2010-11-22 Dirk Pranke <dpranke@chromium.org>
[WebKit.git] / WebKitTools / ChangeLog
1 2010-11-22  Dirk Pranke  <dpranke@chromium.org>
2
3         Reviewed by Tony Chang.
4
5         Fix some spacing issues with log messages with
6         new-run-webkit-tests --verbose.
7
8         https://bugs.webkit.org/show_bug.cgi?id=49936
9
10         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
11
12 2010-11-22  Ojan Vafai  <ojan@chromium.org>
13
14         Reviewed by Adam Barth.
15
16         update scm to work with different remote branches
17         https://bugs.webkit.org/show_bug.cgi?id=49949
18
19         This works if there are multiple svn tracking branches
20         and/or if the tracking branch is refs/remots/origin/master
21         instead of refs/remotes/trunk which is the direction we're trying
22         to head since that makes fetches faster.
23
24         * Scripts/webkitpy/common/checkout/scm.py:
25         * Scripts/webkitpy/common/checkout/scm_unittest.py:
26
27 2010-11-22  Hayato Ito  <hayato@chromium.org>
28
29         Reviewed by Shinichiro Hamaji.
30
31         Ignore reference files which will be used by reftests when collecting
32         test cases.
33         https://bugs.webkit.org/show_bug.cgi?id=49835
34
35         * Scripts/webkitpy/layout_tests/port/test_files.py:
36         * Scripts/webkitpy/layout_tests/port/test_files_unittest.py:
37
38 2010-11-22  Adam Roben  <aroben@apple.com>
39
40         Use paths relative to $WebKitVSPropsRedirectionDir to access shared .vsprops files
41
42         Apple's Windows build allows placing header files and import libraries for WebKit's
43         dependencies (CoreGraphics, CFNetwork, SQLite, etc.) outside the source tree via the
44         $WebKitLibrariesDir environment variable. This is both required for production builds and
45         convenient for Apple-internal developer builds. Apple's production builds also require that
46         WebKit's shared .vsprops files be accessed relative to $WebKitLibrariesDir. In production
47         builds, the files are copied into that directory tree by the
48         WebKitLibraries/win/tools/WinTools.make file. In Apple-internal developer builds, the
49         copying is done by
50         JavaScriptCore/JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.make.
51
52         This .vsprops copying is problematic in one very important case: when a developer updates
53         their source tree and then tries to build. Visual Studio only reads .vsprops files when a
54         project is first loaded. So, when Visual Studio is first opened after the .vsprops files are
55         updated, it reads in the old files that were already residing in $WebKitLibrariesDir. When a
56         build is started, JavaScriptCoreGenerated.make copies the new .vsprops files into
57         $WebKitLibrariesDir, but Visual Studio will not pick up the changes.  The rest of the build
58         will proceed with out-of-date .vsprops files, which will likely result in a build failure.
59
60         To fix this, we now use normal relative paths to access the .vsprops files in the source
61         tree rather than in $WebKitLibrariesDir, but prefix those paths with a new environment
62         variable, $WebKitVSPropsRedirectionDir. In developer builds, this environment variable is
63         unset, so the normal relative paths are used to read the .vsprops files out of the source
64         tree directly. In production builds, this environment variable is set to a fake directory
65         that will cause the .vsprops files in $WebKitLibrariesDir to be found when the relative path
66         is resolved.
67         
68         For example, JavaScriptCore.vcproj uses this path for FeatureDefines.vsprops:
69
70         $(WebKitVSPropsRedirectionDir)..\..\..\WebKitLibraries\win\tools\vsprops\FeatureDefines.vsprops
71
72         In developer builds, where $WebKitVSPropsRedirectionDir is unset, this will point to the
73         files in WebKitLibraries\win\tools\vsprops in the source tree. In production builds,
74         JavaScriptCore.make sets $WebKitVSPropsRedirectionDir to
75         "$(SRCROOT)\AppleInternal\tools\vsprops\OpenSource\1\2\3\", so the full path for
76         FeatureDefines.vsprops becomes:
77
78         $(SRCROOT)\AppleInternal\tools\vsprops\OpenSource\1\2\3\..\..\..\WebKitLibraries\win\tools\vsprops\FeatureDefines.vsprops
79
80         which resolves to:
81
82         $(SRCROOT)\AppleInternal\tools\vsprops\OpenSource\WebKitLibraries\win\tools\vsprops\FeatureDefines.vsprops
83
84         (We rely on the fact that Windows doesn't care whether the directories "1", "2", and "3"
85         actually exist since they are matched by an equal number of ".." path components.)
86
87         Note that Visual Studio still won't pick up changes made to .vsprops files while Visual
88         Studio is open, but that problem hasn't seemed to cause developers many headaches so far.
89
90         Fixes <http://webkit.org/b/49181> Windows build fails mysteriously when .vsprops files are
91         updated
92
93         Reviewed by Dave Hyatt.
94
95         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
96         * DumpRenderTree/win/DumpRenderTree.vcproj:
97         * DumpRenderTree/win/ImageDiff.vcproj:
98         * FindSafari/FindSafari.vcproj:
99         * MiniBrowser/MiniBrowser.vcproj:
100         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
101         * TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj:
102         * WebKitAPITest/WebKitAPITest.vcproj:
103         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
104         * WebKitTestRunner/win/InjectedBundle.vcproj:
105         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
106         * WinLauncher/WinLauncher.vcproj:
107         * record-memory-win/record-memory-win.vcproj:
108         Changed to use paths relative to $WebKitVSPropsRedirectionDir to access shared .vsprops
109         files.
110
111 2010-11-22  Adam Roben  <aroben@apple.com>
112
113         Make it possible to run tests on Windows without Visual Studio or VC++
114         Express installed
115
116         webkitdir.pm::setupCygwinEnv dies when Visual Studio and VC++ Express
117         are not installed. But this function doesn't need to be called when we
118         already have a build available and are just trying to run the tests.
119
120         Fixes <http://webkit.org/b/49932> New Windows 7 bot can't run tests
121         because Visual Studio/VC++ Express aren't installed
122
123         Reviewed by Dave Hyatt.
124
125         * Scripts/webkitdirs.pm:
126         (determineConfigurationForVisualStudio): Don't call setupCygwinEnv, as
127         it is not needed by this function. Also added a FIXME.
128
129         (usingVisualStudioExpress): Call setupCygwinEnv directly rather than
130         relying on determineConfigurationForVisualStudio doing it.
131
132 2010-11-19  Hayato Ito  <hayato@chromium.org>
133
134         Reviewed by Shinichiro Hamaji.
135
136         Simplify Driver.run_test() so that it takes only one parameter, TestInput.
137
138         https://bugs.webkit.org/show_bug.cgi?id=49812
139
140         * Scripts/webkitpy/layout_tests/driver_test.py:
141         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
142         * Scripts/webkitpy/layout_tests/port/base.py:
143         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
144         * Scripts/webkitpy/layout_tests/port/chromium.py:
145         * Scripts/webkitpy/layout_tests/port/dryrun.py:
146         * Scripts/webkitpy/layout_tests/port/test.py:
147         * Scripts/webkitpy/layout_tests/port/webkit.py:
148         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
149         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
150
151 2010-11-22  Hayato Ito  <hayato@chromium.org>
152
153         Reviewed by Tony Chang.
154
155         [NRWT] Retry a few times in reading a png image to avoid a race condition.
156
157         https://bugs.webkit.org/show_bug.cgi?id=49924
158
159         * Scripts/webkitpy/layout_tests/port/chromium.py:
160
161 2010-11-22  João Paulo Rechi Vita  <jprvita@profusion.mobi>
162
163         Reviewed by Kenneth Rohde Christiansen.
164
165         Fix EWebLauncher zoom discretization
166         https://bugs.webkit.org/show_bug.cgi?id=49810
167
168         * EWebLauncher/main.c:
169         (nearest_zoom_level_get):
170         (zoom_level_set):
171         (on_load_finished):
172         (on_key_down):
173
174 2010-11-21  Sam Weinig  <sam@webkit.org>
175
176         Rubber-stamped by Dan Bernstein.
177
178         Fix ~300 WebKit2 tests by enabling developer extras preference.
179
180         * WebKitTestRunner/TestController.cpp:
181         (WTR::TestController::resetStateToConsistentValues):
182
183 2010-11-21  Sam Weinig  <sam@webkit.org>
184
185         Reviewed by Dan Bernstein.
186
187         Textareas should be resizable by default
188         https://bugs.webkit.org/show_bug.cgi?id=49892
189
190         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
191         * TestWebKitAPI/Tests/WebKit2/WKPreferences.cpp: Added.
192         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
193         Add preferences test.
194
195 2010-11-19  Mihai Parparita  <mihaip@chromium.org>
196
197         Reviewed by Tony Chang.
198
199         Rebaseline server: checksum-only differences and prefetching
200         https://bugs.webkit.org/show_bug.cgi?id=49841
201         
202         Add support for image test results where the only difference is in the
203         checksum.
204         
205         Add prefetching of the image results from the next test.
206         
207         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
208         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
209         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
210
211 2010-11-19  Dirk Pranke  <dpranke@chromium.org>
212
213         Reviewed by Tony Chang.
214
215         nrwt multiprocessing - add 'worker number' concept, move stuff to worker thread
216
217         Add the 'worker number' and 'worker name' concepts to the 
218         TestShellThread objects, and move test_types and test_args from
219         the TestRunner to the TestShellThread.
220
221         https://bugs.webkit.org/show_bug.cgi?id=49768
222
223         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
224         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
225
226 2010-11-19  Dirk Pranke  <dpranke@chromium.org>
227
228         Reviewed by Ojan Vafai.
229
230         Do some minor cleanup and bug fixing.
231
232         https://bugs.webkit.org/show_bug.cgi?id=49777
233
234         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
235         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
236
237 2010-11-19  Dirk Pranke  <dpranke@chromium.org>
238
239         Reviewed by Ojan Vafai.
240
241         nrwt - config.build_directory() is busted
242
243         Fixes a typo that was causing us to usually return the top level
244         directory WebKitBuild instead of WebKitBuild/{Debug,Release}. The
245         bug was hidden by test stubs that were too simplistic :(.
246
247         https://bugs.webkit.org/show_bug.cgi?id=49815
248
249         * Scripts/webkitpy/common/system/executive_mock.py:
250         * Scripts/webkitpy/layout_tests/port/config.py:
251         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
252
253 2010-11-19  Steve Falkenburg  <sfalken@apple.com>
254
255         Reviewed by Adam Roben.
256
257         Add Debug_Cairo_CFLite and Release_Cairo_CFLite configurations for all vcproj files
258         https://bugs.webkit.org/show_bug.cgi?id=49819
259
260         * DumpRenderTree/DumpRenderTree.sln:
261         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
262         * DumpRenderTree/win/DumpRenderTree.vcproj:
263         * DumpRenderTree/win/ImageDiff.vcproj:
264         * FindSafari/FindSafari.vcproj:
265         * MiniBrowser/MiniBrowser.vcproj:
266         * TestWebKitAPI/win/TestWebKitAPI.sln:
267         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
268         * TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj:
269         * WebKitAPITest/WebKitAPITest.vcproj:
270         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
271         * WebKitTestRunner/win/InjectedBundle.vcproj:
272         * WebKitTestRunner/win/InjectedBundleGenerated.vcproj:
273         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
274         * WinLauncher/WinLauncher.vcproj:
275         * record-memory-win/record-memory-win.vcproj:
276
277 2010-11-19  Adam Roben  <aroben@apple.com>
278
279         Add some more Windows slaves
280
281         This is probably more than we will be using in the immediate future,
282         but having some extra slave names makes experimenting with different
283         configurations easier.
284
285         Reviewed by Jon Honeycutt.
286
287         * BuildSlaveSupport/build.webkit.org-config/config.json: Added 5 more
288         Windows slaves and distributed them among the three different Test
289         builders.
290
291 2010-11-19  Kinuko Yasuda  <kinuko@chromium.org>
292
293         Reviewed by Ojan Vafai.
294
295         [Chromium] REGRESSION: Record actual test results in the JSON result summary for non-layout tests
296         https://bugs.webkit.org/show_bug.cgi?id=49702
297
298         Re-landing r72357 with a test fix.
299
300         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
301         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
302         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
303
304 2010-11-19  Avi Drissman  <avi@google.com>
305
306         Reviewed by Antonio Gomes.
307
308         Complete support for Unix editing mode
309         https://bugs.webkit.org/show_bug.cgi?id=49757
310
311         * DumpRenderTree/LayoutTestController.cpp:
312         (setEditingBehaviorCallback):
313         * DumpRenderTree/chromium/LayoutTestController.cpp:
314         (LayoutTestController::setEditingBehavior):
315
316 2010-11-19  Adam Roben  <aroben@apple.com>
317
318         Fix run-javascriptcore-tests when there's a space in the path to
319         testapi
320
321         Reviewed by Sam Weinig.
322
323         * Scripts/run-javascriptcore-tests: Use an indirect object when calling
324         system() to force it to interpret spaces in the path correctly.
325
326 2010-11-19  Adam Roben  <aroben@apple.com>
327
328         Update scripts for .exe name changes after r72327
329
330         As of r72327, the "_debug" suffix is only used in Debug_All builds.
331
332         Reviewed by Sam Weinig.
333
334         * Scripts/run-javascriptcore-tests:
335         (testapiPath):
336
337         * Scripts/sunspider-compare-results:
338         (pathToBuiltJSC):
339
340         * Scripts/webkitdirs.pm:
341         (jscPath):
342         (safariPath):
343
344 2010-11-19  Steve Falkenburg  <sfalken@apple.com>
345
346         Reviewed by Darin Adler.
347
348         Normalize Cairo/CFLite project/solution configuration names
349         https://bugs.webkit.org/show_bug.cgi?id=49818
350
351         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
352         * DumpRenderTree/win/DumpRenderTree.vcproj:
353         * DumpRenderTree/win/ImageDiff.vcproj:
354         * MiniBrowser/MiniBrowser.vcproj:
355         * TestWebKitAPI/win/TestWebKitAPI.sln:
356         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
357         * WebKitAPITest/WebKitAPITest.vcproj:
358         * WebKitTestRunner/win/InjectedBundle.vcproj:
359         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
360         * WinLauncher/WinLauncher.vcproj:
361         * Scripts/webkitdirs.pm: Updated cairo build configs.
362
363 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
364
365         Reviewed by Adam Roben.
366
367         Windows vcproj configuration names should be normalized across projects
368         https://bugs.webkit.org/show_bug.cgi?id=49776
369
370         * DumpRenderTree/DumpRenderTree.sln:
371         * FindSafari/FindSafari.vcproj:
372         * TestWebKitAPI/win/TestWebKitAPI.sln:
373         * TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj:
374         * TestWebKitAPI/win/TestWebKitAPIGeneratedCommon.vsprops: Added.
375         * WebKitAPITest/WebKitAPITest.vcproj:
376         * WebKitAPITest/WebKitAPITestCommon.vsprops:
377         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
378         * WebKitTestRunner/WebKitTestRunner.sln:
379         * WebKitTestRunner/win/InjectedBundleGenerated.vcproj:
380         * WebKitTestRunner/win/InjectedBundleGeneratedCommon.vsprops: Added.
381         * record-memory-win/record-memory-win.vcproj:
382
383 2010-11-18  Gavin Barraclough  <barraclough@apple.com>
384
385         Build fix - as per DRT-mac, always enabled developer extras,
386         in win/qt/gtk DRT.
387
388         This is necessary to enable rich exception messages to be generated
389         following https://bugs.webkit.org/show_bug.cgi?id=49708.
390
391         * DumpRenderTree/gtk/DumpRenderTree.cpp:
392         (shouldEnableDeveloperExtras):
393         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
394         (WebCore::shouldEnableDeveloperExtras):
395         * DumpRenderTree/win/DumpRenderTree.cpp:
396         (shouldEnableDeveloperExtras):
397
398 2010-11-18  Sheriff Bot  <webkit.review.bot@gmail.com>
399
400         Unreviewed, rolling out r72357.
401         http://trac.webkit.org/changeset/72357
402         https://bugs.webkit.org/show_bug.cgi?id=49784
403
404         It broke test-webkitpy tests since the patch didn't have the
405         corresponding test update (Requested by kinuko on #webkit).
406
407         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
408         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
409
410 2010-11-18  Mihai Parparita  <mihaip@chromium.org>
411
412         Reviewed by Tony Chang.
413
414         Rebaseline server: add rebaseline queue
415         https://bugs.webkit.org/show_bug.cgi?id=49763
416         
417         Add a UI queue (a multi-item <select>) where tests to be baselined
418         can be enqueued (also supports basic removal). A queue is used instead
419         of immediately doing rebaselines since SCM operations can take a while.
420         It's better to go through lots of tests quickly to mark them as
421         needing rebaselining and then batch the SCM operations.
422         
423         Also adds a simple log where results can be displayed.
424
425         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
426         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
427         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
428         * Scripts/webkitpy/tool/commands/data/rebaselineserver/queue.js: Added.
429         * Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js:
430         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
431
432 2010-11-18  Gavin Barraclough  <barraclough@apple.com>
433
434         Reviewed by Oliver Hunt.
435
436         Bug 49708 - Stop recompiling functions to regenerate exception info.
437
438         Instead only hold info as necessary – keep divot info is the inspector
439         is enabled, line number info is debugging or profiling, and handler
440         info for functions with try/catch.
441
442         * DumpRenderTree/mac/DumpRenderTree.mm:
443         (shouldEnableDeveloperExtras):
444             - always enable the developer tools from DRT, to ensure we
445               produce rich error messages on JavaScript exceptions.
446
447 2010-11-18  Kinuko Yasuda  <kinuko@chromium.org>
448
449         Reviewed by Ojan Vafai.
450
451         [Chromium] REGRESSION: Record actual test results in the JSON result summary for non-layout tests
452         https://bugs.webkit.org/show_bug.cgi?id=49702
453
454         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
455         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
456
457 2010-11-18  Dirk Pranke  <dpranke@chromium.org>
458
459         Reviewed by Tony Chang.
460
461         new-run-webkit-tests: create first part of 'message_broker' class for multiprocessing fixes
462
463         Create the first version of the 'message_broker' package. This
464         class will encapsulate all of the threading/multiprocessing and
465         message-sending details for the communication between the
466         'manager' object and the 'worker' objects. For the moment, it
467         just holds some routines and tests for logging thread stacks.
