[Qt] Fix build with latest Qt 5
[WebKit-https.git] / Tools / ChangeLog
1 2012-09-24  Simon Hausmann  <simon.hausmann@digia.com>
2
3         [Qt] Fix build with latest Qt 5
4
5         Reviewed by Tor Arne Vestbø.
6
7         The variables for syncqt handling are now set in api.pri, but let's add a comment here that we can
8         remove them from default_pre.prf once we're past Qt 5 beta2.
9
10         * qmake/mkspecs/features/default_pre.prf:
11
12 2012-09-22  Dominic Mazzoni  <dmazzoni@google.com>
13
14         AX: Layout tests would be easier to write if AccessibilityController could find an element by id
15         https://bugs.webkit.org/show_bug.cgi?id=97301
16
17         Reviewed by Chris Fleizach.
18
19         Adds a new accessibleElementById method to AccessibilityController, to make it
20         easier to write cross-platform layout tests.
21
22         Implemented for Chromium and for Mac (DRT and WKTR).
23
24         * DumpRenderTree/AccessibilityController.cpp:
25         (getAccessibleElementByIdCallback):
26         (AccessibilityController::getJSClass):
27         * DumpRenderTree/AccessibilityController.h:
28         (AccessibilityController):
29         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
30         * DumpRenderTree/chromium/TestRunner/AccessibilityControllerChromium.cpp:
31         (AccessibilityController::AccessibilityController):
32         (AccessibilityController::findAccessibleElementByIdRecursive):
33         (AccessibilityController::getAccessibleElementById):
34         (AccessibilityController::accessibleElementByIdGetterCallback):
35         * DumpRenderTree/chromium/TestRunner/AccessibilityControllerChromium.h:
36         (AccessibilityController):
37         * DumpRenderTree/gtk/AccessibilityControllerGtk.cpp:
38         (AccessibilityController::accessibleElementById):
39         * DumpRenderTree/mac/AccessibilityCommonMac.h: added
40         * DumpRenderTree/mac/AccessibilityCommonMac.mm: added
41         (+[NSString stringWithJSStringRef:]):
42         (-[NSString createJSStringRef]):
43         * DumpRenderTree/mac/AccessibilityControllerMac.mm:
44         (findAccessibleObjectById):
45         (AccessibilityController::accessibleElementById):
46         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
47         * DumpRenderTree/win/AccessibilityControllerWin.cpp:
48         (AccessibilityController::accessibleElementById):
49         * WebKitTestRunner/InjectedBundle/AccessibilityController.cpp:
50         (WTR::AccessibilityController::accessibleElementById):
51         * WebKitTestRunner/InjectedBundle/AccessibilityController.h:
52         (AccessibilityController):
53         * WebKitTestRunner/InjectedBundle/Bindings/AccessibilityController.idl:
54         * WebKitTestRunner/InjectedBundle/mac/AccessibilityCommonMac.h: added
55         * WebKitTestRunner/InjectedBundle/mac/AccessibilityCommonMac.mm: added
56         (+[NSString stringWithJSStringRef:]):
57         (-[NSString createJSStringRef]):
58         * WebKitTestRunner/InjectedBundle/mac/AccessibilityControllerMac.mm:
59         (WTR::findAccessibleObjectById):
60         (WTR):
61         (WTR::AccessibilityController::accessibleElementById):
62         * WebKitTestRunner/InjectedBundle/mac/AccessibilityUIElementMac.mm:
63         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
64
65 2012-09-22  Zan Dobersek  <zandobersek@gmail.com>
66
67         Unexpected reftest passes are only reported when pixel testing is enabled
68         https://bugs.webkit.org/show_bug.cgi?id=97242
69
70         Reviewed by Dirk Pranke.
71
72         LayoutTestRunner should know whether the finished test it's handling is
73         a reftest. This is necessary when updating result summary for reftests
74         when pixel testing is not enabled - in these circumstances an unexpectedly
75         passing reftest was not reported as such due to pixel testing then being
76         determined by searching for reftest failures in failures list, which were
77         not found in an empty failure list of a passing test. Now the TestResult
78         objects have a member variable indicating the test is a reftest whenever
79         that's the case.
80
81         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner.py:
82         (LayoutTestRunner._update_summary_with_result):
83         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner_unittest.py:
84         (LayoutTestRunnerTests.test_update_summary_with_result):
85         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py:
86         (SingleTestRunner.run):
87         (SingleTestRunner._run_reftest):
88         * Scripts/webkitpy/layout_tests/models/test_results.py:
89         (TestResult.__init__):
90
91 2012-09-21  Sam Weinig  <sam@webkit.org>
92
93         WebProcess XPC services need have their environment set without disrupting all other XPC services
94         https://bugs.webkit.org/show_bug.cgi?id=95161
95
96         Reviewed by Anders Carlsson.
97
98         Remove setting the __XPC_* environment variables.
99
100         * Scripts/webkitdirs.pm:
101         (setUpGuardMallocIfNeeded):
102         (runMacWebKitApp):
103         (execMacWebKitAppForDebugging):
104
105 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
106
107         Fix typo in additional-platform-directory patch just landed
108         https://bugs.webkit.org/show_bug.cgi?id=97380
109         
110         Unreviewed, build fix.
111
112         * Scripts/webkitpy/layout_tests/port/base.py:
113         (Port.relative_test_filename):
114         (Port.relative_perf_test_filename):
115
116 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
117
118         nrwt: don't require additional-platform-directory to be an abspath or live under LayoutTests
119         https://bugs.webkit.org/show_bug.cgi?id=97380
120
121         Reviewed by Ojan Vafai.
122
123         There doesn't seem to be a good reason for this restriction and
124         it's useful to be able to point to directories outside the
125         checkout for results (e.g., for local failures due to a 10.7.4
126         install ;).
127
128         * Scripts/webkitpy/layout_tests/port/base.py:
129         (Port.relative_test_filename):
130         (Port.relative_perf_test_filename):
131         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
132         (ChromiumAndroidDriver._command_from_driver_input):
133         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
134         (_set_up_derived_options):
135         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
136         (MainTest.test_additional_platform_directory):
137
138 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
139
140         webkitpy: drop support for old TestExpectations syntax
141         https://bugs.webkit.org/show_bug.cgi?id=97364
142
143         Reviewed by Ryosuke Niwa.
144
145         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
146         (TestExpectationParser._collect_matching_tests):
147         (TestExpectationParser):
148         (TestExpectationParser._tokenize_line):
149         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
150         (SkippedTests.test_skipped_entry_dont_exist):
151         (ExpectationSyntaxTests.assert_tokenize_exp):
152         (ExpectationSyntaxTests.test_bare_name):
153         (ExpectationSyntaxTests.test_bare_name_and_bugs):
154         (ExpectationSyntaxTests.test_comments):
155         (ExpectationSyntaxTests.test_config_modifiers):
156         (ExpectationSyntaxTests.test_unknown_config):
157         (ExpectationSyntaxTests.test_unknown_expectation):
158         (ExpectationSyntaxTests.test_skip):
159         (ExpectationSyntaxTests.test_slow):
160         (ExpectationSyntaxTests.test_wontfix):
161         (ExpectationSyntaxTests.test_blank_line):
162         (ExpectationSyntaxTests.test_warnings):
163         (RebaseliningTest.test_no_get_rebaselining_failures):
164
165 2012-09-21  Simon Fraser  <simon.fraser@apple.com>
166
167         Improve WTR unresponsiveness output a little
168         https://bugs.webkit.org/show_bug.cgi?id=97370
169
170         Reviewed by Timothy Horton.
171
172         Distinguish between conditions that already set the errorMessage,
173         and unresponsiveness due to slow about:blank loads when WTR
174         reports unresponsiveness.
175
176         * WebKitTestRunner/TestInvocation.cpp:
177         (WTR::TestInvocation::invoke):
178
179 2012-09-21  Chris Rogers  <crogers@google.com>
180
181         Add Web Audio support for deprecated/legacy APIs
182         https://bugs.webkit.org/show_bug.cgi?id=97050
183
184         Reviewed by Eric Carlson.
185
186         * Scripts/webkitperl/FeatureList.pm:
187
188 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
189
190         webkitpy: update remaining tests to use the new expectation syntax
191         https://bugs.webkit.org/show_bug.cgi?id=97362
192
193         Reviewed by Ojan Vafai.
194
195         This patch updates all the unit tests that were still using the
196         old TestExpectations syntax to use the new syntax *except* for
197         the tests that were specifically testing that we parsed the old
198         syntax correctly.
199
200         Also, a block of tests for the new syntax were duplicated, so
201         I've deleted the duplicate.
202
203         Note that the old syntax is still supported so this change should
204         produce no visible changes.
205
206         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
207         (MiscTests.test_multiple_results):
208         (MiscTests.test_category_expectations):
209         (MiscTests.test_error_on_different_platform):
210         (MiscTests.test_error_on_different_build_type):
211         (MiscTests.test_overrides):
212         (MiscTests.test_overrides__directory):
213         (MiscTests.test_overrides__duplicate):
214         (MiscTests.test_more_specific_override_resets_skip):
215         (SkippedTests.check):
216         (SkippedTests.test_duplicate_skipped_test_fails_lint):
217         (SkippedTests.test_skipped_file_overrides_expectations):
218         (SkippedTests.test_skipped_dir_overrides_expectations):
219         (SkippedTests.test_skipped_file_overrides_overrides):
220         (SkippedTests.test_skipped_dir_overrides_overrides):
221         (ExpectationSyntaxTests.disabled_test_missing_expectation):
222         (ExpectationSyntaxTests.disabled_test_missing_colon):
223         (ExpectationSyntaxTests.disabled_test_too_many_colons):
224         (ExpectationSyntaxTests.disabled_test_too_many_equals_signs):
225         (ExpectationSyntaxTests):
226         (ExpectationSyntaxTests.test_unrecognized_expectation):
227         (ExpectationSyntaxTests.test_macro):
228         (SemanticTests.test_bug_format):
229         (SemanticTests.test_bad_bugid):
230         (SemanticTests.test_missing_bugid):
231         (SemanticTests.test_slow_and_timeout):
232         (SemanticTests.test_rebaseline):
233         (test_missing_file):
234         (test_ambiguous):
235         (test_more_modifiers):
236         (test_order_in_file):
237         (test_macro_overrides):
238         (OldExpectationParserTests):
239         (OldExpectationParserTests._tokenize):
240         (OldExpectationParserTests.test_tokenize_extra_colon):
241         (OldExpectationParserTests.test_tokenize_missing_equal):
242         (OldExpectationParserTests.test_tokenize_extra_equal):
243         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
244         (MainTest.test_skip_failing_tests):
245         (MainTest.test_additional_expectations):
246         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
247         (TestExpectationsTestCase.test_valid_expectations):
248         (TestExpectationsTestCase.test_invalid_expectations):
249         (TestExpectationsTestCase.test_tab):
250
251 2012-09-21  Benjamin Poulain  <bpoulain@apple.com>
252
253         [WK2] Add basic testing support for Geolocation
254         https://bugs.webkit.org/show_bug.cgi?id=97278
255
256         Reviewed by Kenneth Rohde Christiansen.
257
258         Update the Geolocation testing to use the proper API in the UIProcess.
259
260         * WebKitTestRunner/CMakeLists.txt:
261         * WebKitTestRunner/GNUmakefile.am:
262
263         * WebKitTestRunner/GeolocationProviderMock.cpp: Added.
264         (WTR::startUpdatingCallback):
265         (WTR::stopUpdatingCallback):
266         (WTR::GeolocationProviderMock::GeolocationProvierMock):
267         (WTR::GeolocationProviderMock::setMockGeolocationPosition):
268         (WTR::GeolocationProviderMock::startUpdating):
269         (WTR::GeolocationProviderMock::stopUpdating):
270         (GeolocationProviderMock):
271         The GeolocationProvider store the location update and deliver them as needed.
272
273         WebCore GeolocationController do not support asynchronous update on start/stop. This is not
274         a problem in this case because all the messages between the WebProcess and the UIProcess are
275         asynchronous. Because of this, unlike GeolocationClientMock, we do not use a timer for event
276         delivery.
277
278         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
279         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
280         (WTR::InjectedBundle::setGeolocationPermission):
281         (WTR::InjectedBundle::setMockGeolocationPosition):
282         * WebKitTestRunner/InjectedBundle/InjectedBundle.h:
283         (InjectedBundle):
284         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
285         (WTR::TestRunner::setGeolocationPermission):
286         (WTR::TestRunner::setMockGeolocationPosition):
287         From the InjectedBundle, we now pass the information to the UIProcess so that
288         GeolocationProvider and the TestController can respond appropriately.
289
290         * WebKitTestRunner/InjectedBundle/TestRunner.h:
291         (TestRunner):
292         * WebKitTestRunner/Target.pri:
293         * WebKitTestRunner/TestController.cpp:
294         (WTR::TestController::TestController):
295         (WTR::decidePolicyForGeolocationPermissionRequest):
296         (WTR::TestController::createOtherPage):
297         (WTR::TestController::initialize):
298         (WTR::TestController::setMockGeolocationPosition):
299         * WebKitTestRunner/TestController.h:
300         (TestController):
301         (WTR::TestController::setGeolocationPermission):
302         (WTR::TestController::isGeolocationPermissionAllowed):
303         * WebKitTestRunner/TestInvocation.cpp:
304         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
305         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
306         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
307
308 2012-09-21  Joanmarie Diggs  <jdiggs@igalia.com>
309
310         REGRESSION (r127882): accessibility/spinbutton-value.html failing on GTK
311         https://bugs.webkit.org/show_bug.cgi?id=96196
312
313         Reviewed by Martin Robinson.
314
315         The "regression" is that a new test was added but the support was missing
316         in the Gtk port for spin buttons.
317
318         * DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
319         (AccessibilityUIElement::valueDescription): Updated the FIXME comment to
320         indicate that this cannot be implemented until it is implemented in ATK.
321         URL of the newly-filed ATK bug included for reference.
322
323 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
324
325         Fix test_skip_and_wontfix failure
326         https://bugs.webkit.org/show_bug.cgi?id=97225
327
328         Unreviewed, build fix.
329
330         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
331         (SemanticTests.test_skip_and_wontfix):
332
333 2012-09-21  Carlos Garcia Campos  <cgarcia@igalia.com>
334
335         [GTK] Implement ViewState methods in PageClientImpl in WebKit2
336         https://bugs.webkit.org/show_bug.cgi?id=97202
337
338         Reviewed by Martin Robinson.
339
340         * WebKitTestRunner/gtk/PlatformWebViewGtk.cpp:
341         (WTR::PlatformWebView::focus): Focus the view.
342
343 2012-09-21  Sudarsana Nagineni  <sudarsana.nagineni@intel.com>
344
345         [WTR] Memory leaks in InjectedBundleController::initialize()
346         https://bugs.webkit.org/show_bug.cgi?id=97329
347
348         Reviewed by Alexey Proskuryakov.
349
350         Fix leaks in InjectedBundleController::initialize() by adopting
351         strings created with WKStringCreateWithUTF8CString().
352
353         * TestWebKitAPI/InjectedBundleController.cpp:
354         (TestWebKitAPI::InjectedBundleController::initialize):
355
356 2012-09-21  Marcelo Lira  <marcelo.lira@openbossa.org>
357
358         [Qt][WK2] Removed duplicated code from EventSenderProxy::keyDown
359         https://bugs.webkit.org/show_bug.cgi?id=97235
360
361         Reviewed by Luiz Agostini.
362
363         * WebKitTestRunner/qt/EventSenderProxyQt.cpp:
364         (WTR::EventSenderProxy::keyDown):
365
366 2012-09-21  Gergely Kis  <gergely@homejinni.com>
367
368         Add MIPS build slave to build.webkit.org
369         https://bugs.webkit.org/show_bug.cgi?id=96713
370
371         Reviewed by Csaba Osztrogonác.
372
373         Added a build slave for MIPS, and enabled a builder for 
374         Qt Linux MIPS32R2 little-endian release build.
375
376         * BuildSlaveSupport/build.webkit.org-config/config.json:
377
378 2012-09-21  Peter Beverloo  <peter@chromium.org>
379
380         [Chromium] Switch back to a fixed fifo path for Android
381         https://bugs.webkit.org/show_bug.cgi?id=97230
382
383         Reviewed by Tony Chang.
384
385         Because not all external storage cards will be formated using a file
386         system that supports named pipes, Chromium has been changed to creating
387         the pipes in a temporary folder on the internal storage. Adapt this in
388         WebKit so we can continue to run layout tests.
389
390         This also includes an *unreviewed* fix for a breakage in the webkitpy
391         tests I made in r129221. Two lines and related to this code, so I decided
392         to include it in this change.
393
394         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
395         (ChromiumAndroidDriver.__init__):
396         (ChromiumAndroidDriver._setup_test):
397         (ChromiumAndroidDriver._update_version):
398
399 2012-09-20  Peter Beverloo  <peter@chromium.org>
400
401         Leverage Chromium's code to set up FIFOs for Chromium Android layout tests
402         https://bugs.webkit.org/show_bug.cgi?id=97227
403
404         Reviewed by Tony Chang.
405
406         We switched Chromium to using FIFOs in order to achieve better consistency,
407         which was done by Marcus in r157541. Remove all custom WebKit code in
408         favor of Chromium's implementation.
409
410         Remove more FIFO code in the test runner itself, including the code in
411         DumpRenderTree that invoked it. We can now switch to Chromium's brand
412         new FIFO-creating code, which is being set-up for all test targets build
413         for Android, including DumpRenderTree, TestWebKitAPI and webkit_unit_tests.
414
415         This also changes the ChromiumAndroidDriver._remove_all_pipes method to
416         delete the files individually. "rm" would fail if one of the earlier files
417         does not exist, and the "-f" argument doesn't seem to be reliable.
418
419         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
420         * DumpRenderTree/chromium/TestShellAndroid.cpp:
421         (platformInit):
422         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
423         (ChromiumAndroidDriver.__init__):
424         (ChromiumAndroidDriver._setup_test):
425         (ChromiumAndroidDriver._get_external_storage):
426         (ChromiumAndroidDriver._drt_cmd_line):
427         (ChromiumAndroidDriver._remove_all_pipes):
428         (ChromiumAndroidDriver.stop):
429         * TestWebKitAPI/TestWebKitAPI.gyp/TestWebKitAPI.gyp:
430
431 2012-09-21  Christophe Dumez  <christophe.dumez@intel.com>
432
433         [WK2][WKTR] EventSender needs to implement scheduleAsynchronousClick
434         https://bugs.webkit.org/show_bug.cgi?id=97326
435
436         Reviewed by Kenneth Rohde Christiansen.
437
438         Implement scheduleAsynchronousClick() in WebKitTestRunner's
439         EventSender by sending a "MouseDown" and a "MouseUp" message
440         asynchronously to the WebProcess.
