REGRESSION (r241988): Switching tabs is slow
[WebKit-https.git] / Tools / ChangeLog-2012-10-02
1 2012-10-02  Daniel Bates  <dbates@webkit.org>
2
3         Define watch list "webkitperl", "SVNScripts", and "XSS" to watch for changes
4         to Tools/Scripts/{VCSUtils.pm, webkitdirs.pm, webkitperl}, Tools/Scripts/svn-*
5         scripts, and files whose name contains the word "XSS", respectively. Subscribe
6         myself to these lists.
7
8         * Scripts/webkitpy/common/config/watchlist:
9
10 2012-10-02  Roger Fong  <roger_fong@apple.com>
11
12         executive_unittest.py fails on Apple Windows bots but not when run locally.
13         https://bugs.webkit.org/show_bug.cgi?id=98196
14
15         Reviewed by Eric Seidel.
16
17         Making the test more flexible to check for cygwin and allow different outputs from process.wait().
18
19         * Scripts/webkitpy/common/system/executive_unittest.py:
20         (ExecutiveTest.serial_test_kill_process):
21
22 2012-10-02  Roger Fong  <roger_fong@apple.com>
23
24         Ignore failing line of SpacebarScrolling API test on Windows.
25         https://bugs.webkit.org/show_bug.cgi?id=84961
26
27         Reviewed by Anders Carlsson.
28
29         Test is failing because of https://bugs.webkit.org/show_bug.cgi?id=97946.
30         Ignoring the line that fails on Windows for now until the larger problem at hand is fixed.
31
32         * TestWebKitAPI/Tests/WebKit2/SpacebarScrolling.cpp:
33         (TestWebKitAPI::TEST):
34
35 2012-10-02  Tony Chang  <tony@chromium.org>
36
37         check-webkit-style can't handle qt-arm, qt-win, qt-mac, qt-5.0, qt-5.0-wk1 and qt-5.0-wk2 TestExpecatations
38         https://bugs.webkit.org/show_bug.cgi?id=98140
39
40         Reviewed by Eric Seidel.
41
42         Fix a bug where we were always using the host OS rather than using what was passed in as part of the
43         port name on Qt.
44
45         There's also a bug with being able to test for qt-5.0 expectations files when on a system with
46         Qt 4.8 installed. I'm not sure how to solve that, so for now, stop emitting a style warning about
47         not checking a file.
48
49         * Scripts/webkitpy/layout_tests/port/qt.py:
50         (QtPort._search_paths): Use passed in OS.
51         * Scripts/webkitpy/style/checkers/test_expectations.py:
52         (TestExpectationsChecker.check): Don't warn if we don't know about the TestExpectations file.
53         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
54         (TestExpectationsTestCase.test_determine_port_from_expectations_path): Added a test case for the OS fix.
55         Add a commented out test for the version case.
56
57 2012-10-02  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
58
59         Remove the last bits of MSYS support from webkitperl.
60         https://bugs.webkit.org/show_bug.cgi?id=98195
61
62         Reviewed by Eric Seidel.
63
64         No ports currently need MinGW/MSYS support, so remove the last
65         hackish bits related to it from our Perl code.
66
67         * Scripts/commit-log-editor:
68         (createCommitMessage):
69         * Scripts/old-run-webkit-tests:
70         * Scripts/webkitdirs.pm:
71         (isAnyWindows):
72         (isWindows):
73
74 2012-10-02  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
75
76         Remove MSYS-related changes to the http testing infrastructure.
77         https://bugs.webkit.org/show_bug.cgi?id=98183
78
79         Reviewed by Eric Seidel.
80
81         In the process of removing the cruft that has accumulated in our
82         Apache configuration code, we can basically revert r77161 which
83         added some MSYS (mingw)-specific code to old-run-webkit-tests,
84         since it is not being used by any port anymore.
85
86         * Scripts/run-webkit-httpd: Remove isMsys()-dependent hacks.
87         * Scripts/webkitperl/httpd.pm:
88         (getHTTPDPath): Remove isMsys() code path.
89         (getDefaultConfigForTestDirectory): Revert isMsys() hacks.
90         (getHTTPDConfigPathForTestDirectory): Do not point to the noew
91         removed apache2-msys-httpd.conf.
92         (convertMsysPath): Removed.
93
94 2012-10-02  Sadrul Habib Chowdhury  <sadrul@chromium.org>
95
96         [chromium] Make sure the touch-points in the touch-events have the correct state.
97         https://bugs.webkit.org/show_bug.cgi?id=98110
98
99         Reviewed by Adam Barth.
100
101         TestWebPlugin can now be configured to print out some details about
102         events. For now, details for only touch events are printed. This is
103         used to verify that touch-points in the touch-events are properly
104         set.
105
106         * DumpRenderTree/chromium/TestWebPlugin.cpp:
107         (pointState):
108         (printTouchList):
109         (printEventDetails):
110         (TestWebPlugin::TestWebPlugin):
111         (TestWebPlugin::handleInputEvent):
112         * DumpRenderTree/chromium/TestWebPlugin.h:
113         (TestWebPlugin):
114
115 2012-10-02  Zan Dobersek  <zandobersek@gmail.com>
116
117         [XvfbDriver] First tests in each worker occasionally crash
118         https://bugs.webkit.org/show_bug.cgi?id=96203
119
120         Reviewed by Dirk Pranke.
121
122         Often in debug builds, the first tests in each worker crash because they
123         can't yet open the Xvfb display assigned to them. This indicates Xvfb hasn't
124         yet been set up.
125
126         To avoid this, the execution should halt for a small amount of time before tests
127         are run so the Xvfb process has enough time to get ready. At the moment this is
128         done right after the Xvfb process is spawned and the execution halts for one
129         second. That value is overriden to zero when testing to not prolong the testing time.
130
131         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py:
132         (XvfbDriver.__init__):
133         (XvfbDriver._start):
134         * Scripts/webkitpy/layout_tests/port/xvfbdriver_unittest.py:
135         (XvfbDriverTest.make_driver):
136
137 2012-10-02  Christophe Dumez  <christophe.dumez@intel.com>
138
139         [EFL] Use XvfbDriver for layout tests
140         https://bugs.webkit.org/show_bug.cgi?id=98162
141
142         Reviewed by Kenneth Rohde Christiansen.
143
144         Use XvfbDriver for layout tests on EFL port.
145
146         * Scripts/webkitpy/layout_tests/port/efl.py:
147         (EflPort._driver_class):
148         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py:
149         (XvfbDriver._start): Set DUMPRENDERTREE_TEMP and LOCAL_RESOURCE_ROOT
150         environment variables in XvfbDriver as those are needed for
151         TestRunner::pathToLocalResource().
152
153 2012-10-02  Zan Dobersek  <zandobersek@gmail.com>
154
155         [Gtk] Several tests launch g_object_ref/unref assertion messages
156         https://bugs.webkit.org/show_bug.cgi?id=95062
157
158         Reviewed by Martin Robinson.
159
160         Only increase and decrease reference count of the accessibility
161         UI element if it actually exists.
162
163         * DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
164         (AccessibilityUIElement::AccessibilityUIElement):
165         (AccessibilityUIElement::~AccessibilityUIElement):
166
167 2012-10-02  Douglas Stockwell  <dstockwell@chromium.org>
168
169         [chromium] running layout tests on mountainlion should use chromium TestExpectations
170         https://bugs.webkit.org/show_bug.cgi?id=98119
171
172         Reviewed by Adam Barth.
173
174         Some basic support for mountainlion has already been committed but it failed to refer
175         to the chromium TestExpectations.
176
177         * Scripts/webkitpy/layout_tests/port/chromium.py:
178         (ChromiumPort):
179         * Scripts/webkitpy/layout_tests/port/chromium_port_testcase.py:
180         (ChromiumPortTestCase.test_all_test_configurations):
181
182 2012-10-02  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
183
184         webkitpy: Stop listening on port 8081 when using Apache.
185         https://bugs.webkit.org/show_bug.cgi?id=98155
186
187         Reviewed by Eric Seidel.
188
189         No test currently uses this port, and the lighttpd configuration
190         does not open it.
191
192         This helps the efforts of bringing some consistency to the Apache
193         configuration at the moment by opening the same ports in the
194         webkitpy side and in the httpd.conf side (the idea is to remove
195         the httpd.conf side later), and also makes the behavior consistent
196         across the two servers we currently support.
197
198         * Scripts/webkitpy/layout_tests/servers/apache_http_server.py:
199         (LayoutTestApacheHttpd.__init__):
200
201 2012-10-02  Csaba Osztrogonác  <ossy@webkit.org>
202
203         [WK2] REGRESSION(r128623): It made layout tests extremely slow
204         https://bugs.webkit.org/show_bug.cgi?id=96862
205
206         Reviewed by Simon Hausmann.
207
208         Disable calling the extremely slow WKBundleSetDatabaseQuota() between tests until proper fix.
209
210         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
211         (WTR::InjectedBundle::beginTesting):
212
213 2012-10-02  Philip Rogers  <pdr@google.com>
214
215         Fix PerfTest standard deviation calculation.
216         https://bugs.webkit.org/show_bug.cgi?id=98115
217
218         Reviewed by Ryosuke Niwa.
219
220         Previously our standard deviation calculation was incorrect. This patch
221         updates perftest.py's algorithm to calculate the sample standard deviation
222         (with Bessel's correction) using Knuth's online algorithm:
223         http://en.wikipedia.org/wiki/Algorithms_for_calculating_variance#Online_algorithm
224         An existing test has been modified to prove our new results.
225
226         This patch also updates runner.js to use Bessel's correction in
227         its sample standard deviation calculation, which is more accurate
228         for small sample sizes.
229
230         Additionally, runner.js has been modified to not calculate
231         the 'sum' statistic, which was not very useful.
232
233         * Scripts/webkitpy/performance_tests/perftest.py:
234
235             The unused variable valueSum has also been removed.
236
237         (PageLoadingPerfTest.run):
238         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
239
240             This test calculates the stdev of {2000, 3000, ..., 20000} which
241             was hand-calculated using a spreadsheet.
242
243         (TestPageLoadingPerfTest.test_run):
244
245 2012-10-01  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
246
247         webkitpy should accept a different httpd.conf specified by the user
248         https://bugs.webkit.org/show_bug.cgi?id=98071
249
250         Reviewed by Dirk Pranke.
251
252         The existing httpd.conf variants (or lighttpd.conf, for that
253         matter) we have do not always suit the user's system. This is
254         particularly true on Linux/Unix, where Apache can be installed in
255         a plethora of ways and the LoadModule calls can fail to specify
256         the proper module paths.
257
258         For now, we start accepting the WEBKIT_HTTP_SERVER_CONF_PATH
259         environment variable, which allows the user to specify the
260         absolute path to another http server configuration file that might
261         work on the user's system.
262
263         In the long term, we should try to generate our configuration file
264         and stop requiring all the different httpd.conf files we have as
265         well as this hack.
266
267         * Scripts/webkitpy/layout_tests/port/base.py:
268         (Port._path_to_apache_config_file):
269         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
270         (test_path_to_apache_config_file):
271
272 2012-10-01  Emil A Eklund  <eae@chromium.org>
273
274         Unreviewed, upgrade eae to reviewer. 
275         http://lists.webkit.org/mailman/private/webkit-committers/2012-October/000191.html
276
277         * Scripts/webkitpy/common/config/committers.py:
278
279 2012-10-01  Ming Xie  <mxie@rim.com>
280
281         [BlackBerry] Add a early return in buildCMakeProjectOrExit()
282         https://bugs.webkit.org/show_bug.cgi?id=98052
283
284         Reviewed by Rob Buis.
285
286         This option allows our build to only run the CMake build configuration
287         step when environment variable GENERATE_CMAKE_PROJECT_ONLY is specified.
288
289         * Scripts/webkitdirs.pm:
290         (buildCMakeProjectOrExit):
291
292 2012-10-01  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
293
294         [webkitpy] Detect the multiple names of the wdiff binary in the Port class.
295         https://bugs.webkit.org/show_bug.cgi?id=98039
296
297         Reviewed by Dirk Pranke.
298
299         Unify the duplicate checks for the wdiff binary present in
300         different ports into the base Port class.
301
302         * Scripts/webkitpy/layout_tests/port/base.py:
303         (Port):
304         (Port._path_to_wdiff):
305         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
306         (ChromiumLinuxPort._path_to_wdiff): Removed.
307         * Scripts/webkitpy/layout_tests/port/gtk.py:
308         (GtkPort._path_to_wdiff): Removed.
309
310 2012-10-01  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
311
312         [chromium] Remove custom implementation of _path_to_apache_config_file in chromium_linux.
313         https://bugs.webkit.org/show_bug.cgi?id=98042
314
315         Reviewed by Dirk Pranke.
316
317         The Chromium-based ports inherit from Port these days, so get rid
318         of the custom implementation of _path_to_apache_config_file() in
319         favor of the more portable one in the Port class.
320
321         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
322         (ChromiumLinuxPort._path_to_apache_config_file):
323
324 2012-10-01  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
325
326         [chromium] Remove duplicated _is_redhat_based() implementation from chromium_linux.py.
327         https://bugs.webkit.org/show_bug.cgi?id=98035
328
329         Reviewed by Dirk Pranke.
330
331         Since the Chromium port classes inherit from Port these days, it
332         can rely on the same implementation of _is_redhat_based() present
333         there instead of duplicating it.
334
335         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
336         (ChromiumLinuxPort._is_redhat_based): Removed.
337
338 2012-10-01  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
339
340         [GTK] Rely on the general implementation for detecting the Apache server path.
341         https://bugs.webkit.org/show_bug.cgi?id=98033
342
343         Reviewed by Martin Robinson.
344
345         The implemenentation of _path_to_apache() in base.py already
346         checks for the same paths gtk.py tries; the only difference is
347         that all of the are tried regardless of the current distro.
348
349         Doing so is more portable, and lets us remove some distro-checking
350         from the code.
351
352         * Scripts/webkitpy/layout_tests/port/gtk.py:
353         (GtkPort._path_to_apache): Remove.
354
355 2012-10-01  Csaba Osztrogonác  <ossy@webkit.org>
356
357         Unreviewed, update the URL of the Szeged SVN mirror.
358
359         * BuildSlaveSupport/build.webkit.org-config/config.json:
360
361 2012-10-01  Yuta Kitamura  <yutak@chromium.org>
362
363         Unreviewed. Create WebSocket watch list, and add myself to that list.
364
365         * Scripts/webkitpy/common/config/watchlist:
366
367 2012-09-30  Taiju Tsuiki  <tzik@chromium.org>
368
369         Unreviewed. Adding myself as a committer.
370
371         * Scripts/webkitpy/common/config/committers.py:
372
373 2012-09-30  Sadrul Habib Chowdhury  <sadrul@chromium.org>
374
375         [chromium] Add a test to verify that a plugin does not receive touch-events unless it explicitly requests for them
376         https://bugs.webkit.org/show_bug.cgi?id=97975
377
378         Reviewed by Adam Barth.
379
380         TestWebPlugin can now be configured to accept touch events or not. Use this
381         configurability to verify that a plugin does not receive touch events unless
382         it explicitly requests for them.
383
384         * DumpRenderTree/chromium/TestWebPlugin.cpp:
385         (TestWebPlugin::TestWebPlugin):
386         (TestWebPlugin::initialize):
387         (TestWebPlugin::parseBoolean):
388         * DumpRenderTree/chromium/TestWebPlugin.h:
389         (TestWebPlugin):
390
391 2012-09-28  Mariusz Grzegorczyk  <mariusz.g@samsung.com>
392
393         [WK2][GTK][EFL] Share WebKit2-GTK plugin process implementation with EFL port
394         https://bugs.webkit.org/show_bug.cgi?id=91844
395
396         Reviewed by Simon Hausmann.
397
398         Turn on Netscape Plugin API option by default.
399
400         * Scripts/webkitperl/FeatureList.pm:
401
402 2012-09-28  Dirk Pranke  <dpranke@chromium.org>
403
404         webkit-patch rebaseline-expectations is broken
405         https://bugs.webkit.org/show_bug.cgi?id=97619
406
407         Reviewed by Ojan Vafai.
408
409         The new expectations format is causing two problems for
410         webkit-patch rebaseline expectations
411
412         1) the transformation into and out of the old syntax (which is
413         still used internally) is somewhat lossy, e.g., we're not
414         preserving the case of Bug(x) identifiers. Also, we can't
415         tell if the input was [ WontFix ] or [ Skip WontFix Pass ]
416         2) the new syntax is more lenient, allowing for multiple ways to
417         specify the same result, e.g., "[ Skip ]" may or may not be
418         missing.
419
420         As a result, we end up rewriting a whole bunch of lines for
421         purely syntactic differences.
422
423         Both of these things are bugs that need to be fixed, but
424         we should also not just rewrite lines that shouldn't be
425         modified.
426
427         This patch fixes the last issue (no rewriting).
428
429         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
430         (TestExpectationLine._format_line):
431         (TestExpectations.remove_rebaselined_tests):
432         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
433         (RebaseliningTest.test_remove):
434
435 2012-09-28  Dirk Pranke  <dpranke@chromium.org>
436
437         virtual directories should just rename the beginning of the test path
438         https://bugs.webkit.org/show_bug.cgi?id=97925
439
440         Reviewed by Ojan Vafai.
441
442         We were replacing all occurrences of the base prefix of a
443         virtual test suite with the virtual name, instead of just the
444         first. So, where we had softwarecompositing running the
445         compositing tests, any compositing test that actually had
446         "compositing" in the test base name would get that expanded
447         incorrectly.
448
449         * Scripts/webkitpy/layout_tests/port/base.py:
450         (Port.populated_virtual_test_suites):
451         (Port.lookup_virtual_test_base):
452         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
453         (PortTest.test_tests):
454         * Scripts/webkitpy/layout_tests/port/test.py:
455
456 2012-09-28  Ojan Vafai  <ojan@chromium.org>
457
458         Garden-o-matic sometimes things buildbot warnings are failures
459         https://bugs.webkit.org/show_bug.cgi?id=97926
460
461         Reviewed by Tony Chang.
462
463         It turns out the JSON clearly tells us if it's a failure.
464         We want to skip all warnings even if they don't have "warning" in the text.
465         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/builders.js:
466         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/builders_unittests.js:
467
468 2012-09-28  Harald Tveit Alvestrand  <harald@alvestrand.no>
469
470         Implement the GetStats interface on PeerConnection
471         https://bugs.webkit.org/show_bug.cgi?id=95193
472
473         Reviewed by Adam Barth.
474
475         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
476         (RTCStatsRequestSucceededTask):
477         (RTCStatsRequestSucceededTask::RTCStatsRequestSucceededTask):
478         (MockWebRTCPeerConnectionHandler::getStats):
479         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.h:
480         (MockWebRTCPeerConnectionHandler):
481
482 2012-09-28  Kristóf Kosztyó  <kkristof@inf.u-szeged.hu>
483
484         [NRWT] XvfbDriver should choose the next free display
485         https://bugs.webkit.org/show_bug.cgi?id=88414
486
487         Reviewed by Dirk Pranke.
488
489         This change how the XvfbDriver choose the next display. Before
490         this the choosing are based on the worker number but it caused
491         errors when several nrwt run on the same time. This add process
492         based free diplay search.
