54b00aba80d4de40ea9de6251399c28c39e89c6d
[WebKit.git] / WebKitTools / ChangeLog
1 2010-11-22  Ojan Vafai  <ojan@chromium.org>
2
3         Reviewed by Adam Barth.
4
5         update scm to work with different remote branches
6         https://bugs.webkit.org/show_bug.cgi?id=49949
7
8         This works if there are multiple svn tracking branches
9         and/or if the tracking branch is refs/remots/origin/master
10         instead of refs/remotes/trunk which is the direction we're trying
11         to head since that makes fetches faster.
12
13         * Scripts/webkitpy/common/checkout/scm.py:
14         * Scripts/webkitpy/common/checkout/scm_unittest.py:
15
16 2010-11-22  Hayato Ito  <hayato@chromium.org>
17
18         Reviewed by Shinichiro Hamaji.
19
20         Ignore reference files which will be used by reftests when collecting
21         test cases.
22         https://bugs.webkit.org/show_bug.cgi?id=49835
23
24         * Scripts/webkitpy/layout_tests/port/test_files.py:
25         * Scripts/webkitpy/layout_tests/port/test_files_unittest.py:
26
27 2010-11-22  Adam Roben  <aroben@apple.com>
28
29         Use paths relative to $WebKitVSPropsRedirectionDir to access shared .vsprops files
30
31         Apple's Windows build allows placing header files and import libraries for WebKit's
32         dependencies (CoreGraphics, CFNetwork, SQLite, etc.) outside the source tree via the
33         $WebKitLibrariesDir environment variable. This is both required for production builds and
34         convenient for Apple-internal developer builds. Apple's production builds also require that
35         WebKit's shared .vsprops files be accessed relative to $WebKitLibrariesDir. In production
36         builds, the files are copied into that directory tree by the
37         WebKitLibraries/win/tools/WinTools.make file. In Apple-internal developer builds, the
38         copying is done by
39         JavaScriptCore/JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreGenerated.make.
40
41         This .vsprops copying is problematic in one very important case: when a developer updates
42         their source tree and then tries to build. Visual Studio only reads .vsprops files when a
43         project is first loaded. So, when Visual Studio is first opened after the .vsprops files are
44         updated, it reads in the old files that were already residing in $WebKitLibrariesDir. When a
45         build is started, JavaScriptCoreGenerated.make copies the new .vsprops files into
46         $WebKitLibrariesDir, but Visual Studio will not pick up the changes.  The rest of the build
47         will proceed with out-of-date .vsprops files, which will likely result in a build failure.
48
49         To fix this, we now use normal relative paths to access the .vsprops files in the source
50         tree rather than in $WebKitLibrariesDir, but prefix those paths with a new environment
51         variable, $WebKitVSPropsRedirectionDir. In developer builds, this environment variable is
52         unset, so the normal relative paths are used to read the .vsprops files out of the source
53         tree directly. In production builds, this environment variable is set to a fake directory
54         that will cause the .vsprops files in $WebKitLibrariesDir to be found when the relative path
55         is resolved.
56         
57         For example, JavaScriptCore.vcproj uses this path for FeatureDefines.vsprops:
58
59         $(WebKitVSPropsRedirectionDir)..\..\..\WebKitLibraries\win\tools\vsprops\FeatureDefines.vsprops
60
61         In developer builds, where $WebKitVSPropsRedirectionDir is unset, this will point to the
62         files in WebKitLibraries\win\tools\vsprops in the source tree. In production builds,
63         JavaScriptCore.make sets $WebKitVSPropsRedirectionDir to
64         "$(SRCROOT)\AppleInternal\tools\vsprops\OpenSource\1\2\3\", so the full path for
65         FeatureDefines.vsprops becomes:
66
67         $(SRCROOT)\AppleInternal\tools\vsprops\OpenSource\1\2\3\..\..\..\WebKitLibraries\win\tools\vsprops\FeatureDefines.vsprops
68
69         which resolves to:
70
71         $(SRCROOT)\AppleInternal\tools\vsprops\OpenSource\WebKitLibraries\win\tools\vsprops\FeatureDefines.vsprops
72
73         (We rely on the fact that Windows doesn't care whether the directories "1", "2", and "3"
74         actually exist since they are matched by an equal number of ".." path components.)
75
76         Note that Visual Studio still won't pick up changes made to .vsprops files while Visual
77         Studio is open, but that problem hasn't seemed to cause developers many headaches so far.
78
79         Fixes <http://webkit.org/b/49181> Windows build fails mysteriously when .vsprops files are
80         updated
81
82         Reviewed by Dave Hyatt.
83
84         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
85         * DumpRenderTree/win/DumpRenderTree.vcproj:
86         * DumpRenderTree/win/ImageDiff.vcproj:
87         * FindSafari/FindSafari.vcproj:
88         * MiniBrowser/MiniBrowser.vcproj:
89         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
90         * TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj:
91         * WebKitAPITest/WebKitAPITest.vcproj:
92         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
93         * WebKitTestRunner/win/InjectedBundle.vcproj:
94         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
95         * WinLauncher/WinLauncher.vcproj:
96         * record-memory-win/record-memory-win.vcproj:
97         Changed to use paths relative to $WebKitVSPropsRedirectionDir to access shared .vsprops
98         files.
99
100 2010-11-22  Adam Roben  <aroben@apple.com>
101
102         Make it possible to run tests on Windows without Visual Studio or VC++
103         Express installed
104
105         webkitdir.pm::setupCygwinEnv dies when Visual Studio and VC++ Express
106         are not installed. But this function doesn't need to be called when we
107         already have a build available and are just trying to run the tests.
108
109         Fixes <http://webkit.org/b/49932> New Windows 7 bot can't run tests
110         because Visual Studio/VC++ Express aren't installed
111
112         Reviewed by Dave Hyatt.
113
114         * Scripts/webkitdirs.pm:
115         (determineConfigurationForVisualStudio): Don't call setupCygwinEnv, as
116         it is not needed by this function. Also added a FIXME.
117
118         (usingVisualStudioExpress): Call setupCygwinEnv directly rather than
119         relying on determineConfigurationForVisualStudio doing it.
120
121 2010-11-19  Hayato Ito  <hayato@chromium.org>
122
123         Reviewed by Shinichiro Hamaji.
124
125         Simplify Driver.run_test() so that it takes only one parameter, TestInput.
126
127         https://bugs.webkit.org/show_bug.cgi?id=49812
128
129         * Scripts/webkitpy/layout_tests/driver_test.py:
130         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
131         * Scripts/webkitpy/layout_tests/port/base.py:
132         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
133         * Scripts/webkitpy/layout_tests/port/chromium.py:
134         * Scripts/webkitpy/layout_tests/port/dryrun.py:
135         * Scripts/webkitpy/layout_tests/port/test.py:
136         * Scripts/webkitpy/layout_tests/port/webkit.py:
137         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
138         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
139
140 2010-11-22  Hayato Ito  <hayato@chromium.org>
141
142         Reviewed by Tony Chang.
143
144         [NRWT] Retry a few times in reading a png image to avoid a race condition.
145
146         https://bugs.webkit.org/show_bug.cgi?id=49924
147
148         * Scripts/webkitpy/layout_tests/port/chromium.py:
149
150 2010-11-22  João Paulo Rechi Vita  <jprvita@profusion.mobi>
151
152         Reviewed by Kenneth Rohde Christiansen.
153
154         Fix EWebLauncher zoom discretization
155         https://bugs.webkit.org/show_bug.cgi?id=49810
156
157         * EWebLauncher/main.c:
158         (nearest_zoom_level_get):
159         (zoom_level_set):
160         (on_load_finished):
161         (on_key_down):
162
163 2010-11-21  Sam Weinig  <sam@webkit.org>
164
165         Rubber-stamped by Dan Bernstein.
166
167         Fix ~300 WebKit2 tests by enabling developer extras preference.
168
169         * WebKitTestRunner/TestController.cpp:
170         (WTR::TestController::resetStateToConsistentValues):
171
172 2010-11-21  Sam Weinig  <sam@webkit.org>
173
174         Reviewed by Dan Bernstein.
175
176         Textareas should be resizable by default
177         https://bugs.webkit.org/show_bug.cgi?id=49892
178
179         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
180         * TestWebKitAPI/Tests/WebKit2/WKPreferences.cpp: Added.
181         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
182         Add preferences test.
183
184 2010-11-19  Mihai Parparita  <mihaip@chromium.org>
185
186         Reviewed by Tony Chang.
187
188         Rebaseline server: checksum-only differences and prefetching
189         https://bugs.webkit.org/show_bug.cgi?id=49841
190         
191         Add support for image test results where the only difference is in the
192         checksum.
193         
194         Add prefetching of the image results from the next test.
195         
196         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
197         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
198         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
199
200 2010-11-19  Dirk Pranke  <dpranke@chromium.org>
201
202         Reviewed by Tony Chang.
203
204         nrwt multiprocessing - add 'worker number' concept, move stuff to worker thread
205
206         Add the 'worker number' and 'worker name' concepts to the 
207         TestShellThread objects, and move test_types and test_args from
208         the TestRunner to the TestShellThread.
209
210         https://bugs.webkit.org/show_bug.cgi?id=49768
211
212         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
213         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
214
215 2010-11-19  Dirk Pranke  <dpranke@chromium.org>
216
217         Reviewed by Ojan Vafai.
218
219         Do some minor cleanup and bug fixing.
220
221         https://bugs.webkit.org/show_bug.cgi?id=49777
222
223         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
224         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
225
226 2010-11-19  Dirk Pranke  <dpranke@chromium.org>
227
228         Reviewed by Ojan Vafai.
229
230         nrwt - config.build_directory() is busted
231
232         Fixes a typo that was causing us to usually return the top level
233         directory WebKitBuild instead of WebKitBuild/{Debug,Release}. The
234         bug was hidden by test stubs that were too simplistic :(.
235
236         https://bugs.webkit.org/show_bug.cgi?id=49815
237
238         * Scripts/webkitpy/common/system/executive_mock.py:
239         * Scripts/webkitpy/layout_tests/port/config.py:
240         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
241
242 2010-11-19  Steve Falkenburg  <sfalken@apple.com>
243
244         Reviewed by Adam Roben.
245
246         Add Debug_Cairo_CFLite and Release_Cairo_CFLite configurations for all vcproj files
247         https://bugs.webkit.org/show_bug.cgi?id=49819
248
249         * DumpRenderTree/DumpRenderTree.sln:
250         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
251         * DumpRenderTree/win/DumpRenderTree.vcproj:
252         * DumpRenderTree/win/ImageDiff.vcproj:
253         * FindSafari/FindSafari.vcproj:
254         * MiniBrowser/MiniBrowser.vcproj:
255         * TestWebKitAPI/win/TestWebKitAPI.sln:
256         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
257         * TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj:
258         * WebKitAPITest/WebKitAPITest.vcproj:
259         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
260         * WebKitTestRunner/win/InjectedBundle.vcproj:
261         * WebKitTestRunner/win/InjectedBundleGenerated.vcproj:
262         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
263         * WinLauncher/WinLauncher.vcproj:
264         * record-memory-win/record-memory-win.vcproj:
265
266 2010-11-19  Adam Roben  <aroben@apple.com>
267
268         Add some more Windows slaves
269
270         This is probably more than we will be using in the immediate future,
271         but having some extra slave names makes experimenting with different
272         configurations easier.
273
274         Reviewed by Jon Honeycutt.
275
276         * BuildSlaveSupport/build.webkit.org-config/config.json: Added 5 more
277         Windows slaves and distributed them among the three different Test
278         builders.
279
280 2010-11-19  Kinuko Yasuda  <kinuko@chromium.org>
281
282         Reviewed by Ojan Vafai.
283
284         [Chromium] REGRESSION: Record actual test results in the JSON result summary for non-layout tests
285         https://bugs.webkit.org/show_bug.cgi?id=49702
286
287         Re-landing r72357 with a test fix.
288
289         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
290         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
291         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
292
293 2010-11-19  Avi Drissman  <avi@google.com>
294
295         Reviewed by Antonio Gomes.
296
297         Complete support for Unix editing mode
298         https://bugs.webkit.org/show_bug.cgi?id=49757
299
300         * DumpRenderTree/LayoutTestController.cpp:
301         (setEditingBehaviorCallback):
302         * DumpRenderTree/chromium/LayoutTestController.cpp:
303         (LayoutTestController::setEditingBehavior):
304
305 2010-11-19  Adam Roben  <aroben@apple.com>
306
307         Fix run-javascriptcore-tests when there's a space in the path to
308         testapi
309
310         Reviewed by Sam Weinig.
311
312         * Scripts/run-javascriptcore-tests: Use an indirect object when calling
313         system() to force it to interpret spaces in the path correctly.
314
315 2010-11-19  Adam Roben  <aroben@apple.com>
316
317         Update scripts for .exe name changes after r72327
318
319         As of r72327, the "_debug" suffix is only used in Debug_All builds.
320
321         Reviewed by Sam Weinig.
322
323         * Scripts/run-javascriptcore-tests:
324         (testapiPath):
325
326         * Scripts/sunspider-compare-results:
327         (pathToBuiltJSC):
328
329         * Scripts/webkitdirs.pm:
330         (jscPath):
331         (safariPath):
332
333 2010-11-19  Steve Falkenburg  <sfalken@apple.com>
334
335         Reviewed by Darin Adler.
336
337         Normalize Cairo/CFLite project/solution configuration names
338         https://bugs.webkit.org/show_bug.cgi?id=49818
339
340         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
341         * DumpRenderTree/win/DumpRenderTree.vcproj:
342         * DumpRenderTree/win/ImageDiff.vcproj:
343         * MiniBrowser/MiniBrowser.vcproj:
344         * TestWebKitAPI/win/TestWebKitAPI.sln:
345         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
346         * WebKitAPITest/WebKitAPITest.vcproj:
347         * WebKitTestRunner/win/InjectedBundle.vcproj:
348         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
349         * WinLauncher/WinLauncher.vcproj:
350         * Scripts/webkitdirs.pm: Updated cairo build configs.
351
352 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
353
354         Reviewed by Adam Roben.
355
356         Windows vcproj configuration names should be normalized across projects
357         https://bugs.webkit.org/show_bug.cgi?id=49776
358
359         * DumpRenderTree/DumpRenderTree.sln:
360         * FindSafari/FindSafari.vcproj:
361         * TestWebKitAPI/win/TestWebKitAPI.sln:
362         * TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj:
363         * TestWebKitAPI/win/TestWebKitAPIGeneratedCommon.vsprops: Added.
364         * WebKitAPITest/WebKitAPITest.vcproj:
365         * WebKitAPITest/WebKitAPITestCommon.vsprops:
366         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
367         * WebKitTestRunner/WebKitTestRunner.sln:
368         * WebKitTestRunner/win/InjectedBundleGenerated.vcproj:
369         * WebKitTestRunner/win/InjectedBundleGeneratedCommon.vsprops: Added.
370         * record-memory-win/record-memory-win.vcproj:
371
372 2010-11-18  Gavin Barraclough  <barraclough@apple.com>
373
374         Build fix - as per DRT-mac, always enabled developer extras,
375         in win/qt/gtk DRT.
376
377         This is necessary to enable rich exception messages to be generated
378         following https://bugs.webkit.org/show_bug.cgi?id=49708.
379
380         * DumpRenderTree/gtk/DumpRenderTree.cpp:
381         (shouldEnableDeveloperExtras):
382         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
383         (WebCore::shouldEnableDeveloperExtras):
384         * DumpRenderTree/win/DumpRenderTree.cpp:
385         (shouldEnableDeveloperExtras):
386
387 2010-11-18  Sheriff Bot  <webkit.review.bot@gmail.com>
388
389         Unreviewed, rolling out r72357.
390         http://trac.webkit.org/changeset/72357
391         https://bugs.webkit.org/show_bug.cgi?id=49784
392
393         It broke test-webkitpy tests since the patch didn't have the
394         corresponding test update (Requested by kinuko on #webkit).
395
396         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
397         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
398
399 2010-11-18  Mihai Parparita  <mihaip@chromium.org>
400
401         Reviewed by Tony Chang.
402
403         Rebaseline server: add rebaseline queue
404         https://bugs.webkit.org/show_bug.cgi?id=49763
405         
406         Add a UI queue (a multi-item <select>) where tests to be baselined
407         can be enqueued (also supports basic removal). A queue is used instead
408         of immediately doing rebaselines since SCM operations can take a while.
409         It's better to go through lots of tests quickly to mark them as
410         needing rebaselining and then batch the SCM operations.
411         
412         Also adds a simple log where results can be displayed.
413
414         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
415         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
416         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
417         * Scripts/webkitpy/tool/commands/data/rebaselineserver/queue.js: Added.
418         * Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js:
419         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
420
421 2010-11-18  Gavin Barraclough  <barraclough@apple.com>
422
423         Reviewed by Oliver Hunt.
424
425         Bug 49708 - Stop recompiling functions to regenerate exception info.
426
427         Instead only hold info as necessary – keep divot info is the inspector
428         is enabled, line number info is debugging or profiling, and handler
429         info for functions with try/catch.
430
431         * DumpRenderTree/mac/DumpRenderTree.mm:
432         (shouldEnableDeveloperExtras):
433             - always enable the developer tools from DRT, to ensure we
434               produce rich error messages on JavaScript exceptions.
435
436 2010-11-18  Kinuko Yasuda  <kinuko@chromium.org>
437
438         Reviewed by Ojan Vafai.
439
440         [Chromium] REGRESSION: Record actual test results in the JSON result summary for non-layout tests
441         https://bugs.webkit.org/show_bug.cgi?id=49702
442
443         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
444         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
445
446 2010-11-18  Dirk Pranke  <dpranke@chromium.org>
447
448         Reviewed by Tony Chang.
449
450         new-run-webkit-tests: create first part of 'message_broker' class for multiprocessing fixes
451
452         Create the first version of the 'message_broker' package. This
453         class will encapsulate all of the threading/multiprocessing and
454         message-sending details for the communication between the
455         'manager' object and the 'worker' objects. For the moment, it
456         just holds some routines and tests for logging thread stacks.
457
458         There should be no functional changes in this patch, just moving stuff
459         around.
460
461         https://bugs.webkit.org/show_bug.cgi?id=49707
462
463         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
464         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread_unittest.py: Removed.
465         * Scripts/webkitpy/layout_tests/layout_package/message_broker.py: Added.
466         * Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py: Added.
467         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
468         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
469
470 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
471
472         Reviewed by Adam Roben.
473
474         Remove leftover Windows Debug_Internal configurations
475         https://bugs.webkit.org/show_bug.cgi?id=49758
476
477         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
478         * DumpRenderTree/win/DumpRenderTree.vcproj:
479         * DumpRenderTree/win/ImageDiff.vcproj:
480         * MiniBrowser/MiniBrowser.vcproj:
481         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
482         * WebKitAPITest/WebKitAPITest.vcproj:
483         * WebKitTestRunner/win/InjectedBundle.vcproj:
484         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
485         * WinLauncher/WinLauncher.vcproj:
486
487 2010-11-18  Adam Roben  <aroben@apple.com>
488
489         Add three new Windows XP build slaves
490
491         Reviewed by Steve Falkenburg.
492
493         * BuildSlaveSupport/build.webkit.org-config/config.json: Added three
494         new slaves, and added them to the Windows XP Debug (Tests) builder.
495
496 2010-11-18  Daniel Bates  <dbates@rim.com>
497
498         Reviewed by Adam Roben.
499
500         update-webkit-support-libs should fall back to existing
501         WebKitSupportLibrary version if there is no internet connectivity
502         https://bugs.webkit.org/show_bug.cgi?id=49503
503
504         Fall back to existing support libraries (if present) when there is
505         no internet connection.
506
507         Currently, update-webkit-support-libs dies with an "out-of-date"
508         error when there is no internet connection because it cannot
509         retrieve versioning information from developer.apple.com. Because
510         update-webkit-support-libs fails, build-webkit fails. Instead,
511         if there is no internet connection and the support libraries are
512         present then we should warn the user and exit() with success so
513         that build-webkit can work without an internet connection.
514
515         * Scripts/update-webkit-support-libs:
516
517 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
518
519         Reviewed by Adam Roben.
520
521         Debug_Internal Windows configuration is unnecessary, should be removed
522         https://bugs.webkit.org/show_bug.cgi?id=49753
523
524         * DumpRenderTree/win/DumpRenderTree.cpp:
525         * MiniBrowser/win/stdafx.h:
526         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp:
527         * WebKitTestRunner/win/TestControllerWin.cpp:
528
529 2010-11-18  Steve Falkenburg  <sfalken@apple.com>
530
531         Reviewed by Adam Roben.
532
533         Clean up vcproj errors
534         https://bugs.webkit.org/show_bug.cgi?id=49741
535
536         * MiniBrowser/MiniBrowser.vcproj:
537         * WebKitAPITest/WebKitAPITest.vcproj:
538         * WebKitTestRunner/win/InjectedBundle.vcproj:
539         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
540
541 2010-11-18  Hayato Ito  <hayato@chromium.org>
542
543         Reviewed by Shinichiro Hamaji.
544
545         [NRWT] Fix a break of '--new-baseline' feature in pixel_test
546         which is caused by r72249.
547
548         https://bugs.webkit.org/show_bug.cgi?id=49751
549
550         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
551
552 2010-11-18  Adam Roben  <aroben@apple.com>
553
554         Implement layoutTestController.waitForPolicyDelegate on Windows
555
556         Fixes <http://webkit.org/b/25038> <rdar://problem/6790213>.
557
558         Reviewed by Simon Fraser.
559
560         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
561         (LayoutTestController::waitForPolicyDelegate): Implemented by porting
562         code from LayoutTestControllerMac.mm.