468
469         There should be no functional changes in this patch, just moving stuff
470         around.
471
472         https://bugs.webkit.org/show_bug.cgi?id=49707
473
474         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
475         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread_unittest.py: Removed.
476         * Scripts/webkitpy/layout_tests/layout_package/message_broker.py: Added.
477         * Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py: Added.
478         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
479         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
480
481 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
482
483         Reviewed by Adam Roben.
484
485         Remove leftover Windows Debug_Internal configurations
486         https://bugs.webkit.org/show_bug.cgi?id=49758
487
488         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
489         * DumpRenderTree/win/DumpRenderTree.vcproj:
490         * DumpRenderTree/win/ImageDiff.vcproj:
491         * MiniBrowser/MiniBrowser.vcproj:
492         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
493         * WebKitAPITest/WebKitAPITest.vcproj:
494         * WebKitTestRunner/win/InjectedBundle.vcproj:
495         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
496         * WinLauncher/WinLauncher.vcproj:
497
498 2010-11-18  Adam Roben  <aroben@apple.com>
499
500         Add three new Windows XP build slaves
501
502         Reviewed by Steve Falkenburg.
503
504         * BuildSlaveSupport/build.webkit.org-config/config.json: Added three
505         new slaves, and added them to the Windows XP Debug (Tests) builder.
506
507 2010-11-18  Daniel Bates  <dbates@rim.com>
508
509         Reviewed by Adam Roben.
510
511         update-webkit-support-libs should fall back to existing
512         WebKitSupportLibrary version if there is no internet connectivity
513         https://bugs.webkit.org/show_bug.cgi?id=49503
514
515         Fall back to existing support libraries (if present) when there is
516         no internet connection.
517
518         Currently, update-webkit-support-libs dies with an "out-of-date"
519         error when there is no internet connection because it cannot
520         retrieve versioning information from developer.apple.com. Because
521         update-webkit-support-libs fails, build-webkit fails. Instead,
522         if there is no internet connection and the support libraries are
523         present then we should warn the user and exit() with success so
524         that build-webkit can work without an internet connection.
525
526         * Scripts/update-webkit-support-libs:
527
528 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
529
530         Reviewed by Adam Roben.
531
532         Debug_Internal Windows configuration is unnecessary, should be removed
533         https://bugs.webkit.org/show_bug.cgi?id=49753
534
535         * DumpRenderTree/win/DumpRenderTree.cpp:
536         * MiniBrowser/win/stdafx.h:
537         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp:
538         * WebKitTestRunner/win/TestControllerWin.cpp:
539
540 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
541
542         Reviewed by Adam Roben.
543
544         Clean up vcproj errors
545         https://bugs.webkit.org/show_bug.cgi?id=49741
546
547         * MiniBrowser/MiniBrowser.vcproj:
548         * WebKitAPITest/WebKitAPITest.vcproj:
549         * WebKitTestRunner/win/InjectedBundle.vcproj:
550         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
551
552 2010-11-18  Hayato Ito  <hayato@chromium.org>
553
554         Reviewed by Shinichiro Hamaji.
555
556         [NRWT] Fix a break of '--new-baseline' feature in pixel_test
557         which is caused by r72249.
558
559         https://bugs.webkit.org/show_bug.cgi?id=49751
560
561         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
562
563 2010-11-18  Adam Roben  <aroben@apple.com>
564
565         Implement layoutTestController.waitForPolicyDelegate on Windows
566
567         Fixes <http://webkit.org/b/25038> <rdar://problem/6790213>.
568
569         Reviewed by Simon Fraser.
570
571         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
572         (LayoutTestController::waitForPolicyDelegate): Implemented by porting
573         code from LayoutTestControllerMac.mm.
574
575 2010-11-18  Hayato Ito  <hayato@chromium.org>
576
577         Reviewed by Shinichiro Hamaji.
578
579         [NRWT] Fix a break of '--new-baseline' feature which is caused by r72249.
580
581         https://bugs.webkit.org/show_bug.cgi?id=49740
582
583         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
584
585 2010-11-18  Tony Chang  <tony@chromium.org>
586
587         Reviewed by Adam Barth.
588
589         run platform/chromium/plugins/refcount-leaks.html on all platforms
590         https://bugs.webkit.org/show_bug.cgi?id=49485
591
592         Add PluginObject.testObjectCount which returns the number of allocated
593         TestObjects. Add PluginObject.testCreateTestObject which allocates
594         and returns a TestObject.
595
596         Add TestObject.refCount which returns the number of refs on the
597         TestObject.
598
599         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
600         (pluginGetProperty):
601         (pluginInvoke):
602         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
603         (getTestObjectCount):
604         (testAllocate):
605         (testDeallocate):
606         (testGetProperty):
607         (testScriptObjectInvoke): Release a ref to a plugin object that was
608             previously leaked.
609         * DumpRenderTree/TestNetscapePlugIn/TestObject.h:
610
611 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
612
613         Reviewed by Adam Roben.
614
615         WebKitTools projects (WebKitTestRunner, record-memory-win) should use vsprops for common build settings
616         https://bugs.webkit.org/show_bug.cgi?id=49711
617
618         * WebKitTestRunner/Configurations/WebKitTestRunnerCFLite.vsprops: Added.
619         * WebKitTestRunner/Configurations/WebKitTestRunnerCommon.vsprops: Added.
620         * WebKitTestRunner/Configurations/WebKitTestRunnerCoreFoundation.vsprops: Added.
621         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
622         * record-memory-win/record-memory-win-common.vsprops: Added.
623         * record-memory-win/record-memory-win.vcproj:
624
625 2010-11-18  Andras Becsi  <abecsi@inf.u-szeged.hu>
626
627         Reviewed by Csaba Osztrogonác.
628
629         [Qt][WK2] Only add user agent strings to the list which aren't listed yet.
630
631         * MiniBrowser/qt/BrowserWindow.cpp:
632         (BrowserWindow::updateUserAgentList):
633
634 2010-11-17  Adam Roben  <aroben@apple.com>
635
636         Don't trigger Windows builds when chromium-win test results change
637
638         Reviewed by Mark Rowe.
639
640         * Scripts/webkitpy/common/config/build.py:
641         (_should_file_trigger_build): Changed the regular expression used to
642         search for directory names to only consider full directory names,
643         rather than matching directory names where the string we care about is
644         a suffix of the name.
645
646         * Scripts/webkitpy/common/config/build_unittest.py:
647         (ShouldBuildTest): Added a test.
648
649 2010-11-17  Hayato Ito  <hayato@chromium.org>
650
651         Reviewed by Shinichiro Hamaji.
652
653         Remove a uri member from TestInput class.
654
655         https://bugs.webkit.org/show_bug.cgi?id=49691
656
657         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
658         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
659
660 2010-11-17  Adam Roben  <aroben@apple.com>
661
662         Move FindSafari's settings into a vsprops file
663
664         Fixes <http://webkit.org/b/49699> FindSafari should use a vsprops file
665
666         Reviewed by Steve Falkenburg.
667
668         * FindSafari/FindSafari.vcproj:
669         * FindSafari/FindSafariCommon.vsprops: Added.
670
671 2010-11-17  Adam Roben  <aroben@apple.com>
672
673         Use a minimum font size of 0 in DRT on Windows
674
675         This is the Windows equivalent of r72164.
676
677         Rubber-stamped by Dave Hyatt.
678
679         * DumpRenderTree/win/DumpRenderTree.cpp:
680         (resetDefaultsToConsistentValues):
681
682 2010-11-17  Dirk Pranke  <dpranke@chromium.org>
683
684         Reviewed by Andreas Kling.
685
686         webkit-patch pretty-diff shouldn't need to be run from a checkout root in git
687         https://bugs.webkit.org/show_bug.cgi?id=49639
688
689         * Scripts/webkitpy/common/checkout/scm.py:
690
691 2010-11-17  Mihai Parparita  <mihaip@chromium.org>
692
693         Reviewed by Tony Chang.
694
695         Rebaseline server: add loupe for image diffs
696         https://bugs.webkit.org/show_bug.cgi?id=49692
697         
698         Add a loupe (magnifiying glass) for inspecting image diffs. Shows an
699         enlarged area of the expected, actual and diff images side by side.
700
701         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
702         * Scripts/webkitpy/tool/commands/data/rebaselineserver/loupe.js: Added.
703         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
704         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
705
706 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
707
708         Reviewed by Adam Roben.
709
710         WebKitLauncherWin should use vsprops for shared build settings
711         https://bugs.webkit.org/show_bug.cgi?id=49696
712
713         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
714         * WebKitLauncherWin/WebKitLauncherWinCommon.vsprops: Added.
715
716 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
717
718         Reviewed by Adam Roben.
719
720         WinLauncher should use vsprops for shared build settings
721         https://bugs.webkit.org/show_bug.cgi?id=49695
722
723         * WinLauncher/WinLauncher.cpp:
724         (_tWinMain):
725         (PrintView):
726         (WndProc):
727         (MyEditProc):
728         * WinLauncher/WinLauncher.vcproj:
729         * WinLauncher/WinLauncherCommon.vsprops: Added.
730
731 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
732
733         Delete unused file unintentionally added in last change.
734
735         * DumpRenderTree/win/DumpRenderTreeCURL.vsprops: Removed.
736
737 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
738
739         Reviewed by Adam Roben.
740
741         DumpRenderTree on Windows should use a vsprops file for shared project settings
742         https://bugs.webkit.org/show_bug.cgi?id=49690
743
744         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
745         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePluginCommon.vsprops: Added.
746         * DumpRenderTree/win/DumpRenderTree.vcproj:
747         * DumpRenderTree/win/DumpRenderTreeApple.vsprops: Added.
748         * DumpRenderTree/win/DumpRenderTreeCFLite.vsprops: Added.
749         * DumpRenderTree/win/DumpRenderTreeCURL.vsprops: Added.
750         * DumpRenderTree/win/DumpRenderTreeCairo.vsprops: Added.
751         * DumpRenderTree/win/DumpRenderTreeCommon.vsprops: Added.
752         * DumpRenderTree/win/ImageDiff.vcproj:
753         * DumpRenderTree/win/ImageDiffCommon.vsprops: Added.
754
755 2010-11-17  Hayato Ito  <hayato@chromium.org>
756
757         Refactor TestTypeBase.compare_output().
758
759         Introduce a TestOutput class and update compare_output() of each test
760         types so that they can take both actual and expected TestOutput objects.
761
762         https://bugs.webkit.org/show_bug.cgi?id=49431
763
764         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
765         * Scripts/webkitpy/layout_tests/layout_package/test_output.py: Added.
766         * Scripts/webkitpy/layout_tests/port/base.py:
767         * Scripts/webkitpy/layout_tests/port/chromium.py:
768         * Scripts/webkitpy/layout_tests/port/dryrun.py:
769         * Scripts/webkitpy/layout_tests/port/test.py:
770         * Scripts/webkitpy/layout_tests/port/webkit.py:
771         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
772         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
773         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
774         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
775
776 2010-11-17  Adam Roben  <aroben@apple.com>
777
778         Make each Windows Test builder use the same OS for all its slaves
779
780         Currently each WebKit1 Windows Test builder has one Windows XP slave
781         and one Windows 7 slave. Having different OSes in a single builder
782         meant that the builders would alternate between red and green when
783         there were OS-specific failures.
784
785         Fixes <http://webkit.org/b/49688> Windows Test builders alternate
786         between red and green because of multiple OS versions
787
788         Reviewed by Mark Rowe.
789
790         * BuildSlaveSupport/build.webkit.org-config/config.json: Renamed
791         "Windows Release (Tests)" to "Windows 7 Release (Tests)", "Windows
792         Debug (Tests)" to "Windows XP Debug (Tests)", and "Windows Release
793         (WebKit2 Tests)" to "Windows 7 Release (WebKit2 Tests)". Updated the
794         builders for the two WebKit1 builders so that they would all use the
795         appropriate OS.
796
797 2010-11-17  James Robinson  <jamesr@chromium.org>
798
799         Reviewed by Dan Bernstein.
800
801         [chromium] Set minimum font size to 0 for DRT
802         https://bugs.webkit.org/show_bug.cgi?id=49677
803
804         Followup for http://trac.webkit.org/changeset/72141.
805
806         Test: http/tests/misc/acid3.html
807
808         * DumpRenderTree/chromium/WebPreferences.cpp:
809         (WebPreferences::reset):
810
811 2010-11-17  Andras Becsi  <abecsi@webkit.org>
812
813         Unreviewed build fix.
814
815         [Qt] Add missing file I forgot in r72220.
816
817         * MiniBrowser/MiniBrowser.qrc: Added.
818
819 2010-11-17  Andras Becsi  <abecsi@webkit.org>
820
821         Reviewed by Andreas Kling.
822
823         [Qt][WK2] Make it possible to set custom user agent strings in MiniBrowser.
824         https://bugs.webkit.org/show_bug.cgi?id=49627
825
826         Add a user agent dialog and a resource file for useragentlist.txt to MiniBrowser.
827
828         * MiniBrowser/MiniBrowser.qrc: Added.
829         Add it here to prevent qmake from detecting it
830         since it needs to be copied to the build directory.
831         * MiniBrowser/qt/BrowserWindow.cpp:
832         (BrowserWindow::BrowserWindow):
833         (BrowserWindow::updateUserAgentList):
834         (BrowserWindow::showUserAgentDialog):
835         * MiniBrowser/qt/BrowserWindow.h:
836         * MiniBrowser/qt/MiniBrowser.pro:
837
838 2010-11-17  Adam Roben  <aroben@apple.com>
839
840         Make changes to the Mac WebKit2 Skipped file trigger Windows builds
841
842         We use that file on Windows!
843
844         Fixes <http://webkit.org/b/49643> Changes to
845         LayoutTests/platform/mac-wk2 should trigger a Windows build, but don't
846
847         Reviewed by Steve Falkenburg.
848
849         * Scripts/webkitpy/common/config/build.py:
850         (_should_file_trigger_build): Added the mac-wk2 directory and make it
851         trigger builds on SnowLeopard and Windows.
852
853         * Scripts/webkitpy/common/config/build_unittest.py:
854         (ShouldBuildTest.test_should_build): Added a test.
855
856 2010-11-17  Satish Sampath  <satish@chromium.org>
857
858         Reviewed by Jeremy Orlow.
859
860         Clear the speech input mock explicitly before each test.
861         https://bugs.webkit.org/show_bug.cgi?id=49660
862
863         * DumpRenderTree/chromium/LayoutTestController.cpp:
864         (LayoutTestController::setMockSpeechInputResult):
865         * DumpRenderTree/chromium/LayoutTestController.h:
866         * DumpRenderTree/chromium/WebViewHost.cpp:
867         (WebViewHost::speechInputController):
868         (WebViewHost::reset): Invoke speech input mock's clearResults method.
869         * DumpRenderTree/chromium/WebViewHost.h:
870         (WebViewHost::speechInputControllerMock): Moved mock from LayoutTestController.
871
872 2010-11-17  Alexander Pavlov  <apavlov@chromium.org>
873
874         Unreviewed, specify IRC nickname in committers.py.
875
876         * Scripts/webkitpy/common/config/committers.py:
877
878 2010-11-17  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
879
880         Reviewed by Tony Chang.
881
882         [NRWT] Make http locking similar to perl implementation
883         https://bugs.webkit.org/show_bug.cgi?id=49187
884
885         * Scripts/webkitpy/common/system/file_lock.py: Added.
886         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
887         * Scripts/webkitpy/layout_tests/port/http_lock.py:
888
889 2010-11-17  MORITA Hajime  <morrita@google.com>
890
891         Reviewed by Kent Tamura.
892
893         [Chromium][DRT] EventSender.contextClick() should aware spellchecking
894         https://bugs.webkit.org/show_bug.cgi?id=49366
895
896         - EvenSender: Checked WebContextMenuData.misspelledWord and added extra context menu entries
897           according to the spellchecker suggestion
898         - MockSpellCheck: add fillSuggestionList to provide fake suggestions.
899
900         test_expectations.txt will be changed after this change is ported to test_shell.
901         
902         * DumpRenderTree/chromium/EventSender.cpp:
903         (makeMenuItemStringsFor):
904         (EventSender::contextClick):
905         * DumpRenderTree/chromium/MockSpellCheck.cpp:
906         (MockSpellCheck::fillSuggestionList):
907         (MockSpellCheck::initializeIfNeeded):
908         * DumpRenderTree/chromium/MockSpellCheck.h:
909         * DumpRenderTree/chromium/WebViewHost.cpp:
910         (WebViewHost::mockSpellCheck):
911         * DumpRenderTree/chromium/WebViewHost.h:
912
913 2010-11-16  Dave Hyatt  <hyatt@apple.com>
914
915         Reviewed by Sam Weinig.
916
917         Make sure the pref for minimum font size is just explicitly set to 0 every time,
918         since the bots have 1 stuck in their plist otherwise.
919
920         * DumpRenderTree/mac/DumpRenderTree.mm:
921         (resetDefaultsToConsistentValues):
922
923 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
924
925         Reviewed by Tony Chang.
926
927         Rebaseline server: compute diffs client-side
928         https://bugs.webkit.org/show_bug.cgi?id=49640
929         
930         The image diff output from the DRT is pretty bad on some ports (at
931         least the Mac one), so it's better to compute diffs on the client by
932         using <canvas>.
933
934         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
935         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
936         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
937
938 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
939
940         Reviewed by Tony Chang.
941
942         Rebaseline server: display test results
943         https://bugs.webkit.org/show_bug.cgi?id=49626
944         
945         Adds basic result display to the rebaseline server. On the Python side
946         this involves:
947         - Parsing the unexpected_results.json into a dictionary.
948         - Serving it as JSON under /results.json.
949         (the JSON -> dict -> JSON transform isn't strictly necessary right now,
950         but I'll need to have access to the parsed results on the Python side
951         for follow-up changes).
