check-webkit-style can't handle qt-arm, qt-win, qt-mac, qt-5.0, qt-5.0-wk1 and qt...
[WebKit-https.git] / Tools / ChangeLog
1 2012-10-02  Tony Chang  <tony@chromium.org>
2
3         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
4         https://bugs.webkit.org/show_bug.cgi?id=98140
5
6         Reviewed by Eric Seidel.
7
8         Fix a bug where we were always using the host OS rather than using what was passed in as part of the
9         port name on Qt.
10
11         There's also a bug with being able to test for qt-5.0 expectations files when on a system with
12         Qt 4.8 installed. I'm not sure how to solve that, so for now, stop emitting a style warning about
13         not checking a file.
14
15         * Scripts/webkitpy/layout_tests/port/qt.py:
16         (QtPort._search_paths): Use passed in OS.
17         * Scripts/webkitpy/style/checkers/test_expectations.py:
18         (TestExpectationsChecker.check): Don't warn if we don't know about the TestExpectations file.
19         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
20         (TestExpectationsTestCase.test_determine_port_from_expectations_path): Added a test case for the OS fix.
21         Add a commented out test for the version case.
22
23 2012-10-02  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
24
25         Remove the last bits of MSYS support from webkitperl.
26         https://bugs.webkit.org/show_bug.cgi?id=98195
27
28         Reviewed by Eric Seidel.
29
30         No ports currently need MinGW/MSYS support, so remove the last
31         hackish bits related to it from our Perl code.
32
33         * Scripts/commit-log-editor:
34         (createCommitMessage):
35         * Scripts/old-run-webkit-tests:
36         * Scripts/webkitdirs.pm:
37         (isAnyWindows):
38         (isWindows):
39
40 2012-10-02  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
41
42         Remove MSYS-related changes to the http testing infrastructure.
43         https://bugs.webkit.org/show_bug.cgi?id=98183
44
45         Reviewed by Eric Seidel.
46
47         In the process of removing the cruft that has accumulated in our
48         Apache configuration code, we can basically revert r77161 which
49         added some MSYS (mingw)-specific code to old-run-webkit-tests,
50         since it is not being used by any port anymore.
51
52         * Scripts/run-webkit-httpd: Remove isMsys()-dependent hacks.
53         * Scripts/webkitperl/httpd.pm:
54         (getHTTPDPath): Remove isMsys() code path.
55         (getDefaultConfigForTestDirectory): Revert isMsys() hacks.
56         (getHTTPDConfigPathForTestDirectory): Do not point to the noew
57         removed apache2-msys-httpd.conf.
58         (convertMsysPath): Removed.
59
60 2012-10-02  Sadrul Habib Chowdhury  <sadrul@chromium.org>
61
62         [chromium] Make sure the touch-points in the touch-events have the correct state.
63         https://bugs.webkit.org/show_bug.cgi?id=98110
64
65         Reviewed by Adam Barth.
66
67         TestWebPlugin can now be configured to print out some details about
68         events. For now, details for only touch events are printed. This is
69         used to verify that touch-points in the touch-events are properly
70         set.
71
72         * DumpRenderTree/chromium/TestWebPlugin.cpp:
73         (pointState):
74         (printTouchList):
75         (printEventDetails):
76         (TestWebPlugin::TestWebPlugin):
77         (TestWebPlugin::handleInputEvent):
78         * DumpRenderTree/chromium/TestWebPlugin.h:
79         (TestWebPlugin):
80
81 2012-10-02  Zan Dobersek  <zandobersek@gmail.com>
82
83         [XvfbDriver] First tests in each worker occasionally crash
84         https://bugs.webkit.org/show_bug.cgi?id=96203
85
86         Reviewed by Dirk Pranke.
87
88         Often in debug builds, the first tests in each worker crash because they
89         can't yet open the Xvfb display assigned to them. This indicates Xvfb hasn't
90         yet been set up.
91
92         To avoid this, the execution should halt for a small amount of time before tests
93         are run so the Xvfb process has enough time to get ready. At the moment this is
94         done right after the Xvfb process is spawned and the execution halts for one
95         second. That value is overriden to zero when testing to not prolong the testing time.
96
97         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py:
98         (XvfbDriver.__init__):
99         (XvfbDriver._start):
100         * Scripts/webkitpy/layout_tests/port/xvfbdriver_unittest.py:
101         (XvfbDriverTest.make_driver):
102
103 2012-10-02  Christophe Dumez  <christophe.dumez@intel.com>
104
105         [EFL] Use XvfbDriver for layout tests
106         https://bugs.webkit.org/show_bug.cgi?id=98162
107
108         Reviewed by Kenneth Rohde Christiansen.
109
110         Use XvfbDriver for layout tests on EFL port.
111
112         * Scripts/webkitpy/layout_tests/port/efl.py:
113         (EflPort._driver_class):
114         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py:
115         (XvfbDriver._start): Set DUMPRENDERTREE_TEMP and LOCAL_RESOURCE_ROOT
116         environment variables in XvfbDriver as those are needed for
117         TestRunner::pathToLocalResource().
118
119 2012-10-02  Zan Dobersek  <zandobersek@gmail.com>
120
121         [Gtk] Several tests launch g_object_ref/unref assertion messages
122         https://bugs.webkit.org/show_bug.cgi?id=95062
123
124         Reviewed by Martin Robinson.
125
126         Only increase and decrease reference count of the accessibility
127         UI element if it actually exists.
128
129         * DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
130         (AccessibilityUIElement::AccessibilityUIElement):
131         (AccessibilityUIElement::~AccessibilityUIElement):
132
133 2012-10-02  Douglas Stockwell  <dstockwell@chromium.org>
134
135         [chromium] running layout tests on mountainlion should use chromium TestExpectations
136         https://bugs.webkit.org/show_bug.cgi?id=98119
137
138         Reviewed by Adam Barth.
139
140         Some basic support for mountainlion has already been committed but it failed to refer
141         to the chromium TestExpectations.
142
143         * Scripts/webkitpy/layout_tests/port/chromium.py:
144         (ChromiumPort):
145         * Scripts/webkitpy/layout_tests/port/chromium_port_testcase.py:
146         (ChromiumPortTestCase.test_all_test_configurations):
147
148 2012-10-02  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
149
150         webkitpy: Stop listening on port 8081 when using Apache.
151         https://bugs.webkit.org/show_bug.cgi?id=98155
152
153         Reviewed by Eric Seidel.
154
155         No test currently uses this port, and the lighttpd configuration
156         does not open it.
157
158         This helps the efforts of bringing some consistency to the Apache
159         configuration at the moment by opening the same ports in the
160         webkitpy side and in the httpd.conf side (the idea is to remove
161         the httpd.conf side later), and also makes the behavior consistent
162         across the two servers we currently support.
163
164         * Scripts/webkitpy/layout_tests/servers/apache_http_server.py:
165         (LayoutTestApacheHttpd.__init__):
166
167 2012-10-02  Csaba Osztrogonác  <ossy@webkit.org>
168
169         [WK2] REGRESSION(r128623): It made layout tests extremely slow
170         https://bugs.webkit.org/show_bug.cgi?id=96862
171
172         Reviewed by Simon Hausmann.
173
174         Disable calling the extremely slow WKBundleSetDatabaseQuota() between tests until proper fix.
175
176         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
177         (WTR::InjectedBundle::beginTesting):
178
179 2012-10-02  Philip Rogers  <pdr@google.com>
180
181         Fix PerfTest standard deviation calculation.
182         https://bugs.webkit.org/show_bug.cgi?id=98115
183
184         Reviewed by Ryosuke Niwa.
185
186         Previously our standard deviation calculation was incorrect. This patch
187         updates perftest.py's algorithm to calculate the sample standard deviation
188         (with Bessel's correction) using Knuth's online algorithm:
189         http://en.wikipedia.org/wiki/Algorithms_for_calculating_variance#Online_algorithm
190         An existing test has been modified to prove our new results.
191
192         This patch also updates runner.js to use Bessel's correction in
193         its sample standard deviation calculation, which is more accurate
194         for small sample sizes.
195
196         Additionally, runner.js has been modified to not calculate
197         the 'sum' statistic, which was not very useful.
198
199         * Scripts/webkitpy/performance_tests/perftest.py:
200
201             The unused variable valueSum has also been removed.
202
203         (PageLoadingPerfTest.run):
204         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
205
206             This test calculates the stdev of {2000, 3000, ..., 20000} which
207             was hand-calculated using a spreadsheet.
208
209         (TestPageLoadingPerfTest.test_run):
210
211 2012-10-01  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
212
213         webkitpy should accept a different httpd.conf specified by the user
214         https://bugs.webkit.org/show_bug.cgi?id=98071
215
216         Reviewed by Dirk Pranke.
217
218         The existing httpd.conf variants (or lighttpd.conf, for that
219         matter) we have do not always suit the user's system. This is
220         particularly true on Linux/Unix, where Apache can be installed in
221         a plethora of ways and the LoadModule calls can fail to specify
222         the proper module paths.
223
224         For now, we start accepting the WEBKIT_HTTP_SERVER_CONF_PATH
225         environment variable, which allows the user to specify the
226         absolute path to another http server configuration file that might
227         work on the user's system.
228
229         In the long term, we should try to generate our configuration file
230         and stop requiring all the different httpd.conf files we have as
231         well as this hack.
232
233         * Scripts/webkitpy/layout_tests/port/base.py:
234         (Port._path_to_apache_config_file):
235         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
236         (test_path_to_apache_config_file):
237
238 2012-10-01  Emil A Eklund  <eae@chromium.org>
239
240         Unreviewed, upgrade eae to reviewer. 
241         http://lists.webkit.org/mailman/private/webkit-committers/2012-October/000191.html
242
243         * Scripts/webkitpy/common/config/committers.py:
244
245 2012-10-01  Ming Xie  <mxie@rim.com>
246
247         [BlackBerry] Add a early return in buildCMakeProjectOrExit()
248         https://bugs.webkit.org/show_bug.cgi?id=98052
249
250         Reviewed by Rob Buis.
251
252         This option allows our build to only run the CMake build configuration
253         step when environment variable GENERATE_CMAKE_PROJECT_ONLY is specified.
254
255         * Scripts/webkitdirs.pm:
256         (buildCMakeProjectOrExit):
257
258 2012-10-01  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
259
260         [webkitpy] Detect the multiple names of the wdiff binary in the Port class.
261         https://bugs.webkit.org/show_bug.cgi?id=98039
262
263         Reviewed by Dirk Pranke.
264
265         Unify the duplicate checks for the wdiff binary present in
266         different ports into the base Port class.
267
268         * Scripts/webkitpy/layout_tests/port/base.py:
269         (Port):
270         (Port._path_to_wdiff):
271         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
272         (ChromiumLinuxPort._path_to_wdiff): Removed.
273         * Scripts/webkitpy/layout_tests/port/gtk.py:
274         (GtkPort._path_to_wdiff): Removed.
275
276 2012-10-01  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
277
278         [chromium] Remove custom implementation of _path_to_apache_config_file in chromium_linux.
279         https://bugs.webkit.org/show_bug.cgi?id=98042
280
281         Reviewed by Dirk Pranke.
282
283         The Chromium-based ports inherit from Port these days, so get rid
284         of the custom implementation of _path_to_apache_config_file() in
285         favor of the more portable one in the Port class.
286
287         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
288         (ChromiumLinuxPort._path_to_apache_config_file):
289
290 2012-10-01  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
291
292         [chromium] Remove duplicated _is_redhat_based() implementation from chromium_linux.py.
293         https://bugs.webkit.org/show_bug.cgi?id=98035
294
295         Reviewed by Dirk Pranke.
296
297         Since the Chromium port classes inherit from Port these days, it
298         can rely on the same implementation of _is_redhat_based() present
299         there instead of duplicating it.
300
301         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
302         (ChromiumLinuxPort._is_redhat_based): Removed.
303
304 2012-10-01  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
305
306         [GTK] Rely on the general implementation for detecting the Apache server path.
