent):
3238
3239 2010-11-01  Dirk Pranke  <dpranke@chromium.org>
3240
3241         Reviewed by James Robinson.
3242
3243         new-run-webkit-tests: use DRT, child-processes=1 for GPU tests by default
3244         https://bugs.webkit.org/show_bug.cgi?id=48790
3245
3246         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
3247         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
3248
3249 2010-11-01  Adam Barth  <abarth@webkit.org>
3250
3251         Reviewed by Eric Seidel.
3252
3253         Teach check-webkit-style how to accept a list of files to diff on the
3254         command line
3255         https://bugs.webkit.org/show_bug.cgi?id=48784
3256
3257         In a future patch, webkit-patch will use this option to improve
3258         performance.  I'm landing this in two pieces to avoid causing a version
3259         skew problem for the style-bot.
3260
3261         * Scripts/check-webkit-style:
3262         * Scripts/webkitpy/style/optparser.py:
3263         * Scripts/webkitpy/style/optparser_unittest.py:
3264         * Scripts/webkitpy/style_references.py:
3265
3266 2010-11-01  Anders Carlsson  <andersca@apple.com>
3267
3268         Reviewed by John Sullivan.
3269
3270         Tear down the related WebProcessProxy when a WebContext is deallocated
3271         https://bugs.webkit.org/show_bug.cgi?id=48769
3272
3273         * TestWebKitAPI/Tests/WebKit2/FailedLoad.cpp:
3274         (TestWebKitAPI::didFailProvisionalLoadWithErrorForFrame):
3275         We don't support empty URLs anymore, update test to expect a null URL instead.
3276
3277 2010-11-01  Søren Gjesse  <sgjesse@chromium.org>
3278
3279         Reviewed by Andreas Kling.
3280
3281         Fix warning when compiling the chromium port of DumpRenderShell
3282         with clang.
3283         https://bugs.webkit.org/show_bug.cgi?id=48414
3284
3285         * DumpRenderTree/chromium/TestShell.h:
3286         (TestShell::javaScriptFlagsForLoad):
3287
3288 2010-11-01  Adam Roben  <aroben@apple.com>
3289
3290         Fix typo from r71022
3291
3292         * BuildSlaveSupport/build.webkit.org-config/config.json:
3293
3294 2010-11-01  Adam Roben  <aroben@apple.com>
3295
3296         Trigger the Windows Release WebKit2 tests when a Release build
3297         finishes, not when a Debug build finishes
3298
3299         Fixes <http://webkit.org/b/48754> Windows Release WebKit2 tests are
3300         triggered at the wrong time
3301
3302         Reviewed by Sam Weinig.
3303
3304         * BuildSlaveSupport/build.webkit.org-config/config.json: Fixed
3305         triggerable name and trigger.
3306
3307 2010-11-01  Mario Sanchez Prada  <msanchez@igalia.com>
3308
3309         Unreviewed. Adding my IRC nickname to the list of committers.
3310
3311         * Scripts/webkitpy/common/config/committers.py:
3312
3313 2010-10-31  Robert Hogan  <robert@webkit.org>
3314
3315         Reviewed by Antonio Gomes.
3316
3317         [Qt] [Gtk] Plug-ins having upper case in mime type are failing to load
3318
3319         Qt and Gtk are case-sensitive when storing the declared mime-type
3320         of plugins. Since plugin mime-types are lowercased prior to searching
3321         for them in the plugin database, ensure they are loaded with the
3322         mime-type in lower case too.
3323
3324         Change the test netscape plugin to declare its mimetype in sentence
3325         case so that the correct behaviour is enforced.
3326
3327         https://bugs.webkit.org/show_bug.cgi?id=36815
3328
3329         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
3330         (NP_GetMIMEDescription):
3331
3332 2010-10-31  Robert Hogan  <robert@webkit.org>
3333
3334         Reviewed by Antonio Gomes.
3335
3336         [Qt] Support nodesFromRect in DRT
3337
3338         https://bugs.webkit.org/show_bug.cgi?id=48716
3339
3340         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3341         (LayoutTestController::nodesFromRect):
3342         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3343
3344 2010-10-30  Andreas Kling  <kling@webkit.org>
3345
3346         Reviewed by Antonio Gomes.
3347
3348         [Qt] QtTestBrowser: Switching view type moves the embedded inspector
3349         https://bugs.webkit.org/show_bug.cgi?id=48705
3350
3351         Reinsert the embedded inspector into the splitter after changing
3352         between QWebView/QGraphicsWebView.