493
494         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py:
495         (XvfbDriver.__init__):
496         (XvfbDriver._next_free_display):
497         (XvfbDriver._start):
498         (XvfbDriver.stop):
499         * Scripts/webkitpy/layout_tests/port/xvfbdriver_unittest.py:
500         (XvfbDriverTest.make_driver):
501         (XvfbDriverTest.test_start_no_pixel_tests):
502         (XvfbDriverTest.test_start_pixel_tests):
503         (XvfbDriverTest.test_start_arbitrary_worker_number):
504         (XvfbDriverTest.test_next_free_display):
505         (XvfbDriverTest):
506         (XvfbDriverTest.test_start_next_worker):
507
508 2012-09-28  Huang Dongsung  <luxtella@company100.net>
509
510         [GTK] Enable CSS Shaders layout LayoutTests on GTK+
511         https://bugs.webkit.org/show_bug.cgi?id=97821
512
513         Reviewed by Martin Robinson.
514
515         This feature is disabled via Settings by default, but for testing,
516         we enable it using layoutTestController.overridePreferences. Add the
517         necessary plumbing for DRT.
518
519         WTR already works because support was added for Apple Mac earlier.
520
521         * DumpRenderTree/gtk/DumpRenderTree.cpp:
522         (resetDefaultsToConsistentValues): Feature is off by default.
523         * DumpRenderTree/gtk/TestRunnerGtk.cpp:
524         (TestRunner::overridePreference): Add handling of WebKitCSSCustomFilterEnabled.
525
526 2012-09-28  Csaba Osztrogonác  <ossy@webkit.org>
527
528         [Qt] Make NRWT use wk2/TestExpectations file
529         https://bugs.webkit.org/show_bug.cgi?id=97894
530
531         Reviewed by Simon Hausmann.
532
533         * Scripts/webkitpy/layout_tests/port/qt.py: Remove unnecessary _skipped_file_search_paths function.
534         (QtPort.expectations_files): Add wk2 if --webkit-test-runner option added.
535         * Scripts/webkitpy/layout_tests/port/qt_unittest.py: Remove unnecessary test__skipped_file_search_paths function.
536         (QtPortTest._assert_search_path): Removed.
537         (QtPortTest.test_expectations_files): Add wk2 if --webkit-test-runner option added, and code cleanup.
538
539 2012-09-28  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
540
541         [EFL][DRT] EFL's layoutTestController.keepWebHistory should enable visited links tracking
542         https://bugs.webkit.org/show_bug.cgi?id=97901
543
544         Reviewed by Kenneth Rohde Christiansen.
545
546         EFL's DRT testRunner.keepWebHistory should enable visited links tracking in order
547         to be consistent with WTR implementation and also several tests rely on that (fast/history).
548
549         * DumpRenderTree/efl/TestRunnerEfl.cpp:
550         (TestRunner::keepWebHistory):
551
552 2012-09-28  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
553
554         [EFL][DRT] EFL DRT should disable visited links tracking by default
555         https://bugs.webkit.org/show_bug.cgi?id=97881
556
557         Reviewed by Kenneth Rohde Christiansen.
558
559         Visited links tracking is disabled by default for each test.
560
561         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
562         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
563
564 2012-09-28  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
565
566         Unreviewed, rolling out r129863.
567         http://trac.webkit.org/changeset/129863
568         https://bugs.webkit.org/show_bug.cgi?id=97173
569
570         Broken debug WK2 layout test
571
572         * MiniBrowser/efl/main.c:
573         (quit):
574         (main):
575         * WebKitTestRunner/efl/PlatformWebViewEfl.cpp:
576         (WTR::initEcoreEvas):
577         (WTR::PlatformWebView::~PlatformWebView):
578         * WebKitTestRunner/efl/main.cpp:
579         (main):
580
581 2012-09-28  Kai Koehne  <kai.koehne@digia.com>
582
583         [Qt] Use qInstallMessageHandler()
584         https://bugs.webkit.org/show_bug.cgi?id=96648
585
586         Reviewed by Jocelyn Turcotte.
587
588         qInstallMsgHandler() got deprecated in Qt 5.
589
590         * DumpRenderTree/qt/DumpRenderTreeMain.cpp:
591         (messageHandler):
592         (main):
593         * WebKitTestRunner/qt/main.cpp:
594         (messageHandler):
595         (main):
596
597 2012-09-28  Eunmi Lee  <eunmi15.lee@samsung.com>
598
599         [EFL][WK2] Refactoring initialization and shutdown codes of EFL libraries.
600         https://bugs.webkit.org/show_bug.cgi?id=97173
601
602         Reviewed by Gyuyoung Kim.
603
604         The initialization and shutdown are done inside the webkit, so we don't
605         have to call them out of the webkit.
606
607         * MiniBrowser/efl/main.c:
608         (quit):
609         (main):
610         * WebKitTestRunner/efl/PlatformWebViewEfl.cpp:
611         (WTR::initEcoreEvas):
612         (WTR::PlatformWebView::~PlatformWebView):
613         * WebKitTestRunner/efl/main.cpp:
614         (main):
615
616 2012-09-27  Alexandre Elias  <aelias@chromium.org>
617
618         [chromium] DumpRenderTree support for software compositing
619         https://bugs.webkit.org/show_bug.cgi?id=96853
620
621         Reviewed by James Robinson.
622
623         This adds software compositor support to DumpRenderTree.  The only
624         thing needed is a minimal OutputSurfaceSoftware that holds a
625         viewport-sized bitmap. Then, I added virtual test suite
626         platform/chromium/virtual/softwarecompositing -- a mirror of the
627         compositing/ directory with --enable-software-compositing, and
628         expectations for not yet supported tests.
629
630         * DumpRenderTree/DumpRenderTree.gypi:
631         * DumpRenderTree/chromium/DumpRenderTree.cpp:
632         (main):
633         * DumpRenderTree/chromium/TestShell.cpp:
634         (TestShell::TestShell):
635         * DumpRenderTree/chromium/TestShell.h:
636         (TestShell::softwareCompositingEnabled):
637         (TestShell::setSoftwareCompositingEnabled):
638         (TestShell):
639         * DumpRenderTree/chromium/WebViewHost.cpp:
640         (WebViewHost::createOutputSurface):
641         * DumpRenderTree/chromium/WebViewHostOutputSurface.cpp:
642         (WebKit::WebViewHostOutputSurface::create3d):
643         (WebKit):
644         (WebKit::WebViewHostOutputSurface::createSoftware):
645         (WebKit::WebViewHostOutputSurface::WebViewHostOutputSurface):
646         (WebKit::WebViewHostOutputSurface::bindToClient):
647         (WebKit::WebViewHostOutputSurface::surfaceSoftware):
648         * DumpRenderTree/chromium/WebViewHostOutputSurface.h:
649         (WebKit):
650         (WebViewHostOutputSurface):
651         * DumpRenderTree/chromium/WebViewHostSoftwareOutputDevice.cpp: Copied from Tools/DumpRenderTree/chromium/WebViewHostOutputSurface.cpp.
652         (WebKit):
653         (WebKit::WebViewHostSoftwareOutputDevice::lockForWrite):
654         (WebKit::WebViewHostSoftwareOutputDevice::unlockForWrite):
655         (WebKit::WebViewHostSoftwareOutputDevice::lockForRead):
656         (WebKit::WebViewHostSoftwareOutputDevice::unlockForRead):
657         (WebKit::WebViewHostSoftwareOutputDevice::viewportChanged):
658         * DumpRenderTree/chromium/WebViewHostSoftwareOutputDevice.h: Copied from Tools/DumpRenderTree/chromium/WebViewHostOutputSurface.cpp.
659         (WebKit):
660         (WebViewHostSoftwareOutputDevice):
661         * Scripts/webkitpy/layout_tests/port/chromium.py:
662         (ChromiumPort.virtual_test_suites):
663
664 2012-09-27  Nasko Oskov  <nasko@chromium.org>
665
666         Move DumpRenderTree to use newer version of the name method of
667         WebFrame and the willCheckAndDispatchMessageEvent method of WebFrameClient
668         https://bugs.webkit.org/show_bug.cgi?id=95254
669
670         Reviewed by Adam Barth.
671
672         The WebFrame name() method is removed and replaced by better named
673         uniqueName() method, so DRT needs to be updated to use this new version.
674         Also the willCheckAndDispatchMessageEvent in WebFrameClient gained
675         one more parameter, which had to be added in DRT.
676
677         * DumpRenderTree/chromium/TestShell.cpp:
678         (dumpFramesAsText):
679         (dumpFramesAsPrintedText):
680         (dumpFrameScrollPosition):
681         * DumpRenderTree/chromium/WebViewHost.cpp:
682         (WebViewHost::unableToImplementPolicyWithError):
683         (WebViewHost::willCheckAndDispatchMessageEvent):
684         (WebViewHost::printFrameDescription):
685         * DumpRenderTree/chromium/WebViewHost.h:
686         (WebViewHost):
687
688 2012-09-27  Julie Parent  <jparent@chromium.org>
689
690         Add webkit_unit_tests to the flakiness dashboard
691         https://bugs.webkit.org/show_bug.cgi?id=97814
692
693         Reviewed by Dirk Pranke.
694
695         webkit_unit_tests run on the same bots as test_shell_tests.
696
697         * TestResultServer/static-dashboards/builders.js:
698         (loadBuildersList):
699         * TestResultServer/static-dashboards/dashboard_base.js:
700         (currentBuilderGroupCategory):
701
702 2012-09-27  Florin Malita  <fmalita@chromium.org>
703
704         [Perftest] PerfTestRunner exception when using nested git checkouts
705         https://bugs.webkit.org/show_bug.cgi?id=97796
706
707         Reviewed by Ryosuke Niwa.
708
709         In order to support separate Chromium vs. WebKit git trees, PerfTestsRunner needs to
710         instantiate per-repository SCM objects when collecting revision information.
711
712         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
713         (PerfTestsRunner._generate_results_dict):
714
715 2012-09-27  Sudarsana Nagineni  <sudarsana.nagineni@intel.com>
716
717         [Mac][DRT] Add support to allow non-array modifier arguments in keyDown()
718         https://bugs.webkit.org/show_bug.cgi?id=97805
719
720         Reviewed by Alexey Proskuryakov.
721
722         Add support for handling a string as modifier argument to keyDown().
723         This support has already been added to other ports, so implementing
724         this on Mac too, to be consistent with the other ports.
725
726         * DumpRenderTree/mac/EventSendingController.mm:
727         (modifierFlags):
728         (buildModifierFlags):
729
730 2012-09-27  Dirk Pranke  <dpranke@chromium.org>
731
732         efl wk2 bots weren't looking at platform/wk2/TestExpectations
733         https://bugs.webkit.org/show_bug.cgi?id=97815
734
735         Unreviewed, build fix.
736
737         Should clean up this logic later so that we're just using the
738         implementation in base ...
739
740         * Scripts/webkitpy/layout_tests/port/efl.py:
741         (EflPort._search_paths):
742         (EflPort.expectations_files):
743
744 2012-09-27  Dirk Pranke  <dpranke@chromium.org>
745
746         webkitpy: make warning about missing BUG identifiers per-port configurable
747         https://bugs.webkit.org/show_bug.cgi?id=97751
748
749         Reviewed by Ojan Vafai.
750
751         Prior to this change linting TestExpectations files (checking their style)
752         would raise warnings if lines were missing a BUG modifier, even
753         though such a modifier is actually optional in the syntax. As we
754         migrate the entries over from the Skipped files, these warnings
755         will generate a lot of noise, and it's not yet clear that we
756         actually want to require the modifiers on all ports.
757
758         So, this change makes this per-port configurable so that they
759         will generate warnings in the Chromium TestExpectations and not
760         anywhere else.
761
762         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
763         (TestExpectationParser._parse_modifiers):
764         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
765         (MiscTests.test_parse_warning):
766         (SemanticTests.test_missing_bugid):
767         * Scripts/webkitpy/layout_tests/port/base.py:
768         (Port.warn_if_bug_missing_in_test_expectations):
769         * Scripts/webkitpy/layout_tests/port/chromium.py:
770         (ChromiumPort.warn_if_bug_missing_in_test_expectations):
771
772 2012-09-27  Xianzhu Wang  <wangxianzhu@chromium.org>
773
774         [Chromium-Android] Push more fonts for layout tests to match chromium-linux
775         https://bugs.webkit.org/show_bug.cgi?id=97746
776
777         Reviewed by Dirk Pranke.
778
779         * DumpRenderTree/chromium/android_fallback_fonts.xml: Added the new fonts into fallback list. Removed the Android system fallback font DroidSansFallback.ttf.
780         * Scripts/webkitpy/layout_tests/port/builders.py: Add chromium-android to pass style check.
781         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
782         (ChromiumAndroidPort.check_build): Added check for md5sum and forwarder utilities to give better error message when they are missing.
783         (ChromiumAndroidPort.check_sys_deps): Allow host font files to come from one of multiple directories.
784         (ChromiumAndroidPort._path_to_md5sum):
785         (ChromiumAndroidDriver._setup_md5sum_and_push_data_if_needed):
786         (ChromiumAndroidDriver._push_fonts):
787         (ChromiumAndroidDriver._push_test_resources):
788
789 2012-09-27  Tommy Widenflycht  <tommyw@google.com>
790
791         MediaStream API: Enhance MediaConstraints to make it easier to get the constraint data
792         https://bugs.webkit.org/show_bug.cgi?id=97559
793
794         Reviewed by Adam Barth.
795
796         Refactoring for the new WebMediaConstraints api.
797
798         * DumpRenderTree/chromium/MockConstraints.cpp:
799         (MockConstraints::verify):
800
801 2012-09-27  Tommy Widenflycht  <tommyw@google.com>
802
803         MediaStream API: Rename MockConstraints::verify to verifyConstraints
804         https://bugs.webkit.org/show_bug.cgi?id=97779
805
806         Reviewed by Csaba Osztrogonác.
807
808         The Mac OS X sdk has a macro called verify :/
809
810         * DumpRenderTree/chromium/MockConstraints.cpp:
811         (MockConstraints::verifyConstraints):
812         * DumpRenderTree/chromium/MockConstraints.h:
813         (MockConstraints):
814         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
815         (MockWebRTCPeerConnectionHandler::initialize):
816         * DumpRenderTree/chromium/WebUserMediaClientMock.cpp:
817         (WebUserMediaClientMock::requestUserMedia):
818
819 2012-09-27  Tommy Widenflycht  <tommyw@google.com>
820
821         MediaStream API: Update getUserMedia to match the latest specification
822         https://bugs.webkit.org/show_bug.cgi?id=97540
823
824         Reviewed by Adam Barth.
825
826         http://dev.w3.org/2011/webrtc/editor/getusermedia.html#navigatorusermedia
827         Navigator.getUserMedia is changed so that the audio and video members can either
828         be a bool or a constraints object.
829
830         * public/WebUserMediaRequest.h:
831         (WebKit):
832         (WebUserMediaRequest):
833         * src/WebUserMediaRequest.cpp:
834         (WebKit::WebUserMediaRequest::audioConstraints):
835         (WebKit):
836         (WebKit::WebUserMediaRequest::videoConstraints):
837
838 2012-09-26  Laszlo Gombos  <l.gombos@samsung.com>
839
840         Unreviewed. Update my e-mail address.
841
842         * Scripts/webkitpy/common/config/committers.py:
843
844 2012-09-26  Dirk Pranke  <dpranke@chromium.org>
845
846         modify old-run-webkit-tests to support TestExpectations files a little
847         https://bugs.webkit.org/show_bug.cgi?id=97276
848
849         Reviewed by Daniel Bates.
850
851         This patch adds minimal support to old-run-webkit-tests so that
852         it can read a TestExpectations file and skip the tests listed
853         there. It will skip *all* tests, regardless of the expectation
854         (even if the test is listed as [ Pass ]), and it will only look
855         at the TestExpectations file if there isn't a Skipped file next
856         to it.
857         
858         This will allow us to switch to using just using
859         TestExpectations files without making old-run-webkit-tests
860         totally useless.
861
862         * Scripts/old-run-webkit-tests:
863         (readSkippedFiles):
864         (processSkippedFileEntry): Added.
865         (startsWith): Added.
866
867 2012-09-26  Dirk Pranke  <dpranke@chromium.org>
868
869         The style bot spams about skia_test_expectations.txt
870         https://bugs.webkit.org/show_bug.cgi?id=97699
871
872         Reviewed by Adam Barth.
873
874         Remove the warning about a missing skia_test_expectations.txt
875         file for now. There are some situations where it would be
876         helpful to get this warning as it might indicate that the user
877         is going to get results different than what they're expecting,
878         but at the moment it looks difficult if not impossible to get
879         those warnings and not also get warnings that we don't care
880         about, so the noise is outweighing the value of the signal.
881
882         This at least will stop the false negatives we're getting on the
883         style bots. If it turns out people do need the warning we can
884         add it back in in some other more conditional way.
885
886         * Scripts/webkitpy/layout_tests/port/chromium.py:
887         (ChromiumPort.expectations_files):
888         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
889         (TestRebaseline.test_rebaseline_updates_expectations_file_noop):
890         (test_rebaseline_updates_expectations_file):
891         (test_rebaseline_does_not_include_overrides):
892         (test_rebaseline_expectations):
893         (_assert_command):
894
895 2012-09-26  David Michael Barr  <davidbarr@chromium.org>
896
897         Unreviewed. Upgrade myself from contributor to committer.
898
899         * Scripts/webkitpy/common/config/committers.py:
900
901 2012-09-26  Ryosuke Niwa  <rniwa@webkit.org>
902
903         Remove deprecated code from perftestrunner.py
904         https://bugs.webkit.org/show_bug.cgi?id=97724
905
906         Reviewed by Adam Barth.
907
908         Remove the code deprecated in r129580 now that we have landed r129597,
909         which updated master.cfg to use new options, and restarted the buildbot master.
910
911         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
912         (PerfTestsRunner._parse_args):
913         (PerfTestsRunner._generate_and_show_results):
914         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
915         (test_run_with_slave_config_json):
916         (test_parse_args):
917
918 2012-09-26  Beth Dakin  <bdakin@apple.com>
919
920         https://bugs.webkit.org/show_bug.cgi?id=97629
921         http/tests/security/cross-frame-access-put.html failing on Mac WK2 bots after 
922         being unskipped
923
924         Reviewed by Tim Horton.
925
926         The test was failing because we were failing to use the fake window origin. We 
927         should really always use the fake origin. This patch eliminates 
928         _shouldUseFakeOrigin in favor of always using it.
