2010-11-24 Dirk Pranke <dpranke@chromium.org>
[WebKit.git] / WebKitTools / ChangeLog
1 2010-11-24  Dirk Pranke  <dpranke@chromium.org>
2
3         Reviewed by Tony Chang.
4
5         This patch implements the first part of the manager side of the
6         Broker objects - it handles creating threads, waiting for them
7         to complete, and running a single-threaded loop as well.
8
9         https://bugs.webkit.org/show_bug.cgi?id=49779
10
11         * Scripts/webkitpy/layout_tests/layout_package/message_broker.py:
12         * Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py:
13         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
14         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
15
16 2010-11-24  Mihai Parparita  <mihaip@chromium.org>
17
18         Reviewed by David Levin.
19
20         DRT should not generate pixel output for text/plain tests
21         https://bugs.webkit.org/show_bug.cgi?id=50002
22         
23         Modify Mac, Windows and GTK DRT implementations to disable pixel output
24         when encountering a text/plain response (the Chromium one already does
25         this, and none of the other ports special-case text/plain output in
26         the first place).
27
28         * DumpRenderTree/gtk/DumpRenderTree.cpp:
29         (dump):
30         * DumpRenderTree/mac/DumpRenderTree.mm:
31         (dump):
32         * DumpRenderTree/win/DumpRenderTree.cpp:
33         (dump):
34
35 2010-11-24  Andras Becsi  <abecsi@inf.u-szeged.hu>
36
37         Unreviewed.
38
39         Add my other email address to committers.py.
40
41         * Scripts/webkitpy/common/config/committers.py:
42
43 2010-11-24  Andras Becsi  <abecsi@inf.u-szeged.hu>
44
45         Reviewed by Csaba Osztrogonác.
46
47         Make build-webkit --minimal build with ENABLE_INSPECTOR=0
48         https://bugs.webkit.org/show_bug.cgi?id=49975
49
50         * Scripts/build-webkit: Add ENABLE_INSPECTOR define to features,
51         to enable the script to define it 0 in case of a minimal build.
52
53 2010-11-24  Martin Robinson  <mrobinson@igalia.com>
54
55         Reviewed by Xan Lopez.
56
57         [GTK] The WebKitWebView should expose a set of signals encapsulating the behavior of the EditorClient
58         https://bugs.webkit.org/show_bug.cgi?id=49143
59
60         * DumpRenderTree/gtk/DumpRenderTree.cpp:
61         (createWebView): Call the new connectEditingCallbacks function.
62         * DumpRenderTree/gtk/EditingCallbacks.cpp: Added.
63         (dumpNodePath):
64         (dumpRange):
65         (insertActionString):
66         (selectionAffinityString):
67         (shouldBeginEditing):
68         (shouldEndEditing):
69         (shouldInsertNode):
70         (shouldInsertText):
71         (shouldDeleteRange):
72         (shouldShowDeleteInterfaceForElement):
73         (shouldChangeSelectedRange):
74         (shouldApplyStyle):
75         (editingBegan):
76         (userChangedContents):
77         (editingEnded):
78         (selectionChanged):
79         (connectEditingCallbacks):
80         * DumpRenderTree/gtk/EditingCallbacks.h: Added.
81         * GNUmakefile.am: Added EditingCallbacks.{h,cpp} to the source list.
82
83 2010-11-24  Yi Shen  <yi.4.shen@nokia.com>
84
85         Reviewed by Andreas Kling.
86
87         [Qt] Make QtTestBrowser remember the selected user agent
88         https://bugs.webkit.org/show_bug.cgi?id=50021
89
90         Set user agent for a new page by using the old page's
91
92         * QtTestBrowser/mainwindow.cpp:
93         (MainWindow::setPage):
94
95 2010-11-24 Patrick Gansterer <paroga@webkit.org>
96
97         Reviewed by Csaba Osztrogonác.
98
99         Remove Bakefile build system files
100         https://bugs.webkit.org/show_bug.cgi?id=49983
101
102         r53757 only removed the content, but not the files.
103         This patch removes that empty files.
104
105         * DumpRenderTree/wx/DumpRenderTree.bkl: Removed.
106         * wx/browser/browser.bkl: Removed.
107         * wx/build-wxwebkit: Removed.
108
109 2010-11-23  Dirk Pranke  <dpranke@chromium.org>
110
111         Reviewed by Tony Chang.
112
113         This patch cleans up the logic used to shard tests into groups a
114         bit and adds the --worker-model flag to NRWT. The flag is only
115         used at the moment to control whether to run single-threaded or
116         not, but eventually will also allow toggling between threads and
117         processes.
118
119         Also add a minor cleanup with _test_is_slow(), which just
120         eliminates some repetition and gives slightly better encapsulation.
121
122         https://bugs.webkit.org/show_bug.cgi?id=49773
123
124         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
125         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
126
127 2010-11-23  Mihai Parparita  <mihaip@chromium.org>
128
129         Reviewed by Tony Chang.
130
131         Rebaseline server: list current baselines and platforms
132         https://bugs.webkit.org/show_bug.cgi?id=49991
133         
134         List existing baselines (with Trac links) for tests.
135         
136         Add dropdowns for choosing with platform baselines to target (and what
137         to do with existing ones).
138
139         Also fix MockFileSystem.join to behave more like os.path.join (unit
140         test was ending up with two consecutive slashes in a layout test
141         path).        
142
143         * Scripts/webkitpy/common/system/filesystem_mock.py:
144         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
145         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
146         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
147         * Scripts/webkitpy/tool/commands/data/rebaselineserver/queue.js:
148         * Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js:
149         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
150         * Scripts/webkitpy/tool/commands/rebaselineserver_unittest.py: Added.
151
152 2010-11-23  Ojan Vafai  <ojan@chromium.org>
153
154         Reviewed by Tony Chang.
155
156         remove timeout argument to urlopen
157         https://bugs.webkit.org/show_bug.cgi?id=49995
158
159         Apparently the version of python we have on the bots doesn't accept a timeout argument.
160
161         * Scripts/webkitpy/layout_tests/layout_package/test_results_uploader.py:
162
163 2010-11-23  Sheriff Bot  <webkit.review.bot@gmail.com>
164
165         Unreviewed, rolling out r72628.
166         http://trac.webkit.org/changeset/72628
167         https://bugs.webkit.org/show_bug.cgi?id=49994
168
169         This patch is causing layout-test failtures on GTK Linux
170         64-bit Debug (Requested by ctguil on #webkit).
171
172         * DumpRenderTree/chromium/AccessibilityUIElement.cpp:
173         (AccessibilityUIElement::AccessibilityUIElement):
174         (AccessibilityUIElement::parentElementCallback):
175         * DumpRenderTree/chromium/AccessibilityUIElement.h:
176         * DumpRenderTree/chromium/CppBoundClass.cpp:
177         * DumpRenderTree/chromium/CppBoundClass.h:
178
179 2010-11-22  Ojan Vafai  <ojan@chromium.org>
180
181         Reviewed by Tony Chang.
182
183         speculative fix for upload errors: stop using mechanize to upload to test-results.appspot.com
184         https://bugs.webkit.org/show_bug.cgi?id=49944
185
186         * Scripts/webkitpy/layout_tests/layout_package/test_results_uploader.py:
187
188 2010-11-23  Chris Guillory  <chris.guillory@google.com>
189
190         Reviewed by Dimitri Glazkov.
191
192         Include the FrameView widget of a RenderWidget in the accessibility tree.
193         https://bugs.webkit.org/show_bug.cgi?id=49106
194
195         * DumpRenderTree/chromium/AccessibilityUIElement.cpp:
196         (AccessibilityUIElement::AccessibilityUIElement):
197         (AccessibilityUIElement::parentElementCallback):
198         (AccessibilityUIElement::isEqualCallback):
199         * DumpRenderTree/chromium/AccessibilityUIElement.h:
200         * DumpRenderTree/chromium/CppBoundClass.cpp:
201         (CppBoundClass::getFromCppVariant):
202         * DumpRenderTree/chromium/CppBoundClass.h:
203
204 2010-11-23  Alex Grilo  <abgrilo@profusion.mobi>
205
206         Reviewed by Kenneth Rohde Christiansen.
207
208         [EFL] Add tiled backing store to EWebLauncher
209         https://bugs.webkit.org/show_bug.cgi?id=45777
210
211         Allow EWebLauncher to choose between single and tiled backing store
212
213         * EWebLauncher/main.c:
214         (on_key_down):
215         (browserCreate):
216         (main):
217
218 2010-11-23  Simon Hausmann  <simon.hausmann@nokia.com>
219
220         Reviewed by Tor Arne Vestbø.
221
222         [Qt] Review the setUserPermission & friends API
223         https://bugs.webkit.org/show_bug.cgi?id=46810
224
225         Renamed requestPermissionFromUser to featurePermissionRequested
226         and cancelRequestsForPermission to featurePermissionRequestCanceled.
227
228         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
229         (WebCore::WebPage::WebPage):
230         * QtTestBrowser/webpage.cpp:
231         (WebPage::WebPage):
232         (WebPage::featurePermissionRequestCanceled):
233         * QtTestBrowser/webpage.h:
234
235 2010-11-23  Simon Hausmann  <simon.hausmann@nokia.com>
236
237         Reviewed by Tor Arne Vestbø.
238
239         [Qt] Review the setUserPermission & friends API
240         https://bugs.webkit.org/show_bug.cgi?id=46810
241
242         Rename setUserPermission to setFeaturePermission
243
244         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
245         (WebCore::WebPage::requestPermission):
246         (WebCore::WebPage::permissionSet):
247         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
248         (LayoutTestController::grantDesktopNotificationPermission):
249         * QtTestBrowser/webpage.cpp:
250         (WebPage::requestPermission):
251
252 2010-11-23  Simon Hausmann  <simon.hausmann@nokia.com>
253
254         Reviewed by Tor Arne Vestbø.
255
256         [Qt] Review the setUserPermission & friends API
257         https://bugs.webkit.org/show_bug.cgi?id=46810
258
259         Rename PermissionDomain to Feature, NotificationsPermissionDomain to Notifications
260         and GeolocationPermissionDomain to Geolocation.
261
262         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
263         (WebCore::WebPage::WebPage):
264         (WebCore::WebPage::requestPermission):
265         (WebCore::WebPage::cancelPermission):
266         (WebCore::WebPage::permissionSet):
267         (WebCore::DumpRenderTree::geolocationPermissionSet):
268         * DumpRenderTree/qt/DumpRenderTreeQt.h:
269         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
270         (LayoutTestController::grantDesktopNotificationPermission):
271         * QtTestBrowser/webpage.cpp:
272         (WebPage::WebPage):
273         (WebPage::requestPermission):
274         (WebPage::cancelRequestsForPermission):
275         * QtTestBrowser/webpage.h:
276
277 2010-11-23  Simon Hausmann  <simon.hausmann@nokia.com>
278
279         Reviewed by Tor Arne Vestbø.
280
281         [Qt] Review the setUserPermission & friends API
282         https://bugs.webkit.org/show_bug.cgi?id=46810
283
284         Add a ByUser suffix to PermissionGranted/Denied. In the future
285         we can add PermissionGrantedByDefault.
286
287
288         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
289         (WebCore::WebPage::requestPermission):
290         (WebCore::WebPage::permissionSet):
291         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
292         (LayoutTestController::grantDesktopNotificationPermission):
293         * QtTestBrowser/webpage.cpp:
294         (WebPage::requestPermission):
295
296 2010-11-18  Satish Sampath  <satish@chromium.org>
297
298         Reviewed by Jeremy Orlow.
299
300         For speech input event, send an event object containing all the recognition results and metadata.
301         https://bugs.webkit.org/show_bug.cgi?id=49736
302
303         Updated DRT by renaming the mock result method on all platforms to the new signature.
304
305         * DumpRenderTree/LayoutTestController.cpp:
306         (addMockSpeechInputResultCallback):
307         (LayoutTestController::staticFunctions):
308         * DumpRenderTree/LayoutTestController.h:
309         * DumpRenderTree/chromium/LayoutTestController.cpp:
310         (LayoutTestController::LayoutTestController):
311         (LayoutTestController::addMockSpeechInputResult):
312         * DumpRenderTree/chromium/LayoutTestController.h:
313         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
314         (LayoutTestController::addMockSpeechInputResult):
315         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
316         (LayoutTestController::addMockSpeechInputResult):
317         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
318         (LayoutTestController::addMockSpeechInputResult):
319         * DumpRenderTree/qt/LayoutTestControllerQt.h:
320         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
321         (LayoutTestController::addMockSpeechInputResult):
322         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
323         (LayoutTestController::addMockSpeechInputResult):
324
325 2010-11-22  Dirk Pranke  <dpranke@chromium.org>
326
327         Reviewed by Tony Chang.
328
329         Fix some spacing issues with log messages with
330         new-run-webkit-tests --verbose.
331
332         https://bugs.webkit.org/show_bug.cgi?id=49936
333
334         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
335
336 2010-11-22  Ojan Vafai  <ojan@chromium.org>
337
338         Reviewed by Adam Barth.
339
340         update scm to work with different remote branches
341         https://bugs.webkit.org/show_bug.cgi?id=49949
342
343         This works if there are multiple svn tracking branches
344         and/or if the tracking branch is refs/remots/origin/master
345         instead of refs/remotes/trunk which is the direction we're trying
346         to head since that makes fetches faster.
347
348         * Scripts/webkitpy/common/checkout/scm.py:
349         * Scripts/webkitpy/common/checkout/scm_unittest.py:
350
351 2010-11-22  Hayato Ito  <hayato@chromium.org>
352
353         Reviewed by Shinichiro Hamaji.
354
355         Ignore reference files which will be used by reftests when collecting
356         test cases.
357         https://bugs.webkit.org/show_bug.cgi?id=49835
358
359         * Scripts/webkitpy/layout_tests/port/test_files.py:
360         * Scripts/webkitpy/layout_tests/port/test_files_unittest.py:
361
362 2010-11-22  Adam Roben  <aroben@apple.com>
363
364         Use paths relative to $WebKitVSPropsRedirectionDir to access shared .vsprops files
365
366         Apple's Windows build allows placing header files and import libraries for WebKit's
367         dependencies (CoreGraphics, CFNetwork, SQLite, etc.) outside the source tree via the
368         $WebKitLibrariesDir environment variable. This is both required for production builds and
369         convenient for Apple-internal developer builds. Apple's production builds also require that
370         WebKit's shared .vsprops files be accessed relative to $WebKitLibrariesDir. In production
371         builds, the files are copied into that directory tree by the
372         WebKitLibraries/win/tools/WinTools.make file. In Apple-internal developer builds, the
373         copying is done by
374         JavaScriptCore/JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.make.
375
376         This .vsprops copying is problematic in one very important case: when a developer updates
377         their source tree and then tries to build. Visual Studio only reads .vsprops files when a
378         project is first loaded. So, when Visual Studio is first opened after the .vsprops files are
379         updated, it reads in the old files that were already residing in $WebKitLibrariesDir. When a
380         build is started, JavaScriptCoreGenerated.make copies the new .vsprops files into
381         $WebKitLibrariesDir, but Visual Studio will not pick up the changes.  The rest of the build
382         will proceed with out-of-date .vsprops files, which will likely result in a build failure.
383
384         To fix this, we now use normal relative paths to access the .vsprops files in the source
385         tree rather than in $WebKitLibrariesDir, but prefix those paths with a new environment
386         variable, $WebKitVSPropsRedirectionDir. In developer builds, this environment variable is
387         unset, so the normal relative paths are used to read the .vsprops files out of the source
388         tree directly. In production builds, this environment variable is set to a fake directory
389         that will cause the .vsprops files in $WebKitLibrariesDir to be found when the relative path
390         is resolved.
391         
392         For example, JavaScriptCore.vcproj uses this path for FeatureDefines.vsprops:
393
394         $(WebKitVSPropsRedirectionDir)..\..\..\WebKitLibraries\win\tools\vsprops\FeatureDefines.vsprops
395
396         In developer builds, where $WebKitVSPropsRedirectionDir is unset, this will point to the
397         files in WebKitLibraries\win\tools\vsprops in the source tree. In production builds,
398         JavaScriptCore.make sets $WebKitVSPropsRedirectionDir to
399         "$(SRCROOT)\AppleInternal\tools\vsprops\OpenSource\1\2\3\", so the full path for
400         FeatureDefines.vsprops becomes:
401
402         $(SRCROOT)\AppleInternal\tools\vsprops\OpenSource\1\2\3\..\..\..\WebKitLibraries\win\tools\vsprops\FeatureDefines.vsprops
403
404         which resolves to:
405
406         $(SRCROOT)\AppleInternal\tools\vsprops\OpenSource\WebKitLibraries\win\tools\vsprops\FeatureDefines.vsprops
407
408         (We rely on the fact that Windows doesn't care whether the directories "1", "2", and "3"
409         actually exist since they are matched by an equal number of ".." path components.)
410
411         Note that Visual Studio still won't pick up changes made to .vsprops files while Visual
412         Studio is open, but that problem hasn't seemed to cause developers many headaches so far.
413
414         Fixes <http://webkit.org/b/49181> Windows build fails mysteriously when .vsprops files are
415         updated
416
417         Reviewed by Dave Hyatt.
418
419         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
420         * DumpRenderTree/win/DumpRenderTree.vcproj:
421         * DumpRenderTree/win/ImageDiff.vcproj:
422         * FindSafari/FindSafari.vcproj:
423         * MiniBrowser/MiniBrowser.vcproj:
424         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
425         * TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj:
426         * WebKitAPITest/WebKitAPITest.vcproj:
427         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
428         * WebKitTestRunner/win/InjectedBundle.vcproj:
429         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
430         * WinLauncher/WinLauncher.vcproj:
431         * record-memory-win/record-memory-win.vcproj:
432         Changed to use paths relative to $WebKitVSPropsRedirectionDir to access shared .vsprops
433         files.
434
435 2010-11-22  Adam Roben  <aroben@apple.com>
436
437         Make it possible to run tests on Windows without Visual Studio or VC++
438         Express installed
439
440         webkitdir.pm::setupCygwinEnv dies when Visual Studio and VC++ Express
441         are not installed. But this function doesn't need to be called when we
442         already have a build available and are just trying to run the tests.