563
564 2010-11-18  Hayato Ito  <hayato@chromium.org>
565
566         Reviewed by Shinichiro Hamaji.
567
568         [NRWT] Fix a break of '--new-baseline' feature which is caused by r72249.
569
570         https://bugs.webkit.org/show_bug.cgi?id=49740
571
572         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
573
574 2010-11-18  Tony Chang  <tony@chromium.org>
575
576         Reviewed by Adam Barth.
577
578         run platform/chromium/plugins/refcount-leaks.html on all platforms
579         https://bugs.webkit.org/show_bug.cgi?id=49485
580
581         Add PluginObject.testObjectCount which returns the number of allocated
582         TestObjects. Add PluginObject.testCreateTestObject which allocates
583         and returns a TestObject.
584
585         Add TestObject.refCount which returns the number of refs on the
586         TestObject.
587
588         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
589         (pluginGetProperty):
590         (pluginInvoke):
591         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
592         (getTestObjectCount):
593         (testAllocate):
594         (testDeallocate):
595         (testGetProperty):
596         (testScriptObjectInvoke): Release a ref to a plugin object that was
597             previously leaked.
598         * DumpRenderTree/TestNetscapePlugIn/TestObject.h:
599
600 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
601
602         Reviewed by Adam Roben.
603
604         WebKitTools projects (WebKitTestRunner, record-memory-win) should use vsprops for common build settings
605         https://bugs.webkit.org/show_bug.cgi?id=49711
606
607         * WebKitTestRunner/Configurations/WebKitTestRunnerCFLite.vsprops: Added.
608         * WebKitTestRunner/Configurations/WebKitTestRunnerCommon.vsprops: Added.
609         * WebKitTestRunner/Configurations/WebKitTestRunnerCoreFoundation.vsprops: Added.
610         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
611         * record-memory-win/record-memory-win-common.vsprops: Added.
612         * record-memory-win/record-memory-win.vcproj:
613
614 2010-11-18  Andras Becsi  <abecsi@inf.u-szeged.hu>
615
616         Reviewed by Csaba Osztrogonác.
617
618         [Qt][WK2] Only add user agent strings to the list which aren't listed yet.
619
620         * MiniBrowser/qt/BrowserWindow.cpp:
621         (BrowserWindow::updateUserAgentList):
622
623 2010-11-17  Adam Roben  <aroben@apple.com>
624
625         Don't trigger Windows builds when chromium-win test results change
626
627         Reviewed by Mark Rowe.
628
629         * Scripts/webkitpy/common/config/build.py:
630         (_should_file_trigger_build): Changed the regular expression used to
631         search for directory names to only consider full directory names,
632         rather than matching directory names where the string we care about is
633         a suffix of the name.
634
635         * Scripts/webkitpy/common/config/build_unittest.py:
636         (ShouldBuildTest): Added a test.
637
638 2010-11-17  Hayato Ito  <hayato@chromium.org>
639
640         Reviewed by Shinichiro Hamaji.
641
642         Remove a uri member from TestInput class.
643
644         https://bugs.webkit.org/show_bug.cgi?id=49691
645
646         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
647         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
648
649 2010-11-17  Adam Roben  <aroben@apple.com>
650
651         Move FindSafari's settings into a vsprops file
652
653         Fixes <http://webkit.org/b/49699> FindSafari should use a vsprops file
654
655         Reviewed by Steve Falkenburg.
656
657         * FindSafari/FindSafari.vcproj:
658         * FindSafari/FindSafariCommon.vsprops: Added.
659
660 2010-11-17  Adam Roben  <aroben@apple.com>
661
662         Use a minimum font size of 0 in DRT on Windows
663
664         This is the Windows equivalent of r72164.
665
666         Rubber-stamped by Dave Hyatt.
667
668         * DumpRenderTree/win/DumpRenderTree.cpp:
669         (resetDefaultsToConsistentValues):
670
671 2010-11-17  Dirk Pranke  <dpranke@chromium.org>
672
673         Reviewed by Andreas Kling.
674
675         webkit-patch pretty-diff shouldn't need to be run from a checkout root in git
676         https://bugs.webkit.org/show_bug.cgi?id=49639
677
678         * Scripts/webkitpy/common/checkout/scm.py:
679
680 2010-11-17  Mihai Parparita  <mihaip@chromium.org>
681
682         Reviewed by Tony Chang.
683
684         Rebaseline server: add loupe for image diffs
685         https://bugs.webkit.org/show_bug.cgi?id=49692
686         
687         Add a loupe (magnifiying glass) for inspecting image diffs. Shows an
688         enlarged area of the expected, actual and diff images side by side.
689
690         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
691         * Scripts/webkitpy/tool/commands/data/rebaselineserver/loupe.js: Added.
692         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
693         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
694
695 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
696
697         Reviewed by Adam Roben.
698
699         WebKitLauncherWin should use vsprops for shared build settings
700         https://bugs.webkit.org/show_bug.cgi?id=49696
701
702         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
703         * WebKitLauncherWin/WebKitLauncherWinCommon.vsprops: Added.
704
705 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
706
707         Reviewed by Adam Roben.
708
709         WinLauncher should use vsprops for shared build settings
710         https://bugs.webkit.org/show_bug.cgi?id=49695
711
712         * WinLauncher/WinLauncher.cpp:
713         (_tWinMain):
714         (PrintView):
715         (WndProc):
716         (MyEditProc):
717         * WinLauncher/WinLauncher.vcproj:
718         * WinLauncher/WinLauncherCommon.vsprops: Added.
719
720 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
721
722         Delete unused file unintentionally added in last change.
723
724         * DumpRenderTree/win/DumpRenderTreeCURL.vsprops: Removed.
725
726 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
727
728         Reviewed by Adam Roben.
729
730         DumpRenderTree on Windows should use a vsprops file for shared project settings
731         https://bugs.webkit.org/show_bug.cgi?id=49690
732
733         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
734         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePluginCommon.vsprops: Added.
735         * DumpRenderTree/win/DumpRenderTree.vcproj:
736         * DumpRenderTree/win/DumpRenderTreeApple.vsprops: Added.
737         * DumpRenderTree/win/DumpRenderTreeCFLite.vsprops: Added.
738         * DumpRenderTree/win/DumpRenderTreeCURL.vsprops: Added.
739         * DumpRenderTree/win/DumpRenderTreeCairo.vsprops: Added.
740         * DumpRenderTree/win/DumpRenderTreeCommon.vsprops: Added.
741         * DumpRenderTree/win/ImageDiff.vcproj:
742         * DumpRenderTree/win/ImageDiffCommon.vsprops: Added.
743
744 2010-11-17  Hayato Ito  <hayato@chromium.org>
745
746         Refactor TestTypeBase.compare_output().
747
748         Introduce a TestOutput class and update compare_output() of each test
749         types so that they can take both actual and expected TestOutput objects.
750
751         https://bugs.webkit.org/show_bug.cgi?id=49431
752
753         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
754         * Scripts/webkitpy/layout_tests/layout_package/test_output.py: Added.
755         * Scripts/webkitpy/layout_tests/port/base.py:
756         * Scripts/webkitpy/layout_tests/port/chromium.py:
757         * Scripts/webkitpy/layout_tests/port/dryrun.py:
758         * Scripts/webkitpy/layout_tests/port/test.py:
759         * Scripts/webkitpy/layout_tests/port/webkit.py:
760         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
761         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
762         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
763         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
764
765 2010-11-17  Adam Roben  <aroben@apple.com>
766
767         Make each Windows Test builder use the same OS for all its slaves
768
769         Currently each WebKit1 Windows Test builder has one Windows XP slave
770         and one Windows 7 slave. Having different OSes in a single builder
771         meant that the builders would alternate between red and green when
772         there were OS-specific failures.
773
774         Fixes <http://webkit.org/b/49688> Windows Test builders alternate
775         between red and green because of multiple OS versions
776
777         Reviewed by Mark Rowe.
778
779         * BuildSlaveSupport/build.webkit.org-config/config.json: Renamed
780         "Windows Release (Tests)" to "Windows 7 Release (Tests)", "Windows
781         Debug (Tests)" to "Windows XP Debug (Tests)", and "Windows Release
782         (WebKit2 Tests)" to "Windows 7 Release (WebKit2 Tests)". Updated the
783         builders for the two WebKit1 builders so that they would all use the
784         appropriate OS.
785
786 2010-11-17  James Robinson  <jamesr@chromium.org>
787
788         Reviewed by Dan Bernstein.
789
790         [chromium] Set minimum font size to 0 for DRT
791         https://bugs.webkit.org/show_bug.cgi?id=49677
792
793         Followup for http://trac.webkit.org/changeset/72141.
794
795         Test: http/tests/misc/acid3.html
796
797         * DumpRenderTree/chromium/WebPreferences.cpp:
798         (WebPreferences::reset):
799
800 2010-11-17  Andras Becsi  <abecsi@webkit.org>
801
802         Unreviewed build fix.
803
804         [Qt] Add missing file I forgot in r72220.
805
806         * MiniBrowser/MiniBrowser.qrc: Added.
807
808 2010-11-17  Andras Becsi  <abecsi@webkit.org>
809
810         Reviewed by Andreas Kling.
811
812         [Qt][WK2] Make it possible to set custom user agent strings in MiniBrowser.
813         https://bugs.webkit.org/show_bug.cgi?id=49627
814
815         Add a user agent dialog and a resource file for useragentlist.txt to MiniBrowser.
816
817         * MiniBrowser/MiniBrowser.qrc: Added.
818         Add it here to prevent qmake from detecting it
819         since it needs to be copied to the build directory.
820         * MiniBrowser/qt/BrowserWindow.cpp:
821         (BrowserWindow::BrowserWindow):
822         (BrowserWindow::updateUserAgentList):
823         (BrowserWindow::showUserAgentDialog):
824         * MiniBrowser/qt/BrowserWindow.h:
825         * MiniBrowser/qt/MiniBrowser.pro:
826
827 2010-11-17  Adam Roben  <aroben@apple.com>
828
829         Make changes to the Mac WebKit2 Skipped file trigger Windows builds
830
831         We use that file on Windows!
832
833         Fixes <http://webkit.org/b/49643> Changes to
834         LayoutTests/platform/mac-wk2 should trigger a Windows build, but don't
835
836         Reviewed by Steve Falkenburg.
837
838         * Scripts/webkitpy/common/config/build.py:
839         (_should_file_trigger_build): Added the mac-wk2 directory and make it
840         trigger builds on SnowLeopard and Windows.
841
842         * Scripts/webkitpy/common/config/build_unittest.py:
843         (ShouldBuildTest.test_should_build): Added a test.
844
845 2010-11-17  Satish Sampath  <satish@chromium.org>
846
847         Reviewed by Jeremy Orlow.
848
849         Clear the speech input mock explicitly before each test.
850         https://bugs.webkit.org/show_bug.cgi?id=49660
851
852         * DumpRenderTree/chromium/LayoutTestController.cpp:
853         (LayoutTestController::setMockSpeechInputResult):
854         * DumpRenderTree/chromium/LayoutTestController.h:
855         * DumpRenderTree/chromium/WebViewHost.cpp:
856         (WebViewHost::speechInputController):
857         (WebViewHost::reset): Invoke speech input mock's clearResults method.
858         * DumpRenderTree/chromium/WebViewHost.h:
859         (WebViewHost::speechInputControllerMock): Moved mock from LayoutTestController.
860
861 2010-11-17  Alexander Pavlov  <apavlov@chromium.org>
862
863         Unreviewed, specify IRC nickname in committers.py.
864
865         * Scripts/webkitpy/common/config/committers.py:
866
867 2010-11-17  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
868
869         Reviewed by Tony Chang.
870
871         [NRWT] Make http locking similar to perl implementation
872         https://bugs.webkit.org/show_bug.cgi?id=49187
873
874         * Scripts/webkitpy/common/system/file_lock.py: Added.
875         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
876         * Scripts/webkitpy/layout_tests/port/http_lock.py:
877
878 2010-11-17  MORITA Hajime  <morrita@google.com>
879
880         Reviewed by Kent Tamura.
881
882         [Chromium][DRT] EventSender.contextClick() should aware spellchecking
883         https://bugs.webkit.org/show_bug.cgi?id=49366
884
885         - EvenSender: Checked WebContextMenuData.misspelledWord and added extra context menu entries
886           according to the spellchecker suggestion
887         - MockSpellCheck: add fillSuggestionList to provide fake suggestions.
888
889         test_expectations.txt will be changed after this change is ported to test_shell.
890         
891         * DumpRenderTree/chromium/EventSender.cpp:
892         (makeMenuItemStringsFor):
893         (EventSender::contextClick):
894         * DumpRenderTree/chromium/MockSpellCheck.cpp:
895         (MockSpellCheck::fillSuggestionList):
896         (MockSpellCheck::initializeIfNeeded):
897         * DumpRenderTree/chromium/MockSpellCheck.h:
898         * DumpRenderTree/chromium/WebViewHost.cpp:
899         (WebViewHost::mockSpellCheck):
900         * DumpRenderTree/chromium/WebViewHost.h:
901
902 2010-11-16  Dave Hyatt  <hyatt@apple.com>
903
904         Reviewed by Sam Weinig.
905
906         Make sure the pref for minimum font size is just explicitly set to 0 every time,
907         since the bots have 1 stuck in their plist otherwise.
908
909         * DumpRenderTree/mac/DumpRenderTree.mm:
910         (resetDefaultsToConsistentValues):
911
912 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
913
914         Reviewed by Tony Chang.
915
916         Rebaseline server: compute diffs client-side
917         https://bugs.webkit.org/show_bug.cgi?id=49640
918         
919         The image diff output from the DRT is pretty bad on some ports (at
920         least the Mac one), so it's better to compute diffs on the client by
921         using <canvas>.
922
923         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
924         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
925         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
926
927 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
928
929         Reviewed by Tony Chang.
930
931         Rebaseline server: display test results
932         https://bugs.webkit.org/show_bug.cgi?id=49626
933         
934         Adds basic result display to the rebaseline server. On the Python side
935         this involves:
936         - Parsing the unexpected_results.json into a dictionary.
937         - Serving it as JSON under /results.json.
938         (the JSON -> dict -> JSON transform isn't strictly necessary right now,
939         but I'll need to have access to the parsed results on the Python side
940         for follow-up changes).
941         
942         On the web UI side this adds:
943         - Markup for display image and text results (expected, actual, diff),
944           and JS for populating it.
945         - Markup for breaking down test results by failure type and directory,
946           and JS for populating it.
947
948         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
949         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
950         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
951         * Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js: Added.
952         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
953
954 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
955
956         Reviewed by Ojan Vafai.
957
958         See Chromium issue http://codereview.chromium.org/5133001/ - we
959         are modifying the build bots to pass in the builder name with
960         the "GPU" string appended instead of appending it in the code.
961
962         https://bugs.webkit.org/show_bug.cgi?id=49636
963
964         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
965
966 2010-11-16  Steve Falkenburg  <sfalken@apple.com>
967
968         Reviewed by Adam Roben.
969
970         Disable LTCG for Windows Release builds. Add new Release_LTCG configuration.
971         https://bugs.webkit.org/show_bug.cgi?id=49632
972
973         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
974         * DumpRenderTree/win/DumpRenderTree.vcproj:
975         * DumpRenderTree/win/ImageDiff.vcproj:
976         * MiniBrowser/MiniBrowser.vcproj:
977         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
978         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
979         * WebKitTestRunner/win/InjectedBundle.vcproj:
980         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
981         * WinLauncher/WinLauncher.vcproj:
982         * record-memory-win/record-memory-win.vcproj:
983
984 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
985
986         Reviewed by Ojan Vafai.
987
988         new-run-webkit-tests: rename TestInfo to TestInput, move image hash to work thread
989
990         Rename the TestInfo class to TestInput to be clearer about its
991         function, and move the checksum-reading code into dump_render_tree_thread
992         to avoid cross-thread access.
993
994         https://bugs.webkit.org/show_bug.cgi?id=49573
995
996         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
997         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
998         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
999
1000 2010-11-16  Dave Hyatt  <hyatt@apple.com>
1001
1002         Fix bustage. Remove the minimum font size pref setting in DumpRenderTree.
1003
1004         * DumpRenderTree/mac/DumpRenderTree.mm:
1005         (resetDefaultsToConsistentValues):
1006
1007 2010-11-16  Stephanie Lewis  <slewis@apple.com>
1008
1009         Reviewed by Geoff Garen.
1010
1011         <rdar://problem/8624267> Leak creating offscreen webview running fast/dom tests
1012
1013         Use shared DeviceOrientationProviderMock. The old allocation method created a new WebDeviceOrientationProviderMock 
1014         for every WebView, and the WebKit API doesn't support that ownership model.
1015
1016         I also fixed a leak in setMockDeviceOrientation by adding a missing release.
1017
1018         * DumpRenderTree/mac/DumpRenderTree.mm:
1019         (createWebViewAndOffscreenWindow):
1020         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1021         (LayoutTestController::setMockDeviceOrientation):
1022
1023 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
1024
1025         Reviewed by Ojan Vafai.
1026
1027         new-run-webkit-tests: log the process id in --verbose mode
1028
1029         As part of the switch to multi-process mode, we should log the
1030         process id when running w/ --verbose.
1031
1032         https://bugs.webkit.org/show_bug.cgi?id=49571
1033
1034         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
1035
1036 2010-11-16  Adam Roben  <aroben@apple.com>
1037
1038         Only trigger Windows builds when files that we actually use on Windows
1039         are changed
1040
1041         A new Scheduler subclass, PlatformSpecificScheduler, has been added.
1042         It uses the new webkitpy.common.config.build module to determine
1043         whether a particular change should trigger a build on a particular
1044         platform. The Windows builders have been switched to use a
1045         PlatformSpecificScheduler.
1046
1047         The logic to determine whether or not a particular change should
1048         trigger a build on a given platform has only been implemented/tested
1049         for Windows. I tried to make it easy to add more platforms in the
1050         future, but I don't have enough familiarity with all platforms to be
1051         able to implement it for them.
1052
1053         Fixes <http://webkit.org/b/49407> Windows builders kick off builds for
1054         lots irrelevant changes (e.g., rebaselining Chromium test results)
1055
1056         Reviewed by Eric Seidel.
1057
1058         * BuildSlaveSupport/build.webkit.org-config/config.json: Use a
1059         PlatformSpecificScheduler for the Windows builders.
1060
1061         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1062         (PlatformSpecificScheduler.__init__): Added. Stores our platform, sets
1063         up a ChangeFilter that filters to our branch and will call through to
1064         our filter method, and calls up to our base class.
1065         (PlatformSpecificScheduler.filter): Calls through to
1066         build.should_build to find out whether we should trigger a build for
1067         this change.
1068
1069         * Scripts/webkitpy/common/net/build.py: Added.
1070         (_should_file_trigger_build): Uses a set of directories and regexp
1071         patterns to determine whether the given file should trigger a build on
1072         the given platform. As mentioned earlier, this has only been
1073         implemented for Windows, though I did try to make some guesses about
1074         other platforms.
1075         (should_build): Returns true if any of the files should trigger a
1076         build on the given platform.
1077
1078         * Scripts/webkitpy/common/net/build_unittest.py: Added.
1079         (ShouldBuildTest.test_should_build): Does some basic testing to make
1080         sure we're triggering builds for the right files. It only tests
1081         Windows for now, though I tried to make some guesses about other
1082         platforms.
1083
1084 2010-11-16  Simon Hausmann  <simon.hausmann@nokia.com>
1085
1086         Reviewed by Andreas Kling.
1087
1088         [Qt] Remove synchronous QWebPage::checkPermissions signal
1089         https://bugs.webkit.org/show_bug.cgi?id=46810
1090
1091         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1092         (WebCore::WebPage::WebPage):
1093         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1094         (LayoutTestController::grantDesktopNotificationPermission): When granting
1095         permission, grant it directly on the QWebPage/Frame, that will remember it.
1096         * QtTestBrowser/webpage.cpp:
1097         (WebPage::WebPage):
1098         * QtTestBrowser/webpage.h:
1099
1100
1101 2010-11-16  Andras Becsi  <abecsi@inf.u-szeged.hu>
1102
1103         Reviewed by Csaba Osztrogonác.
1104
1105         [Qt] Add isPrinting and setPrinting methods to DRT's LayoutTestController.
1106
1107         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1108         (LayoutTestController::reset):
1109         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1110         (LayoutTestController::isPrinting):
1111         (LayoutTestController::setPrinting):
1112
1113 2010-11-16  Balazs Kelemen  <kbalazs@webkit.org>
1114
1115         Reviewed by Andreas Kling.
1116
1117         [Qt][WK2] Add a way to turn on tiled backing store in MiniBrowser
1118         https://bugs.webkit.org/show_bug.cgi?id=49587
1119
1120         * MiniBrowser/qt/BrowserView.cpp:
1121         (BrowserView::BrowserView): Added BackingStoreType
1122         argument to ctor. Propagating it to the QGraphicsWKView.
1123         * MiniBrowser/qt/BrowserView.h:
1124         * MiniBrowser/qt/BrowserWindow.cpp:
1125         (BrowserWindow::BrowserWindow):
1126         (BrowserWindow::newWindow): Create the new BrowserWindow
1127         with the same BackingStoreType that this instance has.
1128         * MiniBrowser/qt/BrowserWindow.h: Added BackingStoreType member.