929         * WebKitTestRunner/mac/PlatformWebViewMac.mm:
930         (-[WebKitTestRunnerWindow setFrameOrigin:]):
931         (-[WebKitTestRunnerWindow setFrame:display:animate:]):
932         (-[WebKitTestRunnerWindow setFrame:display:]):
933         (-[WebKitTestRunnerWindow frameRespectingFakeOrigin]):
934
935 2012-09-26  Tony Chang  <tony@chromium.org>
936
937         C++ style checker should warn when the indentation is wrong
938         https://bugs.webkit.org/show_bug.cgi?id=97602
939
940         Reviewed by Ojan Vafai.
941
942         Rewrite the indentation checker to ensure that indentation is always a factor of 4
943         and that we only indent 4 spaces from the previous line.
944
945         I deleted some of the old indentation checking code since it was filtered out
946         and didn't match the webkit style.
947
948         * Scripts/webkitpy/style/checker.py:
949         * Scripts/webkitpy/style/checkers/cpp.py:
950         (check_indentation_amount): New function that checks indentation. Intentionally runs after
951         other indentation checks.
952         (check_style): Delete the old indentation code since we were already filtering out the whitespace/labels warnings.
953         (CppChecker): Remove the whitespace/labels filter.
954         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
955         (CppStyleTest.test_multi_line_comments): Force 4 space indent.
956         (CppStyleTest.test_explicit_single_argument_constructors):
957         (CppStyleTest.test_explicit_single_argument_constructors.Foo):
958         (CppStyleTest.test_explicit_single_argument_constructors.Qualifier):
959         (CppStyleTest.test_slash_star_comment_on_single_line):
960         (CppStyleTest.test_braces):
961         (CppStyleTest.test_spacing_before_last_semicolon):
962         (CppStyleTest.test_static_or_global_stlstrings):
963         (CppStyleTest.test_indent): Added macro test case.
964         (CppStyleTest.test_build_class):
965         (NoNonVirtualDestructorsTest.test_no_error):
966         (NoNonVirtualDestructorsTest.test_no_error.Foo):
967         (NoNonVirtualDestructorsTest.test_no_error.Qualified):
968         (NoNonVirtualDestructorsTest.test_no_destructor_when_virtual_needed):
969         (NoNonVirtualDestructorsTest.test_no_destructor_when_virtual_needed.Foo):
970         (NoNonVirtualDestructorsTest.test_destructor_non_virtual_when_virtual_needed):
971         (NoNonVirtualDestructorsTest.test_destructor_non_virtual_when_virtual_needed.Foo):
972         (NoNonVirtualDestructorsTest.test_no_warn_when_derived):
973         (NoNonVirtualDestructorsTest.test_no_warn_when_derived.Foo):
974         (NoNonVirtualDestructorsTest.test_internal_braces):
975         (NoNonVirtualDestructorsTest.test_internal_braces.Foo):
976         (NoNonVirtualDestructorsTest.test_inner_class_needs_virtual_destructor):
977         (NoNonVirtualDestructorsTest.test_inner_class_needs_virtual_destructor.Foo):
978         (NoNonVirtualDestructorsTest.test_inner_class_needs_virtual_destructor.Foo.Goo):
979         (NoNonVirtualDestructorsTest.test_outer_class_needs_virtual_destructor):
980         (NoNonVirtualDestructorsTest.test_outer_class_needs_virtual_destructor.Foo):
981         (NoNonVirtualDestructorsTest.test_outer_class_needs_virtual_destructor.Foo.Goo):
982         (NoNonVirtualDestructorsTest.test_qualified_class_needs_virtual_destructor):
983         (NoNonVirtualDestructorsTest.test_qualified_class_needs_virtual_destructor.Qualified):
984         (NoNonVirtualDestructorsTest.test_multi_line_declaration_no_error):
985         (NoNonVirtualDestructorsTest.test_multi_line_declaration_no_error.Foo):
986         (NoNonVirtualDestructorsTest.test_multi_line_declaration_with_error):
987         (NoNonVirtualDestructorsTest.test_multi_line_declaration_with_error.Foo):
988         (WebKitStyleTest.test_indentation):
989         (WebKitStyleTest.test_braces): Add test for call params spanning lines.
990
991 2012-09-26  Adam Barth  <abarth@webkit.org>
992
993         Rollout messages should be clickable in bugzilla
994         https://bugs.webkit.org/show_bug.cgi?id=97711
995
996         Reviewed by Simon Fraser.
997
998         If we say Bug NNNN instead of just NNNN, bugzilla will autolink to the
999         actual bug.
1000
1001         * Scripts/webkitpy/tool/commands/download_unittest.py:
1002         * Scripts/webkitpy/tool/steps/createbug.py:
1003         (CreateBug.run):
1004
1005 2012-09-26  Marcelo Lira  <marcelo.lira@openbossa.org>
1006
1007         run-perf-tests must expand environment variables in user provided paths
1008         https://bugs.webkit.org/show_bug.cgi?id=97686
1009
1010         Reviewed by Ryosuke Niwa.
1011
1012         The run-perf-tests command line options that receive paths from the
1013         user now expand any possible environment variables, because Python's
1014         file handling methods do not handle those.
1015
1016         Examples:
1017         run-perf-tests --platform=qt --release --output-json-path=~/perf-results
1018         run-perf-tests --platform=qt --release --output-json-path=$HOME/perf-results
1019
1020         Also removed unused imports.
1021
1022         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
1023         (PerfTestsRunner._parse_args._expand_path): expand environment variables in a path passed via command line
1024         (PerfTestsRunner._parse_args):
1025
1026 2012-09-26  Sheriff Bot  <webkit.review.bot@gmail.com>
1027
1028         Unreviewed, rolling out r129654.
1029         http://trac.webkit.org/changeset/129654
1030         https://bugs.webkit.org/show_bug.cgi?id=97702
1031
1032         breaks chromium windows build (Requested by schenney on
1033         #webkit).
1034
1035         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
1036         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.h:
1037         (MockWebRTCPeerConnectionHandler):
1038
1039 2012-09-26  Michael Saboff  <msaboff@apple.com>
1040
1041         Unreviewed build change.
1042
1043         Reverting inadvertent change to project file.
1044
1045         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
1046
1047 2012-09-26  Harald Tveit Alvestrand  <harald@alvestrand.no>
1048
1049         Implement the GetStats interface on PeerConnection
1050         https://bugs.webkit.org/show_bug.cgi?id=95193
1051
1052         Reviewed by Adam Barth.
1053
1054         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
1055         (RTCStatsRequestSucceededTask):
1056         (RTCStatsRequestSucceededTask::RTCStatsRequestSucceededTask):
1057         (MockWebRTCPeerConnectionHandler::getStats):
1058         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.h:
1059         (MockWebRTCPeerConnectionHandler):
1060
1061 2012-09-26  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
1062
1063         [WK2][WTR] Policy client: dumping from decidePolicyForResponse callback
1064         https://bugs.webkit.org/show_bug.cgi?id=97034
1065
1066         Reviewed by Kenneth Rohde Christiansen.
1067
1068         Provided dumping from WTR Policy client 'decidePolicyForResponse' callback.
1069
1070         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1071         (WTR::InjectedBundlePage::decidePolicyForResponse):
1072
1073 2012-09-26  Zoltan Arvai  <zarvai@inf.u-szeged.hu>
1074
1075         [Qt][WTR] Build fix for MSVC. 
1076         https://bugs.webkit.org/show_bug.cgi?id=97667
1077
1078         Reviewed by Csaba Osztrogonác.
1079
1080         WTR misses InjectedBundle headers. Adding $$PWD to Target.pri is required for MSVC. Clean build is necessary on Windows.
1081
1082         * WebKitTestRunner/InjectedBundle/Target.pri:
1083
1084 2012-09-26  Xiaobo Wang  <xbwang@torchmobile.com.cn>
1085
1086         [BlackBerry] [DRT] eventSender.keyDown() doesn't support Ctrl key modifier
1087         https://bugs.webkit.org/show_bug.cgi?id=97665
1088
1089         Reviewed by Rob Buis.
1090
1091         This patch fixes test fast/forms/input-text-option-delete.html.
1092
1093         * DumpRenderTree/blackberry/EventSender.cpp:
1094         (keyDownCallback):
1095
1096 2012-09-26  Dinu Jacob  <dinu.s.jacob@intel.com>
1097
1098         Unreviewed. Adding myself as a committer.
1099
1100         * Scripts/webkitpy/common/config/committers.py:
1101
1102 2012-09-26  Carlos Garcia Campos  <cgarcia@igalia.com>
1103
1104         [GTK] Remove color codes characters from run-gtk-tests output when not writing to a TTY
1105         https://bugs.webkit.org/show_bug.cgi?id=97660
1106
1107         Reviewed by Philippe Normand.
1108
1109         Remove also the '\r' characters returned by TTY output to try to fix
1110         the double line endings of the script output in the bots.
1111
1112         * Scripts/run-gtk-tests:
1113         (TestRunner.__init__): Compile a regexp for color codes if stdout
1114         it not a TTY.
1115         (TestRunner._run_test_command.parse_line): Remove color codes and
1116         '\r' characters from the line returned by the pseudo-terminal if
1117         stdout is not a TTY.
1118         * gtk/common.py:
1119         (parse_output_lines): Fix the last chunk check to make sure we
1120         always send entire lines to the callback, otherwise color codes
1121         regexp might fail if the line is not complete.
1122
1123 2012-09-25  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
1124
1125         [DRT][WTR] Support overriding the 'WebKitDisplayImagesKey' preference
1126         https://bugs.webkit.org/show_bug.cgi?id=96883
1127
1128         Reviewed by Kenneth Rohde Christiansen.
1129
1130         * DumpRenderTree/efl/TestRunnerEfl.cpp:
1131         (TestRunner::overridePreference): Support the
1132         'WebKitDisplayImagesKey' preference.
1133
1134 2012-09-26  Zan Dobersek  <zandobersek@gmail.com>
1135
1136         [GTK] Enable some of the unstable CSS features
1137         https://bugs.webkit.org/show_bug.cgi?id=97572
1138
1139         Reviewed by Martin Robinson.
1140
1141         Enable the <style scoped> support via DumpRenderTreeSupportGtk method
1142         whenever resetting defualts to consistent values.
1143
1144         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1145         (resetDefaultsToConsistentValues):
1146
1147 2012-09-26  Zan Dobersek  <zandobersek@gmail.com>
1148
1149         REGRESSION (r126189): No more mismatch-did-not-occur failures in reftests
1150         https://bugs.webkit.org/show_bug.cgi?id=97112
1151
1152         Reviewed by Dirk Pranke.
1153
1154         Don't use the reftest's image hash when gathering output of its reference.
1155         This indirectly forces the driver to return the reference's image output
1156         as well. This specifically solves the problem when the reftest and its
1157         mismatch reference are equal but the unexpected match failure is not
1158         reported due to missing image output of the reference file.
1159
1160         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py:
1161         (SingleTestRunner._run_reftest):
1162         * Scripts/webkitpy/layout_tests/port/test.py:
1163         (TestPort.diff_image):
1164         (TestDriver.run_test):
1165
1166 2012-09-25  Ryosuke Niwa  <rniwa@webkit.org>
1167
1168         Update master.cfg to invoke run-perf-tests with --no-show-results and --reset-results
1169         https://bugs.webkit.org/show_bug.cgi?id=97613
1170
1171         Reviewed by Dirk Pranke.
1172
1173         Use options added in r129580.
1174
1175         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1176         (RunAndUploadPerfTests):
1177
1178 2012-09-25  Peter Rybin  <peter.rybin@gmail.com>
1179
1180         Web Inspector: Fix logging in pure protocol test harness
1181         https://bugs.webkit.org/show_bug.cgi?id=97579
1182
1183         Reviewed by Yury Semikhatsky.
1184
1185         JavaScript alert implementation in DumpRenderTree gets immediate flush to make sure that
1186         the message survives crash or kill by timeout.
1187
1188         * DumpRenderTree/blackberry/DumpRenderTree.cpp:
1189         (BlackBerry::WebKit::DumpRenderTree::runJavaScriptAlert):
1190         * DumpRenderTree/chromium/WebViewHost.cpp:
1191         (WebViewHost::runModalAlertDialog):
1192         * DumpRenderTree/efl/DumpRenderTreeView.cpp:
1193         (onJavaScriptAlert):
1194         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1195         (webViewScriptAlert):
1196         * DumpRenderTree/mac/UIDelegate.mm:
1197         (-[UIDelegate webView:runJavaScriptAlertPanelWithMessage:initiatedByFrame:]):
1198         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1199         (WebCore::WebPage::javaScriptAlert):
1200         * DumpRenderTree/win/UIDelegate.cpp:
1201         (UIDelegate::runJavaScriptAlertPanelWithMessage):
1202
1203 2012-09-25  Dan Bernstein  <mitz@apple.com>
1204
1205         Tools part of <rdar://problem/11455228> [mac] Stop using screen fonts
1206         https://bugs.webkit.org/show_bug.cgi?id=97620
1207
1208         Reviewed by John Sullivan.
1209
1210         * DumpRenderTree/mac/DumpRenderTree.mm:
1211         (resetDefaultsToConsistentValues): Enable screen font substitution when
1212         running the tests.
1213         * WebKitTestRunner/TestController.cpp:
1214         (WTR::TestController::resetStateToConsistentValues): Ditto.
1215
1216 2012-09-25  Ryosuke Niwa  <rniwa@webkit.org>
1217
1218         Build fix. "values" doesn't necessarily contain multiple values but it needs to be always parsed as an array.
1219
1220         * Scripts/webkitpy/performance_tests/perftest.py:
1221         (PerfTest.parse_output):
1222
1223 2012-09-25  Ryosuke Niwa  <rniwa@webkit.org>
1224
1225         run-perf-tests: cleanup options and results generation code
1226         https://bugs.webkit.org/show_bug.cgi?id=97611
1227
1228         Reviewed by Dirk Pranke.
1229
1230         Previously, --test-results-server triggered old JSOn format where the outermost structure was a dictionary
1231         instead of an array, and also implicitly triggered --no-show-results, caused the old outputs not to merge,
1232         and prevented the generation of results page. Also, it was not obvious that --source-json-path is an option
1233         used only on buildbot slaves.
1234
1235         This patch will:
1236         - Remove the old format since perf-o-matic supports new format now.
1237         - Add --reset-results option so that we can explicitly clear existing outputs.
1238         - Add --slave-config-json-path option to replace --source-json-path option.
1239
1240         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
1241         (PerfTestsRunner._parse_args): Added --reset-results and --slave-config-json-path options.
1242         (PerfTestsRunner._generate_and_show_results): Refactored. Also removed the code to strip "values" from
1243         results since perf-o-matic can parse and store these values now.
1244         (PerfTestsRunner._merge_outputs_if_needed): Renamed from _merge_outputs_if_needed.
1245         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
1246         (test_run_with_json_output): Test a harmless behavioral change to generate resuls page even when
1247         --test-results-server is present.
1248         (test_run_with_description):
1249         (test_run_generates_json_by_default):
1250         (test_run_merges_output_by_default): Added.
1251         (test_run_respects_reset_results): Added.
1252         (test_run_with_slave_config_json):
1253         (test_run_with_bad_slave_config_json): Use --slave-config-json-path instead of --source-json-path to make
1254         sure this optioon works as well.
1255         (test_run_with_multiple_repositories):
1256         (test_run_with_upload_json):
1257
1258 2012-09-25  Simon Fraser  <simon.fraser@apple.com>
1259
1260         Fix SnowLeopard build by adding #ifdefs.
1261         
1262         Fix crashes on some tests where the CFURLRef appears to be null.
1263
1264         * WebKitTestRunner/InjectedBundle/mac/InjectedBundlePageMac.mm:
1265         (WTR::testPathFromURL):
1266         (WTR::InjectedBundlePage::platformDidStartProvisionalLoadForFrame):
1267
1268 2012-09-25  Simon Fraser  <simon.fraser@apple.com>
1269
1270         Speculative build fix for Production builds.
1271
1272         * DumpRenderTree/mac/Configurations/Base.xcconfig:
1273         * WebKitTestRunner/Configurations/Base.xcconfig:
1274
1275 2012-09-25  Simon Fraser  <simon.fraser@apple.com>
1276
1277         Fix the build after r129566.
1278
1279         * DumpRenderTree/mac/FrameLoadDelegate.mm:
1280
1281 2012-09-25  Dominic Mazzoni  <dmazzoni@google.com>
1282
1283         AX: AccessibilityUIElementChromium.cpp won't build on Windows without wtf/StringExtras.h
1284         https://bugs.webkit.org/show_bug.cgi?id=97614
1285
1286         Unreviewed, fixing windows breakage.
1287
1288         Added wtf/StringExtras.h to the includes to fix the Chromium/Windows build.
1289
1290         * DumpRenderTree/chromium/TestRunner/AccessibilityUIElementChromium.cpp:
1291
1292 2012-09-24  Simon Fraser  <simon.fraser@apple.com>
1293
1294         <rdar://problem/12351906> Have DumpRenderTree and WebKitTestRunner crash logs show which test crashed
1295
1296         Reviewed by Mark Rowe.
1297
1298         Use a new WebKitSytemInterface function to add data to crash logs about which 
1299         test was running when the crash happened.
1300
1301         * DumpRenderTree/mac/Configurations/DebugRelease.xcconfig: Make sure WEBKIT_SYSTEM_INTERFACE_LIBRARY is defined.
1302         * DumpRenderTree/mac/Configurations/DumpRenderTree.xcconfig: link with the WebKitSystemInterface library.
1303         * DumpRenderTree/mac/FrameLoadDelegate.mm:
1304         (testPathFromURL): Given a (probably) file URL, return the part after "/LayoutTests/"
1305         (-[FrameLoadDelegate webView:didStartProvisionalLoadForFrame:]): Tell Crash Reporter which test was running,
1306         if we're the main frame and the test is active.
1307         * WebKitTestRunner/Configurations/DebugRelease.xcconfig: Make sure WEBKIT_SYSTEM_INTERFACE_LIBRARY is defined.
1308         * WebKitTestRunner/Configurations/InjectedBundle.xcconfig: link the injected bundle with the WebKitSystemInterface library.
1309         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1310         (WTR::InjectedBundlePage::didStartProvisionalLoadForFrame): Call platformDidStartProvisionalLoadForFrame().
1311         (WTR::InjectedBundlePage::platformDidStartProvisionalLoadForFrame): Stub for non-Mac platforms.
1312         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
1313         Add a platformDidStartProvisionalLoadForFrame() for platform-specific work.
1314         * WebKitTestRunner/InjectedBundle/mac/InjectedBundlePageMac.mm: Added.
1315         (WTR::testPathFromURL): Given a (probably) file URL, return the part after "/LayoutTests/"
1316         (WTR::InjectedBundlePage::platformDidStartProvisionalLoadForFrame): Tell Crash Reporter which test was running,
1317         if we're the main frame.
1318         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj: Add InjectedBundlePageMac.mm.