441
442         * WebKitTestRunner/InjectedBundle/Bindings/EventSendingController.idl:
443         * WebKitTestRunner/InjectedBundle/EventSendingController.cpp:
444         (WTR::createMouseMessageBody):
445         (WTR):
446         (WTR::EventSendingController::mouseDown):
447         (WTR::EventSendingController::mouseUp):
448         (WTR::EventSendingController::scheduleAsynchronousClick):
449         * WebKitTestRunner/InjectedBundle/EventSendingController.h:
450         (EventSendingController):
451         * WebKitTestRunner/TestController.cpp:
452         (WTR::TestController::didReceiveMessageFromInjectedBundle):
453
454 2012-09-21  Simon Hausmann  <simon.hausmann@digia.com>
455
456         [Qt] Error out early if we don't have ICU available
457
458         Reviewed by Tor Arne Vestbø.
459
460         * qmake/mkspecs/features/configure.prf:
461
462 2012-09-21  Simon Hausmann  <simon.hausmann@digia.com>
463
464         [Qt] Bail out when trying to build WebKit with Qt != 5
465
466         Reviewed by Tor Arne Vestbø.
467
468         Moved check for Qt version out of default_pre into top-level WebKit.pro,
469         because we never reach default_pre.prf due to the lack of .qmake.conf support
470         in older versions of Qt/QMake.
471
472         * qmake/mkspecs/features/default_pre.prf:
473
474 2012-09-21  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
475
476         WebKitTestRunner needs to print frame load delegate information
477         https://bugs.webkit.org/show_bug.cgi?id=42705
478
479         Reviewed by Kenneth Rohde Christiansen.
480
481         Added missing dumping from WTR::InjectedBundlePage::didFailLoadWithErrorForFrame.
482
483         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
484         (WTR::InjectedBundlePage::didFailLoadWithErrorForFrame):
485
486 2012-09-21  Christophe Dumez  <christophe.dumez@intel.com>
487
488         [EFL] EventSender should mimic CTRL+o emacs shortcut
489         https://bugs.webkit.org/show_bug.cgi?id=97224
490
491         Reviewed by Kenneth Rohde Christiansen.
492
493         Add support for mimicking CTRL+o emacs shortcut in EFL's
494         EventSender in DumpRenderTree and WebKitTestRunner.
495
496         * DumpRenderTree/efl/EventSender.cpp:
497         (sendKeyDown):
498         * WebKitTestRunner/efl/EventSenderProxyEfl.cpp:
499         (WTR::EventSenderProxy::keyDown):
500
501 2012-09-21  Simon Hausmann  <simon.hausmann@digia.com>
502
503         [Qt] Re-fix clean builds
504
505         Reviewed by Tor Arne Vestbø.
506
507         Re-introduce the sanitization for LIBS when creating a module to use
508         LIBS_PRIVATE and otherwise do _not_ use LIBS_PRIVATE. We decided to
509         continue to use QT, LIBS and PKGCONFIG instead of their _PRIVATE variants
510         throughout the code base, so just using LIBS_PRIVATE in linkAgainstLibrary()
511         causes build issues when depending system libraries end up in LIBS before
512         LIBS_PRIVATE.
513
514         * qmake/mkspecs/features/default_post.prf:
515         * qmake/mkspecs/features/functions.prf:
516
517 2012-09-20  Simon Fraser  <simon.fraser@apple.com>
518
519         Comment out a failing webkitpy unit test until Dirk can fix it.
520
521         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
522         (SemanticTests.test_skip_and_wontfix):
523
524 2012-09-20  Dirk Pranke  <dpranke@chromium.org>
525
526         make Skip, WontFix be the only expectations on a line
527         https://bugs.webkit.org/show_bug.cgi?id=97225
528
529         Reviewed by Ojan Vafai.
530
531         It is now incorrect in the new syntax to have a line like:
532
533             foo.html [ WontFix Crash ]
534
535         This will generate a lint warning and be treated as an invalid
536         line. Fixing this caused a whole bunch of unit tests to need updating
537         to no longer be marked as WontFix :). Also, this patch adjusts
538         the warnings so that missing Bug() identifiers will cause lint
539         warnings but will *not* cause the line to be treated as invalid.
540         Fixing these issues also revealed that test_hung_thread was no
541         longer testing the right logic, so I adjusted the timeouts in
542         test.py to make that test work again.
543
544         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner.py:
545         (Worker._run_test_in_another_thread):
546         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
547         (ResultSummaryTest.test_summarized_results_wontfix):
548         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
549         (TestExpectationParser._parse_modifiers):
550         (TestExpectationParser._tokenize_line_using_new_format):
551         (TestExpectationLine.is_invalid):
552         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
553         (BasicTests.test_basic):
554         (test_get_test_set):
555         (test_parse_warning):
556         (test_pixel_tests_flag):
557         (SemanticTests.test_missing_bugid):
558         (SemanticTests):
559         (SemanticTests.test_skip_and_wontfix):
560         * Scripts/webkitpy/layout_tests/port/test.py:
561         (TestDriver.run_test):
562         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
563         (MainTest.test_hung_thread):
564         * Scripts/webkitpy/tool/commands/queries_unittest.py:
565         (PrintExpectationsTest.test_basic):
566         (PrintExpectationsTest.test_multiple):
567         (PrintExpectationsTest.test_full):
568         (PrintExpectationsTest.test_exclude):
569         (PrintExpectationsTest.test_csv):
570
571 2012-09-20  Lucas Forschler  <lforschler@apple.com>
572
573         Unreviewed.  Start running tests on the mac-ews.
574         
575         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
576         (MacEWS):
577
578 2012-09-20  Tony Chang  <tony@chromium.org>
579
580         [chromium] Enable more clang warnings
581         https://bugs.webkit.org/show_bug.cgi?id=97257
582
583         Reviewed by James Robinson.
584
585         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp: Add -Wunused-parameter. Don't add -Wglobal-constructors because
586         there are lots of violations in these files.
587
588 2012-09-19  Ryosuke Niwa  <rniwa@webkit.org>
589
590         run-perf-tests should record individual value instead of statistics
591         https://bugs.webkit.org/show_bug.cgi?id=97155
592
593         Reviewed by Hajime Morita.
594
595         Parse the list of individual value reported by tests and include them as "values".
596         We strip "values" from the output JSON when uploading it to the perf-o-matic
597         since it doesn't know how to parse "values" or ignore it.
598
599         * Scripts/webkitpy/performance_tests/perftest.py:
600         (PerfTest):
601         (PerfTest.parse_output): Parse and report "values".
602         (PageLoadingPerfTest.run): Report indivisual page loading time in "values".
603         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
604         (MainTest.test_parse_output):
605         (MainTest.test_parse_output_with_failing_line):
606         (TestPageLoadingPerfTest.test_run):
607         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
608         (PerfTestsRunner._generate_and_show_results): Strip "values" from each result
609         until we update perf-o-matic.
610         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
611         (test_run_memory_test):
612         (test_run_with_json_output):
613         (test_run_with_description):
614         (test_run_with_slave_config_json):
615         (test_run_with_multiple_repositories):
616
617 2012-09-20  Simon Hausmann  <simon.hausmann@digia.com>
618
619         [Qt] Fix production builds
620
621         Unreviewed trivial fix: Follow up r129119 and avoid clobbering LIBS through the use of LIBS_PRIVATE.
622
623         Otherwise the libQtWebKit.prl file for example contains -lWebKit1 -lWebCore, etc.
624
625         * qmake/mkspecs/features/functions.prf:
626
627 2012-09-20  Dirk Pranke  <dpranke@chromium.org>
628
629         REGRESSION: layout test results doesn't show diffs
630         https://bugs.webkit.org/show_bug.cgi?id=97182
631
632         Reviewed by Ojan Vafai.
633
634         Go back to storing TEXT, AUDIO, and IMAGE+TEXT in results.json
635         so that results.html (and hopefully garden-o-matic) can
636         determine which things actually failed. However, we keep mapping
637         these results to Failure so that we still only have a single
638         expectation type for them.
639
640         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
641         (JSONLayoutResultsGenerator):
642         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
643         (TestExpectationParser):
644         (TestExpectations):
645         (TestExpectations.result_was_expected):
646         * Scripts/webkitpy/layout_tests/models/test_failures.py:
647         (determine_result_type):
648         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
649         (MainTest.test_missing_and_unexpected_results):
650         (MainTest.test_retrying_and_flaky_tests):
651
652 2012-09-20  Tommy Widenflycht  <tommyw@google.com>
653
654         MediaStream API: Extend UserMediaRequest with a ownerDocument method
655         https://bugs.webkit.org/show_bug.cgi?id=97095
656
657         Reviewed by Adam Barth.
658
659         Extending WebUserMediaClientMock to check that the owning document is valid,
660         and that the document has a frame.
661
662         * DumpRenderTree/chromium/WebUserMediaClientMock.cpp:
663         (WebKit::WebUserMediaClientMock::requestUserMedia):
664
665 2012-09-20  Stephen Chenney  <schenney@chromium.org>
666
667         [Chromium] DRT does not support --dump-all-pixels flag
668         https://bugs.webkit.org/show_bug.cgi?id=95098
669
670         Reviewed by Dirk Pranke.
671
672         Add support for the --pixel-tests and shorthand -p option in Chromium DumpRenderTree. Use
673         of this flag causes pixel results to be created for all tests, regardless of
674         individual test options. If an individual test provides a pixel hash it will be used,
675         otherwise the hash will be empty. This replaces a previously defined but unused option
676         --dump-all-pixels, and is useful primarily when debugging DRT instances.
677
678         * DumpRenderTree/chromium/DumpRenderTree.cpp:
679         (runTest): Add a parameter and code to force pixel results for the test.
680         (main): Add parameter handling for --pixels-test and -p, and remove --dump-all-pixels.
681
682 2012-09-20  Carlos Garcia Campos  <cgarcia@igalia.com>
683
684         [GTK] run-api-tests should not buffer test stdout
685         https://bugs.webkit.org/show_bug.cgi?id=88474
686
687         Reviewed by Philippe Normand.
688
689         * Scripts/run-gtk-tests:
690         (TestRunner._run_test_command): Use os.forkpty() instead of
691         subprocess.Popen() so that gtest sends the output with colors to
692         stdout. Use common.parse_output_lines() to parse the output and
693         write it to stdout while it's read.
694         (TestRunner._run_test_command.parse_line): Parse the line to get
695         the test pid and write the line to stdout.
696         (TestRunner._run_test_command.waitpid): Helper function to call
697         waitpid handling EINTR.
698         (TestRunner._run_test_command.return_code_from_exit_status):
699         Helper function to convert exit status of test commands to a
700         return code.
701         * gtk/common.py:
702         (parse_output_lines): Helper function that uses select to read
703         the given file descriptor and call the given callback for every
704         line read.
705
706 2012-09-20  Simon Hausmann  <simon.hausmann@digia.com>
707
708         [Qt] Fix initial build
709
710         Reviewed by Tor Arne Vestbø.
711
712         When building QtWebKit the first time there is no qt_webkit.pri module pri file, and therefore
713         $$QT.webkit.name isn't set and so creating_module isn't set. That has all sorts of implications
714         causing incorrect linking for Makefile.api, etc.
715
716         Fix the determination by simply checking if MODULE is set, which only happens in api.pri.
717
718         * qmake/mkspecs/features/webkit_modules.prf:
719
720 2012-09-20  Sheriff Bot  <webkit.review.bot@gmail.com>
721
722         Unreviewed, rolling out r129091.
723         http://trac.webkit.org/changeset/129091
724         https://bugs.webkit.org/show_bug.cgi?id=97205
725
726         It broke perf tests everywhere (Requested by Ossy on #webkit).
727
728         * Scripts/webkitpy/performance_tests/perftest.py:
729         (PerfTest):
730         (PerfTest.parse_output):
731         (PageLoadingPerfTest.run):
732         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
733         (MainTest.test_parse_output):
734         (MainTest.test_parse_output_with_failing_line):
735         (TestPageLoadingPerfTest.test_run):
736         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
737         (PerfTestsRunner._generate_and_show_results):
738         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
739         (test_run_memory_test):
740         (test_run_with_json_output):
741         (test_run_with_description):
742         (test_run_with_slave_config_json):
743         (test_run_with_multiple_repositories):
744
745 2012-09-20  Simon Hausmann  <simon.hausmann@digia.com>
746
747         [Qt] QtWebKit module header includes private dependencies
748
749         Reviewed by Tor Arne Vestbø.
750
751         * qmake/mkspecs/features/default_post.prf: Don't try to sanitize LIBS, because we can make sure
752         that LIBS_PRIVATE is set from the beginning. Moved the creating_module and PKGCONFIG/QT(_PRIVATE)
753         sanitization into webkit_modules.prf. creating_module determination requires TARGET to be set, so
754         we can't do it in default_pre.
755         * qmake/mkspecs/features/webkit_modules.prf:
756
757 2012-09-20  Simon Hausmann  <simon.hausmann@digia.com>
758
759         [Qt] Fix build with MingW
760
761         Reviewed by Tor Arne Vestbø.
762
763         Don't try to compile WebCore, etc. with debug symbols for production
764         builds, it's just too big.
765
766         * qmake/mkspecs/features/production_build.prf:
767
768 2012-09-20  Michał Pakuła vel Rutka  <m.pakula@samsung.com>
769
770         [EFL][DRT]DumpRenderTree needs to reset focus state when test starts.
771         https://bugs.webkit.org/show_bug.cgi?id=97087
772
773         Reviewed by Gyuyoung Kim.
774
775         Add focusing a main frame on settings reset.
776         After editing/undo/undo-iframe-location-change was executed a frame
777         was left in unfocused state. This caused flakiness in two tests results.
778
779         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
780         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
781
782 2012-09-20  Allan Sandfeld Jensen  <allan.jensen@digia.com>
783
784         [Qt] Add eventSender.gestureTap
785         https://bugs.webkit.org/show_bug.cgi?id=66173
786
787         Reviewed by Kenneth Rohde Christiansen.
788
789         Implement support for sending raw QGesture events.
790
791         * DumpRenderTree/qt/EventSenderQt.cpp:
792         (EventSender::EventSender):
793         (EventSender::gestureTap):
794         * DumpRenderTree/qt/EventSenderQt.h:
795         (EventSender):
796
797 2012-09-19  Sheriff Bot  <webkit.review.bot@gmail.com>
798
799         Unreviewed, rolling out r129007.
800         http://trac.webkit.org/changeset/129007
801         https://bugs.webkit.org/show_bug.cgi?id=97172
802
803         It broke the build on the Qt bots (Requested by Ossy on
804         #webkit).
805
806         * qmake/config.tests/gccdepends/empty.cpp:
807         * qmake/config.tests/gccdepends/gccdepends.pro:
808         * qmake/mkspecs/features/default_pre.prf:
809
810 2012-09-19  Ryosuke Niwa  <rniwa@webkit.org>
811
812         run-perf-tests should record indivisual value instead of statistics
813         https://bugs.webkit.org/show_bug.cgi?id=97155
814
815         Reviewed by Hajime Morita.
816
817         Parse the list of indivisual value reported by tests and include them as "values".
818         We strip "values" from the output JSON when uploading it to the perf-o-matic
819         since it doesn't know how to parse "values" or ignore it.
820
821         * Scripts/webkitpy/performance_tests/perftest.py:
822         (PerfTest):
823         (PerfTest.parse_output): Parse and report "values".
824         (PageLoadingPerfTest.run): Report indivisual page loading time in "values".
825         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
826         (MainTest.test_parse_output):
827         (MainTest.test_parse_output_with_failing_line):
828         (TestPageLoadingPerfTest.test_run):
829         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
830         (PerfTestsRunner._generate_and_show_results): Strip "values" from each result
831         until we update perf-o-matic.
832         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
833         (test_run_memory_test):
834         (test_run_with_json_output):
835         (test_run_with_description):
836         (test_run_with_slave_config_json):
837         (test_run_with_multiple_repositories):
838
839 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
840
841         Fix regex groups for bug matching in flakiness dashboard.
842         https://bugs.webkit.org/show_bug.cgi?id=97152
843
844         Unreviewed, build fix.
845
846         * TestResultServer/static-dashboards/flakiness_dashboard.js:
847         (htmlForBugs):
848
849 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
850
851         nrwt: print unexpected results using new TestExpectations syntax
852         https://bugs.webkit.org/show_bug.cgi?id=97159
853
854         Unreviewed, build fix.
855
856         Change new-run-webkit-tests to print out failures using the new
857         syntax when there are unexpected results, e.g.:
858
859         52 tests ran as expected, 19 didn't:
860
861
862         Regressions: Unexpected failures : (2)
863           failures/flaky/text.html [ Failure ]
864           failures/unexpected/text-image-checksum.html [ Failure ]
865
866         and so forth
867
868         * Scripts/webkitpy/layout_tests/views/printing.py:
869         (Printer._print_unexpected_results):
870
871 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
872
873         update flakiness dashboard after cutover to new test expectations syntax
874         https://bugs.webkit.org/show_bug.cgi?id=97152
875
876         Unreviewed, build fix.
877
878         Handle (??) the new Bug notations as well. Hopefully we don't
879         still need the old ones.
880
881         * TestResultServer/static-dashboards/flakiness_dashboard.js:
882         (filterBugs):
883         (htmlForBugs):
884         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
885
886 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
887
888         update flakiness dashboard after cutover to new test expectations syntax
889         https://bugs.webkit.org/show_bug.cgi?id=97152
890
891         Reviewed by Ryosuke Niwa.
892
893         This change clones the TestExpectation parsing state machine
894         from python into javascript.
895
896         * TestResultServer/static-dashboards/flakiness_dashboard.js:
897         (parsedExpectations.lines.forEach):
898         (parsedExpectations):
899         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
900
901 2012-09-19  Dana Jansens  <danakj@chromium.org>
902
903         Add backer@chromium.org as contributor
904         https://bugs.webkit.org/show_bug.cgi?id=97150
905
906         * Scripts/webkitpy/common/config/committers.py:
907
908 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
909
910         fix MISSING after TestExpectations conversion
911         https://bugs.webkit.org/show_bug.cgi?id=97148
912
913         Unreviewed, expectations change / build fix.
914
915         * Scripts/convert-test-expectations:
916         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
917         (TestExpectationParser):
918
919 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
920
921         nrwt: convert chromium TestExpectations to the new syntax
922         https://bugs.webkit.org/show_bug.cgi?id=97139
923
924         Reviewed by Ryosuke Niwa.
925     
926         Add in a file temporarily that does the conversion of
927         test expectations formats
928
929         * Tools/Scripts/convert-test-expectations: Added.
930
931 2012-09-19  Ryosuke Niwa  <rniwa@webkit.org>
932
933         REGRESSION: run-perf-tests no longer reports the total test time
934         https://bugs.webkit.org/show_bug.cgi?id=97138
935
936         Reviewed by Tony Chang.
937
938         Report the finished time as a info-level log as opposed to a debug level log.
939
940         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
941         (PerfTestsRunner._run_single_test):
942         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
943         (MainTest.normalizeFinishedTime): Added. It replaces all finished times by 0.1 seconds.
944         (test_run_test_pause_before_testing):
945         (test_run_test_set_for_parser_tests):
946         (test_run_memory_test):
947         (_test_run_with_json_output):
948
949 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
950
951         Support new TestExpectations format alongside old one
952         https://bugs.webkit.org/show_bug.cgi?id=96588
953
954         Reviewed by Ojan Vafai.
955
956         This patch adds support for actually reading in lines formatted
957         in the new style, and when we re-serialize/write out the file,
958         all lines will be written in the new style. Note that reading in
959         the old style is still supported, and no updates are being made
960         to the actual TestExpectations files as part of this change.
961
962         This change updates most but not all of the unit tests to use
963         the new syntax. I will update the rest when (or before where
964         possible) I drop support for the old format.