3353
3354         * QtTestBrowser/launcherwindow.cpp:
3355         (LauncherWindow::init):
3356         (LauncherWindow::initializeView):
3357         * QtTestBrowser/webinspector.h:
3358         (WebInspector::WebInspector):
3359
3360 2010-10-28  Antonio Gomes  <agomes@rim.com>
3361
3362         Reviewed by Ojan Vafai.
3363
3364         Needs a "LinuxEditingBehavior", perhaps with a better name
3365         https://bugs.webkit.org/show_bug.cgi?id=36627
3366
3367         Adding support to Mac's, GTK+'s, Windows' and Chromium's LayoutTestController class to test the newly introduced Unix editing behavior.
3368
3369         * DumpRenderTree/chromium/LayoutTestController.cpp:
3370         (LayoutTestController::setEditingBehavior):
3371         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3372         (LayoutTestController::setEditingBehavior):
3373         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3374         (LayoutTestController::setEditingBehavior):
3375         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3376         (LayoutTestController::setEditingBehavior):
3377
3378 2010-10-29  Andreas Kling  <kling@webkit.org>
3379
3380         Reviewed by Kenneth Rohde Christiansen.
3381
3382         [Qt] QtTestBrowser: Add keyboard shortcut to toggle full screen (F11)
3383         https://bugs.webkit.org/show_bug.cgi?id=48695
3384
3385         * QtTestBrowser/launcherwindow.cpp:
3386         (LauncherWindow::createChrome):
3387
3388 2010-10-29  Andreas Kling  <kling@webkit.org>
3389
3390         Reviewed by Kenneth Rohde Christiansen.
3391
3392         [Qt] QtTestBrowser: Fix uninitialized read in FpsTimer
3393         https://bugs.webkit.org/show_bug.cgi?id=48675
3394
3395         FpsTimer::m_timer was never initialized and passed to QObject::killTimer()
3396         on startup with in -graphicsbased mode.
3397
3398         * QtTestBrowser/fpstimer.cpp:
3399         (FpsTimer::FpsTimer):
3400         (FpsTimer::numFrames):
3401         (FpsTimer::stop):
3402         (FpsTimer::timerEvent):
3403         * QtTestBrowser/fpstimer.h:
3404
3405 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
3406
3407         Reviewed by Tony Chang.
3408
3409         new-run-webkit-tests: change TestResults to be serializable
3410
3411         In preparation for changing new-run-webkit-tests from
3412         multithreaded to multiprocess, we need to make sure the data
3413         going between the threads is easily serialized over a socket.
3414
3415         This change adds serialization/pickling for the TestResults and
3416         TestFailure objects (using cPickle).
3417
3418         The TestFailure objects included a "has_wdiff" flag for Text
3419         results, but the flag wasn't being used, so I've removed it,
3420         simplifying the state to basically a set of enum objects with
3421         associated methods.
3422
3423         https://bugs.webkit.org/show_bug.cgi?id=48616
3424
3425         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3426         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
3427         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
3428         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
3429         * Scripts/webkitpy/layout_tests/layout_package/test_results.py: Added.
3430         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:Added. 
3431         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3432         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
3433
3434 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
3435
3436         Reviewed by Ojan Vafai.
3437
3438         new-run-webkit-tests: need to provide separate dashboard results for GPU tests
3439         https://bugs.webkit.org/show_bug.cgi?id=48687
3440
3441         Modify the steps to upload the JSON files to the dashboards so
3442         that the GPU tests don't conflict with the regular tests on a
3443         port. We do this by modifying the --builder-name parameter to
3444         append " - GPU", which should cause the app to treat the results
3445         as a completely new builder. This is a little non-obvious, but
3446         keeps us from having to restructure the app.
3447
3448         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
3449         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
3450         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3451
3452 2010-10-29  Ojan Vafai  <ojan@chromium.org>
3453
3454         Reviewed by Tony Chang.
3455
3456         [chromium] add debug logging to help diagnose flakiness dashboard issues
3457         https://bugs.webkit.org/show_bug.cgi?id=48657
3458
3459         The appengine app thinks it's getting empty files uploaded. Add some logging
3460         to see if new-run-webkit-tests agrees.
3461         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3462
3463 2010-10-29  Dimitri Glazkov  <dglazkov@chromium.org>
3464
3465         Reviewed by Ojan Vafai.
3466
3467         [Chromium/DRT] Add master-name flag to new-run-webkit-tests.
3468         https://bugs.webkit.org/show_bug.cgi?id=48649
3469
3470         * BuildSlaveSupport/build.webkit.org-config/master.cfg: Added flag.