1319
1320 2012-09-25  Julie Parent  <jparent@chromium.org>
1321
1322         Make test_shell_tests show up in flakiness dashboard.
1323         https://bugs.webkit.org/show_bug.cgi?id=97585
1324
1325         Reviewed by Dirk Pranke.
1326
1327         test_shell_tests run only on chromium.org, and on the same bots as
1328         the layout tests, so do not use the generic fallbacks.
1329
1330         * TestResultServer/static-dashboards/builders.js:
1331         (loadBuildersList):
1332         * TestResultServer/static-dashboards/dashboard_base.js:
1333         (currentBuilderGroupCategory):
1334
1335 2012-09-25  Lucas Forschler  <lforschler@apple.com>
1336
1337         Python tests fail on Windows bots when running with more than one child process.  Temporarily use only one child process until the problem can be fixed.
1338         https://bugs.webkit.org/show_bug.cgi?id=97465
1339         
1340         Reviewed by Tim Horton.
1341
1342         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1343         (RunPythonTests.start):
1344
1345 2012-09-25  Mike West  <mkwst@chromium.org>
1346
1347         Unreviewed. Adding myself as a committer.
1348
1349         Yay. :)
1350
1351         * Scripts/webkitpy/common/config/committers.py:
1352
1353 2012-09-25  Beth Dakin  <bdakin@apple.com>
1354
1355         https://bugs.webkit.org/show_bug.cgi?id=95397
1356         Need to merge didFirstVisuallyNonEmptyLayout and 
1357         didNewFirstVisuallyNonEmptyLayout
1358         -and corresponding-
1359         <rdar://problem/10791680>
1360
1361         Reviewed by Sam Weinig.
1362
1363         Account for the new member of the PageLoadClient.
1364         * MiniBrowser/mac/BrowserWindowController.m:
1365         (-[BrowserWindowController awakeFromNib]):
1366         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1367         (WTR::InjectedBundlePage::InjectedBundlePage):
1368         * WebKitTestRunner/TestController.cpp:
1369         (WTR::TestController::initialize):
1370
1371         The NewFirstVisuallyNonEmptyLayoutForImages has existed for a while, 
1372         but it did not have the right target membership.
1373         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
1374
1375         Use new didLayout API instead of old 
1376         didNewFirstVisuallyNonEmptyLayout API.
1377         * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayout.cpp:
1378         (TestWebKitAPI):
1379         (TestWebKitAPI::didLayout):
1380         (TestWebKitAPI::setPageLoaderClient):
1381         (TestWebKitAPI::TEST):
1382         * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayoutFails.cpp:
1383         (TestWebKitAPI):
1384         (TestWebKitAPI::didLayout):
1385         (TestWebKitAPI::setPageLoaderClient):
1386         (TestWebKitAPI::TEST):
1387         * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayoutForImages.cpp:
1388         (TestWebKitAPI):
1389         (TestWebKitAPI::didLayout):
1390         (TestWebKitAPI::setPageLoaderClient):
1391         (TestWebKitAPI::TEST):
1392         * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayoutFrames.cpp:
1393         (TestWebKitAPI):
1394         (TestWebKitAPI::didLayout):
1395         (TestWebKitAPI::setPageLoaderClient):
1396         (TestWebKitAPI::TEST):
1397
1398 2012-09-25  Benjamin Poulain  <bpoulain@apple.com>
1399
1400         Add missing support for Geolocation tests on WebKit2
1401         https://bugs.webkit.org/show_bug.cgi?id=97518
1402
1403         Reviewed by Sam Weinig.
1404
1405         There are two main features added to the tooling to make the test run:
1406         -Add support for error reporting for position errors through setMockGeolocationPositionUnavailableError.
1407         -Support the 3 states permission: undefined-allowed-denied.
1408
1409         * WebKitTestRunner/GeolocationProviderMock.cpp:
1410         (WTR::GeolocationProviderMock::GeolocationProviderMock):
1411         (WTR::GeolocationProviderMock::setPosition):
1412         (WTR::GeolocationProviderMock::setPositionUnavailableError):
1413         (WTR::GeolocationProviderMock::startUpdating):
1414         (WTR::GeolocationProviderMock::sendPositionIfNeeded):
1415         (WTR::GeolocationProviderMock::sendErrorIfNeeded):
1416         * WebKitTestRunner/GeolocationProviderMock.h:
1417         (GeolocationProviderMock):
1418         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
1419         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
1420         (WTR::InjectedBundle::setMockGeolocationPositionUnavailableError):
1421         * WebKitTestRunner/InjectedBundle/InjectedBundle.h:
1422         (InjectedBundle):
1423         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
1424         (WTR::TestRunner::setMockGeolocationPositionUnavailableError):
1425         * WebKitTestRunner/InjectedBundle/TestRunner.h:
1426         (TestRunner):
1427         * WebKitTestRunner/TestController.cpp:
1428         (WTR::TestController::TestController):
1429         (WTR::decidePolicyForGeolocationPermissionRequest):
1430         (WTR::TestController::resetStateToConsistentValues):
1431         (WTR::TestController::setGeolocationPermission):
1432         (WTR::TestController::setMockGeolocationPosition):
1433         (WTR::TestController::setMockGeolocationPositionUnavailableError):
1434         (WTR::TestController::handleGeolocationPermissionRequest):
1435         (WTR::TestController::decidePolicyForGeolocationPermissionRequestIfPossible):
1436         * WebKitTestRunner/TestController.h:
1437         (TestController):
1438         * WebKitTestRunner/TestInvocation.cpp:
1439         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
1440
1441 2012-09-25  Sheriff Bot  <webkit.review.bot@gmail.com>
1442
1443         Unreviewed, rolling out r129517.
1444         http://trac.webkit.org/changeset/129517
1445         https://bugs.webkit.org/show_bug.cgi?id=97582
1446
1447         Link errors in chromium (Requested by alecf on #webkit).
1448
1449         * DumpRenderTree/DumpRenderTree.gypi:
1450         * DumpRenderTree/chromium/MockConstraints.cpp: Removed.
1451         * DumpRenderTree/chromium/MockConstraints.h: Removed.
1452         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
1453         (RTCSessionDescriptionRequestSuccededTask::RTCSessionDescriptionRequestSuccededTask):
1454         (RTCSessionDescriptionRequestFailedTask::RTCSessionDescriptionRequestFailedTask):
1455         (RTCVoidRequestTask::RTCVoidRequestTask):
1456         (isSupportedConstraint):
1457         (isValidConstraint):
1458         (MockWebRTCPeerConnectionHandler::initialize):
1459         * DumpRenderTree/chromium/WebUserMediaClientMock.cpp:
1460         (WebKit::WebUserMediaClientMock::requestUserMedia):
1461         * DumpRenderTree/chromium/WebUserMediaClientMock.h:
1462         (WebKit):
1463         (WebUserMediaClientMock):
1464         * DumpRenderTree/chromium/WebViewHost.h:
1465         (WebKit):
1466         (WebViewHost):
1467
1468 2012-09-25  Dominic Mazzoni  <dmazzoni@google.com>
1469
1470         AX: Chromium doesn't pass accessibility text range and line number tests
1471         https://bugs.webkit.org/show_bug.cgi?id=97246
1472
1473         Reviewed by Chris Fleizach.
1474
1475         Provide a chromium implementation of some AccessibilityUIElement
1476         methods needed by accessibility tests.
1477
1478         Enables four existing tests to pass on Chromium:
1479         Tests: accessibility/textarea-insertion-point-line-number.html
1480                accessibility/textarea-line-for-index.html
1481                accessibility/textarea-selected-text-range.html
1482                accessibility/textbox-role-reports-selection.html
1483
1484         * DumpRenderTree/chromium/TestRunner/AccessibilityUIElementChromium.cpp:
1485         (AccessibilityUIElement::insertionPointLineNumberGetterCallback):
1486         (AccessibilityUIElement::selectedTextRangeGetterCallback):
1487         (AccessibilityUIElement::lineForIndexCallback):
1488         (AccessibilityUIElement::setSelectedTextRangeCallback):
1489
1490 2012-09-25  Peter Beverloo  <peter@chromium.org>
1491
1492         [Chromium] Android's build archives have grown from 38M to 102M in 11 days
1493         https://bugs.webkit.org/show_bug.cgi?id=97336
1494
1495         Reviewed by Adam Barth.
1496
1497         Each WebKit target results in four APKs in the out/{Debug,Release}/ directory,
1498         while we really only need one. Ignoring the unsigned APKs removes six from
1499         the archive will make sure that we only archive a single APK per target,
1500         bringing its size back to rougly 40 megabytes.
1501
1502         * BuildSlaveSupport/built-product-archive:
1503         (archiveBuiltProduct):
1504
1505 2012-09-25  Tommy Widenflycht  <tommyw@google.com>
1506
1507         MediaStream API: Update getUserMedia to match the latest specification
1508         https://bugs.webkit.org/show_bug.cgi?id=97540
1509
1510         Reviewed by Adam Barth.
1511
1512         Adding mock support for the enhanced getUserMedia call.
1513
1514         * DumpRenderTree/DumpRenderTree.gypi:
1515         * DumpRenderTree/chromium/MockConstraints.cpp: Copied from Source/WebCore/platform/chromium/support/WebMediaConstraints.cpp.
1516         (MockConstraints):
1517         (MockConstraints::isSupported):
1518         (MockConstraints::isValid):
1519         (MockConstraints::verify):
1520         * DumpRenderTree/chromium/MockConstraints.h: Copied from Tools/DumpRenderTree/chromium/WebUserMediaClientMock.h.
1521         (WebKit):
1522         (MockConstraints):
1523         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
1524         (RTCSessionDescriptionRequestSuccededTask::RTCSessionDescriptionRequestSuccededTask):
1525         (RTCSessionDescriptionRequestFailedTask::RTCSessionDescriptionRequestFailedTask):
1526         (RTCVoidRequestTask::RTCVoidRequestTask):
1527         (MockWebRTCPeerConnectionHandler::initialize):
1528         * DumpRenderTree/chromium/WebUserMediaClientMock.cpp:
1529         (UserMediaRequestTask):
1530         (UserMediaRequestTask::UserMediaRequestTask):
1531         (WebUserMediaClientMock::requestUserMedia):
1532         * DumpRenderTree/chromium/WebUserMediaClientMock.h:
1533         (WebUserMediaClientMock):
1534         (WebUserMediaClientMock::taskList):
1535         * DumpRenderTree/chromium/WebViewHost.h:
1536         (WebKit):
1537         (WebViewHost):
1538
1539 2012-09-25  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
1540
1541         [EFL] Do not dump inspector output in DRT
1542         https://bugs.webkit.org/show_bug.cgi?id=97018
1543
1544         Reviewed by Kenneth Rohde Christiansen.
1545
1546         Some inspector-related layout tests expect console output from the
1547         inspector view to be suppressed. Do that by setting a property in
1548         the inspector view that makes the onConsoleMessage callback return
1549         early and not output anything.
1550
1551         Using evas_object_data_{get,set} is not very beautiful, but it is
1552         still better than properly subclassing Ewk_View and having to
1553         implement something similar to drt_view_console_message_enable_set().
1554
1555         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
1556         (DumpRenderTreeChrome::createWebInspectorView):
1557         * DumpRenderTree/efl/DumpRenderTreeView.cpp:
1558         (onConsoleMessage):
1559
1560 2012-09-25  Christophe Dumez  <christophe.dumez@intel.com>
1561
1562         [WK2][WKTR] TestRunner needs support for setWillSendRequestReturnsNullOnRedirect
1563         https://bugs.webkit.org/show_bug.cgi?id=97542
1564
1565         Reviewed by Kenneth Rohde Christiansen.
1566
1567         Implement support for setWillSendRequestReturnsNullOnRedirect
1568         in WebKitTestRunner.
1569
1570         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
1571         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1572         (WTR::InjectedBundlePage::willSendRequestForFrame):
1573         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
1574         (WTR::TestRunner::TestRunner):
1575         * WebKitTestRunner/InjectedBundle/TestRunner.h:
1576         (WTR::TestRunner::willSendRequestReturnsNullOnRedirect):
1577         (WTR::TestRunner::setWillSendRequestReturnsNullOnRedirect):
1578         (TestRunner):
1579
1580 2012-09-25  Xiaobo Wang  <xbwang@torchmobile.com.cn>
1581
1582         [BlackBerry] [DRT] fast/forms/textarea-live-pseudo-selectors.html failed
1583         https://bugs.webkit.org/show_bug.cgi?id=97189
1584
1585         Reviewed by Rob Buis.
1586
1587         Translate ASCII key code to BlackBerry key code for BACKSPACE and DELETE key.
1588
1589         * DumpRenderTree/blackberry/EventSender.cpp:
1590         (keyDownCallback):
1591
1592 2012-09-25  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
1593
1594         [WK2][WTR] WTR bundle client loads binary data as text
1595         https://bugs.webkit.org/show_bug.cgi?id=97532
1596
1597         Reviewed by Kenneth Rohde Christiansen.
1598
1599         WTR::InjectedBundlePage::decidePolicyForResponse now checks response MIME type
1600         and returns WKBundlePagePolicyActionPassThrough if it cannot be shown.
1601
1602         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1603         (WTR::InjectedBundlePage::decidePolicyForResponse):
1604
1605 2012-09-25  Zan Dobersek  <zandobersek@gmail.com>
1606
1607         Unreviewed, another test case adjustment after r129459. Python 2.7 doesn't
1608         seem to be a requirement yet so the assertDictEqual method is not yet available
1609         on all builders. Assert that the empty dictionary length equals 0 instead.
1610
1611         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
1612         (ResultSummaryTest.test_interpret_test_failures):
1613
1614 2012-09-24  Simon Hausmann  <simon.hausmann@digia.com>
1615
1616         [Qt] Fix build with latest Qt 5
1617         https://bugs.webkit.org/show_bug.cgi?id=97479
1618
1619         Reviewed by Tor Arne Vestbø.
1620
1621         The variables for syncqt handling are now set in api.pri, but let's add a comment here that we can
1622         remove them from default_pre.prf once we're past Qt 5 beta2.
1623
1624         * qmake/mkspecs/features/default_pre.prf:
1625
1626 2012-09-25  Zan Dobersek  <zandobersek@gmail.com>
1627
1628         Unreviewed, adjusting the test cases after r129459. The JSON data
1629         for reftests now contains the 'reftest_type' key and no more 'is_reftest'
1630         or 'is_mismatch_reftest' keys.
1631
1632         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1633         (EndToEndTest.test_reftest_with_two_notrefs):
1634
1635 2012-09-24  Zan Dobersek  <zandobersek@gmail.com>
1636
1637         Unexpected reftest passes are only reported when pixel testing is enabled in results.html as well
1638         https://bugs.webkit.org/show_bug.cgi?id=97426
1639
1640         Reviewed by Ojan Vafai.
1641
1642         The 'is_reftest' member of the TestResult class has been replaced by the reftest_type
1643         member which is a set containing either '!=', '==', both or none if the test represented
1644         by the object is not a reftest.
1645
1646         When summarizing results, the test dictionary is updated to contain the 'reftest_type' key
1647         if the test is a reftest. The value of this key is a list of all the reftest types of this test.
1648
1649         The test failure interpretation method has been refactored, it now reports only
1650         missing results of various formats and image diff percent in case of image or
1651         reftest mismatch.
1652
1653         Unit tests have been modified accordingly.
1654
1655         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner.py:
1656         (LayoutTestRunner._update_summary_with_result):
1657         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner_unittest.py:
1658         (LayoutTestRunnerTests.test_update_summary_with_result):
1659         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1660         (interpret_test_failures):
1661         (summarize_results):
1662         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
1663         (ResultSummaryTest.test_interpret_test_failures):
1664         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py:
1665         (SingleTestRunner.run):
1666         (SingleTestRunner._run_reftest):
1667         * Scripts/webkitpy/layout_tests/models/test_results.py:
1668         (TestResult.__init__):
1669         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1670         (EndToEndTest.test_reftest_with_two_notrefs):
1671
1672 2012-09-24  Sam Weinig  <sam@webkit.org>
1673
1674         Use NSUserDefaults rather than an environment variable to control whether to use an XPC Service for the WebProcess
1675         https://bugs.webkit.org/show_bug.cgi?id=97514
1676
1677         Reviewed by Anders Carlsson.
1678
1679         * Scripts/webkitdirs.pm:
1680         (argumentsForRunAndDebugMacWebKitApp):
1681         Push "-WebKit2UseXPCServiceForWebProcess YES" as arguments when --use-web-process-xpc-service
1682         is provided to set the user default. Move adding of @ARGV to after custom additions as calling
1683         shouldUseXPCServiceForWebProcess() can change @ARGV.
1684
1685         (runMacWebKitApp):
1686         (execMacWebKitAppForDebugging):
1687         Stop setting the WEBKIT_USE_XPC_SERVICE_FOR_WEB_PROCESS environment variable.
1688
1689 2012-09-24  Laszlo Gombos  <l.gombos@samsung.com>
1690
1691         [GTK][EFL] Make sure cairo is built with fontconfig support
1692         https://bugs.webkit.org/show_bug.cgi?id=97513
1693
1694         Reviewed by Martin Robinson.
1695
1696         The GTK and EFL port of WebKit requires a version of cairo that is built with fontconfig support. 
1697
1698         * efl/jhbuild.modules:
1699         * gtk/jhbuild.modules:
1700
1701 2012-09-24  Benjamin Poulain  <bpoulain@apple.com>
1702
1703         Fix Geolocation error reporting in the test support
1704         https://bugs.webkit.org/show_bug.cgi?id=97386
1705
1706         Reviewed by Sam Weinig.
1707         The TestRunner was exposing internal implementation details of
1708         the Geolocation object through setMockGeolocationError().
1709
1710         In practice, only the error code PositionUnavailable is supported
1711         or even implemented.
1712
1713         Exposing setMockGeolocationError() does not make sense for WebKit2 as
1714         passing generic errors from the UIProcess is a bad idea. This is the
1715         opportunity to clean TestRunner and prepare and interface suitable for WebKit2.
