[Qt] Add top-level .qmake.conf
[WebKit.git] / Tools / ChangeLog
1 2012-07-06  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
2
3         [Qt] Add top-level .qmake.conf
4
5         With Qt5, this makes setting $QMAKEPATH externally unnecessary.
6
7         The magic in the perl scripts to set QMAKEPATH is still there, as it doesn't
8         hurt, and is still required for Qt4.
9
10         https://bugs.webkit.org/show_bug.cgi?id=90461
11
12         Reviewed by Tor Arne Vestbø.
13
14         * qmake/mkspecs/features/default_pre.prf:
15
16 2012-07-06  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
17
18         [Qt] Unify qtFeatureDefaults code paths
19
20         There's no reason to run qmake on features.prf directly anymore, as we
21         can selectivly run configure tests in the project file now.
22
23         https://bugs.webkit.org/show_bug.cgi?id=90461
24
25         Reviewed by Tor Arne Vestbø.
26
27         * Scripts/webkitdirs.pm:
28         (qtFeatureDefaults):
29         * qmake/configure.pri:
30         * qmake/mkspecs/features/features.prf:
31
32 2012-07-06  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
33
34         [Qt] Don't let qt_webkit.pri proclaim its own location
35
36         This won't work any more with recent Qt5 versions, as the forwarding
37         pri is created by qt_module_config, which needs MODULE_PRI to be set
38         up already.
39
40         We also need to load build_config, not qt_module.
41
42         https://bugs.webkit.org/show_bug.cgi?id=90461
43
44         Reviewed by Tor Arne Vestbø.
45
46         * qmake/qt_webkit.pri:
47
48 2012-07-06  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
49
50         [Qt] Remove redundant CONFIG+=module
51
52         qt_module_config takes care of that.
53
54         In api.pri we are actually testing the flag ourselves, so now we need to
55         test a related flag qt_module_config sets instead.
56
57         Reviewed by Tor Arne Vestbø.
58
59         * qmake/mkspecs/features/default_post.prf:
60
61 2012-07-06  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
62
63         [Qt] Switch to new-style Qt 5 configure tests
64
65         Use explicit project file action instead of syncqt magic.
66
67         https://bugs.webkit.org/show_bug.cgi?id=90461
68
69         Reviewed by Tor Arne Vestbø.
70
71         * Scripts/webkitdirs.pm:
72         (buildQMakeProjects):
73         * qmake/configure.pri: Copied from Tools/qmake/configure.pro.
74         * qmake/configure.pro:
75         * qmake/mkspecs/features/default_pre.prf:
76         * qmake/mkspecs/features/features.prf:
77         * qmake/sync.profile: Removed.
78
79 2012-07-05  Jason Liu  <jason.liu@torchmobile.com.cn>
80
81         Add Jason Liu to committers.py.
82         https://bugs.webkit.org/show_bug.cgi?id=90654
83
84         Unreviewed.
85
86         * Scripts/webkitpy/common/config/committers.py:
87
88 2012-07-05  Benjamin Poulain  <bpoulain@apple.com>
89
90         Double release of resources if the load is canceled in a callback of ResourceLoader::didFinishLoading
91         https://bugs.webkit.org/show_bug.cgi?id=90431
92
93         Reviewed by Anders Carlsson.
94
95         Add a Mac API test.
96
97         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
98         * TestWebKitAPI/Tests/mac/CancelLoadFromResourceLoadDelegate.html: Added.
99         * TestWebKitAPI/Tests/mac/CancelLoadFromResourceLoadDelegate.mm: Added.
100         (-[CancelLoadFromResourceLoadDelegate webView:resource:didFinishLoadingFromDataSource:]):
101         (-[CancelLoadFromResourceLoadDelegateFrameLoadDelegate webView:didFinishLoadForFrame:]):
102         (TestWebKitAPI):
103         (TestWebKitAPI::TEST):
104
105 2012-07-05  Dave Tharp  <dtharp@codeaurora.org>
106
107         Adding myself as committer to committers.py
108
109         Unreviewed.
110
111         * Scripts/webkitpy/common/config/committers.py:
112
113 2012-07-05  Alexey Proskuryakov  <ap@apple.com>
114
115         [Mac][WK2] Enable HTTPS tests
116         https://bugs.webkit.org/show_bug.cgi?id=90600
117
118         Reviewed by Brady Eidson.
119
120         * DumpRenderTree/mac/DumpRenderTree.mm: (dumpRenderTree): Removed a misleading comment.
121         It explained why we were doing this here instead of a frame load delegate, but that's
122         just history, not an explanation of why we need this code.
123
124         * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm: (WTR::InjectedBundle::platformInitialize):
125         Ignore certificate errors for localhost and 127.0.0.1. Since this is only needed
126         for tests, it seems appropriate to have platform specific solutions in every WTR,
127         and not add a WebKit2 API.
128
129 2012-07-05  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
130
131         [Qt] Remove rpath-hacks in qt_webkit.pri
132
133         Having it in qt_webkit.pri meant that all client apps would inherit the rpath,
134         even if they were not using QtWebKit.
135
136         This is a regression for people who build client apps against a non-installed
137         QtWebKit, as the client app will be missing a rpath to the standalone WebKit
138         build dir. This can be solved by LD_LIBRARY_PATH or custom QMAKE_RPATHDIR
139         logic, until the real issue is fixed in Qt.
140
141         Reviewed by Csaba Osztrogonác.
142
143         * qmake/mkspecs/features/rpath.prf:
144         * qmake/qt_webkit.pri:
145
146 2012-07-05  Dongwoo Im  <dw.im@samsung.com>
147
148         [EFL] Enable the CUSTOM_SCHEME_HANDLER feature as default.
149         https://bugs.webkit.org/show_bug.cgi?id=88608
150
151         Reviewed by Hajime Morita.
152
153         * Scripts/webkitperl/FeatureList.pm: Enable the CUSTOM_SCHEME_HANDLER feature as default.
154
155 2012-07-05  Ryuan Choi  <ryuan.choi@samsung.com>
156
157         [Wk2][EFL] EFL needs a WebKitTestRunner
158         https://bugs.webkit.org/show_bug.cgi?id=87659
159
160         Reviewed by Kenneth Rohde Christiansen.
161
162         Add an implementation of WebKitTestRunner for Efl port.
163
164         * CMakeLists.txt: Includes WebKitTestRunner build script.
165         * Scripts/build-webkittestrunner: Added knowledge of WebKitTestRunner/Efl.
166         * Scripts/webkitpy/layout_tests/port/efl.py: Ditto.
167         (EflPort.setup_environ_for_server):
168         * WebKitTestRunner/CMakeLists.txt: Added.
169         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
170         * WebKitTestRunner/InjectedBundle/efl/ActivateFontsEfl.cpp: Added.
171         * WebKitTestRunner/InjectedBundle/efl/InjectedBundleEfl.cpp: Added.
172         * WebKitTestRunner/InjectedBundle/efl/LayoutTestControllerEfl.cpp: Added.
173         * WebKitTestRunner/PlatformEfl.cmake: Added.
174         * WebKitTestRunner/PlatformWebView.h:
175         * WebKitTestRunner/TestController.cpp:
176         * WebKitTestRunner/config.h: Added knowledge of WebKitTestRunner/Efl.
177         * WebKitTestRunner/efl/PlatformWebViewEfl.cpp: Added.
178         * WebKitTestRunner/efl/TestControllerEfl.cpp: Added.
179         * WebKitTestRunner/efl/main.cpp: Added.
180         (main):
181
182 2012-07-05  Christophe Dumez  <christophe.dumez@intel.com>
183
184         [WK2][EFL] Ewk_View needs API to load HTML data
185         https://bugs.webkit.org/show_bug.cgi?id=90540
186
187         Reviewed by Kenneth Rohde Christiansen.
188
189         Update EFL MiniBrowser to catch the "load,error" signal
190         on the view and display an error page.
191
192         * MiniBrowser/efl/main.c:
193         (on_error):
194         (browserCreate):
195
196 2012-07-05  Sergio Villar Senin  <svillar@igalia.com>
197
198         Hidden dirs are not copied when creating the built product archive
199         https://bugs.webkit.org/show_bug.cgi?id=90559
200
201         Reviewed by Carlos Garcia Campos.
202
203         Hidden dirs are not copied inside the built product archive
204         because the recursive copy command was ignoring them. We need those
205         hidden directories for the GTK WebKit2 testing bot to work fine.
206
207         * BuildSlaveSupport/built-product-archive:
208         (archiveBuiltProduct): replaced "*" by "." to include hidden
209         dirs/files.
210
211 2012-07-04  Yoshifumi Inoue  <yosin@chromium.org>
212
213         Unreviewed Chromium gardening, Disable always failed python tests on Windows.
214         https://bugs.webkit.org/show_bug.cgi?id=90587
215
216         PortTest.qt_linux and mac_lion are always failed on Chromium Windows at
217         creating child process.
218
219         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
220         (PortTest.disabled_test_qt_linux):
221         (PortTest.disabled_test_mac_lion):
222
223 2012-07-04  Yoshifumi Inoue  <yosin@chromium.org>
224
225         [Tools] webkit_unittest.py got assertion
226         https://bugs.webkit.org/show_bug.cgi?id=90579
227
228         Reviewed by Hajime Morita.
229
230         This patch updates expectation of assertion in test_skipped_directories_for_symbols().
231         r121874 added new (exclude) directory (inspector/styles/variables), however it didn't
232         update one of two assertions.
233
234         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
235
236 2012-07-04  Balazs Ankes  <bank@inf.u-szeged.hu>
237
238         webkit-patch should add reviewer if "Reviewed by NOBODY ..." is missing
239         https://bugs.webkit.org/show_bug.cgi?id=67935
240
241         Reviewed by Ryosuke Niwa.
242
243         * Scripts/webkitpy/common/checkout/changelog.py:
244         (ChangeLog.set_reviewer):
245         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
246         (test_set_reviewer):
247         (test_set_short_description_and_bug_url):
248
249 2012-07-04  Sheriff Bot  <webkit.review.bot@gmail.com>
250
251         Unreviewed, rolling out r121862.
252         http://trac.webkit.org/changeset/121862
253         https://bugs.webkit.org/show_bug.cgi?id=90569
254
255         It broke built-product-archive for GTK, Qt and EFL (Requested
256         by svillar on #webkit).
257
258         * BuildSlaveSupport/built-product-archive:
259         (archiveBuiltProduct):
260
261 2012-07-04  Sergio Villar Senin  <svillar@igalia.com>
262
263         Hidden dirs are not copied when creating the built product archive
264         https://bugs.webkit.org/show_bug.cgi?id=90559
265
266         Reviewed by Carlos Garcia Campos.
267
268         Hidden dirs are not copied inside the built product archive because
269         the recursive copy command was ignoring them. We need those hidden
270         directories for the GTK WebKit2 testing bot to work fine.
271
272         * BuildSlaveSupport/built-product-archive:
273         (archiveBuiltProduct): removed the trailing "*" from the copy
274         command.
275
276 2012-07-04  John Mellor  <johnme@chromium.org>
277
278         Text Autosizing: Add compile flag and runtime setting
279         https://bugs.webkit.org/show_bug.cgi?id=87394
280
281         This patch renames Font Boosting to Text Autosizing.
282
283         Reviewed by Adam Barth.
284
285         * Scripts/webkitperl/FeatureList.pm:
286         * qmake/mkspecs/features/features.pri:
287
288 2012-07-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
289
290         [Qt] Replace deprecated IN_PWD with PWD
291
292         Reviewed by Csaba Osztrogonác.
293
294         * qmake/mkspecs/features/default_pre.prf:
295
296 2012-07-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
297
298         [Qt] Don't try to build WTR twice
299
300         Reviewed by Csaba Osztrogonác.
301
302         * Tools.pro:
303
304 2012-07-04  Sheriff Bot  <webkit.review.bot@gmail.com>
305
306         Unreviewed, rolling out r121821.
307         http://trac.webkit.org/changeset/121821
308         https://bugs.webkit.org/show_bug.cgi?id=90551
309
310         This patch did not receive a high-quality review and has a
311         number of errors (Requested by abarth on #webkit).
312
313         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
314         (Build.results_url):
315         * Scripts/webkitpy/common/net/buildbot/chromiumbuildbot.py:
316         (ChromiumBuilder.accumulated_results_url):
317         * Scripts/webkitpy/layout_tests/port/builders.py:
318         * Scripts/webkitpy/tool/commands/rebaseline.py:
319         (AbstractParallelRebaselineCommand.__init__):
320         (Rebaseline):
321         (Rebaseline._builder_to_pull_from):
322         (Rebaseline._tests_to_update):
323         (Rebaseline.execute):
324         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
325         (test_rebaseline.mock_builder_to_pull_from):
326         (test_rebaseline):
327
328 2012-07-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
329
330         [Qt] Get rid of un-needed QT += declarative for Qt 5
331
332         The declarative module has been renamed to quick1 in Qt 5, and the
333         engine-only module for Qt 5 is named 'qml'. For many of the instances
334         we could just remove 'declarative', since the project file was only
335         used for Qt5/WebKit2 builds. In the other cases the module was wrapped
336         in a haveQt(4) scope.
337
338         Reviewed by Csaba Osztrogonác.
339
340         * MiniBrowser/qt/MiniBrowser.pro:
341         * WebKitTestRunner/InjectedBundle/Target.pri:
342         * WebKitTestRunner/Target.pri:
343
344 2012-07-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
345
346         [Qt] Remove a few un-needed load(features) after r121777
347
348         The features are computed by configure.pro and cached in .qmake.cache.
349
350         Reviewed by Jocelyn Turcotte..
351
352         * MiniBrowser/qt/raw/Target.pri:
353
354 2012-07-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
355
356         [Qt] Use haveQtModule() in project files instead of manual checks
357
358         Reviewed by Jocelyn Turcotte.
359
360         * qmake/mkspecs/features/default_pre.prf:
361         * qmake/mkspecs/features/features.prf:
362
363 2012-07-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
364
365         [Qt] Fix haveQtModule() check
366
367         Without {} brackets we were pulling out the value of the module.name variable.
368
369         Reviewed by Jocelyn Turcotte.
370
371         * qmake/mkspecs/features/functions.prf:
372
373 2012-07-04  Sergio Villar Senin  <svillar@igalia.com>
374
375         [WK2] [GTK] WK2 testing bot unable to find the InjectedBundle library
376         https://bugs.webkit.org/show_bug.cgi?id=90541
377
378         Reviewed by Carlos Garcia Campos.
379
380         We must add the contents of WebKitBuild/*/Libraries/ directory to
381         the file to be uploaded to b.w.o to successfully run WebKit2 tests
382         in the WebKit2 testing bot.
383
384         * BuildSlaveSupport/built-product-archive:
385         (archiveBuiltProduct): appended Libraries to the list of
386         neededDirectories.
387
388 2012-07-04  Sergio Villar Senin  <svillar@igalia.com>
389
390         [GTK] WebKit2 testing bot unable to find the build directory
391         https://bugs.webkit.org/show_bug.cgi?id=90523
392
393         Reviewed by Carlos Garcia Campos.
394
395         The presence of GNUmakefile was used to determine whether a given
396         directory was a valid build directory or not. That's not correct for
397         testing bots as we export only the minimum required stuff to run the
398         tests.
399
400         * gtk/common.py:
401         (get_build_path.is_valid_build_directory): added an extra check.:
402
403 2012-07-04  Joe Thomas  <joethomas@motorola.com>
404
405         Adding myself to committers list.
406
407         Unreviewed.
408
409         * Scripts/webkitpy/common/config/committers.py:
410
411 2012-07-03  Christophe Dumez  <christophe.dumez@intel.com>
412
413         [WK2][EFL] Ewk_View should report the load progress
414         https://bugs.webkit.org/show_bug.cgi?id=90457
415
416         Reviewed by Kenneth Rohde Christiansen.
417
418         Update EFL MiniBrowser so that it listens for the
419         "load,progress" on the Ewk_View and updates its
420         window title accordingly.
421
422         * MiniBrowser/efl/main.c:
423         (title_set):
424         (on_title_changed):
425         (on_progress):
426         (browserCreate):
427
428 2012-07-03  Dirk Pranke  <dpranke@chromium.org>
429
430         disable failing mock-chromium-mac python unit tests
431
432         Unreviewed, build fix.
433
434         Disable these tests until I can figure out why they're failing
435         on the bots.
436
437         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
438         (PortTest.disabled_test_chromium_mac_lion):
439         (PortTest.disabled_test_chromium_mac_lion_in_test_shell_mode):
440
441 2012-07-03  Ian Vollick  <vollick@chromium.org>
442
443         Add vollick to the list of committers
444         https://bugs.webkit.org/show_bug.cgi?id=90512
445
446         Reviewed by Adrienne Walker.
447
448         * Scripts/webkitpy/common/config/committers.py:
449
450 2012-07-03  Ojan Vafai  <ojan@chromium.org>
451
452         Improve webkit-patch rebaseline to work for more cases
453         https://bugs.webkit.org/show_bug.cgi?id=90504
454
455         Reviewed by Dirk Pranke.
456
457         -Makes it work for the build.chromium.org bots.
458         -Lets you rebaseline all builders instead of just one.
459         -Lets you pass in the builder or test to rebaseline.
460
461         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
462         (Build.results_url):
463         Make this work for build.chromium.org builders as well.
464
465         * Scripts/webkitpy/common/net/buildbot/chromiumbuildbot.py:
466         (ChromiumBuilder):
467         (ChromiumBuilder.latest_cached_build):
468         ChromiumBuilder doesn't support large parts of the Builder interface.
469         This provides the bare minimum for this patch to work. Eventually,
470         we should create a single interface that can be supported for both
471         sets of buildbots.
472
473         * Scripts/webkitpy/layout_tests/port/builders.py:
474         (builder_path_from_name):
475         (all_builder_names):
476         memoizing here is incorrect because the test override _exact_matches,
477         so these can return different values. In either case, I'm pretty sure these
478         are not remotely hot codepaths.
479
480         * Scripts/webkitpy/tool/commands/rebaseline.py:
481         (AbstractParallelRebaselineCommand.__init__):
482         (Rebaseline):
483         (Rebaseline.__init__):
484         (Rebaseline._builders_to_pull_from):
485         (Rebaseline._builder_with_name):
486         (Rebaseline._tests_to_update):
487         (Rebaseline.execute):
488         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
489         (test_rebaseline.mock_builders_to_pull_from):
490         (test_rebaseline):
491         (test_rebaseline_command_line_flags):
492         (test_rebaseline_multiple_builders):
493         (test_rebaseline_multiple_builders.mock_builders_to_pull_from):
494         (test_rebaseline_multiple_builders.mock_tests_to_update):
495
496 2012-07-03  Christophe Dumez  <christophe.dumez@intel.com>
497
498         [EFL] Enable CSS variables support at compile time
499         https://bugs.webkit.org/show_bug.cgi?id=90448
500
501         Reviewed by Kenneth Rohde Christiansen.
502
503         Turn on CSS_VARIABLES flag by default on EFL port.
504
505         * Scripts/webkitperl/FeatureList.pm:
506
507 2012-07-03  Dirk Pranke  <dpranke@chromium.org>
508
509         nrwt: clean up exception handling and make sure we log some more failures
510         https://bugs.webkit.org/show_bug.cgi?id=90503
511
512         Reviewed by Ojan Vafai.
513
514         There were several places where exceptions weren't getting
515         logged, most notably if you passed a bad value to --platform.