307         https://bugs.webkit.org/show_bug.cgi?id=98033
308
309         Reviewed by Martin Robinson.
310
311         The implemenentation of _path_to_apache() in base.py already
312         checks for the same paths gtk.py tries; the only difference is
313         that all of the are tried regardless of the current distro.
314
315         Doing so is more portable, and lets us remove some distro-checking
316         from the code.
317
318         * Scripts/webkitpy/layout_tests/port/gtk.py:
319         (GtkPort._path_to_apache): Remove.
320
321 2012-10-01  Csaba Osztrogonác  <ossy@webkit.org>
322
323         Unreviewed, update the URL of the Szeged SVN mirror.
324
325         * BuildSlaveSupport/build.webkit.org-config/config.json:
326
327 2012-10-01  Yuta Kitamura  <yutak@chromium.org>
328
329         Unreviewed. Create WebSocket watch list, and add myself to that list.
330
331         * Scripts/webkitpy/common/config/watchlist:
332
333 2012-09-30  Taiju Tsuiki  <tzik@chromium.org>
334
335         Unreviewed. Adding myself as a committer.
336
337         * Scripts/webkitpy/common/config/committers.py:
338
339 2012-09-30  Sadrul Habib Chowdhury  <sadrul@chromium.org>
340
341         [chromium] Add a test to verify that a plugin does not receive touch-events unless it explicitly requests for them
342         https://bugs.webkit.org/show_bug.cgi?id=97975
343
344         Reviewed by Adam Barth.
345
346         TestWebPlugin can now be configured to accept touch events or not. Use this
347         configurability to verify that a plugin does not receive touch events unless
348         it explicitly requests for them.
349
350         * DumpRenderTree/chromium/TestWebPlugin.cpp:
351         (TestWebPlugin::TestWebPlugin):
352         (TestWebPlugin::initialize):
353         (TestWebPlugin::parseBoolean):
354         * DumpRenderTree/chromium/TestWebPlugin.h:
355         (TestWebPlugin):
356
357 2012-09-28  Mariusz Grzegorczyk  <mariusz.g@samsung.com>
358
359         [WK2][GTK][EFL] Share WebKit2-GTK plugin process implementation with EFL port
360         https://bugs.webkit.org/show_bug.cgi?id=91844
361
362         Reviewed by Simon Hausmann.
363
364         Turn on Netscape Plugin API option by default.
365
366         * Scripts/webkitperl/FeatureList.pm:
367
368 2012-09-28  Dirk Pranke  <dpranke@chromium.org>
369
370         webkit-patch rebaseline-expectations is broken
371         https://bugs.webkit.org/show_bug.cgi?id=97619
372
373         Reviewed by Ojan Vafai.
374
375         The new expectations format is causing two problems for
376         webkit-patch rebaseline expectations
377
378         1) the transformation into and out of the old syntax (which is
379         still used internally) is somewhat lossy, e.g., we're not
380         preserving the case of Bug(x) identifiers. Also, we can't
381         tell if the input was [ WontFix ] or [ Skip WontFix Pass ]
382         2) the new syntax is more lenient, allowing for multiple ways to
383         specify the same result, e.g., "[ Skip ]" may or may not be
384         missing.
385
386         As a result, we end up rewriting a whole bunch of lines for
387         purely syntactic differences.
388
389         Both of these things are bugs that need to be fixed, but
390         we should also not just rewrite lines that shouldn't be
391         modified.
392
393         This patch fixes the last issue (no rewriting).
394
395         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
396         (TestExpectationLine._format_line):
397         (TestExpectations.remove_rebaselined_tests):
398         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
399         (RebaseliningTest.test_remove):
400
401 2012-09-28  Dirk Pranke  <dpranke@chromium.org>
402
403         virtual directories should just rename the beginning of the test path
404         https://bugs.webkit.org/show_bug.cgi?id=97925
405
406         Reviewed by Ojan Vafai.
407
408         We were replacing all occurrences of the base prefix of a
409         virtual test suite with the virtual name, instead of just the
410         first. So, where we had softwarecompositing running the
411         compositing tests, any compositing test that actually had
412         "compositing" in the test base name would get that expanded
413         incorrectly.
414
415         * Scripts/webkitpy/layout_tests/port/base.py:
416         (Port.populated_virtual_test_suites):
417         (Port.lookup_virtual_test_base):
418         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
419         (PortTest.test_tests):
420         * Scripts/webkitpy/layout_tests/port/test.py:
421
422 2012-09-28  Ojan Vafai  <ojan@chromium.org>
423
424         Garden-o-matic sometimes things buildbot warnings are failures
425         https://bugs.webkit.org/show_bug.cgi?id=97926
426
427         Reviewed by Tony Chang.
428
429         It turns out the JSON clearly tells us if it's a failure.
430         We want to skip all warnings even if they don't have "warning" in the text.
431         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/builders.js:
432         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/builders_unittests.js:
433
434 2012-09-28  Harald Tveit Alvestrand  <harald@alvestrand.no>
435
436         Implement the GetStats interface on PeerConnection
437         https://bugs.webkit.org/show_bug.cgi?id=95193
438
439         Reviewed by Adam Barth.
440
441         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
442         (RTCStatsRequestSucceededTask):
443         (RTCStatsRequestSucceededTask::RTCStatsRequestSucceededTask):
444         (MockWebRTCPeerConnectionHandler::getStats):
445         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.h:
446         (MockWebRTCPeerConnectionHandler):
447
448 2012-09-28  Kristóf Kosztyó  <kkristof@inf.u-szeged.hu>
449
450         [NRWT] XvfbDriver should choose the next free display
451         https://bugs.webkit.org/show_bug.cgi?id=88414
452
453         Reviewed by Dirk Pranke.
454
455         This change how the XvfbDriver choose the next display. Before
456         this the choosing are based on the worker number but it caused
457         errors when several nrwt run on the same time. This add process
458         based free diplay search.
459
460         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py:
461         (XvfbDriver.__init__):
462         (XvfbDriver._next_free_display):
463         (XvfbDriver._start):
464         (XvfbDriver.stop):
465         * Scripts/webkitpy/layout_tests/port/xvfbdriver_unittest.py:
466         (XvfbDriverTest.make_driver):
467         (XvfbDriverTest.test_start_no_pixel_tests):
468         (XvfbDriverTest.test_start_pixel_tests):
469         (XvfbDriverTest.test_start_arbitrary_worker_number):
470         (XvfbDriverTest.test_next_free_display):
471         (XvfbDriverTest):
472         (XvfbDriverTest.test_start_next_worker):
473
474 2012-09-28  Huang Dongsung  <luxtella@company100.net>
475
476         [GTK] Enable CSS Shaders layout LayoutTests on GTK+
477         https://bugs.webkit.org/show_bug.cgi?id=97821
478
479         Reviewed by Martin Robinson.
480
481         This feature is disabled via Settings by default, but for testing,
482         we enable it using layoutTestController.overridePreferences. Add the
483         necessary plumbing for DRT.
484
485         WTR already works because support was added for Apple Mac earlier.
486
487         * DumpRenderTree/gtk/DumpRenderTree.cpp:
488         (resetDefaultsToConsistentValues): Feature is off by default.
489         * DumpRenderTree/gtk/TestRunnerGtk.cpp:
490         (TestRunner::overridePreference): Add handling of WebKitCSSCustomFilterEnabled.
491
492 2012-09-28  Csaba Osztrogonác  <ossy@webkit.org>
493
494         [Qt] Make NRWT use wk2/TestExpectations file
495         https://bugs.webkit.org/show_bug.cgi?id=97894
496
497         Reviewed by Simon Hausmann.
498
499         * Scripts/webkitpy/layout_tests/port/qt.py: Remove unnecessary _skipped_file_search_paths function.
500         (QtPort.expectations_files): Add wk2 if --webkit-test-runner option added.
501         * Scripts/webkitpy/layout_tests/port/qt_unittest.py: Remove unnecessary test__skipped_file_search_paths function.
502         (QtPortTest._assert_search_path): Removed.
503         (QtPortTest.test_expectations_files): Add wk2 if --webkit-test-runner option added, and code cleanup.
504
505 2012-09-28  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
506
507         [EFL][DRT] EFL's layoutTestController.keepWebHistory should enable visited links tracking
508         https://bugs.webkit.org/show_bug.cgi?id=97901
509
510         Reviewed by Kenneth Rohde Christiansen.
511
512         EFL's DRT testRunner.keepWebHistory should enable visited links tracking in order
513         to be consistent with WTR implementation and also several tests rely on that (fast/history).
514
515         * DumpRenderTree/efl/TestRunnerEfl.cpp:
516         (TestRunner::keepWebHistory):
517
518 2012-09-28  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
519
520         [EFL][DRT] EFL DRT should disable visited links tracking by default
521         https://bugs.webkit.org/show_bug.cgi?id=97881
522
523         Reviewed by Kenneth Rohde Christiansen.
524
525         Visited links tracking is disabled by default for each test.
526
527         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
528         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
529
530 2012-09-28  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
531
532         Unreviewed, rolling out r129863.
533         http://trac.webkit.org/changeset/129863
534         https://bugs.webkit.org/show_bug.cgi?id=97173
535
536         Broken debug WK2 layout test
537
538         * MiniBrowser/efl/main.c:
539         (quit):
540         (main):
541         * WebKitTestRunner/efl/PlatformWebViewEfl.cpp:
542         (WTR::initEcoreEvas):
543         (WTR::PlatformWebView::~PlatformWebView):
544         * WebKitTestRunner/efl/main.cpp:
545         (main):
546
547 2012-09-28  Kai Koehne  <kai.koehne@digia.com>
548
549         [Qt] Use qInstallMessageHandler()
550         https://bugs.webkit.org/show_bug.cgi?id=96648
551
552         Reviewed by Jocelyn Turcotte.
553
554         qInstallMsgHandler() got deprecated in Qt 5.
555
556         * DumpRenderTree/qt/DumpRenderTreeMain.cpp:
557         (messageHandler):
558         (main):
559         * WebKitTestRunner/qt/main.cpp:
560         (messageHandler):
561         (main):
562
563 2012-09-28  Eunmi Lee  <eunmi15.lee@samsung.com>
564
565         [EFL][WK2] Refactoring initialization and shutdown codes of EFL libraries.
566         https://bugs.webkit.org/show_bug.cgi?id=97173
567
568         Reviewed by Gyuyoung Kim.
569
570         The initialization and shutdown are done inside the webkit, so we don't
571         have to call them out of the webkit.
572
573         * MiniBrowser/efl/main.c:
574         (quit):
575         (main):
576         * WebKitTestRunner/efl/PlatformWebViewEfl.cpp:
577         (WTR::initEcoreEvas):
578         (WTR::PlatformWebView::~PlatformWebView):
579         * WebKitTestRunner/efl/main.cpp:
580         (main):
581
582 2012-09-27  Alexandre Elias  <aelias@chromium.org>
583
584         [chromium] DumpRenderTree support for software compositing
585         https://bugs.webkit.org/show_bug.cgi?id=96853
586
587         Reviewed by James Robinson.
588
589         This adds software compositor support to DumpRenderTree.  The only
590         thing needed is a minimal OutputSurfaceSoftware that holds a
591         viewport-sized bitmap. Then, I added virtual test suite
592         platform/chromium/virtual/softwarecompositing -- a mirror of the
593         compositing/ directory with --enable-software-compositing, and
594         expectations for not yet supported tests.