443
444         Fixes <http://webkit.org/b/49932> New Windows 7 bot can't run tests
445         because Visual Studio/VC++ Express aren't installed
446
447         Reviewed by Dave Hyatt.
448
449         * Scripts/webkitdirs.pm:
450         (determineConfigurationForVisualStudio): Don't call setupCygwinEnv, as
451         it is not needed by this function. Also added a FIXME.
452
453         (usingVisualStudioExpress): Call setupCygwinEnv directly rather than
454         relying on determineConfigurationForVisualStudio doing it.
455
456 2010-11-19  Hayato Ito  <hayato@chromium.org>
457
458         Reviewed by Shinichiro Hamaji.
459
460         Simplify Driver.run_test() so that it takes only one parameter, TestInput.
461
462         https://bugs.webkit.org/show_bug.cgi?id=49812
463
464         * Scripts/webkitpy/layout_tests/driver_test.py:
465         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
466         * Scripts/webkitpy/layout_tests/port/base.py:
467         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
468         * Scripts/webkitpy/layout_tests/port/chromium.py:
469         * Scripts/webkitpy/layout_tests/port/dryrun.py:
470         * Scripts/webkitpy/layout_tests/port/test.py:
471         * Scripts/webkitpy/layout_tests/port/webkit.py:
472         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
473         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
474
475 2010-11-22  Hayato Ito  <hayato@chromium.org>
476
477         Reviewed by Tony Chang.
478
479         [NRWT] Retry a few times in reading a png image to avoid a race condition.
480
481         https://bugs.webkit.org/show_bug.cgi?id=49924
482
483         * Scripts/webkitpy/layout_tests/port/chromium.py:
484
485 2010-11-22  João Paulo Rechi Vita  <jprvita@profusion.mobi>
486
487         Reviewed by Kenneth Rohde Christiansen.
488
489         Fix EWebLauncher zoom discretization
490         https://bugs.webkit.org/show_bug.cgi?id=49810
491
492         * EWebLauncher/main.c:
493         (nearest_zoom_level_get):
494         (zoom_level_set):
495         (on_load_finished):
496         (on_key_down):
497
498 2010-11-21  Sam Weinig  <sam@webkit.org>
499
500         Rubber-stamped by Dan Bernstein.
501
502         Fix ~300 WebKit2 tests by enabling developer extras preference.
503
504         * WebKitTestRunner/TestController.cpp:
505         (WTR::TestController::resetStateToConsistentValues):
506
507 2010-11-21  Sam Weinig  <sam@webkit.org>
508
509         Reviewed by Dan Bernstein.
510
511         Textareas should be resizable by default
512         https://bugs.webkit.org/show_bug.cgi?id=49892
513
514         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
515         * TestWebKitAPI/Tests/WebKit2/WKPreferences.cpp: Added.
516         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
517         Add preferences test.
518
519 2010-11-19  Mihai Parparita  <mihaip@chromium.org>
520
521         Reviewed by Tony Chang.
522
523         Rebaseline server: checksum-only differences and prefetching
524         https://bugs.webkit.org/show_bug.cgi?id=49841
525         
526         Add support for image test results where the only difference is in the
527         checksum.
528         
529         Add prefetching of the image results from the next test.
530         
531         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
532         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
533         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
534
535 2010-11-19  Dirk Pranke  <dpranke@chromium.org>
536
537         Reviewed by Tony Chang.
538
539         nrwt multiprocessing - add 'worker number' concept, move stuff to worker thread
540
541         Add the 'worker number' and 'worker name' concepts to the 
542         TestShellThread objects, and move test_types and test_args from
543         the TestRunner to the TestShellThread.
544
545         https://bugs.webkit.org/show_bug.cgi?id=49768
546
547         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
548         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
549
550 2010-11-19  Dirk Pranke  <dpranke@chromium.org>
551
552         Reviewed by Ojan Vafai.
553
554         Do some minor cleanup and bug fixing.
555
556         https://bugs.webkit.org/show_bug.cgi?id=49777
557
558         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
559         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
560
561 2010-11-19  Dirk Pranke  <dpranke@chromium.org>
562
563         Reviewed by Ojan Vafai.
564
565         nrwt - config.build_directory() is busted
566
567         Fixes a typo that was causing us to usually return the top level
568         directory WebKitBuild instead of WebKitBuild/{Debug,Release}. The
569         bug was hidden by test stubs that were too simplistic :(.
570
571         https://bugs.webkit.org/show_bug.cgi?id=49815
572
573         * Scripts/webkitpy/common/system/executive_mock.py:
574         * Scripts/webkitpy/layout_tests/port/config.py:
575         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
576
577 2010-11-19  Steve Falkenburg  <sfalken@apple.com>
578
579         Reviewed by Adam Roben.
580
581         Add Debug_Cairo_CFLite and Release_Cairo_CFLite configurations for all vcproj files
582         https://bugs.webkit.org/show_bug.cgi?id=49819
583
584         * DumpRenderTree/DumpRenderTree.sln:
585         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
586         * DumpRenderTree/win/DumpRenderTree.vcproj:
587         * DumpRenderTree/win/ImageDiff.vcproj:
588         * FindSafari/FindSafari.vcproj:
589         * MiniBrowser/MiniBrowser.vcproj:
590         * TestWebKitAPI/win/TestWebKitAPI.sln:
591         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
592         * TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj:
593         * WebKitAPITest/WebKitAPITest.vcproj:
594         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
595         * WebKitTestRunner/win/InjectedBundle.vcproj:
596         * WebKitTestRunner/win/InjectedBundleGenerated.vcproj:
597         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
598         * WinLauncher/WinLauncher.vcproj:
599         * record-memory-win/record-memory-win.vcproj:
600
601 2010-11-19  Adam Roben  <aroben@apple.com>
602
603         Add some more Windows slaves
604
605         This is probably more than we will be using in the immediate future,
606         but having some extra slave names makes experimenting with different
607         configurations easier.
608
609         Reviewed by Jon Honeycutt.
610
611         * BuildSlaveSupport/build.webkit.org-config/config.json: Added 5 more
612         Windows slaves and distributed them among the three different Test
613         builders.
614
615 2010-11-19  Kinuko Yasuda  <kinuko@chromium.org>
616
617         Reviewed by Ojan Vafai.
618
619         [Chromium] REGRESSION: Record actual test results in the JSON result summary for non-layout tests
620         https://bugs.webkit.org/show_bug.cgi?id=49702
621
622         Re-landing r72357 with a test fix.
623
624         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
625         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
626         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
627
628 2010-11-19  Avi Drissman  <avi@google.com>
629
630         Reviewed by Antonio Gomes.
631
632         Complete support for Unix editing mode
633         https://bugs.webkit.org/show_bug.cgi?id=49757
634
635         * DumpRenderTree/LayoutTestController.cpp:
636         (setEditingBehaviorCallback):
637         * DumpRenderTree/chromium/LayoutTestController.cpp:
638         (LayoutTestController::setEditingBehavior):
639
640 2010-11-19  Adam Roben  <aroben@apple.com>
641
642         Fix run-javascriptcore-tests when there's a space in the path to
643         testapi
644
645         Reviewed by Sam Weinig.
646
647         * Scripts/run-javascriptcore-tests: Use an indirect object when calling
648         system() to force it to interpret spaces in the path correctly.
649
650 2010-11-19  Adam Roben  <aroben@apple.com>
651
652         Update scripts for .exe name changes after r72327
653
654         As of r72327, the "_debug" suffix is only used in Debug_All builds.
655
656         Reviewed by Sam Weinig.
657
658         * Scripts/run-javascriptcore-tests:
659         (testapiPath):
660
661         * Scripts/sunspider-compare-results:
662         (pathToBuiltJSC):
663
664         * Scripts/webkitdirs.pm:
665         (jscPath):
666         (safariPath):
667
668 2010-11-19  Steve Falkenburg  <sfalken@apple.com>
669
670         Reviewed by Darin Adler.
671
672         Normalize Cairo/CFLite project/solution configuration names
673         https://bugs.webkit.org/show_bug.cgi?id=49818
674
675         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
676         * DumpRenderTree/win/DumpRenderTree.vcproj:
677         * DumpRenderTree/win/ImageDiff.vcproj:
678         * MiniBrowser/MiniBrowser.vcproj:
679         * TestWebKitAPI/win/TestWebKitAPI.sln:
680         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
681         * WebKitAPITest/WebKitAPITest.vcproj:
682         * WebKitTestRunner/win/InjectedBundle.vcproj:
683         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
684         * WinLauncher/WinLauncher.vcproj:
685         * Scripts/webkitdirs.pm: Updated cairo build configs.
686
687 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
688
689         Reviewed by Adam Roben.
690
691         Windows vcproj configuration names should be normalized across projects
692         https://bugs.webkit.org/show_bug.cgi?id=49776
693
694         * DumpRenderTree/DumpRenderTree.sln:
695         * FindSafari/FindSafari.vcproj:
696         * TestWebKitAPI/win/TestWebKitAPI.sln:
697         * TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj:
698         * TestWebKitAPI/win/TestWebKitAPIGeneratedCommon.vsprops: Added.
699         * WebKitAPITest/WebKitAPITest.vcproj:
700         * WebKitAPITest/WebKitAPITestCommon.vsprops:
701         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
702         * WebKitTestRunner/WebKitTestRunner.sln:
703         * WebKitTestRunner/win/InjectedBundleGenerated.vcproj:
704         * WebKitTestRunner/win/InjectedBundleGeneratedCommon.vsprops: Added.
705         * record-memory-win/record-memory-win.vcproj:
706
707 2010-11-18  Gavin Barraclough  <barraclough@apple.com>
708
709         Build fix - as per DRT-mac, always enabled developer extras,
710         in win/qt/gtk DRT.
711
712         This is necessary to enable rich exception messages to be generated
713         following https://bugs.webkit.org/show_bug.cgi?id=49708.
714
715         * DumpRenderTree/gtk/DumpRenderTree.cpp:
716         (shouldEnableDeveloperExtras):
717         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
718         (WebCore::shouldEnableDeveloperExtras):
719         * DumpRenderTree/win/DumpRenderTree.cpp:
720         (shouldEnableDeveloperExtras):
721
722 2010-11-18  Sheriff Bot  <webkit.review.bot@gmail.com>
723
724         Unreviewed, rolling out r72357.
725         http://trac.webkit.org/changeset/72357
726         https://bugs.webkit.org/show_bug.cgi?id=49784
727
728         It broke test-webkitpy tests since the patch didn't have the
729         corresponding test update (Requested by kinuko on #webkit).
730
731         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
732         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
733
734 2010-11-18  Mihai Parparita  <mihaip@chromium.org>
735
736         Reviewed by Tony Chang.
737
738         Rebaseline server: add rebaseline queue
739         https://bugs.webkit.org/show_bug.cgi?id=49763
740         
741         Add a UI queue (a multi-item <select>) where tests to be baselined
742         can be enqueued (also supports basic removal). A queue is used instead
743         of immediately doing rebaselines since SCM operations can take a while.
744         It's better to go through lots of tests quickly to mark them as
745         needing rebaselining and then batch the SCM operations.
746         
747         Also adds a simple log where results can be displayed.
748
749         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
750         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
751         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
752         * Scripts/webkitpy/tool/commands/data/rebaselineserver/queue.js: Added.
753         * Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js:
754         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
755
756 2010-11-18  Gavin Barraclough  <barraclough@apple.com>
757
758         Reviewed by Oliver Hunt.
759
760         Bug 49708 - Stop recompiling functions to regenerate exception info.
761
762         Instead only hold info as necessary – keep divot info is the inspector
763         is enabled, line number info is debugging or profiling, and handler
764         info for functions with try/catch.
765
766         * DumpRenderTree/mac/DumpRenderTree.mm:
767         (shouldEnableDeveloperExtras):
768             - always enable the developer tools from DRT, to ensure we
769               produce rich error messages on JavaScript exceptions.
770
771 2010-11-18  Kinuko Yasuda  <kinuko@chromium.org>
772
773         Reviewed by Ojan Vafai.
774
775         [Chromium] REGRESSION: Record actual test results in the JSON result summary for non-layout tests
776         https://bugs.webkit.org/show_bug.cgi?id=49702
777
778         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
779         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
780
781 2010-11-18  Dirk Pranke  <dpranke@chromium.org>
782
783         Reviewed by Tony Chang.
784
785         new-run-webkit-tests: create first part of 'message_broker' class for multiprocessing fixes
786
787         Create the first version of the 'message_broker' package. This
788         class will encapsulate all of the threading/multiprocessing and
789         message-sending details for the communication between the
790         'manager' object and the 'worker' objects. For the moment, it
791         just holds some routines and tests for logging thread stacks.
792
793         There should be no functional changes in this patch, just moving stuff
794         around.
795
796         https://bugs.webkit.org/show_bug.cgi?id=49707
797
798         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
799         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread_unittest.py: Removed.
800         * Scripts/webkitpy/layout_tests/layout_package/message_broker.py: Added.
801         * Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py: Added.
802         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
803         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
804
805 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
806
807         Reviewed by Adam Roben.
808
809         Remove leftover Windows Debug_Internal configurations
810         https://bugs.webkit.org/show_bug.cgi?id=49758
811
812         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
813         * DumpRenderTree/win/DumpRenderTree.vcproj:
814         * DumpRenderTree/win/ImageDiff.vcproj:
815         * MiniBrowser/MiniBrowser.vcproj:
816         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
817         * WebKitAPITest/WebKitAPITest.vcproj:
818         * WebKitTestRunner/win/InjectedBundle.vcproj:
819         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
820         * WinLauncher/WinLauncher.vcproj:
821
822 2010-11-18  Adam Roben  <aroben@apple.com>
823
824         Add three new Windows XP build slaves
825
826         Reviewed by Steve Falkenburg.
827
828         * BuildSlaveSupport/build.webkit.org-config/config.json: Added three
829         new slaves, and added them to the Windows XP Debug (Tests) builder.
830
831 2010-11-18  Daniel Bates  <dbates@rim.com>
832
833         Reviewed by Adam Roben.
834
835         update-webkit-support-libs should fall back to existing
836         WebKitSupportLibrary version if there is no internet connectivity
837         https://bugs.webkit.org/show_bug.cgi?id=49503
838
839         Fall back to existing support libraries (if present) when there is
840         no internet connection.
841
842         Currently, update-webkit-support-libs dies with an "out-of-date"
843         error when there is no internet connection because it cannot
844         retrieve versioning information from developer.apple.com. Because
845         update-webkit-support-libs fails, build-webkit fails. Instead,
846         if there is no internet connection and the support libraries are
847         present then we should warn the user and exit() with success so
848         that build-webkit can work without an internet connection.
849
850         * Scripts/update-webkit-support-libs:
851
852 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
853
854         Reviewed by Adam Roben.
855
856         Debug_Internal Windows configuration is unnecessary, should be removed
857         https://bugs.webkit.org/show_bug.cgi?id=49753
858
859         * DumpRenderTree/win/DumpRenderTree.cpp:
860         * MiniBrowser/win/stdafx.h:
861         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp:
862         * WebKitTestRunner/win/TestControllerWin.cpp:
863
864 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
865
866         Reviewed by Adam Roben.
867
868         Clean up vcproj errors
869         https://bugs.webkit.org/show_bug.cgi?id=49741
870
871         * MiniBrowser/MiniBrowser.vcproj:
872         * WebKitAPITest/WebKitAPITest.vcproj:
873         * WebKitTestRunner/win/InjectedBundle.vcproj:
874         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
875
876 2010-11-18  Hayato Ito  <hayato@chromium.org>
877
878         Reviewed by Shinichiro Hamaji.
879
880         [NRWT] Fix a break of '--new-baseline' feature in pixel_test
881         which is caused by r72249.
882
883         https://bugs.webkit.org/show_bug.cgi?id=49751
884
885         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
886
887 2010-11-18  Adam Roben  <aroben@apple.com>
888
889         Implement layoutTestController.waitForPolicyDelegate on Windows
890
891         Fixes <http://webkit.org/b/25038> <rdar://problem/6790213>.
892
893         Reviewed by Simon Fraser.
894
895         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
896         (LayoutTestController::waitForPolicyDelegate): Implemented by porting
897         code from LayoutTestControllerMac.mm.
898
899 2010-11-18  Hayato Ito  <hayato@chromium.org>
900
901         Reviewed by Shinichiro Hamaji.
902
903         [NRWT] Fix a break of '--new-baseline' feature which is caused by r72249.
904
905         https://bugs.webkit.org/show_bug.cgi?id=49740
906
907         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
908
909 2010-11-18  Tony Chang  <tony@chromium.org>
910
911         Reviewed by Adam Barth.
912
913         run platform/chromium/plugins/refcount-leaks.html on all platforms
914         https://bugs.webkit.org/show_bug.cgi?id=49485
915
916         Add PluginObject.testObjectCount which returns the number of allocated
917         TestObjects. Add PluginObject.testCreateTestObject which allocates
918         and returns a TestObject.
919
920         Add TestObject.refCount which returns the number of refs on the
921         TestObject.
922
923         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
924         (pluginGetProperty):
925         (pluginInvoke):
926         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
927         (getTestObjectCount):
928         (testAllocate):
929         (testDeallocate):
930         (testGetProperty):
931         (testScriptObjectInvoke): Release a ref to a plugin object that was
932             previously leaked.
933         * DumpRenderTree/TestNetscapePlugIn/TestObject.h:
934
935 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
936
937         Reviewed by Adam Roben.
938
939         WebKitTools projects (WebKitTestRunner, record-memory-win) should use vsprops for common build settings
940         https://bugs.webkit.org/show_bug.cgi?id=49711
941
942         * WebKitTestRunner/Configurations/WebKitTestRunnerCFLite.vsprops: Added.
943         * WebKitTestRunner/Configurations/WebKitTestRunnerCommon.vsprops: Added.
944         * WebKitTestRunner/Configurations/WebKitTestRunnerCoreFoundation.vsprops: Added.
945         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
946         * record-memory-win/record-memory-win-common.vsprops: Added.