965
966         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
967         (TestExpectationParser):
968         (TestExpectationParser._tokenize_line):
969         (TestExpectationParser._tokenize_line_using_new_format):
970         (TestExpectationLine.to_string):
971         (TestExpectationLine._format_line):
972         (TestExpectations):
973         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
974         (test_parse_warning):
975         (SkippedTests.test_skipped_entry_dont_exist):
976         (NewExpectationSyntaxTests.test_warnings):
977         (RemoveConfigurationsTest.test_remove):
978         (test_remove_line):
979         (RebaseliningTest.test_remove):
980         (RebaseliningTest.test_no_get_rebaselining_failures):
981         (TestExpectationParserTests.test_tokenize_blank):
982         (TestExpectationParserTests.test_tokenize_extra_colon):
983         (TestExpectationParserTests.test_tokenize_missing_equal):
984         (TestExpectationParserTests.test_tokenize_extra_equal):
985         (TestExpectationSerializationTests.test_unparsed_to_string):
986         (TestExpectationSerializationTests.test_unparsed_list_to_string):
987         (TestExpectationSerializationTests.test_parsed_to_string):
988         (TestExpectationSerializationTests.test_format_line):
989         (TestExpectationSerializationTests.test_string_roundtrip):
990         (TestExpectationSerializationTests.test_list_roundtrip):
991         (TestExpectationSerializationTests.test_reconstitute_only_these):
992         (TestExpectationSerializationTests.test_string_whitespace_stripping):
993         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
994         (LintTest.test_lint_test_files__errors):
995         * Scripts/webkitpy/tool/commands/queries_unittest.py:
996         (PrintExpectationsTest.test_basic):
997         (PrintExpectationsTest.test_multiple):
998         (PrintExpectationsTest.test_full):
999         (PrintExpectationsTest.test_exclude):
1000         * Scripts/webkitpy/tool/commands/rebaseline.py:
1001         (RebaselineExpectations.execute):
1002         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
1003         (TestRebaseline.test_rebaseline_updates_expectations_file_noop):
1004         (test_rebaseline_updates_expectations_file):
1005         (test_rebaseline_does_not_include_overrides):
1006         (test_rebaseline_expectations_noop):
1007         (test_overrides_are_included_correctly):
1008
1009 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
1010
1011         implement first part of support for the new TestExpectations syntax
1012         https://bugs.webkit.org/show_bug.cgi?id=96569
1013
1014         Reviewed by Ryosuke Niwa.
1015
1016         This patch implements support for parsing a line of the new
1017         format for the TestExpectations file and converting it back into
1018         the old format for compatibility. This routine is not yet used
1019         by anything.
1020
1021         The new format is documented at:
1022           http://trac.webkit.org/wiki/TestExpectations
1023
1024         but, in short:
1025
1026           [bugs] [ "[" modifiers "]" ] test_name [ "[" expectations "]" ]
1027
1028         - Comments are indicated with "#" instead of "//"
1029         - If no expectations are specified we default to Skip for
1030           compatibility with the Skipped files (these two changes make
1031           Skipped files a subset of TestExpectations files)
1032
1033         - All of the tokens are now CamelCase instead of ALLCAPS.
1034         - FAIL -> Failure
1035         - IMAGE -> ImageOnlyFailure
1036         - WONTFIX -> WontFix
1037         - modifiers refer to just the platforms and configurations
1038           (release/debug) that the line applies to.
1039         - WontFix, Rebaseline, Slow, and Skip move to the right-hand side as
1040           expectations
1041         - expectations will typically be written out in lexicographic order
1042         - We use webkit.org/b/12345, crbug.com/12345, and Bug(dpranke)
1043           instead of BUGWK12345, BUGCR12345, and BUGDPRANKE.
1044
1045         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
1046         (TestExpectationParser):
1047         (TestExpectationParser._tokenize_line_using_new_format):
1048         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
1049         (NewExpectationSyntaxTests):
1050         (NewExpectationSyntaxTests.assert_exp):
1051         (NewExpectationSyntaxTests.test_bare_name):
1052         (NewExpectationSyntaxTests.test_bare_name_and_bugs):
1053         (NewExpectationSyntaxTests.test_comments):
1054         (NewExpectationSyntaxTests.test_config_modifiers):
1055         (NewExpectationSyntaxTests.test_unknown_config):
1056         (NewExpectationSyntaxTests.test_unknown_expectation):
1057         (NewExpectationSyntaxTests.test_skip):
1058         (NewExpectationSyntaxTests.test_slow):
1059         (NewExpectationSyntaxTests.test_wontfix):
1060         (NewExpectationSyntaxTests.test_blank_line):
1061         (NewExpectationSyntaxTests.test_warnings):
1062
1063 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
1064
1065         nrwt: replace TEXT, AUDIO, and IMAGE+TEXT with FAIL
1066         https://bugs.webkit.org/show_bug.cgi?id=96845
1067
1068         Reviewed by Ojan Vafai.
1069
1070         In preparation for the new TestExpectations syntax, we replace
1071         all TEXT, IMAGE+TEXT, and AUDIO failures with FAIL. This will
1072         make switching to the new syntax lossless (i.e., we lose
1073         information now, not then).
1074
1075         Note that we can still parse in results.json files that have the
1076         old data for backwards compatibility.
1077         
1078
1079         * Scripts/webkitpy/common/net/resultsjsonparser.py:
1080         (JSONTestResult._failure_types_from_actual_result):
1081         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
1082         (ResultsJSONParserTest):
1083         (test_basic):
1084         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
1085         (ResultSummaryTest.test_summarized_results_wontfix):
1086         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
1087         (JSONLayoutResultsGenerator):
1088         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
1089         (TestExpectations):
1090         (TestExpectations.remove_pixel_failures):
1091         (TestExpectations.has_pixel_failures):
1092         (TestExpectations.suffixes_for_expectations):
1093         (TestExpectations.get_rebaselining_failures):
1094         (TestExpectations.remove_configuration_from_test):
1095         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
1096         (assert_bad_expectations):
1097         (BasicTests):
1098         (BasicTests.test_basic):
1099         (MiscTests):
1100         (MiscTests.test_multiple_results):
1101         (MiscTests.test_result_was_expected):
1102         (MiscTests.test_remove_pixel_failures):
1103         (MiscTests.test_suffixes_for_expectations):
1104         (test_get_expectations_string):
1105         (test_parse_warning):
1106         (test_error_on_different_platform):
1107         (test_error_on_different_build_type):
1108         (test_overrides):
1109         (test_overrides__directory):
1110         (test_overrides__duplicate):
1111         (test_pixel_tests_flag):
1112         (test_more_specific_override_resets_skip):
1113         (SkippedTests.test_skipped_file_overrides_expectations):
1114         (SkippedTests.test_skipped_dir_overrides_expectations):
1115         (SkippedTests.test_skipped_file_overrides_overrides):
1116         (SkippedTests.test_skipped_dir_overrides_overrides):
1117         (ExpectationSyntaxTests.test_missing_colon):
1118         (ExpectationSyntaxTests.test_too_many_equals_signs):
1119         (SemanticTests.test_bug_format):
1120         (SemanticTests.test_bad_bugid):
1121         (SemanticTests.test_missing_bugid):
1122         (SemanticTests.test_rebaseline):
1123         (test_missing_file):
1124         (test_more_modifiers):
1125         (test_order_in_file):
1126         (test_macro_overrides):
1127         (RemoveConfigurationsTest.test_remove):
1128         (test_remove_line):
1129         (RebaseliningTest.test_remove):
1130         (TestExpectationSerializationTests.test_serialize_parsed_expectations):
1131         * Scripts/webkitpy/layout_tests/models/test_failures.py:
1132         (determine_result_type):
1133         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
1134         (test_test_expectations):
1135         * Scripts/webkitpy/layout_tests/port/test.py:
1136         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1137         (MainTest.test_missing_and_unexpected_results):
1138         (MainTest.test_retrying_and_flaky_tests):
1139         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
1140         (TestExpectationsTestCase.test_valid_expectations):
1141         * Scripts/webkitpy/tool/commands/queries_unittest.py:
1142         (PrintExpectationsTest.test_basic):
1143         (PrintExpectationsTest.test_multiple):
1144         (PrintExpectationsTest.test_full):
1145         (PrintExpectationsTest.test_exclude):
1146         (PrintExpectationsTest.test_csv):
1147         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
1148         (test_rebaseline_does_not_include_overrides):
1149         (test_overrides_are_included_correctly):
1150
1151 2012-09-19  Dominic Mazzoni  <dmazzoni@google.com>
1152
1153         AX: A few control types are returning the wrong answer for isReadOnly
1154         https://bugs.webkit.org/show_bug.cgi?id=96735
1155
1156         Reviewed by Chris Fleizach.
1157
1158         Exposing isReadOnly in an AccessibilityObject to DumpRenderTree.
1159
1160         * DumpRenderTree/chromium/TestRunner/AccessibilityUIElementChromium.cpp:
1161         (AccessibilityUIElement::AccessibilityUIElement):
1162         (AccessibilityUIElement::isReadOnlyGetterCallback):
1163         * DumpRenderTree/chromium/TestRunner/AccessibilityUIElementChromium.h:
1164         (AccessibilityUIElement):
1165
1166 2012-09-19  Sudarsana Nagineni  <sudarsana.nagineni@intel.com>
1167
1168         [WTR] Memory leaks in TestRunner::deliverWebIntent()
1169         https://bugs.webkit.org/show_bug.cgi?id=97111
1170
1171         Reviewed by Kenneth Rohde Christiansen.
1172
1173         Fix memory leaks in deliverWebIntent() by adopting strings
1174         created with WKStringCreateWithUTF8CString().
1175
1176         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
1177         (WTR::TestRunner::deliverWebIntent):
1178
1179 2012-09-19  Simon Hausmann  <simon.hausmann@digia.com>
1180
1181         [Qt] Fix incremental builds with all-in-one-files and gccdepends
1182
1183         Reviewed by Tor Arne Vestbø.
1184
1185         Pass -MP to gcc when we use the gcc depends feature, to ensure that implicit rules
1186         are not only created for header files but also for .cpp files. AllInOne.cpp files
1187         include other .cpp files, and when those are removed we need those dummy rules to
1188         avoid a "No rule to make Foo.cpp required by AllInOne.o" error.
1189
1190         * qmake/config.tests/gccdepends/empty.cpp:
1191         (main):
1192         * qmake/config.tests/gccdepends/gccdepends.pro:
1193         * qmake/mkspecs/features/default_pre.prf:
1194
1195 2012-09-19  Rick Byers  <rbyers@chromium.org>
1196
1197         Do touch adjustment on GestureTapDown
1198         https://bugs.webkit.org/show_bug.cgi?id=96677
1199
1200         Reviewed by Antonio Gomes.
1201
1202         Allow radius to be set for GestureTapDown events.
1203         * DumpRenderTree/chromium/TestRunner/EventSender.cpp:
1204         (EventSender::gestureEvent):
1205
1206 2012-09-19  Christophe Dumez  <christophe.dumez@intel.com>
1207
1208         [EFL] EFL's DRT does not support overriding 'WebKitCSSRegionsEnabled' preference
1209         https://bugs.webkit.org/show_bug.cgi?id=97100
1210
1211         Reviewed by Gyuyoung Kim.
1212
1213         EFL's DRT now supports overriding the 'WebKitCSSRegionsEnabled'
1214         preference, in order to match WebKitTestRunner functionality.
1215
1216         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
1217         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
1218         * DumpRenderTree/efl/TestRunnerEfl.cpp:
1219         (TestRunner::overridePreference):
1220
1221 2012-09-18  Kangil Han  <kangil.han@samsung.com>
1222
1223         [WK2][WTR] CodeGeneratorTestRunner could keep original copyright.
1224         https://bugs.webkit.org/show_bug.cgi?id=96181
1225
1226         Reviewed by Daniel Bates.
1227
1228         This patch enabled derived files, in DerivedSources/InjectedBundle, to keep original copyright.
1229
1230         * WebKitTestRunner/InjectedBundle/Bindings/CodeGeneratorTestRunner.pm:
1231         (new):
1232         (_parseLicenseBlock):
1233         (_parseLicenseBlockFromFile):
1234         (_defaultLicenseBlock):
1235         (_licenseBlock):
1236         (_generateHeaderFile):
1237         (_generateImplementationFile):
1238
1239 2012-09-18  Byungwoo Lee  <bw80.lee@samsung.com>
1240
1241         Title string should be changed when document.title is set to ''.
1242         https://bugs.webkit.org/show_bug.cgi?id=96793
1243
1244         Reviewed by Kenneth Rohde Christiansen.
1245
1246         Change dump format of dumpTitleChanges more understandable.
1247         Uses single quotation marks for the title string.
1248
1249         * DumpRenderTree/blackberry/DumpRenderTree.cpp:
1250         (BlackBerry::WebKit::DumpRenderTree::didReceiveTitleForFrame):
1251         * DumpRenderTree/chromium/WebViewHost.cpp:
1252         (WebViewHost::didReceiveTitle):
1253         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
1254         (DumpRenderTreeChrome::onFrameTitleChanged):
1255         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1256         (webViewTitleChanged):
1257         * DumpRenderTree/mac/FrameLoadDelegate.mm:
1258         (-[FrameLoadDelegate webView:didReceiveTitle:forFrame:]):
1259         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1260         (WebCore::DumpRenderTree::titleChanged):
1261         * DumpRenderTree/win/FrameLoadDelegate.cpp:
1262         (FrameLoadDelegate::didReceiveTitle):
1263         * DumpRenderTree/wx/DumpRenderTreeWx.cpp:
1264         (LayoutWebViewEventHandler::OnReceivedTitleEvent):
1265         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1266         (WTR::InjectedBundlePage::didReceiveTitleForFrame):
1267
1268 2012-09-18  Szilard Ledan  <szledan@inf.u-szeged.hu>
1269
1270         EWS shouldn't sleep if there are new patches in its queue
1271         https://bugs.webkit.org/show_bug.cgi?id=83038
1272
1273         Reviewed by Eric Seidel.
1274
1275         EWS tries to process a security patch. Of course it can't, because the EWS isn't
1276         the member of the security group. But the problem is that after it can't process
1277         the attachment, it says that queue is empty (but it isn't!) and it sleeps 2 minutes
1278         and push the security patch to the end of the queue.
1279         Now it stays in the loop until it finds a patch or the queue gets empty.
1280
1281         * Scripts/webkitpy/tool/commands/queues.py:
1282         (AbstractPatchQueue._next_patch):
1283         * Scripts/webkitpy/tool/commands/queues_unittest.py:
1284         (AbstractPatchQueueTest.test_next_patch):
1285
1286 2012-09-18  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
1287
1288         [WK2][WTR] InjectedBundle::booleanForKey() should handle literals effectively
1289         https://bugs.webkit.org/show_bug.cgi?id=97014
1290
1291         Reviewed by Kenneth Rohde Christiansen.
1292
1293         According to http://trac.webkit.org/wiki/EfficientStrings WTF::StringBuilder::appendLiteral() shall
1294         be used for literals rather than WTF::StringBuilder::append().
1295
1296         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
1297         (WTR::InjectedBundle::booleanForKey):
1298
1299 2012-09-18  Simon Hausmann  <simon.hausmann@digia.com>
1300
1301         Update my e-mail address.
1302
1303         * Scripts/webkitpy/common/config/committers.py:
1304
1305 2012-09-18  Andras Becsi  <andras.becsi@digia.com>
1306
1307         Update my e-mail address.
1308
1309         Unreviewed.
1310
1311         * Scripts/webkitpy/common/config/committers.py:
1312
1313 2012-09-18  Andras Becsi  <andras.becsi@digia.com>
1314
1315         [Qt] qt_webkit.pri should not be listed in Tools.pro
1316
1317         Reviewed and rubber-stamped by Simon Hausmann and Tor Arne Vestbø.
1318
1319         Since r128751 the module pri file is auto-generated
1320         but it was still listed in OTHER_FILES.
1321
1322         * Tools.pro: Remove unneeded line.
1323
1324 2012-09-18  Simon Hausmann  <simon.hausmann@digia.com>
1325
1326         [Qt] Fix build with some versions of the gold linker
1327
1328         Reviewed by Tor Arne Vestbø.
1329
1330         Don't unconditionally pass --no-keep-memory to the linker, some versions might not support it.
1331         Instead run a compile/link test first to see if it works.
1332
1333         * qmake/config.tests/gnuld/gnuld.pro: Added.
1334         * qmake/config.tests/gnuld/main.cpp: Added.
1335         (main):
1336         * qmake/mkspecs/features/unix/default_post.prf:
1337
1338 2012-09-18  Simon Hausmann  <simon.hausmann@digia.com>
1339
1340         [Qt] Remove forced use of gold.
1341
1342         Reviewed by Tor Arne Vestbø.
1343
1344         The choice of what linker to use with WebKit should be taken by Qt's build system and ideally the same for all
1345         modules of Qt. Then in turn it is usually up to the administrator of the machine. Recent Debian based systems
1346         often offer the automatic use of gold through a symlink and a dpkg-diversion when installing the gold package.
1347
1348         * qmake/mkspecs/features/unix/default_post.prf:
1349
1350 2012-09-18  Simon Hausmann  <simon.hausmann@digia.com>
1351
1352         [Qt] Prospective Qt/Windows cross-compiling fix
1353
1354         Reviewed by Tor Arne Vestbø.
1355
1356         The win32 scope is not set when cross-compiling from Linux to Windows.
1357
1358         * qmake/mkspecs/features/functions.prf:
1359
1360 2012-09-18  Seokju Kwon  <seokju.kwon@samsung.com>
1361
1362         [EFL] Remove background view on EWebLauncher and MiniBrowser
1363         https://bugs.webkit.org/show_bug.cgi?id=96905
1364
1365         Reviewed by Gyuyoung Kim.
1366
1367         The size of webview was changed after adding url bar.
1368         And the background view is not necessary anymore, since it was used for debugging back in the day.
1369
1370         * EWebLauncher/main.c:
1371         (_ELauncher):
1372         (on_ecore_evas_resize):
1373         (browserCreate):
1374         * MiniBrowser/efl/main.c:
1375         (_MiniBrowser):
1376         (on_ecore_evas_resize):
1377         (browserCreate):
1378
1379 2012-09-18  Sergio Villar Senin  <svillar@igalia.com>
1380
1381         [GTK] run-webkit-tests unable to find TestExpectations for WK2
1382         https://bugs.webkit.org/show_bug.cgi?id=96998
1383
1384         Reviewed by Philippe Normand.
1385
1386         We should look for TestExpectations files in all the locations where
1387         we currently look for Skipped files. This will allow
1388         run-webkit-tests to look for TestExpectations files in
1389         platform/gtk-wk2 and platform/wk2 if the "-2" flag is used.
1390
1391         * Scripts/webkitpy/layout_tests/port/gtk.py:
1392         (GtkPort.expectations_files):
1393
1394 2012-09-18  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
1395
1396         [Qt] Fix build without the QtQuick module
1397
1398         Reviewed by Simon Hausmann.
1399
1400         * qmake/mkspecs/features/features.prf:
1401
1402 2012-09-18  Stephanie Lewis  <slewis@apple.com>
1403
1404         Build fix after http://trac.webkit.org/projects/webkit/changeset/128852.
1405
1406         Unreviewed.
1407
1408         * DumpRenderTree/mac/DumpRenderTree.mm:
1409         (poseAsClass):
1410
1411 2012-09-18  Allan Sandfeld Jensen  <allan.jensen@nokia.com>
1412
1413         Unreviewed update of email addresses for Berlin QtWebKit office.
1414
1415         * Scripts/webkitpy/common/config/committers.py:
1416
1417 2012-09-17  Zan Dobersek  <zandobersek@gmail.com>
1418
1419         [GTK] fast/loader/display-image-unset-can-block-image-and-can-reload-in-place.html failing after r128645
1420         https://bugs.webkit.org/show_bug.cgi?id=96899
1421
1422         Reviewed by Martin Robinson.
1423
1424         When overriding the 'WebKitDisplayImageKey' preference, use the
1425         'auto-load-images' property of WebKitWebSettings as the property which
1426         should be updated with the corresponding preference value.