595
596         * DumpRenderTree/DumpRenderTree.gypi:
597         * DumpRenderTree/chromium/DumpRenderTree.cpp:
598         (main):
599         * DumpRenderTree/chromium/TestShell.cpp:
600         (TestShell::TestShell):
601         * DumpRenderTree/chromium/TestShell.h:
602         (TestShell::softwareCompositingEnabled):
603         (TestShell::setSoftwareCompositingEnabled):
604         (TestShell):
605         * DumpRenderTree/chromium/WebViewHost.cpp:
606         (WebViewHost::createOutputSurface):
607         * DumpRenderTree/chromium/WebViewHostOutputSurface.cpp:
608         (WebKit::WebViewHostOutputSurface::create3d):
609         (WebKit):
610         (WebKit::WebViewHostOutputSurface::createSoftware):
611         (WebKit::WebViewHostOutputSurface::WebViewHostOutputSurface):
612         (WebKit::WebViewHostOutputSurface::bindToClient):
613         (WebKit::WebViewHostOutputSurface::surfaceSoftware):
614         * DumpRenderTree/chromium/WebViewHostOutputSurface.h:
615         (WebKit):
616         (WebViewHostOutputSurface):
617         * DumpRenderTree/chromium/WebViewHostSoftwareOutputDevice.cpp: Copied from Tools/DumpRenderTree/chromium/WebViewHostOutputSurface.cpp.
618         (WebKit):
619         (WebKit::WebViewHostSoftwareOutputDevice::lockForWrite):
620         (WebKit::WebViewHostSoftwareOutputDevice::unlockForWrite):
621         (WebKit::WebViewHostSoftwareOutputDevice::lockForRead):
622         (WebKit::WebViewHostSoftwareOutputDevice::unlockForRead):
623         (WebKit::WebViewHostSoftwareOutputDevice::viewportChanged):
624         * DumpRenderTree/chromium/WebViewHostSoftwareOutputDevice.h: Copied from Tools/DumpRenderTree/chromium/WebViewHostOutputSurface.cpp.
625         (WebKit):
626         (WebViewHostSoftwareOutputDevice):
627         * Scripts/webkitpy/layout_tests/port/chromium.py:
628         (ChromiumPort.virtual_test_suites):
629
630 2012-09-27  Nasko Oskov  <nasko@chromium.org>
631
632         Move DumpRenderTree to use newer version of the name method of
633         WebFrame and the willCheckAndDispatchMessageEvent method of WebFrameClient
634         https://bugs.webkit.org/show_bug.cgi?id=95254
635
636         Reviewed by Adam Barth.
637
638         The WebFrame name() method is removed and replaced by better named
639         uniqueName() method, so DRT needs to be updated to use this new version.
640         Also the willCheckAndDispatchMessageEvent in WebFrameClient gained
641         one more parameter, which had to be added in DRT.
642
643         * DumpRenderTree/chromium/TestShell.cpp:
644         (dumpFramesAsText):
645         (dumpFramesAsPrintedText):
646         (dumpFrameScrollPosition):
647         * DumpRenderTree/chromium/WebViewHost.cpp:
648         (WebViewHost::unableToImplementPolicyWithError):
649         (WebViewHost::willCheckAndDispatchMessageEvent):
650         (WebViewHost::printFrameDescription):
651         * DumpRenderTree/chromium/WebViewHost.h:
652         (WebViewHost):
653
654 2012-09-27  Julie Parent  <jparent@chromium.org>
655
656         Add webkit_unit_tests to the flakiness dashboard
657         https://bugs.webkit.org/show_bug.cgi?id=97814
658
659         Reviewed by Dirk Pranke.
660
661         webkit_unit_tests run on the same bots as test_shell_tests.
662
663         * TestResultServer/static-dashboards/builders.js:
664         (loadBuildersList):
665         * TestResultServer/static-dashboards/dashboard_base.js:
666         (currentBuilderGroupCategory):
667
668 2012-09-27  Florin Malita  <fmalita@chromium.org>
669
670         [Perftest] PerfTestRunner exception when using nested git checkouts
671         https://bugs.webkit.org/show_bug.cgi?id=97796
672
673         Reviewed by Ryosuke Niwa.
674
675         In order to support separate Chromium vs. WebKit git trees, PerfTestsRunner needs to
676         instantiate per-repository SCM objects when collecting revision information.
677
678         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
679         (PerfTestsRunner._generate_results_dict):
680
681 2012-09-27  Sudarsana Nagineni  <sudarsana.nagineni@intel.com>
682
683         [Mac][DRT] Add support to allow non-array modifier arguments in keyDown()
684         https://bugs.webkit.org/show_bug.cgi?id=97805
685
686         Reviewed by Alexey Proskuryakov.
687
688         Add support for handling a string as modifier argument to keyDown().
689         This support has already been added to other ports, so implementing
690         this on Mac too, to be consistent with the other ports.
691
692         * DumpRenderTree/mac/EventSendingController.mm:
693         (modifierFlags):
694         (buildModifierFlags):
695
696 2012-09-27  Dirk Pranke  <dpranke@chromium.org>
697
698         efl wk2 bots weren't looking at platform/wk2/TestExpectations
699         https://bugs.webkit.org/show_bug.cgi?id=97815
700
701         Unreviewed, build fix.
702
703         Should clean up this logic later so that we're just using the
704         implementation in base ...
705
706         * Scripts/webkitpy/layout_tests/port/efl.py:
707         (EflPort._search_paths):
708         (EflPort.expectations_files):
709
710 2012-09-27  Dirk Pranke  <dpranke@chromium.org>
711
712         webkitpy: make warning about missing BUG identifiers per-port configurable
713         https://bugs.webkit.org/show_bug.cgi?id=97751
714
715         Reviewed by Ojan Vafai.
716
717         Prior to this change linting TestExpectations files (checking their style)
718         would raise warnings if lines were missing a BUG modifier, even
719         though such a modifier is actually optional in the syntax. As we
720         migrate the entries over from the Skipped files, these warnings
721         will generate a lot of noise, and it's not yet clear that we
722         actually want to require the modifiers on all ports.
723
724         So, this change makes this per-port configurable so that they
725         will generate warnings in the Chromium TestExpectations and not
726         anywhere else.
727
728         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
729         (TestExpectationParser._parse_modifiers):
730         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
731         (MiscTests.test_parse_warning):
732         (SemanticTests.test_missing_bugid):
733         * Scripts/webkitpy/layout_tests/port/base.py:
734         (Port.warn_if_bug_missing_in_test_expectations):
735         * Scripts/webkitpy/layout_tests/port/chromium.py:
736         (ChromiumPort.warn_if_bug_missing_in_test_expectations):
737
738 2012-09-27  Xianzhu Wang  <wangxianzhu@chromium.org>
739
740         [Chromium-Android] Push more fonts for layout tests to match chromium-linux
741         https://bugs.webkit.org/show_bug.cgi?id=97746
742
743         Reviewed by Dirk Pranke.
744
745         * DumpRenderTree/chromium/android_fallback_fonts.xml: Added the new fonts into fallback list. Removed the Android system fallback font DroidSansFallback.ttf.
746         * Scripts/webkitpy/layout_tests/port/builders.py: Add chromium-android to pass style check.
747         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
748         (ChromiumAndroidPort.check_build): Added check for md5sum and forwarder utilities to give better error message when they are missing.
749         (ChromiumAndroidPort.check_sys_deps): Allow host font files to come from one of multiple directories.
750         (ChromiumAndroidPort._path_to_md5sum):
751         (ChromiumAndroidDriver._setup_md5sum_and_push_data_if_needed):
752         (ChromiumAndroidDriver._push_fonts):
753         (ChromiumAndroidDriver._push_test_resources):
754
755 2012-09-27  Tommy Widenflycht  <tommyw@google.com>
756
757         MediaStream API: Enhance MediaConstraints to make it easier to get the constraint data
758         https://bugs.webkit.org/show_bug.cgi?id=97559
759
760         Reviewed by Adam Barth.
761
762         Refactoring for the new WebMediaConstraints api.
763
764         * DumpRenderTree/chromium/MockConstraints.cpp:
765         (MockConstraints::verify):
766
767 2012-09-27  Tommy Widenflycht  <tommyw@google.com>
768
769         MediaStream API: Rename MockConstraints::verify to verifyConstraints
770         https://bugs.webkit.org/show_bug.cgi?id=97779
771
772         Reviewed by Csaba Osztrogonác.
773
774         The Mac OS X sdk has a macro called verify :/
775
776         * DumpRenderTree/chromium/MockConstraints.cpp:
777         (MockConstraints::verifyConstraints):
778         * DumpRenderTree/chromium/MockConstraints.h:
779         (MockConstraints):
780         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
781         (MockWebRTCPeerConnectionHandler::initialize):
782         * DumpRenderTree/chromium/WebUserMediaClientMock.cpp:
783         (WebUserMediaClientMock::requestUserMedia):
784
785 2012-09-27  Tommy Widenflycht  <tommyw@google.com>
786
787         MediaStream API: Update getUserMedia to match the latest specification
788         https://bugs.webkit.org/show_bug.cgi?id=97540
789
790         Reviewed by Adam Barth.
791
792         http://dev.w3.org/2011/webrtc/editor/getusermedia.html#navigatorusermedia
793         Navigator.getUserMedia is changed so that the audio and video members can either
794         be a bool or a constraints object.
795
796         * public/WebUserMediaRequest.h:
797         (WebKit):
798         (WebUserMediaRequest):
799         * src/WebUserMediaRequest.cpp:
800         (WebKit::WebUserMediaRequest::audioConstraints):
801         (WebKit):
802         (WebKit::WebUserMediaRequest::videoConstraints):
803
804 2012-09-26  Laszlo Gombos  <l.gombos@samsung.com>
805
806         Unreviewed. Update my e-mail address.
807
808         * Scripts/webkitpy/common/config/committers.py:
809
810 2012-09-26  Dirk Pranke  <dpranke@chromium.org>
811
812         modify old-run-webkit-tests to support TestExpectations files a little
813         https://bugs.webkit.org/show_bug.cgi?id=97276
814
815         Reviewed by Daniel Bates.
816
817         This patch adds minimal support to old-run-webkit-tests so that
818         it can read a TestExpectations file and skip the tests listed
819         there. It will skip *all* tests, regardless of the expectation
820         (even if the test is listed as [ Pass ]), and it will only look
821         at the TestExpectations file if there isn't a Skipped file next
822         to it.
823         
824         This will allow us to switch to using just using
825         TestExpectations files without making old-run-webkit-tests
826         totally useless.
827
828         * Scripts/old-run-webkit-tests:
829         (readSkippedFiles):
830         (processSkippedFileEntry): Added.
831         (startsWith): Added.
832
833 2012-09-26  Dirk Pranke  <dpranke@chromium.org>
834
835         The style bot spams about skia_test_expectations.txt
836         https://bugs.webkit.org/show_bug.cgi?id=97699
837
838         Reviewed by Adam Barth.
839
840         Remove the warning about a missing skia_test_expectations.txt
841         file for now. There are some situations where it would be
842         helpful to get this warning as it might indicate that the user
843         is going to get results different than what they're expecting,
844         but at the moment it looks difficult if not impossible to get
845         those warnings and not also get warnings that we don't care
846         about, so the noise is outweighing the value of the signal.
847
848         This at least will stop the false negatives we're getting on the
849         style bots. If it turns out people do need the warning we can
850         add it back in in some other more conditional way.
851
852         * Scripts/webkitpy/layout_tests/port/chromium.py:
853         (ChromiumPort.expectations_files):
854         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
855         (TestRebaseline.test_rebaseline_updates_expectations_file_noop):
856         (test_rebaseline_updates_expectations_file):
857         (test_rebaseline_does_not_include_overrides):
858         (test_rebaseline_expectations):
859         (_assert_command):
860
861 2012-09-26  David Michael Barr  <davidbarr@chromium.org>
862
863         Unreviewed. Upgrade myself from contributor to committer.
864
865         * Scripts/webkitpy/common/config/committers.py:
866
867 2012-09-26  Ryosuke Niwa  <rniwa@webkit.org>
868
869         Remove deprecated code from perftestrunner.py
870         https://bugs.webkit.org/show_bug.cgi?id=97724
871
872         Reviewed by Adam Barth.
873
874         Remove the code deprecated in r129580 now that we have landed r129597,
875         which updated master.cfg to use new options, and restarted the buildbot master.