1716
1717         * DumpRenderTree/TestRunner.cpp:
1718         (setMockGeolocationPositionUnavailableErrorCallback):
1719         (TestRunner::staticFunctions):
1720         * DumpRenderTree/TestRunner.h:
1721         (TestRunner):
1722         * DumpRenderTree/blackberry/TestRunnerBlackBerry.cpp:
1723         (TestRunner::setMockGeolocationPositionUnavailableError):
1724         * DumpRenderTree/chromium/DRTTestRunner.cpp:
1725         (DRTTestRunner::DRTTestRunner):
1726         (DRTTestRunner::setMockGeolocationPositionUnavailableError):
1727         * DumpRenderTree/chromium/DRTTestRunner.h:
1728         (DRTTestRunner):
1729         * DumpRenderTree/efl/TestRunnerEfl.cpp:
1730         (TestRunner::setMockGeolocationPositionUnavailableError):
1731         * DumpRenderTree/gtk/TestRunnerGtk.cpp:
1732         (TestRunner::setMockGeolocationPositionUnavailableError):
1733         * DumpRenderTree/mac/MockGeolocationProvider.h:
1734         * DumpRenderTree/mac/MockGeolocationProvider.mm:
1735         (-[MockGeolocationProvider dealloc]):
1736         (-[MockGeolocationProvider resetError]):
1737         (-[MockGeolocationProvider setPosition:]):
1738         (-[MockGeolocationProvider setPositionUnavailableErrorWithMessage:]):
1739         (-[MockGeolocationProvider lastPosition]):
1740         (-[MockGeolocationProvider timerFired]):
1741         * DumpRenderTree/mac/TestRunnerMac.mm:
1742         (TestRunner::setMockGeolocationPositionUnavailableError):
1743         * DumpRenderTree/qt/TestRunnerQt.cpp:
1744         (TestRunner::setMockGeolocationPositionUnavailableError):
1745         * DumpRenderTree/qt/TestRunnerQt.h:
1746         (TestRunner):
1747         * DumpRenderTree/win/TestRunnerWin.cpp:
1748         (TestRunner::setMockGeolocationPositionUnavailableError):
1749         * DumpRenderTree/wx/TestRunnerWx.cpp:
1750         (TestRunner::setMockGeolocationPositionUnavailableError):
1751
1752 2012-09-24  Nikhil Bansal  <n.bansal@samsung.com>
1753
1754         [EFL][WK2] TestRunner needs touch events support.
1755         https://bugs.webkit.org/show_bug.cgi?id=96465
1756
1757         Reviewed by Kenneth Rohde Christiansen.
1758
1759         Simulate touch events in WebkitTestRunner for touch related
1760         layout tests. This patch uses ewk_view_feed_touch_event() to
1761         generate touch events.
1762
1763         * WebKitTestRunner/EventSenderProxy.h:
1764         (EventSenderProxy):
1765         * WebKitTestRunner/efl/EventSenderProxyEfl.cpp:
1766         (WTR):
1767         (WTR::setEvasModifiers):
1768         (WTR::EventSenderProxy::EventSenderProxy):
1769         (WTR::EventSenderProxy::~EventSenderProxy):
1770         (WTR::EventSenderProxy::sendTouchEvent):
1771         (WTR::EventSenderProxy::addTouchPoint):
1772         (WTR::EventSenderProxy::updateTouchPoint):
1773         (WTR::EventSenderProxy::setTouchModifier):
1774         (WTR::EventSenderProxy::touchStart):
1775         (WTR::EventSenderProxy::touchMove):
1776         (WTR::EventSenderProxy::touchEnd):
1777         (WTR::EventSenderProxy::touchCancel):
1778         (WTR::EventSenderProxy::clearTouchPoints):
1779         (WTR::EventSenderProxy::releaseTouchPoint):
1780         (WTR::EventSenderProxy::cancelTouchPoint):
1781
1782 2012-09-24  Christophe Dumez  <christophe.dumez@intel.com>
1783
1784         [WK2] TestRunner does not support overridePreference with value "0"
1785         https://bugs.webkit.org/show_bug.cgi?id=97457
1786
1787         Reviewed by Alexey Proskuryakov.
1788
1789         testRunner.overridePreference() was marked as taking a boolean
1790         in argument in testRunner.idl even though the test cases are
1791         passing strings to it. The current code works if the test
1792         case wants to enable a setting and therefore passes "1" string.
1793         However, if the test wants to disable a setting and passes "0"
1794         string, the current code also evaluates this to true causing
1795         certain test cases to fail.
1796
1797         The argument to testRunner.overridePreference() is now of string
1798         type and a function has been added to convert the string passed
1799         by the test to a boolean value.
1800
1801         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
1802         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
1803         (WTR::toBool):
1804         (WTR):
1805         (WTR::TestRunner::overridePreference):
1806         * WebKitTestRunner/InjectedBundle/TestRunner.h:
1807         (TestRunner):
1808
1809 2012-09-24  Sheriff Bot  <webkit.review.bot@gmail.com>
1810
1811         Unreviewed, rolling out r129343.
1812         http://trac.webkit.org/changeset/129343
1813         https://bugs.webkit.org/show_bug.cgi?id=97479
1814
1815         It broke the build (Requested by Ossy_NIGHT on #webkit).
1816
1817         * qmake/mkspecs/features/default_pre.prf:
1818
1819 2012-09-24  Zan Dobersek  <zandobersek@gmail.com>
1820
1821         [Gtk] Unskip the Shadow DOM layout tests
1822         https://bugs.webkit.org/show_bug.cgi?id=90776
1823
1824         Reviewed by Ryosuke Niwa.
1825
1826         Enable Shadow DOM through DumpRenderTreeSupport when resetting defaults in between tests.
1827
1828         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1829         (resetDefaultsToConsistentValues):
1830
1831 2012-09-24  Zan Dobersek  <zandobersek@gmail.com>
1832
1833         REGRESSION (r129211-r129218): http/tests/loading/redirect-with-no-location-crash.html failing on GTK Linux 64-bit Release WK2 (Tests)
1834         https://bugs.webkit.org/show_bug.cgi?id=97404
1835
1836         Reviewed by Philippe Normand.
1837
1838         Print out the 'didReceiveTitle' information when dumping frame load callbacks.
1839
1840         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1841         (webViewTitleChanged):
1842
1843 2012-09-24  Simon Hausmann  <simon.hausmann@digia.com>
1844
1845         [Qt] Fix build with latest Qt 5
1846
1847         Reviewed by Tor Arne Vestbø.
1848
1849         The variables for syncqt handling are now set in api.pri, but let's add a comment here that we can
1850         remove them from default_pre.prf once we're past Qt 5 beta2.
1851
1852         * qmake/mkspecs/features/default_pre.prf:
1853
1854 2012-09-22  Dominic Mazzoni  <dmazzoni@google.com>
1855
1856         AX: Layout tests would be easier to write if AccessibilityController could find an element by id
1857         https://bugs.webkit.org/show_bug.cgi?id=97301
1858
1859         Reviewed by Chris Fleizach.
1860
1861         Adds a new accessibleElementById method to AccessibilityController, to make it
1862         easier to write cross-platform layout tests.
1863
1864         Implemented for Chromium and for Mac (DRT and WKTR).
1865
1866         * DumpRenderTree/AccessibilityController.cpp:
1867         (getAccessibleElementByIdCallback):
1868         (AccessibilityController::getJSClass):
1869         * DumpRenderTree/AccessibilityController.h:
1870         (AccessibilityController):
1871         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
1872         * DumpRenderTree/chromium/TestRunner/AccessibilityControllerChromium.cpp:
1873         (AccessibilityController::AccessibilityController):
1874         (AccessibilityController::findAccessibleElementByIdRecursive):
1875         (AccessibilityController::getAccessibleElementById):
1876         (AccessibilityController::accessibleElementByIdGetterCallback):
1877         * DumpRenderTree/chromium/TestRunner/AccessibilityControllerChromium.h:
1878         (AccessibilityController):
1879         * DumpRenderTree/gtk/AccessibilityControllerGtk.cpp:
1880         (AccessibilityController::accessibleElementById):
1881         * DumpRenderTree/mac/AccessibilityCommonMac.h: added
1882         * DumpRenderTree/mac/AccessibilityCommonMac.mm: added
1883         (+[NSString stringWithJSStringRef:]):
1884         (-[NSString createJSStringRef]):
1885         * DumpRenderTree/mac/AccessibilityControllerMac.mm:
1886         (findAccessibleObjectById):
1887         (AccessibilityController::accessibleElementById):
1888         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
1889         * DumpRenderTree/win/AccessibilityControllerWin.cpp:
1890         (AccessibilityController::accessibleElementById):
1891         * WebKitTestRunner/InjectedBundle/AccessibilityController.cpp:
1892         (WTR::AccessibilityController::accessibleElementById):
1893         * WebKitTestRunner/InjectedBundle/AccessibilityController.h:
1894         (AccessibilityController):
1895         * WebKitTestRunner/InjectedBundle/Bindings/AccessibilityController.idl:
1896         * WebKitTestRunner/InjectedBundle/mac/AccessibilityCommonMac.h: added
1897         * WebKitTestRunner/InjectedBundle/mac/AccessibilityCommonMac.mm: added
1898         (+[NSString stringWithJSStringRef:]):
1899         (-[NSString createJSStringRef]):
1900         * WebKitTestRunner/InjectedBundle/mac/AccessibilityControllerMac.mm:
1901         (WTR::findAccessibleObjectById):
1902         (WTR):
1903         (WTR::AccessibilityController::accessibleElementById):
1904         * WebKitTestRunner/InjectedBundle/mac/AccessibilityUIElementMac.mm:
1905         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
1906
1907 2012-09-22  Zan Dobersek  <zandobersek@gmail.com>
1908
1909         Unexpected reftest passes are only reported when pixel testing is enabled
1910         https://bugs.webkit.org/show_bug.cgi?id=97242
1911
1912         Reviewed by Dirk Pranke.
1913
1914         LayoutTestRunner should know whether the finished test it's handling is
1915         a reftest. This is necessary when updating result summary for reftests
1916         when pixel testing is not enabled - in these circumstances an unexpectedly
1917         passing reftest was not reported as such due to pixel testing then being
1918         determined by searching for reftest failures in failures list, which were
1919         not found in an empty failure list of a passing test. Now the TestResult
1920         objects have a member variable indicating the test is a reftest whenever
1921         that's the case.
1922
1923         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner.py:
1924         (LayoutTestRunner._update_summary_with_result):
1925         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner_unittest.py:
1926         (LayoutTestRunnerTests.test_update_summary_with_result):
1927         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py:
1928         (SingleTestRunner.run):
1929         (SingleTestRunner._run_reftest):
1930         * Scripts/webkitpy/layout_tests/models/test_results.py:
1931         (TestResult.__init__):
1932
1933 2012-09-21  Sam Weinig  <sam@webkit.org>
1934
1935         WebProcess XPC services need have their environment set without disrupting all other XPC services
1936         https://bugs.webkit.org/show_bug.cgi?id=95161
1937
1938         Reviewed by Anders Carlsson.
1939
1940         Remove setting the __XPC_* environment variables.
1941
1942         * Scripts/webkitdirs.pm:
1943         (setUpGuardMallocIfNeeded):
1944         (runMacWebKitApp):
1945         (execMacWebKitAppForDebugging):
1946
1947 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
1948
1949         Fix typo in additional-platform-directory patch just landed
1950         https://bugs.webkit.org/show_bug.cgi?id=97380
1951         
1952         Unreviewed, build fix.
1953
1954         * Scripts/webkitpy/layout_tests/port/base.py:
1955         (Port.relative_test_filename):
1956         (Port.relative_perf_test_filename):
1957
1958 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
1959
1960         nrwt: don't require additional-platform-directory to be an abspath or live under LayoutTests
1961         https://bugs.webkit.org/show_bug.cgi?id=97380
1962
1963         Reviewed by Ojan Vafai.
1964
1965         There doesn't seem to be a good reason for this restriction and
1966         it's useful to be able to point to directories outside the
1967         checkout for results (e.g., for local failures due to a 10.7.4
1968         install ;).
1969
1970         * Scripts/webkitpy/layout_tests/port/base.py:
1971         (Port.relative_test_filename):
1972         (Port.relative_perf_test_filename):
1973         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
1974         (ChromiumAndroidDriver._command_from_driver_input):
1975         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1976         (_set_up_derived_options):
1977         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1978         (MainTest.test_additional_platform_directory):
1979
1980 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
1981
1982         webkitpy: drop support for old TestExpectations syntax
1983         https://bugs.webkit.org/show_bug.cgi?id=97364
1984
1985         Reviewed by Ryosuke Niwa.
1986
1987         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
1988         (TestExpectationParser._collect_matching_tests):
1989         (TestExpectationParser):
1990         (TestExpectationParser._tokenize_line):
1991         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
1992         (SkippedTests.test_skipped_entry_dont_exist):
1993         (ExpectationSyntaxTests.assert_tokenize_exp):
1994         (ExpectationSyntaxTests.test_bare_name):
1995         (ExpectationSyntaxTests.test_bare_name_and_bugs):
1996         (ExpectationSyntaxTests.test_comments):
1997         (ExpectationSyntaxTests.test_config_modifiers):
1998         (ExpectationSyntaxTests.test_unknown_config):
1999         (ExpectationSyntaxTests.test_unknown_expectation):
2000         (ExpectationSyntaxTests.test_skip):
2001         (ExpectationSyntaxTests.test_slow):
2002         (ExpectationSyntaxTests.test_wontfix):
2003         (ExpectationSyntaxTests.test_blank_line):
2004         (ExpectationSyntaxTests.test_warnings):
2005         (RebaseliningTest.test_no_get_rebaselining_failures):
2006
2007 2012-09-21  Simon Fraser  <simon.fraser@apple.com>
2008
2009         Improve WTR unresponsiveness output a little
2010         https://bugs.webkit.org/show_bug.cgi?id=97370
2011
2012         Reviewed by Timothy Horton.
2013
2014         Distinguish between conditions that already set the errorMessage,
2015         and unresponsiveness due to slow about:blank loads when WTR
2016         reports unresponsiveness.
2017
2018         * WebKitTestRunner/TestInvocation.cpp:
2019         (WTR::TestInvocation::invoke):
2020
2021 2012-09-21  Chris Rogers  <crogers@google.com>
2022
2023         Add Web Audio support for deprecated/legacy APIs
2024         https://bugs.webkit.org/show_bug.cgi?id=97050
2025
2026         Reviewed by Eric Carlson.
2027
2028         * Scripts/webkitperl/FeatureList.pm:
2029
2030 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
2031
2032         webkitpy: update remaining tests to use the new expectation syntax
2033         https://bugs.webkit.org/show_bug.cgi?id=97362
2034
2035         Reviewed by Ojan Vafai.
2036
2037         This patch updates all the unit tests that were still using the
2038         old TestExpectations syntax to use the new syntax *except* for
2039         the tests that were specifically testing that we parsed the old
2040         syntax correctly.
2041
2042         Also, a block of tests for the new syntax were duplicated, so
2043         I've deleted the duplicate.
2044
2045         Note that the old syntax is still supported so this change should
2046         produce no visible changes.
2047
2048         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2049         (MiscTests.test_multiple_results):
2050         (MiscTests.test_category_expectations):
2051         (MiscTests.test_error_on_different_platform):
2052         (MiscTests.test_error_on_different_build_type):
2053         (MiscTests.test_overrides):
2054         (MiscTests.test_overrides__directory):
2055         (MiscTests.test_overrides__duplicate):
2056         (MiscTests.test_more_specific_override_resets_skip):
2057         (SkippedTests.check):
2058         (SkippedTests.test_duplicate_skipped_test_fails_lint):
2059         (SkippedTests.test_skipped_file_overrides_expectations):
2060         (SkippedTests.test_skipped_dir_overrides_expectations):
2061         (SkippedTests.test_skipped_file_overrides_overrides):
2062         (SkippedTests.test_skipped_dir_overrides_overrides):
2063         (ExpectationSyntaxTests.disabled_test_missing_expectation):
2064         (ExpectationSyntaxTests.disabled_test_missing_colon):
2065         (ExpectationSyntaxTests.disabled_test_too_many_colons):
2066         (ExpectationSyntaxTests.disabled_test_too_many_equals_signs):
2067         (ExpectationSyntaxTests):
2068         (ExpectationSyntaxTests.test_unrecognized_expectation):
2069         (ExpectationSyntaxTests.test_macro):
2070         (SemanticTests.test_bug_format):
2071         (SemanticTests.test_bad_bugid):
2072         (SemanticTests.test_missing_bugid):
2073         (SemanticTests.test_slow_and_timeout):
2074         (SemanticTests.test_rebaseline):
2075         (test_missing_file):
2076         (test_ambiguous):
2077         (test_more_modifiers):
2078         (test_order_in_file):
2079         (test_macro_overrides):
2080         (OldExpectationParserTests):
2081         (OldExpectationParserTests._tokenize):
2082         (OldExpectationParserTests.test_tokenize_extra_colon):
2083         (OldExpectationParserTests.test_tokenize_missing_equal):
2084         (OldExpectationParserTests.test_tokenize_extra_equal):
2085         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2086         (MainTest.test_skip_failing_tests):
2087         (MainTest.test_additional_expectations):
2088         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2089         (TestExpectationsTestCase.test_valid_expectations):
2090         (TestExpectationsTestCase.test_invalid_expectations):
2091         (TestExpectationsTestCase.test_tab):
2092
2093 2012-09-21  Benjamin Poulain  <bpoulain@apple.com>
2094
2095         [WK2] Add basic testing support for Geolocation
2096         https://bugs.webkit.org/show_bug.cgi?id=97278
2097
2098         Reviewed by Kenneth Rohde Christiansen.
2099
2100         Update the Geolocation testing to use the proper API in the UIProcess.
2101
2102         * WebKitTestRunner/CMakeLists.txt:
2103         * WebKitTestRunner/GNUmakefile.am:
2104
2105         * WebKitTestRunner/GeolocationProviderMock.cpp: Added.
2106         (WTR::startUpdatingCallback):
2107         (WTR::stopUpdatingCallback):
2108         (WTR::GeolocationProviderMock::GeolocationProvierMock):
2109         (WTR::GeolocationProviderMock::setMockGeolocationPosition):
2110         (WTR::GeolocationProviderMock::startUpdating):
2111         (WTR::GeolocationProviderMock::stopUpdating):
2112         (GeolocationProviderMock):
2113         The GeolocationProvider store the location update and deliver them as needed.
2114
2115         WebCore GeolocationController do not support asynchronous update on start/stop. This is not
2116         a problem in this case because all the messages between the WebProcess and the UIProcess are
2117         asynchronous. Because of this, unlike GeolocationClientMock, we do not use a timer for event
2118         delivery.
2119
2120         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
2121         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
2122         (WTR::InjectedBundle::setGeolocationPermission):
2123         (WTR::InjectedBundle::setMockGeolocationPosition):
2124         * WebKitTestRunner/InjectedBundle/InjectedBundle.h:
2125         (InjectedBundle):
2126         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
2127         (WTR::TestRunner::setGeolocationPermission):
2128         (WTR::TestRunner::setMockGeolocationPosition):
2129         From the InjectedBundle, we now pass the information to the UIProcess so that
2130         GeolocationProvider and the TestController can respond appropriately.
2131
2132         * WebKitTestRunner/InjectedBundle/TestRunner.h:
2133         (TestRunner):
2134         * WebKitTestRunner/Target.pri:
2135         * WebKitTestRunner/TestController.cpp:
2136         (WTR::TestController::TestController):
2137         (WTR::decidePolicyForGeolocationPermissionRequest):
2138         (WTR::TestController::createOtherPage):
2139         (WTR::TestController::initialize):
2140         (WTR::TestController::setMockGeolocationPosition):
2141         * WebKitTestRunner/TestController.h:
2142         (TestController):
2143         (WTR::TestController::setGeolocationPermission):
2144         (WTR::TestController::isGeolocationPermissionAllowed):
2145         * WebKitTestRunner/TestInvocation.cpp:
2146         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
2147         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
2148         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
2149
2150 2012-09-21  Joanmarie Diggs  <jdiggs@igalia.com>
2151
2152         REGRESSION (r127882): accessibility/spinbutton-value.html failing on GTK
2153         https://bugs.webkit.org/show_bug.cgi?id=96196
2154
2155         Reviewed by Martin Robinson.
2156
2157         The "regression" is that a new test was added but the support was missing
2158         in the Gtk port for spin buttons.