1427
1428         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1429         (resetDefaultsToConsistentValues):
1430         * DumpRenderTree/gtk/TestRunnerGtk.cpp:
1431         (TestRunner::overridePreference):
1432
1433 2012-09-17  Pratik Solanki  <psolanki@apple.com>
1434
1435         DumpRenderTree and WebKitTestRunner should compile with -Wundef on Mac
1436         https://bugs.webkit.org/show_bug.cgi?id=96973
1437
1438         Reviewed by Dan Bernstein.
1439
1440         * DumpRenderTree/mac/Configurations/Base.xcconfig:
1441         * WebKitTestRunner/Configurations/Base.xcconfig:
1442         * WebKitTestRunner/PlatformWebView.h: Use #ifdef __OBJC__ and not #if.
1443
1444 2012-09-17  Dirk Pranke  <dpranke@chromium.org>
1445
1446         nrwt: remove "unexpected EOF" warnings
1447         https://bugs.webkit.org/show_bug.cgi?id=96970
1448
1449         Reviewed by Ojan Vafai.
1450
1451         After debugging this a bit, it looks like there aren't any cases
1452         that I can reproduce where a read() of zero indicates something
1453         actually wrong; either it is a prelude to a crash, or a false
1454         negative. So, I'm removing these warnings and adding a comment.
1455
1456         * Scripts/webkitpy/layout_tests/port/server_process.py:
1457         (ServerProcess._wait_for_data_and_update_buffers_using_select):
1458
1459 2012-09-17  Dirk Pranke  <dpranke@chromium.org>
1460
1461         [chromium] ASAN bot is crashing at the end of the run
1462         https://bugs.webkit.org/show_bug.cgi?id=96967
1463
1464         Reviewed by Abhishek Arya.
1465
1466         The ASAN bot is crashing attempting to decode some output into
1467         UTF-8; there's no reason to do this, so let's not do this and
1468         see if something else is going on as well.
1469
1470         * Scripts/webkitpy/layout_tests/port/chromium.py:
1471         (ChromiumPort._get_crash_log):
1472
1473 2012-09-17  Dirk Pranke  <dpranke@chromium.org>
1474
1475         nrwt: --results-directory isn't getting printed properly
1476         https://bugs.webkit.org/show_bug.cgi?id=96965
1477
1478         Reviewed by Ojan Vafai.
1479
1480         options.results_directory isn't actually initialized with the
1481         default values until after we call print_config(), so this
1482         changes things to print the value directly.
1483
1484         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1485         (run):
1486         * Scripts/webkitpy/layout_tests/views/printing.py:
1487         (Printer.print_config):
1488         * Scripts/webkitpy/layout_tests/views/printing_unittest.py:
1489         (Testprinter.test_print_config):
1490
1491 2012-09-17  Rob Buis  <rbuis@rim.com>
1492
1493         [BlackBerry] Enable VIDEO_TRACK
1494         https://bugs.webkit.org/show_bug.cgi?id=96949
1495
1496         Reviewed by Antonio Gomes.
1497
1498         * Scripts/webkitperl/FeatureList.pm:
1499
1500 2012-09-17  Rick Byers  <rbyers@chromium.org>
1501
1502         Add handling of new GestureTapCancel in DRT
1503
1504         https://bugs.webkit.org/show_bug.cgi?id=96183
1505
1506         Reviewed by Antonio Gomes.
1507
1508         * DumpRenderTree/chromium/TestWebPlugin.cpp:
1509         (TestWebPlugin::handleInputEvent):
1510         * DumpRenderTree/chromium/EventSender.cpp:
1511         (EventSender::gestureTapCancel):
1512
1513 2012-09-17  Philip Rogers  <pdr@google.com>
1514
1515         Teach style checker about preprocessor directive indentation rules
1516         https://bugs.webkit.org/show_bug.cgi?id=96874
1517
1518         Reviewed by Adam Barth.
1519
1520         Preprocessor directives (#ifdef, #include, #define, etc.) should not be indented.
1521         This is not explicit in our style guide but is generally followed in our code.
1522         Searching for violations in our codebase shows these are rarely indented:
1523             #include - indented in 6 files
1524             #ifdef - indented in 0 files
1525             #ifndef - indented in 1 file
1526             #define - indented in 11 files
1527             #if - indented in 7 files
1528
1529         * Scripts/webkitpy/style/checkers/cpp.py:
1530         (check_directive_indentation):
1531
1532             This is the simple test where we look for spaces followed by a #.
1533
1534         (check_style):
1535         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
1536
1537             A few tests needed to be modified because they had unintentionally indented
1538             preprocessor directives.
1539
1540         (CppStyleTest.test_build_class.Foo):
1541         (CppStyleTest.test_build_class):
1542         (CppStyleTest.test_build_class.DERIVE_FROM_GOO):
1543         (WebKitStyleTest.test_line_breaking):
1544         (WebKitStyleTest.test_directive_indentation):
1545
1546 2012-09-17  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
1547
1548         [Qt] Auto-generate the module pri file for QtWebKit
1549
1550         Reviewed by Simon Hausmann.
1551
1552         * qmake/qt_webkit.pri: Removed.
1553
1554 2012-09-14  Dirk Pranke  <dpranke@chromium.org>
1555
1556         nrwt: --additional-platform-dir is broken on chromium ports
1557         https://bugs.webkit.org/show_bug.cgi?id=96840
1558
1559         Reviewed by Ojan Vafai.
1560
1561         Looks like I broke this when cleaning things up as part of
1562         adding ML support. Fixing and adding a test.
1563
1564         * Scripts/webkitpy/layout_tests/port/chromium.py:
1565         (ChromiumPort.default_baseline_search_path):
1566         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
1567         (ChromiumAndroidPort.default_baseline_search_path):
1568         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1569         (ChromiumLinuxPort.default_baseline_search_path):
1570         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1571         (ChromiumMacPort.__init__):
1572         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1573         (ChromiumWinPort.setup_environ_for_server):
1574         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
1575         (test_check_build):
1576         (test_additional_platform_directory):
1577
1578 2012-09-13  Stephanie Lewis  <slewis@apple.com>
1579
1580         mac-future ignores previous platform TestExpectations.
1581         https://bugs.webkit.org/show_bug.cgi?id=96718.
1582
1583         Reviewed by Dirk Pranke.
1584
1585         TestExpectations match an expectation's configuraton against a list of configurations that are 
1586         valid for that port.  That list does not contain mac-future so none of the expectations can be
1587         applied.
1588
1589         * Scripts/webkitpy/layout_tests/port/apple.py:
1590         (ApplePort._generate_all_test_configurations):
1591
1592 2012-09-14  Adam Barth  <abarth@webkit.org>
1593
1594         Remove webkitPostMessage
1595         https://bugs.webkit.org/show_bug.cgi?id=96577
1596
1597         Reviewed by Ojan Vafai.
1598
1599         Add ENABLE_LEGACY_VENDOR_PREFIXES flag.
1600
1601         * Scripts/webkitperl/FeatureList.pm:
1602         * qmake/mkspecs/features/features.pri:
1603
1604 2012-09-14  Alexey Proskuryakov  <ap@apple.com>
1605
1606         [WK2] webarchive/loading/javascript-url-iframe-crash.html fails
1607         https://bugs.webkit.org/show_bug.cgi?id=96817
1608
1609         Reviewed by Kenneth Rohde Christiansen.
1610
1611         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1612         (WTR::InjectedBundlePage::didFailProvisionalLoadWithErrorForFrame): This callback
1613         got overlooked when adding logging elsewhere.
1614
1615 2012-09-14  Peter Beverloo  <peter@chromium.org>
1616
1617         [Chromium] Build fix for DumpRenderTree following r128628
1618         https://bugs.webkit.org/show_bug.cgi?id=96808
1619
1620         Reviewed by Tony Chang.
1621
1622         When building DumpRenderTree from the Chromium project, a compile error
1623         shows up because including the header ForwardIOStreamsAndroid.h cannot
1624         be located. TestWebKitAPI does this correctly.
1625
1626         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
1627
1628 2012-09-14  Dana Jansens  <danakj@chromium.org>
1629
1630         Change cc-bugs@google.com watchlist to cc-bugs@chromium.org
1631         https://bugs.webkit.org/show_bug.cgi?id=96805
1632
1633         Reviewed by Dirk Pranke.
1634
1635         * Scripts/webkitpy/common/config/committers.py:
1636         * Scripts/webkitpy/common/config/watchlist:
1637
1638 2012-09-14  Peter Beverloo  <peter@chromium.org>
1639
1640         [Chromium] Support the --{in,out,err}-fifo arguments on TestWebKitAPI and webkit_unit_tests
1641         https://bugs.webkit.org/show_bug.cgi?id=96687
1642
1643         Reviewed by Tony Chang.
1644
1645         Android's DumpRenderTree currently supports these arguments, implemented
1646         as part of TestShellAndroid:
1647         http://trac.webkit.org/browser/trunk/Tools/DumpRenderTree/chromium/TestShellAndroid.cpp?rev=128496
1648
1649         They're used by the layout test runner to get the STDOUT and STDERR while
1650         a layout test run is in process, which is a safer alternative to parsing
1651         all the logcat output manually. The implementation can be seen here:
1652         http://trac.webkit.org/browser/trunk/Tools/Scripts/webkitpy/layout_tests/port/chromium_android.py?rev=128496#L590
1653
1654         This patch generalizes parsing of and applying the effects of these arguments
1655         so that they can be used for TestWebKitAPI and webkit_unit_tests as well.
1656         After this patch, this will make it possible to pull out the output-reading
1657         code from Android's layout test port and generalize it so it can be re-used
1658         in the new test-runner for the other two test suites.
1659
1660         This has no effect when compiling and running these tests as part of Chromium
1661         code, which has a much more advanced test-runner that does parse complete log
1662         output, but also directly depends on code licensed under Apache 2.
1663
1664         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
1665         * DumpRenderTree/chromium/TestShellAndroid.cpp:
1666         (platformInit):
1667         * TestWebKitAPI/TestWebKitAPI.gyp/TestWebKitAPI.gyp:
1668
1669 2012-09-14  Zoltan Horvath  <zoltan@webkit.org>
1670
1671         check-webkit-style should not warn in case of NONCOPYABLE and FAST_ALLOCATED macros
1672         https://bugs.webkit.org/show_bug.cgi?id=96656
1673
1674         Reviewed by Adam Barth.
1675
1676         We should allow to use WTF_MAKE_NONCOPYABLE(ClassName) and WTF_MAKE_FAST_ALLOCATED
1677         macros in 1 line.
1678
1679         * Scripts/webkitpy/style/checkers/cpp.py:
1680         (check_style.definitions):
1681         (check_style): Add rule.
1682         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
1683         (WebKitStyleTest.test_line_breaking): Add unittest.
1684
1685 2012-09-14  Christophe Dumez  <christophe.dumez@intel.com>
1686
1687         WebKitTestRunner needs layoutTestController.dumpDatabaseCallbacks
1688         https://bugs.webkit.org/show_bug.cgi?id=57570
1689
1690         Reviewed by Kenneth Rohde Christiansen.
1691
1692         Implement testRunner.dumpDatabaseCallbacks in WebKitTestRunner
1693         and dump the information expected by the test cases.
1694
1695         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
1696         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
1697         (WTR::InjectedBundle::beginTesting): Explicitly set the default
1698         database quota.
1699         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1700         (WTR::securityOriginToStr): New convenience function to convert
1701         a security origin to the string expected in test results.
1702         (WTR):
1703         (WTR::InjectedBundlePage::InjectedBundlePage):
1704         (WTR::InjectedBundlePage::didExceedDatabaseQuota):
1705         (WTR::InjectedBundlePage::didReachApplicationCacheOriginQuota): Minor
1706         refactoring to share code with didExceedDatabaseQuota.
1707         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
1708         (InjectedBundlePage):
1709         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
1710         (WTR::TestRunner::TestRunner):
1711         * WebKitTestRunner/InjectedBundle/TestRunner.h:
1712         (WTR::TestRunner::dumpDatabaseCallbacks):
1713         (WTR::TestRunner::shouldDumpDatabaseCallbacks):
1714         (TestRunner):
1715
1716 2012-09-14  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
1717
1718         [Qt] Set force_static_libs_as_shared early enough to be caught by configure
1719
1720         Rubber-stamped by Simon Hausmann.
1721
1722         * qmake/mkspecs/features/default_post.prf:
1723
1724 2012-09-14  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
1725
1726         [Qt] Remove old cruft from the qmake build system
1727
1728         Rubber-stamped by Simon Hausmann.
1729
1730         * DumpRenderTree/qt/DumpRenderTree.pro:
1731         * qmake/mkspecs/features/default_post.prf:
1732
1733 2012-09-14  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
1734
1735         [Qt] Enable force_static_libs_as_shared for development
1736
1737         Not enabled for buildbots or production-builds (part of Qt5)
1738
1739         Reviewed by Simon Hausmann.
1740
1741         * qmake/mkspecs/features/default_post.prf:
1742
1743 2012-09-14  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
1744
1745         [Qt] Make force_static_libs_as_shared work on Mac OS
1746
1747         We had to move a few LIBS += around that were in the wrong place,
1748         and not caught when everything was just linked into the final
1749         QtWebKit library.
1750
1751         Reviewed by Simon Hausmann.
1752
1753         * qmake/mkspecs/features/default_post.prf:
1754         * qmake/mkspecs/features/force_static_libs_as_shared.prf:
1755
1756 2012-09-14  Simon Hausmann  <simon.hausmann@nokia.com>
1757
1758         [Qt] Fix build of WTR on Windows
1759
1760         Reviewed by Tor Arne Vestbø.
1761
1762         WTR/qt/main.cpp includes TestController.h, which lives in just WTR/
1763         With qmake's unix makefile generator as well as with the MingW generator,
1764         the directory where the pro file lives (WTR/) is automatically added to
1765         the include search path, but not so with MSVC. Instead of relying on qmake's
1766         implicit feature here, add WTR/ explicitly to the include search path.
1767
1768         * WebKitTestRunner/Target.pri:
1769
1770 2012-09-14  Christophe Dumez  <christophe.dumez@intel.com>
1771
1772         WebKitTestRunner needs layoutTestController.setStopProvisionalFrameLoads
1773         https://bugs.webkit.org/show_bug.cgi?id=42691
1774
1775         Reviewed by Kenneth Rohde Christiansen.
1776
1777         Add implementation for testrunner.setStopProvisionalFrameLoads()
1778         in WebKitTestRunner.
1779
1780         Original patch by Alexey Proskuryakov.
1781
1782         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
1783         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1784         (WTR::InjectedBundlePage::didStartProvisionalLoadForFrame):
1785         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
1786         (WTR::TestRunner::TestRunner):
1787         * WebKitTestRunner/InjectedBundle/TestRunner.h:
1788         (TestRunner):
1789         (WTR::TestRunner::setStopProvisionalFrameLoads):
1790         (WTR::TestRunner::shouldStopProvisionalFrameLoads):
1791
1792 2012-09-14  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
1793
1794         [Qt] Fix handling of debug/release/debug_and_release/build_all
1795
1796         We now pick up the defaults from Qt, in default_pre (CONFIG already
1797         contains the appropriate values for debug and release). We then let
1798         the command line arguments to qmake, or the project files themselves,
1799         override the configuration, and finally we sanitize the configuration
1800         in default_post.
1801
1802         Reviewed by Simon Hausmann.
1803
1804         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
1805         * Scripts/webkitdirs.pm:
1806         (buildQMakeProjects):
1807         * qmake/mkspecs/features/default_post.prf:
1808         * qmake/mkspecs/features/default_pre.prf:
1809         * qmake/mkspecs/features/mac/default_pre.prf:
1810
1811 2012-09-14  Sudarsana Nagineni  <sudarsana.nagineni@linux.intel.com>
1812
1813         WebKitTestRunner needs layoutTestController.setTabKeyCyclesThroughElements
1814         https://bugs.webkit.org/show_bug.cgi?id=42687
1815
1816         Reviewed by Kenneth Rohde Christiansen.
1817
1818         Implement testRunner.setTabKeyCyclesThroughElements for
1819         WebKitTestRunner.
1820
1821         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
1822         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
1823         (WTR::InjectedBundle::beginTesting):
1824         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
1825         (WTR::TestRunner::setTabKeyCyclesThroughElements):
1826         (WTR):
1827         * WebKitTestRunner/InjectedBundle/TestRunner.h:
1828         (TestRunner):
1829
1830 2012-09-14  Sudarsana Nagineni  <sudarsana.nagineni@linux.intel.com>
1831
1832         EventSendingController::keyDown does not support non-array modifier arguments
1833         https://bugs.webkit.org/show_bug.cgi?id=96727
1834
1835         Reviewed by Kenneth Rohde Christiansen.
1836
1837         Add support for handling a string as modifier argument to
1838         EventSendingController::keyDown.
1839
1840         * WebKitTestRunner/InjectedBundle/EventSendingController.cpp:
1841         (WTR::parseModifierArray):
1842
1843 2012-09-14  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
1844
1845         WebKitTestRunner needs layoutTestController.setCustomPolicyDelegate
1846         https://bugs.webkit.org/show_bug.cgi?id=42546
1847
1848         Reviewed by Kenneth Rohde Christiansen.
1849
1850         Exported TestRunner::setCustomPolicyDelegate() method.
1851         Allowed Policy Delegate dumping even if we're not in 'waitUntilDone' mode.
1852
1853         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
1854         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1855         (WTR::InjectedBundlePage::decidePolicyForNavigationAction):
1856
1857 2012-09-14  Zan Dobersek  <zandobersek@gmail.com>
1858
1859         [GTK] Clear application cache between tests in DumpRenderTree
1860         https://bugs.webkit.org/show_bug.cgi?id=96543
1861
1862         Reviewed by Philippe Normand.
1863
1864         Call the new DumpRenderTreeSupportGtk method to clear application cache
1865         after each test in DumpRenderTree.
1866
1867         Set the XDG_CACHE_HOME environment variable when running DumpRenderTree
1868         as well. It's now being set in XvfbDriver._start so that each driver is assigned
1869         a cache directory that's constructed out of the driver's name and its worker
1870         number. These directories are located in the layout tests results directory.
1871         At least on the builders this ensures a driver-specific empty cache directory
1872         with each test run.
1873
1874         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1875         (runTest):
1876         * Scripts/webkitpy/layout_tests/port/gtk.py:
1877         (GtkPort.setup_environ_for_server):
1878         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py:
1879         (XvfbDriver._start):
1880
1881 2012-09-14  Simon Hausmann  <simon.hausmann@nokia.com>
1882
1883         [Qt] Fix linking on Windows
1884
1885         Reviewed by Tor Arne Vestbø.
1886
1887         When compiling code _after_ QtWebKit.dll, we need to make sure that the QWEBKIT_EXPORT
1888         macro turns into an import macro, which means QT_MAKEDLL should only be set _up until_
1889         the QtWebKit.dll creation in the build process and be off afterwards. This is done in
1890         win32/default_post.prf using a !contains(QT, webkit): DEFINES += QT_MAKEDLL.
