2010-11-18 Dirk Pranke <dpranke@chromium.org>
[WebKit.git] / WebKitTools / ChangeLog
1 2010-11-18  Dirk Pranke  <dpranke@chromium.org>
2
3         Reviewed by Tony Chang.
4
5         new-run-webkit-tests: create first part of 'message_broker' class for multiprocessing fixes
6
7         Create the first version of the 'message_broker' package. This
8         class will encapsulate all of the threading/multiprocessing and
9         message-sending details for the communication between the
10         'manager' object and the 'worker' objects. For the moment, it
11         just holds some routines and tests for logging thread stacks.
12
13         There should be no functional changes in this patch, just moving stuff
14         around.
15
16         https://bugs.webkit.org/show_bug.cgi?id=49707
17
18         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
19         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread_unittest.py: Removed.
20         * Scripts/webkitpy/layout_tests/layout_package/message_broker.py: Added.
21         * Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py: Added.
22         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
23         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
24
25 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
26
27         Reviewed by Adam Roben.
28
29         Remove leftover Windows Debug_Internal configurations
30         https://bugs.webkit.org/show_bug.cgi?id=49758
31
32         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
33         * DumpRenderTree/win/DumpRenderTree.vcproj:
34         * DumpRenderTree/win/ImageDiff.vcproj:
35         * MiniBrowser/MiniBrowser.vcproj:
36         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
37         * WebKitAPITest/WebKitAPITest.vcproj:
38         * WebKitTestRunner/win/InjectedBundle.vcproj:
39         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
40         * WinLauncher/WinLauncher.vcproj:
41
42 2010-11-18  Adam Roben  <aroben@apple.com>
43
44         Add three new Windows XP build slaves
45
46         Reviewed by Steve Falkenburg.
47
48         * BuildSlaveSupport/build.webkit.org-config/config.json: Added three
49         new slaves, and added them to the Windows XP Debug (Tests) builder.
50
51 2010-11-18  Daniel Bates  <dbates@rim.com>
52
53         Reviewed by Adam Roben.
54
55         update-webkit-support-libs should fall back to existing
56         WebKitSupportLibrary version if there is no internet connectivity
57         https://bugs.webkit.org/show_bug.cgi?id=49503
58
59         Fall back to existing support libraries (if present) when there is
60         no internet connection.
61
62         Currently, update-webkit-support-libs dies with an "out-of-date"
63         error when there is no internet connection because it cannot
64         retrieve versioning information from developer.apple.com. Because
65         update-webkit-support-libs fails, build-webkit fails. Instead,
66         if there is no internet connection and the support libraries are
67         present then we should warn the user and exit() with success so
68         that build-webkit can work without an internet connection.
69
70         * Scripts/update-webkit-support-libs:
71
72 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
73
74         Reviewed by Adam Roben.
75
76         Debug_Internal Windows configuration is unnecessary, should be removed
77         https://bugs.webkit.org/show_bug.cgi?id=49753
78
79         * DumpRenderTree/win/DumpRenderTree.cpp:
80         * MiniBrowser/win/stdafx.h:
81         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp:
82         * WebKitTestRunner/win/TestControllerWin.cpp:
83
84 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
85
86         Reviewed by Adam Roben.
87
88         Clean up vcproj errors
89         https://bugs.webkit.org/show_bug.cgi?id=49741
90
91         * MiniBrowser/MiniBrowser.vcproj:
92         * WebKitAPITest/WebKitAPITest.vcproj:
93         * WebKitTestRunner/win/InjectedBundle.vcproj:
94         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
95
96 2010-11-18  Hayato Ito  <hayato@chromium.org>
97
98         Reviewed by Shinichiro Hamaji.
99
100         [NRWT] Fix a break of '--new-baseline' feature in pixel_test
101         which is caused by r72249.
102
103         https://bugs.webkit.org/show_bug.cgi?id=49751
104
105         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
106
107 2010-11-18  Adam Roben  <aroben@apple.com>
108
109         Implement layoutTestController.waitForPolicyDelegate on Windows
110
111         Fixes <http://webkit.org/b/25038> <rdar://problem/6790213>.
112
113         Reviewed by Simon Fraser.
114
115         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
116         (LayoutTestController::waitForPolicyDelegate): Implemented by porting
117         code from LayoutTestControllerMac.mm.
118
119 2010-11-18  Hayato Ito  <hayato@chromium.org>
120
121         Reviewed by Shinichiro Hamaji.
122
123         [NRWT] Fix a break of '--new-baseline' feature which is caused by r72249.
124
125         https://bugs.webkit.org/show_bug.cgi?id=49740
126
127         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
128
129 2010-11-18  Tony Chang  <tony@chromium.org>
130
131         Reviewed by Adam Barth.
132
133         run platform/chromium/plugins/refcount-leaks.html on all platforms
134         https://bugs.webkit.org/show_bug.cgi?id=49485
135
136         Add PluginObject.testObjectCount which returns the number of allocated
137         TestObjects. Add PluginObject.testCreateTestObject which allocates
138         and returns a TestObject.
139
140         Add TestObject.refCount which returns the number of refs on the
141         TestObject.
142
143         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
144         (pluginGetProperty):
145         (pluginInvoke):
146         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
147         (getTestObjectCount):
148         (testAllocate):
149         (testDeallocate):
150         (testGetProperty):
151         (testScriptObjectInvoke): Release a ref to a plugin object that was
152             previously leaked.
153         * DumpRenderTree/TestNetscapePlugIn/TestObject.h:
154
155 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
156
157         Reviewed by Adam Roben.
158
159         WebKitTools projects (WebKitTestRunner, record-memory-win) should use vsprops for common build settings
160         https://bugs.webkit.org/show_bug.cgi?id=49711
161
162         * WebKitTestRunner/Configurations/WebKitTestRunnerCFLite.vsprops: Added.
163         * WebKitTestRunner/Configurations/WebKitTestRunnerCommon.vsprops: Added.
164         * WebKitTestRunner/Configurations/WebKitTestRunnerCoreFoundation.vsprops: Added.
165         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
166         * record-memory-win/record-memory-win-common.vsprops: Added.
167         * record-memory-win/record-memory-win.vcproj:
168
169 2010-11-18  Andras Becsi  <abecsi@inf.u-szeged.hu>
170
171         Reviewed by Csaba Osztrogonác.
172
173         [Qt][WK2] Only add user agent strings to the list which aren't listed yet.
174
175         * MiniBrowser/qt/BrowserWindow.cpp:
176         (BrowserWindow::updateUserAgentList):
177
178 2010-11-17  Adam Roben  <aroben@apple.com>
179
180         Don't trigger Windows builds when chromium-win test results change
181
182         Reviewed by Mark Rowe.
183
184         * Scripts/webkitpy/common/config/build.py:
185         (_should_file_trigger_build): Changed the regular expression used to
186         search for directory names to only consider full directory names,
187         rather than matching directory names where the string we care about is
188         a suffix of the name.
189
190         * Scripts/webkitpy/common/config/build_unittest.py:
191         (ShouldBuildTest): Added a test.
192
193 2010-11-17  Hayato Ito  <hayato@chromium.org>
194
195         Reviewed by Shinichiro Hamaji.
196
197         Remove a uri member from TestInput class.
198
199         https://bugs.webkit.org/show_bug.cgi?id=49691
200
201         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
202         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
203
204 2010-11-17  Adam Roben  <aroben@apple.com>
205
206         Move FindSafari's settings into a vsprops file
207
208         Fixes <http://webkit.org/b/49699> FindSafari should use a vsprops file
209
210         Reviewed by Steve Falkenburg.
211
212         * FindSafari/FindSafari.vcproj:
213         * FindSafari/FindSafariCommon.vsprops: Added.
214
215 2010-11-17  Adam Roben  <aroben@apple.com>
216
217         Use a minimum font size of 0 in DRT on Windows
218
219         This is the Windows equivalent of r72164.
220
221         Rubber-stamped by Dave Hyatt.
222
223         * DumpRenderTree/win/DumpRenderTree.cpp:
224         (resetDefaultsToConsistentValues):
225
226 2010-11-17  Dirk Pranke  <dpranke@chromium.org>
227
228         Reviewed by Andreas Kling.
229
230         webkit-patch pretty-diff shouldn't need to be run from a checkout root in git
231         https://bugs.webkit.org/show_bug.cgi?id=49639
232
233         * Scripts/webkitpy/common/checkout/scm.py:
234
235 2010-11-17  Mihai Parparita  <mihaip@chromium.org>
236
237         Reviewed by Tony Chang.
238
239         Rebaseline server: add loupe for image diffs
240         https://bugs.webkit.org/show_bug.cgi?id=49692
241         
242         Add a loupe (magnifiying glass) for inspecting image diffs. Shows an
243         enlarged area of the expected, actual and diff images side by side.
244
245         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
246         * Scripts/webkitpy/tool/commands/data/rebaselineserver/loupe.js: Added.
247         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
248         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
249
250 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
251
252         Reviewed by Adam Roben.
253
254         WebKitLauncherWin should use vsprops for shared build settings
255         https://bugs.webkit.org/show_bug.cgi?id=49696
256
257         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
258         * WebKitLauncherWin/WebKitLauncherWinCommon.vsprops: Added.
259
260 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
261
262         Reviewed by Adam Roben.
263
264         WinLauncher should use vsprops for shared build settings
265         https://bugs.webkit.org/show_bug.cgi?id=49695
266
267         * WinLauncher/WinLauncher.cpp:
268         (_tWinMain):
269         (PrintView):
270         (WndProc):
271         (MyEditProc):
272         * WinLauncher/WinLauncher.vcproj:
273         * WinLauncher/WinLauncherCommon.vsprops: Added.
274
275 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
276
277         Delete unused file unintentionally added in last change.
278
279         * DumpRenderTree/win/DumpRenderTreeCURL.vsprops: Removed.
280
281 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
282
283         Reviewed by Adam Roben.
284
285         DumpRenderTree on Windows should use a vsprops file for shared project settings
286         https://bugs.webkit.org/show_bug.cgi?id=49690
287
288         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
289         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePluginCommon.vsprops: Added.
290         * DumpRenderTree/win/DumpRenderTree.vcproj:
291         * DumpRenderTree/win/DumpRenderTreeApple.vsprops: Added.
292         * DumpRenderTree/win/DumpRenderTreeCFLite.vsprops: Added.
293         * DumpRenderTree/win/DumpRenderTreeCURL.vsprops: Added.
294         * DumpRenderTree/win/DumpRenderTreeCairo.vsprops: Added.
295         * DumpRenderTree/win/DumpRenderTreeCommon.vsprops: Added.
296         * DumpRenderTree/win/ImageDiff.vcproj:
297         * DumpRenderTree/win/ImageDiffCommon.vsprops: Added.
298
299 2010-11-17  Hayato Ito  <hayato@chromium.org>
300
301         Refactor TestTypeBase.compare_output().
302
303         Introduce a TestOutput class and update compare_output() of each test
304         types so that they can take both actual and expected TestOutput objects.
305
306         https://bugs.webkit.org/show_bug.cgi?id=49431
307
308         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
309         * Scripts/webkitpy/layout_tests/layout_package/test_output.py: Added.
310         * Scripts/webkitpy/layout_tests/port/base.py:
311         * Scripts/webkitpy/layout_tests/port/chromium.py:
312         * Scripts/webkitpy/layout_tests/port/dryrun.py:
313         * Scripts/webkitpy/layout_tests/port/test.py:
314         * Scripts/webkitpy/layout_tests/port/webkit.py:
315         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
316         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
317         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
318         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
319
320 2010-11-17  Adam Roben  <aroben@apple.com>
321
322         Make each Windows Test builder use the same OS for all its slaves
323
324         Currently each WebKit1 Windows Test builder has one Windows XP slave
325         and one Windows 7 slave. Having different OSes in a single builder
326         meant that the builders would alternate between red and green when
327         there were OS-specific failures.
328
329         Fixes <http://webkit.org/b/49688> Windows Test builders alternate
330         between red and green because of multiple OS versions
331
332         Reviewed by Mark Rowe.
333
334         * BuildSlaveSupport/build.webkit.org-config/config.json: Renamed
335         "Windows Release (Tests)" to "Windows 7 Release (Tests)", "Windows
336         Debug (Tests)" to "Windows XP Debug (Tests)", and "Windows Release
337         (WebKit2 Tests)" to "Windows 7 Release (WebKit2 Tests)". Updated the
338         builders for the two WebKit1 builders so that they would all use the
339         appropriate OS.
340
341 2010-11-17  James Robinson  <jamesr@chromium.org>
342
343         Reviewed by Dan Bernstein.
344
345         [chromium] Set minimum font size to 0 for DRT
346         https://bugs.webkit.org/show_bug.cgi?id=49677
347
348         Followup for http://trac.webkit.org/changeset/72141.
349
350         Test: http/tests/misc/acid3.html
351
352         * DumpRenderTree/chromium/WebPreferences.cpp:
353         (WebPreferences::reset):
354
355 2010-11-17  Andras Becsi  <abecsi@webkit.org>
356
357         Unreviewed build fix.
358
359         [Qt] Add missing file I forgot in r72220.
360
361         * MiniBrowser/MiniBrowser.qrc: Added.
362
363 2010-11-17  Andras Becsi  <abecsi@webkit.org>
364
365         Reviewed by Andreas Kling.
366
367         [Qt][WK2] Make it possible to set custom user agent strings in MiniBrowser.
368         https://bugs.webkit.org/show_bug.cgi?id=49627
369
370         Add a user agent dialog and a resource file for useragentlist.txt to MiniBrowser.
371
372         * MiniBrowser/MiniBrowser.qrc: Added.
373         Add it here to prevent qmake from detecting it
374         since it needs to be copied to the build directory.
375         * MiniBrowser/qt/BrowserWindow.cpp:
376         (BrowserWindow::BrowserWindow):
377         (BrowserWindow::updateUserAgentList):
378         (BrowserWindow::showUserAgentDialog):
379         * MiniBrowser/qt/BrowserWindow.h:
380         * MiniBrowser/qt/MiniBrowser.pro:
381
382 2010-11-17  Adam Roben  <aroben@apple.com>
383
384         Make changes to the Mac WebKit2 Skipped file trigger Windows builds
385
386         We use that file on Windows!
387
388         Fixes <http://webkit.org/b/49643> Changes to
389         LayoutTests/platform/mac-wk2 should trigger a Windows build, but don't
390
391         Reviewed by Steve Falkenburg.
392
393         * Scripts/webkitpy/common/config/build.py:
394         (_should_file_trigger_build): Added the mac-wk2 directory and make it
395         trigger builds on SnowLeopard and Windows.
396
397         * Scripts/webkitpy/common/config/build_unittest.py:
398         (ShouldBuildTest.test_should_build): Added a test.
399
400 2010-11-17  Satish Sampath  <satish@chromium.org>
401
402         Reviewed by Jeremy Orlow.
403
404         Clear the speech input mock explicitly before each test.
405         https://bugs.webkit.org/show_bug.cgi?id=49660
406
407         * DumpRenderTree/chromium/LayoutTestController.cpp:
408         (LayoutTestController::setMockSpeechInputResult):
409         * DumpRenderTree/chromium/LayoutTestController.h:
410         * DumpRenderTree/chromium/WebViewHost.cpp:
411         (WebViewHost::speechInputController):
412         (WebViewHost::reset): Invoke speech input mock's clearResults method.
413         * DumpRenderTree/chromium/WebViewHost.h:
414         (WebViewHost::speechInputControllerMock): Moved mock from LayoutTestController.
415
416 2010-11-17  Alexander Pavlov  <apavlov@chromium.org>
417
418         Unreviewed, specify IRC nickname in committers.py.
419
420         * Scripts/webkitpy/common/config/committers.py:
421
422 2010-11-17  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
423
424         Reviewed by Tony Chang.
425
426         [NRWT] Make http locking similar to perl implementation
427         https://bugs.webkit.org/show_bug.cgi?id=49187
428
429         * Scripts/webkitpy/common/system/file_lock.py: Added.
430         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
431         * Scripts/webkitpy/layout_tests/port/http_lock.py:
432
433 2010-11-17  MORITA Hajime  <morrita@google.com>
434
435         Reviewed by Kent Tamura.
436
437         [Chromium][DRT] EventSender.contextClick() should aware spellchecking
438         https://bugs.webkit.org/show_bug.cgi?id=49366
439
440         - EvenSender: Checked WebContextMenuData.misspelledWord and added extra context menu entries
441           according to the spellchecker suggestion
442         - MockSpellCheck: add fillSuggestionList to provide fake suggestions.
443
444         test_expectations.txt will be changed after this change is ported to test_shell.
445         
446         * DumpRenderTree/chromium/EventSender.cpp:
447         (makeMenuItemStringsFor):
448         (EventSender::contextClick):
449         * DumpRenderTree/chromium/MockSpellCheck.cpp:
450         (MockSpellCheck::fillSuggestionList):
451         (MockSpellCheck::initializeIfNeeded):
452         * DumpRenderTree/chromium/MockSpellCheck.h:
453         * DumpRenderTree/chromium/WebViewHost.cpp:
454         (WebViewHost::mockSpellCheck):
455         * DumpRenderTree/chromium/WebViewHost.h:
456
457 2010-11-16  Dave Hyatt  <hyatt@apple.com>
458
459         Reviewed by Sam Weinig.
460
461         Make sure the pref for minimum font size is just explicitly set to 0 every time,
462         since the bots have 1 stuck in their plist otherwise.
463
464         * DumpRenderTree/mac/DumpRenderTree.mm:
465         (resetDefaultsToConsistentValues):
466
467 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
468
469         Reviewed by Tony Chang.
470
471         Rebaseline server: compute diffs client-side
472         https://bugs.webkit.org/show_bug.cgi?id=49640
473         
474         The image diff output from the DRT is pretty bad on some ports (at
475         least the Mac one), so it's better to compute diffs on the client by
476         using <canvas>.