2159
2160         * DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
2161         (AccessibilityUIElement::valueDescription): Updated the FIXME comment to
2162         indicate that this cannot be implemented until it is implemented in ATK.
2163         URL of the newly-filed ATK bug included for reference.
2164
2165 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
2166
2167         Fix test_skip_and_wontfix failure
2168         https://bugs.webkit.org/show_bug.cgi?id=97225
2169
2170         Unreviewed, build fix.
2171
2172         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2173         (SemanticTests.test_skip_and_wontfix):
2174
2175 2012-09-21  Carlos Garcia Campos  <cgarcia@igalia.com>
2176
2177         [GTK] Implement ViewState methods in PageClientImpl in WebKit2
2178         https://bugs.webkit.org/show_bug.cgi?id=97202
2179
2180         Reviewed by Martin Robinson.
2181
2182         * WebKitTestRunner/gtk/PlatformWebViewGtk.cpp:
2183         (WTR::PlatformWebView::focus): Focus the view.
2184
2185 2012-09-21  Sudarsana Nagineni  <sudarsana.nagineni@intel.com>
2186
2187         [WTR] Memory leaks in InjectedBundleController::initialize()
2188         https://bugs.webkit.org/show_bug.cgi?id=97329
2189
2190         Reviewed by Alexey Proskuryakov.
2191
2192         Fix leaks in InjectedBundleController::initialize() by adopting
2193         strings created with WKStringCreateWithUTF8CString().
2194
2195         * TestWebKitAPI/InjectedBundleController.cpp:
2196         (TestWebKitAPI::InjectedBundleController::initialize):
2197
2198 2012-09-21  Marcelo Lira  <marcelo.lira@openbossa.org>
2199
2200         [Qt][WK2] Removed duplicated code from EventSenderProxy::keyDown
2201         https://bugs.webkit.org/show_bug.cgi?id=97235
2202
2203         Reviewed by Luiz Agostini.
2204
2205         * WebKitTestRunner/qt/EventSenderProxyQt.cpp:
2206         (WTR::EventSenderProxy::keyDown):
2207
2208 2012-09-21  Gergely Kis  <gergely@homejinni.com>
2209
2210         Add MIPS build slave to build.webkit.org
2211         https://bugs.webkit.org/show_bug.cgi?id=96713
2212
2213         Reviewed by Csaba Osztrogonác.
2214
2215         Added a build slave for MIPS, and enabled a builder for 
2216         Qt Linux MIPS32R2 little-endian release build.
2217
2218         * BuildSlaveSupport/build.webkit.org-config/config.json:
2219
2220 2012-09-21  Peter Beverloo  <peter@chromium.org>
2221
2222         [Chromium] Switch back to a fixed fifo path for Android
2223         https://bugs.webkit.org/show_bug.cgi?id=97230
2224
2225         Reviewed by Tony Chang.
2226
2227         Because not all external storage cards will be formated using a file
2228         system that supports named pipes, Chromium has been changed to creating
2229         the pipes in a temporary folder on the internal storage. Adapt this in
2230         WebKit so we can continue to run layout tests.
2231
2232         This also includes an *unreviewed* fix for a breakage in the webkitpy
2233         tests I made in r129221. Two lines and related to this code, so I decided
2234         to include it in this change.
2235
2236         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
2237         (ChromiumAndroidDriver.__init__):
2238         (ChromiumAndroidDriver._setup_test):
2239         (ChromiumAndroidDriver._update_version):
2240
2241 2012-09-20  Peter Beverloo  <peter@chromium.org>
2242
2243         Leverage Chromium's code to set up FIFOs for Chromium Android layout tests
2244         https://bugs.webkit.org/show_bug.cgi?id=97227
2245
2246         Reviewed by Tony Chang.
2247
2248         We switched Chromium to using FIFOs in order to achieve better consistency,
2249         which was done by Marcus in r157541. Remove all custom WebKit code in
2250         favor of Chromium's implementation.
2251
2252         Remove more FIFO code in the test runner itself, including the code in
2253         DumpRenderTree that invoked it. We can now switch to Chromium's brand
2254         new FIFO-creating code, which is being set-up for all test targets build
2255         for Android, including DumpRenderTree, TestWebKitAPI and webkit_unit_tests.
2256
2257         This also changes the ChromiumAndroidDriver._remove_all_pipes method to
2258         delete the files individually. "rm" would fail if one of the earlier files
2259         does not exist, and the "-f" argument doesn't seem to be reliable.
2260
2261         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
2262         * DumpRenderTree/chromium/TestShellAndroid.cpp:
2263         (platformInit):
2264         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
2265         (ChromiumAndroidDriver.__init__):
2266         (ChromiumAndroidDriver._setup_test):
2267         (ChromiumAndroidDriver._get_external_storage):
2268         (ChromiumAndroidDriver._drt_cmd_line):
2269         (ChromiumAndroidDriver._remove_all_pipes):
2270         (ChromiumAndroidDriver.stop):
2271         * TestWebKitAPI/TestWebKitAPI.gyp/TestWebKitAPI.gyp:
2272
2273 2012-09-21  Christophe Dumez  <christophe.dumez@intel.com>
2274
2275         [WK2][WKTR] EventSender needs to implement scheduleAsynchronousClick
2276         https://bugs.webkit.org/show_bug.cgi?id=97326
2277
2278         Reviewed by Kenneth Rohde Christiansen.
2279
2280         Implement scheduleAsynchronousClick() in WebKitTestRunner's
2281         EventSender by sending a "MouseDown" and a "MouseUp" message
2282         asynchronously to the WebProcess.
2283
2284         * WebKitTestRunner/InjectedBundle/Bindings/EventSendingController.idl:
2285         * WebKitTestRunner/InjectedBundle/EventSendingController.cpp:
2286         (WTR::createMouseMessageBody):
2287         (WTR):
2288         (WTR::EventSendingController::mouseDown):
2289         (WTR::EventSendingController::mouseUp):
2290         (WTR::EventSendingController::scheduleAsynchronousClick):
2291         * WebKitTestRunner/InjectedBundle/EventSendingController.h:
2292         (EventSendingController):
2293         * WebKitTestRunner/TestController.cpp:
2294         (WTR::TestController::didReceiveMessageFromInjectedBundle):
2295
2296 2012-09-21  Simon Hausmann  <simon.hausmann@digia.com>
2297
2298         [Qt] Error out early if we don't have ICU available
2299
2300         Reviewed by Tor Arne Vestbø.
2301
2302         * qmake/mkspecs/features/configure.prf:
2303
2304 2012-09-21  Simon Hausmann  <simon.hausmann@digia.com>
2305
2306         [Qt] Bail out when trying to build WebKit with Qt != 5
2307
2308         Reviewed by Tor Arne Vestbø.
2309
2310         Moved check for Qt version out of default_pre into top-level WebKit.pro,
2311         because we never reach default_pre.prf due to the lack of .qmake.conf support
2312         in older versions of Qt/QMake.
2313
2314         * qmake/mkspecs/features/default_pre.prf:
2315
2316 2012-09-21  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
2317
2318         WebKitTestRunner needs to print frame load delegate information
2319         https://bugs.webkit.org/show_bug.cgi?id=42705
2320
2321         Reviewed by Kenneth Rohde Christiansen.
2322
2323         Added missing dumping from WTR::InjectedBundlePage::didFailLoadWithErrorForFrame.
2324
2325         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2326         (WTR::InjectedBundlePage::didFailLoadWithErrorForFrame):
2327
2328 2012-09-21  Christophe Dumez  <christophe.dumez@intel.com>
2329
2330         [EFL] EventSender should mimic CTRL+o emacs shortcut
2331         https://bugs.webkit.org/show_bug.cgi?id=97224
2332
2333         Reviewed by Kenneth Rohde Christiansen.
2334
2335         Add support for mimicking CTRL+o emacs shortcut in EFL's
2336         EventSender in DumpRenderTree and WebKitTestRunner.
2337
2338         * DumpRenderTree/efl/EventSender.cpp:
2339         (sendKeyDown):
2340         * WebKitTestRunner/efl/EventSenderProxyEfl.cpp:
2341         (WTR::EventSenderProxy::keyDown):
2342
2343 2012-09-21  Simon Hausmann  <simon.hausmann@digia.com>
2344
2345         [Qt] Re-fix clean builds
2346
2347         Reviewed by Tor Arne Vestbø.
2348
2349         Re-introduce the sanitization for LIBS when creating a module to use
2350         LIBS_PRIVATE and otherwise do _not_ use LIBS_PRIVATE. We decided to
2351         continue to use QT, LIBS and PKGCONFIG instead of their _PRIVATE variants
2352         throughout the code base, so just using LIBS_PRIVATE in linkAgainstLibrary()
2353         causes build issues when depending system libraries end up in LIBS before
2354         LIBS_PRIVATE.
2355
2356         * qmake/mkspecs/features/default_post.prf:
2357         * qmake/mkspecs/features/functions.prf:
2358
2359 2012-09-20  Simon Fraser  <simon.fraser@apple.com>
2360
2361         Comment out a failing webkitpy unit test until Dirk can fix it.
2362
2363         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2364         (SemanticTests.test_skip_and_wontfix):
2365
2366 2012-09-20  Dirk Pranke  <dpranke@chromium.org>
2367
2368         make Skip, WontFix be the only expectations on a line
2369         https://bugs.webkit.org/show_bug.cgi?id=97225
2370
2371         Reviewed by Ojan Vafai.
2372
2373         It is now incorrect in the new syntax to have a line like:
2374
2375             foo.html [ WontFix Crash ]
2376
2377         This will generate a lint warning and be treated as an invalid
2378         line. Fixing this caused a whole bunch of unit tests to need updating
2379         to no longer be marked as WontFix :). Also, this patch adjusts
2380         the warnings so that missing Bug() identifiers will cause lint
2381         warnings but will *not* cause the line to be treated as invalid.
2382         Fixing these issues also revealed that test_hung_thread was no
2383         longer testing the right logic, so I adjusted the timeouts in
2384         test.py to make that test work again.
2385
2386         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner.py:
2387         (Worker._run_test_in_another_thread):
2388         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2389         (ResultSummaryTest.test_summarized_results_wontfix):
2390         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2391         (TestExpectationParser._parse_modifiers):
2392         (TestExpectationParser._tokenize_line_using_new_format):
2393         (TestExpectationLine.is_invalid):
2394         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2395         (BasicTests.test_basic):
2396         (test_get_test_set):
2397         (test_parse_warning):
2398         (test_pixel_tests_flag):
2399         (SemanticTests.test_missing_bugid):
2400         (SemanticTests):
2401         (SemanticTests.test_skip_and_wontfix):
2402         * Scripts/webkitpy/layout_tests/port/test.py:
2403         (TestDriver.run_test):
2404         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2405         (MainTest.test_hung_thread):
2406         * Scripts/webkitpy/tool/commands/queries_unittest.py:
2407         (PrintExpectationsTest.test_basic):
2408         (PrintExpectationsTest.test_multiple):
2409         (PrintExpectationsTest.test_full):
2410         (PrintExpectationsTest.test_exclude):
2411         (PrintExpectationsTest.test_csv):
2412
2413 2012-09-20  Lucas Forschler  <lforschler@apple.com>
2414
2415         Unreviewed.  Start running tests on the mac-ews.
2416         
2417         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
2418         (MacEWS):
2419
2420 2012-09-20  Tony Chang  <tony@chromium.org>
2421
2422         [chromium] Enable more clang warnings
2423         https://bugs.webkit.org/show_bug.cgi?id=97257
2424
2425         Reviewed by James Robinson.
2426
2427         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp: Add -Wunused-parameter. Don't add -Wglobal-constructors because
2428         there are lots of violations in these files.
2429
2430 2012-09-19  Ryosuke Niwa  <rniwa@webkit.org>
2431
2432         run-perf-tests should record individual value instead of statistics
2433         https://bugs.webkit.org/show_bug.cgi?id=97155
2434
2435         Reviewed by Hajime Morita.
2436
2437         Parse the list of individual value reported by tests and include them as "values".
2438         We strip "values" from the output JSON when uploading it to the perf-o-matic
2439         since it doesn't know how to parse "values" or ignore it.
2440
2441         * Scripts/webkitpy/performance_tests/perftest.py:
2442         (PerfTest):
2443         (PerfTest.parse_output): Parse and report "values".
2444         (PageLoadingPerfTest.run): Report indivisual page loading time in "values".
2445         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
2446         (MainTest.test_parse_output):
2447         (MainTest.test_parse_output_with_failing_line):
2448         (TestPageLoadingPerfTest.test_run):
2449         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
2450         (PerfTestsRunner._generate_and_show_results): Strip "values" from each result
2451         until we update perf-o-matic.
2452         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
2453         (test_run_memory_test):
2454         (test_run_with_json_output):
2455         (test_run_with_description):
2456         (test_run_with_slave_config_json):
2457         (test_run_with_multiple_repositories):
2458
2459 2012-09-20  Simon Hausmann  <simon.hausmann@digia.com>
2460
2461         [Qt] Fix production builds
2462
2463         Unreviewed trivial fix: Follow up r129119 and avoid clobbering LIBS through the use of LIBS_PRIVATE.
2464
2465         Otherwise the libQtWebKit.prl file for example contains -lWebKit1 -lWebCore, etc.
2466
2467         * qmake/mkspecs/features/functions.prf:
2468
2469 2012-09-20  Dirk Pranke  <dpranke@chromium.org>
2470
2471         REGRESSION: layout test results doesn't show diffs
2472         https://bugs.webkit.org/show_bug.cgi?id=97182
2473
2474         Reviewed by Ojan Vafai.
2475
2476         Go back to storing TEXT, AUDIO, and IMAGE+TEXT in results.json
2477         so that results.html (and hopefully garden-o-matic) can
2478         determine which things actually failed. However, we keep mapping
2479         these results to Failure so that we still only have a single
2480         expectation type for them.
2481
2482         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
2483         (JSONLayoutResultsGenerator):
2484         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2485         (TestExpectationParser):
2486         (TestExpectations):
2487         (TestExpectations.result_was_expected):
2488         * Scripts/webkitpy/layout_tests/models/test_failures.py:
2489         (determine_result_type):
2490         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2491         (MainTest.test_missing_and_unexpected_results):
2492         (MainTest.test_retrying_and_flaky_tests):
2493
2494 2012-09-20  Tommy Widenflycht  <tommyw@google.com>
2495
2496         MediaStream API: Extend UserMediaRequest with a ownerDocument method
2497         https://bugs.webkit.org/show_bug.cgi?id=97095
2498
2499         Reviewed by Adam Barth.
2500
2501         Extending WebUserMediaClientMock to check that the owning document is valid,
2502         and that the document has a frame.
2503
2504         * DumpRenderTree/chromium/WebUserMediaClientMock.cpp:
2505         (WebKit::WebUserMediaClientMock::requestUserMedia):
2506
2507 2012-09-20  Stephen Chenney  <schenney@chromium.org>
2508
2509         [Chromium] DRT does not support --dump-all-pixels flag
2510         https://bugs.webkit.org/show_bug.cgi?id=95098
2511
2512         Reviewed by Dirk Pranke.
2513
2514         Add support for the --pixel-tests and shorthand -p option in Chromium DumpRenderTree. Use
2515         of this flag causes pixel results to be created for all tests, regardless of
2516         individual test options. If an individual test provides a pixel hash it will be used,
2517         otherwise the hash will be empty. This replaces a previously defined but unused option
2518         --dump-all-pixels, and is useful primarily when debugging DRT instances.
2519
2520         * DumpRenderTree/chromium/DumpRenderTree.cpp:
2521         (runTest): Add a parameter and code to force pixel results for the test.
2522         (main): Add parameter handling for --pixels-test and -p, and remove --dump-all-pixels.
2523
2524 2012-09-20  Carlos Garcia Campos  <cgarcia@igalia.com>
2525
2526         [GTK] run-api-tests should not buffer test stdout
2527         https://bugs.webkit.org/show_bug.cgi?id=88474
2528
2529         Reviewed by Philippe Normand.
2530
2531         * Scripts/run-gtk-tests:
2532         (TestRunner._run_test_command): Use os.forkpty() instead of
2533         subprocess.Popen() so that gtest sends the output with colors to
2534         stdout. Use common.parse_output_lines() to parse the output and
2535         write it to stdout while it's read.
2536         (TestRunner._run_test_command.parse_line): Parse the line to get
2537         the test pid and write the line to stdout.
2538         (TestRunner._run_test_command.waitpid): Helper function to call
2539         waitpid handling EINTR.
2540         (TestRunner._run_test_command.return_code_from_exit_status):
2541         Helper function to convert exit status of test commands to a
2542         return code.
2543         * gtk/common.py:
2544         (parse_output_lines): Helper function that uses select to read
2545         the given file descriptor and call the given callback for every
2546         line read.
2547
2548 2012-09-20  Simon Hausmann  <simon.hausmann@digia.com>
2549
2550         [Qt] Fix initial build
2551
2552         Reviewed by Tor Arne Vestbø.
2553
2554         When building QtWebKit the first time there is no qt_webkit.pri module pri file, and therefore
2555         $$QT.webkit.name isn't set and so creating_module isn't set. That has all sorts of implications
2556         causing incorrect linking for Makefile.api, etc.
2557
2558         Fix the determination by simply checking if MODULE is set, which only happens in api.pri.
2559
2560         * qmake/mkspecs/features/webkit_modules.prf:
2561
2562 2012-09-20  Sheriff Bot  <webkit.review.bot@gmail.com>
2563
2564         Unreviewed, rolling out r129091.
2565         http://trac.webkit.org/changeset/129091
2566         https://bugs.webkit.org/show_bug.cgi?id=97205
2567
2568         It broke perf tests everywhere (Requested by Ossy on #webkit).
2569
2570         * Scripts/webkitpy/performance_tests/perftest.py:
2571         (PerfTest):
2572         (PerfTest.parse_output):
2573         (PageLoadingPerfTest.run):
2574         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
2575         (MainTest.test_parse_output):
2576         (MainTest.test_parse_output_with_failing_line):
2577         (TestPageLoadingPerfTest.test_run):
2578         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
2579         (PerfTestsRunner._generate_and_show_results):
2580         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
2581         (test_run_memory_test):
2582         (test_run_with_json_output):
2583         (test_run_with_description):
2584         (test_run_with_slave_config_json):
2585         (test_run_with_multiple_repositories):
2586
2587 2012-09-20  Simon Hausmann  <simon.hausmann@digia.com>
2588
2589         [Qt] QtWebKit module header includes private dependencies
2590
2591         Reviewed by Tor Arne Vestbø.
2592
2593         * qmake/mkspecs/features/default_post.prf: Don't try to sanitize LIBS, because we can make sure
2594         that LIBS_PRIVATE is set from the beginning. Moved the creating_module and PKGCONFIG/QT(_PRIVATE)
2595         sanitization into webkit_modules.prf. creating_module determination requires TARGET to be set, so
2596         we can't do it in default_pre.