947         * record-memory-win/record-memory-win.vcproj:
948
949 2010-11-18  Andras Becsi  <abecsi@inf.u-szeged.hu>
950
951         Reviewed by Csaba Osztrogonác.
952
953         [Qt][WK2] Only add user agent strings to the list which aren't listed yet.
954
955         * MiniBrowser/qt/BrowserWindow.cpp:
956         (BrowserWindow::updateUserAgentList):
957
958 2010-11-17  Adam Roben  <aroben@apple.com>
959
960         Don't trigger Windows builds when chromium-win test results change
961
962         Reviewed by Mark Rowe.
963
964         * Scripts/webkitpy/common/config/build.py:
965         (_should_file_trigger_build): Changed the regular expression used to
966         search for directory names to only consider full directory names,
967         rather than matching directory names where the string we care about is
968         a suffix of the name.
969
970         * Scripts/webkitpy/common/config/build_unittest.py:
971         (ShouldBuildTest): Added a test.
972
973 2010-11-17  Hayato Ito  <hayato@chromium.org>
974
975         Reviewed by Shinichiro Hamaji.
976
977         Remove a uri member from TestInput class.
978
979         https://bugs.webkit.org/show_bug.cgi?id=49691
980
981         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
982         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
983
984 2010-11-17  Adam Roben  <aroben@apple.com>
985
986         Move FindSafari's settings into a vsprops file
987
988         Fixes <http://webkit.org/b/49699> FindSafari should use a vsprops file
989
990         Reviewed by Steve Falkenburg.
991
992         * FindSafari/FindSafari.vcproj:
993         * FindSafari/FindSafariCommon.vsprops: Added.
994
995 2010-11-17  Adam Roben  <aroben@apple.com>
996
997         Use a minimum font size of 0 in DRT on Windows
998
999         This is the Windows equivalent of r72164.
1000
1001         Rubber-stamped by Dave Hyatt.
1002
1003         * DumpRenderTree/win/DumpRenderTree.cpp:
1004         (resetDefaultsToConsistentValues):
1005
1006 2010-11-17  Dirk Pranke  <dpranke@chromium.org>
1007
1008         Reviewed by Andreas Kling.
1009
1010         webkit-patch pretty-diff shouldn't need to be run from a checkout root in git
1011         https://bugs.webkit.org/show_bug.cgi?id=49639
1012
1013         * Scripts/webkitpy/common/checkout/scm.py:
1014
1015 2010-11-17  Mihai Parparita  <mihaip@chromium.org>
1016
1017         Reviewed by Tony Chang.
1018
1019         Rebaseline server: add loupe for image diffs
1020         https://bugs.webkit.org/show_bug.cgi?id=49692
1021         
1022         Add a loupe (magnifiying glass) for inspecting image diffs. Shows an
1023         enlarged area of the expected, actual and diff images side by side.
1024
1025         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
1026         * Scripts/webkitpy/tool/commands/data/rebaselineserver/loupe.js: Added.
1027         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
1028         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
1029
1030 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
1031
1032         Reviewed by Adam Roben.
1033
1034         WebKitLauncherWin should use vsprops for shared build settings
1035         https://bugs.webkit.org/show_bug.cgi?id=49696
1036
1037         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
1038         * WebKitLauncherWin/WebKitLauncherWinCommon.vsprops: Added.
1039
1040 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
1041
1042         Reviewed by Adam Roben.
1043
1044         WinLauncher should use vsprops for shared build settings
1045         https://bugs.webkit.org/show_bug.cgi?id=49695
1046
1047         * WinLauncher/WinLauncher.cpp:
1048         (_tWinMain):
1049         (PrintView):
1050         (WndProc):
1051         (MyEditProc):
1052         * WinLauncher/WinLauncher.vcproj:
1053         * WinLauncher/WinLauncherCommon.vsprops: Added.
1054
1055 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
1056
1057         Delete unused file unintentionally added in last change.
1058
1059         * DumpRenderTree/win/DumpRenderTreeCURL.vsprops: Removed.
1060
1061 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
1062
1063         Reviewed by Adam Roben.
1064
1065         DumpRenderTree on Windows should use a vsprops file for shared project settings
1066         https://bugs.webkit.org/show_bug.cgi?id=49690
1067
1068         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1069         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePluginCommon.vsprops: Added.
1070         * DumpRenderTree/win/DumpRenderTree.vcproj:
1071         * DumpRenderTree/win/DumpRenderTreeApple.vsprops: Added.
1072         * DumpRenderTree/win/DumpRenderTreeCFLite.vsprops: Added.
1073         * DumpRenderTree/win/DumpRenderTreeCURL.vsprops: Added.
1074         * DumpRenderTree/win/DumpRenderTreeCairo.vsprops: Added.
1075         * DumpRenderTree/win/DumpRenderTreeCommon.vsprops: Added.
1076         * DumpRenderTree/win/ImageDiff.vcproj:
1077         * DumpRenderTree/win/ImageDiffCommon.vsprops: Added.
1078
1079 2010-11-17  Hayato Ito  <hayato@chromium.org>
1080
1081         Refactor TestTypeBase.compare_output().
1082
1083         Introduce a TestOutput class and update compare_output() of each test
1084         types so that they can take both actual and expected TestOutput objects.
1085
1086         https://bugs.webkit.org/show_bug.cgi?id=49431
1087
1088         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1089         * Scripts/webkitpy/layout_tests/layout_package/test_output.py: Added.
1090         * Scripts/webkitpy/layout_tests/port/base.py:
1091         * Scripts/webkitpy/layout_tests/port/chromium.py:
1092         * Scripts/webkitpy/layout_tests/port/dryrun.py:
1093         * Scripts/webkitpy/layout_tests/port/test.py:
1094         * Scripts/webkitpy/layout_tests/port/webkit.py:
1095         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1096         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
1097         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
1098         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
1099
1100 2010-11-17  Adam Roben  <aroben@apple.com>
1101
1102         Make each Windows Test builder use the same OS for all its slaves
1103
1104         Currently each WebKit1 Windows Test builder has one Windows XP slave
1105         and one Windows 7 slave. Having different OSes in a single builder
1106         meant that the builders would alternate between red and green when
1107         there were OS-specific failures.
1108
1109         Fixes <http://webkit.org/b/49688> Windows Test builders alternate
1110         between red and green because of multiple OS versions
1111
1112         Reviewed by Mark Rowe.
1113
1114         * BuildSlaveSupport/build.webkit.org-config/config.json: Renamed
1115         "Windows Release (Tests)" to "Windows 7 Release (Tests)", "Windows
1116         Debug (Tests)" to "Windows XP Debug (Tests)", and "Windows Release
1117         (WebKit2 Tests)" to "Windows 7 Release (WebKit2 Tests)". Updated the
1118         builders for the two WebKit1 builders so that they would all use the
1119         appropriate OS.
1120
1121 2010-11-17  James Robinson  <jamesr@chromium.org>
1122
1123         Reviewed by Dan Bernstein.
1124
1125         [chromium] Set minimum font size to 0 for DRT
1126         https://bugs.webkit.org/show_bug.cgi?id=49677
1127
1128         Followup for http://trac.webkit.org/changeset/72141.
1129
1130         Test: http/tests/misc/acid3.html
1131
1132         * DumpRenderTree/chromium/WebPreferences.cpp:
1133         (WebPreferences::reset):
1134
1135 2010-11-17  Andras Becsi  <abecsi@webkit.org>
1136
1137         Unreviewed build fix.
1138
1139         [Qt] Add missing file I forgot in r72220.
1140
1141         * MiniBrowser/MiniBrowser.qrc: Added.
1142
1143 2010-11-17  Andras Becsi  <abecsi@webkit.org>
1144
1145         Reviewed by Andreas Kling.
1146
1147         [Qt][WK2] Make it possible to set custom user agent strings in MiniBrowser.
1148         https://bugs.webkit.org/show_bug.cgi?id=49627
1149
1150         Add a user agent dialog and a resource file for useragentlist.txt to MiniBrowser.
1151
1152         * MiniBrowser/MiniBrowser.qrc: Added.
1153         Add it here to prevent qmake from detecting it
1154         since it needs to be copied to the build directory.
1155         * MiniBrowser/qt/BrowserWindow.cpp:
1156         (BrowserWindow::BrowserWindow):
1157         (BrowserWindow::updateUserAgentList):
1158         (BrowserWindow::showUserAgentDialog):
1159         * MiniBrowser/qt/BrowserWindow.h:
1160         * MiniBrowser/qt/MiniBrowser.pro:
1161
1162 2010-11-17  Adam Roben  <aroben@apple.com>
1163
1164         Make changes to the Mac WebKit2 Skipped file trigger Windows builds
1165
1166         We use that file on Windows!
1167
1168         Fixes <http://webkit.org/b/49643> Changes to
1169         LayoutTests/platform/mac-wk2 should trigger a Windows build, but don't
1170
1171         Reviewed by Steve Falkenburg.
1172
1173         * Scripts/webkitpy/common/config/build.py:
1174         (_should_file_trigger_build): Added the mac-wk2 directory and make it
1175         trigger builds on SnowLeopard and Windows.
1176
1177         * Scripts/webkitpy/common/config/build_unittest.py:
1178         (ShouldBuildTest.test_should_build): Added a test.
1179
1180 2010-11-17  Satish Sampath  <satish@chromium.org>
1181
1182         Reviewed by Jeremy Orlow.
1183
1184         Clear the speech input mock explicitly before each test.
1185         https://bugs.webkit.org/show_bug.cgi?id=49660
1186
1187         * DumpRenderTree/chromium/LayoutTestController.cpp:
1188         (LayoutTestController::setMockSpeechInputResult):
1189         * DumpRenderTree/chromium/LayoutTestController.h:
1190         * DumpRenderTree/chromium/WebViewHost.cpp:
1191         (WebViewHost::speechInputController):
1192         (WebViewHost::reset): Invoke speech input mock's clearResults method.
1193         * DumpRenderTree/chromium/WebViewHost.h:
1194         (WebViewHost::speechInputControllerMock): Moved mock from LayoutTestController.
1195
1196 2010-11-17  Alexander Pavlov  <apavlov@chromium.org>
1197
1198         Unreviewed, specify IRC nickname in committers.py.
1199
1200         * Scripts/webkitpy/common/config/committers.py:
1201
1202 2010-11-17  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1203
1204         Reviewed by Tony Chang.
1205
1206         [NRWT] Make http locking similar to perl implementation
1207         https://bugs.webkit.org/show_bug.cgi?id=49187
1208
1209         * Scripts/webkitpy/common/system/file_lock.py: Added.
1210         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
1211         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1212
1213 2010-11-17  MORITA Hajime  <morrita@google.com>
1214
1215         Reviewed by Kent Tamura.
1216
1217         [Chromium][DRT] EventSender.contextClick() should aware spellchecking
1218         https://bugs.webkit.org/show_bug.cgi?id=49366
1219
1220         - EvenSender: Checked WebContextMenuData.misspelledWord and added extra context menu entries
1221           according to the spellchecker suggestion
1222         - MockSpellCheck: add fillSuggestionList to provide fake suggestions.
1223
1224         test_expectations.txt will be changed after this change is ported to test_shell.
1225         
1226         * DumpRenderTree/chromium/EventSender.cpp:
1227         (makeMenuItemStringsFor):
1228         (EventSender::contextClick):
1229         * DumpRenderTree/chromium/MockSpellCheck.cpp:
1230         (MockSpellCheck::fillSuggestionList):
1231         (MockSpellCheck::initializeIfNeeded):
1232         * DumpRenderTree/chromium/MockSpellCheck.h:
1233         * DumpRenderTree/chromium/WebViewHost.cpp:
1234         (WebViewHost::mockSpellCheck):
1235         * DumpRenderTree/chromium/WebViewHost.h:
1236
1237 2010-11-16  Dave Hyatt  <hyatt@apple.com>
1238
1239         Reviewed by Sam Weinig.
1240
1241         Make sure the pref for minimum font size is just explicitly set to 0 every time,
1242         since the bots have 1 stuck in their plist otherwise.
1243
1244         * DumpRenderTree/mac/DumpRenderTree.mm:
1245         (resetDefaultsToConsistentValues):
1246
1247 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
1248
1249         Reviewed by Tony Chang.
1250
1251         Rebaseline server: compute diffs client-side
1252         https://bugs.webkit.org/show_bug.cgi?id=49640
1253         
1254         The image diff output from the DRT is pretty bad on some ports (at
1255         least the Mac one), so it's better to compute diffs on the client by
1256         using <canvas>.
1257
1258         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
1259         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
1260         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
1261
1262 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
1263
1264         Reviewed by Tony Chang.
1265
1266         Rebaseline server: display test results
1267         https://bugs.webkit.org/show_bug.cgi?id=49626
1268         
1269         Adds basic result display to the rebaseline server. On the Python side
1270         this involves:
1271         - Parsing the unexpected_results.json into a dictionary.
1272         - Serving it as JSON under /results.json.
1273         (the JSON -> dict -> JSON transform isn't strictly necessary right now,
1274         but I'll need to have access to the parsed results on the Python side
1275         for follow-up changes).
1276         
1277         On the web UI side this adds:
1278         - Markup for display image and text results (expected, actual, diff),
1279           and JS for populating it.
1280         - Markup for breaking down test results by failure type and directory,
1281           and JS for populating it.
1282
1283         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
1284         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
1285         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
1286         * Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js: Added.
1287         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
1288
1289 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
1290
1291         Reviewed by Ojan Vafai.
1292
1293         See Chromium issue http://codereview.chromium.org/5133001/ - we
1294         are modifying the build bots to pass in the builder name with
1295         the "GPU" string appended instead of appending it in the code.
1296
1297         https://bugs.webkit.org/show_bug.cgi?id=49636
1298
1299         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
1300
1301 2010-11-16  Steve Falkenburg  <sfalken@apple.com>
1302
1303         Reviewed by Adam Roben.
1304
1305         Disable LTCG for Windows Release builds. Add new Release_LTCG configuration.
1306         https://bugs.webkit.org/show_bug.cgi?id=49632
1307
1308         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1309         * DumpRenderTree/win/DumpRenderTree.vcproj:
1310         * DumpRenderTree/win/ImageDiff.vcproj:
1311         * MiniBrowser/MiniBrowser.vcproj:
1312         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
1313         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
1314         * WebKitTestRunner/win/InjectedBundle.vcproj:
1315         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
1316         * WinLauncher/WinLauncher.vcproj:
1317         * record-memory-win/record-memory-win.vcproj:
1318
1319 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
1320
1321         Reviewed by Ojan Vafai.
1322
1323         new-run-webkit-tests: rename TestInfo to TestInput, move image hash to work thread
1324
1325         Rename the TestInfo class to TestInput to be clearer about its
1326         function, and move the checksum-reading code into dump_render_tree_thread
1327         to avoid cross-thread access.
1328
1329         https://bugs.webkit.org/show_bug.cgi?id=49573
1330
1331         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1332         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1333         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1334
1335 2010-11-16  Dave Hyatt  <hyatt@apple.com>
1336
1337         Fix bustage. Remove the minimum font size pref setting in DumpRenderTree.
1338
1339         * DumpRenderTree/mac/DumpRenderTree.mm:
1340         (resetDefaultsToConsistentValues):
1341
1342 2010-11-16  Stephanie Lewis  <slewis@apple.com>
1343
1344         Reviewed by Geoff Garen.
1345
1346         <rdar://problem/8624267> Leak creating offscreen webview running fast/dom tests
1347
1348         Use shared DeviceOrientationProviderMock. The old allocation method created a new WebDeviceOrientationProviderMock 
1349         for every WebView, and the WebKit API doesn't support that ownership model.
1350
1351         I also fixed a leak in setMockDeviceOrientation by adding a missing release.
1352
1353         * DumpRenderTree/mac/DumpRenderTree.mm:
1354         (createWebViewAndOffscreenWindow):
1355         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1356         (LayoutTestController::setMockDeviceOrientation):
1357
1358 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
1359
1360         Reviewed by Ojan Vafai.
1361
1362         new-run-webkit-tests: log the process id in --verbose mode
1363
1364         As part of the switch to multi-process mode, we should log the
1365         process id when running w/ --verbose.
1366
1367         https://bugs.webkit.org/show_bug.cgi?id=49571
1368
1369         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
1370
1371 2010-11-16  Adam Roben  <aroben@apple.com>
1372
1373         Only trigger Windows builds when files that we actually use on Windows
1374         are changed
1375
1376         A new Scheduler subclass, PlatformSpecificScheduler, has been added.
1377         It uses the new webkitpy.common.config.build module to determine
1378         whether a particular change should trigger a build on a particular
1379         platform. The Windows builders have been switched to use a
1380         PlatformSpecificScheduler.
1381
1382         The logic to determine whether or not a particular change should
1383         trigger a build on a given platform has only been implemented/tested
1384         for Windows. I tried to make it easy to add more platforms in the
1385         future, but I don't have enough familiarity with all platforms to be
1386         able to implement it for them.
1387
1388         Fixes <http://webkit.org/b/49407> Windows builders kick off builds for
1389         lots irrelevant changes (e.g., rebaselining Chromium test results)
1390
1391         Reviewed by Eric Seidel.
1392
1393         * BuildSlaveSupport/build.webkit.org-config/config.json: Use a
1394         PlatformSpecificScheduler for the Windows builders.
1395
1396         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1397         (PlatformSpecificScheduler.__init__): Added. Stores our platform, sets
1398         up a ChangeFilter that filters to our branch and will call through to
1399         our filter method, and calls up to our base class.
1400         (PlatformSpecificScheduler.filter): Calls through to
1401         build.should_build to find out whether we should trigger a build for
1402         this change.
1403
1404         * Scripts/webkitpy/common/net/build.py: Added.
1405         (_should_file_trigger_build): Uses a set of directories and regexp
1406         patterns to determine whether the given file should trigger a build on
1407         the given platform. As mentioned earlier, this has only been
1408         implemented for Windows, though I did try to make some guesses about
1409         other platforms.
1410         (should_build): Returns true if any of the files should trigger a
1411         build on the given platform.
1412
1413         * Scripts/webkitpy/common/net/build_unittest.py: Added.