477
478         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
479         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
480         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
481
482 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
483
484         Reviewed by Tony Chang.
485
486         Rebaseline server: display test results
487         https://bugs.webkit.org/show_bug.cgi?id=49626
488         
489         Adds basic result display to the rebaseline server. On the Python side
490         this involves:
491         - Parsing the unexpected_results.json into a dictionary.
492         - Serving it as JSON under /results.json.
493         (the JSON -> dict -> JSON transform isn't strictly necessary right now,
494         but I'll need to have access to the parsed results on the Python side
495         for follow-up changes).
496         
497         On the web UI side this adds:
498         - Markup for display image and text results (expected, actual, diff),
499           and JS for populating it.
500         - Markup for breaking down test results by failure type and directory,
501           and JS for populating it.
502
503         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
504         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
505         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
506         * Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js: Added.
507         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
508
509 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
510
511         Reviewed by Ojan Vafai.
512
513         See Chromium issue http://codereview.chromium.org/5133001/ - we
514         are modifying the build bots to pass in the builder name with
515         the "GPU" string appended instead of appending it in the code.
516
517         https://bugs.webkit.org/show_bug.cgi?id=49636
518
519         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
520
521 2010-11-16  Steve Falkenburg  <sfalken@apple.com>
522
523         Reviewed by Adam Roben.
524
525         Disable LTCG for Windows Release builds. Add new Release_LTCG configuration.
526         https://bugs.webkit.org/show_bug.cgi?id=49632
527
528         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
529         * DumpRenderTree/win/DumpRenderTree.vcproj:
530         * DumpRenderTree/win/ImageDiff.vcproj:
531         * MiniBrowser/MiniBrowser.vcproj:
532         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
533         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
534         * WebKitTestRunner/win/InjectedBundle.vcproj:
535         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
536         * WinLauncher/WinLauncher.vcproj:
537         * record-memory-win/record-memory-win.vcproj:
538
539 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
540
541         Reviewed by Ojan Vafai.
542
543         new-run-webkit-tests: rename TestInfo to TestInput, move image hash to work thread
544
545         Rename the TestInfo class to TestInput to be clearer about its
546         function, and move the checksum-reading code into dump_render_tree_thread
547         to avoid cross-thread access.
548
549         https://bugs.webkit.org/show_bug.cgi?id=49573
550
551         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
552         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
553         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
554
555 2010-11-16  Dave Hyatt  <hyatt@apple.com>
556
557         Fix bustage. Remove the minimum font size pref setting in DumpRenderTree.
558
559         * DumpRenderTree/mac/DumpRenderTree.mm:
560         (resetDefaultsToConsistentValues):
561
562 2010-11-16  Stephanie Lewis  <slewis@apple.com>
563
564         Reviewed by Geoff Garen.
565
566         <rdar://problem/8624267> Leak creating offscreen webview running fast/dom tests
567
568         Use shared DeviceOrientationProviderMock. The old allocation method created a new WebDeviceOrientationProviderMock 
569         for every WebView, and the WebKit API doesn't support that ownership model.
570
571         I also fixed a leak in setMockDeviceOrientation by adding a missing release.
572
573         * DumpRenderTree/mac/DumpRenderTree.mm:
574         (createWebViewAndOffscreenWindow):
575         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
576         (LayoutTestController::setMockDeviceOrientation):
577
578 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
579
580         Reviewed by Ojan Vafai.
581
582         new-run-webkit-tests: log the process id in --verbose mode
583
584         As part of the switch to multi-process mode, we should log the
585         process id when running w/ --verbose.
586
587         https://bugs.webkit.org/show_bug.cgi?id=49571
588
589         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
590
591 2010-11-16  Adam Roben  <aroben@apple.com>
592
593         Only trigger Windows builds when files that we actually use on Windows
594         are changed
595
596         A new Scheduler subclass, PlatformSpecificScheduler, has been added.
597         It uses the new webkitpy.common.config.build module to determine
598         whether a particular change should trigger a build on a particular
599         platform. The Windows builders have been switched to use a
600         PlatformSpecificScheduler.
601
602         The logic to determine whether or not a particular change should
603         trigger a build on a given platform has only been implemented/tested
604         for Windows. I tried to make it easy to add more platforms in the
605         future, but I don't have enough familiarity with all platforms to be
606         able to implement it for them.
607
608         Fixes <http://webkit.org/b/49407> Windows builders kick off builds for
609         lots irrelevant changes (e.g., rebaselining Chromium test results)
610
611         Reviewed by Eric Seidel.
612
613         * BuildSlaveSupport/build.webkit.org-config/config.json: Use a
614         PlatformSpecificScheduler for the Windows builders.
615
616         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
617         (PlatformSpecificScheduler.__init__): Added. Stores our platform, sets
618         up a ChangeFilter that filters to our branch and will call through to
619         our filter method, and calls up to our base class.
620         (PlatformSpecificScheduler.filter): Calls through to
621         build.should_build to find out whether we should trigger a build for
622         this change.
623
624         * Scripts/webkitpy/common/net/build.py: Added.
625         (_should_file_trigger_build): Uses a set of directories and regexp
626         patterns to determine whether the given file should trigger a build on
627         the given platform. As mentioned earlier, this has only been
628         implemented for Windows, though I did try to make some guesses about
629         other platforms.
630         (should_build): Returns true if any of the files should trigger a
631         build on the given platform.
632
633         * Scripts/webkitpy/common/net/build_unittest.py: Added.
634         (ShouldBuildTest.test_should_build): Does some basic testing to make
635         sure we're triggering builds for the right files. It only tests
636         Windows for now, though I tried to make some guesses about other
637         platforms.
638
639 2010-11-16  Simon Hausmann  <simon.hausmann@nokia.com>
640
641         Reviewed by Andreas Kling.
642
643         [Qt] Remove synchronous QWebPage::checkPermissions signal
644         https://bugs.webkit.org/show_bug.cgi?id=46810
645
646         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
647         (WebCore::WebPage::WebPage):
648         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
649         (LayoutTestController::grantDesktopNotificationPermission): When granting
650         permission, grant it directly on the QWebPage/Frame, that will remember it.
651         * QtTestBrowser/webpage.cpp:
652         (WebPage::WebPage):
653         * QtTestBrowser/webpage.h:
654
655
656 2010-11-16  Andras Becsi  <abecsi@inf.u-szeged.hu>
657
658         Reviewed by Csaba Osztrogonác.
659
660         [Qt] Add isPrinting and setPrinting methods to DRT's LayoutTestController.
661
662         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
663         (LayoutTestController::reset):
664         * DumpRenderTree/qt/LayoutTestControllerQt.h:
665         (LayoutTestController::isPrinting):
666         (LayoutTestController::setPrinting):
667
668 2010-11-16  Balazs Kelemen  <kbalazs@webkit.org>
669
670         Reviewed by Andreas Kling.
671
672         [Qt][WK2] Add a way to turn on tiled backing store in MiniBrowser
673         https://bugs.webkit.org/show_bug.cgi?id=49587
674
675         * MiniBrowser/qt/BrowserView.cpp:
676         (BrowserView::BrowserView): Added BackingStoreType
677         argument to ctor. Propagating it to the QGraphicsWKView.
678         * MiniBrowser/qt/BrowserView.h:
679         * MiniBrowser/qt/BrowserWindow.cpp:
680         (BrowserWindow::BrowserWindow):
681         (BrowserWindow::newWindow): Create the new BrowserWindow
682         with the same BackingStoreType that this instance has.
683         * MiniBrowser/qt/BrowserWindow.h: Added BackingStoreType member.
684         * MiniBrowser/qt/main.cpp:
685         (main): Use tiled backing store if got -tiled command line argument.
686         Avoid redundant copy of command line args.
687
688 2010-11-16  John Knottenbelt  <jknotten@chromium.org>
689
690         Reviewed by Jeremy Orlow.
691
692         Move DeviceOrientationClientMock from LayoutTestController to WebViewHost.
693         https://bugs.webkit.org/show_bug.cgi?id=48506
694
695         This ensures that there is one mock per page WebView / WebViewHost. This is a
696         design constraint of the DeviceOrientationClientMock because the m_controller
697         field can only store a pointer to a single instance of the corresponding
698         DeviceOrientationController.
699
700         Test: fast/dom/DeviceOrientation/no-page-cache.html
701
702         * DumpRenderTree/chromium/LayoutTestController.cpp:
703         (LayoutTestController::setMockDeviceOrientation):
704         (LayoutTestController::deviceOrientationClientMock):
705         * DumpRenderTree/chromium/LayoutTestController.h:
706         * DumpRenderTree/chromium/WebViewHost.cpp:
707         (WebViewHost::deviceOrientationClientMock):
708         (WebViewHost::deviceOrientationClient):
709         (WebViewHost::reset):
710         * DumpRenderTree/chromium/WebViewHost.h:
711
712 2010-11-15  Andras Becsi  <abecsi@webkit.org>
713
714         Reviewed by Andreas Kling.
715
716         [Qt][WK2] Avoid polling in WebKitTestRunner.
717         https://bugs.webkit.org/show_bug.cgi?id=49542
718
719         Make the WK2 testing session about 10-15% faster and decrease random flakiness
720         resulting from timing skew by removing the millisecond polling from TestControllerQt.
721
722         * WebKitTestRunner/TestController.cpp:
723         (WTR::TestController::didFinishLoadForFrame):
724         * WebKitTestRunner/TestController.h:
725         * WebKitTestRunner/TestInvocation.cpp:
726         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
727         * WebKitTestRunner/mac/TestControllerMac.mm:
728         (WTR::TestController::notifyDone): Add method stub.
729         * WebKitTestRunner/qt/TestControllerQt.cpp:
730         (WTR::TestControllerRunLoop::instance):
731         (WTR::TestControllerRunLoop::start):
732         (WTR::TestControllerRunLoop::stop):
733         (WTR::TestControllerRunLoop::TestControllerRunLoop):
734         (WTR::TestControllerRunLoop::timerEvent):
735         (WTR::TestController::notifyDone):
736         (WTR::TestController::platformRunUntil):
737         * WebKitTestRunner/win/TestControllerWin.cpp:
738         (WTR::TestController::notifyDone): Add method stub.
739
740 2010-11-15  Darin Adler  <darin@apple.com>
741
742         * Scripts/webkitpy/common/net/bugzilla: Added property svn:ignore.
743
744 2010-11-15  Benjamin Poulain  <benjamin.poulain@nokia.com>
745
746         Reviewed by Andreas Kling.
747
748         [Qt] [WK2] Add Ctrl+L shortcut to Qt Minibrowser
749         https://bugs.webkit.org/show_bug.cgi?id=49544
750
751         Add the shortcut to BrowserWindow.
752
753         * MiniBrowser/qt/BrowserWindow.cpp:
754         (BrowserWindow::BrowserWindow):
755         (BrowserWindow::openLocation):
756         * MiniBrowser/qt/BrowserWindow.h:
757
758 2010-11-14  David Levin  <levin@chromium.org>
759
760         Reviewed by Daniel Bates.
761
762         check-webkit-style should detect PassRefPtr usage in functions.
763         https://bugs.webkit.org/show_bug.cgi?id=49513
764
765         * Scripts/webkitpy/style/checkers/cpp.py:
766         (check_for_function_lengths): Revert a comment change that I
767          accidentally made in r71986.
768         (check_pass_ptr_usage): Added the code to do the check.
769         (process_line): Added the call to check_pass_ptr_usage.
770         (CppChecker): Added the new error category.
771         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
772         (CppStyleTestBase::perform_pass_ptr_check): Runs the new check for
773          testing purposes.
774         (PassPtrTest::*): The class/functions to unit test the new
775          functionality.
776
777 2010-11-14  David Levin  <levin@chromium.org>
778
779         Reviewed by Shinichiro Hamaji.
780
781         check-webkit-style function detection and the line count style checks should be separate.
782         https://bugs.webkit.org/show_bug.cgi?id=49512
783
784         * Scripts/webkitpy/style/checkers/cpp.py: Do the separation.
785         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Adjust the test to
786           call the detection function and fix line counts in two places now that
787           the code really only counts the lines in the body of the function.
788
789 2010-11-14  Andreas Kling  <kling@webkit.org>
790
791         Reviewed by Antonio Gomes.
792
793         http/tests/plugins tests print "Unhandled variable" to stderr under Qt
794         https://bugs.webkit.org/show_bug.cgi?id=33438
795
796         Stifle "Unhandled variable" warning to match what the other
797         TestNetscapePlugin does (NPP_GetValue in TestNetscapePlugIn/main.cpp
798         simply returns NPERR_GENERIC_ERROR for unhandled variables.)
799
800         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
801         (webkit_test_plugin_get_value):
802
803 2010-11-13  David Levin  <levin@chromium.org>
804
805         Reviewed by Eric Seidel.
806
807         check-webkit-style function detection crashes on functions in templates.
808         https://bugs.webkit.org/show_bug.cgi?id=49504
809
810         For example "template <bool shouldClose(const Element*)>".
811
812         * Scripts/webkitpy/style/checkers/cpp.py: If not function is detected, bail out.
813         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Modified the complex
814           function detection test to expose this issue.
815
816 2010-11-12  David Levin  <levin@chromium.org>
817
818         Reviewed by Shinichiro Hamaji.
819
820         check-webkit-style function detection doesn't detect indented functions declaractions.
821         https://bugs.webkit.org/show_bug.cgi?id=49446
822
823         Indented function declarations occur inside class definitions, so
824         they are a pretty common (and worth detecting).
825
826         * Scripts/webkitpy/style/checkers/cpp.py:
827           Changed regex to allow indentation.
828           Changed the function start detection to only happen when not in a
829           function.
830           Changed function end detection to work based on matching braces
831           instead of finding a close brace at the beginning of the line.
832           Fixed close_expression to do what it says when it doesn't find
833           the close.
834         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Indented function test.
835
836 2010-11-12  Daniel Bates  <dbates@rim.com>
837
838         Rubber-stamped by Nikolas Zimmermann.
839
840         Make do-file-rename rename files in the directory WebKit2.
841
842         This also makes the list of searched directories in do-file-rename
843         consistent with the list of searched directories in the script do-webcore-rename.
844
845         * Scripts/do-file-rename:
846
847 2010-11-12  Dirk Pranke  <dpranke@chromium.org>
848
849         Reviewed by James Robinson.
850
851         Attempt yet again to land the fix for bug 49360 (respecting
852         set-webkit-configuration). We need to handle the cases where
853         trying to run webkit-build-directory to find out where the
854         default configuration might be fails (that shows up on some
855         Chromium bots that apparently don't have perl installed).
856
857         https://bugs.webkit.org/show_bug.cgi?id=49360
858
859         * Scripts/webkitpy/layout_tests/port/config.py:
860         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Added.
861         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
862
863 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
864
865         Reviewed by Adam Barth.
866
867         webkit-patch land logs commit message twice
868         https://bugs.webkit.org/show_bug.cgi?id=49482
869         
870         Don't log the comment that's used by bugzilla.close_bug_as_fixed, since
871         that's just the commit message, which we now output following the
872         actual commit.
873
874         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
875
876 2010-11-12  Tony Chang  <tony@chromium.org>
877
878         Reviewed by Adam Barth.
879
880         run platform/chromium/plugins/script-object-invoke.html on all platforms
881         https://bugs.webkit.org/show_bug.cgi?id=49280
882
883         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
884         (testScriptObjectInvoke):
885         (pluginInvoke):
886
887 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
888
889         Reviewed by Adam Barth.
890
891         webkit-patch land-cowboy (and other commands) should display the committed SVN revision after landing a patch
892         https://bugs.webkit.org/show_bug.cgi?id=49471
893
894         Right now there's no indication of a successful commit.
895         
896         * Scripts/webkitpy/tool/commands/download_unittest.py:
897         * Scripts/webkitpy/tool/steps/commit.py:
898
899 2010-11-12  Kinuko Yasuda  <kinuko@chromium.org>
900
901         Reviewed by Ojan Vafai.
902
903         Include detailed test modifiers in results.json and enable incremental uploading for non-layout tests
904         https://bugs.webkit.org/show_bug.cgi?id=49354
905
906         Also moved/integrated the upload method from run_webkit_tests.py to json_results_generator.py.
907
908         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
909         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
910         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
911         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
912
913 2010-11-12  Tony Chang  <tony@chromium.org>
914
915         Reviewed by Kent Tamura.
916
917         [chromium] Load a blank page when closing DRT WebViewHosts
918         https://bugs.webkit.org/show_bug.cgi?id=49417
919
920         In test_shell, when we close a window, we first load about:blank
921         and run GC to fire the destruction logic of the page.  In test_shell,
922         this happens in ~TestShell.  In DRT, we manage virtual windows as
923         WebViewHosts, so we need to replicate this logic in ~WebViewHost.
924
925         Managing focus is a bit trickier, since normally you would get window
926         events to reset the focus to the right window.  We do it manually
927         instead.
928
929         This fixes 2 plugin tests on DRT.
930
931         * DumpRenderTree/chromium/TestShell.cpp:
932         (TestShell::~TestShell): Move window close logic to ~WebViewHost.
933         (TestShell::closeWindow): Move widget close() to WVH destructor and
934             manage focus manually.
935         (TestShell::closeRemainingWindows): Comment typo.
936         * DumpRenderTree/chromium/WebViewHost.cpp:
937         (WebViewHost::closeWidget):
938         (invokeCloseWidget):
939         (WebViewHost::closeWidgetSoon): Close the window asynchronously (matches
940             test_shell).
941         (WebViewHost::~WebViewHost):
942         * DumpRenderTree/chromium/WebViewHost.h:
943
944 2010-11-12  David Levin  <levin@chromium.org>
945
946         Reviewed by Adam Barth.
947
948         cpp.py needs some tidying.
949         https://bugs.webkit.org/show_bug.cgi?id=49443
950
951         * Scripts/webkitpy/style/checkers/cpp.py:
952           Fix parameter ordering for iteratively_replace_matches_with_char
953           to mirror other similar functions (and improved comments, etc.)
954           Also, use a python-ism (for/else) to get rid of a variable that
955           was there to determine if something had been found in the loop.