876
877         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
878         (PerfTestsRunner._parse_args):
879         (PerfTestsRunner._generate_and_show_results):
880         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
881         (test_run_with_slave_config_json):
882         (test_parse_args):
883
884 2012-09-26  Beth Dakin  <bdakin@apple.com>
885
886         https://bugs.webkit.org/show_bug.cgi?id=97629
887         http/tests/security/cross-frame-access-put.html failing on Mac WK2 bots after 
888         being unskipped
889
890         Reviewed by Tim Horton.
891
892         The test was failing because we were failing to use the fake window origin. We 
893         should really always use the fake origin. This patch eliminates 
894         _shouldUseFakeOrigin in favor of always using it.
895         * WebKitTestRunner/mac/PlatformWebViewMac.mm:
896         (-[WebKitTestRunnerWindow setFrameOrigin:]):
897         (-[WebKitTestRunnerWindow setFrame:display:animate:]):
898         (-[WebKitTestRunnerWindow setFrame:display:]):
899         (-[WebKitTestRunnerWindow frameRespectingFakeOrigin]):
900
901 2012-09-26  Tony Chang  <tony@chromium.org>
902
903         C++ style checker should warn when the indentation is wrong
904         https://bugs.webkit.org/show_bug.cgi?id=97602
905
906         Reviewed by Ojan Vafai.
907
908         Rewrite the indentation checker to ensure that indentation is always a factor of 4
909         and that we only indent 4 spaces from the previous line.
910
911         I deleted some of the old indentation checking code since it was filtered out
912         and didn't match the webkit style.
913
914         * Scripts/webkitpy/style/checker.py:
915         * Scripts/webkitpy/style/checkers/cpp.py:
916         (check_indentation_amount): New function that checks indentation. Intentionally runs after
917         other indentation checks.
918         (check_style): Delete the old indentation code since we were already filtering out the whitespace/labels warnings.
919         (CppChecker): Remove the whitespace/labels filter.
920         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
921         (CppStyleTest.test_multi_line_comments): Force 4 space indent.
922         (CppStyleTest.test_explicit_single_argument_constructors):
923         (CppStyleTest.test_explicit_single_argument_constructors.Foo):
924         (CppStyleTest.test_explicit_single_argument_constructors.Qualifier):
925         (CppStyleTest.test_slash_star_comment_on_single_line):
926         (CppStyleTest.test_braces):
927         (CppStyleTest.test_spacing_before_last_semicolon):
928         (CppStyleTest.test_static_or_global_stlstrings):
929         (CppStyleTest.test_indent): Added macro test case.
930         (CppStyleTest.test_build_class):
931         (NoNonVirtualDestructorsTest.test_no_error):
932         (NoNonVirtualDestructorsTest.test_no_error.Foo):
933         (NoNonVirtualDestructorsTest.test_no_error.Qualified):
934         (NoNonVirtualDestructorsTest.test_no_destructor_when_virtual_needed):
935         (NoNonVirtualDestructorsTest.test_no_destructor_when_virtual_needed.Foo):
936         (NoNonVirtualDestructorsTest.test_destructor_non_virtual_when_virtual_needed):
937         (NoNonVirtualDestructorsTest.test_destructor_non_virtual_when_virtual_needed.Foo):
938         (NoNonVirtualDestructorsTest.test_no_warn_when_derived):
939         (NoNonVirtualDestructorsTest.test_no_warn_when_derived.Foo):
940         (NoNonVirtualDestructorsTest.test_internal_braces):
941         (NoNonVirtualDestructorsTest.test_internal_braces.Foo):
942         (NoNonVirtualDestructorsTest.test_inner_class_needs_virtual_destructor):
943         (NoNonVirtualDestructorsTest.test_inner_class_needs_virtual_destructor.Foo):
944         (NoNonVirtualDestructorsTest.test_inner_class_needs_virtual_destructor.Foo.Goo):
945         (NoNonVirtualDestructorsTest.test_outer_class_needs_virtual_destructor):
946         (NoNonVirtualDestructorsTest.test_outer_class_needs_virtual_destructor.Foo):
947         (NoNonVirtualDestructorsTest.test_outer_class_needs_virtual_destructor.Foo.Goo):
948         (NoNonVirtualDestructorsTest.test_qualified_class_needs_virtual_destructor):
949         (NoNonVirtualDestructorsTest.test_qualified_class_needs_virtual_destructor.Qualified):
950         (NoNonVirtualDestructorsTest.test_multi_line_declaration_no_error):
951         (NoNonVirtualDestructorsTest.test_multi_line_declaration_no_error.Foo):
952         (NoNonVirtualDestructorsTest.test_multi_line_declaration_with_error):
953         (NoNonVirtualDestructorsTest.test_multi_line_declaration_with_error.Foo):
954         (WebKitStyleTest.test_indentation):
955         (WebKitStyleTest.test_braces): Add test for call params spanning lines.
956
957 2012-09-26  Adam Barth  <abarth@webkit.org>
958
959         Rollout messages should be clickable in bugzilla
960         https://bugs.webkit.org/show_bug.cgi?id=97711
961
962         Reviewed by Simon Fraser.
963
964         If we say Bug NNNN instead of just NNNN, bugzilla will autolink to the
965         actual bug.
966
967         * Scripts/webkitpy/tool/commands/download_unittest.py:
968         * Scripts/webkitpy/tool/steps/createbug.py:
969         (CreateBug.run):
970
971 2012-09-26  Marcelo Lira  <marcelo.lira@openbossa.org>
972
973         run-perf-tests must expand environment variables in user provided paths
974         https://bugs.webkit.org/show_bug.cgi?id=97686
975
976         Reviewed by Ryosuke Niwa.
977
978         The run-perf-tests command line options that receive paths from the
979         user now expand any possible environment variables, because Python's
980         file handling methods do not handle those.
981
982         Examples:
983         run-perf-tests --platform=qt --release --output-json-path=~/perf-results
984         run-perf-tests --platform=qt --release --output-json-path=$HOME/perf-results
985
986         Also removed unused imports.
987
988         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
989         (PerfTestsRunner._parse_args._expand_path): expand environment variables in a path passed via command line
990         (PerfTestsRunner._parse_args):
991
992 2012-09-26  Sheriff Bot  <webkit.review.bot@gmail.com>
993
994         Unreviewed, rolling out r129654.
995         http://trac.webkit.org/changeset/129654
996         https://bugs.webkit.org/show_bug.cgi?id=97702
997
998         breaks chromium windows build (Requested by schenney on
999         #webkit).
1000
1001         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
1002         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.h:
1003         (MockWebRTCPeerConnectionHandler):
1004
1005 2012-09-26  Michael Saboff  <msaboff@apple.com>
1006
1007         Unreviewed build change.
1008
1009         Reverting inadvertent change to project file.
1010
1011         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
1012
1013 2012-09-26  Harald Tveit Alvestrand  <harald@alvestrand.no>
1014
1015         Implement the GetStats interface on PeerConnection
1016         https://bugs.webkit.org/show_bug.cgi?id=95193
1017
1018         Reviewed by Adam Barth.
1019
1020         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
1021         (RTCStatsRequestSucceededTask):
1022         (RTCStatsRequestSucceededTask::RTCStatsRequestSucceededTask):
1023         (MockWebRTCPeerConnectionHandler::getStats):
1024         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.h:
1025         (MockWebRTCPeerConnectionHandler):
1026
1027 2012-09-26  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
1028
1029         [WK2][WTR] Policy client: dumping from decidePolicyForResponse callback
1030         https://bugs.webkit.org/show_bug.cgi?id=97034
1031
1032         Reviewed by Kenneth Rohde Christiansen.
1033
1034         Provided dumping from WTR Policy client 'decidePolicyForResponse' callback.
1035
1036         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1037         (WTR::InjectedBundlePage::decidePolicyForResponse):
1038
1039 2012-09-26  Zoltan Arvai  <zarvai@inf.u-szeged.hu>
1040
1041         [Qt][WTR] Build fix for MSVC. 
1042         https://bugs.webkit.org/show_bug.cgi?id=97667
1043
1044         Reviewed by Csaba Osztrogonác.
1045
1046         WTR misses InjectedBundle headers. Adding $$PWD to Target.pri is required for MSVC. Clean build is necessary on Windows.
1047
1048         * WebKitTestRunner/InjectedBundle/Target.pri:
1049
1050 2012-09-26  Xiaobo Wang  <xbwang@torchmobile.com.cn>
1051
1052         [BlackBerry] [DRT] eventSender.keyDown() doesn't support Ctrl key modifier
1053         https://bugs.webkit.org/show_bug.cgi?id=97665
1054
1055         Reviewed by Rob Buis.
1056
1057         This patch fixes test fast/forms/input-text-option-delete.html.
1058
1059         * DumpRenderTree/blackberry/EventSender.cpp:
1060         (keyDownCallback):
1061
1062 2012-09-26  Dinu Jacob  <dinu.s.jacob@intel.com>
1063
1064         Unreviewed. Adding myself as a committer.
1065
1066         * Scripts/webkitpy/common/config/committers.py:
1067
1068 2012-09-26  Carlos Garcia Campos  <cgarcia@igalia.com>
1069
1070         [GTK] Remove color codes characters from run-gtk-tests output when not writing to a TTY
1071         https://bugs.webkit.org/show_bug.cgi?id=97660
1072
1073         Reviewed by Philippe Normand.
1074
1075         Remove also the '\r' characters returned by TTY output to try to fix
1076         the double line endings of the script output in the bots.
1077
1078         * Scripts/run-gtk-tests:
1079         (TestRunner.__init__): Compile a regexp for color codes if stdout
1080         it not a TTY.
1081         (TestRunner._run_test_command.parse_line): Remove color codes and
1082         '\r' characters from the line returned by the pseudo-terminal if
1083         stdout is not a TTY.
1084         * gtk/common.py:
1085         (parse_output_lines): Fix the last chunk check to make sure we
1086         always send entire lines to the callback, otherwise color codes
1087         regexp might fail if the line is not complete.
1088
1089 2012-09-25  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
1090
1091         [DRT][WTR] Support overriding the 'WebKitDisplayImagesKey' preference
1092         https://bugs.webkit.org/show_bug.cgi?id=96883
1093
1094         Reviewed by Kenneth Rohde Christiansen.
1095
1096         * DumpRenderTree/efl/TestRunnerEfl.cpp:
1097         (TestRunner::overridePreference): Support the
1098         'WebKitDisplayImagesKey' preference.
1099
1100 2012-09-26  Zan Dobersek  <zandobersek@gmail.com>
1101
1102         [GTK] Enable some of the unstable CSS features
1103         https://bugs.webkit.org/show_bug.cgi?id=97572
1104
1105         Reviewed by Martin Robinson.
1106
1107         Enable the <style scoped> support via DumpRenderTreeSupportGtk method
1108         whenever resetting defualts to consistent values.
1109
1110         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1111         (resetDefaultsToConsistentValues):
1112
1113 2012-09-26  Zan Dobersek  <zandobersek@gmail.com>
1114
1115         REGRESSION (r126189): No more mismatch-did-not-occur failures in reftests
1116         https://bugs.webkit.org/show_bug.cgi?id=97112
1117
1118         Reviewed by Dirk Pranke.
1119
1120         Don't use the reftest's image hash when gathering output of its reference.
1121         This indirectly forces the driver to return the reference's image output
1122         as well. This specifically solves the problem when the reftest and its
1123         mismatch reference are equal but the unexpected match failure is not
1124         reported due to missing image output of the reference file.
1125
1126         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py:
1127         (SingleTestRunner._run_reftest):
1128         * Scripts/webkitpy/layout_tests/port/test.py:
1129         (TestPort.diff_image):
1130         (TestDriver.run_test):
1131
1132 2012-09-25  Ryosuke Niwa  <rniwa@webkit.org>
1133
1134         Update master.cfg to invoke run-perf-tests with --no-show-results and --reset-results
1135         https://bugs.webkit.org/show_bug.cgi?id=97613
1136
1137         Reviewed by Dirk Pranke.
1138
1139         Use options added in r129580.
1140
1141         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1142         (RunAndUploadPerfTests):
1143
1144 2012-09-25  Peter Rybin  <peter.rybin@gmail.com>
1145
1146         Web Inspector: Fix logging in pure protocol test harness
1147         https://bugs.webkit.org/show_bug.cgi?id=97579
1148
1149         Reviewed by Yury Semikhatsky.
1150
1151         JavaScript alert implementation in DumpRenderTree gets immediate flush to make sure that
1152         the message survives crash or kill by timeout.