1129         * MiniBrowser/qt/main.cpp:
1130         (main): Use tiled backing store if got -tiled command line argument.
1131         Avoid redundant copy of command line args.
1132
1133 2010-11-16  John Knottenbelt  <jknotten@chromium.org>
1134
1135         Reviewed by Jeremy Orlow.
1136
1137         Move DeviceOrientationClientMock from LayoutTestController to WebViewHost.
1138         https://bugs.webkit.org/show_bug.cgi?id=48506
1139
1140         This ensures that there is one mock per page WebView / WebViewHost. This is a
1141         design constraint of the DeviceOrientationClientMock because the m_controller
1142         field can only store a pointer to a single instance of the corresponding
1143         DeviceOrientationController.
1144
1145         Test: fast/dom/DeviceOrientation/no-page-cache.html
1146
1147         * DumpRenderTree/chromium/LayoutTestController.cpp:
1148         (LayoutTestController::setMockDeviceOrientation):
1149         (LayoutTestController::deviceOrientationClientMock):
1150         * DumpRenderTree/chromium/LayoutTestController.h:
1151         * DumpRenderTree/chromium/WebViewHost.cpp:
1152         (WebViewHost::deviceOrientationClientMock):
1153         (WebViewHost::deviceOrientationClient):
1154         (WebViewHost::reset):
1155         * DumpRenderTree/chromium/WebViewHost.h:
1156
1157 2010-11-15  Andras Becsi  <abecsi@webkit.org>
1158
1159         Reviewed by Andreas Kling.
1160
1161         [Qt][WK2] Avoid polling in WebKitTestRunner.
1162         https://bugs.webkit.org/show_bug.cgi?id=49542
1163
1164         Make the WK2 testing session about 10-15% faster and decrease random flakiness
1165         resulting from timing skew by removing the millisecond polling from TestControllerQt.
1166
1167         * WebKitTestRunner/TestController.cpp:
1168         (WTR::TestController::didFinishLoadForFrame):
1169         * WebKitTestRunner/TestController.h:
1170         * WebKitTestRunner/TestInvocation.cpp:
1171         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
1172         * WebKitTestRunner/mac/TestControllerMac.mm:
1173         (WTR::TestController::notifyDone): Add method stub.
1174         * WebKitTestRunner/qt/TestControllerQt.cpp:
1175         (WTR::TestControllerRunLoop::instance):
1176         (WTR::TestControllerRunLoop::start):
1177         (WTR::TestControllerRunLoop::stop):
1178         (WTR::TestControllerRunLoop::TestControllerRunLoop):
1179         (WTR::TestControllerRunLoop::timerEvent):
1180         (WTR::TestController::notifyDone):
1181         (WTR::TestController::platformRunUntil):
1182         * WebKitTestRunner/win/TestControllerWin.cpp:
1183         (WTR::TestController::notifyDone): Add method stub.
1184
1185 2010-11-15  Darin Adler  <darin@apple.com>
1186
1187         * Scripts/webkitpy/common/net/bugzilla: Added property svn:ignore.
1188
1189 2010-11-15  Benjamin Poulain  <benjamin.poulain@nokia.com>
1190
1191         Reviewed by Andreas Kling.
1192
1193         [Qt] [WK2] Add Ctrl+L shortcut to Qt Minibrowser
1194         https://bugs.webkit.org/show_bug.cgi?id=49544
1195
1196         Add the shortcut to BrowserWindow.
1197
1198         * MiniBrowser/qt/BrowserWindow.cpp:
1199         (BrowserWindow::BrowserWindow):
1200         (BrowserWindow::openLocation):
1201         * MiniBrowser/qt/BrowserWindow.h:
1202
1203 2010-11-14  David Levin  <levin@chromium.org>
1204
1205         Reviewed by Daniel Bates.
1206
1207         check-webkit-style should detect PassRefPtr usage in functions.
1208         https://bugs.webkit.org/show_bug.cgi?id=49513
1209
1210         * Scripts/webkitpy/style/checkers/cpp.py:
1211         (check_for_function_lengths): Revert a comment change that I
1212          accidentally made in r71986.
1213         (check_pass_ptr_usage): Added the code to do the check.
1214         (process_line): Added the call to check_pass_ptr_usage.
1215         (CppChecker): Added the new error category.
1216         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
1217         (CppStyleTestBase::perform_pass_ptr_check): Runs the new check for
1218          testing purposes.
1219         (PassPtrTest::*): The class/functions to unit test the new
1220          functionality.
1221
1222 2010-11-14  David Levin  <levin@chromium.org>
1223
1224         Reviewed by Shinichiro Hamaji.
1225
1226         check-webkit-style function detection and the line count style checks should be separate.
1227         https://bugs.webkit.org/show_bug.cgi?id=49512
1228
1229         * Scripts/webkitpy/style/checkers/cpp.py: Do the separation.
1230         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Adjust the test to
1231           call the detection function and fix line counts in two places now that
1232           the code really only counts the lines in the body of the function.
1233
1234 2010-11-14  Andreas Kling  <kling@webkit.org>
1235
1236         Reviewed by Antonio Gomes.
1237
1238         http/tests/plugins tests print "Unhandled variable" to stderr under Qt
1239         https://bugs.webkit.org/show_bug.cgi?id=33438
1240
1241         Stifle "Unhandled variable" warning to match what the other
1242         TestNetscapePlugin does (NPP_GetValue in TestNetscapePlugIn/main.cpp
1243         simply returns NPERR_GENERIC_ERROR for unhandled variables.)
1244
1245         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
1246         (webkit_test_plugin_get_value):
1247
1248 2010-11-13  David Levin  <levin@chromium.org>
1249
1250         Reviewed by Eric Seidel.
1251
1252         check-webkit-style function detection crashes on functions in templates.
1253         https://bugs.webkit.org/show_bug.cgi?id=49504
1254
1255         For example "template <bool shouldClose(const Element*)>".
1256
1257         * Scripts/webkitpy/style/checkers/cpp.py: If not function is detected, bail out.
1258         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Modified the complex
1259           function detection test to expose this issue.
1260
1261 2010-11-12  David Levin  <levin@chromium.org>
1262
1263         Reviewed by Shinichiro Hamaji.
1264
1265         check-webkit-style function detection doesn't detect indented functions declaractions.
1266         https://bugs.webkit.org/show_bug.cgi?id=49446
1267
1268         Indented function declarations occur inside class definitions, so
1269         they are a pretty common (and worth detecting).
1270
1271         * Scripts/webkitpy/style/checkers/cpp.py:
1272           Changed regex to allow indentation.
1273           Changed the function start detection to only happen when not in a
1274           function.
1275           Changed function end detection to work based on matching braces
1276           instead of finding a close brace at the beginning of the line.
1277           Fixed close_expression to do what it says when it doesn't find
1278           the close.
1279         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Indented function test.
1280
1281 2010-11-12  Daniel Bates  <dbates@rim.com>
1282
1283         Rubber-stamped by Nikolas Zimmermann.
1284
1285         Make do-file-rename rename files in the directory WebKit2.
1286
1287         This also makes the list of searched directories in do-file-rename
1288         consistent with the list of searched directories in the script do-webcore-rename.
1289
1290         * Scripts/do-file-rename:
1291
1292 2010-11-12  Dirk Pranke  <dpranke@chromium.org>
1293
1294         Reviewed by James Robinson.
1295
1296         Attempt yet again to land the fix for bug 49360 (respecting
1297         set-webkit-configuration). We need to handle the cases where
1298         trying to run webkit-build-directory to find out where the
1299         default configuration might be fails (that shows up on some
1300         Chromium bots that apparently don't have perl installed).
1301
1302         https://bugs.webkit.org/show_bug.cgi?id=49360
1303
1304         * Scripts/webkitpy/layout_tests/port/config.py:
1305         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Added.
1306         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1307
1308 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
1309
1310         Reviewed by Adam Barth.
1311
1312         webkit-patch land logs commit message twice
1313         https://bugs.webkit.org/show_bug.cgi?id=49482
1314         
1315         Don't log the comment that's used by bugzilla.close_bug_as_fixed, since
1316         that's just the commit message, which we now output following the
1317         actual commit.
1318
1319         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
1320
1321 2010-11-12  Tony Chang  <tony@chromium.org>
1322
1323         Reviewed by Adam Barth.
1324
1325         run platform/chromium/plugins/script-object-invoke.html on all platforms
1326         https://bugs.webkit.org/show_bug.cgi?id=49280
1327
1328         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1329         (testScriptObjectInvoke):
1330         (pluginInvoke):
1331
1332 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
1333
1334         Reviewed by Adam Barth.
1335
1336         webkit-patch land-cowboy (and other commands) should display the committed SVN revision after landing a patch
1337         https://bugs.webkit.org/show_bug.cgi?id=49471
1338
1339         Right now there's no indication of a successful commit.
1340         
1341         * Scripts/webkitpy/tool/commands/download_unittest.py:
1342         * Scripts/webkitpy/tool/steps/commit.py:
1343
1344 2010-11-12  Kinuko Yasuda  <kinuko@chromium.org>
1345
1346         Reviewed by Ojan Vafai.
1347
1348         Include detailed test modifiers in results.json and enable incremental uploading for non-layout tests
1349         https://bugs.webkit.org/show_bug.cgi?id=49354
1350
1351         Also moved/integrated the upload method from run_webkit_tests.py to json_results_generator.py.
1352
1353         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
1354         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1355         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
1356         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1357
1358 2010-11-12  Tony Chang  <tony@chromium.org>
1359
1360         Reviewed by Kent Tamura.
1361
1362         [chromium] Load a blank page when closing DRT WebViewHosts
1363         https://bugs.webkit.org/show_bug.cgi?id=49417
1364
1365         In test_shell, when we close a window, we first load about:blank
1366         and run GC to fire the destruction logic of the page.  In test_shell,
1367         this happens in ~TestShell.  In DRT, we manage virtual windows as
1368         WebViewHosts, so we need to replicate this logic in ~WebViewHost.
1369
1370         Managing focus is a bit trickier, since normally you would get window
1371         events to reset the focus to the right window.  We do it manually
1372         instead.
1373
1374         This fixes 2 plugin tests on DRT.
1375
1376         * DumpRenderTree/chromium/TestShell.cpp:
1377         (TestShell::~TestShell): Move window close logic to ~WebViewHost.
1378         (TestShell::closeWindow): Move widget close() to WVH destructor and
1379             manage focus manually.
1380         (TestShell::closeRemainingWindows): Comment typo.
1381         * DumpRenderTree/chromium/WebViewHost.cpp:
1382         (WebViewHost::closeWidget):
1383         (invokeCloseWidget):
1384         (WebViewHost::closeWidgetSoon): Close the window asynchronously (matches
1385             test_shell).
1386         (WebViewHost::~WebViewHost):
1387         * DumpRenderTree/chromium/WebViewHost.h:
1388
1389 2010-11-12  David Levin  <levin@chromium.org>
1390
1391         Reviewed by Adam Barth.
1392
1393         cpp.py needs some tidying.
1394         https://bugs.webkit.org/show_bug.cgi?id=49443
1395
1396         * Scripts/webkitpy/style/checkers/cpp.py:
1397           Fix parameter ordering for iteratively_replace_matches_with_char
1398           to mirror other similar functions (and improved comments, etc.)
1399           Also, use a python-ism (for/else) to get rid of a variable that
1400           was there to determine if something had been found in the loop.
1401
1402 2010-11-12  Robert Kroeger  <rjkroege@chromium.org>
1403
1404         Reviewed by James Robinson.
1405
1406         [Chromium/DRT] Make EventSendingController send complete touch events.
1407         https://bugs.webkit.org/show_bug.cgi?id=49285
1408
1409         * DumpRenderTree/chromium/EventSender.cpp:
1410         (EventSender::addTouchPoint):
1411         (EventSender::updateTouchPoint):
1412         (EventSender::sendCurrentTouchEvent):
1413
1414 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
1415
1416         Unreviewed rollout of r71858.
1417
1418         Rollout out r71858 since it breaks new-run-webkit-httpd as used by the
1419         NaCl tests.
1420
1421         * Scripts/webkitpy/layout_tests/port/config.py:
1422         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Removed.
1423         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1424
1425 2010-11-12  Benjamin Poulain  <benjamin.poulain@nokia.com>
1426
1427         Reviewed by Kenneth Rohde Christiansen.
1428
1429         [Qt] TestController::runUntil doesn't honor the timeout parameter
1430         https://bugs.webkit.org/show_bug.cgi?id=48941
1431
1432         Implement the missing timeout of WebKitTestRunner for Qt.
1433
1434         A QElapsedTimer has been added to RunUntilConditionLoop in order
1435         to measure how long the loop has been running. When the timer
1436         is bigger or equal than the timeout, the loop ends even if the
1437         condition is not met.
1438
1439         * WebKitTestRunner/qt/TestControllerQt.cpp:
1440         (WTR::RunUntilConditionLoop::start):
1441         (WTR::RunUntilConditionLoop::run):
1442         (WTR::RunUntilConditionLoop::timerEvent):
1443         (WTR::TestController::platformRunUntil):
1444
1445 2010-11-11  David Levin  <levin@chromium.org>
1446
1447         Reviewed by Shinichiro Hamaji.
1448
1449         check-webkit-style function detection doesn't handle templates with spaces.
1450         https://bugs.webkit.org/show_bug.cgi?id=49427
1451
1452         * Scripts/webkitpy/style/checkers/cpp.py: Stripped out templates when
1453           finding the function name.
1454         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test have a
1455           template with a space.
1456
1457 2010-11-11  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
1458
1459         Reviewed by Shinichiro Hamaji.
1460
1461         Enable check-webkit-style on Qt files
1462         https://bugs.webkit.org/show_bug.cgi?id=48322
1463
1464         Adding exception for _q_... private slots and
1465         ..._data() methods in tests
1466
1467         * Scripts/webkitpy/style/checker.py: enable Qt dirs, add exceptions
1468         * Scripts/webkitpy/style/checker_unittest.py: improve unittests
1469
1470 2010-11-11  David Levin  <levin@chromium.org>
1471
1472         Reviewed by Shinichiro Hamaji.
1473
1474         check-webkit-style function detection doesn't work for templates and destructors.
1475         https://bugs.webkit.org/show_bug.cgi?id=49425
1476
1477         * Scripts/webkitpy/style/checkers/cpp.py: Changed regex for templates and destructors.
1478         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test hit this issue.
1479
1480 2010-11-11  Joone Hur  <joone@kldp.org>
1481
1482         Reviewed by Antonio Gomes.
1483
1484         [GTK][DRT] Implement LayoutTestController::nodesFromRect
1485         https://bugs.webkit.org/show_bug.cgi?id=46598
1486
1487         Support nodesFromRect in DRT
1488
1489         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1490         (LayoutTestController::nodesFromRect):
1491
1492 2010-11-11  MORITA Hajime  <morrita@google.com>
1493
1494         Reviewed by Kent Tamura.
1495
1496         editing/selection/context-menu-on-text.html fails on chromium
1497         https://bugs.webkit.org/show_bug.cgi?id=45898
1498
1499         Enhance Chromium's EventSender::contextClick() to return
1500         array of menu strings. The implementation incomplete because
1501         Although actual context menu is implemented by the browser,
1502         it reflects editability of underlying context thus makes
1503         context-menu-on-text.html passed.
1504         
1505         * DumpRenderTree/chromium/EventSender.cpp:
1506         (makeMenuItemStringsFor): Added.
1507         (EventSender::contextClick): Now returns array of menu strings.
1508         * DumpRenderTree/chromium/WebViewHost.cpp:
1509         (WebViewHost::showContextMenu): Now storing given WebContextMenuData object.
1510         (WebViewHost::clearContextMenuData): Added.
1511         (WebViewHost::lastContextMenuData): Added.
1512         * DumpRenderTree/chromium/WebViewHost.h:
1513
1514 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1515
1516         Unreviewed, build fix.
1517
1518         fix build breakage caused by fix for bug 49380 (r71858).
1519
1520         config_standalone needs to ensure that the packages it needs are
1521         in sys.path.
1522
1523         https://bugs.webkit.org/show_bug.cgi?id=49419
1524
1525         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
1526
1527 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1528
1529         Reviewed by Adam Roben.
1530
1531         Fix NRWT to respect set-webkit-configuration again :(
1532
1533         This change fixes a typo in config.py that was causing the wrong
1534         value to be read initially and us never actually looking into
1535         the filesystem to get the default configuration.
1536         
1537         * Scripts/webkitpy/layout_tests/port/config.py:
1538         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
1539         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1540
1541 2010-11-11  Eric Seidel  <eric@webkit.org>
1542
1543         Reviewed by Adam Barth.
1544
1545         Split out Bug, Attachment and CommitterValidator from bugzilla.py
1546         https://bugs.webkit.org/show_bug.cgi?id=49403
1547
1548         CommitterValidator really had nothing to do with bugzilla.py.
1549         I've put it next to committers.py for now since its rather tied to that file.
1550
1551         * Scripts/webkitpy/common/config/committervalidator.py: Added.
1552         * Scripts/webkitpy/common/config/committervalidator_unittest.py: Added.
1553         * Scripts/webkitpy/common/net/bugzilla/__init__.py:
1554         * Scripts/webkitpy/common/net/bugzilla/attachment.py: Added.
1555         * Scripts/webkitpy/common/net/bugzilla/bug.py: Added.
1556         * Scripts/webkitpy/common/net/bugzilla/bug_unittest.py: Added.
1557         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
1558         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py:
1559         * Scripts/webkitpy/tool/bot/feeders.py:
1560         * Scripts/webkitpy/tool/commands/queues.py:
1561
1562 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1563
1564         Reviewed by Eric Seidel.
1565
1566         fix missing 'pid' argument for http lockfile checking on win32
1567         https://bugs.webkit.org/show_bug.cgi?id=49363
1568
1569         * Scripts/webkitpy/common/system/executive.py:
1570
1571 2010-11-11  Eric Seidel  <eric@webkit.org>
1572
1573         Rubber-stamped by Adam Barth.
1574
1575         Move bugzilla.py into its own module in preparation for splitting one-file-per-class
1576         https://bugs.webkit.org/show_bug.cgi?id=49402
1577
1578         * Scripts/webkitpy/common/net/bugzilla/__init__.py: Added.
1579         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla.py.
1580         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla_unittest.py.
1581
1582 2010-11-10  Chris Guillory  <chris.guillory@google.com>
1583
1584         Reviewed by Chris Fleizach.
1585
1586         Use IAccessibleComparable to compare accessibility objects.
1587         https://bugs.webkit.org/show_bug.cgi?id=49118
1588
1589         * DumpRenderTree/AccessibilityUIElement.cpp:
1590         (AccessibilityUIElement::isEqual):
1591         * DumpRenderTree/AccessibilityUIElement.h:
1592         * DumpRenderTree/win/AccessibilityUIElementWin.cpp:
1593         (comparableObject):
1594         (AccessibilityUIElement::isEqual):
1595
1596 2010-11-10  Csaba Osztrogonác  <ossy@webkit.org>
1597
1598         Reviewed by David Hyatt.
1599
1600         HTML5 Ruby support should be mandatory feature
1601         https://bugs.webkit.org/show_bug.cgi?id=49272
1602
1603         Remove Ruby as optional feature.
1604
1605         * Scripts/build-webkit:
1606
1607 2010-11-10  Adam Roben  <aroben@apple.com>
1608
1609         Print VC++ Express build logs in the same order that the projects are
1610         built
1611
1612         Fixes <http://webkit.org/b/49326> print-vse-failure-logs prints logs
1613         in an unhelpful order
1614
1615         Reviewed by Eric Seidel.
1616
1617         * Scripts/print-vse-failure-logs:
1618         (PrintVisualStudioExpressLogs._build_order): Added. Uses
1619         print-msvc-project-dependencies to figure out the order in which
1620         projects are built.
1621         (PrintVisualStudioExpressLogs._sort_buildlogs): Added. Sorts the logs
1622         based on their build order and project name.
1623         (PrintVisualStudioExpressLogs._obj_directory): Moved code to find the
1624         scripts directory from here...
1625         (PrintVisualStudioExpressLogs._scripts_directory): ...to here.
1626         (PrintVisualStudioExpressLogs.main): Sort the logs before printing
1627         them.
1628
1629 2010-11-10  Adam Roben  <aroben@apple.com>
1630
1631         Only print the interesting text from VC++ Express build logs
1632
1633         Fixes <http://webkit.org/b/49325> It's hard to find the interesting
1634         output from print-vse-failure-logs
1635
1636         Reviewed by Eric Seidel.
1637
1638         * Scripts/print-vse-failure-logs: Removed unnecessary os.path import,
1639         added newly-required imports. Removed a comment that was essentially a
1640         FIXME, and is now fixed.
1641         (PrintVisualStudioExpressLogs._relevant_text): Added. Uses
1642         BeautifulSoup to extract the relevant text from the build log.
1643         (PrintVisualStudioExpressLogs.main): Only print the relevant text.
1644
1645 2010-11-09  Mihai Parparita  <mihaip@chromium.org>
1646
1647         Reviewed by Tony Chang.
1648
1649         [Chromium] http/tests/security/XFrameOptions fail with Chromium DRT
1650         https://bugs.webkit.org/show_bug.cgi?id=49286
1651         
1652         The tests in http/tests/security/XFrameOptions started to fail after
1653         r71297, since it changed the WebViewHost::assignIdentifierToRequest
1654         logic to always assign identifiers to requests, even if we're not going
1655         to be dumping resource load callbacks at that point. These tests
1656         only call LayoutTestController.dumpResourceLoadCallbacks after the
1657         request is started, thus their expectations have "<unknown>" as the
1658         identifier. Change WebViewHost to the previous behavior, which also
1659         match the other ports'.