956
957 2010-11-12  Robert Kroeger  <rjkroege@chromium.org>
958
959         Reviewed by James Robinson.
960
961         [Chromium/DRT] Make EventSendingController send complete touch events.
962         https://bugs.webkit.org/show_bug.cgi?id=49285
963
964         * DumpRenderTree/chromium/EventSender.cpp:
965         (EventSender::addTouchPoint):
966         (EventSender::updateTouchPoint):
967         (EventSender::sendCurrentTouchEvent):
968
969 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
970
971         Unreviewed rollout of r71858.
972
973         Rollout out r71858 since it breaks new-run-webkit-httpd as used by the
974         NaCl tests.
975
976         * Scripts/webkitpy/layout_tests/port/config.py:
977         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Removed.
978         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
979
980 2010-11-12  Benjamin Poulain  <benjamin.poulain@nokia.com>
981
982         Reviewed by Kenneth Rohde Christiansen.
983
984         [Qt] TestController::runUntil doesn't honor the timeout parameter
985         https://bugs.webkit.org/show_bug.cgi?id=48941
986
987         Implement the missing timeout of WebKitTestRunner for Qt.
988
989         A QElapsedTimer has been added to RunUntilConditionLoop in order
990         to measure how long the loop has been running. When the timer
991         is bigger or equal than the timeout, the loop ends even if the
992         condition is not met.
993
994         * WebKitTestRunner/qt/TestControllerQt.cpp:
995         (WTR::RunUntilConditionLoop::start):
996         (WTR::RunUntilConditionLoop::run):
997         (WTR::RunUntilConditionLoop::timerEvent):
998         (WTR::TestController::platformRunUntil):
999
1000 2010-11-11  David Levin  <levin@chromium.org>
1001
1002         Reviewed by Shinichiro Hamaji.
1003
1004         check-webkit-style function detection doesn't handle templates with spaces.
1005         https://bugs.webkit.org/show_bug.cgi?id=49427
1006
1007         * Scripts/webkitpy/style/checkers/cpp.py: Stripped out templates when
1008           finding the function name.
1009         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test have a
1010           template with a space.
1011
1012 2010-11-11  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
1013
1014         Reviewed by Shinichiro Hamaji.
1015
1016         Enable check-webkit-style on Qt files
1017         https://bugs.webkit.org/show_bug.cgi?id=48322
1018
1019         Adding exception for _q_... private slots and
1020         ..._data() methods in tests
1021
1022         * Scripts/webkitpy/style/checker.py: enable Qt dirs, add exceptions
1023         * Scripts/webkitpy/style/checker_unittest.py: improve unittests
1024
1025 2010-11-11  David Levin  <levin@chromium.org>
1026
1027         Reviewed by Shinichiro Hamaji.
1028
1029         check-webkit-style function detection doesn't work for templates and destructors.
1030         https://bugs.webkit.org/show_bug.cgi?id=49425
1031
1032         * Scripts/webkitpy/style/checkers/cpp.py: Changed regex for templates and destructors.
1033         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test hit this issue.
1034
1035 2010-11-11  Joone Hur  <joone@kldp.org>
1036
1037         Reviewed by Antonio Gomes.
1038
1039         [GTK][DRT] Implement LayoutTestController::nodesFromRect
1040         https://bugs.webkit.org/show_bug.cgi?id=46598
1041
1042         Support nodesFromRect in DRT
1043
1044         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1045         (LayoutTestController::nodesFromRect):
1046
1047 2010-11-11  MORITA Hajime  <morrita@google.com>
1048
1049         Reviewed by Kent Tamura.
1050
1051         editing/selection/context-menu-on-text.html fails on chromium
1052         https://bugs.webkit.org/show_bug.cgi?id=45898
1053
1054         Enhance Chromium's EventSender::contextClick() to return
1055         array of menu strings. The implementation incomplete because
1056         Although actual context menu is implemented by the browser,
1057         it reflects editability of underlying context thus makes
1058         context-menu-on-text.html passed.
1059         
1060         * DumpRenderTree/chromium/EventSender.cpp:
1061         (makeMenuItemStringsFor): Added.
1062         (EventSender::contextClick): Now returns array of menu strings.
1063         * DumpRenderTree/chromium/WebViewHost.cpp:
1064         (WebViewHost::showContextMenu): Now storing given WebContextMenuData object.
1065         (WebViewHost::clearContextMenuData): Added.
1066         (WebViewHost::lastContextMenuData): Added.
1067         * DumpRenderTree/chromium/WebViewHost.h:
1068
1069 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1070
1071         Unreviewed, build fix.
1072
1073         fix build breakage caused by fix for bug 49380 (r71858).
1074
1075         config_standalone needs to ensure that the packages it needs are
1076         in sys.path.
1077
1078         https://bugs.webkit.org/show_bug.cgi?id=49419
1079
1080         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
1081
1082 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1083
1084         Reviewed by Adam Roben.
1085
1086         Fix NRWT to respect set-webkit-configuration again :(
1087
1088         This change fixes a typo in config.py that was causing the wrong
1089         value to be read initially and us never actually looking into
1090         the filesystem to get the default configuration.
1091         
1092         * Scripts/webkitpy/layout_tests/port/config.py:
1093         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
1094         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1095
1096 2010-11-11  Eric Seidel  <eric@webkit.org>
1097
1098         Reviewed by Adam Barth.
1099
1100         Split out Bug, Attachment and CommitterValidator from bugzilla.py
1101         https://bugs.webkit.org/show_bug.cgi?id=49403
1102
1103         CommitterValidator really had nothing to do with bugzilla.py.
1104         I've put it next to committers.py for now since its rather tied to that file.
1105
1106         * Scripts/webkitpy/common/config/committervalidator.py: Added.
1107         * Scripts/webkitpy/common/config/committervalidator_unittest.py: Added.
1108         * Scripts/webkitpy/common/net/bugzilla/__init__.py:
1109         * Scripts/webkitpy/common/net/bugzilla/attachment.py: Added.
1110         * Scripts/webkitpy/common/net/bugzilla/bug.py: Added.
1111         * Scripts/webkitpy/common/net/bugzilla/bug_unittest.py: Added.
1112         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
1113         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py:
1114         * Scripts/webkitpy/tool/bot/feeders.py:
1115         * Scripts/webkitpy/tool/commands/queues.py:
1116
1117 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1118
1119         Reviewed by Eric Seidel.
1120
1121         fix missing 'pid' argument for http lockfile checking on win32
1122         https://bugs.webkit.org/show_bug.cgi?id=49363
1123
1124         * Scripts/webkitpy/common/system/executive.py:
1125
1126 2010-11-11  Eric Seidel  <eric@webkit.org>
1127
1128         Rubber-stamped by Adam Barth.
1129
1130         Move bugzilla.py into its own module in preparation for splitting one-file-per-class
1131         https://bugs.webkit.org/show_bug.cgi?id=49402
1132
1133         * Scripts/webkitpy/common/net/bugzilla/__init__.py: Added.
1134         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla.py.
1135         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla_unittest.py.
1136
1137 2010-11-10  Chris Guillory  <chris.guillory@google.com>
1138
1139         Reviewed by Chris Fleizach.
1140
1141         Use IAccessibleComparable to compare accessibility objects.
1142         https://bugs.webkit.org/show_bug.cgi?id=49118
1143
1144         * DumpRenderTree/AccessibilityUIElement.cpp:
1145         (AccessibilityUIElement::isEqual):
1146         * DumpRenderTree/AccessibilityUIElement.h:
1147         * DumpRenderTree/win/AccessibilityUIElementWin.cpp:
1148         (comparableObject):
1149         (AccessibilityUIElement::isEqual):
1150
1151 2010-11-10  Csaba Osztrogonác  <ossy@webkit.org>
1152
1153         Reviewed by David Hyatt.
1154
1155         HTML5 Ruby support should be mandatory feature
1156         https://bugs.webkit.org/show_bug.cgi?id=49272
1157
1158         Remove Ruby as optional feature.
1159
1160         * Scripts/build-webkit:
1161
1162 2010-11-10  Adam Roben  <aroben@apple.com>
1163
1164         Print VC++ Express build logs in the same order that the projects are
1165         built
1166
1167         Fixes <http://webkit.org/b/49326> print-vse-failure-logs prints logs
1168         in an unhelpful order
1169
1170         Reviewed by Eric Seidel.
1171
1172         * Scripts/print-vse-failure-logs:
1173         (PrintVisualStudioExpressLogs._build_order): Added. Uses
1174         print-msvc-project-dependencies to figure out the order in which
1175         projects are built.
1176         (PrintVisualStudioExpressLogs._sort_buildlogs): Added. Sorts the logs
1177         based on their build order and project name.
1178         (PrintVisualStudioExpressLogs._obj_directory): Moved code to find the
1179         scripts directory from here...
1180         (PrintVisualStudioExpressLogs._scripts_directory): ...to here.
1181         (PrintVisualStudioExpressLogs.main): Sort the logs before printing
1182         them.
1183
1184 2010-11-10  Adam Roben  <aroben@apple.com>
1185
1186         Only print the interesting text from VC++ Express build logs
1187
1188         Fixes <http://webkit.org/b/49325> It's hard to find the interesting
1189         output from print-vse-failure-logs
1190
1191         Reviewed by Eric Seidel.
1192
1193         * Scripts/print-vse-failure-logs: Removed unnecessary os.path import,
1194         added newly-required imports. Removed a comment that was essentially a
1195         FIXME, and is now fixed.
1196         (PrintVisualStudioExpressLogs._relevant_text): Added. Uses
1197         BeautifulSoup to extract the relevant text from the build log.
1198         (PrintVisualStudioExpressLogs.main): Only print the relevant text.
1199
1200 2010-11-09  Mihai Parparita  <mihaip@chromium.org>
1201
1202         Reviewed by Tony Chang.
1203
1204         [Chromium] http/tests/security/XFrameOptions fail with Chromium DRT
1205         https://bugs.webkit.org/show_bug.cgi?id=49286
1206         
1207         The tests in http/tests/security/XFrameOptions started to fail after
1208         r71297, since it changed the WebViewHost::assignIdentifierToRequest
1209         logic to always assign identifiers to requests, even if we're not going
1210         to be dumping resource load callbacks at that point. These tests
1211         only call LayoutTestController.dumpResourceLoadCallbacks after the
1212         request is started, thus their expectations have "<unknown>" as the
1213         identifier. Change WebViewHost to the previous behavior, which also
1214         match the other ports'.
1215
1216         * DumpRenderTree/chromium/WebViewHost.cpp:
1217         (WebViewHost::assignIdentifierToRequest):
1218         (WebViewHost::removeIdentifierForRequest):
1219
1220 2010-11-08  Tony Chang  <tony@chromium.org>
1221
1222         Reviewed by Adam Barth.
1223
1224         run platform/chromium/plugins/return-npobject.html on all platforms
1225         https://bugs.webkit.org/show_bug.cgi?id=49103
1226
1227         This tests that we can get an NPObject returned through a method on
1228         an NPAPI Object.
1229
1230         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1231         (testCallbackReturn):
1232         (pluginInvoke):
1233
1234 2010-11-09  Sheriff Bot  <webkit.review.bot@gmail.com>
1235
1236         Unreviewed, rolling out r71672.
1237         http://trac.webkit.org/changeset/71672
1238         https://bugs.webkit.org/show_bug.cgi?id=49276
1239
1240         It doesn't work on Chromium Windows bot (Requested by Ossy on
1241         #webkit).
1242
1243         * Scripts/webkitpy/common/system/file_lock.py: Removed.
1244         * Scripts/webkitpy/common/system/file_lock_unittest.py: Removed.
1245         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1246
1247 2010-11-09  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1248
1249         Reviewed by Tony Chang.
1250
1251         [NRWT] Make http locking similar to perl implementation
1252         https://bugs.webkit.org/show_bug.cgi?id=49187
1253
1254         * Scripts/webkitpy/common/system/file_lock.py: Added.
1255         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
1256         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1257
1258 2010-11-09  James Kozianski  <koz@chromium.org>
1259
1260         Reviewed by Eric Seidel.
1261
1262         Clean up imports for webkit-patch.
1263         https://bugs.webkit.org/show_bug.cgi?id=49083
1264
1265         Move some imports from commands into main.py where they are used.
1266         Move command imports into the __init__ of the commands module.
1267
1268         * Scripts/webkitpy/tool/commands/__init__.py:
1269         * Scripts/webkitpy/tool/commands/download.py:
1270         * Scripts/webkitpy/tool/commands/queues.py:
1271         * Scripts/webkitpy/tool/main.py:
1272
1273 2010-11-09  Yi Shen  <yi.4.shen@nokia.com>
1274
1275         Reviewed by Andreas Kling.
1276
1277         [Qt][QtTestBrowser] Don't erase incorrect url in the Url Bar
1278         https://bugs.webkit.org/show_bug.cgi?id=49047
1279
1280         * QtTestBrowser/mainwindow.cpp:
1281         (MainWindow::setAddressUrl):
1282
1283 2010-11-09  Andras Becsi  <abecsi@webkit.org>
1284
1285         Reviewed by Csaba Osztrogonác.
1286
1287         Improve old-run-webkit-tests --verbose to show DumpTool opening/closing
1288         https://bugs.webkit.org/show_bug.cgi?id=49252
1289
1290         * Scripts/old-run-webkit-tests:
1291
1292 2010-11-09  Csaba Osztrogonác  <ossy@webkit.org>
1293
1294         Reviewed by Tor Arne Vestbø.
1295
1296         Output of old-run-webkit-tests is buggy in verbose mode
1297         https://bugs.webkit.org/show_bug.cgi?id=49249
1298
1299         * Scripts/old-run-webkit-tests: Redundant outputs removed.
1300
1301 2010-11-08  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
1302
1303         Reviewed by Martin Robinson.
1304
1305         [GTK] Link with target name set does not work
1306         https://bugs.webkit.org/show_bug.cgi?id=48865
1307
1308         When a new page is created with a name (target=myFrame), the new
1309         mainFrame could not be found because they where not stored in the
1310         same PageGroup. As PageGroup are not exposed externally so the
1311         simpliest solution is to use a global page group name. This also fixes
1312         issue with visited link coloration across pages. After this change the
1313         private function webkit_web_view_set_group_name() was no longer used
1314         so it was removed completly.
1315
1316         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1317         (createWebView):
1318
1319 2010-11-08  John Knottenbelt  <jknotten@chromium.org>
1320
1321         Reviewed by Kent Tamura.
1322
1323         WebViewHost::reset() uses placement new.
1324         https://bugs.webkit.org/show_bug.cgi?id=49069
1325
1326         WebViewHost is using placement destruction / new to simulate a fresh
1327         WebViewHost object at the same address. This is because the WebView remains
1328         open across tests and maintains a pointer to the WebViewHost.
1329         This change resets member variables explictly instead of the placement new dance.
1330
1331         * DumpRenderTree/chromium/WebViewHost.cpp:
1332         (WebViewHost::WebViewHost):
1333         (WebViewHost::reset):
1334
1335 2010-11-08  Dirk Pranke <dpranke@chromium.org>
1336
1337         Reviewed by Eric Seidel.
1338
1339         Fix the webkit_base_dir logic in webkitpy/layout_tests/port/config.py. 
1340         It turns out that NRWT can't use the code in scm.find_checkout_root()
1341         because the Chromium bots don't do full checkouts of the WebKit
1342         tree; they only check out subdirectories like WebKitTools/Scripts.
1343         Until we can figure out a better approach for this, I've
1344         restored the base_dir-detecting code from NRWT, which works in
1345         any directory tree, scm or no.
1346
1347         This also restores the files modified in r71475 and r71474.
1348
1349         https://bugs.webkit.org/show_bug.cgi?id=49151
1350
1351         * WebKitTools/Scripts/webkitpy/layout_tests/port/config.py
1352         * WebKitTools/Scripts/webkitpy/layout_tests/port/config_unittest.py
1353
1354 2010-11-08  Eric Seidel  <eric@webkit.org>
1355
1356         Reviewed by Adam Barth.
1357
1358         webkit-patch failure-reason explodes if a build is missing
1359         https://bugs.webkit.org/show_bug.cgi?id=49195
1360
1361         This is likely a recent regression.  Adding a unit test for this case.
1362
1363         * Scripts/webkitpy/common/net/buildbot.py:
1364         * Scripts/webkitpy/common/net/buildbot_unittest.py:
1365
1366 2010-11-08  Dan Bernstein  <mitz@apple.com>
1367
1368         Reviewed by Darin Adler.
1369
1370         Ignore the system scrollbar setting.
1371
1372         * DumpRenderTree/mac/DumpRenderTree.mm:
1373         (resetDefaultsToConsistentValues): Set the scrollbar default.
1374
1375 2010-11-05  Adam Roben  <aroben@apple.com>
1376
1377         Make webkitpy.common.system.executive_unittest pass when running under
1378         Win32 Python
1379
1380         Fixes <http://webkit.org/b/49033>.
1381
1382         Reviewed by Dave Levin and Eric Seidel.
1383
1384         * Scripts/webkitpy/common/system/executive.py:
1385         (Executive._run_command_with_teed_output): Pass the arguments through
1386         encode_argument_if_needed rather than using Cygwin-specific code here.
1387         (Executive.run_and_throw_if_fail): Use child_process_encoding to decode
1388         the output.
1389         (Executive.run_command): Use encode_argument_if_needed to encode the
1390         arguments and child_process_encoding to decode the output.
1391         (Executive._child_process_encoding): Returns the encoding that should be
1392         used when communicating with child processes. On Windows we use mbcs,
1393         which maps to the current code page. On all other platforms we use
1394         UTF-8.
1395         (Executive._should_encode_child_process_arguments): Returns True if
1396         arguments to child processes need to be encoded. This is currently
1397         only needed on Cygwin and Win32 Python 2.x.
1398         (Executive._encode_argument_if_needed): Encode the argument using
1399         child_process_encoding if we need to encode arguments to child
1400         processes on this platform.
1401
1402         * Scripts/webkitpy/common/system/executive_unittest.py:
1403         (never_ending_command): Added. Returns arguments to run a command that
1404         will not quit until we kill it. On Windows we use wmic, on other
1405         platforms we use yes.