1891
1892         r128450 sanitized the values of QT and PKGCONFIG to only contain the dependencies we
1893         want for QtWebKit.dll, but unfortunately that code was applied to any binary built also
1894         after the module. Therefore QT did never contain webkit anymore and we always passed
1895         QT_MAKEDLL to DEFINES, resulting in QWEBKIT_EXPORT to _not_ turn into an import macro.
1896
1897         This patch introduces a creating_module configuration that is used to
1898         scope the qt/pkgconfig sanitization.
1899
1900         * qmake/mkspecs/features/default_post.prf:
1901
1902 2012-09-14  Christophe Dumez  <christophe.dumez@intel.com>
1903
1904         [WK2][WKTR] TestRunner is not dumping some frame load callbacks information
1905         https://bugs.webkit.org/show_bug.cgi?id=96740
1906
1907         Reviewed by Kenneth Rohde Christiansen.
1908
1909         Dump information expected by the test cases about the following
1910         frame load callbacks:
1911         "didDisplayInsecureContentForFrame"
1912         "didRunInsecureContentForFrame"
1913         "didDetectXSSForFrame"
1914
1915         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1916         (WTR::InjectedBundlePage::willPerformClientRedirectForFrame): Print a URL
1917         as expected by test results by calling pathSuitableForTestResult() instead
1918         of printing raw URL.
1919         (WTR::InjectedBundlePage::didDisplayInsecureContentForFrame):
1920         (WTR::InjectedBundlePage::didRunInsecureContentForFrame):
1921         (WTR::InjectedBundlePage::didDetectXSSForFrame):
1922
1923 2012-09-14  Peter Beverloo  <peter@chromium.org>
1924
1925         The runtime/unsigned category should be valid in the cpp style-checker.
1926         https://bugs.webkit.org/show_bug.cgi?id=96748
1927
1928         Reviewed by Csaba Osztrogonác.
1929
1930         This is causing the webkitpy tests to fail because it's not listed in the
1931         category array. Furthermore, change an "int" to "long" because it's
1932         throwing a double error in one of the earlier asserts.
1933
1934         * Scripts/webkitpy/style/checkers/cpp.py:
1935         (CppChecker):
1936         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
1937         (WebKitStyleTest.test_names):
1938
1939 2012-09-13  Kenneth Rohde Christiansen  <kenneth@webkit.org>
1940
1941         Evas_Object* is a ref'ed structure, so tread it as such
1942         https://bugs.webkit.org/show_bug.cgi?id=96659
1943
1944         Reviewed by Adam Barth.
1945
1946         Replace OwnPtr<Evas_Object> with RefPtr.
1947
1948         * DumpRenderTree/efl/ImageDiff.cpp:
1949         (calculateDifference):
1950         (printImageDifferences):
1951         (readImageFromStdin):
1952         (main):
1953
1954 2012-09-14  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
1955
1956         Unreviewed, rolling out r128507.
1957         http://trac.webkit.org/changeset/128507
1958         https://bugs.webkit.org/show_bug.cgi?id=96659
1959
1960         Revert. r128507 makes too many crashes in EFL layout test bots.
1961
1962         * DumpRenderTree/efl/ImageDiff.cpp:
1963         (calculateDifference):
1964         (printImageDifferences):
1965         (readImageFromStdin):
1966         (main):
1967
1968 2012-09-13  Kevin Funk  <kevin.funk@kdab.com>
1969
1970         Make compile with both OS(WINCE) and PLATFORM(QT) support
1971         https://bugs.webkit.org/show_bug.cgi?id=95536
1972
1973         Reviewed by Simon Hausmann.
1974
1975         Fix wince support in qmake files
1976
1977         * Tools.pro:
1978         * qmake/mkspecs/features/configure.prf:
1979         * qmake/mkspecs/features/default_post.prf:
1980         * qmake/mkspecs/features/features.prf:
1981         * qmake/mkspecs/features/functions.prf:
1982
1983 2012-09-13  KwangYong Choi  <ky0.choi@samsung.com>
1984
1985         [WK2] Add color picker API support for WebKit2
1986         https://bugs.webkit.org/show_bug.cgi?id=95058
1987
1988         Reviewed by Gyuyoung Kim.
1989
1990         Initialization code for color picker callbacks in UI client.
1991
1992         * MiniBrowser/mac/BrowserWindowController.m:
1993         (-[BrowserWindowController awakeFromNib]):
1994         * MiniBrowser/win/BrowserView.cpp:
1995         (BrowserView::create):
1996         * WebKitTestRunner/TestController.cpp:
1997         (WTR::TestController::createOtherPage):
1998         (WTR::TestController::initialize):
1999
2000 2012-09-13  Yuta Kitamura  <yutak@chromium.org>
2001
2002         Upgrade yutak to reviewer
2003         https://bugs.webkit.org/show_bug.cgi?id=96721
2004
2005         Reviewed by Kent Tamura.
2006
2007         * Scripts/webkitpy/common/config/committers.py:
2008
2009 2012-09-13  Adenilson Cavalcanti  <cavalcantii@gmail.com>
2010
2011         Make the style checker report usage of "unsigned int"
2012         https://bugs.webkit.org/show_bug.cgi?id=96693
2013
2014         Reviewed by Dirk Pranke.
2015
2016         Add a rule to check for uses of unsigned int, as the updated coding style recomends.
2017
2018         * Scripts/webkitpy/style/checkers/cpp.py:
2019         (check_language):
2020
2021 2012-09-13  Philip Rogers  <pdr@google.com>
2022
2023         Add gender-neutral form of webkit-patch land-cowboy
2024         https://bugs.webkit.org/show_bug.cgi?id=96709
2025
2026         Reviewed by Adam Barth.
2027
2028         Add webkit-patch land-cowhand, the confident and courageous form of webkit-patch for everyone.
2029
2030         * Scripts/webkitpy/tool/commands/download.py:
2031         (LandCowhand):
2032
2033 2012-09-13  Seokju Kwon  <seokju.kwon@samsung.com>
2034
2035         [EFL] Remove some parameters in browserCreate()
2036         https://bugs.webkit.org/show_bug.cgi?id=96499
2037
2038         Reviewed by Kenneth Rohde Christiansen.
2039
2040         There are too many parameters in browserCreate(). Lots of it seems like settings from command line argument.
2041         And I have moved some parameters into User_Arguments.
2042
2043         * EWebLauncher/main.c:
2044         (_User_Arguments):
2045         (_ELauncher):
2046         (on_key_down):
2047         (browserCreate):
2048         (findThemePath):
2049         (parseUserArguments):
2050         (main):
2051
2052 2012-09-13  Mark Lam  <mark.lam@apple.com>
2053
2054         Unreviewed. Adding myself to the committers list.
2055
2056         * Scripts/webkitpy/common/config/committers.py:
2057
2058 2012-09-13  Ryosuke Niwa  <rniwa@webkit.org>
2059
2060         run-perf-tests output cryptic error when the config file is missing.
2061         https://bugs.webkit.org/show_bug.cgi?id=96453
2062
2063         Reviewed by Tony Chang.
2064
2065         Add a special error message when a configuration file is missing.
2066
2067         Also update the help message of --source-json-path to signify the fact it specifies
2068         the configuration file on performance tests bots.
2069
2070         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
2071         (PerfTestsRunner._parse_args):
2072         (PerfTestsRunner._generate_and_show_results):
2073         (PerfTestsRunner._merge_slave_config_json):
2074         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
2075         (_test_run_with_json_output): Make upload to fail by default and assert the result
2076         in the function so that we can return logs instead.
2077         (_test_run_with_json_output.mock_upload_json):
2078         (test_run_with_json_output):
2079         (test_run_with_description):
2080         (test_run_respects_no_results):
2081         (test_run_with_slave_config_json):
2082         (test_run_with_bad_slave_config_json):
2083         (test_run_with_multiple_repositories):
2084
2085 2012-09-13  Kenneth Rohde Christiansen  <kenneth@webkit.org>
2086
2087         Evas_Object* is a ref'ed structure, so tread it as such
2088         https://bugs.webkit.org/show_bug.cgi?id=96659
2089
2090         Reviewed by Adam Barth.
2091
2092         Replace OwnPtr<Evas_Object> with RefPtr.
2093
2094         * DumpRenderTree/efl/ImageDiff.cpp:
2095         (calculateDifference):
2096         (printImageDifferences):
2097         (readImageFromStdin):
2098         (main):
2099
2100 2012-09-13  Csaba Osztrogonác  <ossy@webkit.org>
2101
2102         One more unreviewed trivial fix after r128399.
2103
2104         * BuildSlaveSupport/build.webkit.org-config/config.json: Fix URLs of the SVN mirror server.
2105
2106 2012-09-13  Csaba Osztrogonác  <ossy@webkit.org>
2107
2108         Unreviewed trivial fix after r128399.
2109
2110         * BuildSlaveSupport/wait-for-SVN-server.py: Renamed from Tools/BuildSlaveSupport/build.webkit.org-config/wait-for-SVN-server.py.
2111         (getLatestSVNRevision):
2112         (waitForSVNRevision):
2113
2114 2012-09-13  Sudarsana Nagineni  <sudarsana.nagineni@linux.intel.com>
2115
2116         [WK2][WTR] WebKitTestRunner needs testRunner.callShouldCloseOnWebView
2117         https://bugs.webkit.org/show_bug.cgi?id=96366
2118
2119         Reviewed by Anders Carlsson.
2120
2121         Implement testRunner.callShouldCloseOnWebView for WebKitTestRunner
2122         by calling shouldClose() on the FrameLoader.
2123
2124         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
2125         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
2126         (WTR::TestRunner::callShouldCloseOnWebView):
2127         (WTR):
2128         * WebKitTestRunner/InjectedBundle/TestRunner.h:
2129         (TestRunner):
2130
2131 2012-09-13  Sheriff Bot  <webkit.review.bot@gmail.com>
2132
2133         Unreviewed, rolling out r128453.
2134         http://trac.webkit.org/changeset/128453
2135         https://bugs.webkit.org/show_bug.cgi?id=96681
2136
2137         Having tests use the same appcache directory leads to timeouts
2138         (Requested by zdobersek on #webkit).
2139
2140         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2141         (runTest):
2142         * Scripts/webkitpy/layout_tests/port/gtk.py:
2143         (GtkPort.setup_environ_for_server):
2144
2145 2012-09-13  James Robinson  <jamesr@chromium.org>
2146
2147         [chromium] Temporarily remove Android Builder (dbg) from garden-o-matic view
2148         https://bugs.webkit.org/show_bug.cgi?id=96678
2149
2150         Reviewed by Dirk Pranke.
2151
2152         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/config.js:
2153         (.):
2154
2155 2012-09-13  Rob Buis  <rbuis@rim.com>
2156
2157         Unreviewed changes to watchlist.
2158
2159         - Add myself to BlackBerry bugs
2160
2161         * Scripts/webkitpy/common/config/watchlist:
2162
2163 2012-09-13  Stephen Chenney  <schenney@chromium.org>
2164
2165         Add an SVG entry to the watchlist
2166
2167         Unreviewed config change.
2168
2169         * Scripts/webkitpy/common/config/watchlist:
2170         - Created a new SVG watchlist label covering WebCore/svg and WebCore/rendering/svg
2171         - Added myself and pdr as watchers
2172
2173 2012-09-13  Simon Hausmann  <simon.hausmann@nokia.com>
2174
2175         [Qt] Make WebKit2 work on Windows with Qt 5
2176         https://bugs.webkit.org/show_bug.cgi?id=76776
2177
2178         Reviewed by Tor Arne Vestbø.
2179
2180         WebKit2 should compile with Qt on Windows now, so enable it. This patch in fact corrects
2181         the check to disable WebKit when we don't have USE_3D_GRAPHICS available, because we need
2182         that for the GL TextureMapper (it's only built when that feature is set).
2183
2184         * qmake/mkspecs/features/configure.prf:
2185
2186 2012-09-13  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2187
2188         [Qt] Ensure that QT is finalized before loading qt_module.prf
2189
2190         As qt_module.prf does dependency tracking based on the content of the
2191         QT variable. The intermediate WebKit modules modify the variable in
2192         their .pri files, so we have to ensure we've loaded all the modules
2193         before loading qt_module.
2194
2195         Reviewed by Tor Arne Vestbø.
2196         Patch by Oswald Buddenhagen <oswald.buddenhagen@nokia.com> on 2012-09-13
2197
2198         * qmake/mkspecs/features/default_post.prf:
2199         * qmake/mkspecs/features/webkit_modules.prf: Added.
2200
2201 2012-09-13  Simon Hausmann  <simon.hausmann@nokia.com>
2202
2203         [Qt] Fix build of Qt WTR/MiniBrowser with MSVC and nmake
2204         https://bugs.webkit.org/show_bug.cgi?id=96633
2205
2206         Reviewed by Csaba Osztrogonác.
2207
2208         WTR and MiniBrowser both have DRT in their include path (for the shared font stuff). Batch
2209         builds with MSVC and nmake result in nmake picking up main.cpp from the wrong directory.
2210         It's the usual problem and the usual solution is to rename one of the conflicting files.
2211         This patch chooses to rename DRT instead of WTR and MBR.
2212
2213         Fixed include order at the same time to match style guide.
2214
2215         * DumpRenderTree/qt/DumpRenderTree.pro:
2216         * DumpRenderTree/qt/DumpRenderTreeMain.cpp: Renamed from Tools/DumpRenderTree/qt/main.cpp.
2217         (messageHandler):
2218         (isOption):
2219         (takeOptionValue):
2220         (printUsage):
2221         (main):
2222
2223 2012-09-13  Kenneth Rohde Christiansen  <kenneth@webkit.org>
2224
2225         The Script run-launcher should support -2 in combination with --efl
2226         https://bugs.webkit.org/show_bug.cgi?id=96639
2227
2228         Reviewed by Simon Hausmann.
2229
2230         Support running the MiniBrowser then supplied -2 as argument.
2231
2232         * Scripts/run-launcher:
2233
2234 2012-09-13  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
2235
2236         Unreviewed, add my Intel email address to the list.
2237
2238         * Scripts/webkitpy/common/config/committers.py:
2239
2240 2012-09-13  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2241
2242         [Qt] Don't overwrite LIBS_PRIVATE when sanitizing LIBS
2243
2244         Reviewed by Ossy.
2245
2246         * qmake/mkspecs/features/default_post.prf:
2247
2248 2012-09-13  Kenneth Rohde Christiansen  <kenneth@webkit.org>
2249
2250         [EFL][DRT] Implement LayoutTestController::layerTreeAsText
2251         https://bugs.webkit.org/show_bug.cgi?id=82294
2252
2253         Reviewed by Gyuyoung Kim.
2254
2255         Implement layerTreeAsText using the DumpRenderTreeSupportEfl method.
2256
2257         * DumpRenderTree/efl/TestRunnerEfl.cpp:
2258         (TestRunner::layerTreeAsText):
2259
2260 2012-09-13  Zan Dobersek  <zandobersek@gmail.com>
2261
2262         [GTK] Clear application cache between tests in DumpRenderTree
2263         https://bugs.webkit.org/show_bug.cgi?id=96543
2264
2265         Reviewed by Philippe Normand.
2266
2267         Call the new DumpRenderTreeSupportGtk method to clear application cache
2268         after each test.
2269
2270         Set the XDG_CACHE_HOME environment variable when running DumpRenderTree
2271         as well. Currently it points to a subdirectory in the layout tests results
2272         directory. At least on the builders this ensures an empty cache directory
2273         with each test run.
2274
2275         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2276         (runTest):
2277         * Scripts/webkitpy/layout_tests/port/gtk.py:
2278         (GtkPort.setup_environ_for_server):
2279
2280 2012-09-13  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2281
2282         [Qt] Make sure LIBS, PKGCONFIG, and QT, do not affect QtWebKit's prl file
2283
2284         Unless it's actually a dependency of the public QtWebKit API.
2285
2286         We could replace every occurance of LIBS, PKGCONFIG, and QT, in the pri
2287         files with their _PRIVATE equivivalent, but that's likely to break when
2288         someone adds a new QT/LIBS/PKGCONFIG += foo line somewhere. Instead we
2289         clean up the variables in default_post.
2290
2291         This means that the CONFIG -= explicitlib in linkAgainstLibrary() is no
2292         longer needed, as it was a workaround for the intermediate libraries ending
2293         up in the prl file. And, since CONFIG -= staticlib was there to support
2294         the explicitlib option, we can remove that too (as well as the exception
2295         for gprof, since it would be empty). If gprof needs tweaks to always link
2296         statically they should go in gprof.prf.
2297
2298         Reviewed by Simon Hausmann.
2299
2300         * qmake/mkspecs/features/default_post.prf:
2301         * qmake/mkspecs/features/functions.prf:
2302
2303 2012-09-13  Mario Sanchez Prada  <msanchez@igalia.com>
2304
2305         [GTK] "Infinite" loop in AccessibilityUIElementGtk.cpp
2306         https://bugs.webkit.org/show_bug.cgi?id=96632
2307
2308         Reviewed by Carlos Garcia Campos.
2309
2310         Fix this by using atk_object_get_n_accessible_children instead of
2311         calling getChildren() from childCount.
2312
2313         * WebKitTestRunner/InjectedBundle/gtk/AccessibilityUIElementGtk.cpp:
2314         (WTR::AccessibilityUIElement::childrenCount): Avoid the infinite
2315         loop by using atk_object_get_n_accessible_children().
2316
2317 2012-09-13  Simon Hausmann  <simon.hausmann@nokia.com>
2318
2319         [Qt][Win] Enable USE(3D_GRAPHICS)
2320         https://bugs.webkit.org/show_bug.cgi?id=90879
2321
2322         Reviewed by Kenneth Rohde Christiansen.
2323
2324         Enable it on Windows as well as on QNX (bug fixed).
2325
2326         * qmake/mkspecs/features/features.prf:
2327
2328 2012-09-13  Simon Hausmann  <simon.hausmann@nokia.com>
2329
2330         [Qt] The ANGLE build on Windows breaks with GnuWin32's flex because it is too old
2331         https://bugs.webkit.org/show_bug.cgi?id=96359
2332
2333         Reviewed by Tor Arne Vestbø.
2334
2335         GnuWin32's flex is 2.5.4 and that is too old for ANGLE's preprocessor
2336         lexer. The only viable alternative I could find on Windows that provides
2337         binaries is the winflexbison distribution (http://sourceforge.net/projects/winflexbison/),
2338         so let's use that one instead.
2339
2340         * Scripts/webkitdirs.pm:
2341         (checkRequiredSystemConfig):
2342         * qmake/mkspecs/features/default_pre.prf:
2343
2344 2012-09-13  Simon Hausmann  <simon.hausmann@nokia.com>
2345
2346         [Qt] Add configuration for production builds
2347         https://bugs.webkit.org/show_bug.cgi?id=96607
2348
2349         Reviewed by Kenneth Rohde Christiansen.
2350
2351         Add configuration tweaks for production builds, to ease memory pressure with the GNU toolchain
2352         and enable all-in-one-files optimizations.