3471
3472 2010-10-29  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
3473
3474         Reviewed by Martin Robinson.
3475
3476         Enable popup window in GtkLauncher
3477         https://bugs.webkit.org/show_bug.cgi?id=48335
3478
3479         The GtkLauncher application does not support opening new window when
3480         a link with "target=_blank" is clicked or similar call to
3481         window.open(). Instead, GtkLauncher does nothing which breaks
3482         navigation of some websites.
3483
3484         * GtkLauncher/main.c:
3485         (activate_uri_entry_cb):
3486         (update_title):
3487         (link_hover_cb):
3488         (notify_title_cb):
3489         (notify_load_status_cb):
3490         (notify_progress_cb):
3491         (destroy_cb):
3492         (go_back_cb):
3493         (go_forward_cb):
3494         (create_web_view_cb):
3495         (web_view_ready_cb):
3496         (close_web_view_cb):
3497         (create_browser):
3498         (create_statusbar):
3499         (create_toolbar):
3500         (create_window):
3501         (main):
3502
3503 2010-10-29  Adam Roben  <aroben@apple.com>
3504
3505         Teach check-webkit-style about WebKit2's idiosyncracies
3506
3507         Fixes <http://webkit.org/b/48638> Style bot complains about a number
3508         of WebKit2 conventions
3509
3510         Reviewed by Anders Carlsson.
3511
3512         * Scripts/webkitpy/style/checker.py: Excluded some rules for various
3513         WebKit2-related files. Also updated the excluded rules for
3514         WebKitAPITest to match the current code.
3515
3516 2010-10-29  Csaba Osztrogonác  <ossy@webkit.org>
3517
3518         Reviewed by Adam Roben and David Kilzer.
3519
3520         Fix and cleanup of build systems
3521         https://bugs.webkit.org/show_bug.cgi?id=48342
3522
3523         * Scripts/build-webkit:
3524          - Remove unnecessary ENABLE_SANDBOX option.
3525          - Add ENABLE_FULLSCREEN_API option.
3526
3527 2010-10-28  Adam Roben  <aroben@apple.com>
3528
3529         Switch the Windows WebKit2 bot to the Release configuration
3530
3531         We only have one machine testing WebKit2 on Windows right now, and
3532         Debug is just too slow for it to keep up.
3533
3534         Fixes (hopefully!) <http://webkit.org/b/48615> Windows WebKit2 bot is
3535         always way behind
3536
3537         Reviewed by Jon Honeycutt.
3538
3539         * BuildSlaveSupport/build.webkit.org-config/config.json:
3540
3541 2010-10-29  Leandro Gracia Gil  <leandrogracia@google.com>
3542
3543         Reviewed by Jeremy Orlow.
3544
3545         Added a second parameter to setMockSpeechInputResult for
3546         the language used in speech input.
3547         https://bugs.webkit.org/show_bug.cgi?id=47089
3548
3549         * DumpRenderTree/LayoutTestController.cpp:
3550         (setMockSpeechInputResultCallback):
3551         * DumpRenderTree/LayoutTestController.h:
3552         * DumpRenderTree/chromium/LayoutTestController.cpp:
3553         (LayoutTestController::setMockSpeechInputResult):
3554         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3555         (LayoutTestController::setMockSpeechInputResult):
3556         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3557         (LayoutTestController::setMockSpeechInputResult):
3558         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3559         (LayoutTestController::setMockSpeechInputResult):
3560         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3561         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3562         (LayoutTestController::setMockSpeechInputResult):
3563         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3564         (LayoutTestController::setMockSpeechInputResult):
3565         (LayoutTestController::pageSizeAndMarginsInPixels):
3566
3567 2010-10-28  Ojan Vafai  <ojan@chromium.org>
3568
3569         Reviewed by Tony Chang.
3570
3571         [chromium] add a result-small.json file for the test dashboard
3572         https://bugs.webkit.org/show_bug.cgi?id=48547
3573
3574         Output both a results.json file and a results-small.json file.
3575         The dashboard will load results-small.json by default so it loads faster.
3576
3577         * TestResultServer/model/jsonresults.py:
3578         * TestResultServer/model/jsonresults_unittest.py:
3579         Added a bunch of sys.path hackery. Unforunately, this uses hardcoded
3580         paths. That obviously needs to be fixed, but at least this way it
3581         clearly documents what paths are necessary.
3582
3583 2010-10-28  Eric Seidel  <eric@webkit.org>
3584
3585         Reviewed by Adam Barth.
3586