1406         (ExecutiveTest.test_run_command_with_unicode): Changed to expect the
1407         mbcs encoding to be used and for output from the child processes to
1408         have been roundtripped through encode/decode on Win32 Python. When
1409         UTF-8 is the encoding the roundtripping is undetectable, but with mbcs
1410         it's possible that some characters will not be able to be converted
1411         and will be replaced by question marks; the round-tripping allows us
1412         to expect this result.
1413
1414         (ExecutiveTest.test_kill_process):
1415         (ExecutiveTest.test_kill_all):
1416         Use never_ending_command instead of invoking "yes" directly. Expect an
1417         exit code of 1 when using Win32 Python, as that's what seems to happen.
1418
1419 2010-11-08  Adam Roben  <aroben@apple.com>
1420
1421         Roll out r71532
1422
1423         It broke the build for Cygwin 1.7 installs. Cygwin 1.7's default
1424         .bashrc unsets %TEMP%, which broke copy-tools.cmd.
1425
1426         * Scripts/webkitdirs.pm:
1427
1428 2010-11-08  Tony Chang  <tony@chromium.org>
1429
1430         Reviewed by Adam Barth.
1431
1432         run platform/chromium/plugins/nested-plugin-objects.html on all platforms
1433         https://bugs.webkit.org/show_bug.cgi?id=49094
1434
1435         This tests that objects created by plugins are proplery cleaned up.
1436
1437         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1438         (testPassTestObject):
1439         (pluginInvoke):
1440         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
1441         (testAllocate):
1442         (testDeallocate):
1443         (testGetProperty):
1444         (testConstruct):
1445
1446 2010-11-08  Adam Roben  <aroben@apple.com>
1447
1448         Mark Windows builds triggered from Perl as being non-interactive
1449
1450         This affects whether some of our scripts will show alerts vs. printing
1451         to the build log.
1452
1453         Fixes <http://webkit.org/b/49181> Windows build fail mysteriously when
1454         .vsprops files are updated
1455
1456         Reviewed by Steve Falkenburg.
1457
1458         * Scripts/webkitdirs.pm:
1459         (buildVisualStudioProject): Set WEBKIT_NONINTERACTIVE_BUILD to 1.
1460
1461 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1462
1463         Reviewed by Ojan Vafai.
1464
1465         Make http locking default in NRWT.
1466         https://bugs.webkit.org/show_bug.cgi?id=48053
1467
1468         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1469         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1470
1471 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1472
1473         Reviewed by Csaba Osztrogonác.
1474
1475         [NRWT] If the http lock fails we shouldn't do any locking
1476         https://bugs.webkit.org/show_bug.cgi?id=49164
1477
1478         If something goes wrong with the locking, the test should keep going.
1479
1480         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1481
1482 2010-11-08  Adam Roben  <aroben@apple.com>
1483
1484         Switch back to using kCGImageAlphaPremultipliedFirst when generating
1485         pixel dumps on Windows
1486
1487         I changed this behavior in r71418 thinking that it was required for
1488         getting plugins to show up in pixel dumps. But it doesn't seem to be
1489         necessary, and was making it impossible to compare new Windows pixel
1490         dumps with existing Windows or Mac pixel dumps (because ImageDiff won't
1491         compare an image with alpha to an image without alpha).
1492
1493         Fixes <http://webkit.org/b/49172> REGRESION (r71418): Can't compare
1494         new Windows pixel results to existing Windows or Mac results
1495
1496         Reviewed by Antti Koivisto.
1497
1498         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
1499         (createBitmapContextFromWebView): Replaced kCGImageAlphaNoneSkipFirst
1500         with kCGImageAlphaPremultipliedFirst.
1501
1502 2010-11-08  Csaba Osztrogonac  <ossy@webkit.org>
1503
1504         Unreviewed, rolling out r71466.
1505         http://trac.webkit.org/changeset/71466
1506         https://bugs.webkit.org/show_bug.cgi?id=48865
1507
1508         It broke layout tests on GTK bots.
1509
1510         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1511         (createWebView):
1512
1513 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1514
1515         Reviewed by Csaba Osztrogonác.
1516
1517         Enable running of Qt API tests on BuildBot
1518         https://bugs.webkit.org/show_bug.cgi?id=49004
1519
1520         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1521
1522 2010-11-08  Eric Seidel  <eric@webkit.org>
1523
1524         Reviewed by Adam Barth.
1525
1526         Add clean-review-queue command to remove closed bugs from the webkit.org/pending-review
1527         https://bugs.webkit.org/show_bug.cgi?id=49160
1528
1529         Bugzilla doesn't automatically remove r? when a bug gets closed.
1530         This script takes care of that for webkit.org.
1531
1532         * Scripts/webkitpy/common/net/bugzilla.py:
1533         * Scripts/webkitpy/tool/commands/upload.py:
1534
1535 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
1536
1537         Unreviewed, rolling out r71474.
1538         http://trac.webkit.org/changeset/71474
1539         https://bugs.webkit.org/show_bug.cgi?id=48280
1540
1541         breaks chromium webkit tests
1542         https://bugs.webkit.org/show_bug.cgi?id=49151
1543
1544         * Scripts/webkitpy/layout_tests/port/base.py:
1545         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
1546         * Scripts/webkitpy/layout_tests/port/config.py:
1547         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1548         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
1549         * Scripts/webkitpy/layout_tests/port/webkit.py:
1550         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1551         * Scripts/webkitpy/tool/mocktool.py:
1552
1553 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
1554
1555         Unreviewed, rolling out r71475.
1556         http://trac.webkit.org/changeset/71475
1557  
1558         breaks chromium webkit tests
1559         https://bugs.webkit.org/show_bug.cgi?id=49151
1560
1561         * Scripts/webkitpy/common/newstringio.py: Removed.
1562         * Scripts/webkitpy/common/newstringio_unittest.py: Removed.
1563         * Scripts/webkitpy/common/system/executive_mock.py: Removed.
1564         * Scripts/webkitpy/common/system/filesystem_mock.py: Removed.
1565         * Scripts/webkitpy/layout_tests/port/config_mock.py: Removed.
1566
1567 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1568
1569         Unreviewed, build fix.
1570
1571         Add files inexplicably not committed in r71474 as part of the
1572         fix for bug 48280.
1573
1574         * Scripts/webkitpy/common/newstringio.py: Added.
1575         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
1576         * Scripts/webkitpy/common/system/executive_mock.py: Added.
1577         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
1578         * Scripts/webkitpy/layout_tests/port/config_mock.py: Added.
1579
1580 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1581
1582         Reviewed by Eric Siedel.
1583
1584         new-run-webkit-tests: update port/base and port/webkit to use the
1585         new FileSystem and Config abstractions, pulling more logic out of
1586         the base Port classes into separate, mockable objects.
1587
1588         Also create a MockFileSystem object, a MockConfig object, move
1589         MockExecutive into common/system to be next to executive, and
1590         update the config object to use a FileSystem.
1591
1592         https://bugs.webkit.org/show_bug.cgi?id=48280
1593
1594         * Scripts/webkitpy/common/newstringio.py: Added.
1595         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
1596         * Scripts/webkitpy/common/system/executive_mock.py: Added.
1597         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
1598         * Scripts/webkitpy/layout_tests/port/base.py:
1599         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
1600         * Scripts/webkitpy/layout_tests/port/config.py:
1601         * Scripts/webkitpy/layout_tests/port/config_mock.py:
1602         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1603         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
1604         * Scripts/webkitpy/layout_tests/port/webkit.py:
1605         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1606         * Scripts/webkitpy/tool/mocktool.py:
1607
1608 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1609
1610         Unreviewed, build breakage.
1611
1612         Apparently I uploaded the wrong version of the file to fix 49122
1613         and neither Eric or I noticed - it was missing a dirname() call.
1614         Fixing ...
1615
1616         https://bugs.webkit.org/show_bug.cgi?id=49122
1617
1618         * Scripts/webkitpy/common/checkout/scm.py:
1619
1620 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1621
1622         Reviewed by Eric Seidel.
1623
1624         webkitpy/tool/* unittests change cwd and don't clean up properly
1625
1626         https://bugs.webkit.org/show_bug.cgi?id=49122
1627
1628         * Scripts/webkitpy/common/checkout/scm.py:
1629
1630 2010-11-05  Chris Marrin  <cmarrin@apple.com>
1631
1632         Reviewed by Simon Fraser.
1633
1634         suspendAnimations/resumeAnimations not present in WebKit2
1635         https://bugs.webkit.org/show_bug.cgi?id=49109
1636
1637         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
1638         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
1639         (WTR::LayoutTestController::suspendAnimations):
1640         (WTR::LayoutTestController::resumeAnimations):
1641         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
1642
1643 2010-11-05  Tony Chang  <tony@chromium.org>
1644
1645         Reviewed by David Levin.
1646
1647         cleanup style in TestNetscapePlugIn/PluginObject.cpp
1648         https://bugs.webkit.org/show_bug.cgi?id=49044
1649
1650         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1651         (getPluginClass):
1652         (pluginGetProperty):
1653         (pluginSetProperty):
1654         (variantToIdentifier):
1655         (testCallback):
1656         (getURL):
1657         (getURLNotify):
1658         (testInvokeDefault):
1659         (testGetProperty):
1660         (testDocumentOpen):
1661         (testWindowOpen):
1662         (pluginInvoke):
1663         (handleCallback):
1664         (notifyStream):
1665         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
1666
1667 2010-11-05  Eric Seidel  <eric@webkit.org>
1668
1669         Reviewed by David Levin.
1670
1671         Add QueueStatusServer/__init__.py so others can run the QueueStatusServer tests
1672         https://bugs.webkit.org/show_bug.cgi?id=49032
1673
1674         I wrote this file as part of bug 47847, but I forgot to commit it.
1675         No one else noticed it missing because test-webkitpy knows how
1676         to recover in the case where it can't import QueueStatusServer
1677         (which generally occurs due to not having installed the AppEngine SDK).
1678
1679         * QueueStatusServer/__init__.py: Added.
1680         * QueueStatusServer/model/workitems_unittest.py:
1681          - Remove a test which fails.  This was probably landed (by me)
1682            from my other machine, which since this __init__.py was missing
1683            I never noticed the failure and landed this invalid test.
1684            Sadly we can't really test remove_work_item as it depends
1685            on .key() working.  .key() will throw unless the object
1686            has already been saved it seems.
1687            This may be a mis-design in our remove_work_item implementation,
1688            but for now, just removing the test.
1689
1690 2010-11-04  Adam Roben  <aroben@apple.com>
1691
1692         Teach check-webkit-style about TestNetscapePlugIn
1693
1694         Fixes <http://webkit.org/b/49030> check-webkit-style is ignorant of
1695         TestNetscapePlugIn's coding conventions
1696
1697         Reviewed by John Sullivan.
1698
1699         * Scripts/webkitpy/style/checker.py: Lump TestNetscapePlugIn in with
1700         WebKitAPITest and TestWebKitAPI in having different include order and
1701         naming conventions than WebCore. Added some comments about why the
1702         exceptions exist.
1703
1704 2010-11-04  Adam Roben  <aroben@apple.com>
1705
1706         Add a test that shows that windowed plugins are able to paint
1707
1708         Somehow we never had a test for this before.
1709
1710         Fixes <http://webkit.org/b/49029> Should add a test that shows
1711         windowed plugins can paint
1712
1713         Reviewed by Jon Honeycutt.
1714
1715         * DumpRenderTree/TestNetscapePlugIn/Tests/win/DrawsGradient.cpp: Added.
1716         (DrawsGradient::DrawsGradient):
1717         (DrawsGradient::wndProc): We handle the WM_PAINT and WM_PRINTCLIENT messages.
1718
1719         (DrawsGradient::onPaint):
1720         (DrawsGradient::onPrintClient):
1721         These both just call through to paint.
1722
1723         (DrawsGradient::paint): Fills our client area with some gradients.
1724
1725         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Link
1726         against Msimg32.lib for ::GradientFill and added DrawsGradient.
1727
1728         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
1729         (createBitmapContextFromWebView): Use WM_PRINT instead of
1730         WM_PRINTCLIENT so that ::DefWindowProc will send
1731         WM_PRINT/WM_PRINTCLIENT messages to the WebView's child windows.
1732         Replaced kCGImageAlphaPremultipledFirst with kCGImageAlphaNoneSkipFirst
1733         because GDI doesn't support alpha and kCGImageBitmapOrder32Little with
1734         kCGImageBitmapOrder32Host because, who knows, maybe someday Windows
1735         will run on a big-endian platform.
1736
1737 2010-11-04  Adam Roben  <aroben@apple.com>
1738
1739         Extract much of NPNInvalidateRectInvalidatesWindow's code into a
1740         WindowedPluginTest base class
1741
1742         The base class takes care of subclassing the plugin's window so that a
1743         custom WNDPROC is called. This will make it easier to write tests that
1744         need to handle window messages.
1745
1746         Fixes <http://webkit.org/b/49028> It's hard to write a PluginTest with
1747         a custom WNDPROC
1748
1749         Reviewed by Jon Honeycutt.
1750
1751         * DumpRenderTree/TestNetscapePlugIn/Tests/win/NPNInvalidateRectInvalidatesWindow.cpp:
1752         (NPNInvalidateRectInvalidatesWindow::NPNInvalidateRectInvalidatesWindow):
1753         (NPNInvalidateRectInvalidatesWindow::NPP_SetWindow):
1754         (NPNInvalidateRectInvalidatesWindow::wndProc):
1755         (NPNInvalidateRectInvalidatesWindow::testInvalidateRect):
1756         Moved code from here to WindowedPluginTest. Changed to use window()
1757         instead of m_window.
1758
1759         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1760         Added WindowedPluginTest and added TestNetscapePlugIn/win to the
1761         include path. Also reordered the include path to make a little more
1762         sense and simplified the entry that added TestNetscapePlugIn itself to
1763         the include path.
1764
1765         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.cpp: Added.
1766         (WindowedPluginTest::WindowedPluginTest):
1767         (WindowedPluginTest::NPP_SetWindow):
1768         (WindowedPluginTest::staticWndProc):
1769         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.h: Added.
1770         (WindowedPluginTest::window):
1771         Code came from NPNInvalidateRectInvalidatesWindow.
1772
1773 2010-11-04  Adam Roben  <aroben@apple.com>
1774
1775         Add a plugin test to show that windowed plugins are clipped correctly
1776
1777         Fixes <http://webkit.org/b/49024> <rdar://problem/8487847> Windowed
1778         plugins aren't clipped in WebKit2 on Windows
1779
1780         Reviewed by Jon Honeycutt.
1781
1782         * DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowRegionIsSetToClipRect.cpp: Added.
1783         (WindowRegionIsSetToClipRect::WindowRegionIsSetToClipRect): Initialize members.
1784         (WindowRegionIsSetToClipRect::NPP_SetWindow): Check that our window
1785         region matches the clip rect we know we should have based on
1786         window-region-is-set-to-clip-rect.html, and check that our window class
1787         doesn't have the CS_PARENTDC style.
1788
1789         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1790         Added WindowRegionIsSetToClipRect.
1791
1792 2010-11-05  Alejandro G. Castro  <alex@igalia.com>
1793
1794         Reviewed by Martin Robinson.
1795
1796         [GTK] Avoid font hinting in the DRT
1797         https://bugs.webkit.org/show_bug.cgi?id=48548
1798
1799         Change the settings to avoid font hinting, it was causing
1800         different results depending on the distribution.
1801
1802         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1803         (initializeGtkFontSettings):
1804         * DumpRenderTree/gtk/fonts/fonts.conf:
1805
1806 2010-11-05  Eric Seidel  <eric@webkit.org>
1807
1808         Reviewed by Adam Barth.
1809
1810         build-webkit spams system.log by repeatedly checking for PBXProductDirectory in com.apple.Xcode
1811         https://bugs.webkit.org/show_bug.cgi?id=49051
1812
1813         This is a speculative fix.  The unit tests cover these methods, however
1814         I don't know if this will fully stop the system.log spam.
1815
1816         * Scripts/webkitpy/layout_tests/port/base.py:
1817         * Scripts/webkitpy/layout_tests/port/webkit.py:
1818
1819 2010-11-05  Eric Seidel  <eric@webkit.org>
1820
1821         Reviewed by Adam Barth.
1822
1823         Add basic support for showing bot id on /queue-status/ pages
1824         https://bugs.webkit.org/show_bug.cgi?id=49037
1825
1826         This support is really simple.  Eventually we'll want to
1827         show the bot id in the lock table too, but we don't have
1828         that information stored in the server yet.
1829
1830         * QueueStatusServer/handlers/queuestatus.py:
1831         * QueueStatusServer/handlers/queuestatus_unittest.py: Added.
1832         * QueueStatusServer/templates/includes/singlequeuestatus.html:
1833
1834 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
1835
1836         Reviewed by Adam Barth.
1837
1838         Create a filesystem wrapper that we can use to enforce
1839         particular conventions and use for mocking and dependency
1840         injection down the line.
1841
1842         https://bugs.webkit.org/show_bug.cgi?id=48144
1843
1844         * Scripts/webkitpy/common/system/filesystem.py: Added.
1845         * Scripts/webkitpy/common/system/filesystem_unittest.py: Added.
1846
1847 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1848
1849         Reviewed by Tony Chang.
1850
1851         run_webkit_tests_unittest fails under Python 2.5
1852         https://bugs.webkit.org/show_bug.cgi?id=49043
1853         
1854         Switch from itertools.chain.from_iterable (which was added in 2.6)
1855         to using itertools.chain directly.
1856
1857         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1858
1859 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1860
1861         Unreviewed fix to webkit-patch rebaseline-server so that it runs under
1862         Python 2.5 (it needs an import to use the with statement).
1863
1864         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
1865
1866 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
1867
1868         Reviewed by Adam Barth.
1869
1870         new-run-webkit-tests: split out webkit-specific configuration stuff into a new module
1871
1872         The current NRWT code has webkit-specific configuration code (like
1873         _script_path, default configuration, etc.) mixed in with
1874         layout-test-specific stuff in port/base. The configuration code
1875         should be split out into a separate module for easier mocking,
1876         testing, and isolation.