516         This change tests that and cleans things up a bit; more cleanup
517         will be possible when we rework the manager_worker_broker code.
518
519         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
520         (_BrokerConnection.raise_exception):
521         (_InlineWorkerConnection.raise_exception):
522         * Scripts/webkitpy/layout_tests/controllers/worker.py:
523         (Worker.run):
524         (Worker.kill_driver):
525         * Scripts/webkitpy/layout_tests/port/factory.py:
526         (PortFactory.get):
527         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
528         (run):
529         (main):
530         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
531         (MainTest.test_unsupported_platfrom):
532
533 2012-07-03  Dirk Pranke  <dpranke@chromium.org>
534
535         nrwt: fix mock port
536         https://bugs.webkit.org/show_bug.cgi?id=90500
537
538         Reviewed by Ojan Vafai.
539
540         The MockDRT code was never updated when we switched the chromium
541         ports to using "drt mode" by default. This change updates that
542         code, fixes a typo in the chromium port that went undetected
543         (default_test_timeout_ms -> default_timeout_ms), and adds tests
544         that actually exercise some of the mock ports. These tests are
545         useful in that they will exercise the port-specific code in an
546         end-to-end-manner, but they are a bit slow for some reason (>1s
547         each) that I need to look into.
548
549         * Scripts/webkitpy/layout_tests/port/chromium.py:
550         (ChromiumDriver.stop):
551         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
552         (ChromiumAndroidPort.default_timeout_ms):
553         * Scripts/webkitpy/layout_tests/port/chromium_android_unittest.py:
554         (TestChromiumAndroidPort.test_default_timeout_ms):
555         * Scripts/webkitpy/layout_tests/port/mock_drt.py:
556         (main):
557         (parse_options):
558         (MockTestShell):
559         (MockTestShell.output_for_test):
560         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
561         (MockDRTTest.assertTest):
562         (MockTestShellTest):
563         (MockTestShellTest.make_drt):
564         (MockTestShellTest.test_test_shell_parse_options):
565         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
566         (PortTest):
567         (PortTest.assert_mock_port_works):
568         (PortTest.test_chromium_mac_lion):
569         (PortTest.test_chromium_mac_lion_in_test_shell_mode):
570         (PortTest.test_qt_linux):
571         (PortTest.test_mac_lion):
572
573 2012-07-03  Dirk Pranke  <dpranke@chromium.org>
574
575         nrwt: make the worker class stand alone with a cleaner interface
576         https://bugs.webkit.org/show_bug.cgi?id=90409
577
578         Reviewed by Ojan Vafai.
579
580         Currently the Worker class derives from AbstractWorker, which is
581         kind of crufty and awkward; it would be better if we did not
582         rely on shared state.
583
584         This change changes that so that Worker derives from object, and
585         exposes the following interface:
586           __init__() - called in the manager process
587           safe_init() - called in the worker process to initialize
588             unpicklable state
589           handle() - a single routine to handle all messages
590           cleanup() - called so the worker can clean up
591
592         Also, all of the "administrative" messages that are handled by
593         the worker (notification of start/stop/etc.) move into
594         manager_worker_broker - this reduces worker.py to just handling
595         the mechanics of actually running each test.
596
597         For the moment, we do this by creating Yet Another wrapper/proxy
598         class in manager_worker_broker, but this will get simpler
599         shortly when the rest of m_w_b is cleaned up.
600
601         With this change worker is now in its new form but there will be
602         a follow-on change that cleans up some names and other minor
603         things.
604
605         This change is again mostly just moving things around and should
606         be covered by the (updated) existing tests.
607
608         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
609         (get):
610         (AbstractWorker.__init__):
611         (AbstractWorker.run):
612         (AbstractWorker):
613         (AbstractWorker.handle_stop):
614         (AbstractWorker.handle_test_list):
615         (AbstractWorker.yield_to_broker):
616         (AbstractWorker.post_message):
617         (_WorkerConnection.__init__):
618         (_Process.run):
619         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py:
620         (_TestWorker):
621         (_TestWorker.__init__):
622         (_TestWorker.name):
623         (_TestWorker.cleanup):
624         (_TestWorker.handle):
625         (_TestWorker.safe_init):
626         (_TestWorker.stop):
627         (_TestsMixin.handle_finished_test):
628         (_TestsMixin.setUp):
629         (_TestsMixin.test_cancel):
630         (_TestsMixin.test_done):
631         * Scripts/webkitpy/layout_tests/controllers/worker.py:
632         (Worker):
633         (Worker.__init__):
634         (Worker.safe_init):
635         (Worker.handle):
636
637 2012-07-03  Dirk Pranke  <dpranke@chromium.org>
638
639         nrwt: moving child process logging code into manager_worker_broker
640         https://bugs.webkit.org/show_bug.cgi?id=90408
641
642         Reviewed by Ojan Vafai.
643
644         Users of manager_worker_broker should not have to be aware of
645         whether they're in the same process or different processes and
646         configure logging themselves; mwb should hide this complexity.
647         We can't quite do this completely/correctly yet, since the
648         manager expects to get a list of messages to log, but this
649         change fixes the worker side of it, at least.
650
651         This is just moving code around, there is no new functionality
652         and this should be covered by existing tests.
653
654         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
655         (AbstractWorker.__init__):
656         (_WorkerConnection.__init__):
657         (_WorkerConnection.post_message):
658         (_WorkerConnection):
659         (_WorkerConnection.set_up_logging):
660         (_WorkerConnection.clean_up_logging):
661         (_InlineWorkerConnection.run):
662         (_MultiProcessWorkerConnection.run):
663         (_WorkerLogHandler):
664         (_WorkerLogHandler.__init__):
665         (_WorkerLogHandler.emit):
666         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py:
667         (_TestWorker.run):
668         (_TestsMixin.handle_done):
669         * Scripts/webkitpy/layout_tests/controllers/worker.py:
670         (Worker.__init__):
671         (Worker.run):
672         (Worker._run_test):
673         (Worker.cleanup):
674         (Worker.run_single_test):
675
676 2012-07-03  Tony Chang  <tony@chromium.org>
677
678         [chromium] Don't archive build files generated by VS2010
679         https://bugs.webkit.org/show_bug.cgi?id=90493
680
681         Reviewed by Ojan Vafai.
682
683         In particular, the pch (pre compiled header) files are gigantic, making
684         release.zip almost a 1gb in size.
685
686         * BuildSlaveSupport/built-product-archive:
687         (archiveBuiltProduct):
688
689 2012-07-03  Tony Chang  <tony@chromium.org>
690
691         [chromium] Unreviewed, update .gitignore to handle VS2010 files.
692
693         * .gitignore:
694         * DumpRenderTree/DumpRenderTree.gyp/.gitignore:
695         * TestWebKitAPI/TestWebKitAPI.gyp/.gitignore:
696
697 2012-07-03  Ojan Vafai  <ojan@chromium.org>
698
699         Have webkit-patch rebaseline use rebaseline-test-internal
700         https://bugs.webkit.org/show_bug.cgi?id=90491
701
702         Reviewed by Dirk Pranke.
703
704         This lets it handle new files, reduces duplicate code and lays the
705         groundwork for making rebaseline have a richer interface.
706
707         * Scripts/webkitpy/common/net/buildbot/buildbot_mock.py:
708         (MockBuild):
709         (MockBuild.__init__):
710         (MockBuilder.build):
711         * Scripts/webkitpy/tool/commands/rebaseline.py:
712         (AbstractParallelRebaselineCommand._files_to_add):
713         (Rebaseline):
714         (Rebaseline._builder_to_pull_from):
715         (Rebaseline._tests_to_update):
716         (Rebaseline.execute):
717         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
718         (test_overrides_are_included_correctly):
719         (test_rebaseline):
720         (test_rebaseline.mock_builder_to_pull_from):
721         (test_rebaseline.mock_tests_to_update):
722
723 2012-07-03  Ojan Vafai  <ojan@chromium.org>
724
725         Rename rebaseline-test to rebaseline-test-internal
726         https://bugs.webkit.org/show_bug.cgi?id=90485
727
728         Reviewed by Adam Barth.
729
730         It's now only used by other rebaseline commands. It's still useful to leave it
731         as it's own command to aid in debugging when something goes wrong. In a followup
732         patch, I'll make webkit-patch rebaseline cover any use-cases that rebaseline-test
733         might have covered.
734
735         We no longer need the --print-scm-changes option since the only caller always passes
736         that option in.
737
738         Also, make all the arguments command-line flags instead. Simplifies the code
739         a bit in my opinion.
740
741         * Scripts/webkitpy/tool/commands/rebaseline.py:
742         (RebaselineTest):
743         (RebaselineTest.__init__):
744         (RebaselineTest.execute):
745         (AbstractParallelRebaselineCommand._rebaseline_commands):
746         (RebaselineExpectations.execute):
747         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
748         (test_rebaseline_all):
749
750 2012-07-03  Balazs Kelemen  <kbalazs@webkit.org>
751
752         [Qt][WTR] Get rid of using DumpRenderTreeSupportQt
753         https://bugs.webkit.org/show_bug.cgi?id=90262
754
755         Unreviewed build fix.
756
757         Reverting r121550 manually because it caused build
758         break on ARM and Mac. It turned out that the Qt build
759         system is too broken for this change now, first we have
760         to address two issues:
761             * symbol hiding was forbidden accidentally on x86 Linux
762             * Qt lacks a separate library for test support
763
764         * Tools.pro:
765         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
766         (WTR::InjectedBundlePage::resetAfterTest):
767         (WTR::InjectedBundlePage::didClearWindowForFrame):
768         * WebKitTestRunner/InjectedBundle/qt/ActivateFontsQt.cpp:
769
770 2012-07-03  Tony Chang  <tony@chromium.org>
771
772         [chromium] Handle trailing backslash on %VSINSTALLDIR%
773         https://bugs.webkit.org/show_bug.cgi?id=90410
774
775         Reviewed by Ojan Vafai.
776
777         * Scripts/webkitdirs.pm:
778         (buildChromiumVisualStudioProject): Handle VSINSTALLDIR ending in a backslash, the backslash escapes the closing double quote.
779
780 2012-07-02  Ojan Vafai  <ojan@chromium.org>
781
782         Make the skia_test_expectations.txt file optional.
783         https://bugs.webkit.org/show_bug.cgi?id=90400
784
785         Reviewed by Dirk Pranke.
786
787         It used to be optional. This regressed at some point. It's important that it be
788         optional so that webkit-patch commands work in a pure-webkit checkout for chromium bots.
789         Specifically, this was breaking webkit-patch rebaseline-test when it would go to update
790         TestExpectations.
791
792         * Scripts/webkitpy/layout_tests/port/chromium.py:
793         (ChromiumPort.expectations_files):
794         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
795         (ChromiumDriverTest.test_expectations_dict):
796
797 2012-07-03  Raphael Kubo da Costa  <rakuco@webkit.org>
798
799         [jhbuild][EFL] Bump libffi dependency.
800         https://bugs.webkit.org/show_bug.cgi?id=85373
801
802         Reviewed by Gustavo Noronha Silva.
803
804         Update libffi to 3.0.11, which fixes the build on some platforms
805         where 3.0.10 had problems.
806
807         * efl/jhbuild.modules:
808
809 2012-07-03  George Staikos  <staikos@webkit.org>
810
811         [BlackBerry] Enable microdata support for BlackBerry.
812         https://bugs.webkit.org/show_bug.cgi?id=90429
813
814         Reviewed by Rob Buis.
815
816         * Scripts/webkitperl/FeatureList.pm:
817
818 2012-07-03  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
819
820         [Qt] Make use of .qmake.cache for caching features
821
822         Instead of loading() features from the files that need them (and re-running
823         a bunch of checks), we now run feature detection as part of configure.pro,
824         and have build-webkit write the computed feature-defines and CONFIG to
825         .qmake.cache, which is then loaded by qmake _before_ even defaults_pre
826         when building WebKit.pro.
827
828         At some point we'll be able to selectivly prevent running of config tests
829         in configure.pro, which means we don't need a separate code-path for
830         the build-webkit --help case.
831
832         We should also move the code in build-webkit that now uses .webkit.config
833         to detect clean builds, to use .qmake.cache, since we now store the same
834         thing there.
835
836         Original patch by Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
837
838         Reviewed by Tor Arne Vestbø.
839
840         * DumpRenderTree/qt/DumpRenderTree.pro:
841         * QtTestBrowser/QtTestBrowser.pro:
842         * Scripts/webkitdirs.pm:
843         (qtFeatureDefaults):
844         (buildQMakeProjects):
845         * Tools.pro:
846         * WebKitTestRunner/InjectedBundle/DerivedSources.pri:
847         * WebKitTestRunner/InjectedBundle/Target.pri:
848         * WebKitTestRunner/Target.pri:
849         * qmake/.qmake.conf: Added.
850         * qmake/configure.pro:
851         * qmake/mkspecs/features/default_post.prf:
852         * qmake/mkspecs/features/default_pre.prf:
853         * qmake/mkspecs/features/features.prf:
854
855 2012-07-03  Nico Weber  <thakis@chromium.org>
856
857         -Wunused-private-field cleanup, linux edition
858         https://bugs.webkit.org/show_bug.cgi?id=90463
859
860         Reviewed by Stephen White.
861
862         * DumpRenderTree/chromium/DRTDevToolsAgent.h:
863         * DumpRenderTree/chromium/GamepadController.h:
864         (GamepadController):
865         * DumpRenderTree/chromium/NotificationPresenter.h:
866         (NotificationPresenter::NotificationPresenter):
867         (NotificationPresenter):
868
869 2012-07-03  Sheriff Bot  <webkit.review.bot@gmail.com>
870
871         Unreviewed, rolling out r121766.
872         http://trac.webkit.org/changeset/121766
873         https://bugs.webkit.org/show_bug.cgi?id=90465
874
875         It caused flakey build errors on the bots (Requested by Ossy
876         on #webkit).
877
878         * DumpRenderTree/qt/DumpRenderTree.pro:
879         * QtTestBrowser/QtTestBrowser.pro:
880         * Scripts/webkitdirs.pm:
881         (qtFeatureDefaults):
882         (buildQMakeProjects):
883         * Tools.pro:
884         * WebKitTestRunner/InjectedBundle/DerivedSources.pri:
885         * WebKitTestRunner/InjectedBundle/Target.pri:
886         * WebKitTestRunner/Target.pri:
887         * qmake/.qmake.conf: Removed.
888         * qmake/configure.pro:
889         * qmake/mkspecs/features/default_post.prf:
890         * qmake/mkspecs/features/default_pre.prf:
891         * qmake/mkspecs/features/features.prf:
892
893 2012-07-03  George Staikos  <staikos@webkit.org>
894
895         [BlackBerry] Enable Custom Scheme Handlers for BlackBerry.
896         https://bugs.webkit.org/show_bug.cgi?id=90422
897
898         Reviewed by Rob Buis.
899
900         * Scripts/webkitperl/FeatureList.pm:
901
902 2012-07-03  George Staikos  <staikos@webkit.org>
903
904         [BlackBerry] Enable RegisterProtocolHandler for BlackBerry.
905         https://bugs.webkit.org/show_bug.cgi?id=90422
906
907         Reviewed by Rob Buis.
908
909         * Scripts/webkitperl/FeatureList.pm:
910
911 2012-07-03  Thiago Marcos P. Santos  <thiago.santos@intel.com>
912
913         [EFL] Add run API tests as a step on the build bots
914         https://bugs.webkit.org/show_bug.cgi?id=87252
915
916         Reviewed by Daniel Bates.
917
918         Run EFL API test suite on the build bots.
919
920         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
921         (RunEflAPITests):
922         (TestFactory.__init__):
923         (BuildAndTestFactory.__init__):
924         * Scripts/run-efl-tests: Added.
925
926 2012-07-03  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
927
928         [Qt] Make use of .qmake.cache for caching features
929
930         Instead of loading() features from the files that need them (and re-running
931         a bunch of checks), we now run feature detection as part of configure.pro,
932         and have build-webkit write the computed feature-defines and CONFIG to
933         .qmake.cache, which is then loaded by qmake _before_ even defaults_pre
934         when building WebKit.pro.
935
936         At some point we'll be able to selectivly prevent running of config tests
937         in configure.pro, which means we don't need a separate code-path for
938         the build-webkit --help case.
939
940         We should also move the code in build-webkit that now uses .webkit.config
941         to detect clean builds, to use .qmake.cache, since we now store the same
942         thing there.
943
944         Original patch by Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
945
946         Reviewed by Tor Arne Vestbø.
947
948         * DumpRenderTree/qt/DumpRenderTree.pro:
949         * QtTestBrowser/QtTestBrowser.pro:
950         * Scripts/webkitdirs.pm:
951         (qtFeatureDefaults):
952         (buildQMakeProjects):
953         * Tools.pro:
954         * WebKitTestRunner/InjectedBundle/DerivedSources.pri:
955         * WebKitTestRunner/InjectedBundle/Target.pri:
956         * WebKitTestRunner/Target.pri:
957         * qmake/.qmake.conf: Added.
958         * qmake/configure.pro:
959         * qmake/mkspecs/features/default_post.prf:
960         * qmake/mkspecs/features/default_pre.prf:
961         * qmake/mkspecs/features/features.prf:
962
963 2012-07-03  Jocelyn Turcotte  <jocelyn.turcotte@nokia.com>  Joel Dillon <joel.dillon@codethink.co.uk>
964
965         [Qt][Win] Fix broken QtWebKit5.lib linking
966         https://bugs.webkit.org/show_bug.cgi?id=88321
967
968         Reviewed by Kenneth Rohde Christiansen.
969
970         On windows the import/export definition is part of the symbol's signature.
971         Define STATICALLY_LINKED_WITH_$$library for each dependend module
972         that is being linked statically to make sure that they can be linked together
973         later on.
974
975         Also do not compile Assertions.cpp in DumpRenderTree anymore since all the
976         used symbols are exported and it would cause a duplicate symbols error.
977
978         * DumpRenderTree/qt/DumpRenderTree.pro:
979         * qmake/mkspecs/features/default_post.prf:
980
981 2012-07-03  Szilard Ledan  <szledan@inf.u-szeged.hu>
982
983         master.cfg cleanup, pass BuildStep instances instead of BuildStep subclasses
984         https://bugs.webkit.org/show_bug.cgi?id=89001
985
986         Reviewed by Csaba Osztrogonác.
987
988         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
989         (UploadBuiltProduct.__init__):
990         (UploadTestResults.__init__):
991         (ExtractTestResults.__init__):
992         (Factory.__init__):
993         (BuildFactory.__init__):
994         (TestFactory.__init__):
995         (BuildAndTestFactory.__init__):
996         (BuildAndPerfTestFactory.__init__):
997         (BuildAndPerfTestWebKit2Factory.__init__):
998         (DownloadAndPerfTestFactory.__init__):
999
1000 2012-07-03  Ádám Kallai  <kadam@inf.u-szeged.hu>
1001
1002         Update unit test after r121724.
1003         https://bugs.webkit.org/show_bug.cgi?id=90444
1004
1005         Reviewed by Csaba Osztrogonác.
1006
1007         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py:
1008         (GardeningServerTest.test_rebaseline_new_port):
1009
1010 2012-07-03  Sheriff Bot  <webkit.review.bot@gmail.com>
1011
1012         Unreviewed, rolling out r121749.
1013         http://trac.webkit.org/changeset/121749
1014         https://bugs.webkit.org/show_bug.cgi?id=90441
1015
1016         Tests crash on the paralel NRWT Qt bot (Requested by Ossy on
1017         #webkit).