1414         (ShouldBuildTest.test_should_build): Does some basic testing to make
1415         sure we're triggering builds for the right files. It only tests
1416         Windows for now, though I tried to make some guesses about other
1417         platforms.
1418
1419 2010-11-16  Simon Hausmann  <simon.hausmann@nokia.com>
1420
1421         Reviewed by Andreas Kling.
1422
1423         [Qt] Remove synchronous QWebPage::checkPermissions signal
1424         https://bugs.webkit.org/show_bug.cgi?id=46810
1425
1426         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1427         (WebCore::WebPage::WebPage):
1428         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1429         (LayoutTestController::grantDesktopNotificationPermission): When granting
1430         permission, grant it directly on the QWebPage/Frame, that will remember it.
1431         * QtTestBrowser/webpage.cpp:
1432         (WebPage::WebPage):
1433         * QtTestBrowser/webpage.h:
1434
1435
1436 2010-11-16  Andras Becsi  <abecsi@inf.u-szeged.hu>
1437
1438         Reviewed by Csaba Osztrogonác.
1439
1440         [Qt] Add isPrinting and setPrinting methods to DRT's LayoutTestController.
1441
1442         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1443         (LayoutTestController::reset):
1444         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1445         (LayoutTestController::isPrinting):
1446         (LayoutTestController::setPrinting):
1447
1448 2010-11-16  Balazs Kelemen  <kbalazs@webkit.org>
1449
1450         Reviewed by Andreas Kling.
1451
1452         [Qt][WK2] Add a way to turn on tiled backing store in MiniBrowser
1453         https://bugs.webkit.org/show_bug.cgi?id=49587
1454
1455         * MiniBrowser/qt/BrowserView.cpp:
1456         (BrowserView::BrowserView): Added BackingStoreType
1457         argument to ctor. Propagating it to the QGraphicsWKView.
1458         * MiniBrowser/qt/BrowserView.h:
1459         * MiniBrowser/qt/BrowserWindow.cpp:
1460         (BrowserWindow::BrowserWindow):
1461         (BrowserWindow::newWindow): Create the new BrowserWindow
1462         with the same BackingStoreType that this instance has.
1463         * MiniBrowser/qt/BrowserWindow.h: Added BackingStoreType member.
1464         * MiniBrowser/qt/main.cpp:
1465         (main): Use tiled backing store if got -tiled command line argument.
1466         Avoid redundant copy of command line args.
1467
1468 2010-11-16  John Knottenbelt  <jknotten@chromium.org>
1469
1470         Reviewed by Jeremy Orlow.
1471
1472         Move DeviceOrientationClientMock from LayoutTestController to WebViewHost.
1473         https://bugs.webkit.org/show_bug.cgi?id=48506
1474
1475         This ensures that there is one mock per page WebView / WebViewHost. This is a
1476         design constraint of the DeviceOrientationClientMock because the m_controller
1477         field can only store a pointer to a single instance of the corresponding
1478         DeviceOrientationController.
1479
1480         Test: fast/dom/DeviceOrientation/no-page-cache.html
1481
1482         * DumpRenderTree/chromium/LayoutTestController.cpp:
1483         (LayoutTestController::setMockDeviceOrientation):
1484         (LayoutTestController::deviceOrientationClientMock):
1485         * DumpRenderTree/chromium/LayoutTestController.h:
1486         * DumpRenderTree/chromium/WebViewHost.cpp:
1487         (WebViewHost::deviceOrientationClientMock):
1488         (WebViewHost::deviceOrientationClient):
1489         (WebViewHost::reset):
1490         * DumpRenderTree/chromium/WebViewHost.h:
1491
1492 2010-11-15  Andras Becsi  <abecsi@webkit.org>
1493
1494         Reviewed by Andreas Kling.
1495
1496         [Qt][WK2] Avoid polling in WebKitTestRunner.
1497         https://bugs.webkit.org/show_bug.cgi?id=49542
1498
1499         Make the WK2 testing session about 10-15% faster and decrease random flakiness
1500         resulting from timing skew by removing the millisecond polling from TestControllerQt.
1501
1502         * WebKitTestRunner/TestController.cpp:
1503         (WTR::TestController::didFinishLoadForFrame):
1504         * WebKitTestRunner/TestController.h:
1505         * WebKitTestRunner/TestInvocation.cpp:
1506         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
1507         * WebKitTestRunner/mac/TestControllerMac.mm:
1508         (WTR::TestController::notifyDone): Add method stub.
1509         * WebKitTestRunner/qt/TestControllerQt.cpp:
1510         (WTR::TestControllerRunLoop::instance):
1511         (WTR::TestControllerRunLoop::start):
1512         (WTR::TestControllerRunLoop::stop):
1513         (WTR::TestControllerRunLoop::TestControllerRunLoop):
1514         (WTR::TestControllerRunLoop::timerEvent):
1515         (WTR::TestController::notifyDone):
1516         (WTR::TestController::platformRunUntil):
1517         * WebKitTestRunner/win/TestControllerWin.cpp:
1518         (WTR::TestController::notifyDone): Add method stub.
1519
1520 2010-11-15  Darin Adler  <darin@apple.com>
1521
1522         * Scripts/webkitpy/common/net/bugzilla: Added property svn:ignore.
1523
1524 2010-11-15  Benjamin Poulain  <benjamin.poulain@nokia.com>
1525
1526         Reviewed by Andreas Kling.
1527
1528         [Qt] [WK2] Add Ctrl+L shortcut to Qt Minibrowser
1529         https://bugs.webkit.org/show_bug.cgi?id=49544
1530
1531         Add the shortcut to BrowserWindow.
1532
1533         * MiniBrowser/qt/BrowserWindow.cpp:
1534         (BrowserWindow::BrowserWindow):
1535         (BrowserWindow::openLocation):
1536         * MiniBrowser/qt/BrowserWindow.h:
1537
1538 2010-11-14  David Levin  <levin@chromium.org>
1539
1540         Reviewed by Daniel Bates.
1541
1542         check-webkit-style should detect PassRefPtr usage in functions.
1543         https://bugs.webkit.org/show_bug.cgi?id=49513
1544
1545         * Scripts/webkitpy/style/checkers/cpp.py:
1546         (check_for_function_lengths): Revert a comment change that I
1547          accidentally made in r71986.
1548         (check_pass_ptr_usage): Added the code to do the check.
1549         (process_line): Added the call to check_pass_ptr_usage.
1550         (CppChecker): Added the new error category.
1551         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
1552         (CppStyleTestBase::perform_pass_ptr_check): Runs the new check for
1553          testing purposes.
1554         (PassPtrTest::*): The class/functions to unit test the new
1555          functionality.
1556
1557 2010-11-14  David Levin  <levin@chromium.org>
1558
1559         Reviewed by Shinichiro Hamaji.
1560
1561         check-webkit-style function detection and the line count style checks should be separate.
1562         https://bugs.webkit.org/show_bug.cgi?id=49512
1563
1564         * Scripts/webkitpy/style/checkers/cpp.py: Do the separation.
1565         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Adjust the test to
1566           call the detection function and fix line counts in two places now that
1567           the code really only counts the lines in the body of the function.
1568
1569 2010-11-14  Andreas Kling  <kling@webkit.org>
1570
1571         Reviewed by Antonio Gomes.
1572
1573         http/tests/plugins tests print "Unhandled variable" to stderr under Qt
1574         https://bugs.webkit.org/show_bug.cgi?id=33438
1575
1576         Stifle "Unhandled variable" warning to match what the other
1577         TestNetscapePlugin does (NPP_GetValue in TestNetscapePlugIn/main.cpp
1578         simply returns NPERR_GENERIC_ERROR for unhandled variables.)
1579
1580         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
1581         (webkit_test_plugin_get_value):
1582
1583 2010-11-13  David Levin  <levin@chromium.org>
1584
1585         Reviewed by Eric Seidel.
1586
1587         check-webkit-style function detection crashes on functions in templates.
1588         https://bugs.webkit.org/show_bug.cgi?id=49504
1589
1590         For example "template <bool shouldClose(const Element*)>".
1591
1592         * Scripts/webkitpy/style/checkers/cpp.py: If not function is detected, bail out.
1593         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Modified the complex
1594           function detection test to expose this issue.
1595
1596 2010-11-12  David Levin  <levin@chromium.org>
1597
1598         Reviewed by Shinichiro Hamaji.
1599
1600         check-webkit-style function detection doesn't detect indented functions declaractions.
1601         https://bugs.webkit.org/show_bug.cgi?id=49446
1602
1603         Indented function declarations occur inside class definitions, so
1604         they are a pretty common (and worth detecting).
1605
1606         * Scripts/webkitpy/style/checkers/cpp.py:
1607           Changed regex to allow indentation.
1608           Changed the function start detection to only happen when not in a
1609           function.
1610           Changed function end detection to work based on matching braces
1611           instead of finding a close brace at the beginning of the line.
1612           Fixed close_expression to do what it says when it doesn't find
1613           the close.
1614         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Indented function test.
1615
1616 2010-11-12  Daniel Bates  <dbates@rim.com>
1617
1618         Rubber-stamped by Nikolas Zimmermann.
1619
1620         Make do-file-rename rename files in the directory WebKit2.
1621
1622         This also makes the list of searched directories in do-file-rename
1623         consistent with the list of searched directories in the script do-webcore-rename.
1624
1625         * Scripts/do-file-rename:
1626
1627 2010-11-12  Dirk Pranke  <dpranke@chromium.org>
1628
1629         Reviewed by James Robinson.
1630
1631         Attempt yet again to land the fix for bug 49360 (respecting
1632         set-webkit-configuration). We need to handle the cases where
1633         trying to run webkit-build-directory to find out where the
1634         default configuration might be fails (that shows up on some
1635         Chromium bots that apparently don't have perl installed).
1636
1637         https://bugs.webkit.org/show_bug.cgi?id=49360
1638
1639         * Scripts/webkitpy/layout_tests/port/config.py:
1640         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Added.
1641         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1642
1643 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
1644
1645         Reviewed by Adam Barth.
1646
1647         webkit-patch land logs commit message twice
1648         https://bugs.webkit.org/show_bug.cgi?id=49482
1649         
1650         Don't log the comment that's used by bugzilla.close_bug_as_fixed, since
1651         that's just the commit message, which we now output following the
1652         actual commit.
1653
1654         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
1655
1656 2010-11-12  Tony Chang  <tony@chromium.org>
1657
1658         Reviewed by Adam Barth.
1659
1660         run platform/chromium/plugins/script-object-invoke.html on all platforms
1661         https://bugs.webkit.org/show_bug.cgi?id=49280
1662
1663         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1664         (testScriptObjectInvoke):
1665         (pluginInvoke):
1666
1667 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
1668
1669         Reviewed by Adam Barth.
1670
1671         webkit-patch land-cowboy (and other commands) should display the committed SVN revision after landing a patch
1672         https://bugs.webkit.org/show_bug.cgi?id=49471
1673
1674         Right now there's no indication of a successful commit.
1675         
1676         * Scripts/webkitpy/tool/commands/download_unittest.py:
1677         * Scripts/webkitpy/tool/steps/commit.py:
1678
1679 2010-11-12  Kinuko Yasuda  <kinuko@chromium.org>
1680
1681         Reviewed by Ojan Vafai.
1682
1683         Include detailed test modifiers in results.json and enable incremental uploading for non-layout tests
1684         https://bugs.webkit.org/show_bug.cgi?id=49354
1685
1686         Also moved/integrated the upload method from run_webkit_tests.py to json_results_generator.py.
1687
1688         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
1689         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1690         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
1691         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1692
1693 2010-11-12  Tony Chang  <tony@chromium.org>
1694
1695         Reviewed by Kent Tamura.
1696
1697         [chromium] Load a blank page when closing DRT WebViewHosts
1698         https://bugs.webkit.org/show_bug.cgi?id=49417
1699
1700         In test_shell, when we close a window, we first load about:blank
1701         and run GC to fire the destruction logic of the page.  In test_shell,
1702         this happens in ~TestShell.  In DRT, we manage virtual windows as
1703         WebViewHosts, so we need to replicate this logic in ~WebViewHost.
1704
1705         Managing focus is a bit trickier, since normally you would get window
1706         events to reset the focus to the right window.  We do it manually
1707         instead.
1708
1709         This fixes 2 plugin tests on DRT.
1710
1711         * DumpRenderTree/chromium/TestShell.cpp:
1712         (TestShell::~TestShell): Move window close logic to ~WebViewHost.
1713         (TestShell::closeWindow): Move widget close() to WVH destructor and
1714             manage focus manually.
1715         (TestShell::closeRemainingWindows): Comment typo.
1716         * DumpRenderTree/chromium/WebViewHost.cpp:
1717         (WebViewHost::closeWidget):
1718         (invokeCloseWidget):
1719         (WebViewHost::closeWidgetSoon): Close the window asynchronously (matches
1720             test_shell).
1721         (WebViewHost::~WebViewHost):
1722         * DumpRenderTree/chromium/WebViewHost.h:
1723
1724 2010-11-12  David Levin  <levin@chromium.org>
1725
1726         Reviewed by Adam Barth.
1727
1728         cpp.py needs some tidying.
1729         https://bugs.webkit.org/show_bug.cgi?id=49443
1730
1731         * Scripts/webkitpy/style/checkers/cpp.py:
1732           Fix parameter ordering for iteratively_replace_matches_with_char
1733           to mirror other similar functions (and improved comments, etc.)
1734           Also, use a python-ism (for/else) to get rid of a variable that
1735           was there to determine if something had been found in the loop.
1736
1737 2010-11-12  Robert Kroeger  <rjkroege@chromium.org>
1738
1739         Reviewed by James Robinson.
1740
1741         [Chromium/DRT] Make EventSendingController send complete touch events.
1742         https://bugs.webkit.org/show_bug.cgi?id=49285
1743
1744         * DumpRenderTree/chromium/EventSender.cpp:
1745         (EventSender::addTouchPoint):
1746         (EventSender::updateTouchPoint):
1747         (EventSender::sendCurrentTouchEvent):
1748
1749 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
1750
1751         Unreviewed rollout of r71858.
1752
1753         Rollout out r71858 since it breaks new-run-webkit-httpd as used by the
1754         NaCl tests.
1755
1756         * Scripts/webkitpy/layout_tests/port/config.py:
1757         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Removed.
1758         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1759
1760 2010-11-12  Benjamin Poulain  <benjamin.poulain@nokia.com>
1761
1762         Reviewed by Kenneth Rohde Christiansen.
1763
1764         [Qt] TestController::runUntil doesn't honor the timeout parameter
1765         https://bugs.webkit.org/show_bug.cgi?id=48941
1766
1767         Implement the missing timeout of WebKitTestRunner for Qt.
1768
1769         A QElapsedTimer has been added to RunUntilConditionLoop in order
1770         to measure how long the loop has been running. When the timer
1771         is bigger or equal than the timeout, the loop ends even if the
1772         condition is not met.
1773
1774         * WebKitTestRunner/qt/TestControllerQt.cpp:
1775         (WTR::RunUntilConditionLoop::start):
1776         (WTR::RunUntilConditionLoop::run):
1777         (WTR::RunUntilConditionLoop::timerEvent):
1778         (WTR::TestController::platformRunUntil):
1779
1780 2010-11-11  David Levin  <levin@chromium.org>
1781
1782         Reviewed by Shinichiro Hamaji.
1783
1784         check-webkit-style function detection doesn't handle templates with spaces.
1785         https://bugs.webkit.org/show_bug.cgi?id=49427
1786
1787         * Scripts/webkitpy/style/checkers/cpp.py: Stripped out templates when
1788           finding the function name.
1789         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test have a
1790           template with a space.
1791
1792 2010-11-11  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
1793
1794         Reviewed by Shinichiro Hamaji.
1795
1796         Enable check-webkit-style on Qt files
1797         https://bugs.webkit.org/show_bug.cgi?id=48322
1798
1799         Adding exception for _q_... private slots and
1800         ..._data() methods in tests
1801
1802         * Scripts/webkitpy/style/checker.py: enable Qt dirs, add exceptions
1803         * Scripts/webkitpy/style/checker_unittest.py: improve unittests
1804
1805 2010-11-11  David Levin  <levin@chromium.org>
1806
1807         Reviewed by Shinichiro Hamaji.
1808
1809         check-webkit-style function detection doesn't work for templates and destructors.
1810         https://bugs.webkit.org/show_bug.cgi?id=49425
1811
1812         * Scripts/webkitpy/style/checkers/cpp.py: Changed regex for templates and destructors.
1813         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test hit this issue.
1814
1815 2010-11-11  Joone Hur  <joone@kldp.org>
1816
1817         Reviewed by Antonio Gomes.
1818
1819         [GTK][DRT] Implement LayoutTestController::nodesFromRect
1820         https://bugs.webkit.org/show_bug.cgi?id=46598
1821
1822         Support nodesFromRect in DRT
1823
1824         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1825         (LayoutTestController::nodesFromRect):
1826
1827 2010-11-11  MORITA Hajime  <morrita@google.com>
1828
1829         Reviewed by Kent Tamura.
1830
1831         editing/selection/context-menu-on-text.html fails on chromium
1832         https://bugs.webkit.org/show_bug.cgi?id=45898
1833
1834         Enhance Chromium's EventSender::contextClick() to return
1835         array of menu strings. The implementation incomplete because
1836         Although actual context menu is implemented by the browser,
1837         it reflects editability of underlying context thus makes
1838         context-menu-on-text.html passed.
1839         
1840         * DumpRenderTree/chromium/EventSender.cpp:
1841         (makeMenuItemStringsFor): Added.
1842         (EventSender::contextClick): Now returns array of menu strings.
1843         * DumpRenderTree/chromium/WebViewHost.cpp:
1844         (WebViewHost::showContextMenu): Now storing given WebContextMenuData object.
1845         (WebViewHost::clearContextMenuData): Added.
1846         (WebViewHost::lastContextMenuData): Added.
1847         * DumpRenderTree/chromium/WebViewHost.h:
1848
1849 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1850
1851         Unreviewed, build fix.
1852
1853         fix build breakage caused by fix for bug 49380 (r71858).
1854
1855         config_standalone needs to ensure that the packages it needs are
1856         in sys.path.