1153
1154         * DumpRenderTree/blackberry/DumpRenderTree.cpp:
1155         (BlackBerry::WebKit::DumpRenderTree::runJavaScriptAlert):
1156         * DumpRenderTree/chromium/WebViewHost.cpp:
1157         (WebViewHost::runModalAlertDialog):
1158         * DumpRenderTree/efl/DumpRenderTreeView.cpp:
1159         (onJavaScriptAlert):
1160         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1161         (webViewScriptAlert):
1162         * DumpRenderTree/mac/UIDelegate.mm:
1163         (-[UIDelegate webView:runJavaScriptAlertPanelWithMessage:initiatedByFrame:]):
1164         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1165         (WebCore::WebPage::javaScriptAlert):
1166         * DumpRenderTree/win/UIDelegate.cpp:
1167         (UIDelegate::runJavaScriptAlertPanelWithMessage):
1168
1169 2012-09-25  Dan Bernstein  <mitz@apple.com>
1170
1171         Tools part of <rdar://problem/11455228> [mac] Stop using screen fonts
1172         https://bugs.webkit.org/show_bug.cgi?id=97620
1173
1174         Reviewed by John Sullivan.
1175
1176         * DumpRenderTree/mac/DumpRenderTree.mm:
1177         (resetDefaultsToConsistentValues): Enable screen font substitution when
1178         running the tests.
1179         * WebKitTestRunner/TestController.cpp:
1180         (WTR::TestController::resetStateToConsistentValues): Ditto.
1181
1182 2012-09-25  Ryosuke Niwa  <rniwa@webkit.org>
1183
1184         Build fix. "values" doesn't necessarily contain multiple values but it needs to be always parsed as an array.
1185
1186         * Scripts/webkitpy/performance_tests/perftest.py:
1187         (PerfTest.parse_output):
1188
1189 2012-09-25  Ryosuke Niwa  <rniwa@webkit.org>
1190
1191         run-perf-tests: cleanup options and results generation code
1192         https://bugs.webkit.org/show_bug.cgi?id=97611
1193
1194         Reviewed by Dirk Pranke.
1195
1196         Previously, --test-results-server triggered old JSOn format where the outermost structure was a dictionary
1197         instead of an array, and also implicitly triggered --no-show-results, caused the old outputs not to merge,
1198         and prevented the generation of results page. Also, it was not obvious that --source-json-path is an option
1199         used only on buildbot slaves.
1200
1201         This patch will:
1202         - Remove the old format since perf-o-matic supports new format now.
1203         - Add --reset-results option so that we can explicitly clear existing outputs.
1204         - Add --slave-config-json-path option to replace --source-json-path option.
1205
1206         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
1207         (PerfTestsRunner._parse_args): Added --reset-results and --slave-config-json-path options.
1208         (PerfTestsRunner._generate_and_show_results): Refactored. Also removed the code to strip "values" from
1209         results since perf-o-matic can parse and store these values now.
1210         (PerfTestsRunner._merge_outputs_if_needed): Renamed from _merge_outputs_if_needed.
1211         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
1212         (test_run_with_json_output): Test a harmless behavioral change to generate resuls page even when
1213         --test-results-server is present.
1214         (test_run_with_description):
1215         (test_run_generates_json_by_default):
1216         (test_run_merges_output_by_default): Added.
1217         (test_run_respects_reset_results): Added.
1218         (test_run_with_slave_config_json):
1219         (test_run_with_bad_slave_config_json): Use --slave-config-json-path instead of --source-json-path to make
1220         sure this optioon works as well.
1221         (test_run_with_multiple_repositories):
1222         (test_run_with_upload_json):
1223
1224 2012-09-25  Simon Fraser  <simon.fraser@apple.com>
1225
1226         Fix SnowLeopard build by adding #ifdefs.
1227         
1228         Fix crashes on some tests where the CFURLRef appears to be null.
1229
1230         * WebKitTestRunner/InjectedBundle/mac/InjectedBundlePageMac.mm:
1231         (WTR::testPathFromURL):
1232         (WTR::InjectedBundlePage::platformDidStartProvisionalLoadForFrame):
1233
1234 2012-09-25  Simon Fraser  <simon.fraser@apple.com>
1235
1236         Speculative build fix for Production builds.
1237
1238         * DumpRenderTree/mac/Configurations/Base.xcconfig:
1239         * WebKitTestRunner/Configurations/Base.xcconfig:
1240
1241 2012-09-25  Simon Fraser  <simon.fraser@apple.com>
1242
1243         Fix the build after r129566.
1244
1245         * DumpRenderTree/mac/FrameLoadDelegate.mm:
1246
1247 2012-09-25  Dominic Mazzoni  <dmazzoni@google.com>
1248
1249         AX: AccessibilityUIElementChromium.cpp won't build on Windows without wtf/StringExtras.h
1250         https://bugs.webkit.org/show_bug.cgi?id=97614
1251
1252         Unreviewed, fixing windows breakage.
1253
1254         Added wtf/StringExtras.h to the includes to fix the Chromium/Windows build.
1255
1256         * DumpRenderTree/chromium/TestRunner/AccessibilityUIElementChromium.cpp:
1257
1258 2012-09-24  Simon Fraser  <simon.fraser@apple.com>
1259
1260         <rdar://problem/12351906> Have DumpRenderTree and WebKitTestRunner crash logs show which test crashed
1261
1262         Reviewed by Mark Rowe.
1263
1264         Use a new WebKitSytemInterface function to add data to crash logs about which 
1265         test was running when the crash happened.
1266
1267         * DumpRenderTree/mac/Configurations/DebugRelease.xcconfig: Make sure WEBKIT_SYSTEM_INTERFACE_LIBRARY is defined.
1268         * DumpRenderTree/mac/Configurations/DumpRenderTree.xcconfig: link with the WebKitSystemInterface library.
1269         * DumpRenderTree/mac/FrameLoadDelegate.mm:
1270         (testPathFromURL): Given a (probably) file URL, return the part after "/LayoutTests/"
1271         (-[FrameLoadDelegate webView:didStartProvisionalLoadForFrame:]): Tell Crash Reporter which test was running,
1272         if we're the main frame and the test is active.
1273         * WebKitTestRunner/Configurations/DebugRelease.xcconfig: Make sure WEBKIT_SYSTEM_INTERFACE_LIBRARY is defined.
1274         * WebKitTestRunner/Configurations/InjectedBundle.xcconfig: link the injected bundle with the WebKitSystemInterface library.
1275         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1276         (WTR::InjectedBundlePage::didStartProvisionalLoadForFrame): Call platformDidStartProvisionalLoadForFrame().
1277         (WTR::InjectedBundlePage::platformDidStartProvisionalLoadForFrame): Stub for non-Mac platforms.
1278         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
1279         Add a platformDidStartProvisionalLoadForFrame() for platform-specific work.
1280         * WebKitTestRunner/InjectedBundle/mac/InjectedBundlePageMac.mm: Added.
1281         (WTR::testPathFromURL): Given a (probably) file URL, return the part after "/LayoutTests/"
1282         (WTR::InjectedBundlePage::platformDidStartProvisionalLoadForFrame): Tell Crash Reporter which test was running,
1283         if we're the main frame.
1284         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj: Add InjectedBundlePageMac.mm.
1285
1286 2012-09-25  Julie Parent  <jparent@chromium.org>
1287
1288         Make test_shell_tests show up in flakiness dashboard.
1289         https://bugs.webkit.org/show_bug.cgi?id=97585
1290
1291         Reviewed by Dirk Pranke.
1292
1293         test_shell_tests run only on chromium.org, and on the same bots as
1294         the layout tests, so do not use the generic fallbacks.
1295
1296         * TestResultServer/static-dashboards/builders.js:
1297         (loadBuildersList):
1298         * TestResultServer/static-dashboards/dashboard_base.js:
1299         (currentBuilderGroupCategory):
1300
1301 2012-09-25  Lucas Forschler  <lforschler@apple.com>
1302
1303         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.
1304         https://bugs.webkit.org/show_bug.cgi?id=97465
1305         
1306         Reviewed by Tim Horton.
1307
1308         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1309         (RunPythonTests.start):
1310
1311 2012-09-25  Mike West  <mkwst@chromium.org>
1312
1313         Unreviewed. Adding myself as a committer.
1314
1315         Yay. :)
1316
1317         * Scripts/webkitpy/common/config/committers.py:
1318
1319 2012-09-25  Beth Dakin  <bdakin@apple.com>
1320
1321         https://bugs.webkit.org/show_bug.cgi?id=95397
1322         Need to merge didFirstVisuallyNonEmptyLayout and 
1323         didNewFirstVisuallyNonEmptyLayout
1324         -and corresponding-
1325         <rdar://problem/10791680>
1326
1327         Reviewed by Sam Weinig.
1328
1329         Account for the new member of the PageLoadClient.
1330         * MiniBrowser/mac/BrowserWindowController.m:
1331         (-[BrowserWindowController awakeFromNib]):
1332         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1333         (WTR::InjectedBundlePage::InjectedBundlePage):
1334         * WebKitTestRunner/TestController.cpp:
1335         (WTR::TestController::initialize):
1336
1337         The NewFirstVisuallyNonEmptyLayoutForImages has existed for a while, 
1338         but it did not have the right target membership.
1339         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
1340
1341         Use new didLayout API instead of old 
1342         didNewFirstVisuallyNonEmptyLayout API.
1343         * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayout.cpp:
1344         (TestWebKitAPI):
1345         (TestWebKitAPI::didLayout):
1346         (TestWebKitAPI::setPageLoaderClient):
1347         (TestWebKitAPI::TEST):
1348         * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayoutFails.cpp:
1349         (TestWebKitAPI):
1350         (TestWebKitAPI::didLayout):
1351         (TestWebKitAPI::setPageLoaderClient):
1352         (TestWebKitAPI::TEST):
1353         * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayoutForImages.cpp:
1354         (TestWebKitAPI):
1355         (TestWebKitAPI::didLayout):
1356         (TestWebKitAPI::setPageLoaderClient):
1357         (TestWebKitAPI::TEST):
1358         * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayoutFrames.cpp:
1359         (TestWebKitAPI):
1360         (TestWebKitAPI::didLayout):
1361         (TestWebKitAPI::setPageLoaderClient):
1362         (TestWebKitAPI::TEST):
1363
1364 2012-09-25  Benjamin Poulain  <bpoulain@apple.com>
1365
1366         Add missing support for Geolocation tests on WebKit2
1367         https://bugs.webkit.org/show_bug.cgi?id=97518
1368
1369         Reviewed by Sam Weinig.
1370
1371         There are two main features added to the tooling to make the test run:
1372         -Add support for error reporting for position errors through setMockGeolocationPositionUnavailableError.
1373         -Support the 3 states permission: undefined-allowed-denied.