1660
1661         * DumpRenderTree/chromium/WebViewHost.cpp:
1662         (WebViewHost::assignIdentifierToRequest):
1663         (WebViewHost::removeIdentifierForRequest):
1664
1665 2010-11-08  Tony Chang  <tony@chromium.org>
1666
1667         Reviewed by Adam Barth.
1668
1669         run platform/chromium/plugins/return-npobject.html on all platforms
1670         https://bugs.webkit.org/show_bug.cgi?id=49103
1671
1672         This tests that we can get an NPObject returned through a method on
1673         an NPAPI Object.
1674
1675         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1676         (testCallbackReturn):
1677         (pluginInvoke):
1678
1679 2010-11-09  Sheriff Bot  <webkit.review.bot@gmail.com>
1680
1681         Unreviewed, rolling out r71672.
1682         http://trac.webkit.org/changeset/71672
1683         https://bugs.webkit.org/show_bug.cgi?id=49276
1684
1685         It doesn't work on Chromium Windows bot (Requested by Ossy on
1686         #webkit).
1687
1688         * Scripts/webkitpy/common/system/file_lock.py: Removed.
1689         * Scripts/webkitpy/common/system/file_lock_unittest.py: Removed.
1690         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1691
1692 2010-11-09  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1693
1694         Reviewed by Tony Chang.
1695
1696         [NRWT] Make http locking similar to perl implementation
1697         https://bugs.webkit.org/show_bug.cgi?id=49187
1698
1699         * Scripts/webkitpy/common/system/file_lock.py: Added.
1700         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
1701         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1702
1703 2010-11-09  James Kozianski  <koz@chromium.org>
1704
1705         Reviewed by Eric Seidel.
1706
1707         Clean up imports for webkit-patch.
1708         https://bugs.webkit.org/show_bug.cgi?id=49083
1709
1710         Move some imports from commands into main.py where they are used.
1711         Move command imports into the __init__ of the commands module.
1712
1713         * Scripts/webkitpy/tool/commands/__init__.py:
1714         * Scripts/webkitpy/tool/commands/download.py:
1715         * Scripts/webkitpy/tool/commands/queues.py:
1716         * Scripts/webkitpy/tool/main.py:
1717
1718 2010-11-09  Yi Shen  <yi.4.shen@nokia.com>
1719
1720         Reviewed by Andreas Kling.
1721
1722         [Qt][QtTestBrowser] Don't erase incorrect url in the Url Bar
1723         https://bugs.webkit.org/show_bug.cgi?id=49047
1724
1725         * QtTestBrowser/mainwindow.cpp:
1726         (MainWindow::setAddressUrl):
1727
1728 2010-11-09  Andras Becsi  <abecsi@webkit.org>
1729
1730         Reviewed by Csaba Osztrogonác.
1731
1732         Improve old-run-webkit-tests --verbose to show DumpTool opening/closing
1733         https://bugs.webkit.org/show_bug.cgi?id=49252
1734
1735         * Scripts/old-run-webkit-tests:
1736
1737 2010-11-09  Csaba Osztrogonác  <ossy@webkit.org>
1738
1739         Reviewed by Tor Arne Vestbø.
1740
1741         Output of old-run-webkit-tests is buggy in verbose mode
1742         https://bugs.webkit.org/show_bug.cgi?id=49249
1743
1744         * Scripts/old-run-webkit-tests: Redundant outputs removed.
1745
1746 2010-11-08  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
1747
1748         Reviewed by Martin Robinson.
1749
1750         [GTK] Link with target name set does not work
1751         https://bugs.webkit.org/show_bug.cgi?id=48865
1752
1753         When a new page is created with a name (target=myFrame), the new
1754         mainFrame could not be found because they where not stored in the
1755         same PageGroup. As PageGroup are not exposed externally so the
1756         simpliest solution is to use a global page group name. This also fixes
1757         issue with visited link coloration across pages. After this change the
1758         private function webkit_web_view_set_group_name() was no longer used
1759         so it was removed completly.
1760
1761         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1762         (createWebView):
1763
1764 2010-11-08  John Knottenbelt  <jknotten@chromium.org>
1765
1766         Reviewed by Kent Tamura.
1767
1768         WebViewHost::reset() uses placement new.
1769         https://bugs.webkit.org/show_bug.cgi?id=49069
1770
1771         WebViewHost is using placement destruction / new to simulate a fresh
1772         WebViewHost object at the same address. This is because the WebView remains
1773         open across tests and maintains a pointer to the WebViewHost.
1774         This change resets member variables explictly instead of the placement new dance.
1775
1776         * DumpRenderTree/chromium/WebViewHost.cpp:
1777         (WebViewHost::WebViewHost):
1778         (WebViewHost::reset):
1779
1780 2010-11-08  Dirk Pranke <dpranke@chromium.org>
1781
1782         Reviewed by Eric Seidel.
1783
1784         Fix the webkit_base_dir logic in webkitpy/layout_tests/port/config.py. 
1785         It turns out that NRWT can't use the code in scm.find_checkout_root()
1786         because the Chromium bots don't do full checkouts of the WebKit
1787         tree; they only check out subdirectories like WebKitTools/Scripts.
1788         Until we can figure out a better approach for this, I've
1789         restored the base_dir-detecting code from NRWT, which works in
1790         any directory tree, scm or no.
1791
1792         This also restores the files modified in r71475 and r71474.
1793
1794         https://bugs.webkit.org/show_bug.cgi?id=49151
1795
1796         * WebKitTools/Scripts/webkitpy/layout_tests/port/config.py
1797         * WebKitTools/Scripts/webkitpy/layout_tests/port/config_unittest.py
1798
1799 2010-11-08  Eric Seidel  <eric@webkit.org>
1800
1801         Reviewed by Adam Barth.
1802
1803         webkit-patch failure-reason explodes if a build is missing
1804         https://bugs.webkit.org/show_bug.cgi?id=49195
1805
1806         This is likely a recent regression.  Adding a unit test for this case.
1807
1808         * Scripts/webkitpy/common/net/buildbot.py:
1809         * Scripts/webkitpy/common/net/buildbot_unittest.py:
1810
1811 2010-11-08  Dan Bernstein  <mitz@apple.com>
1812
1813         Reviewed by Darin Adler.
1814
1815         Ignore the system scrollbar setting.
1816
1817         * DumpRenderTree/mac/DumpRenderTree.mm:
1818         (resetDefaultsToConsistentValues): Set the scrollbar default.
1819
1820 2010-11-05  Adam Roben  <aroben@apple.com>
1821
1822         Make webkitpy.common.system.executive_unittest pass when running under
1823         Win32 Python
1824
1825         Fixes <http://webkit.org/b/49033>.
1826
1827         Reviewed by Dave Levin and Eric Seidel.
1828
1829         * Scripts/webkitpy/common/system/executive.py:
1830         (Executive._run_command_with_teed_output): Pass the arguments through
1831         encode_argument_if_needed rather than using Cygwin-specific code here.
1832         (Executive.run_and_throw_if_fail): Use child_process_encoding to decode
1833         the output.
1834         (Executive.run_command): Use encode_argument_if_needed to encode the
1835         arguments and child_process_encoding to decode the output.
1836         (Executive._child_process_encoding): Returns the encoding that should be
1837         used when communicating with child processes. On Windows we use mbcs,
1838         which maps to the current code page. On all other platforms we use
1839         UTF-8.
1840         (Executive._should_encode_child_process_arguments): Returns True if
1841         arguments to child processes need to be encoded. This is currently
1842         only needed on Cygwin and Win32 Python 2.x.
1843         (Executive._encode_argument_if_needed): Encode the argument using
1844         child_process_encoding if we need to encode arguments to child
1845         processes on this platform.
1846
1847         * Scripts/webkitpy/common/system/executive_unittest.py:
1848         (never_ending_command): Added. Returns arguments to run a command that
1849         will not quit until we kill it. On Windows we use wmic, on other
1850         platforms we use yes.
1851         (ExecutiveTest.test_run_command_with_unicode): Changed to expect the
1852         mbcs encoding to be used and for output from the child processes to
1853         have been roundtripped through encode/decode on Win32 Python. When
1854         UTF-8 is the encoding the roundtripping is undetectable, but with mbcs
1855         it's possible that some characters will not be able to be converted
1856         and will be replaced by question marks; the round-tripping allows us
1857         to expect this result.
1858
1859         (ExecutiveTest.test_kill_process):
1860         (ExecutiveTest.test_kill_all):
1861         Use never_ending_command instead of invoking "yes" directly. Expect an
1862         exit code of 1 when using Win32 Python, as that's what seems to happen.
1863
1864 2010-11-08  Adam Roben  <aroben@apple.com>
1865
1866         Roll out r71532
1867
1868         It broke the build for Cygwin 1.7 installs. Cygwin 1.7's default
1869         .bashrc unsets %TEMP%, which broke copy-tools.cmd.
1870
1871         * Scripts/webkitdirs.pm:
1872
1873 2010-11-08  Tony Chang  <tony@chromium.org>
1874
1875         Reviewed by Adam Barth.
1876
1877         run platform/chromium/plugins/nested-plugin-objects.html on all platforms
1878         https://bugs.webkit.org/show_bug.cgi?id=49094
1879
1880         This tests that objects created by plugins are proplery cleaned up.
1881
1882         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1883         (testPassTestObject):
1884         (pluginInvoke):
1885         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
1886         (testAllocate):
1887         (testDeallocate):
1888         (testGetProperty):
1889         (testConstruct):
1890
1891 2010-11-08  Adam Roben  <aroben@apple.com>
1892
1893         Mark Windows builds triggered from Perl as being non-interactive
1894
1895         This affects whether some of our scripts will show alerts vs. printing
1896         to the build log.
1897
1898         Fixes <http://webkit.org/b/49181> Windows build fail mysteriously when
1899         .vsprops files are updated
1900
1901         Reviewed by Steve Falkenburg.
1902
1903         * Scripts/webkitdirs.pm:
1904         (buildVisualStudioProject): Set WEBKIT_NONINTERACTIVE_BUILD to 1.
1905
1906 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1907
1908         Reviewed by Ojan Vafai.
1909
1910         Make http locking default in NRWT.
1911         https://bugs.webkit.org/show_bug.cgi?id=48053
1912
1913         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1914         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1915
1916 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1917
1918         Reviewed by Csaba Osztrogonác.
1919
1920         [NRWT] If the http lock fails we shouldn't do any locking
1921         https://bugs.webkit.org/show_bug.cgi?id=49164
1922
1923         If something goes wrong with the locking, the test should keep going.
1924
1925         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1926
1927 2010-11-08  Adam Roben  <aroben@apple.com>
1928
1929         Switch back to using kCGImageAlphaPremultipliedFirst when generating
1930         pixel dumps on Windows
1931
1932         I changed this behavior in r71418 thinking that it was required for
1933         getting plugins to show up in pixel dumps. But it doesn't seem to be
1934         necessary, and was making it impossible to compare new Windows pixel
1935         dumps with existing Windows or Mac pixel dumps (because ImageDiff won't
1936         compare an image with alpha to an image without alpha).
1937
1938         Fixes <http://webkit.org/b/49172> REGRESION (r71418): Can't compare
1939         new Windows pixel results to existing Windows or Mac results
1940
1941         Reviewed by Antti Koivisto.
1942
1943         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
1944         (createBitmapContextFromWebView): Replaced kCGImageAlphaNoneSkipFirst
1945         with kCGImageAlphaPremultipliedFirst.
1946
1947 2010-11-08  Csaba Osztrogonac  <ossy@webkit.org>
1948
1949         Unreviewed, rolling out r71466.
1950         http://trac.webkit.org/changeset/71466
1951         https://bugs.webkit.org/show_bug.cgi?id=48865
1952
1953         It broke layout tests on GTK bots.
1954
1955         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1956         (createWebView):
1957
1958 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1959
1960         Reviewed by Csaba Osztrogonác.
1961
1962         Enable running of Qt API tests on BuildBot
1963         https://bugs.webkit.org/show_bug.cgi?id=49004
1964
1965         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1966
1967 2010-11-08  Eric Seidel  <eric@webkit.org>
1968
1969         Reviewed by Adam Barth.
1970
1971         Add clean-review-queue command to remove closed bugs from the webkit.org/pending-review
1972         https://bugs.webkit.org/show_bug.cgi?id=49160
1973
1974         Bugzilla doesn't automatically remove r? when a bug gets closed.
1975         This script takes care of that for webkit.org.
1976
1977         * Scripts/webkitpy/common/net/bugzilla.py:
1978         * Scripts/webkitpy/tool/commands/upload.py:
1979
1980 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
1981
1982         Unreviewed, rolling out r71474.
1983         http://trac.webkit.org/changeset/71474
1984         https://bugs.webkit.org/show_bug.cgi?id=48280
1985
1986         breaks chromium webkit tests
1987         https://bugs.webkit.org/show_bug.cgi?id=49151
1988
1989         * Scripts/webkitpy/layout_tests/port/base.py:
1990         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
1991         * Scripts/webkitpy/layout_tests/port/config.py:
1992         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1993         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
1994         * Scripts/webkitpy/layout_tests/port/webkit.py:
1995         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1996         * Scripts/webkitpy/tool/mocktool.py:
1997
1998 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
1999
2000         Unreviewed, rolling out r71475.
2001         http://trac.webkit.org/changeset/71475
2002  
2003         breaks chromium webkit tests
2004         https://bugs.webkit.org/show_bug.cgi?id=49151
2005
2006         * Scripts/webkitpy/common/newstringio.py: Removed.
2007         * Scripts/webkitpy/common/newstringio_unittest.py: Removed.
2008         * Scripts/webkitpy/common/system/executive_mock.py: Removed.
2009         * Scripts/webkitpy/common/system/filesystem_mock.py: Removed.
2010         * Scripts/webkitpy/layout_tests/port/config_mock.py: Removed.
2011
2012 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2013
2014         Unreviewed, build fix.
2015
2016         Add files inexplicably not committed in r71474 as part of the
2017         fix for bug 48280.
2018
2019         * Scripts/webkitpy/common/newstringio.py: Added.
2020         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
2021         * Scripts/webkitpy/common/system/executive_mock.py: Added.
2022         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
2023         * Scripts/webkitpy/layout_tests/port/config_mock.py: Added.
2024
2025 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2026
2027         Reviewed by Eric Siedel.
2028
2029         new-run-webkit-tests: update port/base and port/webkit to use the
2030         new FileSystem and Config abstractions, pulling more logic out of
2031         the base Port classes into separate, mockable objects.
2032
2033         Also create a MockFileSystem object, a MockConfig object, move
2034         MockExecutive into common/system to be next to executive, and
2035         update the config object to use a FileSystem.
2036
2037         https://bugs.webkit.org/show_bug.cgi?id=48280
2038
2039         * Scripts/webkitpy/common/newstringio.py: Added.
2040         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
2041         * Scripts/webkitpy/common/system/executive_mock.py: Added.
2042         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
2043         * Scripts/webkitpy/layout_tests/port/base.py:
2044         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2045         * Scripts/webkitpy/layout_tests/port/config.py:
2046         * Scripts/webkitpy/layout_tests/port/config_mock.py:
2047         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
2048         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
2049         * Scripts/webkitpy/layout_tests/port/webkit.py:
2050         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2051         * Scripts/webkitpy/tool/mocktool.py:
2052
2053 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2054
2055         Unreviewed, build breakage.
2056
2057         Apparently I uploaded the wrong version of the file to fix 49122
2058         and neither Eric or I noticed - it was missing a dirname() call.
2059         Fixing ...
2060
2061         https://bugs.webkit.org/show_bug.cgi?id=49122
2062
2063         * Scripts/webkitpy/common/checkout/scm.py:
2064
2065 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
2066
2067         Reviewed by Eric Seidel.
2068
2069         webkitpy/tool/* unittests change cwd and don't clean up properly
2070
2071         https://bugs.webkit.org/show_bug.cgi?id=49122
2072
2073         * Scripts/webkitpy/common/checkout/scm.py:
2074
2075 2010-11-05  Chris Marrin  <cmarrin@apple.com>
2076
2077         Reviewed by Simon Fraser.
2078
2079         suspendAnimations/resumeAnimations not present in WebKit2
2080         https://bugs.webkit.org/show_bug.cgi?id=49109
2081
2082         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
2083         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
2084         (WTR::LayoutTestController::suspendAnimations):
2085         (WTR::LayoutTestController::resumeAnimations):
2086         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
2087
2088 2010-11-05  Tony Chang  <tony@chromium.org>
2089
2090         Reviewed by David Levin.
2091
2092         cleanup style in TestNetscapePlugIn/PluginObject.cpp
2093         https://bugs.webkit.org/show_bug.cgi?id=49044
2094
2095         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
2096         (getPluginClass):
2097         (pluginGetProperty):
2098         (pluginSetProperty):
2099         (variantToIdentifier):
2100         (testCallback):
2101         (getURL):
2102         (getURLNotify):
2103         (testInvokeDefault):
2104         (testGetProperty):
2105         (testDocumentOpen):
2106         (testWindowOpen):
2107         (pluginInvoke):
2108         (handleCallback):
2109         (notifyStream):
2110         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
2111
2112 2010-11-05  Eric Seidel  <eric@webkit.org>
2113
2114         Reviewed by David Levin.
2115
2116         Add QueueStatusServer/__init__.py so others can run the QueueStatusServer tests
2117         https://bugs.webkit.org/show_bug.cgi?id=49032
2118
2119         I wrote this file as part of bug 47847, but I forgot to commit it.
2120         No one else noticed it missing because test-webkitpy knows how
2121         to recover in the case where it can't import QueueStatusServer
2122         (which generally occurs due to not having installed the AppEngine SDK).
2123
2124         * QueueStatusServer/__init__.py: Added.
2125         * QueueStatusServer/model/workitems_unittest.py:
2126          - Remove a test which fails.  This was probably landed (by me)
2127            from my other machine, which since this __init__.py was missing
2128            I never noticed the failure and landed this invalid test.
2129            Sadly we can't really test remove_work_item as it depends
2130            on .key() working.  .key() will throw unless the object
2131            has already been saved it seems.
2132            This may be a mis-design in our remove_work_item implementation,
2133            but for now, just removing the test.
2134
2135 2010-11-04  Adam Roben  <aroben@apple.com>
2136
2137         Teach check-webkit-style about TestNetscapePlugIn
2138
2139         Fixes <http://webkit.org/b/49030> check-webkit-style is ignorant of
2140         TestNetscapePlugIn's coding conventions
2141
2142         Reviewed by John Sullivan.
2143
2144         * Scripts/webkitpy/style/checker.py: Lump TestNetscapePlugIn in with
2145         WebKitAPITest and TestWebKitAPI in having different include order and
2146         naming conventions than WebCore. Added some comments about why the
2147         exceptions exist.
2148
2149 2010-11-04  Adam Roben  <aroben@apple.com>
2150
2151         Add a test that shows that windowed plugins are able to paint
2152
2153         Somehow we never had a test for this before.
2154
2155         Fixes <http://webkit.org/b/49029> Should add a test that shows
2156         windowed plugins can paint
2157
2158         Reviewed by Jon Honeycutt.
2159
2160         * DumpRenderTree/TestNetscapePlugIn/Tests/win/DrawsGradient.cpp: Added.
2161         (DrawsGradient::DrawsGradient):
2162         (DrawsGradient::wndProc): We handle the WM_PAINT and WM_PRINTCLIENT messages.
2163
2164         (DrawsGradient::onPaint):
2165         (DrawsGradient::onPrintClient):
2166         These both just call through to paint.
2167
2168         (DrawsGradient::paint): Fills our client area with some gradients.
2169
2170         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Link
2171         against Msimg32.lib for ::GradientFill and added DrawsGradient.
2172
2173         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
2174         (createBitmapContextFromWebView): Use WM_PRINT instead of
2175         WM_PRINTCLIENT so that ::DefWindowProc will send
2176         WM_PRINT/WM_PRINTCLIENT messages to the WebView's child windows.
2177         Replaced kCGImageAlphaPremultipledFirst with kCGImageAlphaNoneSkipFirst
2178         because GDI doesn't support alpha and kCGImageBitmapOrder32Little with
2179         kCGImageBitmapOrder32Host because, who knows, maybe someday Windows
2180         will run on a big-endian platform.
2181
2182 2010-11-04  Adam Roben  <aroben@apple.com>
2183
2184         Extract much of NPNInvalidateRectInvalidatesWindow's code into a
2185         WindowedPluginTest base class
2186
2187         The base class takes care of subclassing the plugin's window so that a
2188         custom WNDPROC is called. This will make it easier to write tests that
2189         need to handle window messages.
2190
2191         Fixes <http://webkit.org/b/49028> It's hard to write a PluginTest with
2192         a custom WNDPROC
2193
2194         Reviewed by Jon Honeycutt.
2195
2196         * DumpRenderTree/TestNetscapePlugIn/Tests/win/NPNInvalidateRectInvalidatesWindow.cpp:
2197         (NPNInvalidateRectInvalidatesWindow::NPNInvalidateRectInvalidatesWindow):
2198         (NPNInvalidateRectInvalidatesWindow::NPP_SetWindow):
2199         (NPNInvalidateRectInvalidatesWindow::wndProc):
2200         (NPNInvalidateRectInvalidatesWindow::testInvalidateRect):
2201         Moved code from here to WindowedPluginTest. Changed to use window()
2202         instead of m_window.