2353
2354         * qmake/mkspecs/features/production_build.prf: Added.
2355
2356 2012-09-13  Simon Hausmann  <simon.hausmann@nokia.com>
2357
2358         [Qt] GStreamer detection does not work when cross-compiling
2359         https://bugs.webkit.org/show_bug.cgi?id=96608
2360
2361         Reviewed by Kenneth Rohde Christiansen.
2362
2363         When cross-compiling it is not safe to just use system(pkg-config),
2364         because there is a high probability that it will find packages in the
2365         host system instead of in the target one. Instead use qmake's new
2366         packagesExist() function that takes care of this delicate aspect of
2367         cross compilation by setting PKG_CONFIG_SYSROOT_DIR and
2368         PKG_CONFIG_LIBDIR accordingly.
2369
2370         * qmake/mkspecs/features/features.prf:
2371
2372 2012-09-13  Christophe Dumez  <christophe.dumez@intel.com>
2373
2374         [WK2][WTR] More efficient string handling in InjectedBundlePage
2375         https://bugs.webkit.org/show_bug.cgi?id=96609
2376
2377         Reviewed by Kenneth Rohde Christiansen.
2378
2379         Update StringBuilder usage in InjectedBundlePage to follow
2380         the latest recommandations for efficient string handling:
2381         http://trac.webkit.org/wiki/EfficientStrings
2382
2383         This involves using appendLiteral() or appendNumber()
2384         instead of append() whenever possible.
2385
2386         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2387         (WTR::dumpPath):
2388         (WTR::rangeToStr):
2389         (WTR::styleDecToStr):
2390         (WTR::frameToStr):
2391         (WTR::dumpResourceURL):
2392         (WTR):
2393         (WTR::dumpFrameDescriptionSuitableForTestResult):
2394         (WTR::dumpRequestDescriptionSuitableForTestResult):
2395         (WTR::dumpResponseDescriptionSuitableForTestResult):
2396         (WTR::dumpErrorDescriptionSuitableForTestResult):
2397         (WTR::InjectedBundlePage::didReceiveIntentForFrame):
2398         (WTR::InjectedBundlePage::registerIntentServiceForFrame):
2399         (WTR::InjectedBundlePage::didStartProvisionalLoadForFrame):
2400         (WTR::InjectedBundlePage::didReceiveServerRedirectForProvisionalLoadForFrame):
2401         (WTR::InjectedBundlePage::didCommitLoadForFrame):
2402         (WTR::InjectedBundlePage::didFinishProgress):
2403         (WTR::dumpFrameScrollPosition):
2404         (WTR::dumpFrameText):
2405         (WTR::dumpDescendantFramesText):
2406         (WTR::InjectedBundlePage::didFinishLoadForFrame):
2407         (WTR::InjectedBundlePage::didReceiveTitleForFrame):
2408         (WTR::InjectedBundlePage::didCancelClientRedirectForFrame):
2409         (WTR::InjectedBundlePage::willPerformClientRedirectForFrame):
2410         (WTR::InjectedBundlePage::didFinishDocumentLoadForFrame):
2411         (WTR::InjectedBundlePage::didHandleOnloadEventsForFrame):
2412         (WTR::InjectedBundlePage::willSendRequestForFrame):
2413         (WTR::InjectedBundlePage::didReceiveResponseForResource):
2414         (WTR::InjectedBundlePage::didFinishLoadForResource):
2415         (WTR::InjectedBundlePage::didFailLoadForResource):
2416         (WTR::InjectedBundlePage::shouldCacheResponse):
2417         (WTR::InjectedBundlePage::decidePolicyForNavigationAction):
2418         (WTR::InjectedBundlePage::willAddMessageToConsole):
2419         (WTR::InjectedBundlePage::willSetStatusbarText):
2420         (WTR::InjectedBundlePage::willRunJavaScriptAlert):
2421         (WTR::InjectedBundlePage::willRunJavaScriptConfirm):
2422         (WTR::InjectedBundlePage::willRunJavaScriptPrompt):
2423         (WTR::InjectedBundlePage::didReachApplicationCacheOriginQuota):
2424         (WTR::InjectedBundlePage::shouldBeginEditing):
2425         (WTR::InjectedBundlePage::shouldEndEditing):
2426         (WTR::InjectedBundlePage::shouldInsertNode):
2427         (WTR::InjectedBundlePage::shouldInsertText):
2428         (WTR::InjectedBundlePage::shouldDeleteRange):
2429         (WTR::InjectedBundlePage::shouldChangeSelectedRange):
2430         (WTR::InjectedBundlePage::shouldApplyStyle):
2431         (WTR::InjectedBundlePage::didBeginEditing):
2432         (WTR::InjectedBundlePage::didEndEditing):
2433         (WTR::InjectedBundlePage::didChange):
2434         (WTR::InjectedBundlePage::didChangeSelection):
2435         (WTR::InjectedBundlePage::supportsFullScreen):
2436         (WTR::InjectedBundlePage::enterFullScreenForElement):
2437         (WTR::InjectedBundlePage::exitFullScreenForElement):
2438         (WTR::InjectedBundlePage::beganEnterFullScreen):
2439         (WTR::InjectedBundlePage::beganExitFullScreen):
2440         (WTR::InjectedBundlePage::closeFullScreen):
2441         (WTR::dumpBackForwardListItem):
2442         (WTR::InjectedBundlePage::dumpBackForwardList):
2443
2444 2012-09-13  Sudarsana Nagineni  <sudarsana.nagineni@linux.intel.com>
2445
2446         [Qt][WK2] Memory leaks in Qt's TestRunner
2447         https://bugs.webkit.org/show_bug.cgi?id=96603
2448
2449         Reviewed by Csaba Osztrogonác.
2450
2451         Fix memory leaks in Qt's TestRunner code by adopting WKString
2452         created with WKStringCreateWithUTF8CString().
2453
2454         * WebKitTestRunner/qt/TestControllerQt.cpp:
2455         (WTR::TestController::initializeInjectedBundlePath):
2456         (WTR::TestController::initializeTestPluginDirectory):
2457
2458 2012-09-13  Zan Dobersek  <zandobersek@gmail.com>
2459
2460         [GTK][NRWT] Lower timeout value to 6 seconds (12 for debug configuration)
2461         https://bugs.webkit.org/show_bug.cgi?id=96506
2462
2463         Reviewed by Dirk Pranke.
2464
2465         Lower the timeout value to 6 seconds for release configuration and 12 seconds
2466         for debug configuration. These are the same values the Chromium port uses.
2467
2468         Currently the value is lowered only for the DumpRenderTree, WebKitTestRunner (and
2469         the layout tests under WebKit2 generally) requires more attention and gardening
2470         than what the current effort produces.
2471
2472         The unit test is updated as well to test these changes.
2473
2474         * Scripts/webkitpy/layout_tests/port/gtk.py:
2475         (GtkPort.default_timeout_ms):
2476         * Scripts/webkitpy/layout_tests/port/gtk_unittest.py:
2477         (GtkPortTest.test_default_timeout_ms):
2478         (GtkPortTest.assertLinesEqual):
2479
2480 2012-09-12  Csaba Osztrogonác  <ossy@webkit.org>
2481
2482         Add SVN mirror handling feature to build.webkit.org
2483         https://bugs.webkit.org/show_bug.cgi?id=85887
2484
2485         Reviewed by Dirk Pranke.
2486
2487         * BuildSlaveSupport/build.webkit.org-config/config.json: Add SVN mirrors for Qt buildslaves hosted in Szeged.
2488         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
2489         (ConfigureBuild.__init__): Add SVNMirror property to be able to watch which slaves use which mirror on build.webkit.org.
2490         (ConfigureBuild.start):
2491         (CheckOutSource.__init__): Set baseURL to the SVNMirror or the default http://svn.webkit.org/repository/webkit/
2492         (WaitForSVNServer): Add new buildstep for waiting the SVNMirror to be in sync with http://svn.webkit.org/repository/webkit/
2493         (Factory.__init__):
2494         (BuildFactory.__init__):
2495         (TestFactory.__init__):
2496         (BuildAndTestFactory.__init__):
2497         (BuildAndPerfTestFactory.__init__):
2498         (BuildAndPerfTestWebKit2Factory.__init__):
2499         (DownloadAndPerfTestFactory.__init__):
2500         (DownloadAndPerfTestWebKit2Factory.__init__):
2501         (loadBuilderConfig): Use kwargs instead of args to be able to add the optional SVNMirror factory argument.
2502         * BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:
2503         (SVNMirrorTest):
2504         (SVNMirrorTest.setUp): Load config.json once at startup.
2505         (SVNMirrorTest.get_SVNMirrorFromConfig): Get the SVNMirror from config.json for a given buildslave.
2506         (SVNMirrorTest.test_CheckOutSource): Compare CheckOutSource.baseURL with SVNMirror in config.json for all builders
2507         * BuildSlaveSupport/build.webkit.org-config/wait-for-SVN-server.py: Added.
2508         (getLatestSVNRevision): Get the latest SVN revison from the given server.
2509         (waitForSVNRevision): Wait until the given SVN revision is committed to the given server. It doesn't wait if the
2510                               SVN revision is empty (force build triggered without revision) or the server is unavailable.
2511
2512 2012-09-12  Kangil Han  <kangil.han@samsung.com>
2513
2514         [WK2][WTR] Set waitUntilDone watchdog timer value equal to WK1.
2515         https://bugs.webkit.org/show_bug.cgi?id=81606
2516
2517         Reviewed by Tim Horton.
2518
2519         6 seconds is harsh for some jquery test cases.
2520         Therefore, adjust the value to match DumpRenderTree.
2521
2522         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
2523         (WTR):
2524
2525 2012-09-12  KyungTae Kim  <ktf.kim@samsung.com>
2526
2527         [EFL] Support download attribute feature
2528         https://bugs.webkit.org/show_bug.cgi?id=96462
2529
2530         Reviewed by Gyuyoung Kim.
2531
2532         Enable DOWNLOAD_ATTRIBUTE feature for EFL port,
2533         and add "download,request" callback to the DumpRenderTree for EFL.
2534
2535         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
2536         (DumpRenderTreeChrome::createView):
2537         (DumpRenderTreeChrome::onDownloadRequest):
2538         * DumpRenderTree/efl/DumpRenderTreeChrome.h:
2539         (DumpRenderTreeChrome):
2540         * Scripts/webkitperl/FeatureList.pm:
2541
2542 2012-09-12  Stephanie Lewis  <slewis@apple.com>
2543
2544         Reduce parallism on the wk2 testers.
2545         https://bugs.webkit.org/show_bug.cgi?id=95906
2546
2547         Reviewed by Dirk Pranke.
2548
2549         The wk2 testers on Mountain Lion are getting stuck and timing out.  The problem appears to be due to
2550         resource contention.  Reducing the number of processes alleviates the issue.  
2551
2552         Starting by reducing 25%.
2553
2554         * Scripts/webkitpy/layout_tests/port/mac.py:
2555         (MacPort.default_child_processes):
2556
2557 2012-09-12  Brady Eidson  <beidson@apple.com>
2558
2559         Assert in NetscapePlugin::destroy() with async plugin init
2560         <rdar://problem/12277595> and https://bugs.webkit.org/show_bug.cgi?id=96576
2561
2562         Reviewed by Anders Carlsson.
2563
2564         Expose NPN_Invoke to plug-in tests:
2565         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
2566         (PluginTest::NPN_Invoke):
2567         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
2568         (PluginTest):
2569
2570         Add a test that uses NPN_Invoke on the window object from inside NPP_New to remove the plug-in element:
2571         * DumpRenderTree/TestNetscapePlugIn/Tests/InvokeDestroysPluginWithinNPP_New.cpp: Added.
2572         (InvokeDestroysPluginWithinNPP_New):
2573         (InvokeDestroysPluginWithinNPP_New::InvokeDestroysPluginWithinNPP_New):
2574         (InvokeDestroysPluginWithinNPP_New::NPP_New):
2575         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
2576
2577 2012-09-12  Dirk Pranke  <dpranke@chromium.org>
2578
2579         refactor TestExpectations tokenization slightly in preparation for the new syntax
2580         https://bugs.webkit.org/show_bug.cgi?id=96564
2581
2582         Reviewed by Ojan Vafai.
2583
2584         This change changes how we will tokenize/lex the
2585         TestExpectations files, in preparation for supporting both the
2586         old syntax and the new one. This change by itself should be a
2587         no-op.
2588
2589         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2590         (TestExpectationParser._tokenize_line):
2591         (TestExpectationParser):
2592         (TestExpectationParser._tokenize_line_using_old_format):
2593         (TestExpectationParser._tokenize_line_using_new_format):
2594
2595 2012-09-12  Dirk Pranke  <dpranke@chromium.org>
2596
2597         remove unused testexpectations editing code from webkitpy, garden-o-matic
2598         https://bugs.webkit.org/show_bug.cgi?id=96562
2599
2600         Reviewed by Adam Barth.
2601
2602         This code was prototyped but never fully realized. Removing
2603         until it reanimates :).
2604
2605         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/checkout.js:
2606         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/checkout_unittests.js:
2607         * Scripts/webkitpy/layout_tests/controllers/test_expectations_editor.py: Removed.
2608         * Scripts/webkitpy/layout_tests/controllers/test_expectations_editor_unittest.py: Removed.
2609         * Scripts/webkitpy/tool/servers/gardeningserver.py:
2610         (BuildCoverageExtrapolator.extrapolate_test_configurations):
2611         (GardeningHTTPRequestHandler._run_webkit_patch):
2612         (GardeningHTTPRequestHandler.ping):
2613         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py:
2614         (BuildCoverageExtrapolatorTest.test_extrapolate):
2615         (GardeningServerTest.test_rebaseline_new_port):
2616
2617 2012-09-12  Lucas Forschler  <lforschler@apple.com>
2618
2619         Remove unused bot, apple-xserve-1.
2620         https://bugs.webkit.org/show_bug.cgi?id=96570
2621         
2622         Reviewed by Stephanie Lewis.
2623
2624         * BuildSlaveSupport/build.webkit.org-config/config.json:
2625
2626 2012-09-12  Xan Lopez  <xlopez@igalia.com>
2627
2628         [GTK] We attempt to rebase documentation even if it's not present
2629         https://bugs.webkit.org/show_bug.cgi?id=96553
2630
2631         Reviewed by Martin Robinson.
2632
2633         Do not make the documentation rebase step fatal. This allows make
2634         install to succeed when there's no documentation generated.
2635
2636         * gtk/generate-gtkdoc:
2637
2638 2012-09-12  Kenneth Rohde Christiansen  <kenneth@webkit.org>
2639
2640         Respect WEBKITOUTPUTDIR when running EFL tests
2641         https://bugs.webkit.org/show_bug.cgi?id=96528
2642
2643         Reviewed by Dirk Pranke.
2644
2645         Expose user set WEBKITOUTPUTDIR to the web process.
2646
2647         * Scripts/webkitpy/layout_tests/port/driver.py:
2648         (Driver._start): Add WEBKITOUTPUTDIR to the environment
2649         of the web process and its potential jhbuild wrapper.
2650
2651 2012-09-12  Dirk Pranke  <dpranke@chromium.org>
2652
2653         webkitdirs: fix uname version handling for cygwin
2654         https://bugs.webkit.org/show_bug.cgi?id=96436
2655
2656         Reviewed by Jon Honeycutt.
2657
2658         Newer versions of cygwin embed an additional version string
2659         inside parentheses, so you get "1.7.16(0.249/5/3)" instead of "1.7.16".
2660         Update the code to handle that.
2661
2662         * Scripts/webkitdirs.pm:
2663         (setupAppleWinEnv):
2664
2665 2012-09-12  Christophe Dumez  <christophe.dumez@intel.com>
2666
2667         Regression(r128338): Broke Windows build
2668         https://bugs.webkit.org/show_bug.cgi?id=96537
2669
2670         Unreviewed build fix.
2671
2672         Fix Apple-Win build by allocating array dynamically
2673         since its size is not constant.
2674
2675         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
2676         (WTR::stringArrayToJS):
2677
2678 2012-09-12  Alex Sakhartchouk  <alexst@chromium.org>
2679
2680         [chromium] Add a virtual test suite for fast/hidpi
2681         https://bugs.webkit.org/show_bug.cgi?id=90192
2682
2683         Reviewed by Dirk Pranke.
2684
2685         Add a virtual test suite to make sure the pixel tests in fast/hidpi give the same result
2686         on the hardware accelerated path as the software path.
2687
2688         * DumpRenderTree/chromium/WebPreferences.cpp:
2689         (WebPreferences::applyTo):
2690         * Scripts/webkitpy/layout_tests/port/chromium.py:
2691         (ChromiumPort.virtual_test_suites):
2692
2693 2012-09-12  Adam Klein  <adamk@chromium.org>
2694
2695         [chromium] Add content_browsertests to the flakiness dashboard
2696         https://bugs.webkit.org/show_bug.cgi?id=96535
2697
2698         Reviewed by Ojan Vafai.
2699
2700         * TestResultServer/static-dashboards/dashboard_base.js:
2701
2702 2012-09-12  Christophe Dumez  <christophe.dumez@intel.com>
2703
2704         [WK2][WKTR] TestRunner needs to implement originsWithApplicationCache
2705         https://bugs.webkit.org/show_bug.cgi?id=96496
2706
2707         Reviewed by Kenneth Rohde Christiansen.
2708
2709         Add implementation for originsWithApplicationCache to
2710         WebKitTestRunner.
2711
2712         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
2713         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
2714         (WTR::stringArrayToJS):
2715         (WTR):
2716         (WTR::TestRunner::originsWithApplicationCache):
2717         * WebKitTestRunner/InjectedBundle/TestRunner.h:
2718         (TestRunner):
2719
2720 2012-09-12  Simon Hausmann  <simon.hausmann@nokia.com>
2721
2722         Fix Qt/Windows build with Python3
2723         https://bugs.webkit.org/show_bug.cgi?id=96473
2724
2725         Reviewed by Csaba Osztrogonác.
2726
2727         In Python 3 print is a real function, so we must use parentheses around
2728         the function parameters. This is backwards compatible with Python 2.
2729
2730         * Scripts/generate-win32-export-forwards:
2731
2732 2012-09-12  Simon Hausmann  <simon.hausmann@nokia.com>
2733
2734         Unreviewed trivial build fix for Qt/Windows after bug #96358.
2735
2736         As pointed out by Kevin, we should inject the "set PATH=..." statement
2737         to add the GnuWin32 directory only if that directory actually exists.
2738
2739         * qmake/mkspecs/features/default_post.prf:
2740
2741 2012-09-12  Sudarsana Nagineni  <sudarsana.nagineni@linux.intel.com>
2742
2743         [EFL] [WK2] Memory leaks in TestControllerEfl
2744         https://bugs.webkit.org/show_bug.cgi?id=96525
2745
2746         Reviewed by Kenneth Rohde Christiansen.
2747
2748         Fix memory leaks in EFL's TestRunner code by adopting an allocation
2749         of WKString created with WKStringCreateWithUTF8CString().