1374
1375         * WebKitTestRunner/GeolocationProviderMock.cpp:
1376         (WTR::GeolocationProviderMock::GeolocationProviderMock):
1377         (WTR::GeolocationProviderMock::setPosition):
1378         (WTR::GeolocationProviderMock::setPositionUnavailableError):
1379         (WTR::GeolocationProviderMock::startUpdating):
1380         (WTR::GeolocationProviderMock::sendPositionIfNeeded):
1381         (WTR::GeolocationProviderMock::sendErrorIfNeeded):
1382         * WebKitTestRunner/GeolocationProviderMock.h:
1383         (GeolocationProviderMock):
1384         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
1385         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
1386         (WTR::InjectedBundle::setMockGeolocationPositionUnavailableError):
1387         * WebKitTestRunner/InjectedBundle/InjectedBundle.h:
1388         (InjectedBundle):
1389         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
1390         (WTR::TestRunner::setMockGeolocationPositionUnavailableError):
1391         * WebKitTestRunner/InjectedBundle/TestRunner.h:
1392         (TestRunner):
1393         * WebKitTestRunner/TestController.cpp:
1394         (WTR::TestController::TestController):
1395         (WTR::decidePolicyForGeolocationPermissionRequest):
1396         (WTR::TestController::resetStateToConsistentValues):
1397         (WTR::TestController::setGeolocationPermission):
1398         (WTR::TestController::setMockGeolocationPosition):
1399         (WTR::TestController::setMockGeolocationPositionUnavailableError):
1400         (WTR::TestController::handleGeolocationPermissionRequest):
1401         (WTR::TestController::decidePolicyForGeolocationPermissionRequestIfPossible):
1402         * WebKitTestRunner/TestController.h:
1403         (TestController):
1404         * WebKitTestRunner/TestInvocation.cpp:
1405         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
1406
1407 2012-09-25  Sheriff Bot  <webkit.review.bot@gmail.com>
1408
1409         Unreviewed, rolling out r129517.
1410         http://trac.webkit.org/changeset/129517
1411         https://bugs.webkit.org/show_bug.cgi?id=97582
1412
1413         Link errors in chromium (Requested by alecf on #webkit).
1414
1415         * DumpRenderTree/DumpRenderTree.gypi:
1416         * DumpRenderTree/chromium/MockConstraints.cpp: Removed.
1417         * DumpRenderTree/chromium/MockConstraints.h: Removed.
1418         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
1419         (RTCSessionDescriptionRequestSuccededTask::RTCSessionDescriptionRequestSuccededTask):
1420         (RTCSessionDescriptionRequestFailedTask::RTCSessionDescriptionRequestFailedTask):
1421         (RTCVoidRequestTask::RTCVoidRequestTask):
1422         (isSupportedConstraint):
1423         (isValidConstraint):
1424         (MockWebRTCPeerConnectionHandler::initialize):
1425         * DumpRenderTree/chromium/WebUserMediaClientMock.cpp:
1426         (WebKit::WebUserMediaClientMock::requestUserMedia):
1427         * DumpRenderTree/chromium/WebUserMediaClientMock.h:
1428         (WebKit):
1429         (WebUserMediaClientMock):
1430         * DumpRenderTree/chromium/WebViewHost.h:
1431         (WebKit):
1432         (WebViewHost):
1433
1434 2012-09-25  Dominic Mazzoni  <dmazzoni@google.com>
1435
1436         AX: Chromium doesn't pass accessibility text range and line number tests
1437         https://bugs.webkit.org/show_bug.cgi?id=97246
1438
1439         Reviewed by Chris Fleizach.
1440
1441         Provide a chromium implementation of some AccessibilityUIElement
1442         methods needed by accessibility tests.
1443
1444         Enables four existing tests to pass on Chromium:
1445         Tests: accessibility/textarea-insertion-point-line-number.html
1446                accessibility/textarea-line-for-index.html
1447                accessibility/textarea-selected-text-range.html
1448                accessibility/textbox-role-reports-selection.html
1449
1450         * DumpRenderTree/chromium/TestRunner/AccessibilityUIElementChromium.cpp:
1451         (AccessibilityUIElement::insertionPointLineNumberGetterCallback):
1452         (AccessibilityUIElement::selectedTextRangeGetterCallback):
1453         (AccessibilityUIElement::lineForIndexCallback):
1454         (AccessibilityUIElement::setSelectedTextRangeCallback):
1455
1456 2012-09-25  Peter Beverloo  <peter@chromium.org>
1457
1458         [Chromium] Android's build archives have grown from 38M to 102M in 11 days
1459         https://bugs.webkit.org/show_bug.cgi?id=97336
1460
1461         Reviewed by Adam Barth.
1462
1463         Each WebKit target results in four APKs in the out/{Debug,Release}/ directory,
1464         while we really only need one. Ignoring the unsigned APKs removes six from
1465         the archive will make sure that we only archive a single APK per target,
1466         bringing its size back to rougly 40 megabytes.
1467
1468         * BuildSlaveSupport/built-product-archive:
1469         (archiveBuiltProduct):
1470
1471 2012-09-25  Tommy Widenflycht  <tommyw@google.com>
1472
1473         MediaStream API: Update getUserMedia to match the latest specification
1474         https://bugs.webkit.org/show_bug.cgi?id=97540
1475
1476         Reviewed by Adam Barth.
1477
1478         Adding mock support for the enhanced getUserMedia call.
1479
1480         * DumpRenderTree/DumpRenderTree.gypi:
1481         * DumpRenderTree/chromium/MockConstraints.cpp: Copied from Source/WebCore/platform/chromium/support/WebMediaConstraints.cpp.
1482         (MockConstraints):
1483         (MockConstraints::isSupported):
1484         (MockConstraints::isValid):
1485         (MockConstraints::verify):
1486         * DumpRenderTree/chromium/MockConstraints.h: Copied from Tools/DumpRenderTree/chromium/WebUserMediaClientMock.h.
1487         (WebKit):
1488         (MockConstraints):
1489         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
1490         (RTCSessionDescriptionRequestSuccededTask::RTCSessionDescriptionRequestSuccededTask):
1491         (RTCSessionDescriptionRequestFailedTask::RTCSessionDescriptionRequestFailedTask):
1492         (RTCVoidRequestTask::RTCVoidRequestTask):
1493         (MockWebRTCPeerConnectionHandler::initialize):
1494         * DumpRenderTree/chromium/WebUserMediaClientMock.cpp:
1495         (UserMediaRequestTask):
1496         (UserMediaRequestTask::UserMediaRequestTask):
1497         (WebUserMediaClientMock::requestUserMedia):
1498         * DumpRenderTree/chromium/WebUserMediaClientMock.h:
1499         (WebUserMediaClientMock):
1500         (WebUserMediaClientMock::taskList):
1501         * DumpRenderTree/chromium/WebViewHost.h:
1502         (WebKit):
1503         (WebViewHost):
1504
1505 2012-09-25  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
1506
1507         [EFL] Do not dump inspector output in DRT
1508         https://bugs.webkit.org/show_bug.cgi?id=97018
1509
1510         Reviewed by Kenneth Rohde Christiansen.
1511
1512         Some inspector-related layout tests expect console output from the
1513         inspector view to be suppressed. Do that by setting a property in
1514         the inspector view that makes the onConsoleMessage callback return
1515         early and not output anything.
1516
1517         Using evas_object_data_{get,set} is not very beautiful, but it is
1518         still better than properly subclassing Ewk_View and having to
1519         implement something similar to drt_view_console_message_enable_set().
1520
1521         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
1522         (DumpRenderTreeChrome::createWebInspectorView):
1523         * DumpRenderTree/efl/DumpRenderTreeView.cpp:
1524         (onConsoleMessage):
1525
1526 2012-09-25  Christophe Dumez  <christophe.dumez@intel.com>
1527
1528         [WK2][WKTR] TestRunner needs support for setWillSendRequestReturnsNullOnRedirect
1529         https://bugs.webkit.org/show_bug.cgi?id=97542
1530
1531         Reviewed by Kenneth Rohde Christiansen.
1532
1533         Implement support for setWillSendRequestReturnsNullOnRedirect
1534         in WebKitTestRunner.
1535
1536         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
1537         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1538         (WTR::InjectedBundlePage::willSendRequestForFrame):
1539         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
1540         (WTR::TestRunner::TestRunner):
1541         * WebKitTestRunner/InjectedBundle/TestRunner.h:
1542         (WTR::TestRunner::willSendRequestReturnsNullOnRedirect):
1543         (WTR::TestRunner::setWillSendRequestReturnsNullOnRedirect):
1544         (TestRunner):
1545
1546 2012-09-25  Xiaobo Wang  <xbwang@torchmobile.com.cn>
1547
1548         [BlackBerry] [DRT] fast/forms/textarea-live-pseudo-selectors.html failed
1549         https://bugs.webkit.org/show_bug.cgi?id=97189
1550
1551         Reviewed by Rob Buis.
1552
1553         Translate ASCII key code to BlackBerry key code for BACKSPACE and DELETE key.
1554
1555         * DumpRenderTree/blackberry/EventSender.cpp:
1556         (keyDownCallback):
1557
1558 2012-09-25  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
1559
1560         [WK2][WTR] WTR bundle client loads binary data as text
1561         https://bugs.webkit.org/show_bug.cgi?id=97532
1562
1563         Reviewed by Kenneth Rohde Christiansen.
1564
1565         WTR::InjectedBundlePage::decidePolicyForResponse now checks response MIME type
1566         and returns WKBundlePagePolicyActionPassThrough if it cannot be shown.
1567
1568         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1569         (WTR::InjectedBundlePage::decidePolicyForResponse):
1570
1571 2012-09-25  Zan Dobersek  <zandobersek@gmail.com>
1572
1573         Unreviewed, another test case adjustment after r129459. Python 2.7 doesn't
1574         seem to be a requirement yet so the assertDictEqual method is not yet available
1575         on all builders. Assert that the empty dictionary length equals 0 instead.
1576
1577         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
1578         (ResultSummaryTest.test_interpret_test_failures):
1579
1580 2012-09-24  Simon Hausmann  <simon.hausmann@digia.com>
1581
1582         [Qt] Fix build with latest Qt 5
1583         https://bugs.webkit.org/show_bug.cgi?id=97479
1584
1585         Reviewed by Tor Arne Vestbø.
1586
1587         The variables for syncqt handling are now set in api.pri, but let's add a comment here that we can
1588         remove them from default_pre.prf once we're past Qt 5 beta2.
1589
1590         * qmake/mkspecs/features/default_pre.prf:
1591
1592 2012-09-25  Zan Dobersek  <zandobersek@gmail.com>
1593
1594         Unreviewed, adjusting the test cases after r129459. The JSON data
1595         for reftests now contains the 'reftest_type' key and no more 'is_reftest'
1596         or 'is_mismatch_reftest' keys.
1597
1598         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1599         (EndToEndTest.test_reftest_with_two_notrefs):
1600
1601 2012-09-24  Zan Dobersek  <zandobersek@gmail.com>
1602
1603         Unexpected reftest passes are only reported when pixel testing is enabled in results.html as well
1604         https://bugs.webkit.org/show_bug.cgi?id=97426
1605
1606         Reviewed by Ojan Vafai.
1607
1608         The 'is_reftest' member of the TestResult class has been replaced by the reftest_type
1609         member which is a set containing either '!=', '==', both or none if the test represented
1610         by the object is not a reftest.
1611
1612         When summarizing results, the test dictionary is updated to contain the 'reftest_type' key
1613         if the test is a reftest. The value of this key is a list of all the reftest types of this test.
1614
1615         The test failure interpretation method has been refactored, it now reports only
1616         missing results of various formats and image diff percent in case of image or
1617         reftest mismatch.
1618
1619         Unit tests have been modified accordingly.
1620
1621         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner.py:
1622         (LayoutTestRunner._update_summary_with_result):
1623         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner_unittest.py:
1624         (LayoutTestRunnerTests.test_update_summary_with_result):
1625         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1626         (interpret_test_failures):
1627         (summarize_results):
1628         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
1629         (ResultSummaryTest.test_interpret_test_failures):
1630         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py:
1631         (SingleTestRunner.run):
1632         (SingleTestRunner._run_reftest):
1633         * Scripts/webkitpy/layout_tests/models/test_results.py:
1634         (TestResult.__init__):
1635         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1636         (EndToEndTest.test_reftest_with_two_notrefs):
1637
1638 2012-09-24  Sam Weinig  <sam@webkit.org>
1639
1640         Use NSUserDefaults rather than an environment variable to control whether to use an XPC Service for the WebProcess
1641         https://bugs.webkit.org/show_bug.cgi?id=97514
1642
1643         Reviewed by Anders Carlsson.
1644
1645         * Scripts/webkitdirs.pm:
1646         (argumentsForRunAndDebugMacWebKitApp):
1647         Push "-WebKit2UseXPCServiceForWebProcess YES" as arguments when --use-web-process-xpc-service
1648         is provided to set the user default. Move adding of @ARGV to after custom additions as calling
1649         shouldUseXPCServiceForWebProcess() can change @ARGV.