1877
1878         https://bugs.webkit.org/show_bug.cgi?id=48264
1879
1880         * Scripts/webkitpy/layout_tests/port/config.py: Added.
1881         * Scripts/webkitpy/layout_tests/port/config_unittest.py: Added.
1882
1883 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1884
1885         Reviewed by Tony Chang.
1886
1887         Rebaseline server: initial framework
1888         https://bugs.webkit.org/show_bug.cgi?id=48892
1889         
1890         Adds the basic framework for the rebaseline server (details at 
1891         http://webkit.org/b/47761). Includes the rebaseline-server webkit-patch
1892         command, which starts an HTTP server that can serve static files or
1893         invoke handler methods on a class.
1894
1895         * Scripts/webkitpy/tool/commands/__init__.py:
1896         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html: Added.
1897         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css: Added.
1898         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js: Added.
1899         * Scripts/webkitpy/tool/commands/rebaselineserver.py: Added.
1900
1901 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
1902
1903         Reviewed by James Robinson.
1904
1905         new-run-webkit-tests wasn't using DRT by default for
1906         --platform chromium-gpu
1907
1908         The default value was set to False instead of None, which meant
1909         that the platform specific logic wasn't firing to change the
1910         value to True (b/c we were afraid we'd be overriding the user
1911         preference).
1912
1913         https://bugs.webkit.org/show_bug.cgi?id=49038
1914
1915         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1916
1917 2010-11-04  Tony Chang  <tony@chromium.org>
1918
1919         Reviewed by Adam Barth.
1920
1921         make platform/chromium/plugins/multiple-plugins.html pass on all platforms
1922         https://bugs.webkit.org/show_bug.cgi?id=49026
1923
1924         Have the testObject.bar property return the string "bar".  This lets
1925         us run and pass platform/chromium/plugins/multiple-plugins.html on all
1926         platforms.
1927
1928         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
1929         (testGetProperty):
1930
1931 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1932
1933         Reviewed by Tony Chang.
1934
1935         NRWT doesn't have good test coverage for --run-chunk, --batch-size, --run-part, etc.
1936         https://bugs.webkit.org/show_bug.cgi?id=48878
1937         
1938         Add get_tests_run so that it's easy to see which tests get run (and with
1939         what batching) for a given flag combination. Flesh out the various
1940         test cases that have FIXMEs.
1941         
1942         Also fixes an off-by-one error (batch sizes were one larger than 
1943         expected) and makes --run-part also have wraparound behavior, like
1944         --run-chunk.
1945
1946         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1947         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1948         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1949
1950 2010-11-04  Erik Arvidsson  <arv@chromium.org>
1951
1952         Reviewed by Dimitri Glazkov.
1953
1954         Support box-sizing without the vendor prefix
1955         https://bugs.webkit.org/show_bug.cgi?id=36713
1956
1957         Based on patch by Peter Beverloo <peter@lvp-media.com>
1958
1959         * iExploder/htdocs/cssproperties.in: Change -webkit-box-sizing to box-sizing.
1960
1961 2010-11-04  Csaba Osztrogonác  <ossy@webkit.org>
1962
1963         Unreviewed rollout r71340, because it broke Chromium Windows bot.
1964
1965         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1966         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1967
1968 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1969
1970         Reviewed by Ojan Vafai.
1971
1972         Make http locking default in NRWT.
1973         https://bugs.webkit.org/show_bug.cgi?id=48053
1974
1975         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1976         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1977
1978 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1979
1980         Reviewed by Eric Seidel.
1981
1982         [NRWT] Clear invalid http locks on Windows platform as well
1983         https://bugs.webkit.org/show_bug.cgi?id=48515
1984
1985         * Scripts/webkitpy/common/system/executive.py:
1986         * Scripts/webkitpy/common/system/executive_unittest.py:
1987         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1988
1989 2010-11-02  Adam Roben  <aroben@apple.com>
1990
1991         Reduce our dependence on coreutils when running Python tests
1992
1993         This patch introduces versions of the standard echo and cat utilities
1994         implemented in Python. They are probably missing features of their
1995         coreutils equivalents, but they can do what's necessary for our Python
1996         tests. This is useful on Windows, where these utilities typically
1997         aren't available.
1998
1999         Fixes <http://webkit.org/b/48883> executive_unittest relies on echo
2000         and cat utilities from coreutils, which are not present on Windows
2001
2002         Reviewed by Eric Seidel.
2003
2004         * Scripts/webkitpy/common/system/executive_unittest.py: Changed to use
2005         our Python-based echo and cat.
2006
2007         * Scripts/webkitpy/common/system/fileutils.py: Added.
2008         (make_stdout_binary): On Windows, puts sys.stdout into binary mode so
2009         that \n won't be translated into \r\n. I couldn't think of a good way
2010         to test this directly without touching the filesystem, but it is tested
2011         indirectly by echo_unittest.
2012
2013         * Scripts/webkitpy/test/cat.py: Added.
2014         (command_arguments): Returns a list for invoking cat with the given arguments.
2015         (main): Acts like a simplified version of the coreutils cat utility.
2016
2017         * Scripts/webkitpy/test/cat_unittest.py: Added.
2018         (CatTest.assert_cat): Runs cat with the given input and ensures the
2019         output matches the input.
2020         (CatTest.test_basic): Performs a simple test of cat.
2021         (CatTest.test_no_newline): Tests what happens when the input string
2022         doesn't have a trailing newline.
2023         (CatTest.test_unicode): Tests passing a unicode string to cat.
2024         (CatTest.test_as_command): Tests running cat as a separate command.
2025
2026         * Scripts/webkitpy/test/echo.py: Added.
2027         (command_arguments): Returns a list for invoking echo with the given arguments.
2028         (main): Acts like a simplified version of the coreutils echo utility.
2029
2030         * Scripts/webkitpy/test/echo_unittest.py: Added.
2031         (EchoTest.test_basic): Performs a simple test of echo.
2032         (EchoTest.test_no_newline): Tests passing -n to echo to suppress the
2033         trailing newline.
2034         (EchoTest.test_unicode): Tests passing unicode and non-unicode strings
2035         to echo.
2036         (EchoTest.test_argument_order): Tests what happens when -n is not the
2037         first argument.
2038         (EchoTest.test_empty_arguments): Tests what happens when you pass [] to
2039         echo.main.
2040         (EchoTest.test_no_arguments): Tests what happens when you call
2041         echo.main with no arguments.
2042         (EchoTest.test_as_command): Tests running echo as a separate command.
2043
2044 2010-11-04  Renata Hodovan  <reni@inf.u-szeged.hu>
2045
2046         Unreviewed: Add myself to the list of Committers.
2047
2048         * Scripts/webkitpy/common/config/committers.py:
2049
2050 2010-11-04  Andreas Kling  <kling@webkit.org>
2051
2052         Reviewed by Kenneth Rohde Christiansen.
2053
2054         [Qt] Hook into QWebPage::scrollRequested for resizes-to-contents mode
2055
2056         In RTC mode the QGraphicsWebView item is the size of the contents,
2057         scrolling works a bit differently  (we need to react to scrollRequested.)
2058
2059         Normally QGraphicsView will replay the last mouse event when scrolling,
2060         so to prevent WebKit from getting confused by this we temporarily make
2061         the QGraphicsView non-interactive.
2062
2063         * QtTestBrowser/webview.cpp:
2064         (WebViewGraphicsBased::setPage):
2065         (WebViewGraphicsBased::scrollRequested):
2066         * QtTestBrowser/webview.h:
2067
2068 2010-11-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2069
2070         Reviewed by Andreas Kling.
2071
2072         Use OS(MAC_OS_X) rather than PLATFORM(MAC) in TestControllerQt
2073
2074         PLATFORM(MAC) is not defined for the Qt port, as it refers to
2075         the Mac-port, not the Mac OS X operating system.
2076
2077         * WebKitTestRunner/qt/TestControllerQt.cpp:
2078         (WTR::isExistingLibrary):
2079
2080 2010-11-04  Leonid Ebril  <leonid.ebril@nokia.com>
2081
2082         Reviewed by Andreas Kling.
2083
2084         [Qt] Adding iphone user agent string the user agent list for QtTestBrowser
2085         https://bugs.webkit.org/show_bug.cgi?id=48869
2086
2087         * QtTestBrowser/useragentlist.txt:
2088
2089 2010-11-03  Adam Roben  <aroben@apple.com>
2090
2091         Always use uppercase drive names in strings returned by abspath_to_uri
2092
2093         Some versions of cygpath use lowercase drive letters while others use
2094         uppercase, which makes it hard to test the output of code that uses
2095         cygpath.
2096
2097         Fixes <http://webkit.org/b/48914> webkitpy.common.system.path_unittest
2098         fails with Cygwin 1.5
2099
2100         Reviewed by Eric Seidel.
2101
2102         * Scripts/webkitpy/common/system/path.py:
2103         (cygpath): Updated the docstring to indicate that only absolute paths
2104         should be passed for now (though relative paths will work fine).
2105         (_Cygpath.convert): Upper-case the first letter of the converted Windows path.
2106
2107 2010-11-03  George Guo  <George.Guo@Nokia.com>
2108
2109         Reviewed by Andreas Kling.
2110
2111         [Qt] QtTestBrowser : set mmp rule pageddata in Symbian
2112         https://bugs.webkit.org/show_bug.cgi?id=48767
2113     
2114         Paging is needd on Symbian devices to support benchmarks tests like
2115         dromaeo.com and Celtic Kane that need a lot of memory to run  
2116
2117         * QtTestBrowser/QtTestBrowser.pro:
2118
2119 2010-11-03  Jenn Braithwaite  <jennb@chromium.org>
2120
2121         Reviewed by Dmitry Titov.
2122
2123         Chromium: Update resource tracking when moving a frame between documents
2124         https://bugs.webkit.org/show_bug.cgi?id=48363
2125
2126         * DumpRenderTree/chromium/WebViewHost.cpp:
2127         (WebViewHost::assignIdentifierToRequest):
2128         Always put resource id in map so we can make assumptions about its
2129         presence.
2130         (WebViewHost::removeIdentifierForRequest):
2131         Added.
2132         (WebViewHost::didFinishResourceLoad):
2133         (WebViewHost::didFailResourceLoad):
2134         Check existence of resource id before removing from map.
2135         * DumpRenderTree/chromium/WebViewHost.h:
2136
2137 2010-11-03  Victor Wang  <victorw@chromium.org>
2138
2139         Reviewed by Adam Barth.
2140
2141         [Chromium] update buildbot names in chromium rebaseline tool.
2142
2143         https://bugs.webkit.org/show_bug.cgi?id=48881
2144
2145         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2146
2147 2010-11-03  Darin Adler  <darin@apple.com>
2148
2149         Updated Xcode projects by opening them with Xcode 3.2.4.
2150         Updated svn:ignore for Xcode projects.
2151
2152         * MiniBrowser/MiniBrowser.xcodeproj: Added property svn:ignore.
2153         * TestWebKitAPI/TestWebKitAPI.xcodeproj: Added property svn:ignore.
2154         * WebKitLauncher/WebKitLauncher.xcodeproj: Modified property svn:ignore.
2155         * WebKitLauncher/WebKitLauncher.xcodeproj/project.pbxproj: Updated with Xcode 3.2.4.
2156
2157 2010-11-03  Adam Roben  <aroben@apple.com>
2158
2159         Stop waiting for messages from the web process after a timeout elapses
2160
2161         TestController::runUntil now takes a timeout parameter that specifies
2162         how long to wait for the condition to become true. WebKitTestRunner
2163         takes a --timeout flag that specifies how long the timeout should be.
2164         run-webkit-tests passes this flag to WebKitTestRunner so its timeout
2165         will be similar to run-webkit-tests.
2166
2167         Fixes <http://webkit.org/b/43047> <rdar://problem/8365833>
2168         WebKitTestRunner waits forever if the web process crashes
2169
2170         Reviewed by Darin Adler and Anders Carlsson.
2171
2172         * Scripts/old-run-webkit-tests:
2173         (top level): Moved the GuardMalloc timeout adjustment here from
2174         readFromDumpToolWithTimer.
2175         (openDumpTool): Make WTR use a timeout similar to but slightly shorter
2176         than the one that was specified on the command line.
2177
2178         * WebKitTestRunner/TestController.cpp:
2179         (WTR::TestController::TestController): Initialize our timeout values.
2180         (WTR::TestController::initialize): Parse the --timeout flag and use it
2181         to modify our timeout values.
2182         (WTR::TestController::resetStateToConsistentValues): Changed to use a
2183         short timeout while waiting for the web process to reset and to return
2184         a boolean indicating whether we were able to reset the web process.
2185         Uses a 5-second timeout while waiting for the process to be reset.
2186         (WTR::TestController::runTest): Changed to return a boolean indicating
2187         whether we were able to reset the web process (and thus run the test).
2188         (WTR::TestController::runUntil): Call through to platformRunUntil.
2189
2190         (WTR::TestController::runTestingServerLoop):
2191         (WTR::TestController::run):
2192         Changed to bail if any test can't be run. This will cause the process
2193         to exit. (Unfortunately this will make run-webkit-tests think we
2194         crashed; see <http://webkit.org/b/48943>.)
2195
2196         * WebKitTestRunner/TestController.h: Added platformRunUntil,
2197         m_longTimeout, and m_shortTimeout.
2198
2199         * WebKitTestRunner/TestInvocation.cpp:
2200         (WTR::TestInvocation::invoke): Use a short timeout when waiting for the
2201         initial response and a long timeout when waiting for the test to
2202         complete. Check whether runUntil timed out and print an error message
2203         if so.
2204
2205         * WebKitTestRunner/mac/TestControllerMac.mm:
2206         (WTR::TestController::platformRunUntil): Renamed from runUntil. Pass
2207         [NSDate distantPast] to -[NSRunLoop runMode:beforeDate:] so that we
2208         won't block waiting for the run loop. Only loop until the timeout
2209         elapses.
2210
2211         * WebKitTestRunner/qt/TestControllerQt.cpp:
2212         (WTR::TestController::platformRunUntil): Renamed from runUntil. Added a
2213         FIXME about honoring the timeout.
2214
2215         * WebKitTestRunner/win/TestControllerWin.cpp:
2216         (WTR::TestController::platformRunUntil): Renamed from runUntil. Use
2217         ::MsgWaitForMultipleObjectsEx to implement the timeout. Changed to use
2218         ::PeekMessageW so that we don't block waiting for messages to become
2219         available.
2220
2221 2010-11-03  Adam Roben  <aroben@apple.com>
2222
2223         Add a plugin test that evaluates JS after removing the plugin element
2224         from the document
2225
2226         This test replaces platform/win/plugins/plugin-delayed-destroy.html.
2227         That test was made to prevent a crash very similar to this one, but
2228         unfortunately tested only the mechanism that prevented the crash and
2229         not whether the crash itself was prevented. Since WebKit2 uses a
2230         different mechanism to prevent the crash, the test was failing even
2231         though WebKit2 was not vulnerable to the crash. This new test crashes
2232         if there is no mechanism in place to prevent it and passes in both
2233         WebKit1 and WebKit2.
2234
2235         Fixes <http://webkit.org/b/46711> <rdar://problem/8485903>
2236         platform/win/plugins/plugin-delayed-destroy.html fails in WebKit2
2237
2238         Reviewed by Anders Carlsson.
2239
2240         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
2241         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2242         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
2243         * GNUmakefile.am:
2244         Added new file.
2245
2246         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
2247         (pluginDeallocate): Make sure we delete the PluginTest object. This
2248         prevents a leak and also allows us to test the crash.
2249
2250         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
2251         (PluginTest::executeScript): Made this into a non-static member
2252         function.
2253
2254         (PluginTest::waitUntilDone):
2255         (PluginTest::notifyDone):
2256         Updated for changes to executeScript.
2257
2258         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Added executeScript.
2259
2260         * DumpRenderTree/TestNetscapePlugIn/Tests/EvaluateJSAfterRemovingPluginElement.cpp: Added.
2261         (EvaluateJSAfterRemovingPluginElement::EvaluateJSAfterRemovingPluginElement):
2262         Initialize ourselves and tell the test harness to wait.
2263         (EvaluateJSAfterRemovingPluginElement::NPP_DestroyStream): Remove our
2264         plugin element from the document, then execute some JavaScript. If
2265         WebKit does not have appropriate mechanisms in place, we'll be
2266         destroyed inside the first call to executeScript and crash on the
2267         second call.
2268
2269 2010-11-02  Stephen White  <senorblanco@chromium.org>
2270
2271         Reviewed by Tony Chang.
2272
2273         [chromium] Fix LayoutTestController UMRs.
2274         https://bugs.webkit.org/show_bug.cgi?id=48872
2275
2276         * DumpRenderTree/chromium/LayoutTestController.cpp:
2277         (LayoutTestController::LayoutTestController):
2278
2279 2010-11-03  Kent Tamura  <tkent@chromium.org>
2280
2281         Reviewed by Dimitri Glazkov.
2282
2283         REGRESSION: rebaseline-chromium-webkit-tests uses non-zero tolerance for
2284         image dup detection
2285         https://bugs.webkit.org/show_bug.cgi?id=48744
2286
2287         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2288          - Make a function for option parsing for ease of test
2289          - Set 0 to options.tolerance
2290         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
2291          - Add a test for this change
2292
2293 2010-11-02  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
2294
2295         Unreviewed: Add myself to the list of Committers.
2296
2297         * Scripts/webkitpy/common/config/committers.py:
2298
2299 2010-11-02  Anders Carlsson  <andersca@apple.com>
2300
2301         Fix build.
2302
2303         * TestWebKitAPI/PlatformUtilities.cpp:
2304         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
2305
2306 2010-11-02  Sam Weinig  <sam@webkit.org>
2307
2308         Reviewed by Anders Carlsson.
2309
2310         Add a way to send startup messages on the context which can be posted when a process launches
2311         <rdar://problem/8617928>
2312         https://bugs.webkit.org/show_bug.cgi?id=48838
2313
2314         * MiniBrowser/mac/WebBundle/WebBundleMain.m:
2315         (WKBundleInitialize):
2316         * TestWebKitAPI/InjectedBundleController.cpp:
2317         (TestWebKitAPI::InjectedBundleController::initialize):
2318         (TestWebKitAPI::InjectedBundleController::didReceiveMessage):
2319         * TestWebKitAPI/InjectedBundleController.h:
2320         * TestWebKitAPI/InjectedBundleMain.cpp:
2321         (WKBundleInitialize):
2322         * TestWebKitAPI/PlatformUtilities.cpp:
2323         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
2324         * WebKitTestRunner/InjectedBundle/InjectedBundleMain.cpp:
2325         (WKBundleInitialize):
2326         Update implementations of WKBundleInitialize to take an initial userData
2327         argument.  Change TestWebKitAPI to use the new initial userData to initialize
2328         each test's bundle.