1018
1019         * Scripts/webkitpy/layout_tests/port/qt.py:
1020         (QtPort._build_driver):
1021         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py:
1022         (XvfbDriver._start):
1023         (XvfbDriver._start.x_filter):
1024         (XvfbDriver.stop):
1025
1026 2012-07-03  Kristóf Kosztyó  <kkristof@inf.u-szeged.hu>
1027
1028         [NRWT] XvfbDriver should choose the next free display
1029         https://bugs.webkit.org/show_bug.cgi?id=88414
1030
1031         Reviewed by Dirk Pranke.
1032
1033         * Scripts/webkitpy/layout_tests/port/qt.py:
1034         (QtPort._driver_class):
1035         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py:
1036         (XvfbDriver.__init__):
1037         (XvfbDriver._start):
1038         (XvfbDriver._start.next_free_id):
1039         (XvfbDriver.stop):
1040
1041 2012-07-03  Christophe Dumez  <christophe.dumez@intel.com>
1042
1043         [EFL] Enable MICRODATA support
1044         https://bugs.webkit.org/show_bug.cgi?id=90377
1045
1046         Reviewed by Kenneth Rohde Christiansen.
1047
1048         Turn on MICRODATA flag for EFL port by default.
1049
1050         * Scripts/webkitperl/FeatureList.pm:
1051
1052 2012-07-02  George Staikos  <staikos@webkit.org>
1053
1054         [BlackBerry] Enable scoped style for BlackBerry.
1055         https://bugs.webkit.org/show_bug.cgi?id=90418
1056
1057         Reviewed by Rob Buis.
1058
1059         * Scripts/webkitperl/FeatureList.pm:
1060
1061 2012-07-02  Xiaobo Wang  <xbwang@torchmobile.com.cn>
1062
1063         [BlackBerry] Use PUBLIC_BUILD to enable/disable DRT
1064         https://bugs.webkit.org/show_bug.cgi?id=90271
1065
1066         Reviewed by George Staikos.
1067
1068         RIM PR #154707
1069
1070         Currently DRT code will be compiled only if ENABLE_DRT is set, and it's not
1071         defined by default.
1072         We should enable DRT by default unless PUBLIC_BUILD is set. In this way we don't
1073         need to rebuild webkit before running DRT.
1074
1075         * Scripts/webkitdirs.pm:
1076         (blackberryCMakeArguments):
1077
1078 2012-07-02  Ojan Vafai  <ojan@chromium.org>
1079
1080         webkit-patch rebaseline-expectations should share code with rebaseline-all
1081         https://bugs.webkit.org/show_bug.cgi?id=90413
1082
1083         Reviewed by Dirk Pranke.
1084
1085         Make them share code. In addition to reducing code duplication this makes
1086         rebaseline-expectations considerably faster by rebaselining in parallel.
1087
1088         * Scripts/webkitpy/tool/commands/rebaseline.py:
1089         (AbstractParallelRebaselineCommand):
1090         (AbstractParallelRebaselineCommand._run_webkit_patch):
1091         (AbstractParallelRebaselineCommand._rebaseline):
1092         (RebaselineJson):
1093         (RebaselineJson.execute):
1094         (RebaselineExpectations):
1095         (RebaselineExpectations._update_expectations_file):
1096         (RebaselineExpectations._tests_to_rebaseline):
1097         (RebaselineExpectations._add_tests_to_rebaseline_for_port):
1098         (RebaselineExpectations.execute):
1099         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
1100         (test_rebaseline_all):
1101         (test_rebaseline_expectations.run_in_parallel):
1102         (test_rebaseline_expectations):
1103
1104 2012-07-02  Xiaobo Wang  <xbwang@torchmobile.com.cn>
1105
1106         [BlackBerry] Update DumpRenderTree to have it work interactively in parallel
1107         https://bugs.webkit.org/show_bug.cgi?id=88326
1108
1109         Reviewed by Rob Buis.
1110
1111         1. Test name is sent to torch-launcher via PPS message(from host machine).
1112            So we get test list by monitoring and receiving PPS message instead of
1113            reading file index.drt.
1114         2. Torch-launcher create a <test file>.done file when it finished a test.
1115         3. We don't need to search for Ref-tests in DumpRenderTree.cpp any more. NRWT
1116            will get them for us.
1117
1118         * DumpRenderTree/blackberry/DumpRenderTree.cpp:
1119         (BlackBerry::WebKit::DumpRenderTree::DumpRenderTree):
1120         (BlackBerry::WebKit::DumpRenderTree::doneDrt):
1121         (BlackBerry::WebKit::DumpRenderTree::runRemainingTests):
1122         (BlackBerry::WebKit::DumpRenderTree::ensurePPS):
1123         (WebKit):
1124         (BlackBerry::WebKit::DumpRenderTree::handlePPSData):
1125         (BlackBerry::WebKit::DumpRenderTree::waitForTest):
1126         (BlackBerry::WebKit::DumpRenderTree::runTests):
1127         (BlackBerry::WebKit::DumpRenderTree::dump):
1128         * DumpRenderTree/blackberry/DumpRenderTreeBlackBerry.h:
1129         (DumpRenderTree):
1130
1131 2012-07-02  Adam Barth  <abarth@webkit.org>
1132
1133         Remove flashplugin-installer from the EWS image because it causes some tests to crash
1134         https://bugs.webkit.org/show_bug.cgi?id=90403
1135
1136         Reviewed by Tony Chang.
1137
1138         * EWSTools/cold-boot.sh:
1139
1140 2012-07-02  Leandro Gracia Gil  <leandrogracia@chromium.org>
1141
1142         [Chromium] Implement a Layout Test for editing/SurroundingText
1143         https://bugs.webkit.org/show_bug.cgi?id=82461
1144
1145         Reviewed by Ryosuke Niwa.
1146
1147         Add a new method to the layout test controller in order to retrieve the
1148         text surrounding a provided element.
1149
1150         * DumpRenderTree/chromium/LayoutTestController.cpp:
1151         (LayoutTestController::LayoutTestController):
1152         (LayoutTestController::textSurroundingElement):
1153         * DumpRenderTree/chromium/LayoutTestController.h:
1154         (LayoutTestController):
1155
1156 2012-07-02  Ojan Vafai  <ojan@chromium.org>
1157
1158         Delete unused rebaseline method in gardeningserver.py
1159         https://bugs.webkit.org/show_bug.cgi?id=90396
1160
1161         Reviewed by Eric Seidel.
1162
1163         As best I can tell, the only usage was in this unittest.
1164
1165         * Scripts/webkitpy/tool/servers/gardeningserver.py:
1166         (GardeningHTTPRequestHandler.updateexpectations):
1167         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py:
1168         (GardeningServerTest.test_rebaseline_new_port):
1169
1170 2012-07-02  Csaba Osztrogonác  <ossy@webkit.org>
1171
1172         REGRESSION(r121497): It switched off and broke many unittests
1173         https://bugs.webkit.org/show_bug.cgi?id=90371
1174
1175         Reviewed by Adam Barth.
1176
1177         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
1178         (ChromiumMacPortTest): Use snowleopard as os_version instead of leopard, because it isn't supported anymore.
1179         * Scripts/webkitpy/layout_tests/port/chromium_port_testcase.py:
1180         (ChromiumPortTestCase.test_all_test_configurations): Remove leopard testcases, because it isn't supported anymore.
1181         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
1182         (PortTestCase): Inherit class PortTestCase from unittest.TestCase instead of object.
1183
1184 2012-07-02  Ojan Vafai  <ojan@chromium.org>
1185
1186         Move rebaseline-all command from the gardening-server down into webkit-patch
1187         https://bugs.webkit.org/show_bug.cgi?id=90395
1188
1189         Reviewed by Adam Barth.
1190
1191         This is just moving code. It it in preparation for making rebaseline-expectations
1192         use the same code in order to get the parallelism benefits and reduces the amount
1193         of code we have for doing rebaselines.
1194
1195         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
1196         (CheckoutTest.test_apply_patch):
1197         Updated due to the change to executive_mock.
1198         * Scripts/webkitpy/common/system/executive_mock.py:
1199         (MockExecutive.run_command):
1200         Update to print out the input passed to stdin.
1201         * Scripts/webkitpy/tool/commands/download_unittest.py:
1202         Updated due to executive_mock change.
1203         * Scripts/webkitpy/tool/commands/rebaseline.py:
1204         (RebaselineAll):
1205         (RebaselineAll._run_webkit_patch):
1206         (RebaselineAll._builders_to_fetch_from):
1207         (RebaselineAll._rebaseline_commands):
1208         (RebaselineAll._files_to_add):
1209         (RebaselineAll._optimize_baselines):
1210         (RebaselineAll._rebaseline):
1211         (RebaselineAll.execute):
1212         All this code is just copy-pasted except for mechanical changes
1213         (e.g. self.server.tool --> self._tool) and the reading in of the 
1214         JSON from stdin instead of the post body.
1215         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
1216         (test_rebaseline_all):
1217         Copied the test-case out of gardeningserver_unittest.py.
1218         * Scripts/webkitpy/tool/servers/gardeningserver.py:
1219         (GardeningHTTPRequestHandler):
1220         (GardeningHTTPRequestHandler.rebaseline):
1221         (GardeningHTTPRequestHandler.rebaselineall):
1222         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py:
1223         (GardeningServerTest.test_rebaselineall):
1224         (GardeningServerTest.test_rebaselineall.run_command):
1225
1226 2012-07-02  Ojan Vafai  <ojan@chromium.org>
1227
1228         Remove Leopard support from the flakiness dashboard
1229         https://bugs.webkit.org/show_bug.cgi?id=90390
1230
1231         Reviewed by Adam Barth.
1232
1233         * TestResultServer/static-dashboards/flakiness_dashboard.js:
1234         (nonChromiumPlatform):
1235         (chromiumPlatform):
1236         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
1237         (test):
1238         * TestResultServer/static-dashboards/run-embedded-unittests.html:
1239         * TestResultServer/static-dashboards/run-unittests.html:
1240
1241 2012-07-02  Ojan Vafai  <ojan@chromium.org>
1242
1243         Fix posting from garden-o-matic. This broke in moving away from jquery's ajax method
1244         in http://trac.webkit.org/changeset/121617.
1245         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/net.js:
1246
1247 2012-07-02  Carlos Garcia Campos  <cgarcia@igalia.com>
1248
1249         [GTK] Read fonts path when running layout tests from alternative fonts dir when main dir doesn't exist
1250         https://bugs.webkit.org/show_bug.cgi?id=89437
1251
1252         Reviewed by Martin Robinson.
1253
1254         If main fonts directory doesn't exist, try with an alternative
1255         fonts directory at build directory.
1256
1257         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1258         (getOutputDir):
1259         (getFontsPath):
1260         (initializeFonts):
1261         * WebKitTestRunner/InjectedBundle/gtk/ActivateFontsGtk.cpp:
1262         (WTR::getOutputDir):
1263         (WTR):
1264         (WTR::getFontsPath):
1265         (WTR::inititializeFontConfigSetting):
1266
1267 2012-07-02  Carlos Garcia Campos  <cgarcia@igalia.com>
1268
1269         [GTK] Don't run the tests with jhbuild wrapper if it's already running under jhbuild
1270         https://bugs.webkit.org/show_bug.cgi?id=89435
1271
1272         Reviewed by Martin Robinson.
1273
1274         * Scripts/new-run-webkit-tests: Don't run the tests with the
1275         jhbuild wrapper if there isn't a Dependencies directory inside the
1276         build directory.
1277
1278 2012-06-29  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
1279
1280         [Qt] Simplify detection of non-installed module file
1281
1282         Has the additional advantage that we do not rely on additional information.
1283
1284         Reviewed by Tor Arne Vestbø.
1285
1286         * qmake/qt_webkit.pri:
1287
1288 2012-07-01  Christophe Dumez  <christophe.dumez@intel.com>
1289
1290         [EFL] Add Gamepad support
1291         https://bugs.webkit.org/show_bug.cgi?id=90170
1292
1293         Reviewed by Kenneth Rohde Christiansen.
1294
1295         * Scripts/webkitperl/FeatureList.pm: Turn on GAMEPAD flag
1296         by default for EFL port.
1297         * efl/jhbuild.modules: Bump dependency versions of EFL libs
1298         since the latest Eeze is required for gamepad support.
1299
1300 2012-07-01  Adam Barth  <abarth@webkit.org>
1301
1302         Unreviewed.
1303
1304         Using the default start cylinder for fdisk causes a warning about the
1305         partition not starting on physical sector boundary.  The Ubuntu forums
1306         recommend using a number that's divisible by 8, which is what we do in
1307         this patch.
1308
1309         * EWSTools/cold-boot.sh:
1310
1311 2012-07-01  Adam Barth  <abarth@webkit.org>
1312
1313         Unreviewed.
1314
1315         It turns out we need to use Ubuntu 10.04 to get the right image results
1316         for chromium-linux. This patch updates our scripts to be compatible
1317         with Ubuntu 10.04.
1318
1319         * EWSTools/cold-boot.sh:
1320             - fdisk doesn't have p and 1 as default commands in 10.04.
1321         * EWSTools/start-queue.sh:
1322             - git doesn't understanding the -B argument in 10.04. We've been
1323               using this change locally on the EC2 bots for a while.
1324
1325 2012-07-01  Adam Barth  <abarth@webkit.org>
1326
1327         Unreviewed.
1328
1329         * EWSTools/boot.sh:
1330             - We need to start the screen in detached mode so that we can run
1331               it remotely via ssh.
1332
1333 2012-07-01  Adam Barth  <abarth@webkit.org>
1334
1335         Unreviewed.
1336
1337         * EWSTools/cold-boot.sh:
1338             - Turns out we need to sudo this command in order for it to
1339               actually dimiss the EULA screen. :)
1340
1341 2012-06-29  Adam Barth  <abarth@webkit.org>
1342
1343         Turns out we need zip too.
1344
1345         * EWSTools/cold-boot.sh:
1346
1347 2012-06-29  Adam Barth  <abarth@webkit.org>
1348
1349         Add a cold-boot.sh script for the EWS
1350         https://bugs.webkit.org/show_bug.cgi?id=90330
1351
1352         Unreviewed.
1353
1354         * EWSTools/cold-boot.sh: Added.
1355             - This script can take us from a cold GCE instance to a running EWS
1356               bot in one fell swoop.
1357         * EWSTools/start-queue.sh:
1358             - The if-block at the top of this script was causing trouble. I
1359               removed it from the bots a while ago. Now that we're using SVN to
1360               cold-boot the EWS bots, we need to make this change in the repo.
1361
1362 2012-06-29  Luiz Agostini  <luiz.agostini@nokia.com>
1363
1364         [Qt][WK2] Private non-QtQuick API
1365         https://bugs.webkit.org/show_bug.cgi?id=84532
1366
1367         MiniBrowserRaw is an usage example for the QRawWebView API.
1368         It is only an example and is not fully implemented.
1369
1370         Reviewed by Noam Rosenthal.
1371
1372         * MiniBrowser/qt/raw/DerivedSources.pri: Added.
1373         * MiniBrowser/qt/raw/MiniBrowserRaw.pro: Added.
1374         * MiniBrowser/qt/raw/Target.pri: Added.
1375         * MiniBrowser/qt/raw/View.cpp: Added.
1376         * MiniBrowser/qt/raw/View.h: Added.
1377         * Scripts/webkitpy/style/checker.py:
1378         * Tools.pro:
1379
1380 2012-06-29  Ojan Vafai  <ojan@chromium.org>
1381
1382         Remove leopard bots from garden-o-matic
1383         https://bugs.webkit.org/show_bug.cgi?id=90328
1384
1385         Reviewed by Adam Barth.
1386
1387         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/config.js:
1388         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/failures_unittests.js:
1389         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/notifications_unittests.js:
1390
1391 2012-06-29  Ojan Vafai  <ojan@chromium.org>
1392
1393         garden-o-matic broken: TypeError: 'undefined' is not an object (evaluating 'buildLocations[currentIndex].url')
1394         https://bugs.webkit.org/show_bug.cgi?id=90243
1395
1396         Reviewed by Dirk Pranke.
1397
1398         jQuery was trying to be too smart and parsing the jsonp as json because of it's content-type.
1399         Excise jQuery and just use XHR directly since it's easier to maintain something where we control it
1400         all.
1401
1402         Not really sure how to unittest this. I tested it all manually of course.
1403
1404         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/net.js:
1405         Made net.ajax a drop-in replacement for the features of $.ajax that we were using.
1406         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui_unittests.js:
1407         Not related to this patch, but figured I'd update the failing test while I was at it.
1408
1409 2012-06-29  Yaron Friedman  <yfriedman@chromium.org>
1410
1411         Fix layout test runner for Android after https://bugs.webkit.org/show_bug.cgi?id=88134
1412         https://bugs.webkit.org/show_bug.cgi?id=90309
1413
1414         Reviewed by Adam Barth.
1415
1416         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
1417         (ChromiumAndroidPort.start_http_server):
1418
1419 2012-06-29  James Weatherall  <wez@chromium.org>
1420
1421         NPObjectWrapper may not address all window script object lifetime issues
1422         https://bugs.webkit.org/show_bug.cgi?id=85679
1423
1424         TestNetscapePlugin now has a leak-window-scriptable-object test which takes a reference to the window script object, and a second reference to it via the "self" property, and does not release those references. This is used to simulate a leaky plugin in layout tests of the NPAPI scripting interface glue code.
1425
1426         Reviewed by Nate Chapin.
1427
1428         * DumpRenderTree/DumpRenderTree.gypi:
1429         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
1430         (PluginTest::NPN_GetProperty):
1431         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
1432         (PluginTest):
1433         * DumpRenderTree/TestNetscapePlugIn/Tests/LeakWindowScriptableObject.cpp: Added.
1434         (LeakWindowScriptableObject):
1435         (LeakWindowScriptableObject::LeakWindowScriptableObject):
1436         (LeakWindowScriptableObject::NPP_New):
1437
1438 2012-06-29  Dirk Pranke  <dpranke@chromium.org>
1439
1440         webkitpy: add comment about how determine_full_port_name() works for apple ports, fix -wk2 bug
1441         https://bugs.webkit.org/show_bug.cgi?id=90314
1442
1443         Reviewed by Ojan Vafai.
1444
1445         Add comments and fix a bug in how we would handle the mac-wk2
1446         and win-wk2 port names after confusion around in bug 90312 :).
1447
1448         * Scripts/webkitpy/layout_tests/port/apple.py:
1449         (ApplePort.determine_full_port_name):
1450
1451 2012-06-29  Dirk Pranke  <dpranke@chromium.org>
1452
1453         webkitpy: remove support for mac leopard from chromium configurations
1454         https://bugs.webkit.org/show_bug.cgi?id=90313
1455
1456         Reviewed by Tony Chang.
1457
1458         Google has shipped the last version of Chrome that will support
1459         Mac OS 10.5 (Leopard), and we no longer have bots that run this
1460         configuration, so we're removing support for it.