1650
1651         (runMacWebKitApp):
1652         (execMacWebKitAppForDebugging):
1653         Stop setting the WEBKIT_USE_XPC_SERVICE_FOR_WEB_PROCESS environment variable.
1654
1655 2012-09-24  Laszlo Gombos  <l.gombos@samsung.com>
1656
1657         [GTK][EFL] Make sure cairo is built with fontconfig support
1658         https://bugs.webkit.org/show_bug.cgi?id=97513
1659
1660         Reviewed by Martin Robinson.
1661
1662         The GTK and EFL port of WebKit requires a version of cairo that is built with fontconfig support. 
1663
1664         * efl/jhbuild.modules:
1665         * gtk/jhbuild.modules:
1666
1667 2012-09-24  Benjamin Poulain  <bpoulain@apple.com>
1668
1669         Fix Geolocation error reporting in the test support
1670         https://bugs.webkit.org/show_bug.cgi?id=97386
1671
1672         Reviewed by Sam Weinig.
1673         The TestRunner was exposing internal implementation details of
1674         the Geolocation object through setMockGeolocationError().
1675
1676         In practice, only the error code PositionUnavailable is supported
1677         or even implemented.
1678
1679         Exposing setMockGeolocationError() does not make sense for WebKit2 as
1680         passing generic errors from the UIProcess is a bad idea. This is the
1681         opportunity to clean TestRunner and prepare and interface suitable for WebKit2.
1682
1683         * DumpRenderTree/TestRunner.cpp:
1684         (setMockGeolocationPositionUnavailableErrorCallback):
1685         (TestRunner::staticFunctions):
1686         * DumpRenderTree/TestRunner.h:
1687         (TestRunner):
1688         * DumpRenderTree/blackberry/TestRunnerBlackBerry.cpp:
1689         (TestRunner::setMockGeolocationPositionUnavailableError):
1690         * DumpRenderTree/chromium/DRTTestRunner.cpp:
1691         (DRTTestRunner::DRTTestRunner):
1692         (DRTTestRunner::setMockGeolocationPositionUnavailableError):
1693         * DumpRenderTree/chromium/DRTTestRunner.h:
1694         (DRTTestRunner):
1695         * DumpRenderTree/efl/TestRunnerEfl.cpp:
1696         (TestRunner::setMockGeolocationPositionUnavailableError):
1697         * DumpRenderTree/gtk/TestRunnerGtk.cpp:
1698         (TestRunner::setMockGeolocationPositionUnavailableError):
1699         * DumpRenderTree/mac/MockGeolocationProvider.h:
1700         * DumpRenderTree/mac/MockGeolocationProvider.mm:
1701         (-[MockGeolocationProvider dealloc]):
1702         (-[MockGeolocationProvider resetError]):
1703         (-[MockGeolocationProvider setPosition:]):
1704         (-[MockGeolocationProvider setPositionUnavailableErrorWithMessage:]):
1705         (-[MockGeolocationProvider lastPosition]):
1706         (-[MockGeolocationProvider timerFired]):
1707         * DumpRenderTree/mac/TestRunnerMac.mm:
1708         (TestRunner::setMockGeolocationPositionUnavailableError):
1709         * DumpRenderTree/qt/TestRunnerQt.cpp:
1710         (TestRunner::setMockGeolocationPositionUnavailableError):
1711         * DumpRenderTree/qt/TestRunnerQt.h:
1712         (TestRunner):
1713         * DumpRenderTree/win/TestRunnerWin.cpp:
1714         (TestRunner::setMockGeolocationPositionUnavailableError):
1715         * DumpRenderTree/wx/TestRunnerWx.cpp:
1716         (TestRunner::setMockGeolocationPositionUnavailableError):
1717
1718 2012-09-24  Nikhil Bansal  <n.bansal@samsung.com>
1719
1720         [EFL][WK2] TestRunner needs touch events support.
1721         https://bugs.webkit.org/show_bug.cgi?id=96465
1722
1723         Reviewed by Kenneth Rohde Christiansen.
1724
1725         Simulate touch events in WebkitTestRunner for touch related
1726         layout tests. This patch uses ewk_view_feed_touch_event() to
1727         generate touch events.
1728
1729         * WebKitTestRunner/EventSenderProxy.h:
1730         (EventSenderProxy):
1731         * WebKitTestRunner/efl/EventSenderProxyEfl.cpp:
1732         (WTR):
1733         (WTR::setEvasModifiers):
1734         (WTR::EventSenderProxy::EventSenderProxy):
1735         (WTR::EventSenderProxy::~EventSenderProxy):
1736         (WTR::EventSenderProxy::sendTouchEvent):
1737         (WTR::EventSenderProxy::addTouchPoint):
1738         (WTR::EventSenderProxy::updateTouchPoint):
1739         (WTR::EventSenderProxy::setTouchModifier):
1740         (WTR::EventSenderProxy::touchStart):
1741         (WTR::EventSenderProxy::touchMove):
1742         (WTR::EventSenderProxy::touchEnd):
1743         (WTR::EventSenderProxy::touchCancel):
1744         (WTR::EventSenderProxy::clearTouchPoints):
1745         (WTR::EventSenderProxy::releaseTouchPoint):
1746         (WTR::EventSenderProxy::cancelTouchPoint):
1747
1748 2012-09-24  Christophe Dumez  <christophe.dumez@intel.com>
1749
1750         [WK2] TestRunner does not support overridePreference with value "0"
1751         https://bugs.webkit.org/show_bug.cgi?id=97457
1752
1753         Reviewed by Alexey Proskuryakov.
1754
1755         testRunner.overridePreference() was marked as taking a boolean
1756         in argument in testRunner.idl even though the test cases are
1757         passing strings to it. The current code works if the test
1758         case wants to enable a setting and therefore passes "1" string.
1759         However, if the test wants to disable a setting and passes "0"
1760         string, the current code also evaluates this to true causing
1761         certain test cases to fail.
1762
1763         The argument to testRunner.overridePreference() is now of string
1764         type and a function has been added to convert the string passed
1765         by the test to a boolean value.
1766
1767         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
1768         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
1769         (WTR::toBool):
1770         (WTR):
1771         (WTR::TestRunner::overridePreference):
1772         * WebKitTestRunner/InjectedBundle/TestRunner.h:
1773         (TestRunner):
1774
1775 2012-09-24  Sheriff Bot  <webkit.review.bot@gmail.com>
1776
1777         Unreviewed, rolling out r129343.
1778         http://trac.webkit.org/changeset/129343
1779         https://bugs.webkit.org/show_bug.cgi?id=97479
1780
1781         It broke the build (Requested by Ossy_NIGHT on #webkit).
1782
1783         * qmake/mkspecs/features/default_pre.prf:
1784
1785 2012-09-24  Zan Dobersek  <zandobersek@gmail.com>
1786
1787         [Gtk] Unskip the Shadow DOM layout tests
1788         https://bugs.webkit.org/show_bug.cgi?id=90776
1789
1790         Reviewed by Ryosuke Niwa.
1791
1792         Enable Shadow DOM through DumpRenderTreeSupport when resetting defaults in between tests.
1793
1794         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1795         (resetDefaultsToConsistentValues):
1796
1797 2012-09-24  Zan Dobersek  <zandobersek@gmail.com>
1798
1799         REGRESSION (r129211-r129218): http/tests/loading/redirect-with-no-location-crash.html failing on GTK Linux 64-bit Release WK2 (Tests)
1800         https://bugs.webkit.org/show_bug.cgi?id=97404
1801
1802         Reviewed by Philippe Normand.
1803
1804         Print out the 'didReceiveTitle' information when dumping frame load callbacks.
1805
1806         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1807         (webViewTitleChanged):
1808
1809 2012-09-24  Simon Hausmann  <simon.hausmann@digia.com>
1810
1811         [Qt] Fix build with latest Qt 5
1812
1813         Reviewed by Tor Arne Vestbø.
1814
1815         The variables for syncqt handling are now set in api.pri, but let's add a comment here that we can
1816         remove them from default_pre.prf once we're past Qt 5 beta2.
1817
1818         * qmake/mkspecs/features/default_pre.prf:
1819
1820 2012-09-22  Dominic Mazzoni  <dmazzoni@google.com>
1821
1822         AX: Layout tests would be easier to write if AccessibilityController could find an element by id
1823         https://bugs.webkit.org/show_bug.cgi?id=97301
1824
1825         Reviewed by Chris Fleizach.
1826
1827         Adds a new accessibleElementById method to AccessibilityController, to make it
1828         easier to write cross-platform layout tests.
1829
1830         Implemented for Chromium and for Mac (DRT and WKTR).
1831
1832         * DumpRenderTree/AccessibilityController.cpp:
1833         (getAccessibleElementByIdCallback):
1834         (AccessibilityController::getJSClass):
1835         * DumpRenderTree/AccessibilityController.h:
1836         (AccessibilityController):
1837         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
1838         * DumpRenderTree/chromium/TestRunner/AccessibilityControllerChromium.cpp:
1839         (AccessibilityController::AccessibilityController):
1840         (AccessibilityController::findAccessibleElementByIdRecursive):
1841         (AccessibilityController::getAccessibleElementById):
1842         (AccessibilityController::accessibleElementByIdGetterCallback):
1843         * DumpRenderTree/chromium/TestRunner/AccessibilityControllerChromium.h:
1844         (AccessibilityController):
1845         * DumpRenderTree/gtk/AccessibilityControllerGtk.cpp:
1846         (AccessibilityController::accessibleElementById):
1847         * DumpRenderTree/mac/AccessibilityCommonMac.h: added
1848         * DumpRenderTree/mac/AccessibilityCommonMac.mm: added
1849         (+[NSString stringWithJSStringRef:]):
1850         (-[NSString createJSStringRef]):
1851         * DumpRenderTree/mac/AccessibilityControllerMac.mm:
1852         (findAccessibleObjectById):
1853         (AccessibilityController::accessibleElementById):
1854         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
1855         * DumpRenderTree/win/AccessibilityControllerWin.cpp:
1856         (AccessibilityController::accessibleElementById):
1857         * WebKitTestRunner/InjectedBundle/AccessibilityController.cpp:
1858         (WTR::AccessibilityController::accessibleElementById):
1859         * WebKitTestRunner/InjectedBundle/AccessibilityController.h:
1860         (AccessibilityController):
1861         * WebKitTestRunner/InjectedBundle/Bindings/AccessibilityController.idl:
1862         * WebKitTestRunner/InjectedBundle/mac/AccessibilityCommonMac.h: added
1863         * WebKitTestRunner/InjectedBundle/mac/AccessibilityCommonMac.mm: added
1864         (+[NSString stringWithJSStringRef:]):
1865         (-[NSString createJSStringRef]):
1866         * WebKitTestRunner/InjectedBundle/mac/AccessibilityControllerMac.mm:
1867         (WTR::findAccessibleObjectById):
1868         (WTR):
1869         (WTR::AccessibilityController::accessibleElementById):
1870         * WebKitTestRunner/InjectedBundle/mac/AccessibilityUIElementMac.mm:
1871         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
1872
1873 2012-09-22  Zan Dobersek  <zandobersek@gmail.com>
1874
1875         Unexpected reftest passes are only reported when pixel testing is enabled
1876         https://bugs.webkit.org/show_bug.cgi?id=97242
1877
1878         Reviewed by Dirk Pranke.
1879
1880         LayoutTestRunner should know whether the finished test it's handling is
1881         a reftest. This is necessary when updating result summary for reftests
1882         when pixel testing is not enabled - in these circumstances an unexpectedly
1883         passing reftest was not reported as such due to pixel testing then being
1884         determined by searching for reftest failures in failures list, which were
1885         not found in an empty failure list of a passing test. Now the TestResult
1886         objects have a member variable indicating the test is a reftest whenever
1887         that's the case.