2750
2751         * WebKitTestRunner/efl/TestControllerEfl.cpp:
2752         (WTR::TestController::initializeInjectedBundlePath):
2753         (WTR::TestController::initializeTestPluginDirectory):
2754
2755 2012-09-12  Christophe Dumez  <christophe.dumez@intel.com>
2756
2757         [WK2][WKTR] TestRunner needs to implement dumpApplicationCacheDelegateCallbacks
2758         https://bugs.webkit.org/show_bug.cgi?id=96374
2759
2760         Reviewed by Kenneth Rohde Christiansen.
2761
2762         Implement support for dumpApplicationCacheDelegateCallbacks
2763         and disallowIncreaseForApplicationCacheQuota in
2764         WebKitTestRunner and properly dump the information
2765         expected by the tests.
2766
2767         If the application cache quota is reached for a given
2768         security origin, WebKitTestRunner will reset the quota
2769         to its default value, unless intructed not to via
2770         disallowIncreaseForApplicationCacheQuota().
2771
2772         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
2773         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2774         (WTR::InjectedBundlePage::InjectedBundlePage):
2775         (WTR::InjectedBundlePage::didReachApplicationCacheOriginQuota):
2776         (WTR):
2777         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
2778         (InjectedBundlePage):
2779         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
2780         (WTR::TestRunner::TestRunner):
2781         (WTR::TestRunner::disallowIncreaseForApplicationCacheQuota):
2782         (WTR):
2783         * WebKitTestRunner/InjectedBundle/TestRunner.h:
2784         (WTR::TestRunner::dumpApplicationCacheDelegateCallbacks):
2785         (TestRunner):
2786         (WTR::TestRunner::shouldDisallowIncreaseForApplicationCacheQuota):
2787         (WTR::TestRunner::shouldDumpApplicationCacheDelegateCallbacks):
2788
2789 2012-09-12  Jochen Eisinger  <jochen@chromium.org>
2790
2791         [chromium] remove deprecated and unused sets import from chromium_android driver
2792         https://bugs.webkit.org/show_bug.cgi?id=96485
2793
2794         Reviewed by Dirk Pranke.
2795
2796         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
2797
2798 2012-09-12  Zan Dobersek  <zandobersek@gmail.com>
2799
2800         Flakiness dashboard doesn't recognize new Chromium Android test builder
2801         https://bugs.webkit.org/show_bug.cgi?id=96523
2802
2803         Reviewed by Ojan Vafai.
2804
2805         Properly return 'ANDROID' as the Chromium platform for Android builders.
2806
2807         * TestResultServer/static-dashboards/flakiness_dashboard.js:
2808         (chromiumPlatform):
2809         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
2810         (test):
2811
2812 2012-09-12  Simon Hausmann  <simon.hausmann@nokia.com>
2813
2814         [Qt] Build on Windows requires bison/flex in PATH
2815         https://bugs.webkit.org/show_bug.cgi?id=96358
2816
2817         Reviewed by Tor Arne Vestbø.
2818
2819         The build requires flex, bison, etc. and they need to be in the PATH when building. On Mac OS X
2820         and Linux that is rarely a problem given how easily available the tools are. On Windows however
2821         a separate installation of various GNU tools is required as the operating system doesn't come with
2822         them. To make the development more convenient, Qt 5 provides a copy of the most essential tools in
2823         the gnuwin32 directory of the qt5.git top-level repository.
2824
2825         This patch tries to detect the presence of those tools and prepends them to the PATH if found.
2826
2827         This is required in preparation for the elimination of qt5/qtwebkit.pri, which currently expands
2828         PATH before calling build-webkit. It it also required for the upcoming introduction of win_flex
2829         as dependency over flex, which can be done with less hassle when qt5's gnuwin32 directory has been
2830         updated with the new tool.
2831
2832         * Scripts/webkitdirs.pm:
2833         (checkRequiredSystemConfig):
2834         * qmake/mkspecs/features/default_post.prf:
2835
2836 2012-09-12  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2837
2838         [Qt] Update build-jsc after r128174
2839
2840         Reviewed by Ossy.
2841
2842         * Scripts/build-jsc:
2843
2844 2012-09-12  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2845
2846         [Qt] Teach addStrictSubdirOrderBetween to handle more than two targets
2847
2848         By hard-coding the names of the targets we defined we ended up just
2849         redefining the previous target when using addStrictSubdirOrderBetween
2850         more than once in a single project file.
2851
2852         We now embed the two base targets into the target names.
2853
2854         Reviewed by Simon Hausmann.
2855
2856         * qmake/mkspecs/features/functions.prf:
2857
2858 2012-09-12  Philippe Liard  <pliard@google.com>
2859
2860         Depend on {base,net} GYP targets rather than {base,net}_java.
2861         https://bugs.webkit.org/show_bug.cgi?id=95690
2862
2863         Reviewed by Adam Barth.
2864
2865         {base,net}_java became 'private' targets on the Chromium side that
2866         clients should not depend on (see
2867         https://chromiumcodereview.appspot.com/10913083/).
2868         This removes a bunch of trailing white spaces additionally in
2869         TestWebkitAPI.gyp (which includes non-trailing white space changes).
2870
2871         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
2872         * TestWebKitAPI/TestWebKitAPI.gyp/TestWebKitAPI.gyp:
2873
2874 2012-09-12  Simon Hausmann  <simon.hausmann@nokia.com>
2875
2876         [Qt] Fix the build with ENABLE_NETSCAPE_PLUGIN_API=0
2877         https://bugs.webkit.org/show_bug.cgi?id=96494
2878
2879         Reviewed by Tor Arne Vestbø.
2880
2881         WK2's ENABLE_PLUGIN_PROCESS uses the NPAPI functions unconditionally, so disable the
2882         plugin process feature if we don't have NPAPI.
2883
2884         * qmake/mkspecs/features/features.prf:
2885
2886 2012-09-12  Sheriff Bot  <webkit.review.bot@gmail.com>
2887
2888         Unreviewed, rolling out r128280.
2889         http://trac.webkit.org/changeset/128280
2890         https://bugs.webkit.org/show_bug.cgi?id=96498
2891
2892         it broke compilation on windows debug bot (Requested by loislo
2893         on #webkit).
2894
2895         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
2896         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2897         (WTR::InjectedBundlePage::InjectedBundlePage):
2898         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
2899         (InjectedBundlePage):
2900         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
2901         (WTR::TestRunner::TestRunner):
2902         * WebKitTestRunner/InjectedBundle/TestRunner.h:
2903         (TestRunner):
2904
2905 2012-09-12  Simon Hausmann  <simon.hausmann@nokia.com>
2906
2907         [Qt] Make it possible to build with "make release" and "make debug" on Windows
2908         https://bugs.webkit.org/show_bug.cgi?id=96488
2909
2910         Reviewed by Tor Arne Vestbø.
2911
2912         A "make debug" is passed through recursively and currently it aborts at Makefile.DerivedSources
2913         because there are no such targets. We want the generated sources to be independent from release
2914         or debug build configurations, so it is sufficient to provide fake debug and release targets that
2915         redirect to the same general-purpose target (first) of creating the derived sources.
2916
2917         * qmake/mkspecs/features/default_post.prf:
2918
2919 2012-09-12  Kenneth Rohde Christiansen  <kenneth@webkit.org>
2920
2921         [EFL] Make DumpRenderTree smarter at finding the fonts
2922         http://webkit.org/b/96281
2923
2924         Reviewed by Gyuyoung Kim.
2925
2926         Respect WEBKITOUTPUTDIR and expand the font dir from it.
2927         Use CString consistently.
2928
2929         * DumpRenderTree/efl/FontManagement.cpp:
2930         (buildPath):
2931         (getCoreFontFiles):
2932         (addFontDirectory):
2933         (addFontFiles):
2934         (getCustomBuildDir):
2935         (getPlatformFontsPath):
2936         (addFontsToEnvironment):
2937
2938 2012-09-12  Christophe Dumez  <christophe.dumez@intel.com>
2939
2940         [WK2][WKTR] TestRunner needs to implement dumpApplicationCacheDelegateCallbacks
2941         https://bugs.webkit.org/show_bug.cgi?id=96374
2942
2943         Reviewed by Kenneth Rohde Christiansen.
2944
2945         Implement support for dumpApplicationCacheDelegateCallbacks
2946         and disallowIncreaseForApplicationCacheQuota in
2947         WebKitTestRunner and properly dump the information
2948         expected by the tests.
2949
2950         If the application cache quota is reached for a given
2951         security origin, WebKitTestRunner will reset the quota
2952         to its default value, unless intructed not to via
2953         disallowIncreaseForApplicationCacheQuota().
2954
2955         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
2956         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2957         (WTR::InjectedBundlePage::InjectedBundlePage):
2958         (WTR::InjectedBundlePage::didReachApplicationCacheOriginQuota):
2959         (WTR):
2960         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
2961         (InjectedBundlePage):
2962         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
2963         (WTR::TestRunner::TestRunner):
2964         (WTR::TestRunner::disallowIncreaseForApplicationCacheQuota):
2965         (WTR):
2966         * WebKitTestRunner/InjectedBundle/TestRunner.h:
2967         (WTR::TestRunner::dumpApplicationCacheDelegateCallbacks):
2968         (TestRunner):
2969         (WTR::TestRunner::shouldDisallowIncreaseForApplicationCacheQuota):
2970         (WTR::TestRunner::shouldDumpApplicationCacheDelegateCallbacks):
2971
2972 2012-09-12  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
2973
2974         [WK2] [WTR] WebKitTestRunner needs TestRunner.workerThreadCount
2975         https://bugs.webkit.org/show_bug.cgi?id=96388
2976
2977         Reviewed by Kenneth Rohde Christiansen.
2978
2979         Exported TestRunner.workerThreadCount as readonly attribute.
2980
2981         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
2982         * WebKitTestRunner/InjectedBundle/TestRunner.cpp: Added workerThreadCount() method implementation.
2983         (WTR::TestRunner::workerThreadCount): Returns count of worker threads.
2984         (WTR):
2985         * WebKitTestRunner/InjectedBundle/TestRunner.h: Added workerThreadCount() method. 
2986         (TestRunner):
2987
2988 2012-09-12  Jochen Eisinger  <jochen@chromium.org>
2989
2990         [chromium] Consume a user gesture when creating a new view.
2991         https://bugs.webkit.org/show_bug.cgi?id=96373
2992
2993         Reviewed by Adam Barth.
2994
2995         * DumpRenderTree/chromium/WebViewHost.cpp:
2996         (WebViewHost::createView):
2997
2998 2012-09-12  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
2999
3000         [WK2][WTR] Some of TestRunner special options are not reset before testing
3001         https://bugs.webkit.org/show_bug.cgi?id=96384
3002
3003         Reviewed by Kenneth Rohde Christiansen.
3004
3005         Now values of the following special options are reset:
3006         void setAcceptsEditing(in boolean value);
3007         void setCloseRemainingWindowsWhenComplete(in boolean value);
3008         void setXSSAuditorEnabled(in boolean value);
3009         void setAllowFileAccessFromFileURLs(in boolean value);
3010         void setPluginsEnabled(in boolean value);
3011         void setPopupBlockingEnabled(in boolean value);
3012
3013         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
3014         (WTR::InjectedBundle::beginTesting):
3015
3016 2012-09-11  Ryuan Choi  <ryuan.choi@samsung.com>
3017
3018         [CMAKE] Supply feature defines to CodeGeneratorTestRunner.
3019         https://bugs.webkit.org/show_bug.cgi?id=96273
3020
3021         Reviewed by Gyuyoung Kim.
3022
3023         Passed FEATURE_DEFINES_WITH_SPACE_SEPARATOR to CodeGeneratorTestRunner
3024         in order to distinguish compile time options such as ENABLE(TOUCH_EVENTS).
3025
3026         * WebKitTestRunner/CMakeLists.txt:
3027
3028 2012-09-11  Dirk Pranke  <dpranke@chromium.org>
3029
3030         [chromium] rename webkit_resources
3031         https://bugs.webkit.org/show_bug.cgi?id=95874
3032
3033         Reviewed by James Robinson.
3034
3035         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
3036
3037 2012-09-11  Jochen Eisinger  <jochen@chromium.org>
3038
3039         [chromium] don't include content shell builders in the garden-o-matic view
3040         https://bugs.webkit.org/show_bug.cgi?id=96418
3041
3042         Reviewed by James Robinson.
3043
3044         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/config.js:
3045
3046 2012-09-11  Adam Klein  <adamk@chromium.org>
3047
3048         Garden-o-matic should ignore a wider variety of warnings in buildbot json
3049         https://bugs.webkit.org/show_bug.cgi?id=96411
3050
3051         Reviewed by Adam Barth.
3052
3053         Previously only the exact string "warning" was treated as a warning
3054         result. This patch treats any string with "warning" as a substring
3055         as a warning (e.g., "warnings", as seen on the cr-win buildbots).
3056
3057         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/builders.js:
3058         (.):
3059         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/builders_unittests.js:
3060
3061 2012-09-11  Marcelo Lira  <marcelo.lira@openbossa.org>
3062
3063         Restore original value of mock scrollbars enabled in InternalSettings
3064         https://bugs.webkit.org/show_bug.cgi?id=87680
3065
3066         Reviewed by Adam Barth.
3067
3068         Added backup support for mock scrollbars state in
3069         InternalSettings::Backup, and removed support for these
3070         from Qt's DumpRenderTree.
3071
3072         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
3073         (WebCore::WebPage::resetSettings):
3074         * DumpRenderTree/qt/TestRunnerQt.cpp:
3075         (TestRunner::setFrameFlatteningEnabled):
3076         * DumpRenderTree/qt/TestRunnerQt.h:
3077         (TestRunner):
3078
3079 2012-09-11  Dan Bernstein  <mitz@apple.com>
3080
3081         Fixed a bug where kill-old-processes failed to kill clang++.
3082
3083         Rubber-stamped by Alexey Proskuryakov.
3084
3085         * BuildSlaveSupport/kill-old-processes: Escape the + signs in "clang++"
3086         where it is passed as a regular expression to killall.
3087
3088 2012-09-11  Philippe Liard  <pliard@google.com>
3089
3090         Make sure that md5sum is not setup on non-Chromium Android platforms.
3091         https://bugs.webkit.org/show_bug.cgi?id=96393
3092
3093         Reviewed by Adam Barth.
3094
3095         Previously md5sum was setup in ChromiumAndroidDriver's constructor
3096         which is also invoked on non-Chromium Android platforms. The following patch
3097         moves this logic to _startup_test() which is guaranteed to be invoked
3098         only on Android.
3099
3100         This also updates the prompt unit test in chromium_android_unittest.py.
3101
3102         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
3103         (ChromiumAndroidDriver.__init__):
3104         (ChromiumAndroidDriver._setup_md5sum_and_push_data_if_needed):
3105         (ChromiumAndroidDriver._setup_test):
3106
3107 2012-09-11  Raphael Kubo da Costa  <rakuco@webkit.org>
3108
3109         [EFL] Rewrite the EFL-related Find modules
3110         https://bugs.webkit.org/show_bug.cgi?id=95237
3111
3112         Reviewed by Kenneth Rohde Christiansen.
3113
3114         * DumpRenderTree/efl/CMakeLists.txt: Stop setting the LINK_FLAGS
3115         property and include all the Enlightenment Foundation Libraries
3116         required by the target.
3117         * EWebLauncher/CMakeLists.txt: Ditto.
3118         * MiniBrowser/efl/CMakeLists.txt: Ditto.
3119         * TestWebKitAPI/PlatformEfl.cmake: Add missing include directories
3120         now that they are not added implicitly.
3121         * WebKitTestRunner/CMakeLists.txt: Stop setting the LINK_FLAGS
3122         property.
3123         * WebKitTestRunner/PlatformEfl.cmake: Stop setting the LINK_FLAGS
3124         property and include all the Enlightenment Foundation Libraries
3125         required by the target.
3126
3127 2012-09-11  Christophe Dumez  <christophe.dumez@intel.com>
3128
3129         [WK2][WKTR] TestRunner needs to implement setApplicationCacheOriginQuota
3130         https://bugs.webkit.org/show_bug.cgi?id=96379
3131
3132         Reviewed by Kenneth Rohde Christiansen.
3133
3134         Add implementation for setApplicationCacheOriginQuota
3135         to WebKitTestRunner.
3136
3137         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
3138         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
3139         (WTR::TestRunner::setApplicationCacheOriginQuota):
3140         (WTR):
3141         * WebKitTestRunner/InjectedBundle/TestRunner.h:
3142         (TestRunner):
3143
3144 2012-09-11  Luciano Wolf  <luciano.wolf@openbossa.org>
3145
3146         [Qt] [WK2] editing/inserting/typing-tab-designmode tests are failing
3147         https://bugs.webkit.org/show_bug.cgi?id=91988
3148
3149         Reviewed by Kenneth Rohde Christiansen.
3150
3151         Regular key events generated by Qt for the tab key fill both
3152         keyCode and keyText, so we should do the same for our fake events.
3153
3154         * WebKitTestRunner/qt/EventSenderProxyQt.cpp:
3155         (WTR::EventSenderProxy::keyDown):
3156
3157 2012-09-11  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
3158
3159         [WK2][WTR] WebKitTestRunner needs testRunner.setSpatialNavigationEnabled
3160         https://bugs.webkit.org/show_bug.cgi?id=96269
3161
3162         Reviewed by Kenneth Rohde Christiansen.
3163
3164         Exported missing testRunner.setSpatialNavigationEnabled() method.
3165
3166         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
3167         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
3168         (WTR::InjectedBundle::beginTesting):
3169         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
3170         (WTR::TestRunner::setSpatialNavigationEnabled):
3171         (WTR):
3172         * WebKitTestRunner/InjectedBundle/TestRunner.h:
3173         (TestRunner):
3174
3175 2012-09-11  Philippe Liard  <pliard@chromium.org>
3176
3177         Support LayoutTests on non-rooted devices for Chromium Android.
3178         https://bugs.webkit.org/show_bug.cgi?id=95346
3179
3180         Reviewed by Dirk Pranke.
3181
3182         We were requiring root permissions to sync resource files with the
3183         device by comparing host and device files' timestamp after we
3184         synchronized the device and host clocks.
3185         We now rely on md5sum provided by Chromium to handle data files
3186         synchronization.
3187
3188         * Scripts/webkitpy/layout_tests/port/chromium.py:
3189         (ChromiumPort._check_driver_build_up_to_date):
3190         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
3191         (ChromiumAndroidDriver.__init__):
3192         (ChromiumAndroidDriver._setup_test):
3193         (ChromiumAndroidDriver._push_executable):
3194         (ChromiumAndroidDriver._push_test_resources):
3195         (ChromiumAndroidDriver._read_prompt):
3196
3197 2012-09-11  Simon Hausmann  <simon.hausmann@nokia.com>
3198
3199         [Qt] Fix output paths on Windows
3200
3201         Reviewed by Tor Arne Vestbø.
3202
3203         BUILD_ROOT_RELATIVE_TO_OUT_PWD was calculated incorrectly Windows, resulting in an incorrect
3204         DESTDIR transformation at the end of default_post.prf. The calculation should be baased no qmake
3205         paths, which can use drive letters but always use forward slashes. So the use of QMAKE_DIR_SEP is
3206         incorrect here and instead for the calculation we should use forward slashes.
3207
3208         * qmake/mkspecs/features/default_pre.prf:
3209
3210 2012-09-11  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3211
3212         [Qt] Fix passing of defines from the build system
3213
3214         We explicitly have to pass -DENABLE_FOO=0 for every single feature
3215         that's not explicitly enabled in WEBKIT_CONFIG, since Platform.h
3216         will add it's own defaults if the features are not defined.