1857
1858         https://bugs.webkit.org/show_bug.cgi?id=49419
1859
1860         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
1861
1862 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1863
1864         Reviewed by Adam Roben.
1865
1866         Fix NRWT to respect set-webkit-configuration again :(
1867
1868         This change fixes a typo in config.py that was causing the wrong
1869         value to be read initially and us never actually looking into
1870         the filesystem to get the default configuration.
1871         
1872         * Scripts/webkitpy/layout_tests/port/config.py:
1873         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
1874         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1875
1876 2010-11-11  Eric Seidel  <eric@webkit.org>
1877
1878         Reviewed by Adam Barth.
1879
1880         Split out Bug, Attachment and CommitterValidator from bugzilla.py
1881         https://bugs.webkit.org/show_bug.cgi?id=49403
1882
1883         CommitterValidator really had nothing to do with bugzilla.py.
1884         I've put it next to committers.py for now since its rather tied to that file.
1885
1886         * Scripts/webkitpy/common/config/committervalidator.py: Added.
1887         * Scripts/webkitpy/common/config/committervalidator_unittest.py: Added.
1888         * Scripts/webkitpy/common/net/bugzilla/__init__.py:
1889         * Scripts/webkitpy/common/net/bugzilla/attachment.py: Added.
1890         * Scripts/webkitpy/common/net/bugzilla/bug.py: Added.
1891         * Scripts/webkitpy/common/net/bugzilla/bug_unittest.py: Added.
1892         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
1893         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py:
1894         * Scripts/webkitpy/tool/bot/feeders.py:
1895         * Scripts/webkitpy/tool/commands/queues.py:
1896
1897 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1898
1899         Reviewed by Eric Seidel.
1900
1901         fix missing 'pid' argument for http lockfile checking on win32
1902         https://bugs.webkit.org/show_bug.cgi?id=49363
1903
1904         * Scripts/webkitpy/common/system/executive.py:
1905
1906 2010-11-11  Eric Seidel  <eric@webkit.org>
1907
1908         Rubber-stamped by Adam Barth.
1909
1910         Move bugzilla.py into its own module in preparation for splitting one-file-per-class
1911         https://bugs.webkit.org/show_bug.cgi?id=49402
1912
1913         * Scripts/webkitpy/common/net/bugzilla/__init__.py: Added.
1914         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla.py.
1915         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla_unittest.py.
1916
1917 2010-11-10  Chris Guillory  <chris.guillory@google.com>
1918
1919         Reviewed by Chris Fleizach.
1920
1921         Use IAccessibleComparable to compare accessibility objects.
1922         https://bugs.webkit.org/show_bug.cgi?id=49118
1923
1924         * DumpRenderTree/AccessibilityUIElement.cpp:
1925         (AccessibilityUIElement::isEqual):
1926         * DumpRenderTree/AccessibilityUIElement.h:
1927         * DumpRenderTree/win/AccessibilityUIElementWin.cpp:
1928         (comparableObject):
1929         (AccessibilityUIElement::isEqual):
1930
1931 2010-11-10  Csaba Osztrogonác  <ossy@webkit.org>
1932
1933         Reviewed by David Hyatt.
1934
1935         HTML5 Ruby support should be mandatory feature
1936         https://bugs.webkit.org/show_bug.cgi?id=49272
1937
1938         Remove Ruby as optional feature.
1939
1940         * Scripts/build-webkit:
1941
1942 2010-11-10  Adam Roben  <aroben@apple.com>
1943
1944         Print VC++ Express build logs in the same order that the projects are
1945         built
1946
1947         Fixes <http://webkit.org/b/49326> print-vse-failure-logs prints logs
1948         in an unhelpful order
1949
1950         Reviewed by Eric Seidel.
1951
1952         * Scripts/print-vse-failure-logs:
1953         (PrintVisualStudioExpressLogs._build_order): Added. Uses
1954         print-msvc-project-dependencies to figure out the order in which
1955         projects are built.
1956         (PrintVisualStudioExpressLogs._sort_buildlogs): Added. Sorts the logs
1957         based on their build order and project name.
1958         (PrintVisualStudioExpressLogs._obj_directory): Moved code to find the
1959         scripts directory from here...
1960         (PrintVisualStudioExpressLogs._scripts_directory): ...to here.
1961         (PrintVisualStudioExpressLogs.main): Sort the logs before printing
1962         them.
1963
1964 2010-11-10  Adam Roben  <aroben@apple.com>
1965
1966         Only print the interesting text from VC++ Express build logs
1967
1968         Fixes <http://webkit.org/b/49325> It's hard to find the interesting
1969         output from print-vse-failure-logs
1970
1971         Reviewed by Eric Seidel.
1972
1973         * Scripts/print-vse-failure-logs: Removed unnecessary os.path import,
1974         added newly-required imports. Removed a comment that was essentially a
1975         FIXME, and is now fixed.
1976         (PrintVisualStudioExpressLogs._relevant_text): Added. Uses
1977         BeautifulSoup to extract the relevant text from the build log.
1978         (PrintVisualStudioExpressLogs.main): Only print the relevant text.
1979
1980 2010-11-09  Mihai Parparita  <mihaip@chromium.org>
1981
1982         Reviewed by Tony Chang.
1983
1984         [Chromium] http/tests/security/XFrameOptions fail with Chromium DRT
1985         https://bugs.webkit.org/show_bug.cgi?id=49286
1986         
1987         The tests in http/tests/security/XFrameOptions started to fail after
1988         r71297, since it changed the WebViewHost::assignIdentifierToRequest
1989         logic to always assign identifiers to requests, even if we're not going
1990         to be dumping resource load callbacks at that point. These tests
1991         only call LayoutTestController.dumpResourceLoadCallbacks after the
1992         request is started, thus their expectations have "<unknown>" as the
1993         identifier. Change WebViewHost to the previous behavior, which also
1994         match the other ports'.
1995
1996         * DumpRenderTree/chromium/WebViewHost.cpp:
1997         (WebViewHost::assignIdentifierToRequest):
1998         (WebViewHost::removeIdentifierForRequest):
1999
2000 2010-11-08  Tony Chang  <tony@chromium.org>
2001
2002         Reviewed by Adam Barth.
2003
2004         run platform/chromium/plugins/return-npobject.html on all platforms
2005         https://bugs.webkit.org/show_bug.cgi?id=49103
2006
2007         This tests that we can get an NPObject returned through a method on
2008         an NPAPI Object.
2009
2010         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
2011         (testCallbackReturn):
2012         (pluginInvoke):
2013
2014 2010-11-09  Sheriff Bot  <webkit.review.bot@gmail.com>
2015
2016         Unreviewed, rolling out r71672.
2017         http://trac.webkit.org/changeset/71672
2018         https://bugs.webkit.org/show_bug.cgi?id=49276
2019
2020         It doesn't work on Chromium Windows bot (Requested by Ossy on
2021         #webkit).
2022
2023         * Scripts/webkitpy/common/system/file_lock.py: Removed.
2024         * Scripts/webkitpy/common/system/file_lock_unittest.py: Removed.
2025         * Scripts/webkitpy/layout_tests/port/http_lock.py:
2026
2027 2010-11-09  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2028
2029         Reviewed by Tony Chang.
2030
2031         [NRWT] Make http locking similar to perl implementation
2032         https://bugs.webkit.org/show_bug.cgi?id=49187
2033
2034         * Scripts/webkitpy/common/system/file_lock.py: Added.
2035         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
2036         * Scripts/webkitpy/layout_tests/port/http_lock.py:
2037
2038 2010-11-09  James Kozianski  <koz@chromium.org>
2039
2040         Reviewed by Eric Seidel.
2041
2042         Clean up imports for webkit-patch.
2043         https://bugs.webkit.org/show_bug.cgi?id=49083
2044
2045         Move some imports from commands into main.py where they are used.
2046         Move command imports into the __init__ of the commands module.
2047
2048         * Scripts/webkitpy/tool/commands/__init__.py:
2049         * Scripts/webkitpy/tool/commands/download.py:
2050         * Scripts/webkitpy/tool/commands/queues.py:
2051         * Scripts/webkitpy/tool/main.py:
2052
2053 2010-11-09  Yi Shen  <yi.4.shen@nokia.com>
2054
2055         Reviewed by Andreas Kling.
2056
2057         [Qt][QtTestBrowser] Don't erase incorrect url in the Url Bar
2058         https://bugs.webkit.org/show_bug.cgi?id=49047
2059
2060         * QtTestBrowser/mainwindow.cpp:
2061         (MainWindow::setAddressUrl):
2062
2063 2010-11-09  Andras Becsi  <abecsi@webkit.org>
2064
2065         Reviewed by Csaba Osztrogonác.
2066
2067         Improve old-run-webkit-tests --verbose to show DumpTool opening/closing
2068         https://bugs.webkit.org/show_bug.cgi?id=49252
2069
2070         * Scripts/old-run-webkit-tests:
2071
2072 2010-11-09  Csaba Osztrogonác  <ossy@webkit.org>
2073
2074         Reviewed by Tor Arne Vestbø.
2075
2076         Output of old-run-webkit-tests is buggy in verbose mode
2077         https://bugs.webkit.org/show_bug.cgi?id=49249
2078
2079         * Scripts/old-run-webkit-tests: Redundant outputs removed.
2080
2081 2010-11-08  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
2082
2083         Reviewed by Martin Robinson.
2084
2085         [GTK] Link with target name set does not work
2086         https://bugs.webkit.org/show_bug.cgi?id=48865
2087
2088         When a new page is created with a name (target=myFrame), the new
2089         mainFrame could not be found because they where not stored in the
2090         same PageGroup. As PageGroup are not exposed externally so the
2091         simpliest solution is to use a global page group name. This also fixes
2092         issue with visited link coloration across pages. After this change the
2093         private function webkit_web_view_set_group_name() was no longer used
2094         so it was removed completly.
2095
2096         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2097         (createWebView):
2098
2099 2010-11-08  John Knottenbelt  <jknotten@chromium.org>
2100
2101         Reviewed by Kent Tamura.
2102
2103         WebViewHost::reset() uses placement new.
2104         https://bugs.webkit.org/show_bug.cgi?id=49069
2105
2106         WebViewHost is using placement destruction / new to simulate a fresh
2107         WebViewHost object at the same address. This is because the WebView remains
2108         open across tests and maintains a pointer to the WebViewHost.
2109         This change resets member variables explictly instead of the placement new dance.
2110
2111         * DumpRenderTree/chromium/WebViewHost.cpp:
2112         (WebViewHost::WebViewHost):
2113         (WebViewHost::reset):
2114
2115 2010-11-08  Dirk Pranke <dpranke@chromium.org>
2116
2117         Reviewed by Eric Seidel.
2118
2119         Fix the webkit_base_dir logic in webkitpy/layout_tests/port/config.py. 
2120         It turns out that NRWT can't use the code in scm.find_checkout_root()
2121         because the Chromium bots don't do full checkouts of the WebKit
2122         tree; they only check out subdirectories like WebKitTools/Scripts.
2123         Until we can figure out a better approach for this, I've
2124         restored the base_dir-detecting code from NRWT, which works in
2125         any directory tree, scm or no.
2126
2127         This also restores the files modified in r71475 and r71474.
2128
2129         https://bugs.webkit.org/show_bug.cgi?id=49151
2130
2131         * WebKitTools/Scripts/webkitpy/layout_tests/port/config.py
2132         * WebKitTools/Scripts/webkitpy/layout_tests/port/config_unittest.py
2133
2134 2010-11-08  Eric Seidel  <eric@webkit.org>
2135
2136         Reviewed by Adam Barth.
2137
2138         webkit-patch failure-reason explodes if a build is missing
2139         https://bugs.webkit.org/show_bug.cgi?id=49195
2140
2141         This is likely a recent regression.  Adding a unit test for this case.
2142
2143         * Scripts/webkitpy/common/net/buildbot.py:
2144         * Scripts/webkitpy/common/net/buildbot_unittest.py:
2145
2146 2010-11-08  Dan Bernstein  <mitz@apple.com>
2147
2148         Reviewed by Darin Adler.
2149
2150         Ignore the system scrollbar setting.
2151
2152         * DumpRenderTree/mac/DumpRenderTree.mm:
2153         (resetDefaultsToConsistentValues): Set the scrollbar default.
2154
2155 2010-11-05  Adam Roben  <aroben@apple.com>
2156
2157         Make webkitpy.common.system.executive_unittest pass when running under
2158         Win32 Python
2159
2160         Fixes <http://webkit.org/b/49033>.
2161
2162         Reviewed by Dave Levin and Eric Seidel.
2163
2164         * Scripts/webkitpy/common/system/executive.py:
2165         (Executive._run_command_with_teed_output): Pass the arguments through
2166         encode_argument_if_needed rather than using Cygwin-specific code here.
2167         (Executive.run_and_throw_if_fail): Use child_process_encoding to decode
2168         the output.
2169         (Executive.run_command): Use encode_argument_if_needed to encode the
2170         arguments and child_process_encoding to decode the output.
2171         (Executive._child_process_encoding): Returns the encoding that should be
2172         used when communicating with child processes. On Windows we use mbcs,
2173         which maps to the current code page. On all other platforms we use
2174         UTF-8.
2175         (Executive._should_encode_child_process_arguments): Returns True if
2176         arguments to child processes need to be encoded. This is currently
2177         only needed on Cygwin and Win32 Python 2.x.
2178         (Executive._encode_argument_if_needed): Encode the argument using
2179         child_process_encoding if we need to encode arguments to child
2180         processes on this platform.
2181
2182         * Scripts/webkitpy/common/system/executive_unittest.py:
2183         (never_ending_command): Added. Returns arguments to run a command that
2184         will not quit until we kill it. On Windows we use wmic, on other
2185         platforms we use yes.
2186         (ExecutiveTest.test_run_command_with_unicode): Changed to expect the
2187         mbcs encoding to be used and for output from the child processes to
2188         have been roundtripped through encode/decode on Win32 Python. When
2189         UTF-8 is the encoding the roundtripping is undetectable, but with mbcs
2190         it's possible that some characters will not be able to be converted
2191         and will be replaced by question marks; the round-tripping allows us
2192         to expect this result.
2193
2194         (ExecutiveTest.test_kill_process):
2195         (ExecutiveTest.test_kill_all):
2196         Use never_ending_command instead of invoking "yes" directly. Expect an
2197         exit code of 1 when using Win32 Python, as that's what seems to happen.
2198
2199 2010-11-08  Adam Roben  <aroben@apple.com>
2200
2201         Roll out r71532
2202
2203         It broke the build for Cygwin 1.7 installs. Cygwin 1.7's default
2204         .bashrc unsets %TEMP%, which broke copy-tools.cmd.
2205
2206         * Scripts/webkitdirs.pm:
2207
2208 2010-11-08  Tony Chang  <tony@chromium.org>
2209
2210         Reviewed by Adam Barth.
2211
2212         run platform/chromium/plugins/nested-plugin-objects.html on all platforms
2213         https://bugs.webkit.org/show_bug.cgi?id=49094
2214
2215         This tests that objects created by plugins are proplery cleaned up.
2216
2217         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
2218         (testPassTestObject):
2219         (pluginInvoke):
2220         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
2221         (testAllocate):
2222         (testDeallocate):
2223         (testGetProperty):
2224         (testConstruct):
2225
2226 2010-11-08  Adam Roben  <aroben@apple.com>
2227
2228         Mark Windows builds triggered from Perl as being non-interactive
2229
2230         This affects whether some of our scripts will show alerts vs. printing
2231         to the build log.
2232
2233         Fixes <http://webkit.org/b/49181> Windows build fail mysteriously when
2234         .vsprops files are updated
2235
2236         Reviewed by Steve Falkenburg.
2237
2238         * Scripts/webkitdirs.pm:
2239         (buildVisualStudioProject): Set WEBKIT_NONINTERACTIVE_BUILD to 1.
2240
2241 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2242
2243         Reviewed by Ojan Vafai.
2244
2245         Make http locking default in NRWT.
2246         https://bugs.webkit.org/show_bug.cgi?id=48053
2247
2248         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2249         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2250
2251 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2252
2253         Reviewed by Csaba Osztrogonác.
2254
2255         [NRWT] If the http lock fails we shouldn't do any locking
2256         https://bugs.webkit.org/show_bug.cgi?id=49164
2257
2258         If something goes wrong with the locking, the test should keep going.
2259
2260         * Scripts/webkitpy/layout_tests/port/http_lock.py:
2261
2262 2010-11-08  Adam Roben  <aroben@apple.com>
2263
2264         Switch back to using kCGImageAlphaPremultipliedFirst when generating
2265         pixel dumps on Windows
2266
2267         I changed this behavior in r71418 thinking that it was required for
2268         getting plugins to show up in pixel dumps. But it doesn't seem to be
2269         necessary, and was making it impossible to compare new Windows pixel
2270         dumps with existing Windows or Mac pixel dumps (because ImageDiff won't
2271         compare an image with alpha to an image without alpha).
2272
2273         Fixes <http://webkit.org/b/49172> REGRESION (r71418): Can't compare
2274         new Windows pixel results to existing Windows or Mac results
2275
2276         Reviewed by Antti Koivisto.
2277
2278         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
2279         (createBitmapContextFromWebView): Replaced kCGImageAlphaNoneSkipFirst
2280         with kCGImageAlphaPremultipliedFirst.