2329
2330 2010-11-02  Benjamin Kalman  <kalman@google.com>
2331
2332         Reviewed by Ojan Vafai.
2333
2334         new-run-webkit-tests doesn't strip "LayoutTests/" from prefix, unlike old-run-webkit-tests
2335         https://bugs.webkit.org/show_bug.cgi?id=48794
2336
2337         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: Strip the "LayoutTests/" prefix from test argument paths.
2338
2339 2010-11-02  Adam Roben  <aroben@apple.com>
2340
2341         Skip webkitpy.layout_tests.run_webkit_tests_unittest.MainTest on Cygwin
2342         Python 2.5.x
2343
2344         It is known to hang on that version of Python. See
2345         <http://webkit.org/b/48614>.
2346
2347         Reviewed by Adam Barth.
2348
2349         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py: Use
2350         skip_if to skip MainTest on Cygwin Python 2.5.x.
2351
2352         * Scripts/webkitpy/test/skip.py: Added.
2353         (skip_if): If the passed-in condition is false, find all the test_*
2354         methods of the given class and replace them with a function that just
2355         logs that we're skipping these tests. This is loosely based on the
2356         unittest.skip_if decorator added in Python 3.1.
2357         (_skipped_method): Returns a function that just logs that the tests are
2358         being skipped. This is used in place of the actual test_* functions.
2359
2360         * Scripts/webkitpy/test/skip_unittest.py: Added.
2361         (SkipTest.setUp): Create a logger for ourselves and give it a handler
2362         that logs to a private stream.
2363         (SkipTest.tearDown): Reset the logger.
2364         (SkipTest.create_fixture_class): Returns a class that we use to test
2365         skip_if. It has a single test_* method, test_foo, that just calls
2366         through to a callback.
2367         (SkipTest.foo_callback): Record that test_foo was called.
2368         (SkipTest.test_skip_if_false): Pass skip_if a False condition and test
2369         that test_foo does get called.
2370         (SkipTest.test_skip_if_true): Pass skip_if a True condition and test
2371         that test_foo does not get called and the appropriate message gets
2372         logged.
2373
2374 2010-11-02  Adam Barth  <abarth@webkit.org>
2375
2376         Reviewed by Eric Seidel.
2377
2378         webkit-patch should tell check-webkit-style which files were changed so
2379         check-webkit-style doesn't have to stat the whole working copy again
2380         https://bugs.webkit.org/show_bug.cgi?id=48792
2381
2382         * Scripts/webkitpy/tool/mocktool.py:
2383         * Scripts/webkitpy/tool/steps/checkstyle.py:
2384
2385 2010-11-02  Robert Kroeger  <rjkroege@chromium.org>
2386
2387         Reviewed by James Robinson.
2388
2389         [Chromium/DRT] Make EventSendingController honour leapForward for touch events.
2390         https://bugs.webkit.org/show_bug.cgi?id=48777
2391
2392         * DumpRenderTree/chromium/EventSender.cpp:
2393         (EventSender::sendCurrentTouchEvent):
2394
2395 2010-11-02  Adam Roben  <aroben@apple.com>
2396
2397         Only track resource identifiers in DRT when dumpResourceLoadCallbacks
2398         is on
2399
2400         This reverts Windows to our pre-r71097 behavior. That patch made us
2401         track all resource identifiers, including the main resource, so the
2402         main resource's URL started appearing in test output instead of
2403         "<unknown>". Arguably having the main resource's URL is better, but all
2404         other platforms print "<unknown>" and we want to match.
2405
2406         Fixes <http://webkit.org/b/48837> <rdar://problem/8620351> REGRESSION
2407         (r71097): Many http tests failing on Windows
2408
2409         Reviewed by Anders Carlsson.
2410
2411         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
2412         (ResourceLoadDelegate::identifierForInitialRequest): Don't add the
2413         identifier to the URL map if we're not supposed to dump resource load
2414         callbacks.
2415         (ResourceLoadDelegate::removeIdentifierForRequest): Always remove the
2416         identifier from the URL map even if we're already "done". There's no
2417         point in keeping out-of-date identifiers around.
2418
2419 2010-11-01  Jenn Braithwaite  <jennb@chromium.org>
2420
2421         Reviewed by Adam Roben.
2422
2423         Windows: Update resource tracking when moving a frame between documents
2424         https://bugs.webkit.org/show_bug.cgi?id=48364
2425
2426         * DumpRenderTree/win/DumpRenderTree.cpp:
2427         (createWebViewAndOffscreenWindow):
2428         (main):
2429         Give each WebView its own ResourceLoadDelegate instance in order to
2430         make assertions about resource ids on a particular WebView.
2431         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
2432         (ResourceLoadDelegate::identifierForInitialRequest):
2433         Always add id to the map.
2434         (ResourceLoadDelegate::removeIdentifierForRequest):
2435         Added.
2436         (ResourceLoadDelegate::willSendRequest):
2437         (ResourceLoadDelegate::didReceiveAuthenticationChallenge):
2438         (ResourceLoadDelegate::didReceiveResponse):
2439         (ResourceLoadDelegate::didFinishLoadingFromDataSource):
2440         (ResourceLoadDelegate::didFailLoadingWithError):
2441         (ResourceLoadDelegate::descriptionSuitableForTestResult):
2442         Replace static descriptionSuitableForTestResult with static member function to access identifier map.
2443         * DumpRenderTree/win/ResourceLoadDelegate.h:
2444         (ResourceLoadDelegate::urlMap):
2445         Moved within class so that each WebView has its own id map.
2446
2447 2010-11-01  Ojan Vafai  <ojan@chromium.org>
2448
2449         Reviewed by Dimitri Glazkov.
2450
2451         remove debug code from run_webkit_tests.py
2452         https://bugs.webkit.org/show_bug.cgi?id=48800
2453
2454         Remove temporary debug code and make --master-name required
2455         if --test-results-server is set now that all clients set
2456         --master-name.
2457
2458         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2459
2460 2010-11-01  Tony Chang  <tony@chromium.org>
2461
2462         Reviewed by Kent Tamura.
2463
2464         add plugin event logging for linux
2465         https://bugs.webkit.org/show_bug.cgi?id=48779
2466
2467         This is taken from Chromium's fork of the layout test plugin:
2468         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
2469
2470         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
2471         (handleEventX11):
2472         (NPP_HandleEvent):
2473
2474 2010-11-01  Dirk Pranke  <dpranke@chromium.org>
2475
2476         Reviewed by James Robinson.
2477
2478         new-run-webkit-tests: use DRT, child-processes=1 for GPU tests by default
2479         https://bugs.webkit.org/show_bug.cgi?id=48790
2480
2481         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2482         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2483
2484 2010-11-01  Adam Barth  <abarth@webkit.org>
2485
2486         Reviewed by Eric Seidel.
2487
2488         Teach check-webkit-style how to accept a list of files to diff on the
2489         command line
2490         https://bugs.webkit.org/show_bug.cgi?id=48784
2491
2492         In a future patch, webkit-patch will use this option to improve
2493         performance.  I'm landing this in two pieces to avoid causing a version
2494         skew problem for the style-bot.
2495
2496         * Scripts/check-webkit-style:
2497         * Scripts/webkitpy/style/optparser.py:
2498         * Scripts/webkitpy/style/optparser_unittest.py:
2499         * Scripts/webkitpy/style_references.py:
2500
2501 2010-11-01  Anders Carlsson  <andersca@apple.com>
2502
2503         Reviewed by John Sullivan.
2504
2505         Tear down the related WebProcessProxy when a WebContext is deallocated
2506         https://bugs.webkit.org/show_bug.cgi?id=48769
2507
2508         * TestWebKitAPI/Tests/WebKit2/FailedLoad.cpp:
2509         (TestWebKitAPI::didFailProvisionalLoadWithErrorForFrame):
2510         We don't support empty URLs anymore, update test to expect a null URL instead.
2511
2512 2010-11-01  Søren Gjesse  <sgjesse@chromium.org>
2513
2514         Reviewed by Andreas Kling.
2515
2516         Fix warning when compiling the chromium port of DumpRenderShell
2517         with clang.
2518         https://bugs.webkit.org/show_bug.cgi?id=48414
2519
2520         * DumpRenderTree/chromium/TestShell.h:
2521         (TestShell::javaScriptFlagsForLoad):
2522
2523 2010-11-01  Adam Roben  <aroben@apple.com>
2524
2525         Fix typo from r71022
2526
2527         * BuildSlaveSupport/build.webkit.org-config/config.json:
2528
2529 2010-11-01  Adam Roben  <aroben@apple.com>
2530
2531         Trigger the Windows Release WebKit2 tests when a Release build
2532         finishes, not when a Debug build finishes
2533
2534         Fixes <http://webkit.org/b/48754> Windows Release WebKit2 tests are
2535         triggered at the wrong time
2536
2537         Reviewed by Sam Weinig.
2538
2539         * BuildSlaveSupport/build.webkit.org-config/config.json: Fixed
2540         triggerable name and trigger.
2541
2542 2010-11-01  Mario Sanchez Prada  <msanchez@igalia.com>
2543
2544         Unreviewed. Adding my IRC nickname to the list of committers.
2545
2546         * Scripts/webkitpy/common/config/committers.py:
2547
2548 2010-10-31  Robert Hogan  <robert@webkit.org>
2549
2550         Reviewed by Antonio Gomes.
2551
2552         [Qt] [Gtk] Plug-ins having upper case in mime type are failing to load
2553
2554         Qt and Gtk are case-sensitive when storing the declared mime-type
2555         of plugins. Since plugin mime-types are lowercased prior to searching
2556         for them in the plugin database, ensure they are loaded with the
2557         mime-type in lower case too.
2558
2559         Change the test netscape plugin to declare its mimetype in sentence
2560         case so that the correct behaviour is enforced.
2561
2562         https://bugs.webkit.org/show_bug.cgi?id=36815
2563
2564         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
2565         (NP_GetMIMEDescription):
2566
2567 2010-10-31  Robert Hogan  <robert@webkit.org>
2568
2569         Reviewed by Antonio Gomes.
2570
2571         [Qt] Support nodesFromRect in DRT
2572
2573         https://bugs.webkit.org/show_bug.cgi?id=48716
2574
2575         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2576         (LayoutTestController::nodesFromRect):
2577         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2578
2579 2010-10-30  Andreas Kling  <kling@webkit.org>
2580
2581         Reviewed by Antonio Gomes.
2582
2583         [Qt] QtTestBrowser: Switching view type moves the embedded inspector
2584         https://bugs.webkit.org/show_bug.cgi?id=48705
2585
2586         Reinsert the embedded inspector into the splitter after changing
2587         between QWebView/QGraphicsWebView.
2588
2589         * QtTestBrowser/launcherwindow.cpp:
2590         (LauncherWindow::init):
2591         (LauncherWindow::initializeView):
2592         * QtTestBrowser/webinspector.h:
2593         (WebInspector::WebInspector):
2594
2595 2010-10-28  Antonio Gomes  <agomes@rim.com>
2596
2597         Reviewed by Ojan Vafai.
2598
2599         Needs a "LinuxEditingBehavior", perhaps with a better name
2600         https://bugs.webkit.org/show_bug.cgi?id=36627
2601
2602         Adding support to Mac's, GTK+'s, Windows' and Chromium's LayoutTestController class to test the newly introduced Unix editing behavior.
2603
2604         * DumpRenderTree/chromium/LayoutTestController.cpp:
2605         (LayoutTestController::setEditingBehavior):
2606         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2607         (LayoutTestController::setEditingBehavior):
2608         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2609         (LayoutTestController::setEditingBehavior):
2610         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2611         (LayoutTestController::setEditingBehavior):
2612
2613 2010-10-29  Andreas Kling  <kling@webkit.org>
2614
2615         Reviewed by Kenneth Rohde Christiansen.
2616
2617         [Qt] QtTestBrowser: Add keyboard shortcut to toggle full screen (F11)
2618         https://bugs.webkit.org/show_bug.cgi?id=48695
2619
2620         * QtTestBrowser/launcherwindow.cpp:
2621         (LauncherWindow::createChrome):
2622
2623 2010-10-29  Andreas Kling  <kling@webkit.org>
2624
2625         Reviewed by Kenneth Rohde Christiansen.
2626
2627         [Qt] QtTestBrowser: Fix uninitialized read in FpsTimer
2628         https://bugs.webkit.org/show_bug.cgi?id=48675
2629
2630         FpsTimer::m_timer was never initialized and passed to QObject::killTimer()
2631         on startup with in -graphicsbased mode.
2632
2633         * QtTestBrowser/fpstimer.cpp:
2634         (FpsTimer::FpsTimer):
2635         (FpsTimer::numFrames):
2636         (FpsTimer::stop):
2637         (FpsTimer::timerEvent):
2638         * QtTestBrowser/fpstimer.h:
2639
2640 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
2641
2642         Reviewed by Tony Chang.
2643
2644         new-run-webkit-tests: change TestResults to be serializable
2645
2646         In preparation for changing new-run-webkit-tests from
2647         multithreaded to multiprocess, we need to make sure the data
2648         going between the threads is easily serialized over a socket.
2649
2650         This change adds serialization/pickling for the TestResults and
2651         TestFailure objects (using cPickle).
2652
2653         The TestFailure objects included a "has_wdiff" flag for Text
2654         results, but the flag wasn't being used, so I've removed it,
2655         simplifying the state to basically a set of enum objects with
2656         associated methods.
2657
2658         https://bugs.webkit.org/show_bug.cgi?id=48616
2659
2660         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2661         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
2662         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
2663         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
2664         * Scripts/webkitpy/layout_tests/layout_package/test_results.py: Added.
2665         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:Added. 
2666         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2667         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2668
2669 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
2670
2671         Reviewed by Ojan Vafai.
2672
2673         new-run-webkit-tests: need to provide separate dashboard results for GPU tests
2674         https://bugs.webkit.org/show_bug.cgi?id=48687
2675
2676         Modify the steps to upload the JSON files to the dashboards so
2677         that the GPU tests don't conflict with the regular tests on a
2678         port. We do this by modifying the --builder-name parameter to
2679         append " - GPU", which should cause the app to treat the results
2680         as a completely new builder. This is a little non-obvious, but
2681         keeps us from having to restructure the app.
2682
2683         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2684         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2685         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2686
2687 2010-10-29  Ojan Vafai  <ojan@chromium.org>
2688
2689         Reviewed by Tony Chang.
2690
2691         [chromium] add debug logging to help diagnose flakiness dashboard issues
2692         https://bugs.webkit.org/show_bug.cgi?id=48657
2693
2694         The appengine app thinks it's getting empty files uploaded. Add some logging
2695         to see if new-run-webkit-tests agrees.
2696         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2697
2698 2010-10-29  Dimitri Glazkov  <dglazkov@chromium.org>
2699
2700         Reviewed by Ojan Vafai.
2701
2702         [Chromium/DRT] Add master-name flag to new-run-webkit-tests.
2703         https://bugs.webkit.org/show_bug.cgi?id=48649
2704
2705         * BuildSlaveSupport/build.webkit.org-config/master.cfg: Added flag.
2706
2707 2010-10-29  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
2708
2709         Reviewed by Martin Robinson.
2710
2711         Enable popup window in GtkLauncher
2712         https://bugs.webkit.org/show_bug.cgi?id=48335
2713
2714         The GtkLauncher application does not support opening new window when
2715         a link with "target=_blank" is clicked or similar call to
2716         window.open(). Instead, GtkLauncher does nothing which breaks
2717         navigation of some websites.
2718
2719         * GtkLauncher/main.c:
2720         (activate_uri_entry_cb):
2721         (update_title):
2722         (link_hover_cb):
2723         (notify_title_cb):
2724         (notify_load_status_cb):
2725         (notify_progress_cb):
2726         (destroy_cb):
2727         (go_back_cb):
2728         (go_forward_cb):
2729         (create_web_view_cb):
2730         (web_view_ready_cb):
2731         (close_web_view_cb):
2732         (create_browser):
2733         (create_statusbar):
2734         (create_toolbar):
2735         (create_window):
2736         (main):
2737
2738 2010-10-29  Adam Roben  <aroben@apple.com>
2739
2740         Teach check-webkit-style about WebKit2's idiosyncracies
2741
2742         Fixes <http://webkit.org/b/48638> Style bot complains about a number
2743         of WebKit2 conventions
2744
2745         Reviewed by Anders Carlsson.
2746
2747         * Scripts/webkitpy/style/checker.py: Excluded some rules for various
2748         WebKit2-related files. Also updated the excluded rules for
2749         WebKitAPITest to match the current code.
2750
2751 2010-10-29  Csaba Osztrogonác  <ossy@webkit.org>
2752
2753         Reviewed by Adam Roben and David Kilzer.
2754
2755         Fix and cleanup of build systems
2756         https://bugs.webkit.org/show_bug.cgi?id=48342
2757
2758         * Scripts/build-webkit:
2759          - Remove unnecessary ENABLE_SANDBOX option.
2760          - Add ENABLE_FULLSCREEN_API option.