1461
1462         A subsequent change will remove the baselines in platform/chromium-mac-leopard.
1463
1464         * Scripts/webkitpy/common/checkout/baselineoptimizer_unittest.py:
1465         (BaselineOptimizerTest.test_complex_shadowing):
1466         * Scripts/webkitpy/layout_tests/port/builders.py:
1467         * Scripts/webkitpy/layout_tests/port/chromium.py:
1468         (ChromiumPort):
1469         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1470         (ChromiumMacPort):
1471         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
1472         (ChromiumMacPortTest.test_versions):
1473         * Scripts/webkitpy/layout_tests/port/factory_unittest.py:
1474         (FactoryTest.test_mac):
1475         (FactoryTest.test_chromium_mac):
1476         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
1477         (test_rebaseline_and_copy_test_with_lion_result):
1478
1479 2012-06-29  Ojan Vafai  <ojan@chromium.org>
1480
1481         Fix optimize-baselines to not move baselines from win to win-7sp0
1482         https://bugs.webkit.org/show_bug.cgi?id=90312
1483
1484         Reviewed by Dirk Pranke.
1485
1486         It used to consider win-7sp0 as the common directory for all the Apple
1487         windows ports and incorrectly move results out of win.
1488
1489         * Scripts/webkitpy/common/checkout/baselineoptimizer.py:
1490         * Scripts/webkitpy/common/checkout/baselineoptimizer_unittest.py:
1491         (BaselineOptimizerTest.test_win_does_not_drop_to_win_7sp0):
1492         (BaselineOptimizerTest.test_common_directory_includes_root):
1493
1494 2012-06-28  Ojan Vafai  <ojan@chromium.org>
1495
1496         Fix rebaselining for Qt and Apple ports
1497         https://bugs.webkit.org/show_bug.cgi?id=90204
1498
1499         Reviewed by Dirk Pranke.
1500
1501         -Apporpriately put wk2 results in the -wk2 directories.
1502         -Since Qt and Apple-Win don't have bots that correspond to the
1503         platform/qt and platform/win directories, we need to fudge it
1504         and always put the results in those directories for those ports.
1505
1506         * Scripts/webkitpy/layout_tests/port/builders.py:
1507         (rebaseline_override_dir):
1508         * Scripts/webkitpy/layout_tests/port/factory.py:
1509         (_builder_options):
1510         Identify webkit2 builders by the WK2 in the builder name.
1511         * Scripts/webkitpy/tool/commands/rebaseline.py:
1512         (RebaselineTest._baseline_directory):
1513         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
1514         (TestRebaseline.test_baseline_directory):
1515
1516 2012-06-29  Zoltan Horvath  <zoltan@webkit.org>
1517
1518         Add support for --force parameter to run-performance-tests
1519         https://bugs.webkit.org/show_bug.cgi?id=90279
1520
1521         Reviewed by Dirk Pranke.
1522
1523         It's helpful to be able to run tests from the Skipped list of the performance tests.
1524
1525         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
1526         (PerfTestsRunner._parse_args):
1527         (PerfTestsRunner._collect_tests):
1528         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py: Add test.
1529         (test_collect_tests_with_skipped_list):
1530
1531 2012-06-29  Tony Chang  <tony@chromium.org>
1532
1533         [GTK] Enable CSS grid layout LayoutTests on GTK+
1534         https://bugs.webkit.org/show_bug.cgi?id=90226
1535
1536         Reviewed by Martin Robinson.
1537
1538         This feature is disabled via Settings by default, but for testing,
1539         we enable it using layoutTestController.overridePreferences. Add the
1540         necessary plumbing for DRT.
1541
1542         WTR already works because support was added for Apple Mac earlier.
1543
1544         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1545         (resetDefaultsToConsistentValues): Feature is off by default.
1546         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1547         (LayoutTestController::overridePreference): Add handling of WebKitCSSGridLayoutEnabled.
1548
1549 2012-06-29  Ádám Kallai  <kadam@inf.u-szeged.hu>
1550
1551         [Qt] Added Qt port for garden-o-matic.
1552         https://bugs.webkit.org/show_bug.cgi?id=82719
1553
1554         Reviewed by Adam Barth.
1555
1556         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/config.js:
1557         (.):
1558
1559 2012-06-29  Balazs Kelemen  <kbalazs@webkit.org>
1560
1561         [Qt][WTR] Get rid of using DumpRenderTreeSupportQt
1562         https://bugs.webkit.org/show_bug.cgi?id=90262
1563
1564         Reviewed by Alexey Proskuryakov.
1565
1566         Now that we decided to not support v8 in WebKit2
1567         we can get rid of using DumpRenderTreeSupportQt
1568         in WebKitTestRunner.
1569
1570         * Tools.pro:
1571         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1572         (WTR::InjectedBundlePage::resetAfterTest):
1573         (WTR::InjectedBundlePage::didClearWindowForFrame):
1574         * WebKitTestRunner/InjectedBundle/qt/ActivateFontsQt.cpp:
1575
1576 2012-06-29  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
1577
1578         Revert r121540, it broke most Qt builds
1579
1580         * qmake/mkspecs/features/default_post.prf:
1581         * qmake/mkspecs/features/functions.prf:
1582
1583 2012-06-29  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
1584
1585         [Qt] Make build-webkit reject uknown configurations, eg. --profile
1586
1587         The qmake-based buildsystem doesn't support it.
1588
1589         Reviewed by Tor Arne Vestbø.
1590
1591         * Scripts/webkitdirs.pm:
1592         (buildQMakeProjects):
1593
1594 2012-06-29  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
1595
1596         [Qt] Don't add Qt module dependencies in features.prf
1597
1598         The required dependencies are already added in WebCore.pri.
1599
1600         Reviewed by Tor Arne Vestbø.
1601
1602         * qmake/mkspecs/features/features.prf:
1603
1604 2012-06-27  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
1605
1606         [Qt] Use LIBS_PRIVATE instead of putting dependencies into LIBS
1607
1608         Reviewed by Tor Arne Vestbø..
1609
1610         * qmake/mkspecs/features/default_post.prf:
1611         * qmake/mkspecs/features/functions.prf:
1612
1613 2012-06-28  Ryosuke Niwa  <rniwa@webkit.org>
1614
1615         DOMHTMLCollection::item may return a wrong element after namedItem is called
1616         https://bugs.webkit.org/show_bug.cgi?id=90240
1617
1618         Reviewed by Antti Koivisto.
1619
1620         Add a WebKit API test since namedItem is not used in the JS/V8 binding code.
1621
1622         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
1623         * TestWebKitAPI/Tests/mac/HTMLCollectionNamedItem.html: Copied from Tools/TestWebKitAPI/Tests/mac/HTMLFormCollectionNamedItem.html.
1624         * TestWebKitAPI/Tests/mac/HTMLCollectionNamedItem.mm: Copied from Tools/TestWebKitAPI/Tests/mac/HTMLFormCollectionNamedItem.mm.
1625         (TestWebKitAPI::TEST):
1626
1627 2012-06-28  Dirk Pranke  <dpranke@chromium.org>
1628
1629         nrwt: remove the 'google-chrome' port code
1630         https://bugs.webkit.org/show_bug.cgi?id=88824
1631
1632         Reviewed by Ojan Vafai.
1633
1634         NRWT now supports passing additional baseline directories
1635         and expectations files on the command line, so there's no need
1636         to support the concept of a 'google-chrome' port directly.
1637
1638         * Scripts/webkitpy/layout_tests/port/base.py:
1639         (Port.path_to_test_expectations_file):
1640         * Scripts/webkitpy/layout_tests/port/builders.py:
1641         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1642         (ChromiumMacPort.__init__):
1643         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1644         (ChromiumWinPort.__init__):
1645         * Scripts/webkitpy/layout_tests/port/factory.py:
1646         (PortFactory):
1647         * Scripts/webkitpy/layout_tests/port/factory_unittest.py:
1648         (FactoryTest.test_win):
1649         * Scripts/webkitpy/layout_tests/port/google_chrome.py: Removed.
1650         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py: Removed.
1651
1652 2012-06-28  Dirk Pranke  <dpranke@chromium.org>
1653
1654         nrwt: clean up how arguments are passed to workers
1655         https://bugs.webkit.org/show_bug.cgi?id=90126
1656
1657         Reviewed by Ojan Vafai.
1658
1659         The way arguments are passed to workers has been crufty. It
1660         turns out it can be a lot cleaner via two things:
1661         1) using a factory method instead of instantiating objects
1662         directly in manager_worker_broker removes the need for passing
1663         'worker arguments' to the broker.
1664         2) it turns out that since mock hosts and test ports are purely
1665         in-memory constructions, they can be pickled and passed to child
1666         workers, meaning that the worker no longer needs hacky code to
1667         pass the port in a special case or to guess what to do if we
1668         don't have a host - all of the test-specific logic can move to
1669         the test file, where we can stub out the mock host's
1670         port_factory to return the same already-created port when it
1671         needs to be shared.
1672
1673         This change also moves WorkerException to manager_worker_broker.py
1674         where it belongs, and removes several useless tests that were
1675         just a maintenance burden (and would've needed rewriting when we
1676         change the rest of the broker implementation).
1677
1678         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1679         (Manager._run_tests.worker_factory):
1680         (Manager._run_tests):
1681         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
1682         (get):
1683         (WorkerException):
1684         (AbstractWorker.__init__):
1685         (_ManagerConnection.__init__):
1686         (_ManagerConnection.start_worker):
1687         (_InlineManager.__init__):
1688         (_InlineManager.start_worker):
1689         (_MultiProcessManager._can_pickle_host):
1690         (_MultiProcessManager):
1691         (_MultiProcessManager.start_worker):
1692         (_WorkerConnection.__init__):
1693         (_InlineWorkerConnection.__init__):
1694         (_InlineWorkerConnection.join):
1695         (_InlineWorkerConnection.run):
1696         (_Process.run):
1697         (_MultiProcessWorkerConnection.__init__):
1698         (_MultiProcessWorkerConnection.start):
1699         (_MultiProcessWorkerConnection):
1700         (_MultiProcessWorkerConnection.run):
1701         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py:
1702         (_TestWorker.__init__):
1703         (_TestWorker.run):
1704         (_TestsMixin.test_name):
1705         (_TestsMixin.test_cancel):
1706         (_TestsMixin.test_done):
1707         (_TestsMixin.test_unknown_message):
1708         (InlineBrokerTests.setUp):
1709         (MultiProcessBrokerTests.setUp):
1710         * Scripts/webkitpy/layout_tests/controllers/worker.py:
1711         (Worker.__init__):
1712         (Worker.run):
1713         * Scripts/webkitpy/layout_tests/controllers/worker_unittest.py: Removed.
1714         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1715         (passing_run):
1716         (logging_run):
1717         (run_and_capture):
1718         (MainTest.test_child_processes_2):
1719         (MainTest.test_child_processes_min):
1720         (MainTest.test_exception_raised):
1721         (MainTest.test_keyboard_interrupt):
1722         (MainTest.test_retrying_and_flaky_tests):
1723         (MainTest.test_run_order__inline):
1724
1725 2012-06-28  Dirk Pranke  <dpranke@chromium.org>
1726
1727         nrwt: don't try to catch worker exceptions in run_webkit_tests.__main__
1728         https://bugs.webkit.org/show_bug.cgi?id=90125
1729
1730         Reviewed by Ojan Vafai.
1731
1732         It turns out run_webkit_tests.py wasn't really using
1733         WorkerException and the catch we had for it was pointless. I've
1734         removed the symbol import and moved it to the integration tests
1735         where it is needed. Eventually the definition of the exception
1736         moves to the broker module, and so minimizing the number of
1737         users of it is a good thing.
1738
1739         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1740         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1741         (MainTest.test_exception_raised):
1742
1743 2012-06-28  Dirk Pranke  <dpranke@chromium.org>
1744
1745         nrwt: clean up passing of log messages between processes
1746         https://bugs.webkit.org/show_bug.cgi?id=90123
1747
1748         Reviewed by Ojan Vafai.
1749
1750         It turns out log messages are perfectly picklable by themselves
1751         and contain the process id of the process that generated the
1752         message, so if we just pass the record from the worker to the
1753         manager and call logger.handle() in the manager, everything just
1754         works :).
1755
1756         The change is covered by existing tests.
1757
1758         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1759         (Manager._log_messages):
1760         * Scripts/webkitpy/layout_tests/controllers/worker.py:
1761         (_WorkerLogHandler.emit):
1762         * Scripts/webkitpy/layout_tests/views/metered_stream.py:
1763         (MeteredStream.__init__):
1764         (_LogHandler.emit):
1765
1766 2012-06-28  Dirk Pranke  <dpranke@chromium.org>
1767
1768         add a pylint wrapper for linting python code
1769         https://bugs.webkit.org/show_bug.cgi?id=90232
1770
1771         Reviewed by Adam Barth.
1772
1773         Currently we use 'pep8' to check python code in
1774         check-webkit-style. pep8 is fast but simple; pylint is slower
1775         but has much more robust linting capabilities and will catch
1776         variable typos and all sorts of other things. Eventually we
1777         should switch check-webkit-style to use this, but our code is
1778         far from linting now so it needs to be cleaned up first.
1779
1780         This change adds the infrastructure and a wrapper so we can
1781         start doing that.
1782
1783         * Scripts/lint-webkitpy: Added.
1784         * Scripts/webkitpy/pylintrc: Added.
1785         * Scripts/webkitpy/thirdparty/__init__.py:
1786         (AutoinstallImportHook.find_module):
1787         (AutoinstallImportHook._install_pylint):
1788
1789 2012-06-28  Tony Chang  <tony@chromium.org>
1790
1791         [GTK] Use WEBKITOUTPUTDIR to find fonts in DumpRenderTree
1792         https://bugs.webkit.org/show_bug.cgi?id=90215
1793
1794         Reviewed by Martin Robinson.
1795
1796         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1797         (initializeFonts): Check for WEBKITOUTPUTDIR first.
1798         * Scripts/webkitpy/layout_tests/port/gtk.py:
1799         (GtkPort.setup_environ_for_server): Copy the environment variable to the child process.
1800         * WebKitTestRunner/InjectedBundle/gtk/ActivateFontsGtk.cpp:
1801         (WTR::inititializeFontConfigSetting): Check for WEBKITOUTPUTDIR first.
1802
1803 2012-06-28  Dirk Pranke  <dpranke@chromium.org>
1804
1805         derive ChromiumPort from WebKitPort in NRWT in order to support skipping tests if symbols are missing
1806         https://bugs.webkit.org/show_bug.cgi?id=89706
1807
1808         Reviewed by Ojan Vafai.
1809
1810         Try again to land the change first landed in r121363. This patch
1811         adds a bunch more tests and reworks the handling of
1812         port-specific default values for --pixel-tests and --time-out-ms
1813         to be more consistent (adding a default_pixel_tests() method,
1814         pushing the webkit default_timeout_ms() value up into base.py,
1815         and overriding it properly in the chromium and apple mac ports.
1816
1817         Also change the logic in
1818         run_webkit_tests._setup_derived_options() to not second-guess
1819         what the port wants the timeout value to be for debug builds;
1820         computing this in two different places led to several bugs.
1821
1822         This change also changes the Chromium unittest ports to derive
1823         from ChromiumPortTestCase instead of PortTestCase, so that we
1824         ensure that we're running the same tests on all port variants
1825         more easily. There's more cleanup that can be done here, but
1826         this is good enough for now
1827
1828         * Scripts/webkitpy/layout_tests/port/base.py:
1829         (Port.default_pixel_tests):
1830         (Port):
1831         (Port.default_timeout_ms):
1832         * Scripts/webkitpy/layout_tests/port/chromium.py:
1833         (ChromiumPort):
1834         (ChromiumPort.__init__):
1835         (ChromiumPort.default_pixel_tests):
1836         (ChromiumPort.default_timeout_ms):
1837         (ChromiumPort.driver_name):
1838         (ChromiumPort._driver_class):
1839         (ChromiumPort._missing_symbol_to_skipped_tests):
1840         (ChromiumPort.skipped_layout_tests):
1841         (ChromiumPort.setup_test_run):
1842         (ChromiumPort._path_to_image_diff):
1843         (ChromiumPort._convert_path):
1844         * Scripts/webkitpy/layout_tests/port/chromium_android_unittest.py:
1845         (ChromiumAndroidPortTest):
1846         (ChromiumAndroidPortTest.test_expectations_files):
1847         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1848         (ChromiumLinuxPort._modules_to_search_for_symbols):
1849         * Scripts/webkitpy/layout_tests/port/chromium_linux_unittest.py:
1850         (ChromiumLinuxPortTest):
1851         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1852         (ChromiumMacPort._modules_to_search_for_symbols):
1853         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
1854         (ChromiumMacPortTest):
1855         * Scripts/webkitpy/layout_tests/port/chromium_port_testcase.py:
1856         split off from chromium_unittest.
1857         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1858         (ChromiumWinPort._modules_to_search_for_symbols):
1859         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1860         (ChromiumWinTest):
1861         * Scripts/webkitpy/layout_tests/port/mac.py:
1862         (MacPort.default_timeout_ms):
1863         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1864         (MacTest.test_default_timeout_ms):
1865         * Scripts/webkitpy/layout_tests/port/mock_drt.py:
1866         (MockDRTPort.start_http_server):
1867         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
1868         (PortTestCase.test_default_timeout_ms):
1869         (PortTestCase):
1870         (PortTestCase.test_default_pixel_tests):
1871         * Scripts/webkitpy/layout_tests/port/test.py:
1872         (TestPort.default_pixel_tests):
1873         * Scripts/webkitpy/layout_tests/port/webkit.py:
1874         (WebKitPort._modules_to_search_for_symbols):
1875         (WebKitPort):
1876         (WebKitPort._symbols_string):
1877         (WebKitPort._skipped_tests_for_unsupported_features):
1878         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
1879         (TestWebKitPort._symbols_string):
1880         (TestWebKitPort._tests_for_disabled_features):
1881         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1882         (_set_up_derived_options):
1883
1884 2012-06-28  Joshua Bell  <jsbell@chromium.org>
1885
1886         run-bindings-tests should return non-zero exit code on test failure
1887         https://bugs.webkit.org/show_bug.cgi?id=90205
1888
1889         Reviewed by Adam Barth.
1890
1891         * Scripts/run-bindings-tests:
1892         (main):
1893
1894 2012-06-28  Terry Anderson  <tdanderson@chromium.org>
1895
1896         [chromium] Reset the device scale factor to 1 before each test is run
1897         https://bugs.webkit.org/show_bug.cgi?id=90212
1898
1899         Reviewed by Adam Barth.
1900
1901         Some tests change the device scale factor, so this needs to be reset to
1902         1.0 at the start of each test to avoid test flakiness.
1903
1904         * DumpRenderTree/chromium/LayoutTestController.cpp:
1905         (LayoutTestController::reset):
1906
1907 2012-06-28  Ryosuke Niwa  <rniwa@webkit.org>
1908
1909         Cleanup HTMLFormCollection
1910         https://bugs.webkit.org/show_bug.cgi?id=90111
1911
1912         Reviewed by Andreas Kling.
1913
1914         Add a WebKit API test using copy-paste design pattern per kling's request.
1915
1916         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
1917         * TestWebKitAPI/Tests/mac/HTMLFormCollectionNamedItem.html: Added.
1918         * TestWebKitAPI/Tests/mac/HTMLFormCollectionNamedItem.mm: Added.