952         
953         On the web UI side this adds:
954         - Markup for display image and text results (expected, actual, diff),
955           and JS for populating it.
956         - Markup for breaking down test results by failure type and directory,
957           and JS for populating it.
958
959         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
960         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
961         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
962         * Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js: Added.
963         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
964
965 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
966
967         Reviewed by Ojan Vafai.
968
969         See Chromium issue http://codereview.chromium.org/5133001/ - we
970         are modifying the build bots to pass in the builder name with
971         the "GPU" string appended instead of appending it in the code.
972
973         https://bugs.webkit.org/show_bug.cgi?id=49636
974
975         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
976
977 2010-11-16  Steve Falkenburg  <sfalken@apple.com>
978
979         Reviewed by Adam Roben.
980
981         Disable LTCG for Windows Release builds. Add new Release_LTCG configuration.
982         https://bugs.webkit.org/show_bug.cgi?id=49632
983
984         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
985         * DumpRenderTree/win/DumpRenderTree.vcproj:
986         * DumpRenderTree/win/ImageDiff.vcproj:
987         * MiniBrowser/MiniBrowser.vcproj:
988         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
989         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
990         * WebKitTestRunner/win/InjectedBundle.vcproj:
991         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
992         * WinLauncher/WinLauncher.vcproj:
993         * record-memory-win/record-memory-win.vcproj:
994
995 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
996
997         Reviewed by Ojan Vafai.
998
999         new-run-webkit-tests: rename TestInfo to TestInput, move image hash to work thread
1000
1001         Rename the TestInfo class to TestInput to be clearer about its
1002         function, and move the checksum-reading code into dump_render_tree_thread
1003         to avoid cross-thread access.
1004
1005         https://bugs.webkit.org/show_bug.cgi?id=49573
1006
1007         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1008         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1009         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1010
1011 2010-11-16  Dave Hyatt  <hyatt@apple.com>
1012
1013         Fix bustage. Remove the minimum font size pref setting in DumpRenderTree.
1014
1015         * DumpRenderTree/mac/DumpRenderTree.mm:
1016         (resetDefaultsToConsistentValues):
1017
1018 2010-11-16  Stephanie Lewis  <slewis@apple.com>
1019
1020         Reviewed by Geoff Garen.
1021
1022         <rdar://problem/8624267> Leak creating offscreen webview running fast/dom tests
1023
1024         Use shared DeviceOrientationProviderMock. The old allocation method created a new WebDeviceOrientationProviderMock 
1025         for every WebView, and the WebKit API doesn't support that ownership model.
1026
1027         I also fixed a leak in setMockDeviceOrientation by adding a missing release.
1028
1029         * DumpRenderTree/mac/DumpRenderTree.mm:
1030         (createWebViewAndOffscreenWindow):
1031         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1032         (LayoutTestController::setMockDeviceOrientation):
1033
1034 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
1035
1036         Reviewed by Ojan Vafai.
1037
1038         new-run-webkit-tests: log the process id in --verbose mode
1039
1040         As part of the switch to multi-process mode, we should log the
1041         process id when running w/ --verbose.
1042
1043         https://bugs.webkit.org/show_bug.cgi?id=49571
1044
1045         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
1046
1047 2010-11-16  Adam Roben  <aroben@apple.com>
1048
1049         Only trigger Windows builds when files that we actually use on Windows
1050         are changed
1051
1052         A new Scheduler subclass, PlatformSpecificScheduler, has been added.
1053         It uses the new webkitpy.common.config.build module to determine
1054         whether a particular change should trigger a build on a particular
1055         platform. The Windows builders have been switched to use a
1056         PlatformSpecificScheduler.
1057
1058         The logic to determine whether or not a particular change should
1059         trigger a build on a given platform has only been implemented/tested
1060         for Windows. I tried to make it easy to add more platforms in the
1061         future, but I don't have enough familiarity with all platforms to be
1062         able to implement it for them.
1063
1064         Fixes <http://webkit.org/b/49407> Windows builders kick off builds for
1065         lots irrelevant changes (e.g., rebaselining Chromium test results)
1066
1067         Reviewed by Eric Seidel.
1068
1069         * BuildSlaveSupport/build.webkit.org-config/config.json: Use a
1070         PlatformSpecificScheduler for the Windows builders.
1071
1072         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1073         (PlatformSpecificScheduler.__init__): Added. Stores our platform, sets
1074         up a ChangeFilter that filters to our branch and will call through to
1075         our filter method, and calls up to our base class.
1076         (PlatformSpecificScheduler.filter): Calls through to
1077         build.should_build to find out whether we should trigger a build for
1078         this change.
1079
1080         * Scripts/webkitpy/common/net/build.py: Added.
1081         (_should_file_trigger_build): Uses a set of directories and regexp
1082         patterns to determine whether the given file should trigger a build on
1083         the given platform. As mentioned earlier, this has only been
1084         implemented for Windows, though I did try to make some guesses about
1085         other platforms.
1086         (should_build): Returns true if any of the files should trigger a
1087         build on the given platform.
1088
1089         * Scripts/webkitpy/common/net/build_unittest.py: Added.
1090         (ShouldBuildTest.test_should_build): Does some basic testing to make
1091         sure we're triggering builds for the right files. It only tests
1092         Windows for now, though I tried to make some guesses about other
1093         platforms.
1094
1095 2010-11-16  Simon Hausmann  <simon.hausmann@nokia.com>
1096
1097         Reviewed by Andreas Kling.
1098
1099         [Qt] Remove synchronous QWebPage::checkPermissions signal
1100         https://bugs.webkit.org/show_bug.cgi?id=46810
1101
1102         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1103         (WebCore::WebPage::WebPage):
1104         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1105         (LayoutTestController::grantDesktopNotificationPermission): When granting
1106         permission, grant it directly on the QWebPage/Frame, that will remember it.
1107         * QtTestBrowser/webpage.cpp:
1108         (WebPage::WebPage):
1109         * QtTestBrowser/webpage.h:
1110
1111
1112 2010-11-16  Andras Becsi  <abecsi@inf.u-szeged.hu>
1113
1114         Reviewed by Csaba Osztrogonác.
1115
1116         [Qt] Add isPrinting and setPrinting methods to DRT's LayoutTestController.
1117
1118         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1119         (LayoutTestController::reset):
1120         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1121         (LayoutTestController::isPrinting):
1122         (LayoutTestController::setPrinting):
1123
1124 2010-11-16  Balazs Kelemen  <kbalazs@webkit.org>
1125
1126         Reviewed by Andreas Kling.
1127
1128         [Qt][WK2] Add a way to turn on tiled backing store in MiniBrowser
1129         https://bugs.webkit.org/show_bug.cgi?id=49587
1130
1131         * MiniBrowser/qt/BrowserView.cpp:
1132         (BrowserView::BrowserView): Added BackingStoreType
1133         argument to ctor. Propagating it to the QGraphicsWKView.
1134         * MiniBrowser/qt/BrowserView.h:
1135         * MiniBrowser/qt/BrowserWindow.cpp:
1136         (BrowserWindow::BrowserWindow):
1137         (BrowserWindow::newWindow): Create the new BrowserWindow
1138         with the same BackingStoreType that this instance has.
1139         * MiniBrowser/qt/BrowserWindow.h: Added BackingStoreType member.
1140         * MiniBrowser/qt/main.cpp:
1141         (main): Use tiled backing store if got -tiled command line argument.
1142         Avoid redundant copy of command line args.
1143
1144 2010-11-16  John Knottenbelt  <jknotten@chromium.org>
1145
1146         Reviewed by Jeremy Orlow.
1147
1148         Move DeviceOrientationClientMock from LayoutTestController to WebViewHost.
1149         https://bugs.webkit.org/show_bug.cgi?id=48506
1150
1151         This ensures that there is one mock per page WebView / WebViewHost. This is a
1152         design constraint of the DeviceOrientationClientMock because the m_controller
1153         field can only store a pointer to a single instance of the corresponding
1154         DeviceOrientationController.
1155
1156         Test: fast/dom/DeviceOrientation/no-page-cache.html
1157
1158         * DumpRenderTree/chromium/LayoutTestController.cpp:
1159         (LayoutTestController::setMockDeviceOrientation):
1160         (LayoutTestController::deviceOrientationClientMock):
1161         * DumpRenderTree/chromium/LayoutTestController.h:
1162         * DumpRenderTree/chromium/WebViewHost.cpp:
1163         (WebViewHost::deviceOrientationClientMock):
1164         (WebViewHost::deviceOrientationClient):
1165         (WebViewHost::reset):
1166         * DumpRenderTree/chromium/WebViewHost.h:
1167
1168 2010-11-15  Andras Becsi  <abecsi@webkit.org>
1169
1170         Reviewed by Andreas Kling.
1171
1172         [Qt][WK2] Avoid polling in WebKitTestRunner.
1173         https://bugs.webkit.org/show_bug.cgi?id=49542
1174
1175         Make the WK2 testing session about 10-15% faster and decrease random flakiness
1176         resulting from timing skew by removing the millisecond polling from TestControllerQt.
1177
1178         * WebKitTestRunner/TestController.cpp:
1179         (WTR::TestController::didFinishLoadForFrame):
1180         * WebKitTestRunner/TestController.h:
1181         * WebKitTestRunner/TestInvocation.cpp:
1182         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
1183         * WebKitTestRunner/mac/TestControllerMac.mm:
1184         (WTR::TestController::notifyDone): Add method stub.
1185         * WebKitTestRunner/qt/TestControllerQt.cpp:
1186         (WTR::TestControllerRunLoop::instance):
1187         (WTR::TestControllerRunLoop::start):
1188         (WTR::TestControllerRunLoop::stop):
1189         (WTR::TestControllerRunLoop::TestControllerRunLoop):
1190         (WTR::TestControllerRunLoop::timerEvent):
1191         (WTR::TestController::notifyDone):
1192         (WTR::TestController::platformRunUntil):
1193         * WebKitTestRunner/win/TestControllerWin.cpp:
1194         (WTR::TestController::notifyDone): Add method stub.
1195
1196 2010-11-15  Darin Adler  <darin@apple.com>
1197
1198         * Scripts/webkitpy/common/net/bugzilla: Added property svn:ignore.
1199
1200 2010-11-15  Benjamin Poulain  <benjamin.poulain@nokia.com>
1201
1202         Reviewed by Andreas Kling.
1203
1204         [Qt] [WK2] Add Ctrl+L shortcut to Qt Minibrowser
1205         https://bugs.webkit.org/show_bug.cgi?id=49544
1206
1207         Add the shortcut to BrowserWindow.
1208
1209         * MiniBrowser/qt/BrowserWindow.cpp:
1210         (BrowserWindow::BrowserWindow):
1211         (BrowserWindow::openLocation):
1212         * MiniBrowser/qt/BrowserWindow.h:
1213
1214 2010-11-14  David Levin  <levin@chromium.org>
1215
1216         Reviewed by Daniel Bates.
1217
1218         check-webkit-style should detect PassRefPtr usage in functions.
1219         https://bugs.webkit.org/show_bug.cgi?id=49513
1220
1221         * Scripts/webkitpy/style/checkers/cpp.py:
1222         (check_for_function_lengths): Revert a comment change that I
1223          accidentally made in r71986.
1224         (check_pass_ptr_usage): Added the code to do the check.
1225         (process_line): Added the call to check_pass_ptr_usage.
1226         (CppChecker): Added the new error category.
1227         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
1228         (CppStyleTestBase::perform_pass_ptr_check): Runs the new check for
1229          testing purposes.
1230         (PassPtrTest::*): The class/functions to unit test the new
1231          functionality.
1232
1233 2010-11-14  David Levin  <levin@chromium.org>
1234
1235         Reviewed by Shinichiro Hamaji.
1236
1237         check-webkit-style function detection and the line count style checks should be separate.
1238         https://bugs.webkit.org/show_bug.cgi?id=49512
1239
1240         * Scripts/webkitpy/style/checkers/cpp.py: Do the separation.
1241         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Adjust the test to
1242           call the detection function and fix line counts in two places now that
1243           the code really only counts the lines in the body of the function.
1244
1245 2010-11-14  Andreas Kling  <kling@webkit.org>
1246
1247         Reviewed by Antonio Gomes.
1248
1249         http/tests/plugins tests print "Unhandled variable" to stderr under Qt
1250         https://bugs.webkit.org/show_bug.cgi?id=33438
1251
1252         Stifle "Unhandled variable" warning to match what the other
1253         TestNetscapePlugin does (NPP_GetValue in TestNetscapePlugIn/main.cpp
1254         simply returns NPERR_GENERIC_ERROR for unhandled variables.)
1255
1256         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
1257         (webkit_test_plugin_get_value):
1258
1259 2010-11-13  David Levin  <levin@chromium.org>
1260
1261         Reviewed by Eric Seidel.
1262
1263         check-webkit-style function detection crashes on functions in templates.
1264         https://bugs.webkit.org/show_bug.cgi?id=49504
1265
1266         For example "template <bool shouldClose(const Element*)>".
1267
1268         * Scripts/webkitpy/style/checkers/cpp.py: If not function is detected, bail out.
1269         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Modified the complex
1270           function detection test to expose this issue.
1271
1272 2010-11-12  David Levin  <levin@chromium.org>
1273
1274         Reviewed by Shinichiro Hamaji.
1275
1276         check-webkit-style function detection doesn't detect indented functions declaractions.
1277         https://bugs.webkit.org/show_bug.cgi?id=49446
1278
1279         Indented function declarations occur inside class definitions, so
1280         they are a pretty common (and worth detecting).
1281
1282         * Scripts/webkitpy/style/checkers/cpp.py:
1283           Changed regex to allow indentation.
1284           Changed the function start detection to only happen when not in a
1285           function.
1286           Changed function end detection to work based on matching braces
1287           instead of finding a close brace at the beginning of the line.
1288           Fixed close_expression to do what it says when it doesn't find
1289           the close.
1290         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Indented function test.
1291
1292 2010-11-12  Daniel Bates  <dbates@rim.com>
1293
1294         Rubber-stamped by Nikolas Zimmermann.
1295
1296         Make do-file-rename rename files in the directory WebKit2.
1297
1298         This also makes the list of searched directories in do-file-rename
1299         consistent with the list of searched directories in the script do-webcore-rename.
1300
1301         * Scripts/do-file-rename:
1302
1303 2010-11-12  Dirk Pranke  <dpranke@chromium.org>
1304
1305         Reviewed by James Robinson.
1306
1307         Attempt yet again to land the fix for bug 49360 (respecting
1308         set-webkit-configuration). We need to handle the cases where
1309         trying to run webkit-build-directory to find out where the
1310         default configuration might be fails (that shows up on some
1311         Chromium bots that apparently don't have perl installed).
1312
1313         https://bugs.webkit.org/show_bug.cgi?id=49360
1314
1315         * Scripts/webkitpy/layout_tests/port/config.py:
1316         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Added.
1317         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1318
1319 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
1320
1321         Reviewed by Adam Barth.
1322
1323         webkit-patch land logs commit message twice
1324         https://bugs.webkit.org/show_bug.cgi?id=49482
1325         
1326         Don't log the comment that's used by bugzilla.close_bug_as_fixed, since
1327         that's just the commit message, which we now output following the
1328         actual commit.
1329
1330         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
1331
1332 2010-11-12  Tony Chang  <tony@chromium.org>
1333
1334         Reviewed by Adam Barth.
1335
1336         run platform/chromium/plugins/script-object-invoke.html on all platforms
1337         https://bugs.webkit.org/show_bug.cgi?id=49280
1338
1339         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1340         (testScriptObjectInvoke):
1341         (pluginInvoke):
1342
1343 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
1344
1345         Reviewed by Adam Barth.
1346
1347         webkit-patch land-cowboy (and other commands) should display the committed SVN revision after landing a patch
1348         https://bugs.webkit.org/show_bug.cgi?id=49471
1349
1350         Right now there's no indication of a successful commit.
1351         
1352         * Scripts/webkitpy/tool/commands/download_unittest.py:
1353         * Scripts/webkitpy/tool/steps/commit.py:
1354
1355 2010-11-12  Kinuko Yasuda  <kinuko@chromium.org>
1356
1357         Reviewed by Ojan Vafai.
1358
1359         Include detailed test modifiers in results.json and enable incremental uploading for non-layout tests
1360         https://bugs.webkit.org/show_bug.cgi?id=49354
1361
1362         Also moved/integrated the upload method from run_webkit_tests.py to json_results_generator.py.
1363
1364         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
1365         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1366         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
1367         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1368
1369 2010-11-12  Tony Chang  <tony@chromium.org>
1370
1371         Reviewed by Kent Tamura.
1372
1373         [chromium] Load a blank page when closing DRT WebViewHosts
1374         https://bugs.webkit.org/show_bug.cgi?id=49417
1375
1376         In test_shell, when we close a window, we first load about:blank
1377         and run GC to fire the destruction logic of the page.  In test_shell,
1378         this happens in ~TestShell.  In DRT, we manage virtual windows as
1379         WebViewHosts, so we need to replicate this logic in ~WebViewHost.
1380
1381         Managing focus is a bit trickier, since normally you would get window
1382         events to reset the focus to the right window.  We do it manually
1383         instead.
1384
1385         This fixes 2 plugin tests on DRT.
1386
1387         * DumpRenderTree/chromium/TestShell.cpp:
1388         (TestShell::~TestShell): Move window close logic to ~WebViewHost.
1389         (TestShell::closeWindow): Move widget close() to WVH destructor and
1390             manage focus manually.
1391         (TestShell::closeRemainingWindows): Comment typo.
1392         * DumpRenderTree/chromium/WebViewHost.cpp:
1393         (WebViewHost::closeWidget):
1394         (invokeCloseWidget):
1395         (WebViewHost::closeWidgetSoon): Close the window asynchronously (matches
1396             test_shell).
1397         (WebViewHost::~WebViewHost):
1398         * DumpRenderTree/chromium/WebViewHost.h:
1399
1400 2010-11-12  David Levin  <levin@chromium.org>
1401
1402         Reviewed by Adam Barth.
1403
1404         cpp.py needs some tidying.