2203
2204         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2205         Added WindowedPluginTest and added TestNetscapePlugIn/win to the
2206         include path. Also reordered the include path to make a little more
2207         sense and simplified the entry that added TestNetscapePlugIn itself to
2208         the include path.
2209
2210         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.cpp: Added.
2211         (WindowedPluginTest::WindowedPluginTest):
2212         (WindowedPluginTest::NPP_SetWindow):
2213         (WindowedPluginTest::staticWndProc):
2214         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.h: Added.
2215         (WindowedPluginTest::window):
2216         Code came from NPNInvalidateRectInvalidatesWindow.
2217
2218 2010-11-04  Adam Roben  <aroben@apple.com>
2219
2220         Add a plugin test to show that windowed plugins are clipped correctly
2221
2222         Fixes <http://webkit.org/b/49024> <rdar://problem/8487847> Windowed
2223         plugins aren't clipped in WebKit2 on Windows
2224
2225         Reviewed by Jon Honeycutt.
2226
2227         * DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowRegionIsSetToClipRect.cpp: Added.
2228         (WindowRegionIsSetToClipRect::WindowRegionIsSetToClipRect): Initialize members.
2229         (WindowRegionIsSetToClipRect::NPP_SetWindow): Check that our window
2230         region matches the clip rect we know we should have based on
2231         window-region-is-set-to-clip-rect.html, and check that our window class
2232         doesn't have the CS_PARENTDC style.
2233
2234         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2235         Added WindowRegionIsSetToClipRect.
2236
2237 2010-11-05  Alejandro G. Castro  <alex@igalia.com>
2238
2239         Reviewed by Martin Robinson.
2240
2241         [GTK] Avoid font hinting in the DRT
2242         https://bugs.webkit.org/show_bug.cgi?id=48548
2243
2244         Change the settings to avoid font hinting, it was causing
2245         different results depending on the distribution.
2246
2247         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2248         (initializeGtkFontSettings):
2249         * DumpRenderTree/gtk/fonts/fonts.conf:
2250
2251 2010-11-05  Eric Seidel  <eric@webkit.org>
2252
2253         Reviewed by Adam Barth.
2254
2255         build-webkit spams system.log by repeatedly checking for PBXProductDirectory in com.apple.Xcode
2256         https://bugs.webkit.org/show_bug.cgi?id=49051
2257
2258         This is a speculative fix.  The unit tests cover these methods, however
2259         I don't know if this will fully stop the system.log spam.
2260
2261         * Scripts/webkitpy/layout_tests/port/base.py:
2262         * Scripts/webkitpy/layout_tests/port/webkit.py:
2263
2264 2010-11-05  Eric Seidel  <eric@webkit.org>
2265
2266         Reviewed by Adam Barth.
2267
2268         Add basic support for showing bot id on /queue-status/ pages
2269         https://bugs.webkit.org/show_bug.cgi?id=49037
2270
2271         This support is really simple.  Eventually we'll want to
2272         show the bot id in the lock table too, but we don't have
2273         that information stored in the server yet.
2274
2275         * QueueStatusServer/handlers/queuestatus.py:
2276         * QueueStatusServer/handlers/queuestatus_unittest.py: Added.
2277         * QueueStatusServer/templates/includes/singlequeuestatus.html:
2278
2279 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
2280
2281         Reviewed by Adam Barth.
2282
2283         Create a filesystem wrapper that we can use to enforce
2284         particular conventions and use for mocking and dependency
2285         injection down the line.
2286
2287         https://bugs.webkit.org/show_bug.cgi?id=48144
2288
2289         * Scripts/webkitpy/common/system/filesystem.py: Added.
2290         * Scripts/webkitpy/common/system/filesystem_unittest.py: Added.
2291
2292 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2293
2294         Reviewed by Tony Chang.
2295
2296         run_webkit_tests_unittest fails under Python 2.5
2297         https://bugs.webkit.org/show_bug.cgi?id=49043
2298         
2299         Switch from itertools.chain.from_iterable (which was added in 2.6)
2300         to using itertools.chain directly.
2301
2302         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2303
2304 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2305
2306         Unreviewed fix to webkit-patch rebaseline-server so that it runs under
2307         Python 2.5 (it needs an import to use the with statement).
2308
2309         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
2310
2311 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
2312
2313         Reviewed by Adam Barth.
2314
2315         new-run-webkit-tests: split out webkit-specific configuration stuff into a new module
2316
2317         The current NRWT code has webkit-specific configuration code (like
2318         _script_path, default configuration, etc.) mixed in with
2319         layout-test-specific stuff in port/base. The configuration code
2320         should be split out into a separate module for easier mocking,
2321         testing, and isolation.
2322
2323         https://bugs.webkit.org/show_bug.cgi?id=48264
2324
2325         * Scripts/webkitpy/layout_tests/port/config.py: Added.
2326         * Scripts/webkitpy/layout_tests/port/config_unittest.py: Added.
2327
2328 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2329
2330         Reviewed by Tony Chang.
2331
2332         Rebaseline server: initial framework
2333         https://bugs.webkit.org/show_bug.cgi?id=48892
2334         
2335         Adds the basic framework for the rebaseline server (details at 
2336         http://webkit.org/b/47761). Includes the rebaseline-server webkit-patch
2337         command, which starts an HTTP server that can serve static files or
2338         invoke handler methods on a class.
2339
2340         * Scripts/webkitpy/tool/commands/__init__.py:
2341         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html: Added.
2342         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css: Added.
2343         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js: Added.
2344         * Scripts/webkitpy/tool/commands/rebaselineserver.py: Added.
2345
2346 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
2347
2348         Reviewed by James Robinson.
2349
2350         new-run-webkit-tests wasn't using DRT by default for
2351         --platform chromium-gpu
2352
2353         The default value was set to False instead of None, which meant
2354         that the platform specific logic wasn't firing to change the
2355         value to True (b/c we were afraid we'd be overriding the user
2356         preference).
2357
2358         https://bugs.webkit.org/show_bug.cgi?id=49038
2359
2360         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2361
2362 2010-11-04  Tony Chang  <tony@chromium.org>
2363
2364         Reviewed by Adam Barth.
2365
2366         make platform/chromium/plugins/multiple-plugins.html pass on all platforms
2367         https://bugs.webkit.org/show_bug.cgi?id=49026
2368
2369         Have the testObject.bar property return the string "bar".  This lets
2370         us run and pass platform/chromium/plugins/multiple-plugins.html on all
2371         platforms.
2372
2373         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
2374         (testGetProperty):
2375
2376 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
2377
2378         Reviewed by Tony Chang.
2379
2380         NRWT doesn't have good test coverage for --run-chunk, --batch-size, --run-part, etc.
2381         https://bugs.webkit.org/show_bug.cgi?id=48878
2382         
2383         Add get_tests_run so that it's easy to see which tests get run (and with
2384         what batching) for a given flag combination. Flesh out the various
2385         test cases that have FIXMEs.
2386         
2387         Also fixes an off-by-one error (batch sizes were one larger than 
2388         expected) and makes --run-part also have wraparound behavior, like
2389         --run-chunk.
2390
2391         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2392         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2393         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2394
2395 2010-11-04  Erik Arvidsson  <arv@chromium.org>
2396
2397         Reviewed by Dimitri Glazkov.
2398
2399         Support box-sizing without the vendor prefix
2400         https://bugs.webkit.org/show_bug.cgi?id=36713
2401
2402         Based on patch by Peter Beverloo <peter@lvp-media.com>
2403
2404         * iExploder/htdocs/cssproperties.in: Change -webkit-box-sizing to box-sizing.
2405
2406 2010-11-04  Csaba Osztrogonác  <ossy@webkit.org>
2407
2408         Unreviewed rollout r71340, because it broke Chromium Windows bot.
2409
2410         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2411         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2412
2413 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2414
2415         Reviewed by Ojan Vafai.
2416
2417         Make http locking default in NRWT.
2418         https://bugs.webkit.org/show_bug.cgi?id=48053
2419
2420         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2421         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2422
2423 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2424
2425         Reviewed by Eric Seidel.
2426
2427         [NRWT] Clear invalid http locks on Windows platform as well
2428         https://bugs.webkit.org/show_bug.cgi?id=48515
2429
2430         * Scripts/webkitpy/common/system/executive.py:
2431         * Scripts/webkitpy/common/system/executive_unittest.py:
2432         * Scripts/webkitpy/layout_tests/port/http_lock.py:
2433
2434 2010-11-02  Adam Roben  <aroben@apple.com>
2435
2436         Reduce our dependence on coreutils when running Python tests
2437
2438         This patch introduces versions of the standard echo and cat utilities
2439         implemented in Python. They are probably missing features of their
2440         coreutils equivalents, but they can do what's necessary for our Python
2441         tests. This is useful on Windows, where these utilities typically
2442         aren't available.
2443
2444         Fixes <http://webkit.org/b/48883> executive_unittest relies on echo
2445         and cat utilities from coreutils, which are not present on Windows
2446
2447         Reviewed by Eric Seidel.
2448
2449         * Scripts/webkitpy/common/system/executive_unittest.py: Changed to use
2450         our Python-based echo and cat.
2451
2452         * Scripts/webkitpy/common/system/fileutils.py: Added.
2453         (make_stdout_binary): On Windows, puts sys.stdout into binary mode so
2454         that \n won't be translated into \r\n. I couldn't think of a good way
2455         to test this directly without touching the filesystem, but it is tested
2456         indirectly by echo_unittest.
2457
2458         * Scripts/webkitpy/test/cat.py: Added.
2459         (command_arguments): Returns a list for invoking cat with the given arguments.
2460         (main): Acts like a simplified version of the coreutils cat utility.
2461
2462         * Scripts/webkitpy/test/cat_unittest.py: Added.
2463         (CatTest.assert_cat): Runs cat with the given input and ensures the
2464         output matches the input.
2465         (CatTest.test_basic): Performs a simple test of cat.
2466         (CatTest.test_no_newline): Tests what happens when the input string
2467         doesn't have a trailing newline.
2468         (CatTest.test_unicode): Tests passing a unicode string to cat.
2469         (CatTest.test_as_command): Tests running cat as a separate command.
2470
2471         * Scripts/webkitpy/test/echo.py: Added.
2472         (command_arguments): Returns a list for invoking echo with the given arguments.
2473         (main): Acts like a simplified version of the coreutils echo utility.
2474
2475         * Scripts/webkitpy/test/echo_unittest.py: Added.
2476         (EchoTest.test_basic): Performs a simple test of echo.
2477         (EchoTest.test_no_newline): Tests passing -n to echo to suppress the
2478         trailing newline.
2479         (EchoTest.test_unicode): Tests passing unicode and non-unicode strings
2480         to echo.
2481         (EchoTest.test_argument_order): Tests what happens when -n is not the
2482         first argument.
2483         (EchoTest.test_empty_arguments): Tests what happens when you pass [] to
2484         echo.main.
2485         (EchoTest.test_no_arguments): Tests what happens when you call
2486         echo.main with no arguments.
2487         (EchoTest.test_as_command): Tests running echo as a separate command.
2488
2489 2010-11-04  Renata Hodovan  <reni@inf.u-szeged.hu>
2490
2491         Unreviewed: Add myself to the list of Committers.
2492
2493         * Scripts/webkitpy/common/config/committers.py:
2494
2495 2010-11-04  Andreas Kling  <kling@webkit.org>
2496
2497         Reviewed by Kenneth Rohde Christiansen.
2498
2499         [Qt] Hook into QWebPage::scrollRequested for resizes-to-contents mode
2500
2501         In RTC mode the QGraphicsWebView item is the size of the contents,
2502         scrolling works a bit differently  (we need to react to scrollRequested.)
2503
2504         Normally QGraphicsView will replay the last mouse event when scrolling,
2505         so to prevent WebKit from getting confused by this we temporarily make
2506         the QGraphicsView non-interactive.
2507
2508         * QtTestBrowser/webview.cpp:
2509         (WebViewGraphicsBased::setPage):
2510         (WebViewGraphicsBased::scrollRequested):
2511         * QtTestBrowser/webview.h:
2512
2513 2010-11-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2514
2515         Reviewed by Andreas Kling.
2516
2517         Use OS(MAC_OS_X) rather than PLATFORM(MAC) in TestControllerQt
2518
2519         PLATFORM(MAC) is not defined for the Qt port, as it refers to
2520         the Mac-port, not the Mac OS X operating system.
2521
2522         * WebKitTestRunner/qt/TestControllerQt.cpp:
2523         (WTR::isExistingLibrary):
2524
2525 2010-11-04  Leonid Ebril  <leonid.ebril@nokia.com>
2526
2527         Reviewed by Andreas Kling.
2528
2529         [Qt] Adding iphone user agent string the user agent list for QtTestBrowser
2530         https://bugs.webkit.org/show_bug.cgi?id=48869
2531
2532         * QtTestBrowser/useragentlist.txt:
2533
2534 2010-11-03  Adam Roben  <aroben@apple.com>
2535
2536         Always use uppercase drive names in strings returned by abspath_to_uri
2537
2538         Some versions of cygpath use lowercase drive letters while others use
2539         uppercase, which makes it hard to test the output of code that uses
2540         cygpath.
2541
2542         Fixes <http://webkit.org/b/48914> webkitpy.common.system.path_unittest
2543         fails with Cygwin 1.5
2544
2545         Reviewed by Eric Seidel.
2546
2547         * Scripts/webkitpy/common/system/path.py:
2548         (cygpath): Updated the docstring to indicate that only absolute paths
2549         should be passed for now (though relative paths will work fine).
2550         (_Cygpath.convert): Upper-case the first letter of the converted Windows path.
2551
2552 2010-11-03  George Guo  <George.Guo@Nokia.com>
2553
2554         Reviewed by Andreas Kling.
2555
2556         [Qt] QtTestBrowser : set mmp rule pageddata in Symbian
2557         https://bugs.webkit.org/show_bug.cgi?id=48767
2558     
2559         Paging is needd on Symbian devices to support benchmarks tests like
2560         dromaeo.com and Celtic Kane that need a lot of memory to run  
2561
2562         * QtTestBrowser/QtTestBrowser.pro:
2563
2564 2010-11-03  Jenn Braithwaite  <jennb@chromium.org>
2565
2566         Reviewed by Dmitry Titov.
2567
2568         Chromium: Update resource tracking when moving a frame between documents
2569         https://bugs.webkit.org/show_bug.cgi?id=48363
2570
2571         * DumpRenderTree/chromium/WebViewHost.cpp:
2572         (WebViewHost::assignIdentifierToRequest):
2573         Always put resource id in map so we can make assumptions about its
2574         presence.
2575         (WebViewHost::removeIdentifierForRequest):
2576         Added.
2577         (WebViewHost::didFinishResourceLoad):
2578         (WebViewHost::didFailResourceLoad):
2579         Check existence of resource id before removing from map.
2580         * DumpRenderTree/chromium/WebViewHost.h:
2581
2582 2010-11-03  Victor Wang  <victorw@chromium.org>
2583
2584         Reviewed by Adam Barth.
2585
2586         [Chromium] update buildbot names in chromium rebaseline tool.
2587
2588         https://bugs.webkit.org/show_bug.cgi?id=48881
2589
2590         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2591
2592 2010-11-03  Darin Adler  <darin@apple.com>
2593
2594         Updated Xcode projects by opening them with Xcode 3.2.4.
2595         Updated svn:ignore for Xcode projects.
2596
2597         * MiniBrowser/MiniBrowser.xcodeproj: Added property svn:ignore.
2598         * TestWebKitAPI/TestWebKitAPI.xcodeproj: Added property svn:ignore.
2599         * WebKitLauncher/WebKitLauncher.xcodeproj: Modified property svn:ignore.
2600         * WebKitLauncher/WebKitLauncher.xcodeproj/project.pbxproj: Updated with Xcode 3.2.4.
2601
2602 2010-11-03  Adam Roben  <aroben@apple.com>
2603
2604         Stop waiting for messages from the web process after a timeout elapses
2605
2606         TestController::runUntil now takes a timeout parameter that specifies
2607         how long to wait for the condition to become true. WebKitTestRunner
2608         takes a --timeout flag that specifies how long the timeout should be.
2609         run-webkit-tests passes this flag to WebKitTestRunner so its timeout
2610         will be similar to run-webkit-tests.
2611
2612         Fixes <http://webkit.org/b/43047> <rdar://problem/8365833>
2613         WebKitTestRunner waits forever if the web process crashes
2614
2615         Reviewed by Darin Adler and Anders Carlsson.
2616
2617         * Scripts/old-run-webkit-tests:
2618         (top level): Moved the GuardMalloc timeout adjustment here from
2619         readFromDumpToolWithTimer.
2620         (openDumpTool): Make WTR use a timeout similar to but slightly shorter
2621         than the one that was specified on the command line.
2622
2623         * WebKitTestRunner/TestController.cpp:
2624         (WTR::TestController::TestController): Initialize our timeout values.
2625         (WTR::TestController::initialize): Parse the --timeout flag and use it
2626         to modify our timeout values.
2627         (WTR::TestController::resetStateToConsistentValues): Changed to use a
2628         short timeout while waiting for the web process to reset and to return
2629         a boolean indicating whether we were able to reset the web process.
2630         Uses a 5-second timeout while waiting for the process to be reset.
2631         (WTR::TestController::runTest): Changed to return a boolean indicating
2632         whether we were able to reset the web process (and thus run the test).
2633         (WTR::TestController::runUntil): Call through to platformRunUntil.
2634
2635         (WTR::TestController::runTestingServerLoop):
2636         (WTR::TestController::run):
2637         Changed to bail if any test can't be run. This will cause the process
2638         to exit. (Unfortunately this will make run-webkit-tests think we
2639         crashed; see <http://webkit.org/b/48943>.)
2640
2641         * WebKitTestRunner/TestController.h: Added platformRunUntil,
2642         m_longTimeout, and m_shortTimeout.
2643
2644         * WebKitTestRunner/TestInvocation.cpp:
2645         (WTR::TestInvocation::invoke): Use a short timeout when waiting for the
2646         initial response and a long timeout when waiting for the test to
2647         complete. Check whether runUntil timed out and print an error message
2648         if so.
2649
2650         * WebKitTestRunner/mac/TestControllerMac.mm:
2651         (WTR::TestController::platformRunUntil): Renamed from runUntil. Pass
2652         [NSDate distantPast] to -[NSRunLoop runMode:beforeDate:] so that we
2653         won't block waiting for the run loop. Only loop until the timeout
2654         elapses.
2655
2656         * WebKitTestRunner/qt/TestControllerQt.cpp:
2657         (WTR::TestController::platformRunUntil): Renamed from runUntil. Added a
2658         FIXME about honoring the timeout.
2659
2660         * WebKitTestRunner/win/TestControllerWin.cpp:
2661         (WTR::TestController::platformRunUntil): Renamed from runUntil. Use
2662         ::MsgWaitForMultipleObjectsEx to implement the timeout. Changed to use
2663         ::PeekMessageW so that we don't block waiting for messages to become
2664         available.
2665
2666 2010-11-03  Adam Roben  <aroben@apple.com>
2667
2668         Add a plugin test that evaluates JS after removing the plugin element
2669         from the document
2670
2671         This test replaces platform/win/plugins/plugin-delayed-destroy.html.
2672         That test was made to prevent a crash very similar to this one, but
2673         unfortunately tested only the mechanism that prevented the crash and
2674         not whether the crash itself was prevented. Since WebKit2 uses a
2675         different mechanism to prevent the crash, the test was failing even
2676         though WebKit2 was not vulnerable to the crash. This new test crashes
2677         if there is no mechanism in place to prevent it and passes in both
2678         WebKit1 and WebKit2.
2679
2680         Fixes <http://webkit.org/b/46711> <rdar://problem/8485903>
2681         platform/win/plugins/plugin-delayed-destroy.html fails in WebKit2
2682
2683         Reviewed by Anders Carlsson.
2684
2685         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
2686         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2687         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
2688         * GNUmakefile.am:
2689         Added new file.
2690
2691         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
2692         (pluginDeallocate): Make sure we delete the PluginTest object. This
2693         prevents a leak and also allows us to test the crash.
2694
2695         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
2696         (PluginTest::executeScript): Made this into a non-static member
2697         function.
2698
2699         (PluginTest::waitUntilDone):
2700         (PluginTest::notifyDone):
2701         Updated for changes to executeScript.
2702
2703         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Added executeScript.
2704
2705         * DumpRenderTree/TestNetscapePlugIn/Tests/EvaluateJSAfterRemovingPluginElement.cpp: Added.
2706         (EvaluateJSAfterRemovingPluginElement::EvaluateJSAfterRemovingPluginElement):
2707         Initialize ourselves and tell the test harness to wait.
2708         (EvaluateJSAfterRemovingPluginElement::NPP_DestroyStream): Remove our
2709         plugin element from the document, then execute some JavaScript. If
2710         WebKit does not have appropriate mechanisms in place, we'll be
2711         destroyed inside the first call to executeScript and crash on the
2712         second call.
2713
2714 2010-11-02  Stephen White  <senorblanco@chromium.org>
2715
2716         Reviewed by Tony Chang.
2717
2718         [chromium] Fix LayoutTestController UMRs.