1919         (-[HTMLFormCollectionNamedItemTest webView:didFinishLoadForFrame:]):
1920         (TestWebKitAPI):
1921         (TestWebKitAPI::TEST):
1922
1923 2012-06-28  Tim Horton  <timothy_horton@apple.com>
1924
1925         [mac] WKTR windows still don't stay off screen sometimes
1926         https://bugs.webkit.org/show_bug.cgi?id=90214
1927         <rdar://problem/11760263>
1928
1929         Reviewed by Simon Fraser.
1930
1931         In some cases, the system itself will consult [WebKitTestRunnerWindow frame], so we should refrain from
1932         overriding it and instead provide a different method to use when we want the web-facing "fake" window origin
1933         (for PlatformWebView::windowFrame()).
1934
1935         * WebKitTestRunner/mac/PlatformWebViewMac.mm:
1936         (-[WebKitTestRunnerWindow frameRespectingFakeOrigin]):
1937         (WTR::PlatformWebView::windowFrame):
1938
1939 2012-06-28  Martin Robinson  <mrobinson@igalia.com>
1940
1941         [GTK] Add unit tests for GtkInputMethodFilter
1942         https://bugs.webkit.org/show_bug.cgi?id=88698
1943
1944         Reviewed by Carlos Garcia Campos.
1945
1946         Add unit tests for GtkInputMethodFilter in the WebCore platform layer.
1947         This change adds the TestGtk test suite which will be used for all non-API
1948         layer GTK unit tests.
1949
1950         * TestWebKitAPI/GNUmakefile.am: Update the build to include the new tests.
1951         * TestWebKitAPI/Tests/gtk/InputMethodFilter.cpp: Added.
1952         (TestWebKitAPI::PlatformWebView::PlatformWebView): Remove the call to gtk_init here
1953         as it's now in main.cpp.
1954         * TestWebKitAPI/gtk/main.cpp: Change the g_type_init call to gtk_init, because now
1955         a majority of all unit tests depend on GTK+ being initialized.
1956
1957 2012-06-28  Christophe Dumez  <christophe.dumez@intel.com>
1958
1959         [EFL] Enable support for HTML5 datalist
1960         https://bugs.webkit.org/show_bug.cgi?id=90157
1961
1962         Reviewed by Martin Robinson.
1963
1964         Turn on DATALIST flag by default on EFL port to
1965         support HTML5 datalist tag.
1966
1967         * Scripts/webkitperl/FeatureList.pm:
1968
1969 2012-06-28  Tony Chang  <tony@chromium.org>
1970
1971         Enable CSS grid layout LayoutTests on platform Mac
1972         https://bugs.webkit.org/show_bug.cgi?id=90113
1973
1974         Reviewed by Ojan Vafai.
1975
1976         * DumpRenderTree/mac/DumpRenderTree.mm:
1977         (resetDefaultsToConsistentValues): Reset the value to NO between tests.
1978
1979 2012-06-28  Mike Fenton  <mifenton@rim.com>
1980
1981         [BlackBerry] Add watchlist options for Blackberry and editing.
1982         https://bugs.webkit.org/show_bug.cgi?id=90193
1983
1984         Unreviewed.
1985
1986         Add BlackBerry and Editing watchlist and monitor them.
1987
1988         * Scripts/webkitpy/common/config/watchlist:
1989
1990 2012-06-28  Tim Horton  <timothy_horton@apple.com>
1991
1992         Unreviewed, upgrade myself to a reviewer!
1993
1994         http://www.webkit.org/blog/2082/tim-horton-is-now-a-webkit-reviewer/
1995
1996         * Scripts/webkitpy/common/config/committers.py:
1997
1998 2012-06-28  Terry Anderson  <tdanderson@chromium.org>
1999
2000         [Chromium] Chromium's LayoutTestController is missing setBackingScaleFactor
2001         https://bugs.webkit.org/show_bug.cgi?id=83635
2002
2003         Reviewed by Adam Barth.
2004
2005         Added InvokeCallbackTask, a new derived class of MethodTask. When
2006         setBackingScaleFactor is called, a call to setDeviceScaleFactor in
2007         WebView is made and then postTask is used to invoke the callback
2008         function specified in testRunner.setBackingScaleFactor. 
2009
2010         * DumpRenderTree/chromium/LayoutTestController.cpp:
2011         (LayoutTestController::LayoutTestController):
2012         (InvokeCallbackTask):
2013         (InvokeCallbackTask::InvokeCallbackTask):
2014         (InvokeCallbackTask::runIfValid):
2015         (LayoutTestController::setBackingScaleFactor):
2016         * DumpRenderTree/chromium/LayoutTestController.h:
2017         (LayoutTestController):
2018
2019 2012-06-28  Ojan Vafai  <ojan@chromium.org>
2020
2021         Make rebaseline-test and rebaseline-expectations work for non-Chromium ports
2022         https://bugs.webkit.org/show_bug.cgi?id=90186
2023
2024         Reviewed by Adam Barth.
2025
2026         This makes rebaselining work for all ports that have a TestExpectations file
2027         in the tree. I didn't test other ports.
2028
2029         This doesn't address 100% of the problem. The rebaseline code puts the expectations
2030         in the most specific directory and relies on optimize-baselines to merge baselines
2031         appropriately. This only works if every platform directory has an equivalent bot
2032         that runs the tests, which is not true for most ports.
2033
2034         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
2035         (Builder._revision_and_build_for_filename):
2036         Some bots have filenames that aren't revision/build number pairs
2037         e.g. they are random junk like aQhxvx. Handle this gracefully.
2038         (Builder._fetch_revision_to_build_map):
2039         (Builder._file_info_list_to_revision_to_build_list):
2040         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2041         (BuilderTest.test_build_and_revision_for_filename):
2042         (BuilderTest.test_file_info_list_to_revision_to_build_list):
2043         * Scripts/webkitpy/layout_tests/port/builders.py:
2044         Update the list of builders. This list needs to be kept up
2045         to date for the rebaseline tool to work.
2046         * Scripts/webkitpy/tool/commands/rebaseline.py:
2047         (RebaselineTest._results_url):
2048         (RebaselineExpectations._run_webkit_patch):
2049         (RebaselineExpectations._rebaseline_port):
2050         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
2051         Qt port uses qmake to determine the right version. Systems without qmake correctly fallback
2052         to a specific version.
2053
2054 2012-06-28  Csaba Osztrogonác  <ossy@webkit.org>
2055
2056         [Qt][NRWT] Fix baseline and skipped file search path.
2057         https://bugs.webkit.org/show_bug.cgi?id=89882
2058
2059         Unreviewed trivial typo fix after r121430.
2060
2061         * Scripts/webkitpy/layout_tests/port/qt_unittest.py:
2062         (QtPortTest._assert_search_path):
2063         (QtPortTest._assert_skipped_path):
2064
2065 2012-06-28  János Badics  <jbadics@inf.u-szeged.hu>
2066
2067         [Qt][NRWT] Fix baseline and skipped file search path.
2068         https://bugs.webkit.org/show_bug.cgi?id=89882
2069
2070         Reviewed by Csaba Osztrogonác.
2071
2072         * Scripts/webkitpy/layout_tests/port/qt.py:
2073         (QtPort._search_paths):
2074         (QtPort):
2075         (QtPort.baseline_search_path):
2076         (QtPort._skipped_file_search_paths):
2077         * Scripts/webkitpy/layout_tests/port/qt_unittest.py:
2078         (QtPortTest):
2079         (QtPortTest._assert_skipped_path):
2080         (QtPortTest.test_baseline_search_path):
2081         (QtPortTest.test_skipped_file_search_path):
2082
2083 2012-06-28  Alexander Pavlov  <apavlov@chromium.org>
2084
2085         [Qt] inspector/styles/inject-stylesheet.html makes 4 tests flakey (TEXT PASS)
2086         https://bugs.webkit.org/show_bug.cgi?id=90167
2087
2088         Reviewed by Csaba Osztrogonác.
2089
2090         DRT should remove user stylesheets from the page group when resetting before running another test.
2091
2092         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
2093         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
2094
2095 2012-06-28  Csaba Osztrogonác  <ossy@webkit.org>
2096
2097         [Qt][DRT] Reset AcceleratedCompositingEnabled between tests
2098         https://bugs.webkit.org/show_bug.cgi?id=90164
2099
2100         Reviewed by Simon Hausmann.
2101
2102         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
2103         (WebCore::WebPage::resetSettings):
2104
2105 2012-06-28  Csaba Osztrogonác  <ossy@webkit.org>
2106
2107         [Qt] Restore original value of mock scrollbars between tests
2108         https://bugs.webkit.org/show_bug.cgi?id=90155
2109
2110         Reviewed by Simon Hausmann.
2111
2112         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
2113         (WebCore::WebPage::resetSettings):
2114         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2115         (LayoutTestController::setMockScrollbarsEnabled):
2116         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2117         (LayoutTestController):
2118
2119 2012-06-27  Mark Rowe  <mrowe@apple.com>
2120
2121         <http://webkit.org/b/90096> Building within Xcode sometimes does a full rebuild after building via build-webkit
2122
2123         We need to ensure that build-webkit uses the same setting for SHARED_PRECOMPS_DIR
2124         as what Xcode itself will use when building, otherwise switching between the two
2125         can result in the precompiled headers being rebuilt and thus the entire world
2126         rebuilding.
2127
2128         Reviewed by Dan Bernstein.
2129
2130         * Scripts/webkitdirs.pm:
2131         (determineBaseProductDir):
2132
2133 2012-06-27  Sheriff Bot  <webkit.review.bot@gmail.com>
2134
2135         Unreviewed, rolling out r121363, r121367, r121384, and
2136         r121390.
2137         http://trac.webkit.org/changeset/121363
2138         http://trac.webkit.org/changeset/121367
2139         http://trac.webkit.org/changeset/121384
2140         http://trac.webkit.org/changeset/121390
2141         https://bugs.webkit.org/show_bug.cgi?id=90134
2142
2143         It broke debug NRWT on GTK and on Qt (Requested by Ossy_NIGHT
2144         on #webkit).
2145
2146         * Scripts/webkitpy/layout_tests/port/chromium.py:
2147         (ChromiumPort):
2148         (ChromiumPort.__init__):
2149         (ChromiumPort._check_file_exists):
2150         (ChromiumPort.default_results_directory):
2151         (ChromiumPort._driver_class):
2152         (ChromiumPort._build_path):
2153         (ChromiumPort._path_to_image_diff):
2154         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2155         (ChromiumLinuxPort.baseline_search_path):
2156         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
2157         (ChromiumMacPort.baseline_search_path):
2158         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2159         (ChromiumPortTest):
2160         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2161         (ChromiumWinPort.baseline_search_path):
2162         * Scripts/webkitpy/layout_tests/port/webkit.py:
2163         (WebKitPort.__init__):
2164         (WebKitPort._webcore_symbols_string):
2165         (WebKitPort._skipped_tests_for_unsupported_features):
2166         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
2167         (TestWebKitPort._webcore_symbols_string):
2168         (WebKitPortUnitTests.test_default_options):
2169
2170 2012-06-27  Dirk Pranke  <dpranke@chromium.org>
2171
2172         Fix typo in r121384 :(
2173
2174         Unreviewed, build fix.
2175
2176         * Scripts/webkitpy/layout_tests/port/webkit.py:
2177         (WebKitPort.default_test_timeout_ms):
2178
2179 2012-06-27  Dirk Pranke  <dpranke@chromium.org>
2180
2181         nrwt: default timeout for debug bots broke in r121363
2182         https://bugs.webkit.org/show_bug.cgi?id=90112
2183
2184         Unreviewed, build fix.
2185
2186         I forgot to account for release and debug having different
2187         default values :(.
2188
2189         * Scripts/webkitpy/layout_tests/port/chromium.py:
2190         (ChromiumPort.default_test_timeout_ms):
2191         * Scripts/webkitpy/layout_tests/port/webkit.py:
2192         (WebKitPort):
2193         (WebKitPort.default_test_timeout_ms):
2194         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
2195         (WebKitPortUnitTests.test_default_options):
2196
2197 2012-06-27  Dirk Pranke  <dpranke@chromium.org>
2198
2199         webkitpy: fix a couple of issues running under cygwin
2200         https://bugs.webkit.org/show_bug.cgi?id=90035
2201
2202         Reviewed by Eric Seidel.
2203
2204         These were causing unit tests to fail on cygwin (apple win bot).
2205
2206         * Scripts/webkitpy/layout_tests/port/chromium.py:
2207         * Scripts/webkitpy/performance_tests/perftest.py:
2208
2209 2012-06-27  Adam Barth  <abarth@webkit.org>
2210
2211         [Chromium] DumpRenderTree on Android should call SkUseTestFontConfigFile once available
2212         https://bugs.webkit.org/show_bug.cgi?id=89801
2213
2214         Reviewed by Nate Chapin.
2215
2216         Let's call SkUseTestFontConfigFile now that it exists.
2217
2218         * DumpRenderTree/chromium/TestShellAndroid.cpp:
2219         (platformInit):
2220
2221 2012-06-27  Dirk Pranke  <dpranke@chromium.org>
2222
2223         Fix typo introduced in r121363.
2224
2225         Unreviewed, build fix.
2226
2227         * Scripts/webkitpy/layout_tests/port/webkit.py:
2228         (WebKitPort._symbols_string):
2229
2230 2012-06-27  Mark Hahnenberg  <mhahnenberg@apple.com>
2231
2232         Moving from committers to reviewers :-)
2233
2234         Rubber stamped by Filip Pizlo.
2235
2236         * Scripts/webkitpy/common/config/committers.py:
2237
2238 2012-06-27  Dirk Pranke  <dpranke@chromium.org>
2239
2240         Derive ChromiumPort from WebKitPort to add support for missing symbols to skip tests
2241         https://bugs.webkit.org/show_bug.cgi?id=89706
2242
2243         Reviewed by Adam Barth.
2244
2245         Based on the original patch by Raymond Toy.
2246
2247         This patch changes ChromiumPort to derive from webkit.WebKitPort
2248         instead of base.Port. This is a long-awaited change and a
2249         precursor to merging base.Port and webkit.WebKitPort, but is
2250         driven by the desire to dynamically detect whether the MP3 and
2251         AAC codecs are compiled into DRT on Chromium, for which we
2252         wanted to re-use the existing logic in WebKit port for determine
2253         what to skip at compile time.
2254
2255         Most of the changes are shuffling things around so that we don't
2256         change any other logic and so we override the necessary methods
2257         in WebKitPort (and try to follow the same method definition
2258         order where possible).
2259
2260         Also, on the Chromium port the mp3 and aac codecs are actually
2261         defined in a separate library, so scanning webcore isn't
2262         sufficient. This patch generalizes the symbol lookup to handle
2263         multiple libraries, and uses different libraries as appropriate
2264         for Chromium.
2265
2266         The only functional/visible changes should be in the value
2267         returned for skipped_layout_tests().
2268
2269         * Scripts/webkitpy/layout_tests/port/chromium.py:
2270         (ChromiumPort):
2271         (ChromiumPort.__init__):
2272         (ChromiumPort.driver_name):
2273         (ChromiumPort._driver_class):
2274         (ChromiumPort._missing_symbol_to_skipped_tests):
2275         (ChromiumPort.skipped_layout_tests):
2276         (ChromiumPort.setup_test_run):
2277         (ChromiumPort._path_to_image_diff):
2278         (ChromiumPort._convert_path):
2279         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2280         (ChromiumPortTest.test_missing_symbol_to_skipped_tests):
2281         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2282         (ChromiumLinuxPort._modules_to_search_for_symbols):
2283         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
2284         (ChromiumLinuxPort._modules_to_search_for_symbols):
2285         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2286         (ChromiumLinuxPort._modules_to_search_for_symbols):
2287         * Scripts/webkitpy/layout_tests/port/webkit.py:
2288         (WebKitPort):
2289         (WebKitPort.__init__):
2290         (WebKitPort._symbols_string):
2291         (WebKitPort._modules_to_search_for_symbols):
2292
2293 2012-06-27  Csaba Osztrogonác  <ossy@webkit.org>
2294
2295         [Qt] REGRESSION(r121339): It broke the build on the Qt Windows bots
2296         https://bugs.webkit.org/show_bug.cgi?id=90081
2297
2298         Buildfix for Qt 4.8 Windows. Use the former path for Qt 4.8, and the newer one for Qt 5.
2299
2300         Reviewed by Noam Rosenthal.
2301
2302         * qmake/mkspecs/features/features.prf:
2303
2304 2012-06-27  Sergio Villar Senin  <svillar@igalia.com>
2305
2306         [WK2] [GTK] WebKit2 testing bot fails to run tests due to missing files
2307         https://bugs.webkit.org/show_bug.cgi?id=90061
2308
2309         Reviewed by Gustavo Noronha Silva.
2310
2311         Add "-no-install -no-fast-install" to the LDFLAGS in bots. With
2312         this flag libtool tells the linker to set the rpath for the output
2313         file to the full path of the .libs directory, instead of using a
2314         wrapper script to set up the LD_LIBRARY_PATH. This will allow us
2315         to directly reuse builds in the pure testing bots.
2316
2317         * BuildSlaveSupport/gtk/daemontools-buildbot.conf:
2318
2319 2012-06-27  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2320
2321         [Qt] No need to save and restore TEMPLATE in a function
2322
2323         This was a leftover from when the logic was not in its own function scope.
2324
2325         QMAKE_FRAMEWORK_BUNDLE_NAME on the other hand is exported in qtLibraryTarget, which
2326         will surprisingly affect the global scope as well, so we have to save and restore it.
2327
2328         Original patch by Oswald Buddenhagen <oswald.buddenhagen@nokia.com> on 2012-06-27
2329
2330         Reviewed by Tor Arne Vestbø.
2331
2332         * qmake/mkspecs/features/functions.prf:
2333
2334 2012-06-27  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
2335
2336         [Qt] Fix lookup location for sqlite sources
2337
2338         Don't look in the install dir - we are unlikely to find anything there
2339         unless we are doing a developer build.
2340
2341         Reviewed by Tor Arne Vestbø.
2342
2343         * qmake/mkspecs/features/features.prf:
2344
2345 2012-06-27  Zan Dobersek  <zandobersek@gmail.com>
2346
2347         [Gtk] Add support for the Gamepad API
2348         https://bugs.webkit.org/show_bug.cgi?id=87503
2349
2350         Reviewed by Carlos Garcia Campos.
2351
2352         Enable the gamepad support for the GTK port.
2353
2354         * Scripts/webkitperl/FeatureList.pm:
2355
2356 2012-06-27  Ryosuke Niwa  <rniwa@webkit.org>
2357
2358         Fix gcc build after r121302
2359         https://bugs.webkit.org/show_bug.cgi?id=90055
2360
2361         Reviewed by Mark Rowe.
2362
2363         Assume RTTI is always disabled so that gtest builds on XCode 3.2.6.
2364
2365         It appears that gcc doesn't like window.get().* inside a template so replace that by [window.get() *] instead.