1405         https://bugs.webkit.org/show_bug.cgi?id=49443
1406
1407         * Scripts/webkitpy/style/checkers/cpp.py:
1408           Fix parameter ordering for iteratively_replace_matches_with_char
1409           to mirror other similar functions (and improved comments, etc.)
1410           Also, use a python-ism (for/else) to get rid of a variable that
1411           was there to determine if something had been found in the loop.
1412
1413 2010-11-12  Robert Kroeger  <rjkroege@chromium.org>
1414
1415         Reviewed by James Robinson.
1416
1417         [Chromium/DRT] Make EventSendingController send complete touch events.
1418         https://bugs.webkit.org/show_bug.cgi?id=49285
1419
1420         * DumpRenderTree/chromium/EventSender.cpp:
1421         (EventSender::addTouchPoint):
1422         (EventSender::updateTouchPoint):
1423         (EventSender::sendCurrentTouchEvent):
1424
1425 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
1426
1427         Unreviewed rollout of r71858.
1428
1429         Rollout out r71858 since it breaks new-run-webkit-httpd as used by the
1430         NaCl tests.
1431
1432         * Scripts/webkitpy/layout_tests/port/config.py:
1433         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Removed.
1434         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1435
1436 2010-11-12  Benjamin Poulain  <benjamin.poulain@nokia.com>
1437
1438         Reviewed by Kenneth Rohde Christiansen.
1439
1440         [Qt] TestController::runUntil doesn't honor the timeout parameter
1441         https://bugs.webkit.org/show_bug.cgi?id=48941
1442
1443         Implement the missing timeout of WebKitTestRunner for Qt.
1444
1445         A QElapsedTimer has been added to RunUntilConditionLoop in order
1446         to measure how long the loop has been running. When the timer
1447         is bigger or equal than the timeout, the loop ends even if the
1448         condition is not met.
1449
1450         * WebKitTestRunner/qt/TestControllerQt.cpp:
1451         (WTR::RunUntilConditionLoop::start):
1452         (WTR::RunUntilConditionLoop::run):
1453         (WTR::RunUntilConditionLoop::timerEvent):
1454         (WTR::TestController::platformRunUntil):
1455
1456 2010-11-11  David Levin  <levin@chromium.org>
1457
1458         Reviewed by Shinichiro Hamaji.
1459
1460         check-webkit-style function detection doesn't handle templates with spaces.
1461         https://bugs.webkit.org/show_bug.cgi?id=49427
1462
1463         * Scripts/webkitpy/style/checkers/cpp.py: Stripped out templates when
1464           finding the function name.
1465         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test have a
1466           template with a space.
1467
1468 2010-11-11  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
1469
1470         Reviewed by Shinichiro Hamaji.
1471
1472         Enable check-webkit-style on Qt files
1473         https://bugs.webkit.org/show_bug.cgi?id=48322
1474
1475         Adding exception for _q_... private slots and
1476         ..._data() methods in tests
1477
1478         * Scripts/webkitpy/style/checker.py: enable Qt dirs, add exceptions
1479         * Scripts/webkitpy/style/checker_unittest.py: improve unittests
1480
1481 2010-11-11  David Levin  <levin@chromium.org>
1482
1483         Reviewed by Shinichiro Hamaji.
1484
1485         check-webkit-style function detection doesn't work for templates and destructors.
1486         https://bugs.webkit.org/show_bug.cgi?id=49425
1487
1488         * Scripts/webkitpy/style/checkers/cpp.py: Changed regex for templates and destructors.
1489         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test hit this issue.
1490
1491 2010-11-11  Joone Hur  <joone@kldp.org>
1492
1493         Reviewed by Antonio Gomes.
1494
1495         [GTK][DRT] Implement LayoutTestController::nodesFromRect
1496         https://bugs.webkit.org/show_bug.cgi?id=46598
1497
1498         Support nodesFromRect in DRT
1499
1500         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1501         (LayoutTestController::nodesFromRect):
1502
1503 2010-11-11  MORITA Hajime  <morrita@google.com>
1504
1505         Reviewed by Kent Tamura.
1506
1507         editing/selection/context-menu-on-text.html fails on chromium
1508         https://bugs.webkit.org/show_bug.cgi?id=45898
1509
1510         Enhance Chromium's EventSender::contextClick() to return
1511         array of menu strings. The implementation incomplete because
1512         Although actual context menu is implemented by the browser,
1513         it reflects editability of underlying context thus makes
1514         context-menu-on-text.html passed.
1515         
1516         * DumpRenderTree/chromium/EventSender.cpp:
1517         (makeMenuItemStringsFor): Added.
1518         (EventSender::contextClick): Now returns array of menu strings.
1519         * DumpRenderTree/chromium/WebViewHost.cpp:
1520         (WebViewHost::showContextMenu): Now storing given WebContextMenuData object.
1521         (WebViewHost::clearContextMenuData): Added.
1522         (WebViewHost::lastContextMenuData): Added.
1523         * DumpRenderTree/chromium/WebViewHost.h:
1524
1525 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1526
1527         Unreviewed, build fix.
1528
1529         fix build breakage caused by fix for bug 49380 (r71858).
1530
1531         config_standalone needs to ensure that the packages it needs are
1532         in sys.path.
1533
1534         https://bugs.webkit.org/show_bug.cgi?id=49419
1535
1536         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
1537
1538 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1539
1540         Reviewed by Adam Roben.
1541
1542         Fix NRWT to respect set-webkit-configuration again :(
1543
1544         This change fixes a typo in config.py that was causing the wrong
1545         value to be read initially and us never actually looking into
1546         the filesystem to get the default configuration.
1547         
1548         * Scripts/webkitpy/layout_tests/port/config.py:
1549         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
1550         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1551
1552 2010-11-11  Eric Seidel  <eric@webkit.org>
1553
1554         Reviewed by Adam Barth.
1555
1556         Split out Bug, Attachment and CommitterValidator from bugzilla.py
1557         https://bugs.webkit.org/show_bug.cgi?id=49403
1558
1559         CommitterValidator really had nothing to do with bugzilla.py.
1560         I've put it next to committers.py for now since its rather tied to that file.
1561
1562         * Scripts/webkitpy/common/config/committervalidator.py: Added.
1563         * Scripts/webkitpy/common/config/committervalidator_unittest.py: Added.
1564         * Scripts/webkitpy/common/net/bugzilla/__init__.py:
1565         * Scripts/webkitpy/common/net/bugzilla/attachment.py: Added.
1566         * Scripts/webkitpy/common/net/bugzilla/bug.py: Added.
1567         * Scripts/webkitpy/common/net/bugzilla/bug_unittest.py: Added.
1568         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
1569         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py:
1570         * Scripts/webkitpy/tool/bot/feeders.py:
1571         * Scripts/webkitpy/tool/commands/queues.py:
1572
1573 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1574
1575         Reviewed by Eric Seidel.
1576
1577         fix missing 'pid' argument for http lockfile checking on win32
1578         https://bugs.webkit.org/show_bug.cgi?id=49363
1579
1580         * Scripts/webkitpy/common/system/executive.py:
1581
1582 2010-11-11  Eric Seidel  <eric@webkit.org>
1583
1584         Rubber-stamped by Adam Barth.
1585
1586         Move bugzilla.py into its own module in preparation for splitting one-file-per-class
1587         https://bugs.webkit.org/show_bug.cgi?id=49402
1588
1589         * Scripts/webkitpy/common/net/bugzilla/__init__.py: Added.
1590         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla.py.
1591         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla_unittest.py.
1592
1593 2010-11-10  Chris Guillory  <chris.guillory@google.com>
1594
1595         Reviewed by Chris Fleizach.
1596
1597         Use IAccessibleComparable to compare accessibility objects.
1598         https://bugs.webkit.org/show_bug.cgi?id=49118
1599
1600         * DumpRenderTree/AccessibilityUIElement.cpp:
1601         (AccessibilityUIElement::isEqual):
1602         * DumpRenderTree/AccessibilityUIElement.h:
1603         * DumpRenderTree/win/AccessibilityUIElementWin.cpp:
1604         (comparableObject):
1605         (AccessibilityUIElement::isEqual):
1606
1607 2010-11-10  Csaba Osztrogonác  <ossy@webkit.org>
1608
1609         Reviewed by David Hyatt.
1610
1611         HTML5 Ruby support should be mandatory feature
1612         https://bugs.webkit.org/show_bug.cgi?id=49272
1613
1614         Remove Ruby as optional feature.
1615
1616         * Scripts/build-webkit:
1617
1618 2010-11-10  Adam Roben  <aroben@apple.com>
1619
1620         Print VC++ Express build logs in the same order that the projects are
1621         built
1622
1623         Fixes <http://webkit.org/b/49326> print-vse-failure-logs prints logs
1624         in an unhelpful order
1625
1626         Reviewed by Eric Seidel.
1627
1628         * Scripts/print-vse-failure-logs:
1629         (PrintVisualStudioExpressLogs._build_order): Added. Uses
1630         print-msvc-project-dependencies to figure out the order in which
1631         projects are built.
1632         (PrintVisualStudioExpressLogs._sort_buildlogs): Added. Sorts the logs
1633         based on their build order and project name.
1634         (PrintVisualStudioExpressLogs._obj_directory): Moved code to find the
1635         scripts directory from here...
1636         (PrintVisualStudioExpressLogs._scripts_directory): ...to here.
1637         (PrintVisualStudioExpressLogs.main): Sort the logs before printing
1638         them.
1639
1640 2010-11-10  Adam Roben  <aroben@apple.com>
1641
1642         Only print the interesting text from VC++ Express build logs
1643
1644         Fixes <http://webkit.org/b/49325> It's hard to find the interesting
1645         output from print-vse-failure-logs
1646
1647         Reviewed by Eric Seidel.
1648
1649         * Scripts/print-vse-failure-logs: Removed unnecessary os.path import,
1650         added newly-required imports. Removed a comment that was essentially a
1651         FIXME, and is now fixed.
1652         (PrintVisualStudioExpressLogs._relevant_text): Added. Uses
1653         BeautifulSoup to extract the relevant text from the build log.
1654         (PrintVisualStudioExpressLogs.main): Only print the relevant text.
1655
1656 2010-11-09  Mihai Parparita  <mihaip@chromium.org>
1657
1658         Reviewed by Tony Chang.
1659
1660         [Chromium] http/tests/security/XFrameOptions fail with Chromium DRT
1661         https://bugs.webkit.org/show_bug.cgi?id=49286
1662         
1663         The tests in http/tests/security/XFrameOptions started to fail after
1664         r71297, since it changed the WebViewHost::assignIdentifierToRequest
1665         logic to always assign identifiers to requests, even if we're not going
1666         to be dumping resource load callbacks at that point. These tests
1667         only call LayoutTestController.dumpResourceLoadCallbacks after the
1668         request is started, thus their expectations have "<unknown>" as the
1669         identifier. Change WebViewHost to the previous behavior, which also
1670         match the other ports'.
1671
1672         * DumpRenderTree/chromium/WebViewHost.cpp:
1673         (WebViewHost::assignIdentifierToRequest):
1674         (WebViewHost::removeIdentifierForRequest):
1675
1676 2010-11-08  Tony Chang  <tony@chromium.org>
1677
1678         Reviewed by Adam Barth.
1679
1680         run platform/chromium/plugins/return-npobject.html on all platforms
1681         https://bugs.webkit.org/show_bug.cgi?id=49103
1682
1683         This tests that we can get an NPObject returned through a method on
1684         an NPAPI Object.
1685
1686         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1687         (testCallbackReturn):
1688         (pluginInvoke):
1689
1690 2010-11-09  Sheriff Bot  <webkit.review.bot@gmail.com>
1691
1692         Unreviewed, rolling out r71672.
1693         http://trac.webkit.org/changeset/71672
1694         https://bugs.webkit.org/show_bug.cgi?id=49276
1695
1696         It doesn't work on Chromium Windows bot (Requested by Ossy on
1697         #webkit).
1698
1699         * Scripts/webkitpy/common/system/file_lock.py: Removed.
1700         * Scripts/webkitpy/common/system/file_lock_unittest.py: Removed.
1701         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1702
1703 2010-11-09  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1704
1705         Reviewed by Tony Chang.
1706
1707         [NRWT] Make http locking similar to perl implementation
1708         https://bugs.webkit.org/show_bug.cgi?id=49187
1709
1710         * Scripts/webkitpy/common/system/file_lock.py: Added.
1711         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
1712         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1713
1714 2010-11-09  James Kozianski  <koz@chromium.org>
1715
1716         Reviewed by Eric Seidel.
1717
1718         Clean up imports for webkit-patch.
1719         https://bugs.webkit.org/show_bug.cgi?id=49083
1720
1721         Move some imports from commands into main.py where they are used.
1722         Move command imports into the __init__ of the commands module.
1723
1724         * Scripts/webkitpy/tool/commands/__init__.py:
1725         * Scripts/webkitpy/tool/commands/download.py:
1726         * Scripts/webkitpy/tool/commands/queues.py:
1727         * Scripts/webkitpy/tool/main.py:
1728
1729 2010-11-09  Yi Shen  <yi.4.shen@nokia.com>
1730
1731         Reviewed by Andreas Kling.
1732
1733         [Qt][QtTestBrowser] Don't erase incorrect url in the Url Bar
1734         https://bugs.webkit.org/show_bug.cgi?id=49047
1735
1736         * QtTestBrowser/mainwindow.cpp:
1737         (MainWindow::setAddressUrl):
1738
1739 2010-11-09  Andras Becsi  <abecsi@webkit.org>
1740
1741         Reviewed by Csaba Osztrogonác.
1742
1743         Improve old-run-webkit-tests --verbose to show DumpTool opening/closing
1744         https://bugs.webkit.org/show_bug.cgi?id=49252
1745
1746         * Scripts/old-run-webkit-tests:
1747
1748 2010-11-09  Csaba Osztrogonác  <ossy@webkit.org>
1749
1750         Reviewed by Tor Arne Vestbø.
1751
1752         Output of old-run-webkit-tests is buggy in verbose mode
1753         https://bugs.webkit.org/show_bug.cgi?id=49249
1754
1755         * Scripts/old-run-webkit-tests: Redundant outputs removed.
1756
1757 2010-11-08  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
1758
1759         Reviewed by Martin Robinson.
1760
1761         [GTK] Link with target name set does not work
1762         https://bugs.webkit.org/show_bug.cgi?id=48865
1763
1764         When a new page is created with a name (target=myFrame), the new
1765         mainFrame could not be found because they where not stored in the
1766         same PageGroup. As PageGroup are not exposed externally so the
1767         simpliest solution is to use a global page group name. This also fixes
1768         issue with visited link coloration across pages. After this change the
1769         private function webkit_web_view_set_group_name() was no longer used
1770         so it was removed completly.
1771
1772         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1773         (createWebView):
1774
1775 2010-11-08  John Knottenbelt  <jknotten@chromium.org>
1776
1777         Reviewed by Kent Tamura.
1778
1779         WebViewHost::reset() uses placement new.
1780         https://bugs.webkit.org/show_bug.cgi?id=49069
1781
1782         WebViewHost is using placement destruction / new to simulate a fresh
1783         WebViewHost object at the same address. This is because the WebView remains
1784         open across tests and maintains a pointer to the WebViewHost.
1785         This change resets member variables explictly instead of the placement new dance.
1786
1787         * DumpRenderTree/chromium/WebViewHost.cpp:
1788         (WebViewHost::WebViewHost):
1789         (WebViewHost::reset):
1790
1791 2010-11-08  Dirk Pranke <dpranke@chromium.org>
1792
1793         Reviewed by Eric Seidel.
1794
1795         Fix the webkit_base_dir logic in webkitpy/layout_tests/port/config.py. 
1796         It turns out that NRWT can't use the code in scm.find_checkout_root()
1797         because the Chromium bots don't do full checkouts of the WebKit
1798         tree; they only check out subdirectories like WebKitTools/Scripts.
1799         Until we can figure out a better approach for this, I've
1800         restored the base_dir-detecting code from NRWT, which works in
1801         any directory tree, scm or no.
1802
1803         This also restores the files modified in r71475 and r71474.
1804
1805         https://bugs.webkit.org/show_bug.cgi?id=49151
1806
1807         * WebKitTools/Scripts/webkitpy/layout_tests/port/config.py
1808         * WebKitTools/Scripts/webkitpy/layout_tests/port/config_unittest.py
1809
1810 2010-11-08  Eric Seidel  <eric@webkit.org>
1811
1812         Reviewed by Adam Barth.
1813
1814         webkit-patch failure-reason explodes if a build is missing
1815         https://bugs.webkit.org/show_bug.cgi?id=49195
1816
1817         This is likely a recent regression.  Adding a unit test for this case.
1818
1819         * Scripts/webkitpy/common/net/buildbot.py:
1820         * Scripts/webkitpy/common/net/buildbot_unittest.py:
1821
1822 2010-11-08  Dan Bernstein  <mitz@apple.com>
1823
1824         Reviewed by Darin Adler.
1825
1826         Ignore the system scrollbar setting.
1827
1828         * DumpRenderTree/mac/DumpRenderTree.mm:
1829         (resetDefaultsToConsistentValues): Set the scrollbar default.
1830
1831 2010-11-05  Adam Roben  <aroben@apple.com>
1832
1833         Make webkitpy.common.system.executive_unittest pass when running under
1834         Win32 Python
1835
1836         Fixes <http://webkit.org/b/49033>.
1837
1838         Reviewed by Dave Levin and Eric Seidel.
1839
1840         * Scripts/webkitpy/common/system/executive.py:
1841         (Executive._run_command_with_teed_output): Pass the arguments through
1842         encode_argument_if_needed rather than using Cygwin-specific code here.
1843         (Executive.run_and_throw_if_fail): Use child_process_encoding to decode
1844         the output.
1845         (Executive.run_command): Use encode_argument_if_needed to encode the
1846         arguments and child_process_encoding to decode the output.
1847         (Executive._child_process_encoding): Returns the encoding that should be
1848         used when communicating with child processes. On Windows we use mbcs,
1849         which maps to the current code page. On all other platforms we use
1850         UTF-8.
1851         (Executive._should_encode_child_process_arguments): Returns True if
1852         arguments to child processes need to be encoded. This is currently
1853         only needed on Cygwin and Win32 Python 2.x.