2761
2762 2010-10-28  Adam Roben  <aroben@apple.com>
2763
2764         Switch the Windows WebKit2 bot to the Release configuration
2765
2766         We only have one machine testing WebKit2 on Windows right now, and
2767         Debug is just too slow for it to keep up.
2768
2769         Fixes (hopefully!) <http://webkit.org/b/48615> Windows WebKit2 bot is
2770         always way behind
2771
2772         Reviewed by Jon Honeycutt.
2773
2774         * BuildSlaveSupport/build.webkit.org-config/config.json:
2775
2776 2010-10-29  Leandro Gracia Gil  <leandrogracia@google.com>
2777
2778         Reviewed by Jeremy Orlow.
2779
2780         Added a second parameter to setMockSpeechInputResult for
2781         the language used in speech input.
2782         https://bugs.webkit.org/show_bug.cgi?id=47089
2783
2784         * DumpRenderTree/LayoutTestController.cpp:
2785         (setMockSpeechInputResultCallback):
2786         * DumpRenderTree/LayoutTestController.h:
2787         * DumpRenderTree/chromium/LayoutTestController.cpp:
2788         (LayoutTestController::setMockSpeechInputResult):
2789         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2790         (LayoutTestController::setMockSpeechInputResult):
2791         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2792         (LayoutTestController::setMockSpeechInputResult):
2793         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2794         (LayoutTestController::setMockSpeechInputResult):
2795         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2796         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2797         (LayoutTestController::setMockSpeechInputResult):
2798         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2799         (LayoutTestController::setMockSpeechInputResult):
2800         (LayoutTestController::pageSizeAndMarginsInPixels):
2801
2802 2010-10-28  Ojan Vafai  <ojan@chromium.org>
2803
2804         Reviewed by Tony Chang.
2805
2806         [chromium] add a result-small.json file for the test dashboard
2807         https://bugs.webkit.org/show_bug.cgi?id=48547
2808
2809         Output both a results.json file and a results-small.json file.
2810         The dashboard will load results-small.json by default so it loads faster.
2811
2812         * TestResultServer/model/jsonresults.py:
2813         * TestResultServer/model/jsonresults_unittest.py:
2814         Added a bunch of sys.path hackery. Unforunately, this uses hardcoded
2815         paths. That obviously needs to be fixed, but at least this way it
2816         clearly documents what paths are necessary.
2817
2818 2010-10-28  Eric Seidel  <eric@webkit.org>
2819
2820         Reviewed by Adam Barth.
2821
2822         webkit-patch upload calls changed_files more often than it should
2823         https://bugs.webkit.org/show_bug.cgi?id=48567
2824
2825         Passing changed_files around everywhere isn't a very elegant solution
2826         but it's the one we have for the moment.  I think keeping an explicit
2827         cache on Checkout (or making StepState() a real class) is a better
2828         long-term option.
2829
2830         Previously bug_id_for_this_commit was calling changed_files and the
2831         result was never getting cached on the state.  Now we're explicitly
2832         caching the result on the state and passing that to the bug_id_for_this_commit call.
2833
2834         I looked into building unit tests for this.  Doing so would require
2835         using a real Checkout object with a MockSCM and overriding the appropriate
2836         calls on SCM to count how often we're stating the file system.
2837         That's a useful set of tests to build for a separate change.
2838
2839         * Scripts/webkitpy/common/checkout/api.py:
2840         * Scripts/webkitpy/tool/commands/download.py:
2841         * Scripts/webkitpy/tool/commands/upload.py:
2842         * Scripts/webkitpy/tool/mocktool.py:
2843
2844 2010-10-28  Eric Seidel  <eric@webkit.org>
2845
2846         Reviewed by Adam Barth.
2847
2848         Make suggest-reviewers slightly faster
2849         https://bugs.webkit.org/show_bug.cgi?id=48562
2850
2851         Add @memoized to one more common call.
2852
2853         * Scripts/webkitpy/common/checkout/api.py:
2854
2855 2010-10-28  Eric Seidel  <eric@webkit.org>
2856
2857         Reviewed by Adam Barth.
2858
2859         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is still very slow
2860         https://bugs.webkit.org/show_bug.cgi?id=48536
2861
2862         This patch makes it about 40 seconds faster, but it still
2863         takes 1:40.  This will require more refinement.
2864
2865         The suggested reviewers list appears to be the same.
2866
2867         I think the next step may be to have it stop the search after
2868         5 reviewers are found.  We never want to suggest 30 people.
2869
2870         * Scripts/webkitpy/common/checkout/scm.py:
2871          - Using --remove-empty to theoretically stop lookups past
2872            when a file is removed.  I'm not entirely clear that the option
2873            does what it says it does.  Example:
2874            git log --pretty=format:%H -5 --remove-empty -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
2875            returns only one commit
2876            vs.
2877            git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
2878            which returns 5.  I was not aware that wx files were ever removed from the repository?
2879
2880 2010-10-28  Kinuko Yasuda  <kinuko@chromium.org>
2881
2882         Reviewed by David Levin.
2883
2884         [Chromium] Support FileSystem in chromium DRT
2885         https://bugs.webkit.org/show_bug.cgi?id=47643
2886
2887         * DumpRenderTree/chromium/WebViewHost.cpp:
2888         (WebViewHost::openFileSystem): Added.
2889         * DumpRenderTree/chromium/WebViewHost.h:
2890         (WebViewHost::openFileSystem): Added.
2891
2892 2010-10-28  Sam Weinig  <sam@webkit.org>
2893
2894         Reviewed by Anders Carlsson.
2895
2896         WKURLRefs should be allowed to be null
2897         <rdar://problem/8575621>
2898         https://bugs.webkit.org/show_bug.cgi?id=48535
2899
2900         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
2901         * TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp:
2902         (TestWebKitAPI::didStartProvisionalLoadForFrame):
2903         (TestWebKitAPI::didCommitLoadForFrame):
2904         (TestWebKitAPI::didFinishLoadForFrame):
2905         Test that URLs are null pointers when unset.
2906
2907 2010-10-28  Chris Fleizach  <cfleizach@apple.com>
2908
2909         Reviewed by Adele Peterson.
2910
2911         AX: multi select group option does not handle setting of AXSelectedChildren correctly
2912         https://bugs.webkit.org/show_bug.cgi?id=48464
2913
2914         Add support for querying information about selected children to DRT, including:
2915            selectedChildAtIndex
2916            selectedChildrenCount
2917            setSelectedChild
2918
2919         * DumpRenderTree/AccessibilityUIElement.cpp:
2920         (selectedChildAtIndexCallback):
2921         (setSelectedChildCallback):
2922         (selectedChildrenCountCallback):
2923         (AccessibilityUIElement::setSelectedChild):
2924         (AccessibilityUIElement::selectedChildrenCount):
2925         (AccessibilityUIElement::selectedChildAtIndex):
2926         (AccessibilityUIElement::getJSClass):
2927         * DumpRenderTree/AccessibilityUIElement.h:
2928         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
2929         (AccessibilityUIElement::selectedChildAtIndex):
2930         (AccessibilityUIElement::selectedChildrenCount):
2931         (AccessibilityUIElement::setSelectedChild):
2932
2933 2010-10-27  Sam Weinig  <sam@webkit.org>
2934
2935         Reviewed by Anders Carlsson.
2936
2937         Add WebKit2 API for window feature getter/setters
2938         <rdar://problem/8590373>
2939         https://bugs.webkit.org/show_bug.cgi?id=48496
2940
2941         * MiniBrowser/mac/BrowserWindowController.m:
2942         (-[BrowserWindowController awakeFromNib]):
2943         * MiniBrowser/win/BrowserView.cpp:
2944         (BrowserView::create):
2945         * WebKitTestRunner/TestController.cpp:
2946         (WTR::createOtherPage):
2947         (WTR::TestController::initialize):
2948
2949 2010-10-28  Søren Gjesse  <sgjesse@chromium.org>
2950
2951         Reviewed by Tony Chang.
2952
2953         Added support for the DumpRenderTree flags --multiple-loads and --js-flags to the Python test runner.
2954         https://bugs.webkit.org/show_bug.cgi?id=48236
2955
2956         * Scripts/webkitpy/layout_tests/port/chromium.py:
2957         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2958
2959 2010-10-28  Sergio Villar Senin  <svillar@igalia.com>
2960
2961         Reviewed by Martin Robinson.
2962
2963         [GTK] WebKitWebFrame's load-status is not properly notified to the tests
2964         https://bugs.webkit.org/show_bug.cgi?id=48048
2965
2966         DRT now listens to WebKitWebFrame load-status signals for each
2967         created frame instead of the load-status signal of the
2968         WebKitWebView that only notifies about changes in the main frame.
2969
2970         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2971         (webViewLoadFinished):
2972         (webFrameLoadStatusNotified):
2973         (frameCreatedCallback):
2974         (createWebView):
2975         (main):
2976
2977 2010-10-28  Tony Chang  <tony@chromium.org>
2978
2979         Reviewed by Kent Tamura.
2980
2981         [chromium] fix textInputController.{selectedRange,markedRange}
2982         https://bugs.webkit.org/show_bug.cgi?id=48487
2983
2984         * DumpRenderTree/chromium/TextInputController.cpp:
2985         (TextInputController::markedRange): Return arrays of ints, rather than a string
2986         (TextInputController::selectedRange): Ditto.
2987
2988 2010-10-27  Ojan Vafai  <ojan@chromium.org>
2989
2990         Reviewed by Tony Chang.
2991
2992         [chromium] add a master-name flag to new-run-webkit-tests
2993         https://bugs.webkit.org/show_bug.cgi?id=48488
2994
2995         The test results server now allows adding a master name to
2996         the uploaded files. This lets us distinguish bots that have
2997         the same name, but are on different masters.
2998
2999         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3000
3001 2010-10-28  Adam Roben  <aroben@apple.com>
3002
3003         Don't append a newline to the test output if the frame has no document
3004         element in WebKitTestRunner
3005
3006         Fixes <http://webkit.org/b/48526> Extra trailing newline when running
3007         plugins/document-open.html in WebKitTestRunner
3008
3009         Reviewed by Anders Carlsson.
3010
3011         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3012         (WTR::toJS): Added. Turns a UTF-8 C string into a JSStringRef.
3013         (WTR::hasDocumentElement): Added. Uses the JSC API to figure out
3014         whether the frame has a document element.
3015         (WTR::dumpFrameText): Match DRT by bailing (rather than appending an
3016         empty string and a newline) if the frame has no document element.
3017
3018 2010-10-28  Adam Roben  <aroben@apple.com>
3019
3020         Skip npn-invalidate-rect-invalidates-window.html on headless XP
3021         machines
3022
3023         TestNetscapePlugIn never receives a WM_PAINT message on headless XP
3024         machines, so this test times out. Fixing the test is covered by
3025         <http://webkit.org/b/48333>.
3026
3027         * Scripts/old-run-webkit-tests: Skip
3028         npn-invalidate-rect-invalidates-window.html on Windows if accelerated
3029         compositing support is disabled, which likely means we're on a headless
3030         XP machine.
3031
3032 2010-10-28  Kimmo Kinnunen  <kimmok@iki.fi>
3033
3034         Adding myself as a committer.
3035
3036         * Scripts/webkitpy/common/config/committers.py:
3037
3038 2010-10-27  Eric Seidel  <eric@webkit.org>
3039
3040         Reviewed by Adam Barth.
3041
3042         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is very slow
3043         https://bugs.webkit.org/show_bug.cgi?id=48500
3044
3045         This doesn't fix the problem, but it makes things slightly better.
3046         Each git svn find-rev call takes about .25 seconds on my desktop.
3047         This patch uses a new memoized class to avoid those calls when possible.
3048
3049         The real slowness is still git log on some files, like:
3050         git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
3051         I'm not yet sure how to make the pathological git logs better.
3052
3053         * Scripts/webkitpy/common/checkout/scm.py:
3054         * Scripts/webkitpy/common/memoized.py: Added.
3055         * Scripts/webkitpy/common/memoized_unittest.py: Added.
3056
3057 2010-10-27  Eric Seidel  <eric@webkit.org>
3058
3059         Reviewed by Ojan Vafai.
3060
3061         EWS bots should not use --quiet when running build-webkit
3062         https://bugs.webkit.org/show_bug.cgi?id=48482
3063
3064         --quiet is only correct when the sub-process does the error reporting.
3065         In the case of _can_build() the parent process is reporting the error.
3066         We'd like the full build log at queues.webkit.org so someone can look
3067         at the log and understand why the EWS is failing to build trunk.
3068
3069         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
3070         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
3071         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
3072         * Scripts/webkitpy/tool/mocktool.py:
3073
3074 2010-10-27  Brian Weinstein  <bweinstein@apple.com>
3075
3076         More Windows build fixage. Rename a variable that was named string.
3077
3078         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3079         (WTR::hasPrefix):
3080
3081 2010-10-27  Ojan Vafai  <ojan@chromium.org>
3082
3083         Reviewed by Dimitri Glazkov.
3084
3085         [chromium] Make the test results server store which master the bot is on
3086         https://bugs.webkit.org/show_bug.cgi?id=48478
3087
3088         The chromium bots recently changed so that there are multiple slaves with
3089         the same name on different masters. Up till now, the test results server
3090         assumed slave names were unique. Adds a master field to the file in order
3091         to distinguish.
3092
3093         Also, for files that currently lack a master or testtype, set them appropriately.
3094
3095         * TestResultServer/handlers/testfilehandler.py:
3096         * TestResultServer/index.yaml:
3097         * TestResultServer/model/jsonresults.py:
3098         * TestResultServer/model/testfile.py:
3099         * TestResultServer/templates/showfilelist.html:
3100         * TestResultServer/templates/uploadform.html:
3101
3102 2010-10-26  Darin Adler  <darin@apple.com>
3103
3104         Reviewed by Sam Weinig.
3105
3106         WebKitTestRunner needs to support layoutTestController.dumpBackForwardList
3107         https://bugs.webkit.org/show_bug.cgi?id=42322
3108         rdar://problem/8193631
3109
3110         WebKitTestRunner needs to support layoutTestController.clearBackForwardList
3111         https://bugs.webkit.org/show_bug.cgi?id=42333
3112         rdar://problem/8193643
3113
3114         * WebKitTestRunner/Configurations/InjectedBundle.xcconfig:
3115         Renamed the product to WebKitTestRunnerInjectedBundle to avoid
3116         name conflicts in the build directory.
3117
3118         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
3119         Added dumpBackForwardList and clearBackForwardList.
3120
3121         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
3122         (WTR::InjectedBundle::InjectedBundle): Removed initialization of m_mainPage.
3123         (WTR::InjectedBundle::didCreatePage): Removed unneeded underscore.
3124         (WTR::InjectedBundle::willDestroyPage): Ditto.
3125         (WTR::InjectedBundle::didReceiveMessage): Ditto.
3126         (WTR::InjectedBundle::initialize): Ditto.
3127         (WTR::InjectedBundle::didCreatePage): Changed code to use m_pages
3128         instead of m_mainPage and m_otherPages.
3129         (WTR::InjectedBundle::willDestroyPage): Ditto.
3130         (WTR::InjectedBundle::page): Ditto.
3131         (WTR::InjectedBundle::beginTesting): Ditto.
3132         (WTR::InjectedBundle::done): Ditto.
3133         (WTR::InjectedBundle::closeOtherPages): Ditto.
3134         (WTR::InjectedBundle::dumpBackForwardListsForAllPages): Added.
3135
3136         * WebKitTestRunner/InjectedBundle/InjectedBundle.h: Changed
3137         page function to longer be inline, and pageCount function
3138         to use m_pages. Added dumpBackForwardListsForAllPages, and
3139         removed some underscores. Replaced m_mainPage and m_otherPageas
3140         with m_pages.
3141
3142         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3143         (WTR::adoptWK): Added.
3144         (WTR::hasPrefix): Added.
3145         (WTR::InjectedBundlePage::reset): Added code to set up
3146         m_previousTestBackForwardListItem.
3147         (WTR::InjectedBundlePage::dump): Added code to call
3148         dumpBackForwardListsForAllPages.
3149         (WTR::compareByTargetName): Added.
3150         (WTR::dumpBackForwardListItem): Added.
3151         (WTR::InjectedBundlePage::dumpBackForwardList): Added.
3152
3153         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h: Added
3154         dumpBackForwardList and m_previousTestBackForwardListItem.
3155
3156         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
3157         (WTR::LayoutTestController::LayoutTestController): Initialize
3158         m_shouldDumpBackForwardListsForAllWindows to false.
3159         (WTR::LayoutTestController::clearBackForwardList): Added.
3160
3161         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
3162         Added dumpBackForwardList, clearBackForwardList,
3163         shouldDumpBackForwardListsForAllWindows, and
3164         m_shouldDumpBackForwardListsForAllWindows.
3165
3166         * WebKitTestRunner/TestController.cpp:
3167         (WTR::TestController::initialize): Set up
3168         didReceiveSynchronousMessageFromInjectedBundle.
3169         (WTR::TestController::didReceiveSynchronousMessageFromInjectedBundle):
3170         Added.
3171         * WebKitTestRunner/TestController.h: Ditto.
3172
3173         * WebKitTestRunner/TestInvocation.cpp:
3174         (WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):
3175         Added.
3176         * WebKitTestRunner/TestInvocation.h: Ditto.
3177
3178         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
3179         Renamed bundle to WebKitTestRunnerInjectedBundle (see above).
3180         * WebKitTestRunner/mac/TestControllerMac.mm:
3181         (WTR::TestController::initializeInjectedBundlePath): Ditto.
3182
3183 2010-10-27  Chris Rogers  <crogers@google.com>
3184
3185         Reviewed by Chris Marrin.
3186
3187         Add ENABLE_WEB_AUDIO feature enable flag (initially disabled) to build-webkit
3188         https://bugs.webkit.org/show_bug.cgi?id=48279
3189
3190         * Scripts/build-webkit:
3191
3192 2010-10-27  Eric Seidel  <eric@webkit.org>
3193
3194         Unreviewed.
3195
3196         build-webkit should collect Visual Studio Express logs and display them
3197         https://bugs.webkit.org/show_bug.cgi?id=39199
3198
3199         It turns out my previous patch wasn't actually working on the win-ews
3200         machine.  So I've fixed my mistakes from before.