2281
2282 2010-11-08  Csaba Osztrogonac  <ossy@webkit.org>
2283
2284         Unreviewed, rolling out r71466.
2285         http://trac.webkit.org/changeset/71466
2286         https://bugs.webkit.org/show_bug.cgi?id=48865
2287
2288         It broke layout tests on GTK bots.
2289
2290         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2291         (createWebView):
2292
2293 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2294
2295         Reviewed by Csaba Osztrogonác.
2296
2297         Enable running of Qt API tests on BuildBot
2298         https://bugs.webkit.org/show_bug.cgi?id=49004
2299
2300         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
2301
2302 2010-11-08  Eric Seidel  <eric@webkit.org>
2303
2304         Reviewed by Adam Barth.
2305
2306         Add clean-review-queue command to remove closed bugs from the webkit.org/pending-review
2307         https://bugs.webkit.org/show_bug.cgi?id=49160
2308
2309         Bugzilla doesn't automatically remove r? when a bug gets closed.
2310         This script takes care of that for webkit.org.
2311
2312         * Scripts/webkitpy/common/net/bugzilla.py:
2313         * Scripts/webkitpy/tool/commands/upload.py:
2314
2315 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
2316
2317         Unreviewed, rolling out r71474.
2318         http://trac.webkit.org/changeset/71474
2319         https://bugs.webkit.org/show_bug.cgi?id=48280
2320
2321         breaks chromium webkit tests
2322         https://bugs.webkit.org/show_bug.cgi?id=49151
2323
2324         * Scripts/webkitpy/layout_tests/port/base.py:
2325         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2326         * Scripts/webkitpy/layout_tests/port/config.py:
2327         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
2328         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
2329         * Scripts/webkitpy/layout_tests/port/webkit.py:
2330         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2331         * Scripts/webkitpy/tool/mocktool.py:
2332
2333 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
2334
2335         Unreviewed, rolling out r71475.
2336         http://trac.webkit.org/changeset/71475
2337  
2338         breaks chromium webkit tests
2339         https://bugs.webkit.org/show_bug.cgi?id=49151
2340
2341         * Scripts/webkitpy/common/newstringio.py: Removed.
2342         * Scripts/webkitpy/common/newstringio_unittest.py: Removed.
2343         * Scripts/webkitpy/common/system/executive_mock.py: Removed.
2344         * Scripts/webkitpy/common/system/filesystem_mock.py: Removed.
2345         * Scripts/webkitpy/layout_tests/port/config_mock.py: Removed.
2346
2347 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2348
2349         Unreviewed, build fix.
2350
2351         Add files inexplicably not committed in r71474 as part of the
2352         fix for bug 48280.
2353
2354         * Scripts/webkitpy/common/newstringio.py: Added.
2355         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
2356         * Scripts/webkitpy/common/system/executive_mock.py: Added.
2357         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
2358         * Scripts/webkitpy/layout_tests/port/config_mock.py: Added.
2359
2360 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2361
2362         Reviewed by Eric Siedel.
2363
2364         new-run-webkit-tests: update port/base and port/webkit to use the
2365         new FileSystem and Config abstractions, pulling more logic out of
2366         the base Port classes into separate, mockable objects.
2367
2368         Also create a MockFileSystem object, a MockConfig object, move
2369         MockExecutive into common/system to be next to executive, and
2370         update the config object to use a FileSystem.
2371
2372         https://bugs.webkit.org/show_bug.cgi?id=48280
2373
2374         * Scripts/webkitpy/common/newstringio.py: Added.
2375         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
2376         * Scripts/webkitpy/common/system/executive_mock.py: Added.
2377         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
2378         * Scripts/webkitpy/layout_tests/port/base.py:
2379         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2380         * Scripts/webkitpy/layout_tests/port/config.py:
2381         * Scripts/webkitpy/layout_tests/port/config_mock.py:
2382         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
2383         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
2384         * Scripts/webkitpy/layout_tests/port/webkit.py:
2385         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2386         * Scripts/webkitpy/tool/mocktool.py:
2387
2388 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2389
2390         Unreviewed, build breakage.
2391
2392         Apparently I uploaded the wrong version of the file to fix 49122
2393         and neither Eric or I noticed - it was missing a dirname() call.
2394         Fixing ...
2395
2396         https://bugs.webkit.org/show_bug.cgi?id=49122
2397
2398         * Scripts/webkitpy/common/checkout/scm.py:
2399
2400 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2401
2402         Reviewed by Eric Seidel.
2403
2404         webkitpy/tool/* unittests change cwd and don't clean up properly
2405
2406         https://bugs.webkit.org/show_bug.cgi?id=49122
2407
2408         * Scripts/webkitpy/common/checkout/scm.py:
2409
2410 2010-11-05  Chris Marrin  <cmarrin@apple.com>
2411
2412         Reviewed by Simon Fraser.
2413
2414         suspendAnimations/resumeAnimations not present in WebKit2
2415         https://bugs.webkit.org/show_bug.cgi?id=49109
2416
2417         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
2418         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
2419         (WTR::LayoutTestController::suspendAnimations):
2420         (WTR::LayoutTestController::resumeAnimations):
2421         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
2422
2423 2010-11-05  Tony Chang  <tony@chromium.org>
2424
2425         Reviewed by David Levin.
2426
2427         cleanup style in TestNetscapePlugIn/PluginObject.cpp
2428         https://bugs.webkit.org/show_bug.cgi?id=49044
2429
2430         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
2431         (getPluginClass):
2432         (pluginGetProperty):
2433         (pluginSetProperty):
2434         (variantToIdentifier):
2435         (testCallback):
2436         (getURL):
2437         (getURLNotify):
2438         (testInvokeDefault):
2439         (testGetProperty):
2440         (testDocumentOpen):
2441         (testWindowOpen):
2442         (pluginInvoke):
2443         (handleCallback):
2444         (notifyStream):
2445         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
2446
2447 2010-11-05  Eric Seidel  <eric@webkit.org>
2448
2449         Reviewed by David Levin.
2450
2451         Add QueueStatusServer/__init__.py so others can run the QueueStatusServer tests
2452         https://bugs.webkit.org/show_bug.cgi?id=49032
2453
2454         I wrote this file as part of bug 47847, but I forgot to commit it.
2455         No one else noticed it missing because test-webkitpy knows how
2456         to recover in the case where it can't import QueueStatusServer
2457         (which generally occurs due to not having installed the AppEngine SDK).
2458
2459         * QueueStatusServer/__init__.py: Added.
2460         * QueueStatusServer/model/workitems_unittest.py:
2461          - Remove a test which fails.  This was probably landed (by me)
2462            from my other machine, which since this __init__.py was missing
2463            I never noticed the failure and landed this invalid test.
2464            Sadly we can't really test remove_work_item as it depends
2465            on .key() working.  .key() will throw unless the object
2466            has already been saved it seems.
2467            This may be a mis-design in our remove_work_item implementation,
2468            but for now, just removing the test.
2469
2470 2010-11-04  Adam Roben  <aroben@apple.com>
2471
2472         Teach check-webkit-style about TestNetscapePlugIn
2473
2474         Fixes <http://webkit.org/b/49030> check-webkit-style is ignorant of
2475         TestNetscapePlugIn's coding conventions
2476
2477         Reviewed by John Sullivan.
2478
2479         * Scripts/webkitpy/style/checker.py: Lump TestNetscapePlugIn in with
2480         WebKitAPITest and TestWebKitAPI in having different include order and
2481         naming conventions than WebCore. Added some comments about why the
2482         exceptions exist.
2483
2484 2010-11-04  Adam Roben  <aroben@apple.com>
2485
2486         Add a test that shows that windowed plugins are able to paint
2487
2488         Somehow we never had a test for this before.
2489
2490         Fixes <http://webkit.org/b/49029> Should add a test that shows
2491         windowed plugins can paint
2492
2493         Reviewed by Jon Honeycutt.
2494
2495         * DumpRenderTree/TestNetscapePlugIn/Tests/win/DrawsGradient.cpp: Added.
2496         (DrawsGradient::DrawsGradient):
2497         (DrawsGradient::wndProc): We handle the WM_PAINT and WM_PRINTCLIENT messages.
2498
2499         (DrawsGradient::onPaint):
2500         (DrawsGradient::onPrintClient):
2501         These both just call through to paint.
2502
2503         (DrawsGradient::paint): Fills our client area with some gradients.
2504
2505         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Link
2506         against Msimg32.lib for ::GradientFill and added DrawsGradient.
2507
2508         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
2509         (createBitmapContextFromWebView): Use WM_PRINT instead of
2510         WM_PRINTCLIENT so that ::DefWindowProc will send
2511         WM_PRINT/WM_PRINTCLIENT messages to the WebView's child windows.
2512         Replaced kCGImageAlphaPremultipledFirst with kCGImageAlphaNoneSkipFirst
2513         because GDI doesn't support alpha and kCGImageBitmapOrder32Little with
2514         kCGImageBitmapOrder32Host because, who knows, maybe someday Windows
2515         will run on a big-endian platform.
2516
2517 2010-11-04  Adam Roben  <aroben@apple.com>
2518
2519         Extract much of NPNInvalidateRectInvalidatesWindow's code into a
2520         WindowedPluginTest base class
2521
2522         The base class takes care of subclassing the plugin's window so that a
2523         custom WNDPROC is called. This will make it easier to write tests that
2524         need to handle window messages.
2525
2526         Fixes <http://webkit.org/b/49028> It's hard to write a PluginTest with
2527         a custom WNDPROC
2528
2529         Reviewed by Jon Honeycutt.
2530
2531         * DumpRenderTree/TestNetscapePlugIn/Tests/win/NPNInvalidateRectInvalidatesWindow.cpp:
2532         (NPNInvalidateRectInvalidatesWindow::NPNInvalidateRectInvalidatesWindow):
2533         (NPNInvalidateRectInvalidatesWindow::NPP_SetWindow):
2534         (NPNInvalidateRectInvalidatesWindow::wndProc):
2535         (NPNInvalidateRectInvalidatesWindow::testInvalidateRect):
2536         Moved code from here to WindowedPluginTest. Changed to use window()
2537         instead of m_window.
2538
2539         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2540         Added WindowedPluginTest and added TestNetscapePlugIn/win to the
2541         include path. Also reordered the include path to make a little more
2542         sense and simplified the entry that added TestNetscapePlugIn itself to
2543         the include path.
2544
2545         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.cpp: Added.
2546         (WindowedPluginTest::WindowedPluginTest):
2547         (WindowedPluginTest::NPP_SetWindow):
2548         (WindowedPluginTest::staticWndProc):
2549         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.h: Added.
2550         (WindowedPluginTest::window):
2551         Code came from NPNInvalidateRectInvalidatesWindow.
2552
2553 2010-11-04  Adam Roben  <aroben@apple.com>
2554
2555         Add a plugin test to show that windowed plugins are clipped correctly
2556
2557         Fixes <http://webkit.org/b/49024> <rdar://problem/8487847> Windowed
2558         plugins aren't clipped in WebKit2 on Windows
2559
2560         Reviewed by Jon Honeycutt.
2561
2562         * DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowRegionIsSetToClipRect.cpp: Added.
2563         (WindowRegionIsSetToClipRect::WindowRegionIsSetToClipRect): Initialize members.
2564         (WindowRegionIsSetToClipRect::NPP_SetWindow): Check that our window
2565         region matches the clip rect we know we should have based on
2566         window-region-is-set-to-clip-rect.html, and check that our window class
2567         doesn't have the CS_PARENTDC style.
2568
2569         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2570         Added WindowRegionIsSetToClipRect.
2571
2572 2010-11-05  Alejandro G. Castro  <alex@igalia.com>
2573
2574         Reviewed by Martin Robinson.
2575
2576         [GTK] Avoid font hinting in the DRT
2577         https://bugs.webkit.org/show_bug.cgi?id=48548
2578
2579         Change the settings to avoid font hinting, it was causing
2580         different results depending on the distribution.
2581
2582         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2583         (initializeGtkFontSettings):
2584         * DumpRenderTree/gtk/fonts/fonts.conf:
2585
2586 2010-11-05  Eric Seidel  <eric@webkit.org>
2587
2588         Reviewed by Adam Barth.
2589
2590         build-webkit spams system.log by repeatedly checking for PBXProductDirectory in com.apple.Xcode
2591         https://bugs.webkit.org/show_bug.cgi?id=49051
2592
2593         This is a speculative fix.  The unit tests cover these methods, however
2594         I don't know if this will fully stop the system.log spam.
2595
2596         * Scripts/webkitpy/layout_tests/port/base.py:
2597         * Scripts/webkitpy/layout_tests/port/webkit.py:
2598
2599 2010-11-05  Eric Seidel  <eric@webkit.org>
2600
2601         Reviewed by Adam Barth.
2602
2603         Add basic support for showing bot id on /queue-status/ pages
2604         https://bugs.webkit.org/show_bug.cgi?id=49037
2605
2606         This support is really simple.  Eventually we'll want to
2607         show the bot id in the lock table too, but we don't have
2608         that information stored in the server yet.
2609
2610         * QueueStatusServer/handlers/queuestatus.py:
2611         * QueueStatusServer/handlers/queuestatus_unittest.py: Added.
2612         * QueueStatusServer/templates/includes/singlequeuestatus.html:
2613
2614 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
2615
2616         Reviewed by Adam Barth.
2617
2618         Create a filesystem wrapper that we can use to enforce
2619         particular conventions and use for mocking and dependency
2620         injection down the line.
2621
2622         https://bugs.webkit.org/show_bug.cgi?id=48144
2623
2624         * Scripts/webkitpy/common/system/filesystem.py: Added.
2625         * Scripts/webkitpy/common/system/filesystem_unittest.py: Added.
2626
2627 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2628
2629         Reviewed by Tony Chang.
2630
2631         run_webkit_tests_unittest fails under Python 2.5
2632         https://bugs.webkit.org/show_bug.cgi?id=49043
2633         
2634         Switch from itertools.chain.from_iterable (which was added in 2.6)
2635         to using itertools.chain directly.
2636
2637         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2638
2639 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2640
2641         Unreviewed fix to webkit-patch rebaseline-server so that it runs under
2642         Python 2.5 (it needs an import to use the with statement).
2643
2644         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
2645
2646 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
2647
2648         Reviewed by Adam Barth.
2649
2650         new-run-webkit-tests: split out webkit-specific configuration stuff into a new module
2651
2652         The current NRWT code has webkit-specific configuration code (like
2653         _script_path, default configuration, etc.) mixed in with
2654         layout-test-specific stuff in port/base. The configuration code
2655         should be split out into a separate module for easier mocking,
2656         testing, and isolation.
2657
2658         https://bugs.webkit.org/show_bug.cgi?id=48264
2659
2660         * Scripts/webkitpy/layout_tests/port/config.py: Added.
2661         * Scripts/webkitpy/layout_tests/port/config_unittest.py: Added.
2662
2663 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2664
2665         Reviewed by Tony Chang.
2666
2667         Rebaseline server: initial framework
2668         https://bugs.webkit.org/show_bug.cgi?id=48892
2669         
2670         Adds the basic framework for the rebaseline server (details at 
2671         http://webkit.org/b/47761). Includes the rebaseline-server webkit-patch
2672         command, which starts an HTTP server that can serve static files or
2673         invoke handler methods on a class.
2674
2675         * Scripts/webkitpy/tool/commands/__init__.py:
2676         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html: Added.
2677         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css: Added.
2678         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js: Added.
2679         * Scripts/webkitpy/tool/commands/rebaselineserver.py: Added.
2680
2681 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
2682
2683         Reviewed by James Robinson.
2684
2685         new-run-webkit-tests wasn't using DRT by default for
2686         --platform chromium-gpu
2687
2688         The default value was set to False instead of None, which meant
2689         that the platform specific logic wasn't firing to change the
2690         value to True (b/c we were afraid we'd be overriding the user
2691         preference).
2692
2693         https://bugs.webkit.org/show_bug.cgi?id=49038
2694
2695         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2696
2697 2010-11-04  Tony Chang  <tony@chromium.org>
2698
2699         Reviewed by Adam Barth.
2700
2701         make platform/chromium/plugins/multiple-plugins.html pass on all platforms
2702         https://bugs.webkit.org/show_bug.cgi?id=49026
2703
2704         Have the testObject.bar property return the string "bar".  This lets
2705         us run and pass platform/chromium/plugins/multiple-plugins.html on all
2706         platforms.
2707
2708         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
2709         (testGetProperty):
2710
2711 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2712
2713         Reviewed by Tony Chang.
2714
2715         NRWT doesn't have good test coverage for --run-chunk, --batch-size, --run-part, etc.
2716         https://bugs.webkit.org/show_bug.cgi?id=48878
2717         
2718         Add get_tests_run so that it's easy to see which tests get run (and with
2719         what batching) for a given flag combination. Flesh out the various
2720         test cases that have FIXMEs.
2721         
2722         Also fixes an off-by-one error (batch sizes were one larger than 
2723         expected) and makes --run-part also have wraparound behavior, like
2724         --run-chunk.
2725
2726         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2727         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2728         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2729
2730 2010-11-04  Erik Arvidsson  <arv@chromium.org>
2731
2732         Reviewed by Dimitri Glazkov.
2733
2734         Support box-sizing without the vendor prefix
2735         https://bugs.webkit.org/show_bug.cgi?id=36713
2736
2737         Based on patch by Peter Beverloo <peter@lvp-media.com>
2738
2739         * iExploder/htdocs/cssproperties.in: Change -webkit-box-sizing to box-sizing.
2740
2741 2010-11-04  Csaba Osztrogonác  <ossy@webkit.org>
2742
2743         Unreviewed rollout r71340, because it broke Chromium Windows bot.
2744
2745         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2746         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2747
2748 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2749
2750         Reviewed by Ojan Vafai.
2751
2752         Make http locking default in NRWT.
2753         https://bugs.webkit.org/show_bug.cgi?id=48053
2754
2755         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2756         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2757
2758 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2759
2760         Reviewed by Eric Seidel.
2761
2762         [NRWT] Clear invalid http locks on Windows platform as well
2763         https://bugs.webkit.org/show_bug.cgi?id=48515
2764
2765         * Scripts/webkitpy/common/system/executive.py:
2766         * Scripts/webkitpy/common/system/executive_unittest.py:
2767         * Scripts/webkitpy/layout_tests/port/http_lock.py:
2768
2769 2010-11-02  Adam Roben  <aroben@apple.com>
2770
2771         Reduce our dependence on coreutils when running Python tests
2772
2773         This patch introduces versions of the standard echo and cat utilities
2774         implemented in Python. They are probably missing features of their
2775         coreutils equivalents, but they can do what's necessary for our Python
2776         tests. This is useful on Windows, where these utilities typically
2777         aren't available.
2778
2779         Fixes <http://webkit.org/b/48883> executive_unittest relies on echo
2780         and cat utilities from coreutils, which are not present on Windows
2781
2782         Reviewed by Eric Seidel.
2783
2784         * Scripts/webkitpy/common/system/executive_unittest.py: Changed to use
2785         our Python-based echo and cat.