2719         https://bugs.webkit.org/show_bug.cgi?id=48872
2720
2721         * DumpRenderTree/chromium/LayoutTestController.cpp:
2722         (LayoutTestController::LayoutTestController):
2723
2724 2010-11-03  Kent Tamura  <tkent@chromium.org>
2725
2726         Reviewed by Dimitri Glazkov.
2727
2728         REGRESSION: rebaseline-chromium-webkit-tests uses non-zero tolerance for
2729         image dup detection
2730         https://bugs.webkit.org/show_bug.cgi?id=48744
2731
2732         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2733          - Make a function for option parsing for ease of test
2734          - Set 0 to options.tolerance
2735         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
2736          - Add a test for this change
2737
2738 2010-11-02  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
2739
2740         Unreviewed: Add myself to the list of Committers.
2741
2742         * Scripts/webkitpy/common/config/committers.py:
2743
2744 2010-11-02  Anders Carlsson  <andersca@apple.com>
2745
2746         Fix build.
2747
2748         * TestWebKitAPI/PlatformUtilities.cpp:
2749         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
2750
2751 2010-11-02  Sam Weinig  <sam@webkit.org>
2752
2753         Reviewed by Anders Carlsson.
2754
2755         Add a way to send startup messages on the context which can be posted when a process launches
2756         <rdar://problem/8617928>
2757         https://bugs.webkit.org/show_bug.cgi?id=48838
2758
2759         * MiniBrowser/mac/WebBundle/WebBundleMain.m:
2760         (WKBundleInitialize):
2761         * TestWebKitAPI/InjectedBundleController.cpp:
2762         (TestWebKitAPI::InjectedBundleController::initialize):
2763         (TestWebKitAPI::InjectedBundleController::didReceiveMessage):
2764         * TestWebKitAPI/InjectedBundleController.h:
2765         * TestWebKitAPI/InjectedBundleMain.cpp:
2766         (WKBundleInitialize):
2767         * TestWebKitAPI/PlatformUtilities.cpp:
2768         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
2769         * WebKitTestRunner/InjectedBundle/InjectedBundleMain.cpp:
2770         (WKBundleInitialize):
2771         Update implementations of WKBundleInitialize to take an initial userData
2772         argument.  Change TestWebKitAPI to use the new initial userData to initialize
2773         each test's bundle.
2774
2775 2010-11-02  Benjamin Kalman  <kalman@google.com>
2776
2777         Reviewed by Ojan Vafai.
2778
2779         new-run-webkit-tests doesn't strip "LayoutTests/" from prefix, unlike old-run-webkit-tests
2780         https://bugs.webkit.org/show_bug.cgi?id=48794
2781
2782         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: Strip the "LayoutTests/" prefix from test argument paths.
2783
2784 2010-11-02  Adam Roben  <aroben@apple.com>
2785
2786         Skip webkitpy.layout_tests.run_webkit_tests_unittest.MainTest on Cygwin
2787         Python 2.5.x
2788
2789         It is known to hang on that version of Python. See
2790         <http://webkit.org/b/48614>.
2791
2792         Reviewed by Adam Barth.
2793
2794         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py: Use
2795         skip_if to skip MainTest on Cygwin Python 2.5.x.
2796
2797         * Scripts/webkitpy/test/skip.py: Added.
2798         (skip_if): If the passed-in condition is false, find all the test_*
2799         methods of the given class and replace them with a function that just
2800         logs that we're skipping these tests. This is loosely based on the
2801         unittest.skip_if decorator added in Python 3.1.
2802         (_skipped_method): Returns a function that just logs that the tests are
2803         being skipped. This is used in place of the actual test_* functions.
2804
2805         * Scripts/webkitpy/test/skip_unittest.py: Added.
2806         (SkipTest.setUp): Create a logger for ourselves and give it a handler
2807         that logs to a private stream.
2808         (SkipTest.tearDown): Reset the logger.
2809         (SkipTest.create_fixture_class): Returns a class that we use to test
2810         skip_if. It has a single test_* method, test_foo, that just calls
2811         through to a callback.
2812         (SkipTest.foo_callback): Record that test_foo was called.
2813         (SkipTest.test_skip_if_false): Pass skip_if a False condition and test
2814         that test_foo does get called.
2815         (SkipTest.test_skip_if_true): Pass skip_if a True condition and test
2816         that test_foo does not get called and the appropriate message gets
2817         logged.
2818
2819 2010-11-02  Adam Barth  <abarth@webkit.org>
2820
2821         Reviewed by Eric Seidel.
2822
2823         webkit-patch should tell check-webkit-style which files were changed so
2824         check-webkit-style doesn't have to stat the whole working copy again
2825         https://bugs.webkit.org/show_bug.cgi?id=48792
2826
2827         * Scripts/webkitpy/tool/mocktool.py:
2828         * Scripts/webkitpy/tool/steps/checkstyle.py:
2829
2830 2010-11-02  Robert Kroeger  <rjkroege@chromium.org>
2831
2832         Reviewed by James Robinson.
2833
2834         [Chromium/DRT] Make EventSendingController honour leapForward for touch events.
2835         https://bugs.webkit.org/show_bug.cgi?id=48777
2836
2837         * DumpRenderTree/chromium/EventSender.cpp:
2838         (EventSender::sendCurrentTouchEvent):
2839
2840 2010-11-02  Adam Roben  <aroben@apple.com>
2841
2842         Only track resource identifiers in DRT when dumpResourceLoadCallbacks
2843         is on
2844
2845         This reverts Windows to our pre-r71097 behavior. That patch made us
2846         track all resource identifiers, including the main resource, so the
2847         main resource's URL started appearing in test output instead of
2848         "<unknown>". Arguably having the main resource's URL is better, but all
2849         other platforms print "<unknown>" and we want to match.
2850
2851         Fixes <http://webkit.org/b/48837> <rdar://problem/8620351> REGRESSION
2852         (r71097): Many http tests failing on Windows
2853
2854         Reviewed by Anders Carlsson.
2855
2856         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
2857         (ResourceLoadDelegate::identifierForInitialRequest): Don't add the
2858         identifier to the URL map if we're not supposed to dump resource load
2859         callbacks.
2860         (ResourceLoadDelegate::removeIdentifierForRequest): Always remove the
2861         identifier from the URL map even if we're already "done". There's no
2862         point in keeping out-of-date identifiers around.
2863
2864 2010-11-01  Jenn Braithwaite  <jennb@chromium.org>
2865
2866         Reviewed by Adam Roben.
2867
2868         Windows: Update resource tracking when moving a frame between documents
2869         https://bugs.webkit.org/show_bug.cgi?id=48364
2870
2871         * DumpRenderTree/win/DumpRenderTree.cpp:
2872         (createWebViewAndOffscreenWindow):
2873         (main):
2874         Give each WebView its own ResourceLoadDelegate instance in order to
2875         make assertions about resource ids on a particular WebView.
2876         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
2877         (ResourceLoadDelegate::identifierForInitialRequest):
2878         Always add id to the map.
2879         (ResourceLoadDelegate::removeIdentifierForRequest):
2880         Added.
2881         (ResourceLoadDelegate::willSendRequest):
2882         (ResourceLoadDelegate::didReceiveAuthenticationChallenge):
2883         (ResourceLoadDelegate::didReceiveResponse):
2884         (ResourceLoadDelegate::didFinishLoadingFromDataSource):
2885         (ResourceLoadDelegate::didFailLoadingWithError):
2886         (ResourceLoadDelegate::descriptionSuitableForTestResult):
2887         Replace static descriptionSuitableForTestResult with static member function to access identifier map.
2888         * DumpRenderTree/win/ResourceLoadDelegate.h:
2889         (ResourceLoadDelegate::urlMap):
2890         Moved within class so that each WebView has its own id map.
2891
2892 2010-11-01  Ojan Vafai  <ojan@chromium.org>
2893
2894         Reviewed by Dimitri Glazkov.
2895
2896         remove debug code from run_webkit_tests.py
2897         https://bugs.webkit.org/show_bug.cgi?id=48800
2898
2899         Remove temporary debug code and make --master-name required
2900         if --test-results-server is set now that all clients set
2901         --master-name.
2902
2903         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2904
2905 2010-11-01  Tony Chang  <tony@chromium.org>
2906
2907         Reviewed by Kent Tamura.
2908
2909         add plugin event logging for linux
2910         https://bugs.webkit.org/show_bug.cgi?id=48779
2911
2912         This is taken from Chromium's fork of the layout test plugin:
2913         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
2914
2915         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
2916         (handleEventX11):
2917         (NPP_HandleEvent):
2918
2919 2010-11-01  Dirk Pranke  <dpranke@chromium.org>
2920
2921         Reviewed by James Robinson.
2922
2923         new-run-webkit-tests: use DRT, child-processes=1 for GPU tests by default
2924         https://bugs.webkit.org/show_bug.cgi?id=48790
2925
2926         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2927         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2928
2929 2010-11-01  Adam Barth  <abarth@webkit.org>
2930
2931         Reviewed by Eric Seidel.
2932
2933         Teach check-webkit-style how to accept a list of files to diff on the
2934         command line
2935         https://bugs.webkit.org/show_bug.cgi?id=48784
2936
2937         In a future patch, webkit-patch will use this option to improve
2938         performance.  I'm landing this in two pieces to avoid causing a version
2939         skew problem for the style-bot.
2940
2941         * Scripts/check-webkit-style:
2942         * Scripts/webkitpy/style/optparser.py:
2943         * Scripts/webkitpy/style/optparser_unittest.py:
2944         * Scripts/webkitpy/style_references.py:
2945
2946 2010-11-01  Anders Carlsson  <andersca@apple.com>
2947
2948         Reviewed by John Sullivan.
2949
2950         Tear down the related WebProcessProxy when a WebContext is deallocated
2951         https://bugs.webkit.org/show_bug.cgi?id=48769
2952
2953         * TestWebKitAPI/Tests/WebKit2/FailedLoad.cpp:
2954         (TestWebKitAPI::didFailProvisionalLoadWithErrorForFrame):
2955         We don't support empty URLs anymore, update test to expect a null URL instead.
2956
2957 2010-11-01  Søren Gjesse  <sgjesse@chromium.org>
2958
2959         Reviewed by Andreas Kling.
2960
2961         Fix warning when compiling the chromium port of DumpRenderShell
2962         with clang.
2963         https://bugs.webkit.org/show_bug.cgi?id=48414
2964
2965         * DumpRenderTree/chromium/TestShell.h:
2966         (TestShell::javaScriptFlagsForLoad):
2967
2968 2010-11-01  Adam Roben  <aroben@apple.com>
2969
2970         Fix typo from r71022
2971
2972         * BuildSlaveSupport/build.webkit.org-config/config.json:
2973
2974 2010-11-01  Adam Roben  <aroben@apple.com>
2975
2976         Trigger the Windows Release WebKit2 tests when a Release build
2977         finishes, not when a Debug build finishes
2978
2979         Fixes <http://webkit.org/b/48754> Windows Release WebKit2 tests are
2980         triggered at the wrong time
2981
2982         Reviewed by Sam Weinig.
2983
2984         * BuildSlaveSupport/build.webkit.org-config/config.json: Fixed
2985         triggerable name and trigger.
2986
2987 2010-11-01  Mario Sanchez Prada  <msanchez@igalia.com>
2988
2989         Unreviewed. Adding my IRC nickname to the list of committers.
2990
2991         * Scripts/webkitpy/common/config/committers.py:
2992
2993 2010-10-31  Robert Hogan  <robert@webkit.org>
2994
2995         Reviewed by Antonio Gomes.
2996
2997         [Qt] [Gtk] Plug-ins having upper case in mime type are failing to load
2998
2999         Qt and Gtk are case-sensitive when storing the declared mime-type
3000         of plugins. Since plugin mime-types are lowercased prior to searching
3001         for them in the plugin database, ensure they are loaded with the
3002         mime-type in lower case too.
3003
3004         Change the test netscape plugin to declare its mimetype in sentence
3005         case so that the correct behaviour is enforced.
3006
3007         https://bugs.webkit.org/show_bug.cgi?id=36815
3008
3009         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
3010         (NP_GetMIMEDescription):
3011
3012 2010-10-31  Robert Hogan  <robert@webkit.org>
3013
3014         Reviewed by Antonio Gomes.
3015
3016         [Qt] Support nodesFromRect in DRT
3017
3018         https://bugs.webkit.org/show_bug.cgi?id=48716
3019
3020         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3021         (LayoutTestController::nodesFromRect):
3022         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3023
3024 2010-10-30  Andreas Kling  <kling@webkit.org>
3025
3026         Reviewed by Antonio Gomes.
3027
3028         [Qt] QtTestBrowser: Switching view type moves the embedded inspector
3029         https://bugs.webkit.org/show_bug.cgi?id=48705
3030
3031         Reinsert the embedded inspector into the splitter after changing
3032         between QWebView/QGraphicsWebView.
3033
3034         * QtTestBrowser/launcherwindow.cpp:
3035         (LauncherWindow::init):
3036         (LauncherWindow::initializeView):
3037         * QtTestBrowser/webinspector.h:
3038         (WebInspector::WebInspector):
3039
3040 2010-10-28  Antonio Gomes  <agomes@rim.com>
3041
3042         Reviewed by Ojan Vafai.
3043
3044         Needs a "LinuxEditingBehavior", perhaps with a better name
3045         https://bugs.webkit.org/show_bug.cgi?id=36627
3046
3047         Adding support to Mac's, GTK+'s, Windows' and Chromium's LayoutTestController class to test the newly introduced Unix editing behavior.
3048
3049         * DumpRenderTree/chromium/LayoutTestController.cpp:
3050         (LayoutTestController::setEditingBehavior):
3051         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3052         (LayoutTestController::setEditingBehavior):
3053         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3054         (LayoutTestController::setEditingBehavior):
3055         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3056         (LayoutTestController::setEditingBehavior):
3057
3058 2010-10-29  Andreas Kling  <kling@webkit.org>
3059
3060         Reviewed by Kenneth Rohde Christiansen.
3061
3062         [Qt] QtTestBrowser: Add keyboard shortcut to toggle full screen (F11)
3063         https://bugs.webkit.org/show_bug.cgi?id=48695
3064
3065         * QtTestBrowser/launcherwindow.cpp:
3066         (LauncherWindow::createChrome):
3067
3068 2010-10-29  Andreas Kling  <kling@webkit.org>
3069
3070         Reviewed by Kenneth Rohde Christiansen.
3071
3072         [Qt] QtTestBrowser: Fix uninitialized read in FpsTimer
3073         https://bugs.webkit.org/show_bug.cgi?id=48675
3074
3075         FpsTimer::m_timer was never initialized and passed to QObject::killTimer()
3076         on startup with in -graphicsbased mode.
3077
3078         * QtTestBrowser/fpstimer.cpp:
3079         (FpsTimer::FpsTimer):
3080         (FpsTimer::numFrames):
3081         (FpsTimer::stop):
3082         (FpsTimer::timerEvent):
3083         * QtTestBrowser/fpstimer.h:
3084
3085 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
3086
3087         Reviewed by Tony Chang.
3088
3089         new-run-webkit-tests: change TestResults to be serializable
3090
3091         In preparation for changing new-run-webkit-tests from
3092         multithreaded to multiprocess, we need to make sure the data
3093         going between the threads is easily serialized over a socket.
3094
3095         This change adds serialization/pickling for the TestResults and
3096         TestFailure objects (using cPickle).
3097
3098         The TestFailure objects included a "has_wdiff" flag for Text
3099         results, but the flag wasn't being used, so I've removed it,
3100         simplifying the state to basically a set of enum objects with
3101         associated methods.
3102
3103         https://bugs.webkit.org/show_bug.cgi?id=48616
3104
3105         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3106         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
3107         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
3108         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
3109         * Scripts/webkitpy/layout_tests/layout_package/test_results.py: Added.
3110         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:Added. 
3111         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3112         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
3113
3114 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
3115
3116         Reviewed by Ojan Vafai.
3117
3118         new-run-webkit-tests: need to provide separate dashboard results for GPU tests
3119         https://bugs.webkit.org/show_bug.cgi?id=48687
3120
3121         Modify the steps to upload the JSON files to the dashboards so
3122         that the GPU tests don't conflict with the regular tests on a
3123         port. We do this by modifying the --builder-name parameter to
3124         append " - GPU", which should cause the app to treat the results
3125         as a completely new builder. This is a little non-obvious, but
3126         keeps us from having to restructure the app.
3127
3128         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
3129         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
3130         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3131
3132 2010-10-29  Ojan Vafai  <ojan@chromium.org>
3133
3134         Reviewed by Tony Chang.
3135
3136         [chromium] add debug logging to help diagnose flakiness dashboard issues
3137         https://bugs.webkit.org/show_bug.cgi?id=48657
3138
3139         The appengine app thinks it's getting empty files uploaded. Add some logging
3140         to see if new-run-webkit-tests agrees.
3141         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3142
3143 2010-10-29  Dimitri Glazkov  <dglazkov@chromium.org>
3144
3145         Reviewed by Ojan Vafai.
3146
3147         [Chromium/DRT] Add master-name flag to new-run-webkit-tests.
3148         https://bugs.webkit.org/show_bug.cgi?id=48649
3149
3150         * BuildSlaveSupport/build.webkit.org-config/master.cfg: Added flag.
3151
3152 2010-10-29  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
3153
3154         Reviewed by Martin Robinson.
3155
3156         Enable popup window in GtkLauncher
3157         https://bugs.webkit.org/show_bug.cgi?id=48335
3158
3159         The GtkLauncher application does not support opening new window when
3160         a link with "target=_blank" is clicked or similar call to
3161         window.open(). Instead, GtkLauncher does nothing which breaks
3162         navigation of some websites.
3163
3164         * GtkLauncher/main.c:
3165         (activate_uri_entry_cb):
3166         (update_title):
3167         (link_hover_cb):
3168         (notify_title_cb):
3169         (notify_load_status_cb):
3170         (notify_progress_cb):
3171         (destroy_cb):
3172         (go_back_cb):
3173         (go_forward_cb):
3174         (create_web_view_cb):
3175         (web_view_ready_cb):
3176         (close_web_view_cb):
3177         (create_browser):
3178         (create_statusbar):
3179         (create_toolbar):
3180         (create_window):
3181         (main):
3182
3183 2010-10-29  Adam Roben  <aroben@apple.com>
3184
3185         Teach check-webkit-style about WebKit2's idiosyncracies
3186
3187         Fixes <http://webkit.org/b/48638> Style bot complains about a number
3188         of WebKit2 conventions
3189
3190         Reviewed by Anders Carlsson.
3191
3192         * Scripts/webkitpy/style/checker.py: Excluded some rules for various
3193         WebKit2-related files. Also updated the excluded rules for
3194         WebKitAPITest to match the current code.
3195
3196 2010-10-29  Csaba Osztrogonác  <ossy@webkit.org>
3197
3198         Reviewed by Adam Roben and David Kilzer.
3199
3200         Fix and cleanup of build systems
3201         https://bugs.webkit.org/show_bug.cgi?id=48342
3202
3203         * Scripts/build-webkit:
3204          - Remove unnecessary ENABLE_SANDBOX option.
3205          - Add ENABLE_FULLSCREEN_API option.
3206
3207 2010-10-28  Adam Roben  <aroben@apple.com>
3208
3209         Switch the Windows WebKit2 bot to the Release configuration
3210
3211         We only have one machine testing WebKit2 on Windows right now, and
3212         Debug is just too slow for it to keep up.
3213
3214         Fixes (hopefully!) <http://webkit.org/b/48615> Windows WebKit2 bot is
3215         always way behind
3216
3217         Reviewed by Jon Honeycutt.
3218
3219         * BuildSlaveSupport/build.webkit.org-config/config.json:
3220
3221 2010-10-29  Leandro Gracia Gil  <leandrogracia@google.com>
3222
3223         Reviewed by Jeremy Orlow.
3224
3225         Added a second parameter to setMockSpeechInputResult for
3226         the language used in speech input.
3227         https://bugs.webkit.org/show_bug.cgi?id=47089
3228
3229         * DumpRenderTree/LayoutTestController.cpp:
3230         (setMockSpeechInputResultCallback):
3231         * DumpRenderTree/LayoutTestController.h:
3232         * DumpRenderTree/chromium/LayoutTestController.cpp:
3233         (LayoutTestController::setMockSpeechInputResult):
3234         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3235         (LayoutTestController::setMockSpeechInputResult):
3236         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3237         (LayoutTestController::setMockSpeechInputResult):
3238         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3239         (LayoutTestController::setMockSpeechInputResult):
3240         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3241         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3242         (LayoutTestController::setMockSpeechInputResult):
3243         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3244         (LayoutTestController::setMockSpeechInputResult):
3245         (LayoutTestController::pageSizeAndMarginsInPixels):
3246
3247 2010-10-28  Ojan Vafai  <ojan@chromium.org>
3248
3249         Reviewed by Tony Chang.
3250
3251         [chromium] add a result-small.json file for the test dashboard
3252         https://bugs.webkit.org/show_bug.cgi?id=48547
3253
3254         Output both a results.json file and a results-small.json file.
3255         The dashboard will load results-small.json by default so it loads faster.
3256
3257         * TestResultServer/model/jsonresults.py:
3258         * TestResultServer/model/jsonresults_unittest.py:
3259         Added a bunch of sys.path hackery. Unforunately, this uses hardcoded
3260         paths. That obviously needs to be fixed, but at least this way it
3261         clearly documents what paths are necessary.
3262
3263 2010-10-28  Eric Seidel  <eric@webkit.org>
3264
3265         Reviewed by Adam Barth.
3266
3267         webkit-patch upload calls changed_files more often than it should
3268         https://bugs.webkit.org/show_bug.cgi?id=48567
3269
3270         Passing changed_files around everywhere isn't a very elegant solution
3271         but it's the one we have for the moment.  I think keeping an explicit
3272         cache on Checkout (or making StepState() a real class) is a better
3273         long-term option.