2366
2367         * TestWebKitAPI/Configurations/Base.xcconfig:
2368         * TestWebKitAPI/Tests/mac/AcceptsFirstMouse.mm:
2369         (TestWebKitAPI::AcceptsFirstMouse::runTest):
2370
2371 2012-06-27  Csaba Osztrogonác  <ossy@webkit.org>
2372
2373         Add master.cfg unittest to help migration - pass BuildStep instances instead of BuildStep subclasses
2374         https://bugs.webkit.org/show_bug.cgi?id=89564
2375
2376         Reviewed by Tony Chang.
2377
2378         * BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:
2379         (BuildStepsConstructorTest):
2380         (BuildStepsConstructorTest.generateTests):
2381         (BuildStepsConstructorTest.createTest):
2382         (BuildStepsConstructorTest.createTest.doTest):
2383
2384 2012-06-26  Mark Hahnenberg  <mhahnenberg@apple.com>
2385
2386         Add support for preciseTime() to WebKitTestRunner
2387         https://bugs.webkit.org/show_bug.cgi?id=90027
2388
2389         Reviewed by Darin Adler.
2390
2391         It would be nice to be able to use preciseTime() in WebKitTestRunner like we can in DumpRenderTree.
2392
2393         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
2394         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: 
2395         (WTR::LayoutTestController::preciseTime): 
2396         (WTR):
2397         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
2398         (LayoutTestController):
2399
2400 2012-06-26  Lucas Forschler  <lforschler@apple.com>
2401
2402       Teach the Apple port how to build the test tools in build-webkit
2403       https://bugs.webkit.org/show_bug.cgi?id=89540
2404
2405       Reviewed by Jon Lee & Simon Fraser.
2406
2407       * BuildSlaveSupport/build.webkit.org-config/master.cfg:
2408       (RunWebKitTests.start): pass --no-build since tools should now be in the downloaded archive
2409       (RunUnitTests.start): ditto
2410       * Scripts/build-webkit: add tools to the projects build list
2411
2412 2012-06-26  Ojan Vafai  <ojan@chromium.org>
2413
2414         Only show the platform-appropriate builders for non-layout test failures in garden-o-matic
2415         https://bugs.webkit.org/show_bug.cgi?id=90025
2416
2417         Reviewed by Simon Fraser.
2418
2419         Move the chromium-specific filtering code into config.js and replace it with a method on each platform
2420         config. Also, let the webkit test step name be webkit_tests (build.chromium.org) or layout-test (build.webkit.org).
2421
2422         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/builders.js:
2423         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/builders_unittests.js:
2424         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/config.js:
2425
2426 2012-06-26  Ojan Vafai  <ojan@chromium.org>
2427
2428         Fix failing garden-o-matic unittests
2429         https://bugs.webkit.org/show_bug.cgi?id=90021
2430
2431         Reviewed by Adam Barth.
2432
2433         These had all just fallen out of date.
2434
2435         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/checkout_unittests.js:
2436         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/notifications_unittests.js:
2437         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui_unittests.js:
2438
2439 2012-06-26  Ojan Vafai  <ojan@chromium.org>
2440
2441         Fix platform picker change handler in garden-o-matic
2442         https://bugs.webkit.org/show_bug.cgi?id=90010
2443
2444         Reviewed by Simon Fraser.
2445
2446         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui.js:
2447         The old code never worked. This is hard to test because change handlers require
2448         a user-initiated action and the code is changing the window's location, which would
2449         navigate the test page.
2450
2451         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/styles/onebar.css:
2452         Fix the CSS so it doesn't cause the tabstrip to be disconnected from the tabs on Linux
2453         due to the large margin-bottom.
2454
2455 2012-06-26  Thiago Marcos P. Santos  <thiago.santos@intel.com>
2456
2457         webkitpy: Make webkit-patch patches-to-review useful
2458         https://bugs.webkit.org/show_bug.cgi?id=89470
2459
2460         Reviewed by Eric Seidel.
2461
2462         webkit-patch patches-to-review will now output the list of
2463         bugs with patches pending for review that has the user on CC,
2464         excluding patches with cq-, sorted by the age of the patch.
2465
2466         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
2467         (BugzillaQueries.fetch_bugs_from_review_queue):
2468         * Scripts/webkitpy/common/net/bugzilla/bugzilla_mock.py:
2469         (MockBugzillaQueries.fetch_bugs_from_review_queue):
2470         (MockBugzilla.__init__):
2471         (MockBugzilla.authenticate):
2472         * Scripts/webkitpy/tool/commands/queries.py:
2473         (PatchesToReview):
2474         (PatchesToReview.__init__):
2475         (PatchesToReview._print_report):
2476         (PatchesToReview._generate_report):
2477         (PatchesToReview.execute):
2478         * Scripts/webkitpy/tool/commands/queries_unittest.py:
2479         (QueryCommandsTest.test_patches_to_review):
2480
2481 2012-06-26  Dirk Pranke  <dpranke@chromium.org>
2482
2483         nrwt: broken for chromium on vista
2484         https://bugs.webkit.org/show_bug.cgi?id=89988
2485
2486         Reviewed by Tony Chang.
2487
2488         r121194 removed support for 'chromium-win-vista' as a separate
2489         port, but this actually prevented the code from running on vista
2490         at all, which is unduly harsh and broke the websocket tests on
2491         the (non-webkit) chromium bots that are still running on vista.
2492
2493         It's probably good enough to pretend that vista is win7 instead;
2494         some layout tests will still fail but at least things'll run.
2495
2496         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2497         (ChromiumWinPort.determine_full_port_name):
2498         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2499         (ChromiumWinTest.test_versions):
2500
2501 2012-06-26  Csaba Osztrogonác  <ossy@webkit.org>
2502
2503         master.cfg cleanup: Pass CheckOutSource instance instead of class to BuildStep.addStep
2504         https://bugs.webkit.org/show_bug.cgi?id=89215
2505
2506         We need it because it is deprecated and will be dropped in buildbot 0.8.7
2507
2508         Reviewed by Tony Chang.
2509
2510         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
2511         (CheckOutSource.__init__):
2512         (Factory.__init__):
2513
2514 2012-06-26  Simon Hausmann  <simon.hausmann@nokia.com>
2515
2516         [Qt][Win] Symbols are not exported in QtWebKit5.dll
2517         https://bugs.webkit.org/show_bug.cgi?id=88873
2518
2519         Reviewed by Tor Arne Vestbø.
2520
2521         When linking the target dll make sure to re-export the symbols from
2522         the static libraries marked as export, with the help of a little python
2523         script and a qmake extra compiler.
2524
2525         * Scripts/generate-win32-export-forwards: Added.
2526         * qmake/mkspecs/features/win32/default_post.prf:
2527
2528 2012-06-25  Jocelyn Turcotte  <turcotte.j@gmail.com>
2529
2530         Add a note about hostname completion not working well with --cc completion
2531
2532         Reviewed by Simon Hausmann.
2533
2534         Hostname completion tries to resolve anything after an @ sign which is present
2535         in the completed list of contributor emails to CC.
2536
2537         * Scripts/webkit-tools-completion.sh:
2538
2539 2012-06-26  Sheriff Bot  <webkit.review.bot@gmail.com>
2540
2541         Unreviewed, rolling out r121244.
2542         http://trac.webkit.org/changeset/121244
2543         https://bugs.webkit.org/show_bug.cgi?id=89966
2544
2545         skip list path is incorrect now. (Requested by Ossy on
2546         #webkit).
2547
2548         * Scripts/webkitpy/layout_tests/port/qt.py:
2549         (QtPort.baseline_search_path):
2550         (QtPort._skipped_file_search_paths):
2551         * Scripts/webkitpy/layout_tests/port/qt_unittest.py:
2552         (QtPortTest.test_baseline_search_path):
2553
2554 2012-06-26  János Badics  <jbadics@inf.u-szeged.hu>
2555
2556         [Qt][NRWT] Fix baseline and skipped file search path.
2557         https://bugs.webkit.org/show_bug.cgi?id=89882
2558
2559         Reviewed by Csaba Osztrogonác.
2560
2561         * Scripts/webkitpy/layout_tests/port/qt.py:
2562         (QtPort.baseline_search_path):
2563         * Scripts/webkitpy/layout_tests/port/qt_unittest.py:
2564         (QtPortTest.test_baseline_search_path):
2565
2566 2012-06-26  Tony Chang  <tony@chromium.org>
2567
2568         [Qt] Enable grid layout LayoutTests
2569         https://bugs.webkit.org/show_bug.cgi?id=89909
2570
2571         Reviewed by Csaba Osztrogonác.
2572
2573         These tests pass, we just need to hook up the overridePreference.
2574
2575         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
2576         (WebCore::WebPage::resetSettings): Reset grid layout and regions between tests.
2577         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2578         (LayoutTestController::overridePreference): Add WebKitCSSGridLayoutEnabled.
2579
2580 2012-06-26  Sheriff Bot  <webkit.review.bot@gmail.com>
2581
2582         Unreviewed, rolling out r121236.
2583         http://trac.webkit.org/changeset/121236
2584         https://bugs.webkit.org/show_bug.cgi?id=89956
2585
2586         It's broke the nrwt on qt and gtk platform (Requested by
2587         kkristof on #webkit).
2588
2589         * Scripts/webkitpy/layout_tests/port/qt.py:
2590         (QtPort._build_driver):
2591         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py:
2592         (XvfbDriver._start):
2593         (XvfbDriver._start.x_filter):
2594         (XvfbDriver.stop):
2595
2596 2012-06-26  Kristóf Kosztyó  <kkristof@inf.u-szeged.hu>
2597
2598         [NRWT] XvfbDriver should choose the next free display
2599         https://bugs.webkit.org/show_bug.cgi?id=88414
2600
2601         Reviewed by Dirk Pranke.
2602
2603         * Scripts/webkitpy/layout_tests/port/qt.py:
2604         (QtPort._driver_class):
2605         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py:
2606         (XvfbDriver.__init__):
2607         (XvfbDriver._start):
2608         (XvfbDriver._start.next_free_id):
2609         (XvfbDriver.stop):
2610
2611 2012-06-25  Simon Hausmann  <simon.hausmann@nokia.com>
2612
2613         [Qt] Make it possible to build WebKit without QtWidgets
2614         https://bugs.webkit.org/show_bug.cgi?id=78109
2615
2616         Reviewed by Tor Arne Vestbø.
2617
2618         * Tools.pro: Disable various tools that depend on WK1.
2619         * qmake/mkspecs/features/default_post.prf: Add WK1 as supported static lib in WEBKIT
2620         * qmake/mkspecs/features/default_pre.prf: Disable WK1 if QtWidgets is not built.
2621         * qmake/mkspecs/modules/webkit1.prf: Added.
2622
2623 2012-06-25  Dirk Pranke  <dpranke@chromium.org>
2624
2625         put vista back for chromium (non-webkit) bots ...
2626         https://bugs.webkit.org/show_bug.cgi?id=89929
2627
2628         Unreviewed, build fix.
2629
2630         Turns out we've only gotten rid of vista on the webkit tests so
2631         far. Put vista back for the other chromium bots ...
2632
2633         * TestResultServer/static-dashboards/flakiness_dashboard.js:
2634         (chromiumPlatform):
2635         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
2636         (test):
2637
2638 2012-06-25  Dirk Pranke  <dpranke@chromium.org>
2639
2640         update flakiness dashboard after removing chromium-vista bots
2641         https://bugs.webkit.org/show_bug.cgi?id=89925
2642
2643         Unreviewed, build fix.
2644
2645         More changes missed in r121194 :(.
2646
2647         * TestResultServer/static-dashboards/flakiness_dashboard.js:
2648         (chromiumPlatform):
2649         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
2650         (test):
2651
2652 2012-06-25  Dirk Pranke  <dpranke@chromium.org>
2653
2654         remove support for chromium vista from tools
2655         https://bugs.webkit.org/show_bug.cgi?id=89915
2656
2657         Reviewed by Tony Chang.
2658
2659         This change is all deleting code and updating tests to handle
2660         the configuration being gone. All tests now pass except for a
2661         few unittests for the flakiness dashboard which were failing
2662         before this patch.
2663
2664         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/builders_unittests.js:
2665         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/checkout_unittests.js:
2666         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/config.js:
2667         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/notifications_unittests.js:
2668         * Scripts/webkitpy/common/checkout/baselineoptimizer_unittest.py:
2669         (BaselineOptimizerTest.test_complex_shadowing):
2670         * Scripts/webkitpy/layout_tests/port/builders.py:
2671         * Scripts/webkitpy/layout_tests/port/chromium.py:
2672         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2673         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2674         (ChromiumWinPort):
2675         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2676         (ChromiumWinTest.test_versions):
2677         (ChromiumWinTest.test_baseline_path):
2678         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
2679         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py:
2680         (BuildCoverageExtrapolatorTest.test_extrapolate):
2681
2682 2012-06-25  Scott Graham  <scottmg@chromium.org>
2683
2684         [Chromium] add empty impl of WebThemeEngine::getSize() for DRT
2685         https://bugs.webkit.org/show_bug.cgi?id=89907
2686
2687         Not reviewed. Fix build in Chromium.
2688
2689         * DumpRenderTree/chromium/WebThemeEngineDRTWin.cpp:
2690         (WebThemeEngineDRTWin::getSize):
2691         * DumpRenderTree/chromium/WebThemeEngineDRTWin.h:
2692         (WebThemeEngineDRTWin):
2693
2694 2012-06-25  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
2695
2696         [EFL] Implement Network Information API
2697         https://bugs.webkit.org/show_bug.cgi?id=87067
2698
2699         Reviewed by Kenneth Rohde Christiansen.
2700
2701         * efl/jhbuild.modules: Add eeze library dependency.
2702
2703 2012-06-25  Simon Hausmann  <simon.hausmann@nokia.com>
2704
2705         [Qt] Fix Win32 debug build
2706
2707         Reviewed by Tor Arne Vestbø.
2708
2709         Consolidate use_all_in_one_files into mac/unix/win32 default_pre.prf.
2710
2711         * qmake/mkspecs/features/unix/default_pre.prf:
2712         * qmake/mkspecs/features/win32/default_pre.prf: Added.
2713
2714 2012-06-24  Thiago Marcos P. Santos  <thiago.santos@intel.com>
2715
2716         [Qt] Enable <input type="color"> support by default
2717         https://bugs.webkit.org/show_bug.cgi?id=89653
2718
2719         Reviewed by Kenneth Rohde Christiansen.
2720
2721         This feature is complete for Qt WebKit2. Enabling by default
2722         will help to mature the implementation.
2723
2724         * Scripts/webkitperl/FeatureList.pm:
2725         * qmake/mkspecs/features/features.pri:
2726
2727 2012-06-24  Dan Bernstein  <mitz@apple.com>
2728
2729         Made debug-{minibrowser,safari,test-runner} work with LLDB.
2730
2731         Reviewed by Sam Weinig.
2732
2733         * Scripts/debug-minibrowser: Pass INCLUDE_OPTIONS_FOR_DEBUGGING to
2734         printHelpAndExitForRunAndDebugWebKitAppIfNeeded().
2735         * Scripts/debug-safari: Ditto.
2736         * Scripts/debug-test-runner: Ditto.
2737         * Scripts/webkitdirs.pm:
2738         (debugger): Added. Calls determineDebugger() if needed and returns the chosen debugger.
2739         (determineDebugger): Added. Sets the debugger to "lldb" if the --use-lldb switch is present,
2740         and to "gdb" otherwise.
2741         (printHelpAndExitForRunAndDebugWebKitAppIfNeeded): Changed to print help for the
2742         --target-web-process and --use-lldb switches if passed INCLUDE_OPTIONS_FOR_DEBUGGING.
2743         (execMacWebKitAppForDebugging): Changed to use the chosen debugger.
2744
2745 2012-06-24  Adam Barth  <abarth@webkit.org>
2746
2747         [Chromium] Release media resources after each LayoutTest on Android
2748         https://bugs.webkit.org/show_bug.cgi?id=89720
2749
2750         Reviewed by Eric Carlson.
2751
2752         The LayoutTests can hang when run on Android because we sometimes run
2753         out of media resources. This patch causes us to release our media
2754         resources after each test in order to avoid running out of this
2755         resource. In production, there are other mechanisms that manage this
2756         resource.
2757
2758         * DumpRenderTree/chromium/TestShell.cpp:
2759         (TestShell::resetTestController):
2760
2761 2012-06-23  Zan Dobersek  <zandobersek@gmail.com>
2762
2763         [Gtk] REGRESSION(r120918): Causes flaky DND tests
2764         https://bugs.webkit.org/show_bug.cgi?id=89770
2765
2766         Reviewed by Martin Robinson.
2767
2768         Connect to the run-file-chooser signal of the WebKitWebView and
2769         return TRUE so no file chooser dialog is run. This prevents certain
2770         test failures.
2771
2772         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2773         (webViewRunFileChooser):
2774         (createWebView):
2775
2776 2012-06-23  Dan Bernstein  <mitz@apple.com>
2777
2778         debug-{minibrowser,safari,test-runner} fail when gdb is not installed in /usr/bin.
2779
2780         Reviewed by Sam Weinig.
2781
2782         * Scripts/webkitdirs.pm:
2783         (execMacWebKitAppForDebugging): Use xcrun to locate gdb.
2784
2785 2012-06-23  Carlos Garcia Campos  <cgarcia@igalia.com>
2786
2787         Unreviewed. Skip GTK+ unit test /webkit2/WebKitFindController/hide.
2788
2789         It fails always when running it in Xvfb.
2790
2791         * gtk/run-api-tests:
2792         (TestRunner):
2793
2794 2012-06-23  Wajahat Siddiqui  <mdwajahatali.siddiqui@motorola.com>
2795
2796         [GTK][WK2]Mis-spelt defaultWindowTitle
2797         https://bugs.webkit.org/show_bug.cgi?id=89808
2798
2799         Reviewed by Carlos Garcia Campos.
2800
2801         * MiniBrowser/gtk/BrowserWindow.c:
2802
2803 2012-06-23  Sergio Villar Senin  <svillar@igalia.com>
2804
2805         [GTK] Add a new webkit2 tests slave bot
2806         https://bugs.webkit.org/show_bug.cgi?id=89336
2807
2808         Reviewed by Csaba Osztrogonác.
2809
2810         Slave configuration for a new GTK 64 bit release bot that will run
2811         WebKit2 tests.
2812
2813         Also BuildAndTest now accepts the "triggers" parametter,
2814         this means that apart from building and testing it can trigger
2815         some other build/test sequences in other slaves.
2816
2817         * BuildSlaveSupport/build.webkit.org-config/config.json:
2818         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
2819         (BuildAndTestFactory.__init__):
2820
2821 2012-06-23  Kwang Yul Seo  <skyul@company100.net>
2822
2823         Unreviewed. Update my email.
2824
2825         * Scripts/webkitpy/common/config/committers.py:
2826
2827 2012-06-22  Adam Barth  <abarth@webkit.org>
2828
2829         [Chromium] DumpRenderTree on Android needs to configure fonts for testing
2830         https://bugs.webkit.org/show_bug.cgi?id=89721
2831
2832         Reviewed by Nate Chapin.
2833
2834         This patch teaches DumpRenderTree to configure Skia to use the fallback
2835         fonts that the LayoutTests assume. This patch reduces the number result
2836         differences between chromium-android and chromium-linux.
2837
2838         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
2839         * DumpRenderTree/chromium/TestShellAndroid.cpp:
2840         (platformInit):
2841         * DumpRenderTree/chromium/android_fallback_fonts.xml: Added.
2842         * DumpRenderTree/chromium/android_main_fonts.xml: Added.