3201
3202         * Scripts/build-webkit:
3203          - Windows VSE builds change the CWD while building.  Why?  Who knows.
3204         * Scripts/print-vse-failure-logs:
3205          - windows VSE builds don't use Debug/Release as I expected, so I've
3206            moved off of --configuration to --top-level and added the /obj
3207            optimization while I was there.
3208
3209 2010-10-27  Dimitri Glazkov  <dglazkov@chromium.org>
3210
3211         Unreviewed, rolling out r70674.
3212         http://trac.webkit.org/changeset/70674
3213         https://bugs.webkit.org/show_bug.cgi?id=48053
3214
3215         Broke Chromium Windows build.
3216
3217         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3218         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3219
3220 2010-10-25  Tony Chang  <tony@chromium.org>
3221
3222         Reviewed by Anders Carlsson.
3223
3224         compile TestNetscapePlugIn on chromium linux
3225         https://bugs.webkit.org/show_bug.cgi?id=48274
3226
3227         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
3228         (NP_Initialize): On Linux, plugin funcs are set in initialize.
3229         (NPP_New): Mark the plugin as windowless.
3230         (NPP_GetValue): Handle mime type values.
3231         (NP_GetMIMEDescription):
3232         (NP_GetValue):
3233         * DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npapi.h: define TRUE and FALSE, which are in webkit's npapi.h.
3234         * DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npfunctions.h: Pull in npapi.h to get TRUE/FALSE (matches webkit's npfunctions.h)
3235
3236 2010-10-27  Kenneth Rohde Christiansen  <kenneth@webkit.org>
3237
3238         Reviewed by Anders Carlsson.
3239
3240         Remove contentSizeChanged callbacks as it is no longer
3241         part of the public UIClient.
3242
3243         Make WKPageContentsSizeChangedCallback be a private API
3244         https://bugs.webkit.org/show_bug.cgi?id=48409
3245
3246         * MiniBrowser/mac/BrowserWindowController.m:
3247         (-[BrowserWindowController awakeFromNib]):
3248         * MiniBrowser/win/BrowserView.cpp:
3249         (BrowserView::create):
3250         * WebKitTestRunner/TestController.cpp:
3251         (WTR::createOtherPage):
3252         (WTR::TestController::initialize):
3253
3254 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3255
3256         Reviewed by Ojan Vafai.
3257
3258         Make http locking default in NRWT.
3259         https://bugs.webkit.org/show_bug.cgi?id=48053
3260
3261         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3262         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3263
3264 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3265
3266         Reviewed by Tony Chang.
3267
3268         [NRWT] Fix http lock on Windows platform
3269         https://bugs.webkit.org/show_bug.cgi?id=48321
3270
3271         * Scripts/webkitpy/layout_tests/port/http_lock.py:
3272
3273 2010-10-27  Satish Sampath  <satish@chromium.org>
3274
3275         Unreviewed, rolling out r70665.
3276         http://trac.webkit.org/changeset/70665
3277         https://bugs.webkit.org/show_bug.cgi?id=47089
3278
3279         Need to address Alexey's review comments.
3280
3281         * DumpRenderTree/LayoutTestController.cpp:
3282         (setMockSpeechInputResultCallback):
3283         * DumpRenderTree/LayoutTestController.h:
3284         * DumpRenderTree/chromium/LayoutTestController.cpp:
3285         (LayoutTestController::setMockSpeechInputResult):
3286         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3287         (LayoutTestController::setMockSpeechInputResult):
3288         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3289         (LayoutTestController::setMockSpeechInputResult):
3290         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3291         (LayoutTestController::setMockSpeechInputResult):
3292         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3293         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3294         (LayoutTestController::setMockSpeechInputResult):
3295         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3296         (LayoutTestController::setMockSpeechInputResult):
3297
3298 2010-10-27  Leandro Gracia Gil  <leandrogracia@google.com>
3299
3300         Reviewed by Jeremy Orlow.
3301
3302         Added a second parameter to setMockSpeechInputResult for
3303         the language used in speech input.
3304         https://bugs.webkit.org/show_bug.cgi?id=47089
3305
3306         * DumpRenderTree/LayoutTestController.cpp:
3307         (setMockSpeechInputResultCallback):
3308         * DumpRenderTree/LayoutTestController.h:
3309         * DumpRenderTree/chromium/LayoutTestController.cpp:
3310         (LayoutTestController::setMockSpeechInputResult):
3311         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3312         (LayoutTestController::setMockSpeechInputResult):
3313         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3314         (LayoutTestController::setMockSpeechInputResult):
3315         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3316         (LayoutTestController::setMockSpeechInputResult):
3317         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3318         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3319         (LayoutTestController::setMockSpeechInputResult):
3320         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3321         (LayoutTestController::setMockSpeechInputResult):
3322         (LayoutTestController::pageSizeAndMarginsInPixels):
3323
3324 2010-10-27  Adam Roben  <aroben@apple.com>
3325
3326         Reset TestNetscapePlugIn's NPP_GetValue pointer when
3327         NullNPPGetValuePointer finishes running
3328
3329         Fixes <http://webkit.org/b/48435> REGRESSION (r70655): Many plugins
3330         tests are failing on Qt
3331
3332         Reviewed by Anders Carlsson.
3333
3334         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
3335         (PluginTest::NPP_Destroy):
3336         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
3337         Added a do-nothing NPP_Destroy implementation.
3338
3339         * DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp:
3340         (NullNPPGetValuePointer::NullNPPGetValuePointer): Save the original
3341         NPP_GetValue pointer in m_originalNPPGetValuePointer so we can restore
3342         it later.
3343         (NullNPPGetValuePointer::NPP_Destroy): Added. Restores the original
3344         NPP_GetValue pointer so it can be used in other tests.
3345
3346         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
3347         (NPP_Destroy):
3348         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
3349         (webkit_test_plugin_destroy_instance):
3350         Call through to the PluginTest when NPP_Destroy is called.
3351
3352 2010-10-26  Brian Weinstein  <bweinstein@apple.com>
3353
3354         Reviewed by Adam Roben.
3355
3356         WebKit2 shouldn't try to send an empty user agent
3357         https://bugs.webkit.org/show_bug.cgi?id=48397
3358         
3359         Add a test that when we set our custom user agent to an empty string, we don't send
3360         an empty user agent.
3361
3362         * TestWebKitAPI/Tests/WebKit2/PreventEmptyUserAgent.cpp: Added.
3363         (TestWebKitAPI::didRunJavaScript): Make sure that the result of navigator.userAgent isn't empty.
3364         (TestWebKitAPI::TEST): Set our custom user agent to the empty string, and run navigator.userAgent.
3365         
3366         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Add the new file.
3367         * TestWebKitAPI/win/TestWebKitAPI.vcproj: Ditto.
3368
3369 2010-10-27  Adam Roben  <aroben@apple.com>
3370
3371         Check in file I forgot in r70653
3372
3373         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
3374         (NP_Initialize): Save the NPPluginFuncs struct the browser passed to us
3375         so we can be naughty and modify it later.
3376
3377 2010-10-27  Adam Roben  <aroben@apple.com>
3378
3379         Test that WebKit doesn't crash if the plugin passes 0 for its
3380         NPP_GetValue pointer
3381
3382         Test for <http://webkit.org/b/48433> Crash in
3383         NetscapePlugin::shouldLoadSrcURL when using Shockwave Director 10.3 in
3384         WebKit2 on Windows
3385
3386         Reviewed by Eric Carlson.
3387
3388         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3389         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
3390         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
3391         * GNUmakefile.am:
3392         Added NullNPPGetValuePointer.cpp.
3393
3394         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
3395         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
3396         Set up a place to store the NPPluginFuncs struct the browser passed to us.
3397
3398         * DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp: Added.
3399         (NullNPPGetValuePointer::NullNPPGetValuePointer): Null out the
3400         NPP_GetValue pointer we passed to the browser to simulate a plugin that
3401         doesn't implement NPP_GetValue.
3402         (NullNPPGetValuePointer::NPP_GetValue): Print an error message. If this
3403         function is called, it means that WebKit has changed in a way that
3404         makes this test invalid.
3405
3406         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
3407         (NP_GetEntryPoints): Save the NPPluginFuncs struct the browser passed
3408         to us so we can be naughty and modify it later.
3409
3410 2010-10-27  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
3411
3412         Reviewed by Andreas Kling.
3413
3414         Remove references to ancient QGVLauncher and QtLauncher
3415         https://bugs.webkit.org/show_bug.cgi?id=48430
3416
3417         QtTestBrowser substitutes both and has checks enabled.
3418
3419         * Scripts/webkitpy/style/checker.py: remove references
3420         * Scripts/webkitpy/style/checker_unittest.py: ditto
3421
3422 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3423
3424         Reviewed by Ojan Vafai.
3425
3426         [NRWT] Don't use image hash when it's no need in single test mode.
3427         https://bugs.webkit.org/show_bug.cgi?id=48326
3428
3429         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3430
3431 2010-10-27  David Kilzer  <ddkilzer@apple.com>
3432
3433         Fix leak of CFMutableDictionaryRef in createXMLStringFromWebArchiveData()
3434
3435         Reviewed by Adam Roben.
3436
3437         Follow-up fix for: <https://bugs.webkit.org/show_bug.cgi?id=48278>
3438
3439         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
3440         (createXMLStringFromWebArchiveData): Use RetainPtr<> to fix a
3441         leak introduced in r70613.
3442
3443 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3444
3445         Reviewed by Csaba Osztrogonác.
3446
3447         [NRWT] Add platform specific baseline search paths for Qt port
3448         https://bugs.webkit.org/show_bug.cgi?id=48428
3449
3450         * Scripts/webkitpy/layout_tests/port/qt.py:
3451
3452 2010-10-27  Adam Roben  <aroben@apple.com>
3453
3454         Catch exceptions when checking if we're inside a git working directory
3455
3456         Fixes <http://webkit.org/b/48420> REGRESSION (r70562): test-webkitpy
3457         fails on systems without git installed
3458
3459         Reviewed by Anders Carlsson.
3460
3461         * Scripts/webkitpy/common/net/credentials.py:
3462         (Credentials._credentials_from_git): Put the call to
3463         Git.in_working_directory inside the try/except since it, too, attempts
3464         to execute git and thus will throw on systems that don't have git
3465         installed.
3466
3467 2010-10-27  Nikolas Zimmermann  <nzimmermann@rim.com>
3468
3469         Reviewed by David Kilzer.
3470
3471         Convert DumpRenderTree webarchive code to CoreFoundation
3472         https://bugs.webkit.org/show_bug.cgi?id=48278
3473
3474         CFPropertyListCreateWithData and CFPropertyListCreateData are only available in 10.6+.
3475         Replace CFPropertyListCreateWithData by a combination of CFReadStreamCreateWithBytesNoCopy and CFPropertyListCreateFromStream.
3476         Replace CFPropertyListCreateData by CFPropertyListCreateXMLData.
3477
3478         These changes are wrapped in BUILDING_ON_LEOPARD, as the methods are deprecated on 10.6+.
3479
3480         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
3481         (createXMLStringFromWebArchiveData):
3482
3483 2010-10-26  Antonio Gomes  <agomes@rim.com>
3484
3485         Reviewed by Martin Robinson.
3486
3487         Remove the absolute path used to include DumpRenderTreeSupportGtk.h from LayoutTestController.h
3488
3489         It turns out this is rather unneeded since WebKitTools/GNUMakefile.am has WebKit/gtk/ in its include
3490         path, and then we can just do #include "WebCoreSupport/DumpRenderTreeSupportGtk.h"
3491
3492         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3493
3494 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3495
3496         Rename WebArchiveDumpSupport.mm to WebArchiveDumpSupport.cpp
3497
3498         Reviewed by Adam Roben.
3499
3500         Part 4 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3501
3502         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3503         Updated for file moves and renames.
3504         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.mm.
3505         * DumpRenderTree/cf/WebArchiveDumpSupport.h: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.h.
3506
3507 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3508
3509         Convert WebArchiveDumpSupport.mm from NS objects to CF types
3510
3511         Reviewed by Adam Roben.
3512
3513         Part 3 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3514
3515         In order to share WebArchive code between the Mac and Windows
3516         ports, the code in WebArchiveDumpSupport.mm was converted from
3517         Cocoa to C++ using CoreFoundation (CF) types.
3518
3519         * DumpRenderTree/mac/DumpRenderTree.mm:
3520         (dump): Renamed serializeWebArchiveToXML() to
3521         createXMLStringFromWebArchiveData() and added HardAutorelease()
3522         to prevent leaks.
3523         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
3524         (createXMLStringFromWebArchiveData): Renamed from
3525         serializeWebArchiveToXML().  Changed to use CF types.
3526         (createCFURLResponseFromResponseData): Renamed from
3527         unarchiveNSURLResponseFromResponseData().  Changed to use CF
3528         types for its parameter and return type.
3529         * DumpRenderTree/mac/WebArchiveDumpSupport.mm: Replaced use of
3530         NS objects with CF types.  It will be renamed to *.cpp in a
3531         future commit.
3532         (convertMIMEType): Changed to use case-insensitive string
3533         comparisons.
3534         (convertWebResourceDataToString):
3535         (normalizeHTTPResponseHeaderFields):
3536         (normalizeWebResourceURL):
3537         (convertWebResourceResponseToDictionary):
3538         (compareResourceURLs):
3539         (createXMLStringFromWebArchiveData):
3540         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
3541         (createCFURLResponseFromResponseData): Renamed from
3542         unarchiveNSURLResponseFromResponseData().  Changed to use CF
3543         types for its parameter and return type.
3544
3545 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3546
3547         Extract use of NSKeyedUnarchiver from WebArchiveDumpSupport.mm
3548
3549         Reviewed by Adam Roben.
3550
3551         Part 2 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3552
3553         There is no equivalent to NSKeyedUnarchiver in CoreFoundation,
3554         so extract it into a platform-specific source file.
3555
3556         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
3557         (unarchiveNSURLResponseFromResponseData): Added declaration.
3558         * DumpRenderTree/mac/WebArchiveDumpSupport.mm:
3559         (convertWebResourceResponseToDictionary): Extracted code to
3560         unarchiveNSURLResponseFromResponseData() in
3561         WebArchiveDumpSupportMac.mm.  Updated logic to return early if
3562         nil is returned from unarchiveNSURLResponseFromResponseData().
3563         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
3564         (unarchiveNSURLResponseFromResponseData): Added.  Extracted
3565         code from convertWebResourceResponseToDictionary() in
3566         WebArchiveDumpSupport.mm.
3567
3568 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3569
3570         Extract call to -[WebHTMLRepresentation supportedNonImageMIMETypes] from WebArchiveDumpSupport.mm
3571
3572         Reviewed by Adam Roben.
3573
3574         Part 1 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3575
3576         The call to -[WebHTMLRepresentation supportedNonImageMIMETypes]
3577         is not cross-platform between Mac and Windows, so extract it
3578         into a platform-specific source file.
3579
3580         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Added
3581         WebArchiveDumpSupportMac.mm to the project.
3582         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
3583         (supportedNonImageMIMETypes): Added declaration.
3584         * DumpRenderTree/mac/WebArchiveDumpSupport.mm:
3585         (convertWebResourceDataToString): Replaced call to
3586         -[WebHTMLRepresentation supportedNonImageMIMETypes] with
3587         supportedNonImageMIMETypes().
3588         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm: Added.
3589         (supportedNonImageMIMETypes): Added.  Extracted from
3590         WebArchiveDumpSupport.mm.
3591
3592 2010-10-26  Antonio Gomes  <agomes@rim.com>
3593
3594         Reviewed by Martin Robinson.
3595
3596         [GTK] Implement DumpRenderTreeSupportGtk (similarly to DumpRenderTreeSupportQt idea)
3597         https://bugs.webkit.org/show_bug.cgi?id=48199
3598
3599         Implements support to WebKitTabToLinksPreferenceKey through LayoutTestController::overridePreference.
3600         The corresponding DumpRenderTreeSupportGtk method is called in the DRT context only.
3601
3602         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3603         (resetDefaultsToConsistentValues):
3604         (createWebView):
3605         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3606         (LayoutTestController::overridePreference):
3607
3608 2010-10-26  Ariya Hidayat  <ariya@sencha.com>
3609
3610         Unreviewed, change the order of my emails for bugzilla autocompletion.
3611
3612         * Scripts/webkitpy/common/config/committers.py:
3613
3614 2010-10-26  Eric Seidel  <eric@webkit.org>
3615
3616         Reviewed by David Kilzer.
3617
3618         build-webkit should collect Visual Studio Express logs and display them
3619         https://bugs.webkit.org/show_bug.cgi?id=39199
3620
3621         * Scripts/build-webkit:
3622         * Scripts/print-vse-failure-logs: Added.
3623         * Scripts/webkitdirs.pm:
3624
3625 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3626
3627         <http://webkit.org/b/48224> build-webkit: add support for --meter-tag switch
3628
3629         Reviewed by Csaba Osztrogonác.
3630
3631         * Scripts/build-webkit: Added support for --meter-tag switch.
3632         It should have been added with r60820.
3633
3634 2010-10-26  Ademar de Souza Reis Jr.  <ademar.reis@openbossa.org>
3635
3636         Reviewed by Dumitru Daniliuc.
3637
3638         check-webkit-style fails on operator+=, operator-=, ... methods
3639         https://bugs.webkit.org/show_bug.cgi?id=48258
3640
3641         * Scripts/webkitpy/style/checkers/cpp.py: Added exceptions
3642         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added unit-tests
3643
3644 2010-10-26  Ojan Vafai  <ojan@chromium.org>
3645
3646         Reviewed by Tony Chang.
3647
3648         remove DEFER support from new-run-webkit-tests
3649         https://bugs.webkit.org/show_bug.cgi?id=48387
3650
3651         DEFER was needed when we were trying to ship Chrome beta.
3652         Now it's just extra complication.
3653
3654         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
3655         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
3656  &