2786
2787         * Scripts/webkitpy/common/system/fileutils.py: Added.
2788         (make_stdout_binary): On Windows, puts sys.stdout into binary mode so
2789         that \n won't be translated into \r\n. I couldn't think of a good way
2790         to test this directly without touching the filesystem, but it is tested
2791         indirectly by echo_unittest.
2792
2793         * Scripts/webkitpy/test/cat.py: Added.
2794         (command_arguments): Returns a list for invoking cat with the given arguments.
2795         (main): Acts like a simplified version of the coreutils cat utility.
2796
2797         * Scripts/webkitpy/test/cat_unittest.py: Added.
2798         (CatTest.assert_cat): Runs cat with the given input and ensures the
2799         output matches the input.
2800         (CatTest.test_basic): Performs a simple test of cat.
2801         (CatTest.test_no_newline): Tests what happens when the input string
2802         doesn't have a trailing newline.
2803         (CatTest.test_unicode): Tests passing a unicode string to cat.
2804         (CatTest.test_as_command): Tests running cat as a separate command.
2805
2806         * Scripts/webkitpy/test/echo.py: Added.
2807         (command_arguments): Returns a list for invoking echo with the given arguments.
2808         (main): Acts like a simplified version of the coreutils echo utility.
2809
2810         * Scripts/webkitpy/test/echo_unittest.py: Added.
2811         (EchoTest.test_basic): Performs a simple test of echo.
2812         (EchoTest.test_no_newline): Tests passing -n to echo to suppress the
2813         trailing newline.
2814         (EchoTest.test_unicode): Tests passing unicode and non-unicode strings
2815         to echo.
2816         (EchoTest.test_argument_order): Tests what happens when -n is not the
2817         first argument.
2818         (EchoTest.test_empty_arguments): Tests what happens when you pass [] to
2819         echo.main.
2820         (EchoTest.test_no_arguments): Tests what happens when you call
2821         echo.main with no arguments.
2822         (EchoTest.test_as_command): Tests running echo as a separate command.
2823
2824 2010-11-04  Renata Hodovan  <reni@inf.u-szeged.hu>
2825
2826         Unreviewed: Add myself to the list of Committers.
2827
2828         * Scripts/webkitpy/common/config/committers.py:
2829
2830 2010-11-04  Andreas Kling  <kling@webkit.org>
2831
2832         Reviewed by Kenneth Rohde Christiansen.
2833
2834         [Qt] Hook into QWebPage::scrollRequested for resizes-to-contents mode
2835
2836         In RTC mode the QGraphicsWebView item is the size of the contents,
2837         scrolling works a bit differently  (we need to react to scrollRequested.)
2838
2839         Normally QGraphicsView will replay the last mouse event when scrolling,
2840         so to prevent WebKit from getting confused by this we temporarily make
2841         the QGraphicsView non-interactive.
2842
2843         * QtTestBrowser/webview.cpp:
2844         (WebViewGraphicsBased::setPage):
2845         (WebViewGraphicsBased::scrollRequested):
2846         * QtTestBrowser/webview.h:
2847
2848 2010-11-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2849
2850         Reviewed by Andreas Kling.
2851
2852         Use OS(MAC_OS_X) rather than PLATFORM(MAC) in TestControllerQt
2853
2854         PLATFORM(MAC) is not defined for the Qt port, as it refers to
2855         the Mac-port, not the Mac OS X operating system.
2856
2857         * WebKitTestRunner/qt/TestControllerQt.cpp:
2858         (WTR::isExistingLibrary):
2859
2860 2010-11-04  Leonid Ebril  <leonid.ebril@nokia.com>
2861
2862         Reviewed by Andreas Kling.
2863
2864         [Qt] Adding iphone user agent string the user agent list for QtTestBrowser
2865         https://bugs.webkit.org/show_bug.cgi?id=48869
2866
2867         * QtTestBrowser/useragentlist.txt:
2868
2869 2010-11-03  Adam Roben  <aroben@apple.com>
2870
2871         Always use uppercase drive names in strings returned by abspath_to_uri
2872
2873         Some versions of cygpath use lowercase drive letters while others use
2874         uppercase, which makes it hard to test the output of code that uses
2875         cygpath.
2876
2877         Fixes <http://webkit.org/b/48914> webkitpy.common.system.path_unittest
2878         fails with Cygwin 1.5
2879
2880         Reviewed by Eric Seidel.
2881
2882         * Scripts/webkitpy/common/system/path.py:
2883         (cygpath): Updated the docstring to indicate that only absolute paths
2884         should be passed for now (though relative paths will work fine).
2885         (_Cygpath.convert): Upper-case the first letter of the converted Windows path.
2886
2887 2010-11-03  George Guo  <George.Guo@Nokia.com>
2888
2889         Reviewed by Andreas Kling.
2890
2891         [Qt] QtTestBrowser : set mmp rule pageddata in Symbian
2892         https://bugs.webkit.org/show_bug.cgi?id=48767
2893     
2894         Paging is needd on Symbian devices to support benchmarks tests like
2895         dromaeo.com and Celtic Kane that need a lot of memory to run  
2896
2897         * QtTestBrowser/QtTestBrowser.pro:
2898
2899 2010-11-03  Jenn Braithwaite  <jennb@chromium.org>
2900
2901         Reviewed by Dmitry Titov.
2902
2903         Chromium: Update resource tracking when moving a frame between documents
2904         https://bugs.webkit.org/show_bug.cgi?id=48363
2905
2906         * DumpRenderTree/chromium/WebViewHost.cpp:
2907         (WebViewHost::assignIdentifierToRequest):
2908         Always put resource id in map so we can make assumptions about its
2909         presence.
2910         (WebViewHost::removeIdentifierForRequest):
2911         Added.
2912         (WebViewHost::didFinishResourceLoad):
2913         (WebViewHost::didFailResourceLoad):
2914         Check existence of resource id before removing from map.
2915         * DumpRenderTree/chromium/WebViewHost.h:
2916
2917 2010-11-03  Victor Wang  <victorw@chromium.org>
2918
2919         Reviewed by Adam Barth.
2920
2921         [Chromium] update buildbot names in chromium rebaseline tool.
2922
2923         https://bugs.webkit.org/show_bug.cgi?id=48881
2924
2925         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2926
2927 2010-11-03  Darin Adler  <darin@apple.com>
2928
2929         Updated Xcode projects by opening them with Xcode 3.2.4.
2930         Updated svn:ignore for Xcode projects.
2931
2932         * MiniBrowser/MiniBrowser.xcodeproj: Added property svn:ignore.
2933         * TestWebKitAPI/TestWebKitAPI.xcodeproj: Added property svn:ignore.
2934         * WebKitLauncher/WebKitLauncher.xcodeproj: Modified property svn:ignore.
2935         * WebKitLauncher/WebKitLauncher.xcodeproj/project.pbxproj: Updated with Xcode 3.2.4.
2936
2937 2010-11-03  Adam Roben  <aroben@apple.com>
2938
2939         Stop waiting for messages from the web process after a timeout elapses
2940
2941         TestController::runUntil now takes a timeout parameter that specifies
2942         how long to wait for the condition to become true. WebKitTestRunner
2943         takes a --timeout flag that specifies how long the timeout should be.
2944         run-webkit-tests passes this flag to WebKitTestRunner so its timeout
2945         will be similar to run-webkit-tests.
2946
2947         Fixes <http://webkit.org/b/43047> <rdar://problem/8365833>
2948         WebKitTestRunner waits forever if the web process crashes
2949
2950         Reviewed by Darin Adler and Anders Carlsson.
2951
2952         * Scripts/old-run-webkit-tests:
2953         (top level): Moved the GuardMalloc timeout adjustment here from
2954         readFromDumpToolWithTimer.
2955         (openDumpTool): Make WTR use a timeout similar to but slightly shorter
2956         than the one that was specified on the command line.
2957
2958         * WebKitTestRunner/TestController.cpp:
2959         (WTR::TestController::TestController): Initialize our timeout values.
2960         (WTR::TestController::initialize): Parse the --timeout flag and use it
2961         to modify our timeout values.
2962         (WTR::TestController::resetStateToConsistentValues): Changed to use a
2963         short timeout while waiting for the web process to reset and to return
2964         a boolean indicating whether we were able to reset the web process.
2965         Uses a 5-second timeout while waiting for the process to be reset.
2966         (WTR::TestController::runTest): Changed to return a boolean indicating
2967         whether we were able to reset the web process (and thus run the test).
2968         (WTR::TestController::runUntil): Call through to platformRunUntil.
2969
2970         (WTR::TestController::runTestingServerLoop):
2971         (WTR::TestController::run):
2972         Changed to bail if any test can't be run. This will cause the process
2973         to exit. (Unfortunately this will make run-webkit-tests think we
2974         crashed; see <http://webkit.org/b/48943>.)
2975
2976         * WebKitTestRunner/TestController.h: Added platformRunUntil,
2977         m_longTimeout, and m_shortTimeout.
2978
2979         * WebKitTestRunner/TestInvocation.cpp:
2980         (WTR::TestInvocation::invoke): Use a short timeout when waiting for the
2981         initial response and a long timeout when waiting for the test to
2982         complete. Check whether runUntil timed out and print an error message
2983         if so.
2984
2985         * WebKitTestRunner/mac/TestControllerMac.mm:
2986         (WTR::TestController::platformRunUntil): Renamed from runUntil. Pass
2987         [NSDate distantPast] to -[NSRunLoop runMode:beforeDate:] so that we
2988         won't block waiting for the run loop. Only loop until the timeout
2989         elapses.
2990
2991         * WebKitTestRunner/qt/TestControllerQt.cpp:
2992         (WTR::TestController::platformRunUntil): Renamed from runUntil. Added a
2993         FIXME about honoring the timeout.
2994
2995         * WebKitTestRunner/win/TestControllerWin.cpp:
2996         (WTR::TestController::platformRunUntil): Renamed from runUntil. Use
2997         ::MsgWaitForMultipleObjectsEx to implement the timeout. Changed to use
2998         ::PeekMessageW so that we don't block waiting for messages to become
2999         available.
3000
3001 2010-11-03  Adam Roben  <aroben@apple.com>
3002
3003         Add a plugin test that evaluates JS after removing the plugin element
3004         from the document
3005
3006         This test replaces platform/win/plugins/plugin-delayed-destroy.html.
3007         That test was made to prevent a crash very similar to this one, but
3008         unfortunately tested only the mechanism that prevented the crash and
3009         not whether the crash itself was prevented. Since WebKit2 uses a
3010         different mechanism to prevent the crash, the test was failing even
3011         though WebKit2 was not vulnerable to the crash. This new test crashes
3012         if there is no mechanism in place to prevent it and passes in both
3013         WebKit1 and WebKit2.
3014
3015         Fixes <http://webkit.org/b/46711> <rdar://problem/8485903>
3016         platform/win/plugins/plugin-delayed-destroy.html fails in WebKit2
3017
3018         Reviewed by Anders Carlsson.
3019
3020         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3021         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
3022         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
3023         * GNUmakefile.am:
3024         Added new file.
3025
3026         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
3027         (pluginDeallocate): Make sure we delete the PluginTest object. This
3028         prevents a leak and also allows us to test the crash.
3029
3030         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
3031         (PluginTest::executeScript): Made this into a non-static member
3032         function.
3033
3034         (PluginTest::waitUntilDone):
3035         (PluginTest::notifyDone):
3036         Updated for changes to executeScript.
3037
3038         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Added executeScript.
3039
3040         * DumpRenderTree/TestNetscapePlugIn/Tests/EvaluateJSAfterRemovingPluginElement.cpp: Added.
3041         (EvaluateJSAfterRemovingPluginElement::EvaluateJSAfterRemovingPluginElement):
3042         Initialize ourselves and tell the test harness to wait.
3043         (EvaluateJSAfterRemovingPluginElement::NPP_DestroyStream): Remove our
3044         plugin element from the document, then execute some JavaScript. If
3045         WebKit does not have appropriate mechanisms in place, we'll be
3046         destroyed inside the first call to executeScript and crash on the
3047         second call.
3048
3049 2010-11-02  Stephen White  <senorblanco@chromium.org>
3050
3051         Reviewed by Tony Chang.
3052
3053         [chromium] Fix LayoutTestController UMRs.
3054         https://bugs.webkit.org/show_bug.cgi?id=48872
3055
3056         * DumpRenderTree/chromium/LayoutTestController.cpp:
3057         (LayoutTestController::LayoutTestController):
3058
3059 2010-11-03  Kent Tamura  <tkent@chromium.org>
3060
3061         Reviewed by Dimitri Glazkov.
3062
3063         REGRESSION: rebaseline-chromium-webkit-tests uses non-zero tolerance for
3064         image dup detection
3065         https://bugs.webkit.org/show_bug.cgi?id=48744
3066
3067         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
3068          - Make a function for option parsing for ease of test
3069          - Set 0 to options.tolerance
3070         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
3071          - Add a test for this change
3072
3073 2010-11-02  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
3074
3075         Unreviewed: Add myself to the list of Committers.
3076
3077         * Scripts/webkitpy/common/config/committers.py:
3078
3079 2010-11-02  Anders Carlsson  <andersca@apple.com>
3080
3081         Fix build.
3082
3083         * TestWebKitAPI/PlatformUtilities.cpp:
3084         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
3085
3086 2010-11-02  Sam Weinig  <sam@webkit.org>
3087
3088         Reviewed by Anders Carlsson.
3089
3090         Add a way to send startup messages on the context which can be posted when a process launches
3091         <rdar://problem/8617928>
3092         https://bugs.webkit.org/show_bug.cgi?id=48838
3093
3094         * MiniBrowser/mac/WebBundle/WebBundleMain.m:
3095         (WKBundleInitialize):
3096         * TestWebKitAPI/InjectedBundleController.cpp:
3097         (TestWebKitAPI::InjectedBundleController::initialize):
3098         (TestWebKitAPI::InjectedBundleController::didReceiveMessage):
3099         * TestWebKitAPI/InjectedBundleController.h:
3100         * TestWebKitAPI/InjectedBundleMain.cpp:
3101         (WKBundleInitialize):
3102         * TestWebKitAPI/PlatformUtilities.cpp:
3103         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
3104         * WebKitTestRunner/InjectedBundle/InjectedBundleMain.cpp:
3105         (WKBundleInitialize):
3106         Update implementations of WKBundleInitialize to take an initial userData
3107         argument.  Change TestWebKitAPI to use the new initial userData to initialize
3108         each test's bundle.
3109
3110 2010-11-02  Benjamin Kalman  <kalman@google.com>
3111
3112         Reviewed by Ojan Vafai.
3113
3114         new-run-webkit-tests doesn't strip "LayoutTests/" from prefix, unlike old-run-webkit-tests
3115         https://bugs.webkit.org/show_bug.cgi?id=48794
3116
3117         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: Strip the "LayoutTests/" prefix from test argument paths.
3118
3119 2010-11-02  Adam Roben  <aroben@apple.com>
3120
3121         Skip webkitpy.layout_tests.run_webkit_tests_unittest.MainTest on Cygwin
3122         Python 2.5.x
3123
3124         It is known to hang on that version of Python. See
3125         <http://webkit.org/b/48614>.
3126
3127         Reviewed by Adam Barth.
3128
3129         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py: Use
3130         skip_if to skip MainTest on Cygwin Python 2.5.x.
3131
3132         * Scripts/webkitpy/test/skip.py: Added.
3133         (skip_if): If the passed-in condition is false, find all the test_*
3134         methods of the given class and replace them with a function that just
3135         logs that we're skipping these tests. This is loosely based on the
3136         unittest.skip_if decorator added in Python 3.1.
3137         (_skipped_method): Returns a function that just logs that the tests are
3138         being skipped. This is used in place of the actual test_* functions.
3139
3140         * Scripts/webkitpy/test/skip_unittest.py: Added.
3141         (SkipTest.setUp): Create a logger for ourselves and give it a handler
3142         that logs to a private stream.
3143         (SkipTest.tearDown): Reset the logger.
3144         (SkipTest.create_fixture_class): Returns a class that we use to test
3145         skip_if. It has a single test_* method, test_foo, that just calls
3146         through to a callback.
3147         (SkipTest.foo_callback): Record that test_foo was called.
3148         (SkipTest.test_skip_if_false): Pass skip_if a False condition and test
3149         that test_foo does get called.
3150         (SkipTest.test_skip_if_true): Pass skip_if a True condition and test
3151         that test_foo does not get called and the appropriate message gets
3152         logged.
3153
3154 2010-11-02  Adam Barth  <abarth@webkit.org>
3155
3156         Reviewed by Eric Seidel.
3157
3158         webkit-patch should tell check-webkit-style which files were changed so
3159         check-webkit-style doesn't have to stat the whole working copy again
3160         https://bugs.webkit.org/show_bug.cgi?id=48792
3161
3162         * Scripts/webkitpy/tool/mocktool.py:
3163         * Scripts/webkitpy/tool/steps/checkstyle.py:
3164
3165 2010-11-02  Robert Kroeger  <rjkroege@chromium.org>
3166
3167         Reviewed by James Robinson.
3168
3169         [Chromium/DRT] Make EventSendingController honour leapForward for touch events.
3170         https://bugs.webkit.org/show_bug.cgi?id=48777
3171
3172         * DumpRenderTree/chromium/EventSender.cpp:
3173         (EventSender::sendCurrentTouchEvent):
3174
3175 2010-11-02  Adam Roben  <aroben@apple.com>
3176
3177         Only track resource identifiers in DRT when dumpResourceLoadCallbacks
3178         is on
3179
3180         This reverts Windows to our pre-r71097 behavior. That patch made us
3181         track all resource identifiers, including the main resource, so the
3182         main resource's URL started appearing in test output instead of
3183         "<unknown>". Arguably having the main resource's URL is better, but all
3184         other platforms print "<unknown>" and we want to match.
3185
3186         Fixes <http://webkit.org/b/48837> <rdar://problem/8620351> REGRESSION
3187         (r71097): Many http tests failing on Windows
3188
3189         Reviewed by Anders Carlsson.
3190
3191         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
3192         (ResourceLoadDelegate::identifierForInitialRequest): Don't add the
3193         identifier to the URL map if we're not supposed to dump resource load
3194         callbacks.