1854         (Executive._encode_argument_if_needed): Encode the argument using
1855         child_process_encoding if we need to encode arguments to child
1856         processes on this platform.
1857
1858         * Scripts/webkitpy/common/system/executive_unittest.py:
1859         (never_ending_command): Added. Returns arguments to run a command that
1860         will not quit until we kill it. On Windows we use wmic, on other
1861         platforms we use yes.
1862         (ExecutiveTest.test_run_command_with_unicode): Changed to expect the
1863         mbcs encoding to be used and for output from the child processes to
1864         have been roundtripped through encode/decode on Win32 Python. When
1865         UTF-8 is the encoding the roundtripping is undetectable, but with mbcs
1866         it's possible that some characters will not be able to be converted
1867         and will be replaced by question marks; the round-tripping allows us
1868         to expect this result.
1869
1870         (ExecutiveTest.test_kill_process):
1871         (ExecutiveTest.test_kill_all):
1872         Use never_ending_command instead of invoking "yes" directly. Expect an
1873         exit code of 1 when using Win32 Python, as that's what seems to happen.
1874
1875 2010-11-08  Adam Roben  <aroben@apple.com>
1876
1877         Roll out r71532
1878
1879         It broke the build for Cygwin 1.7 installs. Cygwin 1.7's default
1880         .bashrc unsets %TEMP%, which broke copy-tools.cmd.
1881
1882         * Scripts/webkitdirs.pm:
1883
1884 2010-11-08  Tony Chang  <tony@chromium.org>
1885
1886         Reviewed by Adam Barth.
1887
1888         run platform/chromium/plugins/nested-plugin-objects.html on all platforms
1889         https://bugs.webkit.org/show_bug.cgi?id=49094
1890
1891         This tests that objects created by plugins are proplery cleaned up.
1892
1893         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1894         (testPassTestObject):
1895         (pluginInvoke):
1896         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
1897         (testAllocate):
1898         (testDeallocate):
1899         (testGetProperty):
1900         (testConstruct):
1901
1902 2010-11-08  Adam Roben  <aroben@apple.com>
1903
1904         Mark Windows builds triggered from Perl as being non-interactive
1905
1906         This affects whether some of our scripts will show alerts vs. printing
1907         to the build log.
1908
1909         Fixes <http://webkit.org/b/49181> Windows build fail mysteriously when
1910         .vsprops files are updated
1911
1912         Reviewed by Steve Falkenburg.
1913
1914         * Scripts/webkitdirs.pm:
1915         (buildVisualStudioProject): Set WEBKIT_NONINTERACTIVE_BUILD to 1.
1916
1917 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1918
1919         Reviewed by Ojan Vafai.
1920
1921         Make http locking default in NRWT.
1922         https://bugs.webkit.org/show_bug.cgi?id=48053
1923
1924         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1925         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1926
1927 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1928
1929         Reviewed by Csaba Osztrogonác.
1930
1931         [NRWT] If the http lock fails we shouldn't do any locking
1932         https://bugs.webkit.org/show_bug.cgi?id=49164
1933
1934         If something goes wrong with the locking, the test should keep going.
1935
1936         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1937
1938 2010-11-08  Adam Roben  <aroben@apple.com>
1939
1940         Switch back to using kCGImageAlphaPremultipliedFirst when generating
1941         pixel dumps on Windows
1942
1943         I changed this behavior in r71418 thinking that it was required for
1944         getting plugins to show up in pixel dumps. But it doesn't seem to be
1945         necessary, and was making it impossible to compare new Windows pixel
1946         dumps with existing Windows or Mac pixel dumps (because ImageDiff won't
1947         compare an image with alpha to an image without alpha).
1948
1949         Fixes <http://webkit.org/b/49172> REGRESION (r71418): Can't compare
1950         new Windows pixel results to existing Windows or Mac results
1951
1952         Reviewed by Antti Koivisto.
1953
1954         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
1955         (createBitmapContextFromWebView): Replaced kCGImageAlphaNoneSkipFirst
1956         with kCGImageAlphaPremultipliedFirst.
1957
1958 2010-11-08  Csaba Osztrogonac  <ossy@webkit.org>
1959
1960         Unreviewed, rolling out r71466.
1961         http://trac.webkit.org/changeset/71466
1962         https://bugs.webkit.org/show_bug.cgi?id=48865
1963
1964         It broke layout tests on GTK bots.
1965
1966         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1967         (createWebView):
1968
1969 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1970
1971         Reviewed by Csaba Osztrogonác.
1972
1973         Enable running of Qt API tests on BuildBot
1974         https://bugs.webkit.org/show_bug.cgi?id=49004
1975
1976         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1977
1978 2010-11-08  Eric Seidel  <eric@webkit.org>
1979
1980         Reviewed by Adam Barth.
1981
1982         Add clean-review-queue command to remove closed bugs from the webkit.org/pending-review
1983         https://bugs.webkit.org/show_bug.cgi?id=49160
1984
1985         Bugzilla doesn't automatically remove r? when a bug gets closed.
1986         This script takes care of that for webkit.org.
1987
1988         * Scripts/webkitpy/common/net/bugzilla.py:
1989         * Scripts/webkitpy/tool/commands/upload.py:
1990
1991 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
1992
1993         Unreviewed, rolling out r71474.
1994         http://trac.webkit.org/changeset/71474
1995         https://bugs.webkit.org/show_bug.cgi?id=48280
1996
1997         breaks chromium webkit tests
1998         https://bugs.webkit.org/show_bug.cgi?id=49151
1999
2000         * Scripts/webkitpy/layout_tests/port/base.py:
2001         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2002         * Scripts/webkitpy/layout_tests/port/config.py:
2003         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
2004         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
2005         * Scripts/webkitpy/layout_tests/port/webkit.py:
2006         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2007         * Scripts/webkitpy/tool/mocktool.py:
2008
2009 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
2010
2011         Unreviewed, rolling out r71475.
2012         http://trac.webkit.org/changeset/71475
2013  
2014         breaks chromium webkit tests
2015         https://bugs.webkit.org/show_bug.cgi?id=49151
2016
2017         * Scripts/webkitpy/common/newstringio.py: Removed.
2018         * Scripts/webkitpy/common/newstringio_unittest.py: Removed.
2019         * Scripts/webkitpy/common/system/executive_mock.py: Removed.
2020         * Scripts/webkitpy/common/system/filesystem_mock.py: Removed.
2021         * Scripts/webkitpy/layout_tests/port/config_mock.py: Removed.
2022
2023 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2024
2025         Unreviewed, build fix.
2026
2027         Add files inexplicably not committed in r71474 as part of the
2028         fix for bug 48280.
2029
2030         * Scripts/webkitpy/common/newstringio.py: Added.
2031         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
2032         * Scripts/webkitpy/common/system/executive_mock.py: Added.
2033         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
2034         * Scripts/webkitpy/layout_tests/port/config_mock.py: Added.
2035
2036 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2037
2038         Reviewed by Eric Siedel.
2039
2040         new-run-webkit-tests: update port/base and port/webkit to use the
2041         new FileSystem and Config abstractions, pulling more logic out of
2042         the base Port classes into separate, mockable objects.
2043
2044         Also create a MockFileSystem object, a MockConfig object, move
2045         MockExecutive into common/system to be next to executive, and
2046         update the config object to use a FileSystem.
2047
2048         https://bugs.webkit.org/show_bug.cgi?id=48280
2049
2050         * Scripts/webkitpy/common/newstringio.py: Added.
2051         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
2052         * Scripts/webkitpy/common/system/executive_mock.py: Added.
2053         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
2054         * Scripts/webkitpy/layout_tests/port/base.py:
2055         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2056         * Scripts/webkitpy/layout_tests/port/config.py:
2057         * Scripts/webkitpy/layout_tests/port/config_mock.py:
2058         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
2059         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
2060         * Scripts/webkitpy/layout_tests/port/webkit.py:
2061         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2062         * Scripts/webkitpy/tool/mocktool.py:
2063
2064 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2065
2066         Unreviewed, build breakage.
2067
2068         Apparently I uploaded the wrong version of the file to fix 49122
2069         and neither Eric or I noticed - it was missing a dirname() call.
2070         Fixing ...
2071
2072         https://bugs.webkit.org/show_bug.cgi?id=49122
2073
2074         * Scripts/webkitpy/common/checkout/scm.py:
2075
2076 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2077
2078         Reviewed by Eric Seidel.
2079
2080         webkitpy/tool/* unittests change cwd and don't clean up properly
2081
2082         https://bugs.webkit.org/show_bug.cgi?id=49122
2083
2084         * Scripts/webkitpy/common/checkout/scm.py:
2085
2086 2010-11-05  Chris Marrin  <cmarrin@apple.com>
2087
2088         Reviewed by Simon Fraser.
2089
2090         suspendAnimations/resumeAnimations not present in WebKit2
2091         https://bugs.webkit.org/show_bug.cgi?id=49109
2092
2093         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
2094         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
2095         (WTR::LayoutTestController::suspendAnimations):
2096         (WTR::LayoutTestController::resumeAnimations):
2097         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
2098
2099 2010-11-05  Tony Chang  <tony@chromium.org>
2100
2101         Reviewed by David Levin.
2102
2103         cleanup style in TestNetscapePlugIn/PluginObject.cpp
2104         https://bugs.webkit.org/show_bug.cgi?id=49044
2105
2106         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
2107         (getPluginClass):
2108         (pluginGetProperty):
2109         (pluginSetProperty):
2110         (variantToIdentifier):
2111         (testCallback):
2112         (getURL):
2113         (getURLNotify):
2114         (testInvokeDefault):
2115         (testGetProperty):
2116         (testDocumentOpen):
2117         (testWindowOpen):
2118         (pluginInvoke):
2119         (handleCallback):
2120         (notifyStream):
2121         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
2122
2123 2010-11-05  Eric Seidel  <eric@webkit.org>
2124
2125         Reviewed by David Levin.
2126
2127         Add QueueStatusServer/__init__.py so others can run the QueueStatusServer tests
2128         https://bugs.webkit.org/show_bug.cgi?id=49032
2129
2130         I wrote this file as part of bug 47847, but I forgot to commit it.
2131         No one else noticed it missing because test-webkitpy knows how
2132         to recover in the case where it can't import QueueStatusServer
2133         (which generally occurs due to not having installed the AppEngine SDK).
2134
2135         * QueueStatusServer/__init__.py: Added.
2136         * QueueStatusServer/model/workitems_unittest.py:
2137          - Remove a test which fails.  This was probably landed (by me)
2138            from my other machine, which since this __init__.py was missing
2139            I never noticed the failure and landed this invalid test.
2140            Sadly we can't really test remove_work_item as it depends
2141            on .key() working.  .key() will throw unless the object
2142            has already been saved it seems.
2143            This may be a mis-design in our remove_work_item implementation,
2144            but for now, just removing the test.
2145
2146 2010-11-04  Adam Roben  <aroben@apple.com>
2147
2148         Teach check-webkit-style about TestNetscapePlugIn
2149
2150         Fixes <http://webkit.org/b/49030> check-webkit-style is ignorant of
2151         TestNetscapePlugIn's coding conventions
2152
2153         Reviewed by John Sullivan.
2154
2155         * Scripts/webkitpy/style/checker.py: Lump TestNetscapePlugIn in with
2156         WebKitAPITest and TestWebKitAPI in having different include order and
2157         naming conventions than WebCore. Added some comments about why the
2158         exceptions exist.
2159
2160 2010-11-04  Adam Roben  <aroben@apple.com>
2161
2162         Add a test that shows that windowed plugins are able to paint
2163
2164         Somehow we never had a test for this before.
2165
2166         Fixes <http://webkit.org/b/49029> Should add a test that shows
2167         windowed plugins can paint
2168
2169         Reviewed by Jon Honeycutt.
2170
2171         * DumpRenderTree/TestNetscapePlugIn/Tests/win/DrawsGradient.cpp: Added.
2172         (DrawsGradient::DrawsGradient):
2173         (DrawsGradient::wndProc): We handle the WM_PAINT and WM_PRINTCLIENT messages.
2174
2175         (DrawsGradient::onPaint):
2176         (DrawsGradient::onPrintClient):
2177         These both just call through to paint.
2178
2179         (DrawsGradient::paint): Fills our client area with some gradients.
2180
2181         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Link
2182         against Msimg32.lib for ::GradientFill and added DrawsGradient.
2183
2184         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
2185         (createBitmapContextFromWebView): Use WM_PRINT instead of
2186         WM_PRINTCLIENT so that ::DefWindowProc will send
2187         WM_PRINT/WM_PRINTCLIENT messages to the WebView's child windows.
2188         Replaced kCGImageAlphaPremultipledFirst with kCGImageAlphaNoneSkipFirst
2189         because GDI doesn't support alpha and kCGImageBitmapOrder32Little with
2190         kCGImageBitmapOrder32Host because, who knows, maybe someday Windows
2191         will run on a big-endian platform.
2192
2193 2010-11-04  Adam Roben  <aroben@apple.com>
2194
2195         Extract much of NPNInvalidateRectInvalidatesWindow's code into a
2196         WindowedPluginTest base class
2197
2198         The base class takes care of subclassing the plugin's window so that a
2199         custom WNDPROC is called. This will make it easier to write tests that
2200         need to handle window messages.
2201
2202         Fixes <http://webkit.org/b/49028> It's hard to write a PluginTest with
2203         a custom WNDPROC
2204
2205         Reviewed by Jon Honeycutt.
2206
2207         * DumpRenderTree/TestNetscapePlugIn/Tests/win/NPNInvalidateRectInvalidatesWindow.cpp:
2208         (NPNInvalidateRectInvalidatesWindow::NPNInvalidateRectInvalidatesWindow):
2209         (NPNInvalidateRectInvalidatesWindow::NPP_SetWindow):
2210         (NPNInvalidateRectInvalidatesWindow::wndProc):
2211         (NPNInvalidateRectInvalidatesWindow::testInvalidateRect):
2212         Moved code from here to WindowedPluginTest. Changed to use window()
2213         instead of m_window.
2214
2215         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2216         Added WindowedPluginTest and added TestNetscapePlugIn/win to the
2217         include path. Also reordered the include path to make a little more
2218         sense and simplified the entry that added TestNetscapePlugIn itself to
2219         the include path.
2220
2221         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.cpp: Added.
2222         (WindowedPluginTest::WindowedPluginTest):
2223         (WindowedPluginTest::NPP_SetWindow):
2224         (WindowedPluginTest::staticWndProc):
2225         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.h: Added.
2226         (WindowedPluginTest::window):
2227         Code came from NPNInvalidateRectInvalidatesWindow.
2228
2229 2010-11-04  Adam Roben  <aroben@apple.com>
2230
2231         Add a plugin test to show that windowed plugins are clipped correctly
2232
2233         Fixes <http://webkit.org/b/49024> <rdar://problem/8487847> Windowed
2234         plugins aren't clipped in WebKit2 on Windows
2235
2236         Reviewed by Jon Honeycutt.
2237
2238         * DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowRegionIsSetToClipRect.cpp: Added.
2239         (WindowRegionIsSetToClipRect::WindowRegionIsSetToClipRect): Initialize members.
2240         (WindowRegionIsSetToClipRect::NPP_SetWindow): Check that our window
2241         region matches the clip rect we know we should have based on
2242         window-region-is-set-to-clip-rect.html, and check that our window class
2243         doesn't have the CS_PARENTDC style.
2244
2245         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2246         Added WindowRegionIsSetToClipRect.
2247
2248 2010-11-05  Alejandro G. Castro  <alex@igalia.com>
2249
2250         Reviewed by Martin Robinson.
2251
2252         [GTK] Avoid font hinting in the DRT
2253         https://bugs.webkit.org/show_bug.cgi?id=48548
2254
2255         Change the settings to avoid font hinting, it was causing
2256         different results depending on the distribution.
2257
2258         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2259         (initializeGtkFontSettings):
2260         * DumpRenderTree/gtk/fonts/fonts.conf:
2261
2262 2010-11-05  Eric Seidel  <eric@webkit.org>
2263
2264         Reviewed by Adam Barth.
2265
2266         build-webkit spams system.log by repeatedly checking for PBXProductDirectory in com.apple.Xcode
2267         https://bugs.webkit.org/show_bug.cgi?id=49051
2268
2269         This is a speculative fix.  The unit tests cover these methods, however
2270         I don't know if this will fully stop the system.log spam.
2271
2272         * Scripts/webkitpy/layout_tests/port/base.py:
2273         * Scripts/webkitpy/layout_tests/port/webkit.py:
2274
2275 2010-11-05  Eric Seidel  <eric@webkit.org>
2276
2277         Reviewed by Adam Barth.
2278
2279         Add basic support for showing bot id on /queue-status/ pages
2280         https://bugs.webkit.org/show_bug.cgi?id=49037
2281
2282         This support is really simple.  Eventually we'll want to
2283         show the bot id in the lock table too, but we don't have
2284         that information stored in the server yet.
2285
2286         * QueueStatusServer/handlers/queuestatus.py:
2287         * QueueStatusServer/handlers/queuestatus_unittest.py: Added.
2288         * QueueStatusServer/templates/includes/singlequeuestatus.html:
2289
2290 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
2291
2292         Reviewed by Adam Barth.
2293
2294         Create a filesystem wrapper that we can use to enforce
2295         particular conventions and use for mocking and dependency
2296         injection down the line.
2297
2298         https://bugs.webkit.org/show_bug.cgi?id=48144
2299
2300         * Scripts/webkitpy/common/system/filesystem.py: Added.
2301         * Scripts/webkitpy/common/system/filesystem_unittest.py: Added.
2302
2303 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2304
2305         Reviewed by Tony Chang.
2306
2307         run_webkit_tests_unittest fails under Python 2.5
2308         https://bugs.webkit.org/show_bug.cgi?id=49043
2309         
2310         Switch from itertools.chain.from_iterable (which was added in 2.6)
2311         to using itertools.chain directly.