2597         * qmake/mkspecs/features/webkit_modules.prf:
2598
2599 2012-09-20  Simon Hausmann  <simon.hausmann@digia.com>
2600
2601         [Qt] Fix build with MingW
2602
2603         Reviewed by Tor Arne Vestbø.
2604
2605         Don't try to compile WebCore, etc. with debug symbols for production
2606         builds, it's just too big.
2607
2608         * qmake/mkspecs/features/production_build.prf:
2609
2610 2012-09-20  Michał Pakuła vel Rutka  <m.pakula@samsung.com>
2611
2612         [EFL][DRT]DumpRenderTree needs to reset focus state when test starts.
2613         https://bugs.webkit.org/show_bug.cgi?id=97087
2614
2615         Reviewed by Gyuyoung Kim.
2616
2617         Add focusing a main frame on settings reset.
2618         After editing/undo/undo-iframe-location-change was executed a frame
2619         was left in unfocused state. This caused flakiness in two tests results.
2620
2621         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
2622         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
2623
2624 2012-09-20  Allan Sandfeld Jensen  <allan.jensen@digia.com>
2625
2626         [Qt] Add eventSender.gestureTap
2627         https://bugs.webkit.org/show_bug.cgi?id=66173
2628
2629         Reviewed by Kenneth Rohde Christiansen.
2630
2631         Implement support for sending raw QGesture events.
2632
2633         * DumpRenderTree/qt/EventSenderQt.cpp:
2634         (EventSender::EventSender):
2635         (EventSender::gestureTap):
2636         * DumpRenderTree/qt/EventSenderQt.h:
2637         (EventSender):
2638
2639 2012-09-19  Sheriff Bot  <webkit.review.bot@gmail.com>
2640
2641         Unreviewed, rolling out r129007.
2642         http://trac.webkit.org/changeset/129007
2643         https://bugs.webkit.org/show_bug.cgi?id=97172
2644
2645         It broke the build on the Qt bots (Requested by Ossy on
2646         #webkit).
2647
2648         * qmake/config.tests/gccdepends/empty.cpp:
2649         * qmake/config.tests/gccdepends/gccdepends.pro:
2650         * qmake/mkspecs/features/default_pre.prf:
2651
2652 2012-09-19  Ryosuke Niwa  <rniwa@webkit.org>
2653
2654         run-perf-tests should record indivisual value instead of statistics
2655         https://bugs.webkit.org/show_bug.cgi?id=97155
2656
2657         Reviewed by Hajime Morita.
2658
2659         Parse the list of indivisual value reported by tests and include them as "values".
2660         We strip "values" from the output JSON when uploading it to the perf-o-matic
2661         since it doesn't know how to parse "values" or ignore it.
2662
2663         * Scripts/webkitpy/performance_tests/perftest.py:
2664         (PerfTest):
2665         (PerfTest.parse_output): Parse and report "values".
2666         (PageLoadingPerfTest.run): Report indivisual page loading time in "values".
2667         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
2668         (MainTest.test_parse_output):
2669         (MainTest.test_parse_output_with_failing_line):
2670         (TestPageLoadingPerfTest.test_run):
2671         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
2672         (PerfTestsRunner._generate_and_show_results): Strip "values" from each result
2673         until we update perf-o-matic.
2674         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
2675         (test_run_memory_test):
2676         (test_run_with_json_output):
2677         (test_run_with_description):
2678         (test_run_with_slave_config_json):
2679         (test_run_with_multiple_repositories):
2680
2681 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2682
2683         Fix regex groups for bug matching in flakiness dashboard.
2684         https://bugs.webkit.org/show_bug.cgi?id=97152
2685
2686         Unreviewed, build fix.
2687
2688         * TestResultServer/static-dashboards/flakiness_dashboard.js:
2689         (htmlForBugs):
2690
2691 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2692
2693         nrwt: print unexpected results using new TestExpectations syntax
2694         https://bugs.webkit.org/show_bug.cgi?id=97159
2695
2696         Unreviewed, build fix.
2697
2698         Change new-run-webkit-tests to print out failures using the new
2699         syntax when there are unexpected results, e.g.:
2700
2701         52 tests ran as expected, 19 didn't:
2702
2703
2704         Regressions: Unexpected failures : (2)
2705           failures/flaky/text.html [ Failure ]
2706           failures/unexpected/text-image-checksum.html [ Failure ]
2707
2708         and so forth
2709
2710         * Scripts/webkitpy/layout_tests/views/printing.py:
2711         (Printer._print_unexpected_results):
2712
2713 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2714
2715         update flakiness dashboard after cutover to new test expectations syntax
2716         https://bugs.webkit.org/show_bug.cgi?id=97152
2717
2718         Unreviewed, build fix.
2719
2720         Handle (??) the new Bug notations as well. Hopefully we don't
2721         still need the old ones.
2722
2723         * TestResultServer/static-dashboards/flakiness_dashboard.js:
2724         (filterBugs):
2725         (htmlForBugs):
2726         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
2727
2728 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2729
2730         update flakiness dashboard after cutover to new test expectations syntax
2731         https://bugs.webkit.org/show_bug.cgi?id=97152
2732
2733         Reviewed by Ryosuke Niwa.
2734
2735         This change clones the TestExpectation parsing state machine
2736         from python into javascript.
2737
2738         * TestResultServer/static-dashboards/flakiness_dashboard.js:
2739         (parsedExpectations.lines.forEach):
2740         (parsedExpectations):
2741         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
2742
2743 2012-09-19  Dana Jansens  <danakj@chromium.org>
2744
2745         Add backer@chromium.org as contributor
2746         https://bugs.webkit.org/show_bug.cgi?id=97150
2747
2748         * Scripts/webkitpy/common/config/committers.py:
2749
2750 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2751
2752         fix MISSING after TestExpectations conversion
2753         https://bugs.webkit.org/show_bug.cgi?id=97148
2754
2755         Unreviewed, expectations change / build fix.
2756
2757         * Scripts/convert-test-expectations:
2758         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2759         (TestExpectationParser):
2760
2761 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2762
2763         nrwt: convert chromium TestExpectations to the new syntax
2764         https://bugs.webkit.org/show_bug.cgi?id=97139
2765
2766         Reviewed by Ryosuke Niwa.
2767     
2768         Add in a file temporarily that does the conversion of
2769         test expectations formats
2770
2771         * Tools/Scripts/convert-test-expectations: Added.
2772
2773 2012-09-19  Ryosuke Niwa  <rniwa@webkit.org>
2774
2775         REGRESSION: run-perf-tests no longer reports the total test time
2776         https://bugs.webkit.org/show_bug.cgi?id=97138
2777
2778         Reviewed by Tony Chang.
2779
2780         Report the finished time as a info-level log as opposed to a debug level log.
2781
2782         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
2783         (PerfTestsRunner._run_single_test):
2784         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
2785         (MainTest.normalizeFinishedTime): Added. It replaces all finished times by 0.1 seconds.
2786         (test_run_test_pause_before_testing):
2787         (test_run_test_set_for_parser_tests):
2788         (test_run_memory_test):
2789         (_test_run_with_json_output):
2790
2791 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2792
2793         Support new TestExpectations format alongside old one
2794         https://bugs.webkit.org/show_bug.cgi?id=96588
2795
2796         Reviewed by Ojan Vafai.
2797
2798         This patch adds support for actually reading in lines formatted
2799         in the new style, and when we re-serialize/write out the file,
2800         all lines will be written in the new style. Note that reading in
2801         the old style is still supported, and no updates are being made
2802         to the actual TestExpectations files as part of this change.
2803
2804         This change updates most but not all of the unit tests to use
2805         the new syntax. I will update the rest when (or before where
2806         possible) I drop support for the old format.
2807
2808         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2809         (TestExpectationParser):
2810         (TestExpectationParser._tokenize_line):
2811         (TestExpectationParser._tokenize_line_using_new_format):
2812         (TestExpectationLine.to_string):
2813         (TestExpectationLine._format_line):
2814         (TestExpectations):
2815         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2816         (test_parse_warning):
2817         (SkippedTests.test_skipped_entry_dont_exist):
2818         (NewExpectationSyntaxTests.test_warnings):
2819         (RemoveConfigurationsTest.test_remove):
2820         (test_remove_line):
2821         (RebaseliningTest.test_remove):
2822         (RebaseliningTest.test_no_get_rebaselining_failures):
2823         (TestExpectationParserTests.test_tokenize_blank):
2824         (TestExpectationParserTests.test_tokenize_extra_colon):
2825         (TestExpectationParserTests.test_tokenize_missing_equal):
2826         (TestExpectationParserTests.test_tokenize_extra_equal):
2827         (TestExpectationSerializationTests.test_unparsed_to_string):
2828         (TestExpectationSerializationTests.test_unparsed_list_to_string):
2829         (TestExpectationSerializationTests.test_parsed_to_string):
2830         (TestExpectationSerializationTests.test_format_line):
2831         (TestExpectationSerializationTests.test_string_roundtrip):
2832         (TestExpectationSerializationTests.test_list_roundtrip):
2833         (TestExpectationSerializationTests.test_reconstitute_only_these):
2834         (TestExpectationSerializationTests.test_string_whitespace_stripping):
2835         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2836         (LintTest.test_lint_test_files__errors):
2837         * Scripts/webkitpy/tool/commands/queries_unittest.py:
2838         (PrintExpectationsTest.test_basic):
2839         (PrintExpectationsTest.test_multiple):
2840         (PrintExpectationsTest.test_full):
2841         (PrintExpectationsTest.test_exclude):
2842         * Scripts/webkitpy/tool/commands/rebaseline.py:
2843         (RebaselineExpectations.execute):
2844         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
2845         (TestRebaseline.test_rebaseline_updates_expectations_file_noop):
2846         (test_rebaseline_updates_expectations_file):
2847         (test_rebaseline_does_not_include_overrides):
2848         (test_rebaseline_expectations_noop):
2849         (test_overrides_are_included_correctly):
2850
2851 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2852
2853         implement first part of support for the new TestExpectations syntax
2854         https://bugs.webkit.org/show_bug.cgi?id=96569
2855
2856         Reviewed by Ryosuke Niwa.
2857
2858         This patch implements support for parsing a line of the new
2859         format for the TestExpectations file and converting it back into
2860         the old format for compatibility. This routine is not yet used
2861         by anything.
2862
2863         The new format is documented at:
2864           http://trac.webkit.org/wiki/TestExpectations
2865
2866         but, in short:
2867
2868           [bugs] [ "[" modifiers "]" ] test_name [ "[" expectations "]" ]
2869
2870         - Comments are indicated with "#" instead of "//"
2871         - If no expectations are specified we default to Skip for
2872           compatibility with the Skipped files (these two changes make
2873           Skipped files a subset of TestExpectations files)
2874
2875         - All of the tokens are now CamelCase instead of ALLCAPS.
2876         - FAIL -> Failure
2877         - IMAGE -> ImageOnlyFailure
2878         - WONTFIX -> WontFix
2879         - modifiers refer to just the platforms and configurations
2880           (release/debug) that the line applies to.
2881         - WontFix, Rebaseline, Slow, and Skip move to the right-hand side as
2882           expectations
2883         - expectations will typically be written out in lexicographic order
2884         - We use webkit.org/b/12345, crbug.com/12345, and Bug(dpranke)
2885           instead of BUGWK12345, BUGCR12345, and BUGDPRANKE.
2886
2887         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2888         (TestExpectationParser):
2889         (TestExpectationParser._tokenize_line_using_new_format):
2890         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2891         (NewExpectationSyntaxTests):
2892         (NewExpectationSyntaxTests.assert_exp):
2893         (NewExpectationSyntaxTests.test_bare_name):
2894         (NewExpectationSyntaxTests.test_bare_name_and_bugs):
2895         (NewExpectationSyntaxTests.test_comments):
2896         (NewExpectationSyntaxTests.test_config_modifiers):
2897         (NewExpectationSyntaxTests.test_unknown_config):
2898         (NewExpectationSyntaxTests.test_unknown_expectation):
2899         (NewExpectationSyntaxTests.test_skip):
2900         (NewExpectationSyntaxTests.test_slow):
2901         (NewExpectationSyntaxTests.test_wontfix):
2902         (NewExpectationSyntaxTests.test_blank_line):
2903         (NewExpectationSyntaxTests.test_warnings):
2904
2905 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2906
2907         nrwt: replace TEXT, AUDIO, and IMAGE+TEXT with FAIL
2908         https://bugs.webkit.org/show_bug.cgi?id=96845
2909
2910         Reviewed by Ojan Vafai.
2911
2912         In preparation for the new TestExpectations syntax, we replace
2913         all TEXT, IMAGE+TEXT, and AUDIO failures with FAIL. This will
2914         make switching to the new syntax lossless (i.e., we lose
2915         information now, not then).
2916
2917         Note that we can still parse in results.json files that have the
2918         old data for backwards compatibility.
2919         
2920
2921         * Scripts/webkitpy/common/net/resultsjsonparser.py:
2922         (JSONTestResult._failure_types_from_actual_result):
2923         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
2924         (ResultsJSONParserTest):
2925         (test_basic):
2926         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2927         (ResultSummaryTest.test_summarized_results_wontfix):
2928         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
2929         (JSONLayoutResultsGenerator):
2930         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2931         (TestExpectations):
2932         (TestExpectations.remove_pixel_failures):
2933         (TestExpectations.has_pixel_failures):
2934         (TestExpectations.suffixes_for_expectations):
2935         (TestExpectations.get_rebaselining_failures):
2936         (TestExpectations.remove_configuration_from_test):
2937         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2938         (assert_bad_expectations):
2939         (BasicTests):
2940         (BasicTests.test_basic):
2941         (MiscTests):
2942         (MiscTests.test_multiple_results):
2943         (MiscTests.test_result_was_expected):
2944         (MiscTests.test_remove_pixel_failures):
2945         (MiscTests.test_suffixes_for_expectations):
2946         (test_get_expectations_string):
2947         (test_parse_warning):
2948         (test_error_on_different_platform):
2949         (test_error_on_different_build_type):
2950         (test_overrides):
2951         (test_overrides__directory):
2952         (test_overrides__duplicate):
2953         (test_pixel_tests_flag):
2954         (test_more_specific_override_resets_skip):
2955         (SkippedTests.test_skipped_file_overrides_expectations):
2956         (SkippedTests.test_skipped_dir_overrides_expectations):
2957         (SkippedTests.test_skipped_file_overrides_overrides):
2958         (SkippedTests.test_skipped_dir_overrides_overrides):
2959         (ExpectationSyntaxTests.test_missing_colon):
2960         (ExpectationSyntaxTests.test_too_many_equals_signs):
2961         (SemanticTests.test_bug_format):
2962         (SemanticTests.test_bad_bugid):
2963         (SemanticTests.test_missing_bugid):
2964         (SemanticTests.test_rebaseline):
2965         (test_missing_file):
2966         (test_more_modifiers):
2967         (test_order_in_file):
2968         (test_macro_overrides):
2969         (RemoveConfigurationsTest.test_remove):
2970         (test_remove_line):
2971         (RebaseliningTest.test_remove):
2972         (TestExpectationSerializationTests.test_serialize_parsed_expectations):
2973         * Scripts/webkitpy/layout_tests/models/test_failures.py:
2974         (determine_result_type):
2975         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
2976         (test_test_expectations):
2977         * Scripts/webkitpy/layout_tests/port/test.py:
2978         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2979         (MainTest.test_missing_and_unexpected_results):
2980         (MainTest.test_retrying_and_flaky_tests):
2981         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2982         (TestExpectationsTestCase.test_valid_expectations):
2983         * Scripts/webkitpy/tool/commands/queries_unittest.py:
2984         (PrintExpectationsTest.test_basic):
2985         (PrintExpectationsTest.test_multiple):
2986         (PrintExpectationsTest.test_full):
2987         (PrintExpectationsTest.test_exclude):
2988         (PrintExpectationsTest.test_csv):
2989         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
2990         (test_rebaseline_does_not_include_overrides):
2991         (test_overrides_are_included_correctly):
2992
2993 2012-09-19  Dominic Mazzoni  <dmazzoni@google.com>
2994
2995         AX: A few control types are returning the wrong answer for isReadOnly
2996         https://bugs.webkit.org/show_bug.cgi?id=96735
2997
2998         Reviewed by Chris Fleizach.
2999
3000         Exposing isReadOnly in an AccessibilityObject to DumpRenderTree.
3001
3002         * DumpRenderTree/chromium/TestRunner/AccessibilityUIElementChromium.cpp:
3003         (AccessibilityUIElement::AccessibilityUIElement):
3004         (AccessibilityUIElement::isReadOnlyGetterCallback):
3005         * DumpRenderTree/chromium/TestRunner/AccessibilityUIElementChromium.h:
3006         (AccessibilityUIElement):
3007
3008 2012-09-19  Sudarsana Nagineni  <sudarsana.nagineni@intel.com>
3009
3010         [WTR] Memory leaks in TestRunner::deliverWebIntent()
3011         https://bugs.webkit.org/show_bug.cgi?id=97111
3012
3013         Reviewed by Kenneth Rohde Christiansen.
3014
3015         Fix memory leaks in deliverWebIntent() by adopting strings
3016         created with WKStringCreateWithUTF8CString().
3017
3018         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
3019         (WTR::TestRunner::deliverWebIntent):
3020
3021 2012-09-19  Simon Hausmann  <simon.hausmann@digia.com>
3022
3023         [Qt] Fix incremental builds with all-in-one-files and gccdepends
3024
3025         Reviewed by Tor Arne Vestbø.
3026
3027         Pass -MP to gcc when we use the gcc depends feature, to ensure that implicit rules
3028         are not only created for header files but also for .cpp files. AllInOne.cpp files
3029         include other .cpp files, and when those are removed we need those dummy rules to
3030         avoid a "No rule to make Foo.cpp required by AllInOne.o" error.