2843         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
2844
2845 2012-06-22  Alexandru Chiculita  <achicu@adobe.com>
2846
2847         [CSS Shaders] Re-enable the CSS Shaders compile time flag on Safari Mac
2848         https://bugs.webkit.org/show_bug.cgi?id=89781
2849
2850         Reviewed by Dean Jackson.
2851
2852         Added CSS Shaders as enabled by default on Safari for Mac.
2853
2854         * Scripts/webkitperl/FeatureList.pm:
2855
2856 2012-06-22  Adam Barth  <abarth@webkit.org>
2857
2858         Add support for test_expectations_android.txt for overriding test expecations on the chromium-android branch
2859         https://bugs.webkit.org/show_bug.cgi?id=89791
2860
2861         Reviewed by Dirk Pranke.
2862
2863         Downstream, the chromium-android port maintains a test expectations
2864         file to keep track of which tests pass or fail. That causes them to
2865         keep a patch in chromium_android.py in their branch.
2866
2867         This patch teaches chromium_android.py to use test_expectations_android.txt.
2868         We don't intend to use this file upstream, but having this code patch
2869         upstream lets us unfork webkitpy.
2870
2871         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
2872         (ChromiumAndroidPort.expectations_files):
2873
2874 2012-06-22  Dirk Pranke  <dpranke@chromium.org>
2875
2876         run_webkit_tests.py failed with AttributeError(''NoneType' object has no attribute 'pid'')
2877         https://bugs.webkit.org/show_bug.cgi?id=89734
2878
2879         Reviewed by Ryosuke Niwa.
2880
2881         Fix a crash in ServerProcess if you called .pid() after it
2882         crashed during a write(). We had a test for this case but the
2883         test wasn't calling pid(), just has_crashed(). Fixed the problem
2884         and the test.
2885
2886         * Scripts/webkitpy/layout_tests/port/server_process.py:
2887         (ServerProcess.__init__):
2888         (ServerProcess.pid):
2889         (ServerProcess._start):
2890         (ServerProcess.stop):
2891         * Scripts/webkitpy/layout_tests/port/server_process_unittest.py:
2892         (FakeServerProcess._start):
2893         (TestServerProcess.test_broken_pipe):
2894
2895 2012-06-22  Peter Beverloo  <peter@chromium.org>
2896
2897         [Chromium] Disable c++0x compatibility warnings in JavaScriptCore.gyp when building for Android
2898         https://bugs.webkit.org/show_bug.cgi?id=88853
2899
2900         Reviewed by Steve Block.
2901
2902         The Android exclusions were necessary to fix a gyp generation error, as
2903         the gcc_version variable wasn't being defined for Android. Remove these
2904         exceptions when Chromium is able to define the gcc_version variable.
2905
2906         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
2907
2908 2012-06-22  Mario Sanchez Prada  <msanchez@igalia.com>
2909
2910         Unreviewed gardening. Unskipping test passing after fixing bug 88419.
2911
2912         * gtk/run-api-tests:
2913         (TestRunner): Unskip WTF.HashMap.
2914
2915 2012-06-22  Zoltan Horvath  <zoltan@webkit.org>
2916
2917         [Qt] Allow DumpRenderTree to dump about:blank
2918         https://bugs.webkit.org/show_bug.cgi?id=89685
2919
2920         Reviewed by Ryosuke Niwa.
2921
2922         We need to allow DumpRenderTree to dump about:blank page, then it will be consistent with other ports,
2923         additionaly this behavior is required for running WTR performance tests.
2924
2925         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
2926         (WebCore::DumpRenderTree::processLine):
2927
2928 2012-06-21  Adam Barth  <abarth@webkit.org>
2929
2930         [Chromium] Reset mediaPlaybackRequiresUserGesture WebSettings after each test
2931         https://bugs.webkit.org/show_bug.cgi?id=89718
2932
2933         Reviewed by Kent Tamura.
2934
2935         We should reset this WebSetting to its default value so that it behaves
2936         consistently across tests.
2937
2938         * DumpRenderTree/chromium/WebPreferences.cpp:
2939         (WebPreferences::reset):
2940         (WebPreferences::applyTo):
2941         * DumpRenderTree/chromium/WebPreferences.h:
2942         (WebPreferences):
2943
2944 2012-06-21  Zoltan Horvath  <zoltan@webkit.org>
2945
2946         DRT/WTR python interface handles about:blank incorrectly
2947         https://bugs.webkit.org/show_bug.cgi?id=89563
2948
2949         Reviewed by Dirk Pranke.
2950
2951         Currently DRT/WTR python interface's _command_from_driver_input function
2952         puts the test directory path before about:blank. Remove this incorrect behavior.
2953
2954         * Scripts/webkitpy/layout_tests/port/webkit.py:
2955         (WebKitDriver._command_from_driver_input):
2956
2957 2012-06-21  Dirk Pranke  <dpranke@chromium.org>
2958
2959         reenable perf tests on win
2960         https://bugs.webkit.org/show_bug.cgi?id=89690
2961
2962         Reviewed by Ryosuke Niwa.
2963
2964         Only the replay tests don't work, and those are disabled by
2965         default. This change also fixes the undefined _log reference
2966         that was causing a crash.
2967
2968         * Scripts/run-perf-tests:
2969
2970 2012-06-21  Sheriff Bot  <webkit.review.bot@gmail.com>
2971
2972         Unreviewed, rolling out r120937.
2973         http://trac.webkit.org/changeset/120937
2974         https://bugs.webkit.org/show_bug.cgi?id=89679
2975
2976         This patch brought buildbot master down (Requested by svillar
2977         on #webkit).
2978
2979         * BuildSlaveSupport/build.webkit.org-config/config.json:
2980
2981 2012-06-21  Martin Robinson  <mrobinson@igalia.com>
2982
2983         [GTK] Combine WebKit API tests into fewer binaries
2984         https://bugs.webkit.org/show_bug.cgi?id=88458
2985
2986         Reviewed by Carlos Garcia Campos.
2987
2988         Instead of creating one binary per-test file, create binaries for each
2989         category of tests. Right now this includes WTF and the WebKit2 C API,
2990         but later tests can be added for the GTK+ platform layer and the WebKit1
2991         and WebKit2 API layers.
2992
2993         * TestWebKitAPI/GNUmakefile.am: Compile only two test binaries, one for the WebKit2
2994           C API and one for WTF. Refresh the source list to ensure that new test are active.
2995         * gtk/run-api-tests: Change the way that tests are skipped by splitting out the
2996         concept of skipping a test and skipping a suite (program) of tests. Test cases are
2997         skipped because of legitimate failures, but entire programs are skipped because of
2998         problems in the harness. As of right now a test program is only skipped if the
2999         accessibility bus cannot be started.
3000         (SkippedTest.__init__): Make the test case a required argument and have one skipped
3001         test case per SkippedTest instance.
3002         (SkippedTest.__str__): Ditto.
3003         (TestRunner): Update the directory list. We only have two gtest binaries now and they are
3004         in one directory. Reformat the test list to make it slightly easier to read.
3005         (TestRunner.__init__): Add the _skipped_test_program member, which handles entire
3006         test programs that are skipped.
3007         (TestRunner._setup_testing_environment): Use the new member.
3008         (TestRunner._test_cases_to_skip): Collect all skipped test cases now instead of just
3009         the first SkippedTest that matches.
3010         (TestRunner._should_run_test_program): Take a look at the new member to make this
3011         decision.
3012         (TestRunner._run_test_command): Use the name test_program instead of test
3013         to disambiguate between test cases and test suites.
3014         (TestRunner._run_test_glib): ditto.
3015         (TestRunner._run_test_google): Ditto.
3016         (TestRunner._run_test): Ditto.
3017         (TestRunner.run_tests): Ditto.
3018
3019 2012-06-21  Yong Li  <yoli@rim.com>
3020
3021         Unreviewed. Moving myself from committer to reviewer.
3022
3023         * Scripts/webkitpy/common/config/committers.py:
3024
3025 2012-06-21  Sergio Villar Senin  <svillar@igalia.com>
3026
3027         [GTK] Add a new webkit2 tests slave bot
3028         https://bugs.webkit.org/show_bug.cgi?id=89336
3029
3030         Reviewed by Philippe Normand.
3031
3032         Slave configuration for a new GTK bot that will run WebKit2 tests.
3033
3034         * BuildSlaveSupport/build.webkit.org-config/config.json:
3035
3036 2012-06-21  Thiago Marcos P. Santos  <thiago.santos@intel.com>
3037
3038         [Qt] Add a custom Color Chooser widget to MiniBrowser
3039         https://bugs.webkit.org/show_bug.cgi?id=87988
3040
3041         Reviewed by Alexis Menard.
3042
3043         Added custom color chooser example for the MiniBrowser.
3044
3045         * MiniBrowser/qt/MiniBrowser.qrc:
3046         * MiniBrowser/qt/qml/BrowserWindow.qml:
3047         * MiniBrowser/qt/qml/ColorChooser.qml: Added.
3048
3049 2012-06-21  Vivek Galatage  <vivekgalatage@gmail.com>
3050
3051         Remove the warning "File not found" in MiniBrowser post-build event
3052         https://bugs.webkit.org/show_bug.cgi?id=89601
3053
3054         Reviewed by Adam Roben.
3055
3056         * MiniBrowser/MiniBrowserPostBuild.cmd:
3057
3058 2012-06-21  Balazs Kelemen  <kbalazs@webkit.org>
3059
3060         [Qt] DRT in standalone mode hangs after the first test
3061         https://bugs.webkit.org/show_bug.cgi?id=89613
3062
3063         Reviewed by Csaba Osztrogonác.
3064
3065         Break an inline signal slot chain that ends up in setting
3066         LayoutTestController::m_hasDumped too early. In LayoutTestController::maybeDump()
3067         we emit done() which starts the chain that finishes in DumpRenderTree::open().
3068         The next line in maybeDump sets m_hasDumped to true. We could simply reorder
3069         these two lines but it seems to be wrong in general to start the next load
3070         from this call chain. The server mode (when DRT is runned by the test harness)
3071         also uses a queued connection to read the next test from stdin.
3072
3073         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
3074         (WebCore::DumpRenderTree::processArgsLine):
3075
3076 2012-06-20  Hans Wennborg  <hans@chromium.org>
3077
3078         Speech JavaScript API: SpeechRecognition should hook up with ActiveDOMObject more
3079         https://bugs.webkit.org/show_bug.cgi?id=89217
3080
3081         Reviewed by Adam Barth.
3082
3083         Add a method for checking whether the mock speech
3084         recognition was aborted.
3085
3086         Also redo the way the mock posts tasks. Instead of posting them all at once,
3087         maintain an internal queue of task objects, and call postTask() for them
3088         once at the time. This means that for example when the page is navigated
3089         away and abort() is called, that call doesn't end up after a bunch
3090         of previously posted events on the event loop.
3091
3092         * DumpRenderTree/chromium/LayoutTestController.cpp:
3093         (LayoutTestController::LayoutTestController):
3094         (LayoutTestController::wasMockSpeechRecognitionAborted):
3095         * DumpRenderTree/chromium/LayoutTestController.h:
3096         (LayoutTestController):
3097         * DumpRenderTree/chromium/MockWebSpeechRecognizer.cpp:
3098         (WebKit::ClientCallTask::ClientCallTask):
3099         (WebKit::ResultTask::ResultTask):
3100         (WebKit::NoMatchTask::NoMatchTask):
3101         (WebKit::ErrorTask::ErrorTask):
3102         (MockWebSpeechRecognizer::start):
3103         (MockWebSpeechRecognizer::abort):
3104         (MockWebSpeechRecognizer::setError):
3105         (MockWebSpeechRecognizer::MockWebSpeechRecognizer):
3106         (MockWebSpeechRecognizer::startTaskQueue):
3107         (MockWebSpeechRecognizer::StepTask::runIfValid):
3108         * DumpRenderTree/chromium/MockWebSpeechRecognizer.h:
3109         (MockWebSpeechRecognizer::hasBeenAborted):
3110         (MockWebSpeechRecognizer):
3111         (MockWebSpeechRecognizer::taskList):
3112         (Task):
3113         (MockWebSpeechRecognizer::Task::Task):
3114         (MockWebSpeechRecognizer::Task::~Task):
3115         (StepTask):
3116         (MockWebSpeechRecognizer::StepTask::StepTask):
3117
3118 2012-06-21  Mario Sanchez Prada  <msanchez@igalia.com>
3119
3120         [GTK] Add support for window.showModalDialog in WebKit2GTK+
3121         https://bugs.webkit.org/show_bug.cgi?id=79500
3122
3123         Reviewed by Carlos Garcia Campos.
3124
3125         Add support for modal dialogs in GTK's MiniBrowser.
3126
3127         * MiniBrowser/gtk/BrowserWindow.c:
3128         (webViewRunAsModal):
3129         (webViewCreate):
3130         (webViewDecidePolicy):
3131         (browser_window_new):
3132         * MiniBrowser/gtk/BrowserWindow.h:
3133         * MiniBrowser/gtk/main.c:
3134         (createBrowserWindow):
3135
3136 2012-06-20  Brent Fulgham  <bfulgham@webkit.org>
3137
3138         [WinCairo] Unreviewed build fix.
3139         The wrapper script that launched new- and old-run-webkit-tests
3140         was discarding the --wincairo argument needed to get the proper
3141         test infrastructure to build under WinCairo.
3142
3143         * Scripts/run-webkit-tests: Emulate the Qt, Wx, Chromium, etc.,
3144         behavior to chain the --wincairo flag through to the new build
3145         and test scripts.
3146
3147 2012-06-20  Mark Rowe  <mrowe@apple.com>
3148
3149         <http://webkit.org/b/89606> Teach run-safari and debug-safari to work with a Safari.app that has entitlements
3150
3151         Reviewed by Dan Bernstein.
3152
3153         * Scripts/webkitdirs.pm:
3154         (executableHasEntitlements):
3155         (safariPathFromSafariBundle):
3156
3157 2012-06-20  Sheriff Bot  <webkit.review.bot@gmail.com>
3158
3159         Unreviewed, rolling out r120821.
3160         http://trac.webkit.org/changeset/120821
3161         https://bugs.webkit.org/show_bug.cgi?id=89605
3162
3163         It made duplicated reviewer entries (Requested by Ossy on
3164         #webkit).
3165
3166         * Scripts/webkitpy/common/checkout/changelog.py:
3167         (ChangeLog.set_reviewer):
3168         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
3169         (test_set_reviewer):
3170         (test_set_short_description_and_bug_url):
3171
3172 2012-06-20  Dirk Pranke  <dpranke@chromium.org>
3173
3174         nrwt outputs empty files for wdiff output if wdiff is not installed
3175         https://bugs.webkit.org/show_bug.cgi?id=88709
3176
3177         Reviewed by Tony Chang.
3178
3179         Reviewed by Tony Chang.
3180
3181         Don't write -wdiff or -pretty.html files if wdiff or prettypatch
3182         aren't available, and clean up the handling for them in the port
3183         code.
3184
3185         * Scripts/webkitpy/run_webkit_tests_integrationtest.py:
3186         (MainTest.test_output_diffs):
3187         * Scripts/webkitpy/layout_tests/controllers/test_result_writer.py:
3188         (TestResultWriter.create_text_diff_and_write_result):
3189         * Scripts/webkitpy/layout_tests/port/base.py:
3190         (Port.check_pretty_patch):
3191         (Port.check_wdiff):
3192         (Port._wdiff_missing_message):
3193         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
3194         (ChromiumLinuxPort.check_build):
3195         (ChromiumLinuxPort._wdiff_missing_message):
3196         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
3197         (ChromiumMacPort.check_build):
3198         (ChromiumMacPort):
3199         (ChromiumMacPort._wdiff_missing_message):
3200         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
3201         (ChromiumMacPortTest):
3202
3203 2012-06-20  Dirk Pranke  <dpranke@chromium.org>
3204
3205         nrwt: fix unit tests for ensuring svn revision is correct
3206         https://bugs.webkit.org/show_bug.cgi?id=89498
3207
3208         Reviewed by Eric Seidel.
3209
3210         Reviewed by Eric Seidel.
3211
3212         Fix the unit tests for testing that we are embedding the SVN
3213         revision in the results json for NRWT properly; this was broken
3214         in r120646 but because the MockHost worked differently than the
3215         real one (by always initializing the MockSCM object) we didn't
3216         notice. Unfortunately, just changing the default breaks all
3217         sorts of unit tests ...
3218
3219         * Scripts/webkitpy/common/host_mock.py:
3220         (MockHost.__init__):
3221         (MockHost._initialize_scm):
3222         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
3223         (ResultSummaryTest.test_no_svn_revision):
3224         (ResultSummaryTest.test_svn_revision):
3225
3226 2012-06-20  Dirk Pranke  <dpranke@chromium.org>
3227
3228         new-run-webkit-tests appends "/Debug" or "/Release" to $WEBKITOUTPUTDIR
3229         https://bugs.webkit.org/show_bug.cgi?id=69360
3230
3231         Reviewed by Eric Seidel.
3232
3233         Reviewed by Eric Seidel.
3234
3235         Propagate the 'port_implementation' part of the platform (i.e.,
3236         gtk,qt,chromium) to webkit-build-directory so that we can pick
3237         up the gtk-specific handling of WEBKITOUTPUTDIR ...
3238
3239         I didn't write any additional tests for this; testing it
3240         properly is an integration test between the python code and the
3241         perl code, which I verified by hand.
3242
3243         * Scripts/webkitpy/layout_tests/port/base.py:
3244         (Port.__init__):
3245         * Scripts/webkitpy/layout_tests/port/config.py:
3246         (Config.__init__):
3247         (Config.build_directory):
3248         * Scripts/webkitpy/layout_tests/port/config_mock.py:
3249         (MockConfig.__init__):
3250         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
3251         (ConfigTest.test_build_directory_passes_port_implementation):
3252
3253 2012-06-20  Dirk Pranke  <dpranke@chromium.org>
3254
3255         Fix import sorting missed in r120846
3256
3257         Reviewed by Tony Chang.
3258
3259         * Scripts/webkitpy/layout_tests/servers/apache_http_server_unittest.py:
3260
3261 2012-06-20  Dirk Pranke  <dpranke@chromium.org>
3262
3263         Fix chromium win http servers after breakage introduced in r120846.
3264
3265         Unreviewed, build fix.
3266
3267         * Scripts/webkitpy/layout_tests/port/base.py:
3268         (Port.to.start_http_server):
3269         * Scripts/webkitpy/layout_tests/servers/http_server.py:
3270         (Lighttpd.__init__):
3271
3272 2012-06-20  Dirk Pranke  <dpranke@chromium.org>
3273
3274         tweak output of webkit-patch print-{baselines,expectations}
3275         https://bugs.webkit.org/show_bug.cgi?id=89588
3276
3277         Reviewed by Tony Chang.
3278
3279         Minor tweaks to the output to make it more readable, including
3280         turning off csv by default for print-baselines with multiple ports.
3281
3282         * Scripts/webkitpy/layout_tests/port/test.py:
3283         (TestPort):
3284         * Scripts/webkitpy/tool/commands/queries.py:
3285         (PrintExpectations.execute):
3286         (PrintBaselines.execute):
3287         * Scripts/webkitpy/tool/commands/queries_unittest.py:
3288         (PrintExpectationsTest.run_test):
3289         (PrintExpectationsTest.test_multiple):
3290         (PrintBaselinesTest.setUp):
3291         (PrintBaselinesTest.test_multiple):
3292
3293 2012-06-18  Dirk Pranke  <dpranke@chromium.org>
3294
3295         new-run-webkit-tests should spin-up enough httpd processes to avoid timeouts
3296         https://bugs.webkit.org/show_bug.cgi?id=88134
3297
3298         Reviewed by Tony Chang.
3299
3300         Change NRWT to spin up 2*min(child_processes, locked_shards)
3301         http servers by default so that we are less likely to get a
3302         bunch of http timeouts at the beginning of a test run.