3217
3218         At some point we might want to run a configure-test to inspect what
3219         Platform.h will do, so we can sync up the build system's view of
3220         the feature-defines with the compile-time situation.
3221
3222         Reviewed by Simon Hausmann.
3223
3224         * qmake/mkspecs/features/functions.prf:
3225
3226 2012-09-11  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3227
3228         [Qt] Fix --qt option to build-webkit
3229
3230         Reviewed by Ossy.
3231
3232         * Scripts/webkitdirs.pm:
3233         (buildQMakeProjects):
3234
3235 2012-09-11  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3236
3237         [Qt] Add a configure step to the Qt build system
3238
3239         This allows building the Qt port using just 'qmake WebKit.pro'. Using
3240         the build-webkit script is still supported, and will add slightly more
3241         logic to the build, such as detecting the need for clean or incremental
3242         builds.
3243
3244         Internally, the build system now uses a new variable, WEBKIT_CONFIG, for
3245         all things related to configuring the build, such as use/have/enable flags,
3246         and these are translated to defines in default_post. Project files should
3247         no longer check contains(DEFINES,...) to detect features, but use the new
3248         enable?(), use?(), and have?() functions. The no_webkit1 and no_webkit2
3249         options have been translated into WEBKIT_CONFIG options as well, and can
3250         be checked using build?().
3251
3252         Reviewed by Simon Hausmann.
3253
3254         * DumpRenderTree/qt/DumpRenderTree.pro:
3255         * MiniBrowser/qt/MiniBrowser.pro:
3256         * QtTestBrowser/QtTestBrowser.pro:
3257         * Scripts/build-webkit:
3258         * Scripts/webkitdirs.pm:
3259         (qtFeatureDefaults):
3260         (buildQMakeProjects):
3261         * Tools.pro:
3262         * WebKitTestRunner/InjectedBundle/DerivedSources.pri:
3263         * WebKitTestRunner/InjectedBundle/Target.pri:
3264         * WebKitTestRunner/Target.pri:
3265         * qmake/.qmake.cache: Added.
3266         * qmake/configure.pri: Removed.
3267         * qmake/configure.pro: Removed.
3268         * qmake/dump-features: Added.
3269         * qmake/mkspecs/features/configure.prf: Added.
3270         * qmake/mkspecs/features/default_post.prf:
3271         * qmake/mkspecs/features/default_pre.prf:
3272         * qmake/mkspecs/features/features.prf:
3273         * qmake/mkspecs/features/features.pri:
3274         * qmake/mkspecs/features/functions.prf:
3275         * qmake/mkspecs/features/gprof.prf:
3276         * qmake/mkspecs/features/mac/default_pre.prf:
3277         * qmake/mkspecs/features/win32/default_pre.prf:
3278
3279 2012-09-11  Christophe Dumez  <christophe.dumez@intel.com>
3280
3281         [WK2][WKTR] TestRunner needs to implement clearApplicationCacheForOrigin
3282         https://bugs.webkit.org/show_bug.cgi?id=96372
3283
3284         Reviewed by Kenneth Rohde Christiansen.
3285
3286         Implement clearApplicationCacheForOrigin() in WebKitTestRunner.
3287
3288         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
3289         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
3290         (WTR::TestRunner::clearApplicationCacheForOrigin):
3291         (WTR):
3292         * WebKitTestRunner/InjectedBundle/TestRunner.h:
3293         (TestRunner):
3294
3295 2012-09-11  Peter Beverloo  <peter@chromium.org>
3296
3297         Don't ignore .pak files when making an archive for Android
3298         https://bugs.webkit.org/show_bug.cgi?id=96375
3299
3300         Reviewed by Jochen Eisinger.
3301
3302         The .pak files do need to be pushed to the device, which is currently
3303         preventing the layout test runner from correctly pushing all files.
3304
3305         * BuildSlaveSupport/built-product-archive:
3306         (archiveBuiltProduct):
3307
3308 2012-09-11  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
3309
3310         [WK2][WTR] WebKitTestRunner needs layoutTestController.setMinimumTimerInterval
3311         https://bugs.webkit.org/show_bug.cgi?id=96256
3312
3313         Reviewed by Kenneth Rohde Christiansen.
3314
3315         Exported TestRunner::setMinimumTimerInterval() method.
3316
3317         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
3318         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
3319         (WTR::InjectedBundle::beginTesting):
3320         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
3321         (WTR::TestRunner::setMinimumTimerInterval):
3322         (WTR):
3323         * WebKitTestRunner/InjectedBundle/TestRunner.h:
3324         (TestRunner):
3325
3326 2012-09-11  Tommy Widenflycht  <tommyw@google.com>
3327
3328         MediaStream API: add RTCPeerConnection::onnegotiationneeded
3329         https://bugs.webkit.org/show_bug.cgi?id=96097
3330
3331         Reviewed by Adam Barth.
3332
3333         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
3334         (MockWebRTCPeerConnectionHandler::addStream):
3335         (MockWebRTCPeerConnectionHandler::removeStream):
3336
3337 2012-09-11  Peter Beverloo  <peter@chromium.org>
3338
3339         Recognize the "--chromium-android" argument in run-webkit-tests
3340         https://bugs.webkit.org/show_bug.cgi?id=96369
3341
3342         Reviewed by Jochen Eisinger.
3343
3344         Chromium for the Android platform cannot determine the platform based on
3345         the host itself, so we'll use the --chromium-android argument as the build
3346         master passes to run-webkit-tests. Make sure the layout test runner
3347         can pick the right port when doing so.
3348
3349         * Scripts/run-webkit-tests:
3350         * Scripts/webkitpy/layout_tests/port/factory.py:
3351         (port_options):
3352
3353 2012-09-11  Simon Hausmann  <simon.hausmann@nokia.com>
3354
3355         [QT] Fix build with newer qmake from Qt 5
3356
3357         Rubber-stamped by Tor Arne Vestbø.
3358
3359         Support for "qmake -query QMAKE_MKSPECS" was removed from qmake. Replace
3360         it with the use of QT_HOST_DATA/mkspecs.
3361         Based on patch by Oswald Buddenhagen <oswald.buddenhagen@nokia.com>.
3362
3363         * Scripts/webkitdirs.pm:
3364         (builtDylibPathForName):
3365         (qtMakeCommand):
3366
3367 2012-09-11  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3368
3369         [Qt] Remove Qt 4 syncqt, no longer needed
3370
3371         Reviewed by Simon Hausmann.
3372
3373         * qmake/syncqt-4.8: Removed.
3374
3375 2012-09-11  Andras Becsi  <andras.becsi@nokia.com>
3376
3377         [Qt][WK2] Fix the build with recent Qt5
3378         https://bugs.webkit.org/show_bug.cgi?id=96283
3379
3380         Reviewed by Simon Hausmann.
3381
3382         We should not use deprecated Qt API because doing that results
3383         in build failure since all the deprecated API was disabled
3384         in qtbase (https://codereview.qt-project.org/#change,24890).
3385
3386         * DumpRenderTree/qt/main.cpp:
3387         (main):
3388         * MiniBrowser/qt/MiniBrowserApplication.cpp:
3389         (MiniBrowserApplication::notify):
3390         * MiniBrowser/qt/raw/View.cpp:
3391         (View::exposeEvent):
3392         (main):
3393
3394 2012-09-11  Christophe Dumez  <christophe.dumez@intel.com>
3395
3396         WTR needs an implementation of applicationCacheDiskUsageForOrigin
3397         https://bugs.webkit.org/show_bug.cgi?id=87188
3398
3399         Reviewed by Kenneth Rohde Christiansen.
3400
3401         Add implementation for applicationCacheDiskUsageForOrigin()
3402         in WebKitTestRunner.
3403
3404         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
3405         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
3406         (WTR::TestRunner::applicationCacheDiskUsageForOrigin):
3407         (WTR):
3408         * WebKitTestRunner/InjectedBundle/TestRunner.h:
3409         (TestRunner):
3410
3411 2012-09-10  Joanmarie Diggs  <jdiggs@igalia.com>
3412
3413         Properly expose <legend> elements to ATs
3414         https://bugs.webkit.org/show_bug.cgi?id=84137
3415
3416         Reviewed by Chris Fleizach.
3417
3418         Created a new WebCore Accessibility Role, LegendRole. Used it to map to
3419         the expected platform role, ATK_ROLE_LABEL. Also established the needed
3420         AtkRelation pair, label-for/labelled-by between the legend and fieldset.
3421
3422         * DumpRenderTree/chromium/TestRunner/AccessibilityUIElementChromium.cpp:
3423         (roleToString): added "Legend" string for WebAccessibilityRoleLegend
3424
3425 2012-09-10  Jer Noble  <jer.noble@apple.com>
3426
3427         Unreviewed; rolling out r128081.
3428         http://trac.webkit.org/changeset/128081
3429         https://bugs.webkit.org/show_bug.cgi?id=96134
3430
3431         New test http/tests/media/video-donottrack.html fails to run on Apple test bots.
3432
3433         * DumpRenderTree/TestRunner.cpp:
3434         (TestRunner::TestRunner):
3435         (TestRunner::staticFunctions):
3436         * DumpRenderTree/TestRunner.h:
3437         (TestRunner):
3438         * DumpRenderTree/mac/FrameLoadDelegate.mm:
3439         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
3440         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3441         (WTR::InjectedBundlePage::InjectedBundlePage):
3442         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
3443         (InjectedBundlePage):
3444         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
3445         (WTR::TestRunner::TestRunner):
3446         * WebKitTestRunner/InjectedBundle/TestRunner.h:
3447         (TestRunner):
3448
3449 2012-09-10  Christophe Dumez  <christophe.dumez@intel.com>
3450
3451         TestExpectationsChecker._determine_port_from_expectations_path() does not support cascaded TestExpectations
3452         https://bugs.webkit.org/show_bug.cgi?id=96205
3453
3454         Reviewed by Tony Chang.
3455
3456         _determine_port_from_expectations_path() was calling port.path_to_test_expectations_file()
3457         internally, which means that it would support only 1 TestExpectations file per port. This
3458         is an issue for ports such as EFL that support cascased TestExpectations (efl-wk2 -> efl
3459         and efl-wk1 -> efl).
3460
3461         This patch makes _determine_port_from_expectations_path() call port.expectations_files()
3462         instead so that all the ports TestExpectations are recognized. The ports are also
3463         constructed twice, with "webkit_test_runner" option set to True and False so that we
3464         retrieve the TestExpectations paths for both WebKit1 and WebKit2.
3465
3466         * Scripts/webkitpy/style/checkers/test_expectations.py:
3467         (TestExpectationsChecker._determine_port_from_expectations_path):
3468         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py: Add corresponding unit test.
3469         (TestExpectationsTestCase.test_determine_port_from_expectations_path):
3470
3471 2012-09-10  Ryuan Choi  <ryuan.choi@samsung.com>
3472
3473         [WTR] Generated source files should include config.h
3474         https://bugs.webkit.org/show_bug.cgi?id=96254
3475
3476         Reviewed by Benjamin Poulain.
3477
3478         Current coding style has encouraged to include config.h at the first position,
3479         But derived sources, generated by CodeGeneratorTestRunner.pm, don't follow
3480         this rule.
3481         Therefore, this patch added it.
3482
3483         * WebKitTestRunner/InjectedBundle/Bindings/CodeGeneratorTestRunner.pm:
3484         (_generateImplementationFile):
3485
3486 2012-09-10  Rick Byers  <rbyers@chromium.org>
3487
3488         Add handling of new GestureTapCancel in DRT
3489         
3490         https://bugs.webkit.org/show_bug.cgi?id=96183
3491
3492         Reviewed by Adam Barth.
3493
3494         * DumpRenderTree/chromium/TestWebPlugin.cpp:
3495         (TestWebPlugin::handleInputEvent):
3496
3497 2012-09-10  Brady Eidson  <beidson@apple.com>
3498
3499         Javascript in foreground tabs should not wait synchronously for plug-ins to load
3500         <rdar://problem/12067415> and https://bugs.webkit.org/show_bug.cgi?id=96167
3501
3502         Reviewed by Geoff Garen.
3503
3504         Enhance the "Slow NPP_New" plug-in to also be able to return properties to javascript.
3505
3506         * DumpRenderTree/TestNetscapePlugIn/Tests/SlowNPPNew.cpp:
3507         (PluginObject):
3508         (SlowNPPNew::PluginObject::PluginObject):
3509         (SlowNPPNew::PluginObject::~PluginObject):
3510         (SlowNPPNew::PluginObject::hasProperty):
3511         (SlowNPPNew::PluginObject::getProperty):
3512         (SlowNPPNew::NPP_GetValue):
3513         (SlowNPPNew):
3514
3515 2012-09-10  Tim Horton  <timothy_horton@apple.com>
3516
3517         WKTR often reports an unresponsive WebProcess on Mac bots
3518         https://bugs.webkit.org/show_bug.cgi?id=95906
3519
3520         Reviewed by Dan Bernstein.
3521
3522         Attempt to work around resource contention issues by increasing the NRWT
3523         timeout when running against WKTR, as well as WKTR's own WebProcess timeout.
3524
3525         * Scripts/webkitpy/layout_tests/port/base.py:
3526         (Port.default_timeout_ms): Default timeout when using WKTR -> 80 seconds.
3527         * WebKitTestRunner/TestController.cpp:
3528         (WTR): LongTimeout (used when waiting for WebProcess to finish) -> 60 seconds.
3529
3530 2012-09-07  Jer Noble  <jer.noble@apple.com>
3531
3532         <audio> and <video> should send Do Not Track when appropriate
3533         https://bugs.webkit.org/show_bug.cgi?id=96134
3534
3535         Reviewed by Eric Carlson.
3536
3537         Add a testRunner JS API for calling setShouldSendDoNotTrackHTTPHeader to DRT and 
3538         WTR. Add FrameLoaderClient client methods for retrieving that value.
3539
3540         * DumpRenderTree/TestRunner.cpp:
3541         (setShouldSendDoNotTrackHTTPHeaderCallback): Pass through to the TestRunner.
3542         (TestRunner::staticFunctions): Add the setShouldSendDoNotTrackHTTPHeader function.
3543         * DumpRenderTree/TestRunner.h:
3544         (TestRunner::setShouldSendDoNotTrackHTTPHeader): Simple setter.
3545         (TestRunner::shouldSendDoNotTrackHTTPHeader): Simple getter.
3546         * DumpRenderTree/mac/FrameLoadDelegate.mm:
3547         (-[FrameLoadDelegate webViewShouldSendDoNotTrackHTTPHeader:]): 
3548             Pass the question on to the TestRunner.
3549         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
3550         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3551         (WTR::InjectedBundlePage::InjectedBundlePage):
3552             Add the setShouldSendDoNotTrackHTTPHeader client function.
3553         (WTR::InjectedBundlePage::shouldSendDoNotTrackHTTPHeader):
3554             Pass the question on to the TestRunner.
3555         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
3556         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
3557         * WebKitTestRunner/InjectedBundle/TestRunner.h:
3558         (WTR::TestRunner::shouldSendDoNotTrackHTTPHeader): Simple getter.
3559         (WTR::TestRunner::setShouldSendDoNotTrackHTTPHeader): Simple Setter.
3560
3561 2012-09-10  Ojan Vafai  <ojan@chromium.org>
3562
3563         Make EFL bots work on the flakiness dashboard
3564         https://bugs.webkit.org/show_bug.cgi?id=96164
3565
3566         Reviewed by Tony Chang.
3567
3568         Also, update outdated unittests that had started failing.
3569
3570         * TestResultServer/static-dashboards/builders.js:
3571         (isWebkitTestRunner):
3572         * TestResultServer/static-dashboards/flakiness_dashboard.js:
3573         (nonChromiumPlatform):
3574         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
3575         (test):
3576
3577 2012-09-10  Tommy Widenflycht  <tommyw@google.com>
3578
3579         [chromium] MediaStream API: Remove the Descriptor postfix
3580         https://bugs.webkit.org/show_bug.cgi?id=96268
3581
3582         Reviewed by Adam Barth.
3583
3584         After consideration I realized that there is no need whatsoever in using the Descriptor prefix
3585         in the WekKit embedder API. Removed from WebRTCSessionDescription and WebRTCICECandidate to start with.
3586
3587         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
3588         (RTCSessionDescriptionRequestSuccededTask::RTCSessionDescriptionRequestSuccededTask):
3589         (MockWebRTCPeerConnectionHandler::createOffer):
3590         (MockWebRTCPeerConnectionHandler::createAnswer):
3591         (MockWebRTCPeerConnectionHandler::setLocalDescription):
3592         (MockWebRTCPeerConnectionHandler::setRemoteDescription):
3593         (MockWebRTCPeerConnectionHandler::localDescription):
3594         (MockWebRTCPeerConnectionHandler::remoteDescription):
3595         (MockWebRTCPeerConnectionHandler::addICECandidate):
3596         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.h:
3597         (MockWebRTCPeerConnectionHandler):
3598
3599 2012-09-10  Arnaud Renevier  <a.renevier@sisa.samsung.com>
3600
3601         [Gtk] failure to build jhbuild dependencies with cups 1.6
3602         https://bugs.webkit.org/show_bug.cgi?id=95991
3603
3604         Reviewed by Gustavo Noronha Silva.
3605
3606         Bump gtk+ and gnome-themes-standard dependencies to 3.4.2 to allow
3607         building with cups >= 1.6 [GNOME #670373].
3608
3609         * gtk/jhbuild.modules:
3610
3611 2012-09-10  Christophe Dumez  <christophe.dumez@intel.com>
3612
3613         [EFL][DRT] Encoding issue in printed console messages
3614         https://bugs.webkit.org/show_bug.cgi?id=96263
3615
3616         Reviewed by Kenneth Rohde Christiansen.
3617
3618         Correctly handle CONSOLE messages that contain non-ascii
3619         characters in EFL's DRT. We need to use String::fromUTF8()
3620         to construct a WTF::String from a raw string.
3621
3622         * DumpRenderTree/efl/DumpRenderTreeView.cpp:
3623         (onConsoleMessage):
3624
3625 2012-09-10  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
3626
3627         [WK2][WTR] User style sheet location should be reset before each test
3628         https://bugs.webkit.org/show_bug.cgi?id=96262
3629
3630         Reviewed by Kenneth Rohde Christiansen.
3631
3632         User style sheet location is reset before each test now.
3633
3634         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
3635         (WTR::InjectedBundle::beginTesting):
3636
3637 2012-09-10  Simon Hausmann  <simon.hausmann@nokia.com>
3638
3639         Fix build of WTR with Qt on Windows
3640         https://bugs.webkit.org/show_bug.cgi?id=96249
3641
3642         Reviewed by Kenneth Rohde Christiansen.
3643
3644         Windows loves to define min and max as macros, which beautifully
3645         clashes in many places where min and max are used as variable or method
3646         names. It can be convinced not to do this when min and max are already
3647         define as macros (expanding to themselves). We have this kind of
3648         workaround in config.h files all over WebKit and as it turns out when
3649         building WTR with Qt on Windows we do end up with such a clash. Therefore
3650         we need the same workaround.
3651
3652         * WebKitTestRunner/config.h:
3653
3654 2012-09-10  Christophe Dumez  <christophe.dumez@intel.com>
3655
3656         [WK2] New fast/forms/number/number-interactive-validation-required.html fails
3657         https://bugs.webkit.org/show_bug.cgi?id=95936
3658
3659         Reviewed by Kenneth Rohde Christiansen.
3660