3195         (ResourceLoadDelegate::removeIdentifierForRequest): Always remove the
3196         identifier from the URL map even if we're already "done". There's no
3197         point in keeping out-of-date identifiers around.
3198
3199 2010-11-01  Jenn Braithwaite  <jennb@chromium.org>
3200
3201         Reviewed by Adam Roben.
3202
3203         Windows: Update resource tracking when moving a frame between documents
3204         https://bugs.webkit.org/show_bug.cgi?id=48364
3205
3206         * DumpRenderTree/win/DumpRenderTree.cpp:
3207         (createWebViewAndOffscreenWindow):
3208         (main):
3209         Give each WebView its own ResourceLoadDelegate instance in order to
3210         make assertions about resource ids on a particular WebView.
3211         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
3212         (ResourceLoadDelegate::identifierForInitialRequest):
3213         Always add id to the map.
3214         (ResourceLoadDelegate::removeIdentifierForRequest):
3215         Added.
3216         (ResourceLoadDelegate::willSendRequest):
3217         (ResourceLoadDelegate::didReceiveAuthenticationChallenge):
3218         (ResourceLoadDelegate::didReceiveResponse):
3219         (ResourceLoadDelegate::didFinishLoadingFromDataSource):
3220         (ResourceLoadDelegate::didFailLoadingWithError):
3221         (ResourceLoadDelegate::descriptionSuitableForTestResult):
3222         Replace static descriptionSuitableForTestResult with static member function to access identifier map.
3223         * DumpRenderTree/win/ResourceLoadDelegate.h:
3224         (ResourceLoadDelegate::urlMap):
3225         Moved within class so that each WebView has its own id map.
3226
3227 2010-11-01  Ojan Vafai  <ojan@chromium.org>
3228
3229         Reviewed by Dimitri Glazkov.
3230
3231         remove debug code from run_webkit_tests.py
3232         https://bugs.webkit.org/show_bug.cgi?id=48800
3233
3234         Remove temporary debug code and make --master-name required
3235         if --test-results-server is set now that all clients set
3236         --master-name.
3237
3238         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3239
3240 2010-11-01  Tony Chang  <tony@chromium.org>
3241
3242         Reviewed by Kent Tamura.
3243
3244         add plugin event logging for linux
3245         https://bugs.webkit.org/show_bug.cgi?id=48779
3246
3247         This is taken from Chromium's fork of the layout test plugin:
3248         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
3249
3250         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
3251         (handleEventX11):
3252         (NPP_HandleEvent):
3253
3254 2010-11-01  Dirk Pranke  <dpranke@chromium.org>
3255
3256         Reviewed by James Robinson.
3257
3258         new-run-webkit-tests: use DRT, child-processes=1 for GPU tests by default
3259         https://bugs.webkit.org/show_bug.cgi?id=48790
3260
3261         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
3262         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
3263
3264 2010-11-01  Adam Barth  <abarth@webkit.org>
3265
3266         Reviewed by Eric Seidel.
3267
3268         Teach check-webkit-style how to accept a list of files to diff on the
3269         command line
3270         https://bugs.webkit.org/show_bug.cgi?id=48784
3271
3272         In a future patch, webkit-patch will use this option to improve
3273         performance.  I'm landing this in two pieces to avoid causing a version
3274         skew problem for the style-bot.
3275
3276         * Scripts/check-webkit-style:
3277         * Scripts/webkitpy/style/optparser.py:
3278         * Scripts/webkitpy/style/optparser_unittest.py:
3279         * Scripts/webkitpy/style_references.py:
3280
3281 2010-11-01  Anders Carlsson  <andersca@apple.com>
3282
3283         Reviewed by John Sullivan.
3284
3285         Tear down the related WebProcessProxy when a WebContext is deallocated
3286         https://bugs.webkit.org/show_bug.cgi?id=48769
3287
3288         * TestWebKitAPI/Tests/WebKit2/FailedLoad.cpp:
3289         (TestWebKitAPI::didFailProvisionalLoadWithErrorForFrame):
3290         We don't support empty URLs anymore, update test to expect a null URL instead.
3291
3292 2010-11-01  Søren Gjesse  <sgjesse@chromium.org>
3293
3294         Reviewed by Andreas Kling.
3295
3296         Fix warning when compiling the chromium port of DumpRenderShell
3297         with clang.
3298         https://bugs.webkit.org/show_bug.cgi?id=48414
3299
3300         * DumpRenderTree/chromium/TestShell.h:
3301         (TestShell::javaScriptFlagsForLoad):
3302
3303 2010-11-01  Adam Roben  <aroben@apple.com>
3304
3305         Fix typo from r71022
3306
3307         * BuildSlaveSupport/build.webkit.org-config/config.json:
3308
3309 2010-11-01  Adam Roben  <aroben@apple.com>
3310
3311         Trigger the Windows Release WebKit2 tests when a Release build
3312         finishes, not when a Debug build finishes
3313
3314         Fixes <http://webkit.org/b/48754> Windows Release WebKit2 tests are
3315         triggered at the wrong time
3316
3317         Reviewed by Sam Weinig.
3318
3319         * BuildSlaveSupport/build.webkit.org-config/config.json: Fixed
3320         triggerable name and trigger.
3321
3322 2010-11-01  Mario Sanchez Prada  <msanchez@igalia.com>
3323
3324         Unreviewed. Adding my IRC nickname to the list of committers.
3325
3326         * Scripts/webkitpy/common/config/committers.py:
3327
3328 2010-10-31  Robert Hogan  <robert@webkit.org>
3329
3330         Reviewed by Antonio Gomes.
3331
3332         [Qt] [Gtk] Plug-ins having upper case in mime type are failing to load
3333
3334         Qt and Gtk are case-sensitive when storing the declared mime-type
3335         of plugins. Since plugin mime-types are lowercased prior to searching
3336         for them in the plugin database, ensure they are loaded with the
3337         mime-type in lower case too.
3338
3339         Change the test netscape plugin to declare its mimetype in sentence
3340         case so that the correct behaviour is enforced.
3341
3342         https://bugs.webkit.org/show_bug.cgi?id=36815
3343
3344         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
3345         (NP_GetMIMEDescription):
3346
3347 2010-10-31  Robert Hogan  <robert@webkit.org>
3348
3349         Reviewed by Antonio Gomes.
3350
3351         [Qt] Support nodesFromRect in DRT
3352
3353         https://bugs.webkit.org/show_bug.cgi?id=48716
3354
3355         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3356         (LayoutTestController::nodesFromRect):
3357         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3358
3359 2010-10-30  Andreas Kling  <kling@webkit.org>
3360
3361         Reviewed by Antonio Gomes.
3362
3363         [Qt] QtTestBrowser: Switching view type moves the embedded inspector
3364         https://bugs.webkit.org/show_bug.cgi?id=48705
3365
3366         Reinsert the embedded inspector into the splitter after changing
3367         between QWebView/QGraphicsWebView.
3368
3369         * QtTestBrowser/launcherwindow.cpp:
3370         (LauncherWindow::init):
3371         (LauncherWindow::initializeView):
3372         * QtTestBrowser/webinspector.h:
3373         (WebInspector::WebInspector):
3374
3375 2010-10-28  Antonio Gomes  <agomes@rim.com>
3376
3377         Reviewed by Ojan Vafai.
3378
3379         Needs a "LinuxEditingBehavior", perhaps with a better name
3380         https://bugs.webkit.org/show_bug.cgi?id=36627
3381
3382         Adding support to Mac's, GTK+'s, Windows' and Chromium's LayoutTestController class to test the newly introduced Unix editing behavior.
3383
3384         * DumpRenderTree/chromium/LayoutTestController.cpp:
3385         (LayoutTestController::setEditingBehavior):
3386         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3387         (LayoutTestController::setEditingBehavior):
3388         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3389         (LayoutTestController::setEditingBehavior):
3390         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3391         (LayoutTestController::setEditingBehavior):
3392
3393 2010-10-29  Andreas Kling  <kling@webkit.org>
3394
3395         Reviewed by Kenneth Rohde Christiansen.
3396
3397         [Qt] QtTestBrowser: Add keyboard shortcut to toggle full screen (F11)
3398         https://bugs.webkit.org/show_bug.cgi?id=48695
3399
3400         * QtTestBrowser/launcherwindow.cpp:
3401         (LauncherWindow::createChrome):
3402
3403 2010-10-29  Andreas Kling  <kling@webkit.org>
3404
3405         Reviewed by Kenneth Rohde Christiansen.
3406
3407         [Qt] QtTestBrowser: Fix uninitialized read in FpsTimer
3408         https://bugs.webkit.org/show_bug.cgi?id=48675
3409
3410         FpsTimer::m_timer was never initialized and passed to QObject::killTimer()
3411         on startup with in -graphicsbased mode.
3412
3413         * QtTestBrowser/fpstimer.cpp:
3414         (FpsTimer::FpsTimer):
3415         (FpsTimer::numFrames):
3416         (FpsTimer::stop):
3417         (FpsTimer::timerEvent):
3418         * QtTestBrowser/fpstimer.h:
3419
3420 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
3421
3422         Reviewed by Tony Chang.
3423
3424         new-run-webkit-tests: change TestResults to be serializable
3425
3426         In preparation for changing new-run-webkit-tests from
3427         multithreaded to multiprocess, we need to make sure the data
3428         going between the threads is easily serialized over a socket.
3429
3430         This change adds serialization/pickling for the TestResults and
3431         TestFailure objects (using cPickle).
3432
3433         The TestFailure objects included a "has_wdiff" flag for Text
3434         results, but the flag wasn't being used, so I've removed it,
3435         simplifying the state to basically a set of enum objects with
3436         associated methods.
3437
3438         https://bugs.webkit.org/show_bug.cgi?id=48616
3439
3440         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3441         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
3442         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
3443         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
3444         * Scripts/webkitpy/layout_tests/layout_package/test_results.py: Added.
3445         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:Added. 
3446         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3447         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
3448
3449 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
3450
3451         Reviewed by Ojan Vafai.
3452
3453         new-run-webkit-tests: need to provide separate dashboard results for GPU tests
3454         https://bugs.webkit.org/show_bug.cgi?id=48687
3455
3456         Modify the steps to upload the JSON files to the dashboards so
3457         that the GPU tests don't conflict with the regular tests on a
3458         port. We do this by modifying the --builder-name parameter to
3459         append " - GPU", which should cause the app to treat the results
3460         as a completely new builder. This is a little non-obvious, but
3461         keeps us from having to restructure the app.
3462
3463         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
3464         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
3465         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3466
3467 2010-10-29  Ojan Vafai  <ojan@chromium.org>
3468
3469         Reviewed by Tony Chang.
3470
3471         [chromium] add debug logging to help diagnose flakiness dashboard issues
3472         https://bugs.webkit.org/show_bug.cgi?id=48657
3473
3474         The appengine app thinks it's getting empty files uploaded. Add some logging
3475         to see if new-run-webkit-tests agrees.
3476         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3477
3478 2010-10-29  Dimitri Glazkov  <dglazkov@chromium.org>
3479
3480         Reviewed by Ojan Vafai.
3481
3482         [Chromium/DRT] Add master-name flag to new-run-webkit-tests.
3483         https://bugs.webkit.org/show_bug.cgi?id=48649
3484
3485         * BuildSlaveSupport/build.webkit.org-config/master.cfg: Added flag.
3486
3487 2010-10-29  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
3488
3489         Reviewed by Martin Robinson.
3490
3491         Enable popup window in GtkLauncher
3492         https://bugs.webkit.org/show_bug.cgi?id=48335
3493
3494         The GtkLauncher application does not support opening new window when
3495         a link with "target=_blank" is clicked or similar call to
3496         window.open(). Instead, GtkLauncher does nothing which breaks
3497         navigation of some websites.
3498
3499         * GtkLauncher/main.c:
3500         (activate_uri_entry_cb):
3501         (update_title):
3502         (link_hover_cb):
3503         (notify_title_cb):
3504         (notify_load_status_cb):
3505         (notify_progress_cb):
3506         (destroy_cb):
3507         (go_back_cb):
3508         (go_forward_cb):
3509         (create_web_view_cb):
3510         (web_view_ready_cb):
3511         (close_web_view_cb):
3512         (create_browser):
3513         (create_statusbar):
3514         (create_toolbar):
3515         (create_window):
3516         (main):
3517
3518 2010-10-29  Adam Roben  <aroben@apple.com>
3519
3520         Teach check-webkit-style about WebKit2's idiosyncracies
3521
3522         Fixes <http://webkit.org/b/48638> Style bot complains about a number
3523         of WebKit2 conventions
3524
3525         Reviewed by Anders Carlsson.
3526
3527         * Scripts/webkitpy/style/checker.py: Excluded some rules for various
3528         WebKit2-related files. Also updated the excluded rules for
3529         WebKitAPITest to match the current code.
3530
3531 2010-10-29  Csaba Osztrogonác  <ossy@webkit.org>
3532
3533         Reviewed by Adam Roben and David Kilzer.
3534
3535         Fix and cleanup of build systems
3536         https://bugs.webkit.org/show_bug.cgi?id=48342
3537
3538         * Scripts/build-webkit:
3539          - Remove unnecessary ENABLE_SANDBOX option.
3540          - Add ENABLE_FULLSCREEN_API option.
3541
3542 2010-10-28  Adam Roben  <aroben@apple.com>
3543
3544         Switch the Windows WebKit2 bot to the Release configuration
3545
3546         We only have one machine testing WebKit2 on Windows right now, and
3547         Debug is just too slow for it to keep up.
3548
3549         Fixes (hopefully!) <http://webkit.org/b/48615> Windows WebKit2 bot is
3550         always way behind
3551
3552         Reviewed by Jon Honeycutt.
3553
3554         * BuildSlaveSupport/build.webkit.org-config/config.json:
3555
3556 2010-10-29  Leandro Gracia Gil  <leandrogracia@google.com>
3557
3558         Reviewed by Jeremy Orlow.
3559
3560         Added a second parameter to setMockSpeechInputResult for
3561         the language used in speech input.
3562         https://bugs.webkit.org/show_bug.cgi?id=47089
3563
3564         * DumpRenderTree/LayoutTestController.cpp:
3565         (setMockSpeechInputResultCallback):
3566         * DumpRenderTree/LayoutTestController.h:
3567         * DumpRenderTree/chromium/LayoutTestController.cpp:
3568         (LayoutTestController::setMockSpeechInputResult):
3569         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3570         (LayoutTestController::setMockSpeechInputResult):
3571         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3572         (LayoutTestController::setMockSpeechInputResult):
3573         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3574         (LayoutTestController::setMockSpeechInputResult):
3575         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3576         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3577         (LayoutTestController::setMockSpeechInputResult):
3578         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3579         (LayoutTestController::setMockSpeechInputResult):
3580         (LayoutTestController::pageSizeAndMarginsInPixels):
3581
3582 2010-10-28  Ojan Vafai  <ojan@chromium.org>
3583
3584         Reviewed by Tony Chang.
3585
3586         [chromium] add a result-small.json file for the test dashboard
3587         https://bugs.webkit.org/show_bug.cgi?id=48547
3588
3589         Output both a results.json file and a results-small.json file.
3590         The dashboard will load results-small.json by default so it loads faster.
3591
3592         * TestResultServer/model/jsonresults.py:
3593         * TestResultServer/model/jsonresults_unittest.py:
3594         Added a bunch of sys.path hackery. Unforunately, this uses hardcoded
3595         paths. That obviously needs to be fixed, but at least this way it
3596         clearly documents what paths are necessary.
3597
3598 2010-10-28  Eric Seidel  <eric@webkit.org>
3599
3600         Reviewed by Adam Barth.
3601
3602         webkit-patch upload calls changed_files more often than it should
3603         https://bugs.webkit.org/show_bug.cgi?id=48567
3604
3605         Passing changed_files around everywhere isn't a very elegant solution
3606         but it's the one we have for the moment.  I think keeping an explicit
3607         cache on Checkout (or making StepState() a real class) is a better
3608         long-term option.
3609
3610         Previously bug_id_for_this_commit was calling changed_files and the
3611         result was never getting cached on the state.  Now we're explicitly
3612         caching the result on the state and passing that to the bug_id_for_this_commit call.
3613
3614         I looked into building unit tests for this.  Doing so would require
3615         using a real Checkout object with a MockSCM and overriding the appropriate
3616         calls on SCM to count how often we're stating the file system.
3617         That's a useful set of tests to build for a separate change.
3618
3619         * Scripts/webkitpy/common/checkout/api.py:
3620         * Scripts/webkitpy/tool/commands/download.py:
3621         * Scripts/webkitpy/tool/commands/upload.py:
3622         * Scripts/webkitpy/tool/mocktool.py:
3623
3624 2010-10-28  Eric Seidel  <eric@webkit.org>
3625
3626         Reviewed by Adam Barth.
3627
3628         Make suggest-reviewers slightly faster
3629         https://bugs.webkit.org/show_bug.cgi?id=48562
3630
3631         Add @memoized to one more common call.
3632
3633         * Scripts/webkitpy/common/checkout/api.py:
3634
3635 2010-10-28  Eric Seidel  <eric@webkit.org>
3636
3637         Reviewed by Adam Barth.
3638
3639         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is still very slow
3640         https://bugs.webkit.org/show_bug.cgi?id=48536
3641
3642         This patch makes it about 40 seconds faster, but it still
3643         takes 1:40.  This will require more refinement.
3644
3645         The suggested reviewers list appears to be the same.
3646
3647         I think the next step may be to have it stop the search after
3648         5 reviewers are found.  We never want to suggest 30 people.
3649
3650         * Scripts/webkitpy/common/checkout/scm.py:
3651          - Using --remove-empty to theoretically stop lookups past
3652            when a file is removed.  I'm not entirely clear that the option
3653            does what it says it does.  Example:
3654            git log --pretty=format:%H -5 --remove-empty -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
3655            returns only one commit
3656            vs.
3657            git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
3658            which returns 5.  I was not aware that wx files were ever removed from the repository?