2312
2313         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2314
2315 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2316
2317         Unreviewed fix to webkit-patch rebaseline-server so that it runs under
2318         Python 2.5 (it needs an import to use the with statement).
2319
2320         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
2321
2322 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
2323
2324         Reviewed by Adam Barth.
2325
2326         new-run-webkit-tests: split out webkit-specific configuration stuff into a new module
2327
2328         The current NRWT code has webkit-specific configuration code (like
2329         _script_path, default configuration, etc.) mixed in with
2330         layout-test-specific stuff in port/base. The configuration code
2331         should be split out into a separate module for easier mocking,
2332         testing, and isolation.
2333
2334         https://bugs.webkit.org/show_bug.cgi?id=48264
2335
2336         * Scripts/webkitpy/layout_tests/port/config.py: Added.
2337         * Scripts/webkitpy/layout_tests/port/config_unittest.py: Added.
2338
2339 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2340
2341         Reviewed by Tony Chang.
2342
2343         Rebaseline server: initial framework
2344         https://bugs.webkit.org/show_bug.cgi?id=48892
2345         
2346         Adds the basic framework for the rebaseline server (details at 
2347         http://webkit.org/b/47761). Includes the rebaseline-server webkit-patch
2348         command, which starts an HTTP server that can serve static files or
2349         invoke handler methods on a class.
2350
2351         * Scripts/webkitpy/tool/commands/__init__.py:
2352         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html: Added.
2353         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css: Added.
2354         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js: Added.
2355         * Scripts/webkitpy/tool/commands/rebaselineserver.py: Added.
2356
2357 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
2358
2359         Reviewed by James Robinson.
2360
2361         new-run-webkit-tests wasn't using DRT by default for
2362         --platform chromium-gpu
2363
2364         The default value was set to False instead of None, which meant
2365         that the platform specific logic wasn't firing to change the
2366         value to True (b/c we were afraid we'd be overriding the user
2367         preference).
2368
2369         https://bugs.webkit.org/show_bug.cgi?id=49038
2370
2371         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2372
2373 2010-11-04  Tony Chang  <tony@chromium.org>
2374
2375         Reviewed by Adam Barth.
2376
2377         make platform/chromium/plugins/multiple-plugins.html pass on all platforms
2378         https://bugs.webkit.org/show_bug.cgi?id=49026
2379
2380         Have the testObject.bar property return the string "bar".  This lets
2381         us run and pass platform/chromium/plugins/multiple-plugins.html on all
2382         platforms.
2383
2384         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
2385         (testGetProperty):
2386
2387 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2388
2389         Reviewed by Tony Chang.
2390
2391         NRWT doesn't have good test coverage for --run-chunk, --batch-size, --run-part, etc.
2392         https://bugs.webkit.org/show_bug.cgi?id=48878
2393         
2394         Add get_tests_run so that it's easy to see which tests get run (and with
2395         what batching) for a given flag combination. Flesh out the various
2396         test cases that have FIXMEs.
2397         
2398         Also fixes an off-by-one error (batch sizes were one larger than 
2399         expected) and makes --run-part also have wraparound behavior, like
2400         --run-chunk.
2401
2402         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2403         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2404         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2405
2406 2010-11-04  Erik Arvidsson  <arv@chromium.org>
2407
2408         Reviewed by Dimitri Glazkov.
2409
2410         Support box-sizing without the vendor prefix
2411         https://bugs.webkit.org/show_bug.cgi?id=36713
2412
2413         Based on patch by Peter Beverloo <peter@lvp-media.com>
2414
2415         * iExploder/htdocs/cssproperties.in: Change -webkit-box-sizing to box-sizing.
2416
2417 2010-11-04  Csaba Osztrogonác  <ossy@webkit.org>
2418
2419         Unreviewed rollout r71340, because it broke Chromium Windows bot.
2420
2421         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2422         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2423
2424 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2425
2426         Reviewed by Ojan Vafai.
2427
2428         Make http locking default in NRWT.
2429         https://bugs.webkit.org/show_bug.cgi?id=48053
2430
2431         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2432         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2433
2434 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2435
2436         Reviewed by Eric Seidel.
2437
2438         [NRWT] Clear invalid http locks on Windows platform as well
2439         https://bugs.webkit.org/show_bug.cgi?id=48515
2440
2441         * Scripts/webkitpy/common/system/executive.py:
2442         * Scripts/webkitpy/common/system/executive_unittest.py:
2443         * Scripts/webkitpy/layout_tests/port/http_lock.py:
2444
2445 2010-11-02  Adam Roben  <aroben@apple.com>
2446
2447         Reduce our dependence on coreutils when running Python tests
2448
2449         This patch introduces versions of the standard echo and cat utilities
2450         implemented in Python. They are probably missing features of their
2451         coreutils equivalents, but they can do what's necessary for our Python
2452         tests. This is useful on Windows, where these utilities typically
2453         aren't available.
2454
2455         Fixes <http://webkit.org/b/48883> executive_unittest relies on echo
2456         and cat utilities from coreutils, which are not present on Windows
2457
2458         Reviewed by Eric Seidel.
2459
2460         * Scripts/webkitpy/common/system/executive_unittest.py: Changed to use
2461         our Python-based echo and cat.
2462
2463         * Scripts/webkitpy/common/system/fileutils.py: Added.
2464         (make_stdout_binary): On Windows, puts sys.stdout into binary mode so
2465         that \n won't be translated into \r\n. I couldn't think of a good way
2466         to test this directly without touching the filesystem, but it is tested
2467         indirectly by echo_unittest.
2468
2469         * Scripts/webkitpy/test/cat.py: Added.
2470         (command_arguments): Returns a list for invoking cat with the given arguments.
2471         (main): Acts like a simplified version of the coreutils cat utility.
2472
2473         * Scripts/webkitpy/test/cat_unittest.py: Added.
2474         (CatTest.assert_cat): Runs cat with the given input and ensures the
2475         output matches the input.
2476         (CatTest.test_basic): Performs a simple test of cat.
2477         (CatTest.test_no_newline): Tests what happens when the input string
2478         doesn't have a trailing newline.
2479         (CatTest.test_unicode): Tests passing a unicode string to cat.
2480         (CatTest.test_as_command): Tests running cat as a separate command.
2481
2482         * Scripts/webkitpy/test/echo.py: Added.
2483         (command_arguments): Returns a list for invoking echo with the given arguments.
2484         (main): Acts like a simplified version of the coreutils echo utility.
2485
2486         * Scripts/webkitpy/test/echo_unittest.py: Added.
2487         (EchoTest.test_basic): Performs a simple test of echo.
2488         (EchoTest.test_no_newline): Tests passing -n to echo to suppress the
2489         trailing newline.
2490         (EchoTest.test_unicode): Tests passing unicode and non-unicode strings
2491         to echo.
2492         (EchoTest.test_argument_order): Tests what happens when -n is not the
2493         first argument.
2494         (EchoTest.test_empty_arguments): Tests what happens when you pass [] to
2495         echo.main.
2496         (EchoTest.test_no_arguments): Tests what happens when you call
2497         echo.main with no arguments.
2498         (EchoTest.test_as_command): Tests running echo as a separate command.
2499
2500 2010-11-04  Renata Hodovan  <reni@inf.u-szeged.hu>
2501
2502         Unreviewed: Add myself to the list of Committers.
2503
2504         * Scripts/webkitpy/common/config/committers.py:
2505
2506 2010-11-04  Andreas Kling  <kling@webkit.org>
2507
2508         Reviewed by Kenneth Rohde Christiansen.
2509
2510         [Qt] Hook into QWebPage::scrollRequested for resizes-to-contents mode
2511
2512         In RTC mode the QGraphicsWebView item is the size of the contents,
2513         scrolling works a bit differently  (we need to react to scrollRequested.)
2514
2515         Normally QGraphicsView will replay the last mouse event when scrolling,
2516         so to prevent WebKit from getting confused by this we temporarily make
2517         the QGraphicsView non-interactive.
2518
2519         * QtTestBrowser/webview.cpp:
2520         (WebViewGraphicsBased::setPage):
2521         (WebViewGraphicsBased::scrollRequested):
2522         * QtTestBrowser/webview.h:
2523
2524 2010-11-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2525
2526         Reviewed by Andreas Kling.
2527
2528         Use OS(MAC_OS_X) rather than PLATFORM(MAC) in TestControllerQt
2529
2530         PLATFORM(MAC) is not defined for the Qt port, as it refers to
2531         the Mac-port, not the Mac OS X operating system.
2532
2533         * WebKitTestRunner/qt/TestControllerQt.cpp:
2534         (WTR::isExistingLibrary):
2535
2536 2010-11-04  Leonid Ebril  <leonid.ebril@nokia.com>
2537
2538         Reviewed by Andreas Kling.
2539
2540         [Qt] Adding iphone user agent string the user agent list for QtTestBrowser
2541         https://bugs.webkit.org/show_bug.cgi?id=48869
2542
2543         * QtTestBrowser/useragentlist.txt:
2544
2545 2010-11-03  Adam Roben  <aroben@apple.com>
2546
2547         Always use uppercase drive names in strings returned by abspath_to_uri
2548
2549         Some versions of cygpath use lowercase drive letters while others use
2550         uppercase, which makes it hard to test the output of code that uses
2551         cygpath.
2552
2553         Fixes <http://webkit.org/b/48914> webkitpy.common.system.path_unittest
2554         fails with Cygwin 1.5
2555
2556         Reviewed by Eric Seidel.
2557
2558         * Scripts/webkitpy/common/system/path.py:
2559         (cygpath): Updated the docstring to indicate that only absolute paths
2560         should be passed for now (though relative paths will work fine).
2561         (_Cygpath.convert): Upper-case the first letter of the converted Windows path.
2562
2563 2010-11-03  George Guo  <George.Guo@Nokia.com>
2564
2565         Reviewed by Andreas Kling.
2566
2567         [Qt] QtTestBrowser : set mmp rule pageddata in Symbian
2568         https://bugs.webkit.org/show_bug.cgi?id=48767
2569     
2570         Paging is needd on Symbian devices to support benchmarks tests like
2571         dromaeo.com and Celtic Kane that need a lot of memory to run  
2572
2573         * QtTestBrowser/QtTestBrowser.pro:
2574
2575 2010-11-03  Jenn Braithwaite  <jennb@chromium.org>
2576
2577         Reviewed by Dmitry Titov.
2578
2579         Chromium: Update resource tracking when moving a frame between documents
2580         https://bugs.webkit.org/show_bug.cgi?id=48363
2581
2582         * DumpRenderTree/chromium/WebViewHost.cpp:
2583         (WebViewHost::assignIdentifierToRequest):
2584         Always put resource id in map so we can make assumptions about its
2585         presence.
2586         (WebViewHost::removeIdentifierForRequest):
2587         Added.
2588         (WebViewHost::didFinishResourceLoad):
2589         (WebViewHost::didFailResourceLoad):
2590         Check existence of resource id before removing from map.
2591         * DumpRenderTree/chromium/WebViewHost.h:
2592
2593 2010-11-03  Victor Wang  <victorw@chromium.org>
2594
2595         Reviewed by Adam Barth.
2596
2597         [Chromium] update buildbot names in chromium rebaseline tool.
2598
2599         https://bugs.webkit.org/show_bug.cgi?id=48881
2600
2601         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2602
2603 2010-11-03  Darin Adler  <darin@apple.com>
2604
2605         Updated Xcode projects by opening them with Xcode 3.2.4.
2606         Updated svn:ignore for Xcode projects.
2607
2608         * MiniBrowser/MiniBrowser.xcodeproj: Added property svn:ignore.
2609         * TestWebKitAPI/TestWebKitAPI.xcodeproj: Added property svn:ignore.
2610         * WebKitLauncher/WebKitLauncher.xcodeproj: Modified property svn:ignore.
2611         * WebKitLauncher/WebKitLauncher.xcodeproj/project.pbxproj: Updated with Xcode 3.2.4.
2612
2613 2010-11-03  Adam Roben  <aroben@apple.com>
2614
2615         Stop waiting for messages from the web process after a timeout elapses
2616
2617         TestController::runUntil now takes a timeout parameter that specifies
2618         how long to wait for the condition to become true. WebKitTestRunner
2619         takes a --timeout flag that specifies how long the timeout should be.
2620         run-webkit-tests passes this flag to WebKitTestRunner so its timeout
2621         will be similar to run-webkit-tests.
2622
2623         Fixes <http://webkit.org/b/43047> <rdar://problem/8365833>
2624         WebKitTestRunner waits forever if the web process crashes
2625
2626         Reviewed by Darin Adler and Anders Carlsson.
2627
2628         * Scripts/old-run-webkit-tests:
2629         (top level): Moved the GuardMalloc timeout adjustment here from
2630         readFromDumpToolWithTimer.
2631         (openDumpTool): Make WTR use a timeout similar to but slightly shorter
2632         than the one that was specified on the command line.
2633
2634         * WebKitTestRunner/TestController.cpp:
2635         (WTR::TestController::TestController): Initialize our timeout values.
2636         (WTR::TestController::initialize): Parse the --timeout flag and use it
2637         to modify our timeout values.
2638         (WTR::TestController::resetStateToConsistentValues): Changed to use a
2639         short timeout while waiting for the web process to reset and to return
2640         a boolean indicating whether we were able to reset the web process.
2641         Uses a 5-second timeout while waiting for the process to be reset.
2642         (WTR::TestController::runTest): Changed to return a boolean indicating
2643         whether we were able to reset the web process (and thus run the test).
2644         (WTR::TestController::runUntil): Call through to platformRunUntil.
2645
2646         (WTR::TestController::runTestingServerLoop):
2647         (WTR::TestController::run):
2648         Changed to bail if any test can't be run. This will cause the process
2649         to exit. (Unfortunately this will make run-webkit-tests think we
2650         crashed; see <http://webkit.org/b/48943>.)
2651
2652         * WebKitTestRunner/TestController.h: Added platformRunUntil,
2653         m_longTimeout, and m_shortTimeout.
2654
2655         * WebKitTestRunner/TestInvocation.cpp:
2656         (WTR::TestInvocation::invoke): Use a short timeout when waiting for the
2657         initial response and a long timeout when waiting for the test to
2658         complete. Check whether runUntil timed out and print an error message
2659         if so.
2660
2661         * WebKitTestRunner/mac/TestControllerMac.mm:
2662         (WTR::TestController::platformRunUntil): Renamed from runUntil. Pass
2663         [NSDate distantPast] to -[NSRunLoop runMode:beforeDate:] so that we
2664         won't block waiting for the run loop. Only loop until the timeout
2665         elapses.
2666
2667         * WebKitTestRunner/qt/TestControllerQt.cpp:
2668         (WTR::TestController::platformRunUntil): Renamed from runUntil. Added a
2669         FIXME about honoring the timeout.
2670
2671         * WebKitTestRunner/win/TestControllerWin.cpp:
2672         (WTR::TestController::platformRunUntil): Renamed from runUntil. Use
2673         ::MsgWaitForMultipleObjectsEx to implement the timeout. Changed to use
2674         ::PeekMessageW so that we don't block waiting for messages to become
2675         available.
2676
2677 2010-11-03  Adam Roben  <aroben@apple.com>
2678
2679         Add a plugin test that evaluates JS after removing the plugin element
2680         from the document
2681
2682         This test replaces platform/win/plugins/plugin-delayed-destroy.html.
2683         That test was made to prevent a crash very similar to this one, but
2684         unfortunately tested only the mechanism that prevented the crash and
2685         not whether the crash itself was prevented. Since WebKit2 uses a
2686         different mechanism to prevent the crash, the test was failing even
2687         though WebKit2 was not vulnerable to the crash. This new test crashes
2688         if there is no mechanism in place to prevent it and passes in both
2689         WebKit1 and WebKit2.
2690
2691         Fixes <http://webkit.org/b/46711> <rdar://problem/8485903>
2692         platform/win/plugins/plugin-delayed-destroy.html fails in WebKit2
2693
2694         Reviewed by Anders Carlsson.
2695
2696         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
2697         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2698         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
2699         * GNUmakefile.am:
2700         Added new file.
2701
2702         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
2703         (pluginDeallocate): Make sure we delete the PluginTest object. This
2704         prevents a leak and also allows us to test the crash.
2705
2706         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
2707         (PluginTest::executeScript): Made this into a non-static member
2708         function.
2709
2710         (PluginTest::waitUntilDone):
2711         (PluginTest::notifyDone):
2712         Updated for changes to executeScript.
2713
2714         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Added executeScript.
2715
2716         * DumpRenderTree/TestNetscapePlugIn/Tests/EvaluateJSAfterRemovingPluginElement.cpp: Added.
2717         (EvaluateJSAfterRemovingPluginElement::EvaluateJSAfterRemovingPluginElement):
2718         Initialize ourselves and tell the test harness to wait.
2719         (EvaluateJSAfterRemovingPluginElement::NPP_DestroyStream): Remove our
2720         plugin element from the document, then execute some JavaScript. If
2721         WebKit does not have appropriate mechanisms in place, we'll be
2722         destroyed inside the first call to executeScript and crash on the
2723         second call.
2724
2725 2010-11-02  Stephen White  <senorblanco@chromium.org>
2726
2727         Reviewed by Tony Chang.
2728
2729         [chromium] Fix LayoutTestController UMRs.