3274
3275         Previously bug_id_for_this_commit was calling changed_files and the
3276         result was never getting cached on the state.  Now we're explicitly
3277         caching the result on the state and passing that to the bug_id_for_this_commit call.
3278
3279         I looked into building unit tests for this.  Doing so would require
3280         using a real Checkout object with a MockSCM and overriding the appropriate
3281         calls on SCM to count how often we're stating the file system.
3282         That's a useful set of tests to build for a separate change.
3283
3284         * Scripts/webkitpy/common/checkout/api.py:
3285         * Scripts/webkitpy/tool/commands/download.py:
3286         * Scripts/webkitpy/tool/commands/upload.py:
3287         * Scripts/webkitpy/tool/mocktool.py:
3288
3289 2010-10-28  Eric Seidel  <eric@webkit.org>
3290
3291         Reviewed by Adam Barth.
3292
3293         Make suggest-reviewers slightly faster
3294         https://bugs.webkit.org/show_bug.cgi?id=48562
3295
3296         Add @memoized to one more common call.
3297
3298         * Scripts/webkitpy/common/checkout/api.py:
3299
3300 2010-10-28  Eric Seidel  <eric@webkit.org>
3301
3302         Reviewed by Adam Barth.
3303
3304         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is still very slow
3305         https://bugs.webkit.org/show_bug.cgi?id=48536
3306
3307         This patch makes it about 40 seconds faster, but it still
3308         takes 1:40.  This will require more refinement.
3309
3310         The suggested reviewers list appears to be the same.
3311
3312         I think the next step may be to have it stop the search after
3313         5 reviewers are found.  We never want to suggest 30 people.
3314
3315         * Scripts/webkitpy/common/checkout/scm.py:
3316          - Using --remove-empty to theoretically stop lookups past
3317            when a file is removed.  I'm not entirely clear that the option
3318            does what it says it does.  Example:
3319            git log --pretty=format:%H -5 --remove-empty -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
3320            returns only one commit
3321            vs.
3322            git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
3323            which returns 5.  I was not aware that wx files were ever removed from the repository?
3324
3325 2010-10-28  Kinuko Yasuda  <kinuko@chromium.org>
3326
3327         Reviewed by David Levin.
3328
3329         [Chromium] Support FileSystem in chromium DRT
3330         https://bugs.webkit.org/show_bug.cgi?id=47643
3331
3332         * DumpRenderTree/chromium/WebViewHost.cpp:
3333         (WebViewHost::openFileSystem): Added.
3334         * DumpRenderTree/chromium/WebViewHost.h:
3335         (WebViewHost::openFileSystem): Added.
3336
3337 2010-10-28  Sam Weinig  <sam@webkit.org>
3338
3339         Reviewed by Anders Carlsson.
3340
3341         WKURLRefs should be allowed to be null
3342         <rdar://problem/8575621>
3343         https://bugs.webkit.org/show_bug.cgi?id=48535
3344
3345         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
3346         * TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp:
3347         (TestWebKitAPI::didStartProvisionalLoadForFrame):
3348         (TestWebKitAPI::didCommitLoadForFrame):
3349         (TestWebKitAPI::didFinishLoadForFrame):
3350         Test that URLs are null pointers when unset.
3351
3352 2010-10-28  Chris Fleizach  <cfleizach@apple.com>
3353
3354         Reviewed by Adele Peterson.
3355
3356         AX: multi select group option does not handle setting of AXSelectedChildren correctly
3357         https://bugs.webkit.org/show_bug.cgi?id=48464
3358
3359         Add support for querying information about selected children to DRT, including:
3360            selectedChildAtIndex
3361            selectedChildrenCount
3362            setSelectedChild
3363
3364         * DumpRenderTree/AccessibilityUIElement.cpp:
3365         (selectedChildAtIndexCallback):
3366         (setSelectedChildCallback):
3367         (selectedChildrenCountCallback):
3368         (AccessibilityUIElement::setSelectedChild):
3369         (AccessibilityUIElement::selectedChildrenCount):
3370         (AccessibilityUIElement::selectedChildAtIndex):
3371         (AccessibilityUIElement::getJSClass):
3372         * DumpRenderTree/AccessibilityUIElement.h:
3373         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
3374         (AccessibilityUIElement::selectedChildAtIndex):
3375         (AccessibilityUIElement::selectedChildrenCount):
3376         (AccessibilityUIElement::setSelectedChild):
3377
3378 2010-10-27  Sam Weinig  <sam@webkit.org>
3379
3380         Reviewed by Anders Carlsson.
3381
3382         Add WebKit2 API for window feature getter/setters
3383         <rdar://problem/8590373>
3384         https://bugs.webkit.org/show_bug.cgi?id=48496
3385
3386         * MiniBrowser/mac/BrowserWindowController.m:
3387         (-[BrowserWindowController awakeFromNib]):
3388         * MiniBrowser/win/BrowserView.cpp:
3389         (BrowserView::create):
3390         * WebKitTestRunner/TestController.cpp:
3391         (WTR::createOtherPage):
3392         (WTR::TestController::initialize):
3393
3394 2010-10-28  Søren Gjesse  <sgjesse@chromium.org>
3395
3396         Reviewed by Tony Chang.
3397
3398         Added support for the DumpRenderTree flags --multiple-loads and --js-flags to the Python test runner.
3399         https://bugs.webkit.org/show_bug.cgi?id=48236
3400
3401         * Scripts/webkitpy/layout_tests/port/chromium.py:
3402         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3403
3404 2010-10-28  Sergio Villar Senin  <svillar@igalia.com>
3405
3406         Reviewed by Martin Robinson.
3407
3408         [GTK] WebKitWebFrame's load-status is not properly notified to the tests
3409         https://bugs.webkit.org/show_bug.cgi?id=48048
3410
3411         DRT now listens to WebKitWebFrame load-status signals for each
3412         created frame instead of the load-status signal of the
3413         WebKitWebView that only notifies about changes in the main frame.
3414
3415         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3416         (webViewLoadFinished):
3417         (webFrameLoadStatusNotified):
3418         (frameCreatedCallback):
3419         (createWebView):
3420         (main):
3421
3422 2010-10-28  Tony Chang  <tony@chromium.org>
3423
3424         Reviewed by Kent Tamura.
3425
3426         [chromium] fix textInputController.{selectedRange,markedRange}
3427         https://bugs.webkit.org/show_bug.cgi?id=48487
3428
3429         * DumpRenderTree/chromium/TextInputController.cpp:
3430         (TextInputController::markedRange): Return arrays of ints, rather than a string
3431         (TextInputController::selectedRange): Ditto.
3432
3433 2010-10-27  Ojan Vafai  <ojan@chromium.org>
3434
3435         Reviewed by Tony Chang.
3436
3437         [chromium] add a master-name flag to new-run-webkit-tests
3438         https://bugs.webkit.org/show_bug.cgi?id=48488
3439
3440         The test results server now allows adding a master name to
3441         the uploaded files. This lets us distinguish bots that have
3442         the same name, but are on different masters.
3443
3444         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3445
3446 2010-10-28  Adam Roben  <aroben@apple.com>
3447
3448         Don't append a newline to the test output if the frame has no document
3449         element in WebKitTestRunner
3450
3451         Fixes <http://webkit.org/b/48526> Extra trailing newline when running
3452         plugins/document-open.html in WebKitTestRunner
3453
3454         Reviewed by Anders Carlsson.
3455
3456         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3457         (WTR::toJS): Added. Turns a UTF-8 C string into a JSStringRef.
3458         (WTR::hasDocumentElement): Added. Uses the JSC API to figure out
3459         whether the frame has a document element.
3460         (WTR::dumpFrameText): Match DRT by bailing (rather than appending an
3461         empty string and a newline) if the frame has no document element.
3462
3463 2010-10-28  Adam Roben  <aroben@apple.com>
3464
3465         Skip npn-invalidate-rect-invalidates-window.html on headless XP
3466         machines
3467
3468         TestNetscapePlugIn never receives a WM_PAINT message on headless XP
3469         machines, so this test times out. Fixing the test is covered by
3470         <http://webkit.org/b/48333>.
3471
3472         * Scripts/old-run-webkit-tests: Skip
3473         npn-invalidate-rect-invalidates-window.html on Windows if accelerated
3474         compositing support is disabled, which likely means we're on a headless
3475         XP machine.
3476
3477 2010-10-28  Kimmo Kinnunen  <kimmok@iki.fi>
3478
3479         Adding myself as a committer.
3480
3481         * Scripts/webkitpy/common/config/committers.py:
3482
3483 2010-10-27  Eric Seidel  <eric@webkit.org>
3484
3485         Reviewed by Adam Barth.
3486
3487         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is very slow
3488         https://bugs.webkit.org/show_bug.cgi?id=48500
3489
3490         This doesn't fix the problem, but it makes things slightly better.
3491         Each git svn find-rev call takes about .25 seconds on my desktop.
3492         This patch uses a new memoized class to avoid those calls when possible.
3493
3494         The real slowness is still git log on some files, like:
3495         git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
3496         I'm not yet sure how to make the pathological git logs better.
3497
3498         * Scripts/webkitpy/common/checkout/scm.py:
3499         * Scripts/webkitpy/common/memoized.py: Added.
3500         * Scripts/webkitpy/common/memoized_unittest.py: Added.
3501
3502 2010-10-27  Eric Seidel  <eric@webkit.org>
3503
3504         Reviewed by Ojan Vafai.
3505
3506         EWS bots should not use --quiet when running build-webkit
3507         https://bugs.webkit.org/show_bug.cgi?id=48482
3508
3509         --quiet is only correct when the sub-process does the error reporting.
3510         In the case of _can_build() the parent process is reporting the error.
3511         We'd like the full build log at queues.webkit.org so someone can look
3512         at the log and understand why the EWS is failing to build trunk.
3513
3514         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
3515         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
3516         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
3517         * Scripts/webkitpy/tool/mocktool.py:
3518
3519 2010-10-27  Brian Weinstein  <bweinstein@apple.com>
3520
3521         More Windows build fixage. Rename a variable that was named string.
3522
3523         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3524         (WTR::hasPrefix):
3525
3526 2010-10-27  Ojan Vafai  <ojan@chromium.org>
3527
3528         Reviewed by Dimitri Glazkov.
3529
3530         [chromium] Make the test results server store which master the bot is on
3531         https://bugs.webkit.org/show_bug.cgi?id=48478
3532
3533         The chromium bots recently changed so that there are multiple slaves with
3534         the same name on different masters. Up till now, the test results server
3535         assumed slave names were unique. Adds a master field to the file in order
3536         to distinguish.
3537
3538         Also, for files that currently lack a master or testtype, set them appropriately.
3539
3540         * TestResultServer/handlers/testfilehandler.py:
3541         * TestResultServer/index.yaml:
3542         * TestResultServer/model/jsonresults.py:
3543         * TestResultServer/model/testfile.py:
3544         * TestResultServer/templates/showfilelist.html:
3545         * TestResultServer/templates/uploadform.html:
3546
3547 2010-10-26  Darin Adler  <darin@apple.com>
3548
3549         Reviewed by Sam Weinig.
3550
3551         WebKitTestRunner needs to support layoutTestController.dumpBackForwardList
3552         https://bugs.webkit.org/show_bug.cgi?id=42322
3553         rdar://problem/8193631
3554
3555         WebKitTestRunner needs to support layoutTestController.clearBackForwardList
3556         https://bugs.webkit.org/show_bug.cgi?id=42333
3557         rdar://problem/8193643
3558
3559         * WebKitTestRunner/Configurations/InjectedBundle.xcconfig:
3560         Renamed the product to WebKitTestRunnerInjectedBundle to avoid
3561         name conflicts in the build directory.
3562
3563         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
3564         Added dumpBackForwardList and clearBackForwardList.
3565
3566         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
3567         (WTR::InjectedBundle::InjectedBundle): Removed initialization of m_mainPage.
3568         (WTR::InjectedBundle::didCreatePage): Removed unneeded underscore.
3569         (WTR::InjectedBundle::willDestroyPage): Ditto.
3570         (WTR::InjectedBundle::didReceiveMessage): Ditto.
3571         (WTR::InjectedBundle::initialize): Ditto.
3572         (WTR::InjectedBundle::didCreatePage): Changed code to use m_pages
3573         instead of m_mainPage and m_otherPages.
3574         (WTR::InjectedBundle::willDestroyPage): Ditto.
3575         (WTR::InjectedBundle::page): Ditto.
3576         (WTR::InjectedBundle::beginTesting): Ditto.
3577         (WTR::InjectedBundle::done): Ditto.
3578         (WTR::InjectedBundle::closeOtherPages): Ditto.
3579         (WTR::InjectedBundle::dumpBackForwardListsForAllPages): Added.
3580
3581         * WebKitTestRunner/InjectedBundle/InjectedBundle.h: Changed
3582         page function to longer be inline, and pageCount function
3583         to use m_pages. Added dumpBackForwardListsForAllPages, and
3584         removed some underscores. Replaced m_mainPage and m_otherPageas
3585         with m_pages.
3586
3587         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3588         (WTR::adoptWK): Added.
3589         (WTR::hasPrefix): Added.
3590         (WTR::InjectedBundlePage::reset): Added code to set up
3591         m_previousTestBackForwardListItem.
3592         (WTR::InjectedBundlePage::dump): Added code to call
3593         dumpBackForwardListsForAllPages.
3594         (WTR::compareByTargetName): Added.
3595         (WTR::dumpBackForwardListItem): Added.
3596         (WTR::InjectedBundlePage::dumpBackForwardList): Added.
3597
3598         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h: Added
3599         dumpBackForwardList and m_previousTestBackForwardListItem.
3600
3601         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
3602         (WTR::LayoutTestController::LayoutTestController): Initialize
3603         m_shouldDumpBackForwardListsForAllWindows to false.
3604         (WTR::LayoutTestController::clearBackForwardList): Added.
3605
3606         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
3607         Added dumpBackForwardList, clearBackForwardList,
3608         shouldDumpBackForwardListsForAllWindows, and
3609         m_shouldDumpBackForwardListsForAllWindows.
3610
3611         * WebKitTestRunner/TestController.cpp:
3612         (WTR::TestController::initialize): Set up
3613         didReceiveSynchronousMessageFromInjectedBundle.
3614         (WTR::TestController::didReceiveSynchronousMessageFromInjectedBundle):
3615         Added.
3616         * WebKitTestRunner/TestController.h: Ditto.
3617
3618         * WebKitTestRunner/TestInvocation.cpp:
3619         (WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):
3620         Added.
3621         * WebKitTestRunner/TestInvocation.h: Ditto.
3622
3623         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
3624         Renamed bundle to WebKitTestRunnerInjectedBundle (see above).
3625         * WebKitTestRunner/mac/TestControllerMac.mm:
3626         (WTR::TestController::initializeInjectedBundlePath): Ditto.
3627
3628 2010-10-27  Chris Rogers  <crogers@google.com>
3629
3630         Reviewed by Chris Marrin.
3631
3632         Add ENABLE_WEB_AUDIO feature enable flag (initially disabled) to build-webkit
3633         https://bugs.webkit.org/show_bug.cgi?id=48279
3634
3635         * Scripts/build-webkit:
3636
3637 2010-10-27  Eric Seidel  <eric@webkit.org>
3638
3639         Unreviewed.
3640
3641         build-webkit should collect Visual Studio Express logs and display them
3642         https://bugs.webkit.org/show_bug.cgi?id=39199
3643
3644         It turns out my previous patch wasn't actually working on the win-ews
3645         machine.  So I've fixed my mistakes from before.
3646
3647         * Scripts/build-webkit:
3648          - Windows VSE builds change the CWD while building.  Why?  Who knows.
3649         * Scripts/print-vse-failure-logs:
3650          - windows VSE builds don't use Debug/Release as I expected, so I've
3651            moved off of --configuration to --top-level and added the /obj
3652            optimization while I was there.
3653
3654 2010-10-27  Dimitri Glazkov  <dglazkov@chromium.org>
3655
3656         Unreviewed, rolling out r70674.
3657         http://trac.webkit.org/changeset/70674
3658         https://bugs.webkit.org/show_bug.cgi?id=48053
3659
3660         Broke Chromium Windows build.
3661
3662         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3663         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3664
3665 2010-10-25  Tony Chang  <tony@chromium.org>
3666
3667         Reviewed by Anders Carlsson.
3668
3669         compile TestNetscapePlugIn on chromium linux
3670         https://bugs.webkit.org/show_bug.cgi?id=48274
3671
3672         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
3673         (NP_Initialize): On Linux, plugin funcs are set in initialize.
3674         (NPP_New): Mark the plugin as windowless.
3675         (NPP_GetValue): Handle mime type values.
3676         (NP_GetMIMEDescription):
3677         (NP_GetValue):
3678         * DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npapi.h: define TRUE and FALSE, which are in webkit's npapi.h.
3679         * DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npfunctions.h: Pull in npapi.h to get TRUE/FALSE (matches webkit's npfunctions.h)
3680
3681 2010-10-27  Kenneth Rohde Christiansen  <kenneth@webkit.org>
3682
3683         Reviewed by Anders Carlsson.
3684
3685         Remove contentSizeChanged callbacks as it is no longer
3686         part of the public UIClient.
3687
3688         Make WKPageContentsSizeChangedCallback be a private API
3689         https://bugs.webkit.org/show_bug.cgi?id=48409
3690
3691         * MiniBrowser/mac/BrowserWindowController.m:
3692         (-[BrowserWindowController awakeFromNib]):
3693         * MiniBrowser/win/BrowserView.cpp:
3694         (BrowserView::create):
3695         * WebKitTestRunner/TestController.cpp:
3696         (WTR::createOtherPage):
3697         (WTR::TestController::initialize):
3698
3699 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3700
3701         Reviewed by Ojan Vafai.
3702
3703         Make http locking default in NRWT.
3704         https://bugs.webkit.org/show_bug.cgi?id=48053
3705
3706         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3707         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3708
3709 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3710
3711         Reviewed by Tony Chang.
3712
3713         [NRWT] Fix http lock on Windows platform
3714         https://bugs.webkit.org/show_bug.cgi?id=48321
3715
3716         * Scripts/webkitpy/layout_tests/port/http_lock.py:
3717
3718 2010-10-27  Satish Sampath  <satish@chromium.org>
3719
3720         Unreviewed, rolling out r70665.
3721         http://trac.webkit.org/changeset/70665
3722         https://bugs.webkit.org/show_bug.cgi?id=47089
3723
3724         Need to address Alexey's review comments.
3725
3726         * DumpRenderTree/LayoutTestController.cpp:
3727         (setMockSpeechInputResultCallback):
3728         * DumpRenderTree/LayoutTestController.h:
3729         * DumpRenderTree/chromium/LayoutTestController.cpp:
3730         (LayoutTestController::setMockSpeechInputResult):
3731         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3732         (LayoutTestController::setMockSpeechInputResult):
3733         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3734         (LayoutTestController::setMockSpeechInputResult):
3735         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3736         (LayoutTestController::setMockSpeechInputResult):
3737         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3738         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3739         (LayoutTestController::setMockSpeechInputResult):
3740         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3741         (LayoutTestController::setMockSpeechInputResult):
3742
3743 2010-10-27  Leandro Gracia Gil  <leandrogracia@google.com>
3744
3745         Reviewed by Jeremy Orlow.
3746
3747         Added a second parameter to setMockSpeechInputResult for
3748         the language used in speech input.
3749         https://bugs.webkit.org/show_bug.cgi?id=47089
3750
3751         * DumpRenderTree/LayoutTestController.cpp:
3752         (setMockSpeechInputResultCallback):
3753         * DumpRenderTree/LayoutTestController.h:
3754         * DumpRenderTree/chromium/LayoutTestController.cpp:
3755         (LayoutTestController::setMockSpeechInputResult):
3756         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3757         (LayoutTestController::setMockSpeechInputResult):
3758         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3759         (LayoutTestController::setMockSpeechInputResult):
3760         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3761         (LayoutTestController::setMockSpeechInputResult):
3762         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3763         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3764         (LayoutTestController::setMockSpeechInputResult):
3765         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3766         (LayoutTestController::setMockSpeechInputResult):
3767         (LayoutTestController::pageSizeAndMarginsInPixels):
3768
3769 2010-10-27  Adam Roben  <aroben@apple.com>
3770
3771         Reset TestNetscapePlugIn's NPP_GetValue pointer when
3772         NullNPPGetValuePointer finishes running
3773
3774         Fixes <http://webkit.org/b/48435> REGRESSION (r70655): Many plugins
3775         tests are failing on Qt
3776
3777         Reviewed by Anders Carlsson.
3778
3779         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
3780         (PluginTest::NPP_Destroy):
3781         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
3782         Added a do-nothing NPP_Destroy implementation.
3783
3784         * DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp:
3785         (NullNPPGetValuePointer::NullNPPGetValuePointer): Save the original
3786         NPP_GetValue pointer in m_originalNPPGetValuePointer so we can restore
3787         it later.
3788         (NullNPPGetValuePointer::NPP_Destroy): Added. Restores the original
3789         NPP_GetValue pointer so it can be used in other tests.
3790
3791         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
3792         (NPP_Destroy):
3793         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
3794         (webkit_test_plugin_destroy_instance):
3795         Call through to the PluginTest when NPP_Destroy is called.
3796
3797 2010-10-26  Brian Weinstein  <bweinstein@apple.com>
3798
3799         Reviewed by Adam Roben.
3800
3801         WebKit2 shouldn't try to send an empty user agent
3802         https://bugs.webkit.org/show_bug.cgi?id=48397
3803         
3804         Add a test that when we set our custom user agent to an empty string, we don't send
3805         an empty user agent.