3303
3304         Note that I had to tweak executive_mock to support mocked stderr
3305         because the apache_http_server code reads stderr when starting a
3306         process to ensure it started okay.
3307
3308         * Scripts/webkitpy/common/system/executive_mock.py:
3309         (MockProcess.__init__):
3310         * Scripts/webkitpy/common/system/outputcapture.py:
3311         * Scripts/webkitpy/layout_tests/controllers/manager.py:
3312         (Manager._run_tests):
3313         (Manager.start_servers_with_lock):
3314         * Scripts/webkitpy/layout_tests/port/base.py:
3315         (Port.to.start_http_server):
3316         * Scripts/webkitpy/layout_tests/port/test.py:
3317         (TestPort.start_http_server):
3318         (TestPort._path_to_apache):
3319         (TestPort):
3320         (TestPort._path_to_apache_config_file):
3321         * Scripts/webkitpy/layout_tests/servers/apache_http_server.py:
3322         (LayoutTestApacheHttpd.__init__):
3323         * Scripts/webkitpy/layout_tests/servers/apache_http_server_unittest.py: Added
3324         (LayoutTestApacheHttpd.__init__):
3325         * Scripts/webkitpy/layout_tests/servers/http_server.py:
3326         (Lighttpd.__init__):
3327         * Scripts/webkitpy/layout_tests/servers/http_server_base.py:
3328         (HttpServerBase.__init__):
3329
3330 2012-06-20  Tom Zakrajsek  <tomz@codeaurora.org>
3331
3332         Unreviewed: Back out accidentally checked in debug print which broke a test
3333
3334         * Scripts/webkitpy/performance_tests/perftest.py:
3335         (PerfTest.parse_output):
3336
3337 2012-06-20  Tom Zakrajsek  <tomz@codeaurora.org>
3338
3339         Import themaninblue.com/experiment/AnimationBenchmark/ as performance tests
3340         https://bugs.webkit.org/show_bug.cgi?id=78789
3341
3342         Reviewed by Ryosuke Niwa.
3343
3344         Updated parser to include fps as a valid unit.
3345
3346         * Scripts/webkitpy/performance_tests/perftest.py:
3347         (PerfTest):
3348         (PerfTest.parse_output):
3349
3350 2012-06-20  Balazs Ankes  <bank@inf.u-szeged.hu>
3351
3352         webkit-patch should add reviewer if "Reviewed by NOBODY ..." is missing
3353         https://bugs.webkit.org/show_bug.cgi?id=67935
3354
3355         Reviewed by Ryosuke Niwa.
3356
3357         * Scripts/webkitpy/common/checkout/changelog.py:
3358         (ChangeLog.set_reviewer):
3359         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
3360         (test_set_reviewer):
3361         (test_set_short_description_and_bug_url):
3362
3363 2012-06-20  Simon Hausmann  <simon.hausmann@nokia.com>
3364
3365         Unreviewed trivial permissions fix.
3366
3367         * qmake/mkspecs/features/features.prf: Don't mark this file as executable.
3368
3369 2012-06-19  Ryuan Choi  <ryuan.choi@samsung.com>
3370
3371         [EFL][Regression] Build break after r120786
3372         https://bugs.webkit.org/show_bug.cgi?id=89549
3373
3374         Unreviewed build fix for Efl port.
3375
3376         * DumpRenderTree/efl/CMakeLists.txt: Added disassembler in includes.
3377
3378 2012-06-19  Dirk Pranke  <dpranke@chromium.org>
3379
3380         NRWT spins up and down the WebSocket server when running a single HTTP test from the command line
3381         https://bugs.webkit.org/show_bug.cgi?id=64489
3382
3383         Reviewed by Tony Chang.
3384
3385         This change changes NRWT so that we don't start the websocket
3386         server if we aren't running websocket tests and don't start the
3387         HTTP server if we aren't running http tests.
3388         
3389         * Scripts/webkitpy/layout_tests/controllers/manager.py:
3390         (Manager._websocket_tests):
3391         (Manager.start_servers_with_lock):
3392         (Manager.stop_servers_with_lock):
3393         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
3394         (ManagerTest.test_servers_started):
3395
3396 2012-06-19  Dirk Pranke  <dpranke@chromium.org>
3397
3398         test-webkitpy: add a -p flag to pass through captured output to enable debugging
3399         https://bugs.webkit.org/show_bug.cgi?id=89158
3400
3401         Reviewed by Eric Seidel.
3402
3403         Normally when outputcapture runs it intercepts stdout/stderr,
3404         which makes running under the debugger difficult and mostly
3405         pointless (since you can't see any output). This change adds a
3406         flag to test-webkitpy (-p) that will cause outputcapture to pass
3407         through the output as well as capture it.
3408
3409         * Scripts/webkitpy/common/system/outputcapture.py:
3410         (OutputCapture.stream_wrapper):
3411         (OutputCapture._capture_output_with_name):
3412         * Scripts/webkitpy/common/system/outputcapture_unittest.py:
3413         (OutputCaptureTest.setUp):
3414         * Scripts/webkitpy/test/main.py:
3415         (Tester._parse_args):
3416         (Tester._run_tests):
3417         (Tester._log_exception):
3418         (_CaptureAndPassThroughStream):
3419         (_CaptureAndPassThroughStream.__init__):
3420         (_CaptureAndPassThroughStream.write):
3421         (_CaptureAndPassThroughStream._message_is_from_pdb):
3422         (_CaptureAndPassThroughStream.flush):
3423         (_CaptureAndPassThroughStream.getvalue):
3424
3425 2012-06-19  Adam Barth  <abarth@webkit.org>
3426
3427         garden-o-matic fails to rebaseline tests with MISSING results
3428         https://bugs.webkit.org/show_bug.cgi?id=89521
3429
3430         Reviewed by Kenneth Russell.
3431
3432         If the bots report "MISSING", we should try to find PNG and TXT results on the server.
3433
3434         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/results.js:
3435
3436 2012-06-19  Adam Barth  <abarth@webkit.org>
3437
3438         svn.delete_list fails to delete empty parent directories
3439         https://bugs.webkit.org/show_bug.cgi?id=89520
3440
3441         Reviewed by Dirk Pranke.
3442
3443         These functions were calling scm.delete and scm.add, which already
3444         delete the parent directory. We'd get an exception when we tried to
3445         delete the parent directory again.
3446
3447         * Scripts/webkitpy/common/checkout/scm/svn.py:
3448         (SVN._add_parent_directories):
3449         (SVN._delete_parent_directories):
3450
3451 2012-06-19  Thiago Marcos P. Santos  <thiago.santos@intel.com>
3452
3453         webkitpy: Simplify fetch_bugs_matching_search()
3454         https://bugs.webkit.org/show_bug.cgi?id=89497
3455
3456         Reviewed by Adam Barth.
3457
3458         Removed parameter with default value. It's not being used
3459         and the handling of the parameter was wrong. Saves a check.
3460
3461         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
3462         (BugzillaQueries.fetch_bugs_matching_search):
3463         * Scripts/webkitpy/common/net/bugzilla/bugzilla_mock.py:
3464         (MockBugzillaQueries.fetch_bugs_matching_search):
3465
3466 2012-06-19  Dirk Pranke  <dpranke@chromium.org>
3467
3468         webkitpy: remove unneeded imports in layout_tests/port/chromium.py
3469         https://bugs.webkit.org/show_bug.cgi?id=89522
3470
3471         Reviewed by Eric Seidel.
3472
3473         They were no longer being used. No tests necessary.
3474
3475         * Scripts/webkitpy/layout_tests/port/chromium.py:
3476
3477 2012-06-19  Adam Barth  <abarth@webkit.org>
3478
3479         Missing results aren't shown in garden-o-matic
3480
3481         Reviewed by Kenneth Russell.
3482
3483         Previously we didn't know to look for text and image results when the
3484         bot told us that results were missing.
3485
3486         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/results.js:
3487
3488 2012-06-19  Dirk Pranke  <dpranke@chromium.org>
3489
3490         Fix regression introduced in r120646 where the svn revision isn't saved in results.json
3491
3492         Unreviewed, build fix.
3493
3494         It turns out that we store the svn revision two different ways
3495         in the JSON, and I only caught one of them :(. This should fix
3496         the other.
3497
3498         Also, this part of the code is poorly tested. I will investigate
3499         adding at least a basic test for this in a separate fix.
3500
3501         * Scripts/webkitpy/layout_tests/controllers/manager.py:
3502         (summarize_results):
3503
3504 2012-06-19  Dirk Pranke  <dpranke@chromium.org>
3505
3506         REGRESSION (NRWT): Results for new non-text-only tests are always put in the most-specific platform directory
3507         https://bugs.webkit.org/show_bug.cgi?id=78127
3508
3509         Reviewed by Ryosuke Niwa.
3510
3511         Add the --add-platform-exceptions flag from ORWT, and ensure
3512         that --new-baseline is equivalent to --reset-results
3513         --add-platform-exceptions. Also fix the default logic for where to
3514         put PNGs and render tree dumps if --new-test-results is true:
3515         if --add-platform-exceptions is False, baselines should go in
3516         the least-specific platform directory (e.g., platform/mac)
3517         rather than the most-specific (platform/mac-snowleopard).
3518
3519         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py:
3520         (SingleTestRunner):
3521         (SingleTestRunner.run):
3522         (SingleTestRunner._run_rebaseline):
3523         (SingleTestRunner._add_missing_baselines):
3524         (SingleTestRunner._location_for_new_baseline):
3525         (SingleTestRunner._overwrite_baselines):
3526         (SingleTestRunner._save_baseline_data):
3527         * Scripts/webkitpy/layout_tests/port/base.py:
3528         (Port.baseline_path):
3529         (Port):
3530         (Port.baseline_platform_dir):
3531         (Port.baseline_version_dir):
3532         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3533         (_set_up_derived_options):
3534         (parse_args):
3535         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
3536         (RebaselineTest.assertBaselines):
3537         (RebaselineTest.test_reset_results):
3538         (RebaselineTest.test_missing_results):
3539         (RebaselineTest.test_new_baseline):
3540
3541 2012-06-19  Dirk Pranke  <dpranke@chromium.org>
3542
3543         NRWT should not take memory used as disk cache into account when deciding how many processes to launch
3544         https://bugs.webkit.org/show_bug.cgi?id=81379
3545
3546         Reviewed by Eric Seidel.
3547
3548         The 'free memory' calculation we were running on the mac seemed
3549         to underestimate how many children we can run in parallel, and
3550         it was complex. This patch replaces that calculation with a
3551         simpler one that reserves 2GB for overhead and assumes 256MB per
3552         DRT/WTR, so if we had 4GB of RAM we can run up to 8 DRTs.
3553
3554         Also, there was a bug where we were truncating the memory
3555         installed on the machine to 4GB by casting to an int instead of
3556         a long; this was probably the source of some of the earlier
3557         problems when using total memory.
3558
3559         This patch also removes the no-longer-needed restrictions on the
3560         number of workers on beefy Lion boxes for both Apple Mac and Chromium
3561         Mac; we should now use all of the cores by default.
3562
3563         The memory calculations have only been implemented on the mac;
3564         having the calculation in base.default_child_processes() was IMO
3565         misleading, and so this patch also moves the computation into
3566         the MacPort. I have not heard of the # of workers being an issue
3567         on any other ports, so this should be fine.
3568
3569         * Scripts/webkitpy/common/system/platforminfo.py:
3570         (PlatformInfo.total_bytes_memory):
3571         * Scripts/webkitpy/common/system/platforminfo_mock.py:
3572         (PlatformInfo.total_bytes_memory):
3573         (PlatformInfo.total_bytes_memory):
3574         (PlatformInfo._win_version_tuple_from_cmd):
3575         * Scripts/webkitpy/common/system/platforminfo_unittest.py:
3576         (TestPlatformInfo.test_total_bytes_memory):
3577         * Scripts/webkitpy/layout_tests/port/base.py:
3578         (Port.default_child_processes):
3579         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
3580         (PortTest.test_default_child_processes):
3581         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
3582         (ChromiumMacPort.operating_system):
3583         * Scripts/webkitpy/layout_tests/port/mac.py:
3584         (MacPort.default_child_processes):
3585         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
3586         (TestMacPort.test_default_child_processes):
3587
3588 2012-06-19  Dirk Pranke  <dpranke@chromium.org>
3589
3590         new-run-webkit-tests reports unexpected pass of pixel tests when pixel testing is disabled
3591         https://bugs.webkit.org/show_bug.cgi?id=85446
3592
3593         Reviewed by Simon Fraser.
3594
3595         Embed whether pixel testing was enabled into the results.json.
3596
3597         * Scripts/webkitpy/layout_tests/controllers/manager.py:
3598         (summarize_results):
3599
3600 2012-06-19  Zoltan Horvath  <zoltan@webkit.org>
3601
3602         [Qt] Modify HTTPS port to 8443 for performance tests
3603         https://bugs.webkit.org/show_bug.cgi?id=89442
3604
3605         Reviewed by Ryosuke Niwa.
3606
3607         We should be consistent with the default HTTPS port and with the changelog of r119188.
3608
3609         * Scripts/webkitpy/performance_tests/perftest.py:
3610         (ReplayServer.__init__):
3611
3612 2012-06-19  Mike West  <mkwst@chromium.org>
3613
3614         Introduce ENABLE_CSP_NEXT configuration flag.
3615         https://bugs.webkit.org/show_bug.cgi?id=89300
3616
3617         Reviewed by Adam Barth.
3618
3619         The 1.0 draft of the Content Security Policy spec is just about to
3620         move to Last Call. We'll hide work on the upcoming 1.1 spec behind
3621         this ENABLE flag, disabled by default.
3622
3623         Spec: https://dvcs.w3.org/hg/content-security-policy/raw-file/tip/csp-specification.dev.html
3624
3625         * Scripts/webkitperl/FeatureList.pm:
3626         * qmake/mkspecs/features/features.pri:
3627
3628 2012-06-19  Christophe Dumez  <christophe.dumez@intel.com>
3629
3630         [EFL] EFL's LayoutTestController does not support setTextDirection
3631         https://bugs.webkit.org/show_bug.cgi?id=87481
3632
3633         Reviewed by Hajime Morita.
3634
3635         Implement setTextDirection in EFL's LayoutTestController and
3636         properly reset its value between test cases to avoid flakiness.
3637
3638         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
3639         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
3640         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp:
3641         (LayoutTestController::setTextDirection):
3642
3643 2012-06-19  Kristóf Kosztyó  <kkristof@inf.u-szeged.hu>
3644
3645         [Qt] Temporarily disable xvfb driver in nrwt
3646         https://bugs.webkit.org/show_bug.cgi?id=88414
3647
3648         Reviewed by Csaba Osztrogonác.
3649
3650         * Scripts/webkitpy/layout_tests/port/qt.py:
3651         (QtPort._build_driver):
3652
3653 2012-06-18  Joone Hur  <joone.hur@intel.com>
3654
3655         [EFL] Fontconfig can't be linked properly
3656         https://bugs.webkit.org/show_bug.cgi?id=89418
3657
3658         Unreviewed build fix.
3659
3660         Make Fontconfig be linked correctly.
3661
3662         * MiniBrowser/efl/CMakeLists.txt: add FONTCONFIG_LIBRARIES to MiniBrowser_LIBRARIES.
3663
3664 2012-06-18  Joone Hur  <joone.hur@intel.com>
3665
3666         Unreviewed. Updating email for committers.py script.
3667
3668         * Scripts/webkitpy/common/config/committers.py:
3669
3670 2012-06-18  Simon Fraser  <simon.fraser@apple.com>
3671
3672         Fix a typo that caused TestFailures to have a syntax error.
3673         
3674         Sadly TestFailures is broken for another reason now too:
3675         https://bugs.webkit.org/show_bug.cgi?id=89419
3676
3677         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ViewController.js:
3678         (ViewController.prototype._domForBuildName):
3679
3680 2012-06-18  Simon Fraser  <simon.fraser@apple.com>
3681
3682         Provide bug links for suspicious commits
3683         https://bugs.webkit.org/show_bug.cgi?id=89408
3684         
3685         Reviewed by Adam Barth.
3686
3687         Add links to bugs for the suspicious commits if we have bug data.
3688
3689         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/Bugzilla.js:
3690         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/notifications.js:
3691         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/styles/notifications.css:
3692         (ol.notifications>li ul.causes>li>div.description>span>span.bugID::before):
3693         (ol.notifications>li ul.causes>li>div.description>span>span.bugID>a):
3694
3695 2012-06-18  Amy Ousterhout  <aousterh@chromium.org>
3696
3697         [Chromium] DeviceOrientation Cleanup
3698         https://bugs.webkit.org/show_bug.cgi?id=89354
3699
3700         Reviewed by Kent Tamura.
3701
3702         * DumpRenderTree/chromium/WebViewHost.h: added OVERRIDE specifier for virtual function
3703         (WebViewHost):
3704
3705 2012-06-18  James Robinson  <jamesr@chromium.org>
3706
3707         [chromium] Stop passing deprecated 'direct' parameter to webkit_support::CreateGraphicsContext3D
3708         https://bugs.webkit.org/show_bug.cgi?id=89254
3709
3710         Reviewed by Adrienne Walker.
3711
3712         This parameter doesn't mean anything since this codepath is only used for onscreen contexts and is deprecated
3713         upstream.
3714
3715         * DumpRenderTree/chromium/WebViewHost.cpp:
3716         (WebViewHost::createGraphicsContext3D):
3717
3718 2012-06-18  Simon Fraser  <simon.fraser@apple.com>
3719
3720         Minor fix suggested by Adam Barth.
3721
3722         Use $(this._what).empty(); rather than removing children one at a time.
3723         For some reason removeAllChildren() is undefined on this node.
3724         
3725         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/notifications.js:
3726
3727 2012-06-18  Simon Fraser  <simon.fraser@apple.com>
3728
3729         Show TOT revision in garden-o-matic, and make the revisions into links
3730         https://bugs.webkit.org/show_bug.cgi?id=89396
3731
3732         Reviewed by Adam Barth.
3733
3734         The "latest revision" notification now reads "Latest revision processed by every bot: NN (trunk is at NN)"
3735         where the revisions are links to trac.
3736         
3737         To avoid setting innerHTML, I made base.createLinkNode() and used it in a bunch of places, which had
3738         the knock-on effect of flipping the target and href attributes for anchors in some test output.
3739
3740         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ViewController.js:
3741         (ViewController.prototype._displayTesters.updateList.list):
3742         (ViewController.prototype._displayTesters.updateList):
3743         (ViewController.prototype._domForRegressionRange.trac.commitDataForRevisionRange):
3744         (ViewController.prototype._domForRegressionRange):
3745         (ViewController.prototype._domForAuxiliaryUIElements):
3746         (ViewController.prototype._domForBuildName):
3747         (ViewController.prototype):
3748         (ViewController.prototype.):
3749         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/base.js:
3750         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/garden-o-matic.js:
3751         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/model.js:
3752         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/notifications.js:
3753         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/notifications_unittests.js:
3754