2730         https://bugs.webkit.org/show_bug.cgi?id=48872
2731
2732         * DumpRenderTree/chromium/LayoutTestController.cpp:
2733         (LayoutTestController::LayoutTestController):
2734
2735 2010-11-03  Kent Tamura  <tkent@chromium.org>
2736
2737         Reviewed by Dimitri Glazkov.
2738
2739         REGRESSION: rebaseline-chromium-webkit-tests uses non-zero tolerance for
2740         image dup detection
2741         https://bugs.webkit.org/show_bug.cgi?id=48744
2742
2743         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2744          - Make a function for option parsing for ease of test
2745          - Set 0 to options.tolerance
2746         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
2747          - Add a test for this change
2748
2749 2010-11-02  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
2750
2751         Unreviewed: Add myself to the list of Committers.
2752
2753         * Scripts/webkitpy/common/config/committers.py:
2754
2755 2010-11-02  Anders Carlsson  <andersca@apple.com>
2756
2757         Fix build.
2758
2759         * TestWebKitAPI/PlatformUtilities.cpp:
2760         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
2761
2762 2010-11-02  Sam Weinig  <sam@webkit.org>
2763
2764         Reviewed by Anders Carlsson.
2765
2766         Add a way to send startup messages on the context which can be posted when a process launches
2767         <rdar://problem/8617928>
2768         https://bugs.webkit.org/show_bug.cgi?id=48838
2769
2770         * MiniBrowser/mac/WebBundle/WebBundleMain.m:
2771         (WKBundleInitialize):
2772         * TestWebKitAPI/InjectedBundleController.cpp:
2773         (TestWebKitAPI::InjectedBundleController::initialize):
2774         (TestWebKitAPI::InjectedBundleController::didReceiveMessage):
2775         * TestWebKitAPI/InjectedBundleController.h:
2776         * TestWebKitAPI/InjectedBundleMain.cpp:
2777         (WKBundleInitialize):
2778         * TestWebKitAPI/PlatformUtilities.cpp:
2779         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
2780         * WebKitTestRunner/InjectedBundle/InjectedBundleMain.cpp:
2781         (WKBundleInitialize):
2782         Update implementations of WKBundleInitialize to take an initial userData
2783         argument.  Change TestWebKitAPI to use the new initial userData to initialize
2784         each test's bundle.
2785
2786 2010-11-02  Benjamin Kalman  <kalman@google.com>
2787
2788         Reviewed by Ojan Vafai.
2789
2790         new-run-webkit-tests doesn't strip "LayoutTests/" from prefix, unlike old-run-webkit-tests
2791         https://bugs.webkit.org/show_bug.cgi?id=48794
2792
2793         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: Strip the "LayoutTests/" prefix from test argument paths.
2794
2795 2010-11-02  Adam Roben  <aroben@apple.com>
2796
2797         Skip webkitpy.layout_tests.run_webkit_tests_unittest.MainTest on Cygwin
2798         Python 2.5.x
2799
2800         It is known to hang on that version of Python. See
2801         <http://webkit.org/b/48614>.
2802
2803         Reviewed by Adam Barth.
2804
2805         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py: Use
2806         skip_if to skip MainTest on Cygwin Python 2.5.x.
2807
2808         * Scripts/webkitpy/test/skip.py: Added.
2809         (skip_if): If the passed-in condition is false, find all the test_*
2810         methods of the given class and replace them with a function that just
2811         logs that we're skipping these tests. This is loosely based on the
2812         unittest.skip_if decorator added in Python 3.1.
2813         (_skipped_method): Returns a function that just logs that the tests are
2814         being skipped. This is used in place of the actual test_* functions.
2815
2816         * Scripts/webkitpy/test/skip_unittest.py: Added.
2817         (SkipTest.setUp): Create a logger for ourselves and give it a handler
2818         that logs to a private stream.
2819         (SkipTest.tearDown): Reset the logger.
2820         (SkipTest.create_fixture_class): Returns a class that we use to test
2821         skip_if. It has a single test_* method, test_foo, that just calls
2822         through to a callback.
2823         (SkipTest.foo_callback): Record that test_foo was called.
2824         (SkipTest.test_skip_if_false): Pass skip_if a False condition and test
2825         that test_foo does get called.
2826         (SkipTest.test_skip_if_true): Pass skip_if a True condition and test
2827         that test_foo does not get called and the appropriate message gets
2828         logged.
2829
2830 2010-11-02  Adam Barth  <abarth@webkit.org>
2831
2832         Reviewed by Eric Seidel.
2833
2834         webkit-patch should tell check-webkit-style which files were changed so
2835         check-webkit-style doesn't have to stat the whole working copy again
2836         https://bugs.webkit.org/show_bug.cgi?id=48792
2837
2838         * Scripts/webkitpy/tool/mocktool.py:
2839         * Scripts/webkitpy/tool/steps/checkstyle.py:
2840
2841 2010-11-02  Robert Kroeger  <rjkroege@chromium.org>
2842
2843         Reviewed by James Robinson.
2844
2845         [Chromium/DRT] Make EventSendingController honour leapForward for touch events.
2846         https://bugs.webkit.org/show_bug.cgi?id=48777
2847
2848         * DumpRenderTree/chromium/EventSender.cpp:
2849         (EventSender::sendCurrentTouchEvent):
2850
2851 2010-11-02  Adam Roben  <aroben@apple.com>
2852
2853         Only track resource identifiers in DRT when dumpResourceLoadCallbacks
2854         is on
2855
2856         This reverts Windows to our pre-r71097 behavior. That patch made us
2857         track all resource identifiers, including the main resource, so the
2858         main resource's URL started appearing in test output instead of
2859         "<unknown>". Arguably having the main resource's URL is better, but all
2860         other platforms print "<unknown>" and we want to match.
2861
2862         Fixes <http://webkit.org/b/48837> <rdar://problem/8620351> REGRESSION
2863         (r71097): Many http tests failing on Windows
2864
2865         Reviewed by Anders Carlsson.
2866
2867         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
2868         (ResourceLoadDelegate::identifierForInitialRequest): Don't add the
2869         identifier to the URL map if we're not supposed to dump resource load
2870         callbacks.
2871         (ResourceLoadDelegate::removeIdentifierForRequest): Always remove the
2872         identifier from the URL map even if we're already "done". There's no
2873         point in keeping out-of-date identifiers around.
2874
2875 2010-11-01  Jenn Braithwaite  <jennb@chromium.org>
2876
2877         Reviewed by Adam Roben.
2878
2879         Windows: Update resource tracking when moving a frame between documents
2880         https://bugs.webkit.org/show_bug.cgi?id=48364
2881
2882         * DumpRenderTree/win/DumpRenderTree.cpp:
2883         (createWebViewAndOffscreenWindow):
2884         (main):
2885         Give each WebView its own ResourceLoadDelegate instance in order to
2886         make assertions about resource ids on a particular WebView.
2887         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
2888         (ResourceLoadDelegate::identifierForInitialRequest):
2889         Always add id to the map.
2890         (ResourceLoadDelegate::removeIdentifierForRequest):
2891         Added.
2892         (ResourceLoadDelegate::willSendRequest):
2893         (ResourceLoadDelegate::didReceiveAuthenticationChallenge):
2894         (ResourceLoadDelegate::didReceiveResponse):
2895         (ResourceLoadDelegate::didFinishLoadingFromDataSource):
2896         (ResourceLoadDelegate::didFailLoadingWithError):
2897         (ResourceLoadDelegate::descriptionSuitableForTestResult):
2898         Replace static descriptionSuitableForTestResult with static member function to access identifier map.
2899         * DumpRenderTree/win/ResourceLoadDelegate.h:
2900         (ResourceLoadDelegate::urlMap):
2901         Moved within class so that each WebView has its own id map.
2902
2903 2010-11-01  Ojan Vafai  <ojan@chromium.org>
2904
2905         Reviewed by Dimitri Glazkov.
2906
2907         remove debug code from run_webkit_tests.py
2908         https://bugs.webkit.org/show_bug.cgi?id=48800
2909
2910         Remove temporary debug code and make --master-name required
2911         if --test-results-server is set now that all clients set
2912         --master-name.
2913
2914         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2915
2916 2010-11-01  Tony Chang  <tony@chromium.org>
2917
2918         Reviewed by Kent Tamura.
2919
2920         add plugin event logging for linux
2921         https://bugs.webkit.org/show_bug.cgi?id=48779
2922
2923         This is taken from Chromium's fork of the layout test plugin:
2924         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
2925
2926         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
2927         (handleEventX11):
2928         (NPP_HandleEvent):
2929
2930 2010-11-01  Dirk Pranke  <dpranke@chromium.org>
2931
2932         Reviewed by James Robinson.
2933
2934         new-run-webkit-tests: use DRT, child-processes=1 for GPU tests by default
2935         https://bugs.webkit.org/show_bug.cgi?id=48790
2936
2937         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2938         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2939
2940 2010-11-01  Adam Barth  <abarth@webkit.org>
2941
2942         Reviewed by Eric Seidel.
2943
2944         Teach check-webkit-style how to accept a list of files to diff on the
2945         command line
2946         https://bugs.webkit.org/show_bug.cgi?id=48784
2947
2948         In a future patch, webkit-patch will use this option to improve
2949         performance.  I'm landing this in two pieces to avoid causing a version
2950         skew problem for the style-bot.
2951
2952         * Scripts/check-webkit-style:
2953         * Scripts/webkitpy/style/optparser.py:
2954         * Scripts/webkitpy/style/optparser_unittest.py:
2955         * Scripts/webkitpy/style_references.py:
2956
2957 2010-11-01  Anders Carlsson  <andersca@apple.com>
2958
2959         Reviewed by John Sullivan.
2960
2961         Tear down the related WebProcessProxy when a WebContext is deallocated
2962         https://bugs.webkit.org/show_bug.cgi?id=48769
2963
2964         * TestWebKitAPI/Tests/WebKit2/FailedLoad.cpp:
2965         (TestWebKitAPI::didFailProvisionalLoadWithErrorForFrame):
2966         We don't support empty URLs anymore, update test to expect a null URL instead.
2967
2968 2010-11-01  Søren Gjesse  <sgjesse@chromium.org>
2969
2970         Reviewed by Andreas Kling.
2971
2972         Fix warning when compiling the chromium port of DumpRenderShell
2973         with clang.
2974         https://bugs.webkit.org/show_bug.cgi?id=48414
2975
2976         * DumpRenderTree/chromium/TestShell.h:
2977         (TestShell::javaScriptFlagsForLoad):
2978
2979 2010-11-01  Adam Roben  <aroben@apple.com>
2980
2981         Fix typo from r71022
2982
2983         * BuildSlaveSupport/build.webkit.org-config/config.json:
2984
2985 2010-11-01  Adam Roben  <aroben@apple.com>
2986
2987         Trigger the Windows Release WebKit2 tests when a Release build
2988         finishes, not when a Debug build finishes
2989
2990         Fixes <http://webkit.org/b/48754> Windows Release WebKit2 tests are
2991         triggered at the wrong time
2992
2993         Reviewed by Sam Weinig.
2994
2995         * BuildSlaveSupport/build.webkit.org-config/config.json: Fixed
2996         triggerable name and trigger.
2997
2998 2010-11-01  Mario Sanchez Prada  <msanchez@igalia.com>
2999
3000         Unreviewed. Adding my IRC nickname to the list of committers.
3001
3002         * Scripts/webkitpy/common/config/committers.py:
3003
3004 2010-10-31  Robert Hogan  <robert@webkit.org>
3005
3006         Reviewed by Antonio Gomes.
3007
3008         [Qt] [Gtk] Plug-ins having upper case in mime type are failing to load
3009
3010         Qt and Gtk are case-sensitive when storing the declared mime-type
3011         of plugins. Since plugin mime-types are lowercased prior to searching
3012         for them in the plugin database, ensure they are loaded with the
3013         mime-type in lower case too.
3014
3015         Change the test netscape plugin to declare its mimetype in sentence
3016         case so that the correct behaviour is enforced.
3017
3018         https://bugs.webkit.org/show_bug.cgi?id=36815
3019
3020         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
3021         (NP_GetMIMEDescription):
3022
3023 2010-10-31  Robert Hogan  <robert@webkit.org>
3024
3025         Reviewed by Antonio Gomes.
3026
3027         [Qt] Support nodesFromRect in DRT
3028
3029         https://bugs.webkit.org/show_bug.cgi?id=48716
3030
3031         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3032         (LayoutTestController::nodesFromRect):
3033         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3034
3035 2010-10-30  Andreas Kling  <kling@webkit.org>
3036
3037         Reviewed by Antonio Gomes.
3038
3039         [Qt] QtTestBrowser: Switching view type moves the embedded inspector
3040         https://bugs.webkit.org/show_bug.cgi?id=48705
3041
3042         Reinsert the embedded inspector into the splitter after changing
3043         between QWebView/QGraphicsWebView.
3044
3045         * QtTestBrowser/launcherwindow.cpp:
3046         (LauncherWindow::init):
3047         (LauncherWindow::initializeView):
3048         * QtTestBrowser/webinspector.h:
3049         (WebInspector::WebInspector):
3050
3051 2010-10-28  Antonio Gomes  <agomes@rim.com>
3052
3053         Reviewed by Ojan Vafai.
3054
3055         Needs a "LinuxEditingBehavior", perhaps with a better name
3056         https://bugs.webkit.org/show_bug.cgi?id=36627
3057
3058         Adding support to Mac's, GTK+'s, Windows' and Chromium's LayoutTestController class to test the newly introduced Unix editing behavior.
3059
3060         * DumpRenderTree/chromium/LayoutTestController.cpp:
3061         (LayoutTestController::setEditingBehavior):
3062         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3063         (LayoutTestController::setEditingBehavior):
3064         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3065         (LayoutTestController::setEditingBehavior):
3066         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3067         (LayoutTestController::setEditingBehavior):
3068
3069 2010-10-29  Andreas Kling  <kling@webkit.org>
3070
3071         Reviewed by Kenneth Rohde Christiansen.
3072
3073         [Qt] QtTestBrowser: Add keyboard shortcut to toggle full screen (F11)
3074         https://bugs.webkit.org/show_bug.cgi?id=48695
3075
3076         * QtTestBrowser/launcherwindow.cpp:
3077         (LauncherWindow::createChrome):
3078
3079 2010-10-29  Andreas Kling  <kling@webkit.org>
3080
3081         Reviewed by Kenneth Rohde Christiansen.
3082
3083         [Qt] QtTestBrowser: Fix uninitialized read in FpsTimer
3084         https://bugs.webkit.org/show_bug.cgi?id=48675
3085
3086         FpsTimer::m_timer was never initialized and passed to QObject::killTimer()
3087         on startup with in -graphicsbased mode.
3088
3089         * QtTestBrowser/fpstimer.cpp:
3090         (FpsTimer::FpsTimer):
3091         (FpsTimer::numFrames):
3092         (FpsTimer::stop):
3093         (FpsTimer::timerEvent):
3094         * QtTestBrowser/fpstimer.h:
3095
3096 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
3097
3098         Reviewed by Tony Chang.
3099
3100         new-run-webkit-tests: change TestResults to be serializable
3101
3102         In preparation for changing new-run-webkit-tests from
3103         multithreaded to multiprocess, we need to make sure the data
3104         going between the threads is easily serialized over a socket.
3105
3106         This change adds serialization/pickling for the TestResults and
3107         TestFailure objects (using cPickle).
3108
3109         The TestFailure objects included a "has_wdiff" flag for Text
3110         results, but the flag wasn't being used, so I've removed it,
3111         simplifying the state to basically a set of enum objects with
3112         associated methods.
3113
3114         https://bugs.webkit.org/show_bug.cgi?id=48616
3115
3116         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3117         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
3118         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
3119         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
3120         * Scripts/webkitpy/layout_tests/layout_package/test_results.py: Added.
3121         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:Added. 
3122         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3123         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
3124
3125 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
3126
3127         Reviewed by Ojan Vafai.
3128
3129         new-run-webkit-tests: need to provide separate dashboard results for GPU tests
3130         https://bugs.webkit.org/show_bug.cgi?id=48687
3131
3132         Modify the steps to upload the JSON files to the dashboards so
3133         that the GPU tests don't conflict with the regular tests on a
3134         port. We do this by modifying the --builder-name parameter to
3135         append " - GPU", which should cause the app to treat the results
3136         as a completely new builder. This is a little non-obvious, but
3137         keeps us from having to restructure the app.
3138
3139         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
3140         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
3141         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3142
3143 2010-10-29  Ojan Vafai  <ojan@chromium.org>
3144
3145         Reviewed by Tony Chang.
3146
3147         [chromium] add debug logging to help diagnose flakiness dashboard issues
3148         https://bugs.webkit.org/show_bug.cgi?id=48657
3149
3150         The appengine app thinks it's getting empty files uploaded. Add some logging
3151         to see if new-run-webkit-tests agrees.
3152         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3153
3154 2010-10-29  Dimitri Glazkov  <dglazkov@chromium.org>
3155
3156         Reviewed by Ojan Vafai.
3157
3158         [Chromium/DRT] Add master-name flag to new-run-webkit-tests.
3159         https://bugs.webkit.org/show_bug.cgi?id=48649
3160
3161         * BuildSlaveSupport/build.webkit.org-config/master.cfg: Added flag.
3162
3163 2010-10-29  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
3164
3165         Reviewed by Martin Robinson.
3166
3167         Enable popup window in GtkLauncher
3168         https://bugs.webkit.org/show_bug.cgi?id=48335
3169
3170         The GtkLauncher application does not support opening new window when
3171         a link with "target=_blank" is clicked or similar call to
3172         window.open(). Instead, GtkLauncher does nothing which breaks
3173         navigation of some websites.
3174
3175         * GtkLauncher/main.c:
3176         (activate_uri_entry_cb):
3177         (update_title):
3178         (link_hover_cb):
3179         (notify_title_cb):
3180         (notify_load_status_cb):
3181         (notify_progress_cb):
3182         (destroy_cb):
3183         (go_back_cb):
3184         (go_forward_cb):
3185         (create_web_view_cb):
3186         (web_view_ready_cb):
3187         (close_web_view_cb):
3188         (create_browser):
3189         (create_statusbar):
3190         (create_toolbar):
3191         (create_window):
3192         (main):
3193
3194 2010-10-29  Adam Roben  <aroben@apple.com>
3195
3196         Teach check-webkit-style about WebKit2's idiosyncracies
3197
3198         Fixes <http://webkit.org/b/48638> Style bot complains about a number
3199         of WebKit2 conventions
3200
3201         Reviewed by Anders Carlsson.
3202
3203         * Scripts/webkitpy/style/checker.py: Excluded some rules for various
3204         WebKit2-related files. Also updated the excluded rules for
3205         WebKitAPITest to match the current code.
3206
3207 2010-10-29  Csaba Osztrogonác  <ossy@webkit.org>
3208
3209         Reviewed by Adam Roben and David Kilzer.
3210
3211         Fix and cleanup of build systems
3212         https://bugs.webkit.org/show_bug.cgi?id=48342
3213
3214         * Scripts/build-webkit:
3215          - Remove unnecessary ENABLE_SANDBOX option.
3216          - Add ENABLE_FULLSCREEN_API option.