3806
3807         * TestWebKitAPI/Tests/WebKit2/PreventEmptyUserAgent.cpp: Added.
3808         (TestWebKitAPI::didRunJavaScript): Make sure that the result of navigator.userAgent isn't empty.
3809         (TestWebKitAPI::TEST): Set our custom user agent to the empty string, and run navigator.userAgent.
3810         
3811         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Add the new file.
3812         * TestWebKitAPI/win/TestWebKitAPI.vcproj: Ditto.
3813
3814 2010-10-27  Adam Roben  <aroben@apple.com>
3815
3816         Check in file I forgot in r70653
3817
3818         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
3819         (NP_Initialize): Save the NPPluginFuncs struct the browser passed to us
3820         so we can be naughty and modify it later.
3821
3822 2010-10-27  Adam Roben  <aroben@apple.com>
3823
3824         Test that WebKit doesn't crash if the plugin passes 0 for its
3825         NPP_GetValue pointer
3826
3827         Test for <http://webkit.org/b/48433> Crash in
3828         NetscapePlugin::shouldLoadSrcURL when using Shockwave Director 10.3 in
3829         WebKit2 on Windows
3830
3831         Reviewed by Eric Carlson.
3832
3833         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3834         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
3835         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
3836         * GNUmakefile.am:
3837         Added NullNPPGetValuePointer.cpp.
3838
3839         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
3840         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
3841         Set up a place to store the NPPluginFuncs struct the browser passed to us.
3842
3843         * DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp: Added.
3844         (NullNPPGetValuePointer::NullNPPGetValuePointer): Null out the
3845         NPP_GetValue pointer we passed to the browser to simulate a plugin that
3846         doesn't implement NPP_GetValue.
3847         (NullNPPGetValuePointer::NPP_GetValue): Print an error message. If this
3848         function is called, it means that WebKit has changed in a way that
3849         makes this test invalid.
3850
3851         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
3852         (NP_GetEntryPoints): Save the NPPluginFuncs struct the browser passed
3853         to us so we can be naughty and modify it later.
3854
3855 2010-10-27  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
3856
3857         Reviewed by Andreas Kling.
3858
3859         Remove references to ancient QGVLauncher and QtLauncher
3860         https://bugs.webkit.org/show_bug.cgi?id=48430
3861
3862         QtTestBrowser substitutes both and has checks enabled.
3863
3864         * Scripts/webkitpy/style/checker.py: remove references
3865         * Scripts/webkitpy/style/checker_unittest.py: ditto
3866
3867 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3868
3869         Reviewed by Ojan Vafai.
3870
3871         [NRWT] Don't use image hash when it's no need in single test mode.
3872         https://bugs.webkit.org/show_bug.cgi?id=48326
3873
3874         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3875
3876 2010-10-27  David Kilzer  <ddkilzer@apple.com>
3877
3878         Fix leak of CFMutableDictionaryRef in createXMLStringFromWebArchiveData()
3879
3880         Reviewed by Adam Roben.
3881
3882         Follow-up fix for: <https://bugs.webkit.org/show_bug.cgi?id=48278>
3883
3884         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
3885         (createXMLStringFromWebArchiveData): Use RetainPtr<> to fix a
3886         leak introduced in r70613.
3887
3888 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3889
3890         Reviewed by Csaba Osztrogonác.
3891
3892         [NRWT] Add platform specific baseline search paths for Qt port
3893         https://bugs.webkit.org/show_bug.cgi?id=48428
3894
3895         * Scripts/webkitpy/layout_tests/port/qt.py:
3896
3897 2010-10-27  Adam Roben  <aroben@apple.com>
3898
3899         Catch exceptions when checking if we're inside a git working directory
3900
3901         Fixes <http://webkit.org/b/48420> REGRESSION (r70562): test-webkitpy
3902         fails on systems without git installed
3903
3904         Reviewed by Anders Carlsson.
3905
3906         * Scripts/webkitpy/common/net/credentials.py:
3907         (Credentials._credentials_from_git): Put the call to
3908         Git.in_working_directory inside the try/except since it, too, attempts
3909         to execute git and thus will throw on systems that don't have git
3910         installed.
3911
3912 2010-10-27  Nikolas Zimmermann  <nzimmermann@rim.com>
3913
3914         Reviewed by David Kilzer.
3915
3916         Convert DumpRenderTree webarchive code to CoreFoundation
3917         https://bugs.webkit.org/show_bug.cgi?id=48278
3918
3919         CFPropertyListCreateWithData and CFPropertyListCreateData are only available in 10.6+.
3920         Replace CFPropertyListCreateWithData by a combination of CFReadStreamCreateWithBytesNoCopy and CFPropertyListCreateFromStream.
3921         Replace CFPropertyListCreateData by CFPropertyListCreateXMLData.
3922
3923         These changes are wrapped in BUILDING_ON_LEOPARD, as the methods are deprecated on 10.6+.
3924
3925         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
3926         (createXMLStringFromWebArchiveData):
3927
3928 2010-10-26  Antonio Gomes  <agomes@rim.com>
3929
3930         Reviewed by Martin Robinson.
3931
3932         Remove the absolute path used to include DumpRenderTreeSupportGtk.h from LayoutTestController.h
3933
3934         It turns out this is rather unneeded since WebKitTools/GNUMakefile.am has WebKit/gtk/ in its include
3935         path, and then we can just do #include "WebCoreSupport/DumpRenderTreeSupportGtk.h"
3936
3937         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3938
3939 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3940
3941         Rename WebArchiveDumpSupport.mm to WebArchiveDumpSupport.cpp
3942
3943         Reviewed by Adam Roben.
3944
3945         Part 4 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3946
3947         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3948         Updated for file moves and renames.
3949         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.mm.
3950         * DumpRenderTree/cf/WebArchiveDumpSupport.h: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.h.
3951
3952 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3953
3954         Convert WebArchiveDumpSupport.mm from NS objects to CF types
3955
3956         Reviewed by Adam Roben.
3957
3958         Part 3 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3959
3960         In order to share WebArchive code between the Mac and Windows
3961         ports, the code in WebArchiveDumpSupport.mm was converted from
3962         Cocoa to C++ using CoreFoundation (CF) types.
3963
3964         * DumpRenderTree/mac/DumpRenderTree.mm:
3965         (dump): Renamed serializeWebArchiveToXML() to
3966         createXMLStringFromWebArchiveData() and added HardAutorelease()
3967         to prevent leaks.
3968         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
3969         (createXMLStringFromWebArchiveData): Renamed from
3970         serializeWebArchiveToXML().  Changed to use CF types.
3971         (createCFURLResponseFromResponseData): Renamed from
3972         unarchiveNSURLResponseFromResponseData().  Changed to use CF
3973         types for its parameter and return type.
3974         * DumpRenderTree/mac/WebArchiveDumpSupport.mm: Replaced use of
3975         NS objects with CF types.  It will be renamed to *.cpp in a
3976         future commit.
3977         (convertMIMEType): Changed to use case-insensitive string
3978         comparisons.
3979         (convertWebResourceDataToString):
3980         (normalizeHTTPResponseHeaderFields):
3981         (normalizeWebResourceURL):
3982         (convertWebResourceResponseToDictionary):
3983         (compareResourceURLs):
3984         (createXMLStringFromWebArchiveData):
3985         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
3986         (createCFURLResponseFromResponseData): Renamed from
3987         unarchiveNSURLResponseFromResponseData().  Changed to use CF
3988         types for its parameter and return type.
3989
3990 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3991
3992         Extract use of NSKeyedUnarchiver from WebArchiveDumpSupport.mm
3993
3994         Reviewed by Adam Roben.
3995
3996         Part 2 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3997
3998         There is no equivalent to NSKeyedUnarchiver in CoreFoundation,
3999         so extract it into a platform-specific source file.
4000
4001         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
4002         (unarchiveNSURLResponseFromResponseData): Added declaration.
4003         * DumpRenderTree/mac/WebArchiveDumpSupport.mm:
4004         (convertWebResourceResponseToDictionary): Extracted code to
4005         unarchiveNSURLResponseFromResponseData() in
4006         WebArchiveDumpSupportMac.mm.  Updated logic to return early if
4007         nil is returned from unarchiveNSURLResponseFromResponseData().
4008         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
4009         (unarchiveNSURLResponseFromResponseData): Added.  Extracted
4010         code from convertWebResourceResponseToDictionary() in
4011         WebArchiveDumpSupport.mm.
4012
4013 2010-10-26  David Kilzer  <ddkilzer@apple.com>
4014
4015         Extract call to -[WebHTMLRepresentation supportedNonImageMIMETypes] from WebArchiveDumpSupport.mm
4016
4017         Reviewed by Adam Roben.
4018
4019         Part 1 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
4020
4021         The call to -[WebHTMLRepresentation supportedNonImageMIMETypes]
4022         is not cross-platform between Mac and Windows, so extract it
4023         into a platform-specific source file.
4024
4025         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Added
4026         WebArchiveDumpSupportMac.mm to the project.
4027         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
4028         (supportedNonImageMIMETypes): Added declaration.
4029         * DumpRenderTree/mac/WebArchiveDumpSupport.mm:
4030         (convertWebResourceDataToString): Replaced call to
4031         -[WebHTMLRepresentation supportedNonImageMIMETypes] with
4032         supportedNonImageMIMETypes().
4033         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm: Added.
4034         (supportedNonImageMIMETypes): Added.  Extracted from
4035         WebArchiveDumpSupport.mm.
4036
4037 2010-10-26  Antonio Gomes  <agomes@rim.com>
4038
4039         Reviewed by Martin Robinson.
4040
4041         [GTK] Implement DumpRenderTreeSupportGtk (similarly to DumpRenderTreeSupportQt idea)
4042         https://bugs.webkit.org/show_bug.cgi?id=48199
4043
4044         Implements support to WebKitTabToLinksPreferenceKey through LayoutTestController::overridePreference.
4045         The corresponding DumpRenderTreeSupportGtk method is called in the DRT context only.
4046
4047         * DumpRenderTree/gtk/DumpRenderTree.cpp:
4048         (resetDefaultsToConsistentValues):
4049         (createWebView):
4050         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
4051         (LayoutTestController::overridePreference):
4052
4053 2010-10-26  Ariya Hidayat  <ariya@sencha.com>
4054
4055         Unreviewed, change the order of my emails for bugzilla autocompletion.
4056
4057         * Scripts/webkitpy/common/config/committers.py:
4058
4059 2010-10-26  Eric Seidel  <eric@webkit.org>
4060
4061         Reviewed by David Kilzer.
4062
4063         build-webkit should collect Visual Studio Express logs and display them
4064         https://bugs.webkit.org/show_bug.cgi?id=39199
4065
4066         * Scripts/build-webkit:
4067         * Scripts/print-vse-failure-logs: Added.
4068         * Scripts/webkitdirs.pm:
4069
4070 2010-10-26  David Kilzer  <ddkilzer@apple.com>
4071
4072         <http://webkit.org/b/48224> build-webkit: add support for --meter-tag switch
4073
4074         Reviewed by Csaba Osztrogonác.
4075
4076         * Scripts/build-webkit: Added support for --meter-tag switch.
4077         It should have been added with r60820.
4078
4079 2010-10-26  Ademar de Souza Reis Jr.  <ademar.reis@openbossa.org>
4080
4081         Reviewed by Dumitru Daniliuc.
4082
4083         check-webkit-style fails on operator+=, operator-=, ... methods
4084         https://bugs.webkit.org/show_bug.cgi?id=48258
4085
4086         * Scripts/webkitpy/style/checkers/cpp.py: Added exceptions
4087         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added unit-tests
4088
4089 2010-10-26  Ojan Vafai  <ojan@chromium.org>
4090
4091         Reviewed by Tony Chang.
4092
4093         remove DEFER support from new-run-webkit-tests
4094         https://bugs.webkit.org/show_bug.cgi?id=48387
4095
4096         DEFER was needed when we were trying to ship Chrome beta.
4097         Now it's just extra complication.
4098
4099         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
4100         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
4101         * Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
4102         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
4103         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
4104         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
4105
4106 2010-10-26  Eric Seidel  <eric@webkit.org>
4107
4108         Unreviewed.  Adding import with_statement to fix python 2.5.
4109
4110         Teach webkit-patch how to read credentials from the environment
4111         https://bugs.webkit.org/show_bug.cgi?id=48275
4112
4113         * Scripts/webkitpy/common/net/credentials_unittest.py:
4114          - import with_statement to fix python 2.5.
4115
4116 2010-10-26  Alexey Proskuryakov  <ap@apple.com>
4117
4118         Reviewed by Sam Weinig.
4119
4120         https://bugs.webkit.org/show_bug.cgi?id=48375
4121         <rdar://problem/8392724> Need delegate calls in PageLoaderClient to indicate if we have
4122         loaded insecure content
4123
4124         Updated for WebKit2 changes.
4125
4126         * MiniBrowser/mac/BrowserWindowController.m:
4127         (didDisplayInsecureContentForFrame):
4128         (didRunInsecureContentForFrame):
4129         (-[BrowserWindowController awakeFromNib]):
4130         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
4131         (WTR::InjectedBundlePage::InjectedBundlePage):
4132         (WTR::InjectedBundlePage::didDisplayInsecureContentForFrame):
4133         (WTR::InjectedBundlePage::didRunInsecureContentForFrame):
4134         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
4135         * WebKitTestRunner/TestController.cpp:
4136         (WTR::TestController::initialize):
4137
4138 2010-10-26  Adam Roben  <aroben@apple.com>
4139
4140         Clear up confusion between "3D rendering" and "accelerated compositing"
4141
4142         WebKit has no software-based 3D rendering implementation, so 3D
4143         rendering can only be enabled when accelerated compositing is. But DRT
4144         was falsely reporting that 3D rendering was available on machines that
4145         don't support accelerated compositing, leading to much confusion.
4146
4147         Reviewed by Darin Adler.
4148
4149         Fixes <http://webkit.org/b/48370> REGRESSION (r70540): Many
4150         transforms/3d tests are failing on the XP bots
4151
4152         * DumpRenderTree/win/DumpRenderTree.cpp:
4153         (main): When ENABLE(3D_RENDERING) is turned on, only report that 3D
4154         rendering is available when acclerated compositing is also available.
4155
4156         * Scripts/old-run-webkit-tests: Skip tests that use the
4157         -webkit-transform-3d media query when 3D rendering is disabled,
4158         rather than when accelerated compositing is disabled.
4159
4160 2010-10-26  Eric Seidel  <eric@webkit.org>
4161
4162         Reviewed by Ojan Vafai.
4163
4164         Remove rietveld code now that it's unused
4165         https://bugs.webkit.org/show_bug.cgi?id=48359
4166
4167         Was cool that we added this support, but now that it's
4168         not used it makes little sense to keep it around.
4169         We can always restore this code from SVN if we need it.
4170
4171         * Scripts/webkitpy/common/config/__init__.py:
4172         * Scripts/webkitpy/common/net/bugzilla.py:
4173         * Scripts/webkitpy/common/net/bugzilla_unittest.py:
4174         * Scripts/webkitpy/common/net/rietveld.py: Removed.
4175         * Scripts/webkitpy/common/net/rietveld_unittest.py: Removed.
4176         * Scripts/webkitpy/thirdparty/__init__.py:
4177         * Scripts/webkitpy/tool/commands/download.py:
4178         * Scripts/webkitpy/tool/commands/download_unittest.py:
4179         * Scripts/webkitpy/tool/commands/queues.py:
4180         * Scripts/webkitpy/tool/commands/queues_unittest.py:
4181         * Scripts/webkitpy/tool/main.py:
4182         * Scripts/webkitpy/tool/mocktool.py:
4183         * Scripts/webkitpy/tool/steps/__init__.py:
4184         * Scripts/webkitpy/tool/steps/postcodereview.py: Removed.
4185
4186 2010-10-26  Anders Carlsson  <andersca@apple.com>
4187
4188         Reviewed by Sam Weinig.
4189
4190         pageDidScroll callback should be on the UI process client rather than (or in addition to) the web process client
4191         https://bugs.webkit.org/show_bug.cgi?id=48366
4192         <rdar://problem/8595202>
4193
4194         * MiniBrowser/mac/BrowserWindowController.m:
4195         (-[BrowserWindowController awakeFromNib]):
4196         * WebKitTestRunner/TestController.cpp:
4197         (WTR::createOtherPage):
4198         (WTR::TestController::initialize):
4199
4200 2010-10-26  Eric Seidel  <eric@webkit.org>
4201
4202         Reviewed by Tony Chang.
4203
4204         Teach webkit-patch how to read credentials from the environment
4205         https://bugs.webkit.org/show_bug.cgi?id=48275
4206
4207         This makes it possible for svn users to have their bugzilla credentials
4208         stored in their environment instead of typing them every time.
4209
4210         We need this for making it easy to run the win-ews bot (which currently
4211         uses svn instead of git).
4212
4213         * Scripts/webkitpy/common/net/credentials.py:
4214         * Scripts/webkitpy/common/net/credentials_unittest.py:
4215
4216 2010-10-26  Kenichi Ishibashi  <bashi@google.com>
4217
4218         Reviewed by Kent Tamura.
4219
4220         Input Method inserts conversion candidates unexpectedly
4221         https://bugs.webkit.org/show_bug.cgi?id=46868
4222
4223         Adds setComposition() to TextInputController to make DRT emulate
4224         an input method behavior.
4225
4226         * DumpRenderTree/chromium/TextInputController.cpp:
4227         (TextInputController::TextInputController):
4228         (TextInputController::setComposition): Added.
4229         * DumpRenderTree/chromium/TextInputController.h:
4230
4231 2010-10-26  Simon Fraser  <simon.fraser@apple.com>
4232
4233         Reviewed by Sam Weinig.
4234
4235         Support layoutTestController.layerTreeAsText in WebKitTestRunner
4236         https://bugs.webkit.org/show_bug.cgi?id=42145
4237         
4238         Implement layerTreeAsText() in WebKitTestRunner.
4239
4240         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
4241         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
4242         (WTR::LayoutTestController::layerTreeAsText):
4243         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
4244
4245 2010-10-26  Adam Barth  <abarth@webkit.org>
4246
4247         Reviewed by Eric Seidel.
4248
4249         webkit-patch rollout produces incorrect patch when using svn move
4250         https://bugs.webkit.org/show_bug.cgi?id=48244
4251
4252         We need to flush our caches when we modify the working copy.
4253
4254         * Scripts/webkitpy/tool/steps/revertrevision.py:
4255
4256 2010-10-26  Eric Seidel  <eric@webkit.org>
4257
4258         Reviewed by Adam Barth.
4259
4260         commit-queue flaky test message can list the same author more than once
4261         https://bugs.webkit.org/show_bug.cgi?id=48268
4262
4263         tonikitoo reported to me over IRC this morning that he's seen
4264         the commit-queue report flaky tests with author lists like:
4265         "adam, adam and adam", suggesting we're not uniquing authors
4266         before writing the message.
4267
4268         I fixed the uniquing and added a bunch more unit testing.
4269
4270         * Scripts/webkitpy/tool/commands/queues.py:
4271         * Scripts/webkitpy/tool/commands/queues_unittest.py:
4272         * Scripts/webkitpy/tool/mocktool.py:
4273
4274 2010-10-26  Adam Roben  <aroben@apple.com>
4275
4276         Pull in the FeatureDefines*.vsprops files when building DRT
4277
4278         This ensures that various ENABLE() macros will get set correctly.
4279         (Prior to r70320 we were picking up the ENABLE(3D_RENDERING) flag
4280         through wtf/Platform.h.)
4281
4282         Fixes <http://webkit.org/b/48343> REGRESSION (r70320): DumpRenderTree
4283         says 3D_RENDERING is disabled even when it is enabled
4284
4285         Reviewed by Ada Chan.
4286
4287         * DumpRenderTree/win/DumpRenderTree.vcproj: Added
4288         FeaturesDefines*.vsprops to all configurations.
4289
4290         * DumpRenderTree/win/DumpRenderTreeWin.h: Touched to force a rebuild.
4291
4292 2010-10-26  Søren Gjesse  <sgjesse@chromium.org>
4293
4294         Reviewed by Dimitri Glazkov.
4295
4296         Added options --multiple-loads and --js-flags to chromium DumpRenderTree. The option --multiple-loads=X
4297         is used to have DumpRenderTree load each test it runs X times. To be able to have more fine-grained control
4298         of how the JavaScript engine behaves for each load the flag --js-flags can specify a list of flag-sets like this
4299
4300           --js-flags="--xxx,--noxxx --yyy,--noyyy"
4301
4302         First load will run with --xxx, the second with --yyy and the third without any (the 'no' prefix is handled by
4303         V8 to turn off the flag).
4304
4305         The changes to the Python test runner will be in a separate change.
4306
4307
4308         * DumpRenderTree/chromium/DumpRenderTree.cpp:
4309         (runTest):
4310         (main):
4311         * DumpRenderTree/chromium/TestShell.cpp:
4312         (TestShell::TestShell):
4313         (TestShell::runFileTest):
4314         (TestShell::testFinished):
4315         * DumpRenderTree/chromium/TestShell.h:
4316         (TestShell::loadCount):
4317         (TestShell::setLoadCount):
4318         (TestShell::javaScriptFlagsForLoad):
4319         (TestShell::setJ