3031
3032         * qmake/config.tests/gccdepends/empty.cpp:
3033         (main):
3034         * qmake/config.tests/gccdepends/gccdepends.pro:
3035         * qmake/mkspecs/features/default_pre.prf:
3036
3037 2012-09-19  Rick Byers  <rbyers@chromium.org>
3038
3039         Do touch adjustment on GestureTapDown
3040         https://bugs.webkit.org/show_bug.cgi?id=96677
3041
3042         Reviewed by Antonio Gomes.
3043
3044         Allow radius to be set for GestureTapDown events.
3045         * DumpRenderTree/chromium/TestRunner/EventSender.cpp:
3046         (EventSender::gestureEvent):
3047
3048 2012-09-19  Christophe Dumez  <christophe.dumez@intel.com>
3049
3050         [EFL] EFL's DRT does not support overriding 'WebKitCSSRegionsEnabled' preference
3051         https://bugs.webkit.org/show_bug.cgi?id=97100
3052
3053         Reviewed by Gyuyoung Kim.
3054
3055         EFL's DRT now supports overriding the 'WebKitCSSRegionsEnabled'
3056         preference, in order to match WebKitTestRunner functionality.
3057
3058         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
3059         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
3060         * DumpRenderTree/efl/TestRunnerEfl.cpp:
3061         (TestRunner::overridePreference):
3062
3063 2012-09-18  Kangil Han  <kangil.han@samsung.com>
3064
3065         [WK2][WTR] CodeGeneratorTestRunner could keep original copyright.
3066         https://bugs.webkit.org/show_bug.cgi?id=96181
3067
3068         Reviewed by Daniel Bates.
3069
3070         This patch enabled derived files, in DerivedSources/InjectedBundle, to keep original copyright.
3071
3072         * WebKitTestRunner/InjectedBundle/Bindings/CodeGeneratorTestRunner.pm:
3073         (new):
3074         (_parseLicenseBlock):
3075         (_parseLicenseBlockFromFile):
3076         (_defaultLicenseBlock):
3077         (_licenseBlock):
3078         (_generateHeaderFile):
3079         (_generateImplementationFile):
3080
3081 2012-09-18  Byungwoo Lee  <bw80.lee@samsung.com>
3082
3083         Title string should be changed when document.title is set to ''.
3084         https://bugs.webkit.org/show_bug.cgi?id=96793
3085
3086         Reviewed by Kenneth Rohde Christiansen.
3087
3088         Change dump format of dumpTitleChanges more understandable.
3089         Uses single quotation marks for the title string.
3090
3091         * DumpRenderTree/blackberry/DumpRenderTree.cpp:
3092         (BlackBerry::WebKit::DumpRenderTree::didReceiveTitleForFrame):
3093         * DumpRenderTree/chromium/WebViewHost.cpp:
3094         (WebViewHost::didReceiveTitle):
3095         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
3096         (DumpRenderTreeChrome::onFrameTitleChanged):
3097         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3098         (webViewTitleChanged):
3099         * DumpRenderTree/mac/FrameLoadDelegate.mm:
3100         (-[FrameLoadDelegate webView:didReceiveTitle:forFrame:]):
3101         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
3102         (WebCore::DumpRenderTree::titleChanged):
3103         * DumpRenderTree/win/FrameLoadDelegate.cpp:
3104         (FrameLoadDelegate::didReceiveTitle):
3105         * DumpRenderTree/wx/DumpRenderTreeWx.cpp:
3106         (LayoutWebViewEventHandler::OnReceivedTitleEvent):
3107         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3108         (WTR::InjectedBundlePage::didReceiveTitleForFrame):
3109
3110 2012-09-18  Szilard Ledan  <szledan@inf.u-szeged.hu>
3111
3112         EWS shouldn't sleep if there are new patches in its queue
3113         https://bugs.webkit.org/show_bug.cgi?id=83038
3114
3115         Reviewed by Eric Seidel.
3116
3117         EWS tries to process a security patch. Of course it can't, because the EWS isn't
3118         the member of the security group. But the problem is that after it can't process
3119         the attachment, it says that queue is empty (but it isn't!) and it sleeps 2 minutes
3120         and push the security patch to the end of the queue.
3121         Now it stays in the loop until it finds a patch or the queue gets empty.
3122
3123         * Scripts/webkitpy/tool/commands/queues.py:
3124         (AbstractPatchQueue._next_patch):
3125         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3126         (AbstractPatchQueueTest.test_next_patch):
3127
3128 2012-09-18  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
3129
3130         [WK2][WTR] InjectedBundle::booleanForKey() should handle literals effectively
3131         https://bugs.webkit.org/show_bug.cgi?id=97014
3132
3133         Reviewed by Kenneth Rohde Christiansen.
3134
3135         According to http://trac.webkit.org/wiki/EfficientStrings WTF::StringBuilder::appendLiteral() shall
3136         be used for literals rather than WTF::StringBuilder::append().
3137
3138         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
3139         (WTR::InjectedBundle::booleanForKey):
3140
3141 2012-09-18  Simon Hausmann  <simon.hausmann@digia.com>
3142
3143         Update my e-mail address.
3144
3145         * Scripts/webkitpy/common/config/committers.py:
3146
3147 2012-09-18  Andras Becsi  <andras.becsi@digia.com>
3148
3149         Update my e-mail address.
3150
3151         Unreviewed.
3152
3153         * Scripts/webkitpy/common/config/committers.py:
3154
3155 2012-09-18  Andras Becsi  <andras.becsi@digia.com>
3156
3157         [Qt] qt_webkit.pri should not be listed in Tools.pro
3158
3159         Reviewed and rubber-stamped by Simon Hausmann and Tor Arne Vestbø.
3160
3161         Since r128751 the module pri file is auto-generated
3162         but it was still listed in OTHER_FILES.
3163
3164         * Tools.pro: Remove unneeded line.
3165
3166 2012-09-18  Simon Hausmann  <simon.hausmann@digia.com>
3167
3168         [Qt] Fix build with some versions of the gold linker
3169
3170         Reviewed by Tor Arne Vestbø.
3171
3172         Don't unconditionally pass --no-keep-memory to the linker, some versions might not support it.
3173         Instead run a compile/link test first to see if it works.
3174
3175         * qmake/config.tests/gnuld/gnuld.pro: Added.
3176         * qmake/config.tests/gnuld/main.cpp: Added.
3177         (main):
3178         * qmake/mkspecs/features/unix/default_post.prf:
3179
3180 2012-09-18  Simon Hausmann  <simon.hausmann@digia.com>
3181
3182         [Qt] Remove forced use of gold.
3183
3184         Reviewed by Tor Arne Vestbø.
3185
3186         The choice of what linker to use with WebKit should be taken by Qt's build system and ideally the same for all
3187         modules of Qt. Then in turn it is usually up to the administrator of the machine. Recent Debian based systems
3188         often offer the automatic use of gold through a symlink and a dpkg-diversion when installing the gold package.
3189
3190         * qmake/mkspecs/features/unix/default_post.prf:
3191
3192 2012-09-18  Simon Hausmann  <simon.hausmann@digia.com>
3193
3194         [Qt] Prospective Qt/Windows cross-compiling fix
3195
3196         Reviewed by Tor Arne Vestbø.
3197
3198         The win32 scope is not set when cross-compiling from Linux to Windows.
3199
3200         * qmake/mkspecs/features/functions.prf:
3201
3202 2012-09-18  Seokju Kwon  <seokju.kwon@samsung.com>
3203
3204         [EFL] Remove background view on EWebLauncher and MiniBrowser
3205         https://bugs.webkit.org/show_bug.cgi?id=96905
3206
3207         Reviewed by Gyuyoung Kim.
3208
3209         The size of webview was changed after adding url bar.
3210         And the background view is not necessary anymore, since it was used for debugging back in the day.
3211
3212         * EWebLauncher/main.c:
3213         (_ELauncher):
3214         (on_ecore_evas_resize):
3215         (browserCreate):
3216         * MiniBrowser/efl/main.c:
3217         (_MiniBrowser):
3218         (on_ecore_evas_resize):
3219         (browserCreate):
3220
3221 2012-09-18  Sergio Villar Senin  <svillar@igalia.com>
3222
3223         [GTK] run-webkit-tests unable to find TestExpectations for WK2
3224         https://bugs.webkit.org/show_bug.cgi?id=96998
3225
3226         Reviewed by Philippe Normand.
3227
3228         We should look for TestExpectations files in all the locations where
3229         we currently look for Skipped files. This will allow
3230         run-webkit-tests to look for TestExpectations files in
3231         platform/gtk-wk2 and platform/wk2 if the "-2" flag is used.
3232
3233         * Scripts/webkitpy/layout_tests/port/gtk.py:
3234         (GtkPort.expectations_files):
3235
3236 2012-09-18  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3237
3238         [Qt] Fix build without the QtQuick module
3239
3240         Reviewed by Simon Hausmann.
3241
3242         * qmake/mkspecs/features/features.prf:
3243
3244 2012-09-18  Stephanie Lewis  <slewis@apple.com>
3245
3246         Build fix after http://trac.webkit.org/projects/webkit/changeset/128852.
3247
3248         Unreviewed.
3249
3250         * DumpRenderTree/mac/DumpRenderTree.mm:
3251         (poseAsClass):
3252
3253 2012-09-18  Allan Sandfeld Jensen  <allan.jensen@nokia.com>
3254
3255         Unreviewed update of email addresses for Berlin QtWebKit office.
3256
3257         * Scripts/webkitpy/common/config/committers.py:
3258
3259 2012-09-17  Zan Dobersek  <zandobersek@gmail.com>
3260
3261         [GTK] fast/loader/display-image-unset-can-block-image-and-can-reload-in-place.html failing after r128645
3262         https://bugs.webkit.org/show_bug.cgi?id=96899
3263
3264         Reviewed by Martin Robinson.
3265
3266         When overriding the 'WebKitDisplayImageKey' preference, use the
3267         'auto-load-images' property of WebKitWebSettings as the property which
3268         should be updated with the corresponding preference value.
3269
3270         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3271         (resetDefaultsToConsistentValues):
3272         * DumpRenderTree/gtk/TestRunnerGtk.cpp:
3273         (TestRunner::overridePreference):
3274
3275 2012-09-17  Pratik Solanki  <psolanki@apple.com>
3276
3277         DumpRenderTree and WebKitTestRunner should compile with -Wundef on Mac
3278         https://bugs.webkit.org/show_bug.cgi?id=96973
3279
3280         Reviewed by Dan Bernstein.
3281
3282         * DumpRenderTree/mac/Configurations/Base.xcconfig:
3283         * WebKitTestRunner/Configurations/Base.xcconfig:
3284         * WebKitTestRunner/PlatformWebView.h: Use #ifdef __OBJC__ and not #if.
3285
3286 2012-09-17  Dirk Pranke  <dpranke@chromium.org>
3287
3288         nrwt: remove "unexpected EOF" warnings
3289         https://bugs.webkit.org/show_bug.cgi?id=96970
3290
3291         Reviewed by Ojan Vafai.
3292
3293         After debugging this a bit, it looks like there aren't any cases
3294         that I can reproduce where a read() of zero indicates something
3295         actually wrong; either it is a prelude to a crash, or a false
3296         negative. So, I'm removing these warnings and adding a comment.
3297
3298         * Scripts/webkitpy/layout_tests/port/server_process.py:
3299         (ServerProcess._wait_for_data_and_update_buffers_using_select):
3300
3301 2012-09-17  Dirk Pranke  <dpranke@chromium.org>
3302
3303         [chromium] ASAN bot is crashing at the end of the run
3304         https://bugs.webkit.org/show_bug.cgi?id=96967
3305
3306         Reviewed by Abhishek Arya.
3307
3308         The ASAN bot is crashing attempting to decode some output into
3309         UTF-8; there's no reason to do this, so let's not do this and
3310         see if something else is going on as well.
3311
3312         * Scripts/webkitpy/layout_tests/port/chromium.py:
3313         (ChromiumPort._get_crash_log):
3314
3315 2012-09-17  Dirk Pranke  <dpranke@chromium.org>
3316
3317         nrwt: --results-directory isn't getting printed properly
3318         https://bugs.webkit.org/show_bug.cgi?id=96965
3319
3320         Reviewed by Ojan Vafai.
3321
3322         options.results_directory isn't actually initialized with the
3323         default values until after we call print_config(), so this
3324         changes things to print the value directly.
3325
3326         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3327         (run):
3328         * Scripts/webkitpy/layout_tests/views/printing.py:
3329         (Printer.print_config):
3330         * Scripts/webkitpy/layout_tests/views/printing_unittest.py:
3331         (Testprinter.test_print_config):
3332
3333 2012-09-17  Rob Buis  <rbuis@rim.com>
3334
3335         [BlackBerry] Enable VIDEO_TRACK
3336         https://bugs.webkit.org/show_bug.cgi?id=96949
3337
3338         Reviewed by Antonio Gomes.
3339
3340         * Scripts/webkitperl/FeatureList.pm:
3341
3342 2012-09-17  Rick Byers  <rbyers@chromium.org>
3343
3344         Add handling of new GestureTapCancel in DRT
3345
3346         https://bugs.webkit.org/show_bug.cgi?id=96183
3347
3348         Reviewed by Antonio Gomes.
3349
3350         * DumpRenderTree/chromium/TestWebPlugin.cpp:
3351         (TestWebPlugin::handleInputEvent):
3352         * DumpRenderTree/chromium/EventSender.cpp:
3353         (EventSender::gestureTapCancel):
3354
3355 2012-09-17  Philip Rogers  <pdr@google.com>
3356
3357         Teach style checker about preprocessor directive indentation rules
3358         https://bugs.webkit.org/show_bug.cgi?id=96874
3359
3360         Reviewed by Adam Barth.
3361
3362         Preprocessor directives (#ifdef, #include, #define, etc.) should not be indented.
3363         This is not explicit in our style guide but is generally followed in our code.
3364         Searching for violations in our codebase shows these are rarely indented:
3365             #include - indented in 6 files
3366             #ifdef - indented in 0 files
3367             #ifndef - indented in 1 file
3368             #define - indented in 11 files
3369             #if - indented in 7 files
3370
3371         * Scripts/webkitpy/style/checkers/cpp.py:
3372         (check_directive_indentation):
3373
3374             This is the simple test where we look for spaces followed by a #.
3375
3376         (check_style):
3377         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
3378
3379             A few tests needed to be modified because they had unintentionally indented
3380             preprocessor directives.
3381
3382         (CppStyleTest.test_build_class.Foo):
3383         (CppStyleTest.test_build_class):
3384         (CppStyleTest.test_build_class.DERIVE_FROM_GOO):
3385         (WebKitStyleTest.test_line_breaking):
3386         (WebKitStyleTest.test_directive_indentation):
3387
3388 2012-09-17  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3389
3390         [Qt] Auto-generate the module pri file for QtWebKit
3391
3392         Reviewed by Simon Hausmann.
3393
3394         * qmake/qt_webkit.pri: Removed.
3395
3396 2012-09-14  Dirk Pranke  <dpranke@chromium.org>
3397
3398         nrwt: --additional-platform-dir is broken on chromium ports
3399         https://bugs.webkit.org/show_bug.cgi?id=96840
3400
3401         Reviewed by Ojan Vafai.
3402
3403         Looks like I broke this when cleaning things up as part of
3404         adding ML support. Fixing and adding a test.
3405
3406         * Scripts/webkitpy/layout_tests/port/chromium.py:
3407         (ChromiumPort.default_baseline_search_path):
3408         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
3409         (ChromiumAndroidPort.default_baseline_search_path):
3410         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
3411         (ChromiumLinuxPort.default_baseline_search_path):
3412         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
3413         (ChromiumMacPort.__init__):
3414         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
3415         (ChromiumWinPort.setup_environ_for_server):
3416         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
3417         (test_check_build):
3418         (test_additional_platform_directory):
3419
3420 2012-09-13  Stephanie Lewis  <slewis@apple.com>
3421
3422         mac-future ignores previous platform TestExpectations.
3423         https://bugs.webkit.org/show_bug.cgi?id=96718.
3424
3425         Reviewed by Dirk Pranke.
3426
3427         TestExpectations match an expectation's configuraton against a list of configurations that are 
3428         valid for that port.  That list does not contain mac-future so none of the expectations can be
3429         applied.
3430
3431         * Scripts/webkitpy/layout_tests/port/apple.py:
3432         (ApplePort._generate_all_test_configurations):
3433
3434 2012-09-14  Adam Barth  <abarth@webkit.org>
3435
3436         Remove webkitPostMessage
3437         https://bugs.webkit.org/show_bug.cgi?id=96577
3438
3439         Reviewed by Ojan Vafai.
3440
3441         Add ENABLE_LEGACY_VENDOR_PREFIXES flag.
3442
3443         * Scripts/webkitperl/FeatureList.pm:
3444         * qmake/mkspecs/features/features.pri:
3445
3446 2012-09-14  Alexey Proskuryakov  <ap@apple.com>
3447
3448         [WK2] webarchive/loading/javascript-url-iframe-crash.html fails
3449         https://bugs.webkit.org/show_bug.cgi?id=96817
3450
3451         Reviewed by Kenneth Rohde Christiansen.
3452
3453         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3454         (WTR::InjectedBundlePage::didFailProvisionalLoadWithErrorForFrame): This callback
3455         got overlooked when adding logging elsewhere.
3456
3457 2012-09-14  Peter Beverloo  <peter@chromium.org>
3458
3459         [Chromium] Build fix for DumpRenderTree following r128628
3460         https://bugs.webkit.org/show_bug.cgi?id=96808
3461
3462         Reviewed by Tony Chang.
3463
3464         When building DumpRenderTree from the Chromium project, a compile error
3465         shows up because including the header ForwardIOStreamsAndroid.h cannot
3466         be located. TestWebKitAPI does this correctly.
3467
3468         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
3469
3470 2012-09-14  Dana Jansens  <danakj@chromium.org>
3471
3472         Change cc-bugs@google.com watchlist to cc-bugs@chromium.org
3473         https://bugs.webkit.org/show_bug.cgi?id=96805
3474
3475         Reviewed by Dirk Pranke.
3476
3477         * Scripts/webkitpy/common/config/committers.py:
3478         * Scripts/webkitpy/common/config/watchlist:
3479
3480 2012-09-14  Peter Beverloo  <peter@chromium.org>
3481
3482         [Chromium] Support the --{in,out,err}-fifo arguments on TestWebKitAPI and webkit_unit_tests
3483         https://bugs.webkit.org/show_bug.cgi?id=96687
3484
3485         Reviewed by Tony Chang.
3486
3487         Android's DumpRenderTree currently supports these arguments, implemented
3488         as part of TestShellAndroid:
3489         http://trac.webkit.org/browser/trunk/Tools/DumpRenderTree/chromium/TestShellAndroid.cpp?rev=128496
3490
3491         They're used by the layout test runner to get the STDOUT and STDERR while
3492         a layout test run is in process, which is a safer alternative to parsing
3493         all the logcat output manually. The implementation can be seen here:
3494         http://trac.webkit.org/browser/trunk/Tools/Scripts/webkitpy/layout_tests/port/chromium_android.py?rev=128496#L590
3495
3496         This patch generalizes parsing of and applying the effects of these arguments
3497         so that they can be used for TestWebKitAPI and webkit_unit_tests as well.
3498         After this patch, this will make it possible to pull out the output-reading
3499         code from Android's layout test port and generalize it so it can be re-used