3217
3218 2010-10-28  Adam Roben  <aroben@apple.com>
3219
3220         Switch the Windows WebKit2 bot to the Release configuration
3221
3222         We only have one machine testing WebKit2 on Windows right now, and
3223         Debug is just too slow for it to keep up.
3224
3225         Fixes (hopefully!) <http://webkit.org/b/48615> Windows WebKit2 bot is
3226         always way behind
3227
3228         Reviewed by Jon Honeycutt.
3229
3230         * BuildSlaveSupport/build.webkit.org-config/config.json:
3231
3232 2010-10-29  Leandro Gracia Gil  <leandrogracia@google.com>
3233
3234         Reviewed by Jeremy Orlow.
3235
3236         Added a second parameter to setMockSpeechInputResult for
3237         the language used in speech input.
3238         https://bugs.webkit.org/show_bug.cgi?id=47089
3239
3240         * DumpRenderTree/LayoutTestController.cpp:
3241         (setMockSpeechInputResultCallback):
3242         * DumpRenderTree/LayoutTestController.h:
3243         * DumpRenderTree/chromium/LayoutTestController.cpp:
3244         (LayoutTestController::setMockSpeechInputResult):
3245         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3246         (LayoutTestController::setMockSpeechInputResult):
3247         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3248         (LayoutTestController::setMockSpeechInputResult):
3249         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3250         (LayoutTestController::setMockSpeechInputResult):
3251         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3252         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3253         (LayoutTestController::setMockSpeechInputResult):
3254         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3255         (LayoutTestController::setMockSpeechInputResult):
3256         (LayoutTestController::pageSizeAndMarginsInPixels):
3257
3258 2010-10-28  Ojan Vafai  <ojan@chromium.org>
3259
3260         Reviewed by Tony Chang.
3261
3262         [chromium] add a result-small.json file for the test dashboard
3263         https://bugs.webkit.org/show_bug.cgi?id=48547
3264
3265         Output both a results.json file and a results-small.json file.
3266         The dashboard will load results-small.json by default so it loads faster.
3267
3268         * TestResultServer/model/jsonresults.py:
3269         * TestResultServer/model/jsonresults_unittest.py:
3270         Added a bunch of sys.path hackery. Unforunately, this uses hardcoded
3271         paths. That obviously needs to be fixed, but at least this way it
3272         clearly documents what paths are necessary.
3273
3274 2010-10-28  Eric Seidel  <eric@webkit.org>
3275
3276         Reviewed by Adam Barth.
3277
3278         webkit-patch upload calls changed_files more often than it should
3279         https://bugs.webkit.org/show_bug.cgi?id=48567
3280
3281         Passing changed_files around everywhere isn't a very elegant solution
3282         but it's the one we have for the moment.  I think keeping an explicit
3283         cache on Checkout (or making StepState() a real class) is a better
3284         long-term option.
3285
3286         Previously bug_id_for_this_commit was calling changed_files and the
3287         result was never getting cached on the state.  Now we're explicitly
3288         caching the result on the state and passing that to the bug_id_for_this_commit call.
3289
3290         I looked into building unit tests for this.  Doing so would require
3291         using a real Checkout object with a MockSCM and overriding the appropriate
3292         calls on SCM to count how often we're stating the file system.
3293         That's a useful set of tests to build for a separate change.
3294
3295         * Scripts/webkitpy/common/checkout/api.py:
3296         * Scripts/webkitpy/tool/commands/download.py:
3297         * Scripts/webkitpy/tool/commands/upload.py:
3298         * Scripts/webkitpy/tool/mocktool.py:
3299
3300 2010-10-28  Eric Seidel  <eric@webkit.org>
3301
3302         Reviewed by Adam Barth.
3303
3304         Make suggest-reviewers slightly faster
3305         https://bugs.webkit.org/show_bug.cgi?id=48562
3306
3307         Add @memoized to one more common call.
3308
3309         * Scripts/webkitpy/common/checkout/api.py:
3310
3311 2010-10-28  Eric Seidel  <eric@webkit.org>
3312
3313         Reviewed by Adam Barth.
3314
3315         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is still very slow
3316         https://bugs.webkit.org/show_bug.cgi?id=48536
3317
3318         This patch makes it about 40 seconds faster, but it still
3319         takes 1:40.  This will require more refinement.
3320
3321         The suggested reviewers list appears to be the same.
3322
3323         I think the next step may be to have it stop the search after
3324         5 reviewers are found.  We never want to suggest 30 people.
3325
3326         * Scripts/webkitpy/common/checkout/scm.py:
3327          - Using --remove-empty to theoretically stop lookups past
3328            when a file is removed.  I'm not entirely clear that the option
3329            does what it says it does.  Example:
3330            git log --pretty=format:%H -5 --remove-empty -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
3331            returns only one commit
3332            vs.
3333            git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
3334            which returns 5.  I was not aware that wx files were ever removed from the repository?
3335
3336 2010-10-28  Kinuko Yasuda  <kinuko@chromium.org>
3337
3338         Reviewed by David Levin.
3339
3340         [Chromium] Support FileSystem in chromium DRT
3341         https://bugs.webkit.org/show_bug.cgi?id=47643
3342
3343         * DumpRenderTree/chromium/WebViewHost.cpp:
3344         (WebViewHost::openFileSystem): Added.
3345         * DumpRenderTree/chromium/WebViewHost.h:
3346         (WebViewHost::openFileSystem): Added.
3347
3348 2010-10-28  Sam Weinig  <sam@webkit.org>
3349
3350         Reviewed by Anders Carlsson.
3351
3352         WKURLRefs should be allowed to be null
3353         <rdar://problem/8575621>
3354         https://bugs.webkit.org/show_bug.cgi?id=48535
3355
3356         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
3357         * TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp:
3358         (TestWebKitAPI::didStartProvisionalLoadForFrame):
3359         (TestWebKitAPI::didCommitLoadForFrame):
3360         (TestWebKitAPI::didFinishLoadForFrame):
3361         Test that URLs are null pointers when unset.
3362
3363 2010-10-28  Chris Fleizach  <cfleizach@apple.com>
3364
3365         Reviewed by Adele Peterson.
3366
3367         AX: multi select group option does not handle setting of AXSelectedChildren correctly
3368         https://bugs.webkit.org/show_bug.cgi?id=48464
3369
3370         Add support for querying information about selected children to DRT, including:
3371            selectedChildAtIndex
3372            selectedChildrenCount
3373            setSelectedChild
3374
3375         * DumpRenderTree/AccessibilityUIElement.cpp:
3376         (selectedChildAtIndexCallback):
3377         (setSelectedChildCallback):
3378         (selectedChildrenCountCallback):
3379         (AccessibilityUIElement::setSelectedChild):
3380         (AccessibilityUIElement::selectedChildrenCount):
3381         (AccessibilityUIElement::selectedChildAtIndex):
3382         (AccessibilityUIElement::getJSClass):
3383         * DumpRenderTree/AccessibilityUIElement.h:
3384         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
3385         (AccessibilityUIElement::selectedChildAtIndex):
3386         (AccessibilityUIElement::selectedChildrenCount):
3387         (AccessibilityUIElement::setSelectedChild):
3388
3389 2010-10-27  Sam Weinig  <sam@webkit.org>
3390
3391         Reviewed by Anders Carlsson.
3392
3393         Add WebKit2 API for window feature getter/setters
3394         <rdar://problem/8590373>
3395         https://bugs.webkit.org/show_bug.cgi?id=48496
3396
3397         * MiniBrowser/mac/BrowserWindowController.m:
3398         (-[BrowserWindowController awakeFromNib]):
3399         * MiniBrowser/win/BrowserView.cpp:
3400         (BrowserView::create):
3401         * WebKitTestRunner/TestController.cpp:
3402         (WTR::createOtherPage):
3403         (WTR::TestController::initialize):
3404
3405 2010-10-28  Søren Gjesse  <sgjesse@chromium.org>
3406
3407         Reviewed by Tony Chang.
3408
3409         Added support for the DumpRenderTree flags --multiple-loads and --js-flags to the Python test runner.
3410         https://bugs.webkit.org/show_bug.cgi?id=48236
3411
3412         * Scripts/webkitpy/layout_tests/port/chromium.py:
3413         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3414
3415 2010-10-28  Sergio Villar Senin  <svillar@igalia.com>
3416
3417         Reviewed by Martin Robinson.
3418
3419         [GTK] WebKitWebFrame's load-status is not properly notified to the tests
3420         https://bugs.webkit.org/show_bug.cgi?id=48048
3421
3422         DRT now listens to WebKitWebFrame load-status signals for each
3423         created frame instead of the load-status signal of the
3424         WebKitWebView that only notifies about changes in the main frame.
3425
3426         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3427         (webViewLoadFinished):
3428         (webFrameLoadStatusNotified):
3429         (frameCreatedCallback):
3430         (createWebView):
3431         (main):
3432
3433 2010-10-28  Tony Chang  <tony@chromium.org>
3434
3435         Reviewed by Kent Tamura.
3436
3437         [chromium] fix textInputController.{selectedRange,markedRange}
3438         https://bugs.webkit.org/show_bug.cgi?id=48487
3439
3440         * DumpRenderTree/chromium/TextInputController.cpp:
3441         (TextInputController::markedRange): Return arrays of ints, rather than a string
3442         (TextInputController::selectedRange): Ditto.
3443
3444 2010-10-27  Ojan Vafai  <ojan@chromium.org>
3445
3446         Reviewed by Tony Chang.
3447
3448         [chromium] add a master-name flag to new-run-webkit-tests
3449         https://bugs.webkit.org/show_bug.cgi?id=48488
3450
3451         The test results server now allows adding a master name to
3452         the uploaded files. This lets us distinguish bots that have
3453         the same name, but are on different masters.
3454
3455         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3456
3457 2010-10-28  Adam Roben  <aroben@apple.com>
3458
3459         Don't append a newline to the test output if the frame has no document
3460         element in WebKitTestRunner
3461
3462         Fixes <http://webkit.org/b/48526> Extra trailing newline when running
3463         plugins/document-open.html in WebKitTestRunner
3464
3465         Reviewed by Anders Carlsson.
3466
3467         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3468         (WTR::toJS): Added. Turns a UTF-8 C string into a JSStringRef.
3469         (WTR::hasDocumentElement): Added. Uses the JSC API to figure out
3470         whether the frame has a document element.
3471         (WTR::dumpFrameText): Match DRT by bailing (rather than appending an
3472         empty string and a newline) if the frame has no document element.
3473
3474 2010-10-28  Adam Roben  <aroben@apple.com>
3475
3476         Skip npn-invalidate-rect-invalidates-window.html on headless XP
3477         machines
3478
3479         TestNetscapePlugIn never receives a WM_PAINT message on headless XP
3480         machines, so this test times out. Fixing the test is covered by
3481         <http://webkit.org/b/48333>.
3482
3483         * Scripts/old-run-webkit-tests: Skip
3484         npn-invalidate-rect-invalidates-window.html on Windows if accelerated
3485         compositing support is disabled, which likely means we're on a headless
3486         XP machine.
3487
3488 2010-10-28  Kimmo Kinnunen  <kimmok@iki.fi>
3489
3490         Adding myself as a committer.
3491
3492         * Scripts/webkitpy/common/config/committers.py:
3493
3494 2010-10-27  Eric Seidel  <eric@webkit.org>
3495
3496         Reviewed by Adam Barth.
3497
3498         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is very slow
3499         https://bugs.webkit.org/show_bug.cgi?id=48500
3500
3501         This doesn't fix the problem, but it makes things slightly better.
3502         Each git svn find-rev call takes about .25 seconds on my desktop.
3503         This patch uses a new memoized class to avoid those calls when possible.
3504
3505         The real slowness is still git log on some files, like:
3506         git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
3507         I'm not yet sure how to make the pathological git logs better.
3508
3509         * Scripts/webkitpy/common/checkout/scm.py:
3510         * Scripts/webkitpy/common/memoized.py: Added.
3511         * Scripts/webkitpy/common/memoized_unittest.py: Added.
3512
3513 2010-10-27  Eric Seidel  <eric@webkit.org>
3514
3515         Reviewed by Ojan Vafai.
3516
3517         EWS bots should not use --quiet when running build-webkit
3518         https://bugs.webkit.org/show_bug.cgi?id=48482
3519
3520         --quiet is only correct when the sub-process does the error reporting.
3521         In the case of _can_build() the parent process is reporting the error.
3522         We'd like the full build log at queues.webkit.org so someone can look
3523         at the log and understand why the EWS is failing to build trunk.
3524
3525         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
3526         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
3527         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
3528         * Scripts/webkitpy/tool/mocktool.py:
3529
3530 2010-10-27  Brian Weinstein  <bweinstein@apple.com>
3531
3532         More Windows build fixage. Rename a variable that was named string.
3533
3534         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3535         (WTR::hasPrefix):
3536
3537 2010-10-27  Ojan Vafai  <ojan@chromium.org>
3538
3539         Reviewed by Dimitri Glazkov.
3540
3541         [chromium] Make the test results server store which master the bot is on
3542         https://bugs.webkit.org/show_bug.cgi?id=48478
3543
3544         The chromium bots recently changed so that there are multiple slaves with
3545         the same name on different masters. Up till now, the test results server
3546         assumed slave names were unique. Adds a master field to the file in order
3547         to distinguish.
3548
3549         Also, for files that currently lack a master or testtype, set them appropriately.
3550
3551         * TestResultServer/handlers/testfilehandler.py:
3552         * TestResultServer/index.yaml:
3553         * TestResultServer/model/jsonresults.py:
3554         * TestResultServer/model/testfile.py:
3555         * TestResultServer/templates/showfilelist.html:
3556         * TestResultServer/templates/uploadform.html:
3557
3558 2010-10-26  Darin Adler  <darin@apple.com>
3559
3560         Reviewed by Sam Weinig.
3561
3562         WebKitTestRunner needs to support layoutTestController.dumpBackForwardList
3563         https://bugs.webkit.org/show_bug.cgi?id=42322
3564         rdar://problem/8193631
3565
3566         WebKitTestRunner needs to support layoutTestController.clearBackForwardList
3567         https://bugs.webkit.org/show_bug.cgi?id=42333
3568         rdar://problem/8193643
3569
3570         * WebKitTestRunner/Configurations/InjectedBundle.xcconfig:
3571         Renamed the product to WebKitTestRunnerInjectedBundle to avoid
3572         name conflicts in the build directory.
3573
3574         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
3575         Added dumpBackForwardList and clearBackForwardList.
3576
3577         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
3578         (WTR::InjectedBundle::InjectedBundle): Removed initialization of m_mainPage.
3579         (WTR::InjectedBundle::didCreatePage): Removed unneeded underscore.
3580         (WTR::InjectedBundle::willDestroyPage): Ditto.
3581         (WTR::InjectedBundle::didReceiveMessage): Ditto.
3582         (WTR::InjectedBundle::initialize): Ditto.
3583         (WTR::InjectedBundle::didCreatePage): Changed code to use m_pages
3584         instead of m_mainPage and m_otherPages.
3585         (WTR::InjectedBundle::willDestroyPage): Ditto.
3586         (WTR::InjectedBundle::page): Ditto.
3587         (WTR::InjectedBundle::beginTesting): Ditto.
3588         (WTR::InjectedBundle::done): Ditto.
3589         (WTR::InjectedBundle::closeOtherPages): Ditto.
3590         (WTR::InjectedBundle::dumpBackForwardListsForAllPages): Added.
3591
3592         * WebKitTestRunner/InjectedBundle/InjectedBundle.h: Changed
3593         page function to longer be inline, and pageCount function
3594         to use m_pages. Added dumpBackForwardListsForAllPages, and
3595         removed some underscores. Replaced m_mainPage and m_otherPageas
3596         with m_pages.
3597
3598         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3599         (WTR::adoptWK): Added.
3600         (WTR::hasPrefix): Added.
3601         (WTR::InjectedBundlePage::reset): Added code to set up
3602         m_previousTestBackForwardListItem.
3603         (WTR::InjectedBundlePage::dump): Added code to call
3604         dumpBackForwardListsForAllPages.
3605         (WTR::compareByTargetName): Added.
3606         (WTR::dumpBackForwardListItem): Added.
3607         (WTR::InjectedBundlePage::dumpBackForwardList): Added.
3608
3609         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h: Added
3610         dumpBackForwardList and m_previousTestBackForwardListItem.
3611
3612         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
3613         (WTR::LayoutTestController::LayoutTestController): Initialize
3614         m_shouldDumpBackForwardListsForAllWindows to false.
3615         (WTR::LayoutTestController::clearBackForwardList): Added.
3616
3617         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
3618         Added dumpBackForwardList, clearBackForwardList,
3619         shouldDumpBackForwardListsForAllWindows, and
3620         m_shouldDumpBackForwardListsForAllWindows.
3621
3622         * WebKitTestRunner/TestController.cpp:
3623         (WTR::TestController::initialize): Set up
3624         didReceiveSynchronousMessageFromInjectedBundle.
3625         (WTR::TestController::didReceiveSynchronousMessageFromInjectedBundle):
3626         Added.
3627         * WebKitTestRunner/TestController.h: Ditto.
3628
3629         * WebKitTestRunner/TestInvocation.cpp:
3630         (WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):
3631         Added.
3632         * WebKitTestRunner/TestInvocation.h: Ditto.
3633
3634         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
3635         Renamed bundle to WebKitTestRunnerInjectedBundle (see above).
3636         * WebKitTestRunner/mac/TestControllerMac.mm:
3637         (WTR::TestController::initializeInjectedBundlePath): Ditto.
3638
3639 2010-10-27  Chris Rogers  <crogers@google.com>
3640
3641         Reviewed by Chris Marrin.
3642
3643         Add ENABLE_WEB_AUDIO feature enable flag (initially disabled) to build-webkit
3644         https://bugs.webkit.org/show_bug.cgi?id=48279
3645
3646         * Scripts/build-webkit:
3647
3648 2010-10-27  Eric Seidel  <eric@webkit.org>
3649
3650         Unreviewed.
3651
3652         build-webkit should collect Visual Studio Express logs and display them
3653         https://bugs.webkit.org/show_bug.cgi?id=39199