1888
1889         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner.py:
1890         (LayoutTestRunner._update_summary_with_result):
1891         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner_unittest.py:
1892         (LayoutTestRunnerTests.test_update_summary_with_result):
1893         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py:
1894         (SingleTestRunner.run):
1895         (SingleTestRunner._run_reftest):
1896         * Scripts/webkitpy/layout_tests/models/test_results.py:
1897         (TestResult.__init__):
1898
1899 2012-09-21  Sam Weinig  <sam@webkit.org>
1900
1901         WebProcess XPC services need have their environment set without disrupting all other XPC services
1902         https://bugs.webkit.org/show_bug.cgi?id=95161
1903
1904         Reviewed by Anders Carlsson.
1905
1906         Remove setting the __XPC_* environment variables.
1907
1908         * Scripts/webkitdirs.pm:
1909         (setUpGuardMallocIfNeeded):
1910         (runMacWebKitApp):
1911         (execMacWebKitAppForDebugging):
1912
1913 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
1914
1915         Fix typo in additional-platform-directory patch just landed
1916         https://bugs.webkit.org/show_bug.cgi?id=97380
1917         
1918         Unreviewed, build fix.
1919
1920         * Scripts/webkitpy/layout_tests/port/base.py:
1921         (Port.relative_test_filename):
1922         (Port.relative_perf_test_filename):
1923
1924 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
1925
1926         nrwt: don't require additional-platform-directory to be an abspath or live under LayoutTests
1927         https://bugs.webkit.org/show_bug.cgi?id=97380
1928
1929         Reviewed by Ojan Vafai.
1930
1931         There doesn't seem to be a good reason for this restriction and
1932         it's useful to be able to point to directories outside the
1933         checkout for results (e.g., for local failures due to a 10.7.4
1934         install ;).
1935
1936         * Scripts/webkitpy/layout_tests/port/base.py:
1937         (Port.relative_test_filename):
1938         (Port.relative_perf_test_filename):
1939         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
1940         (ChromiumAndroidDriver._command_from_driver_input):
1941         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1942         (_set_up_derived_options):
1943         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1944         (MainTest.test_additional_platform_directory):
1945
1946 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
1947
1948         webkitpy: drop support for old TestExpectations syntax
1949         https://bugs.webkit.org/show_bug.cgi?id=97364
1950
1951         Reviewed by Ryosuke Niwa.
1952
1953         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
1954         (TestExpectationParser._collect_matching_tests):
1955         (TestExpectationParser):
1956         (TestExpectationParser._tokenize_line):
1957         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
1958         (SkippedTests.test_skipped_entry_dont_exist):
1959         (ExpectationSyntaxTests.assert_tokenize_exp):
1960         (ExpectationSyntaxTests.test_bare_name):
1961         (ExpectationSyntaxTests.test_bare_name_and_bugs):
1962         (ExpectationSyntaxTests.test_comments):
1963         (ExpectationSyntaxTests.test_config_modifiers):
1964         (ExpectationSyntaxTests.test_unknown_config):
1965         (ExpectationSyntaxTests.test_unknown_expectation):
1966         (ExpectationSyntaxTests.test_skip):
1967         (ExpectationSyntaxTests.test_slow):
1968         (ExpectationSyntaxTests.test_wontfix):
1969         (ExpectationSyntaxTests.test_blank_line):
1970         (ExpectationSyntaxTests.test_warnings):
1971         (RebaseliningTest.test_no_get_rebaselining_failures):
1972
1973 2012-09-21  Simon Fraser  <simon.fraser@apple.com>
1974
1975         Improve WTR unresponsiveness output a little
1976         https://bugs.webkit.org/show_bug.cgi?id=97370
1977
1978         Reviewed by Timothy Horton.
1979
1980         Distinguish between conditions that already set the errorMessage,
1981         and unresponsiveness due to slow about:blank loads when WTR
1982         reports unresponsiveness.
1983
1984         * WebKitTestRunner/TestInvocation.cpp:
1985         (WTR::TestInvocation::invoke):
1986
1987 2012-09-21  Chris Rogers  <crogers@google.com>
1988
1989         Add Web Audio support for deprecated/legacy APIs
1990         https://bugs.webkit.org/show_bug.cgi?id=97050
1991
1992         Reviewed by Eric Carlson.
1993
1994         * Scripts/webkitperl/FeatureList.pm:
1995
1996 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
1997
1998         webkitpy: update remaining tests to use the new expectation syntax
1999         https://bugs.webkit.org/show_bug.cgi?id=97362
2000
2001         Reviewed by Ojan Vafai.
2002
2003         This patch updates all the unit tests that were still using the
2004         old TestExpectations syntax to use the new syntax *except* for
2005         the tests that were specifically testing that we parsed the old
2006         syntax correctly.
2007
2008         Also, a block of tests for the new syntax were duplicated, so
2009         I've deleted the duplicate.
2010
2011         Note that the old syntax is still supported so this change should
2012         produce no visible changes.
2013
2014         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2015         (MiscTests.test_multiple_results):
2016         (MiscTests.test_category_expectations):
2017         (MiscTests.test_error_on_different_platform):
2018         (MiscTests.test_error_on_different_build_type):
2019         (MiscTests.test_overrides):
2020         (MiscTests.test_overrides__directory):
2021         (MiscTests.test_overrides__duplicate):
2022         (MiscTests.test_more_specific_override_resets_skip):
2023         (SkippedTests.check):
2024         (SkippedTests.test_duplicate_skipped_test_fails_lint):
2025         (SkippedTests.test_skipped_file_overrides_expectations):
2026         (SkippedTests.test_skipped_dir_overrides_expectations):
2027         (SkippedTests.test_skipped_file_overrides_overrides):
2028         (SkippedTests.test_skipped_dir_overrides_overrides):
2029         (ExpectationSyntaxTests.disabled_test_missing_expectation):
2030         (ExpectationSyntaxTests.disabled_test_missing_colon):
2031         (ExpectationSyntaxTests.disabled_test_too_many_colons):
2032         (ExpectationSyntaxTests.disabled_test_too_many_equals_signs):
2033         (ExpectationSyntaxTests):
2034         (ExpectationSyntaxTests.test_unrecognized_expectation):
2035         (ExpectationSyntaxTests.test_macro):
2036         (SemanticTests.test_bug_format):
2037         (SemanticTests.test_bad_bugid):
2038         (SemanticTests.test_missing_bugid):
2039         (SemanticTests.test_slow_and_timeout):
2040         (SemanticTests.test_rebaseline):
2041         (test_missing_file):
2042         (test_ambiguous):
2043         (test_more_modifiers):
2044         (test_order_in_file):
2045         (test_macro_overrides):
2046         (OldExpectationParserTests):
2047         (OldExpectationParserTests._tokenize):
2048         (OldExpectationParserTests.test_tokenize_extra_colon):
2049         (OldExpectationParserTests.test_tokenize_missing_equal):
2050         (OldExpectationParserTests.test_tokenize_extra_equal):
2051         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2052         (MainTest.test_skip_failing_tests):
2053         (MainTest.test_additional_expectations):
2054         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2055         (TestExpectationsTestCase.test_valid_expectations):
2056         (TestExpectationsTestCase.test_invalid_expectations):
2057         (TestExpectationsTestCase.test_tab):
2058
2059 2012-09-21  Benjamin Poulain  <bpoulain@apple.com>
2060
2061         [WK2] Add basic testing support for Geolocation
2062         https://bugs.webkit.org/show_bug.cgi?id=97278
2063
2064         Reviewed by Kenneth Rohde Christiansen.
2065
2066         Update the Geolocation testing to use the proper API in the UIProcess.
2067
2068         * WebKitTestRunner/CMakeLists.txt:
2069         * WebKitTestRunner/GNUmakefile.am:
2070
2071         * WebKitTestRunner/GeolocationProviderMock.cpp: Added.
2072         (WTR::startUpdatingCallback):
2073         (WTR::stopUpdatingCallback):
2074         (WTR::GeolocationProviderMock::GeolocationProvierMock):
2075         (WTR::GeolocationProviderMock::setMockGeolocationPosition):
2076         (WTR::GeolocationProviderMock::startUpdating):
2077         (WTR::GeolocationProviderMock::stopUpdating):
2078         (GeolocationProviderMock):
2079         The GeolocationProvider store the location update and deliver them as needed.
2080
2081         WebCore GeolocationController do not support asynchronous update on start/stop. This is not
2082         a problem in this case because all the messages between the WebProcess and the UIProcess are
2083         asynchronous. Because of this, unlike GeolocationClientMock, we do not use a timer for event
2084         delivery.
2085
2086         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
2087         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
2088         (WTR::InjectedBundle::setGeolocationPermission):
2089         (WTR::InjectedBundle::setMockGeolocationPosition):
2090         * WebKitTestRunner/InjectedBundle/InjectedBundle.h:
2091         (InjectedBundle):
2092         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
2093         (WTR::TestRunner::setGeolocationPermission):
2094         (WTR::TestRunner::setMockGeolocationPosition):
2095         From the InjectedBundle, we now pass the information to the UIProcess so that
2096         GeolocationProvider and the TestController can respond appropriately.
2097
2098         * WebKitTestRunner/InjectedBundle/TestRunner.h:
2099         (TestRunner):
2100         * WebKitTestRunner/Target.pri:
2101         * WebKitTestRunner/TestController.cpp:
2102         (WTR::TestController::TestController):
2103         (WTR::decidePolicyForGeolocationPermissionRequest):
2104         (WTR::TestController::createOtherPage):
2105         (WTR::TestController::initialize):
2106         (WTR::TestController::setMockGeolocationPosition):
2107         * WebKitTestRunner/TestController.h:
2108         (TestController):
2109         (WTR::TestController::setGeolocationPermission):
2110         (WTR::TestController::isGeolocationPermissionAllowed):
2111         * WebKitTestRunner/TestInvocation.cpp:
2112         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
2113         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
2114         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
2115
2116 2012-09-21  Joanmarie Diggs  <jdiggs@igalia.com>
2117
2118         REGRESSION (r127882): accessibility/spinbutton-value.html failing on GTK
2119         https://bugs.webkit.org/show_bug.cgi?id=96196
2120
2121         Reviewed by Martin Robinson.
2122
2123         The "regression" is that a new test was added but the support was missing
2124         in the Gtk port for spin buttons.
2125
2126         * DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
2127         (AccessibilityUIElement::valueDescription): Updated the FIXME comment to
2128         indicate that this cannot be implemented until it is implemented in ATK.
2129         URL of the newly-filed ATK bug included for reference.
2130
2131 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
2132
2133         Fix test_skip_and_wontfix failure
2134         https://bugs.webkit.org/show_bug.cgi?id=97225
2135
2136         Unreviewed, build fix.
2137
2138         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2139         (SemanticTests.test_skip_and_wontfix):
2140
2141 2012-09-21  Carlos Garcia Campos  <cgarcia@igalia.com>
2142
2143         [GTK] Implement ViewState methods in PageClientImpl in WebKit2
2144         https://bugs.webkit.org/show_bug.cgi?id=97202
2145
2146         Reviewed by Martin Robinson.
2147
2148         * WebKitTestRunner/gtk/PlatformWebViewGtk.cpp:
2149         (WTR::PlatformWebView::focus): Focus the view.
2150
2151 2012-09-21  Sudarsana Nagineni  <sudarsana.nagineni@intel.com>
2152
2153         [WTR] Memory leaks in InjectedBundleController::initialize()
2154         https://bugs.webkit.org/show_bug.cgi?id=97329
2155
2156         Reviewed by Alexey Proskuryakov.
2157
2158         Fix leaks in InjectedBundleController::initialize() by adopting
2159         strings created with WKStringCreateWithUTF8CString().
2160
2161         * TestWebKitAPI/InjectedBundleController.cpp:
2162         (TestWebKitAPI::InjectedBundleController::initialize):
2163
2164 2012-09-21  Marcelo Lira  <marcelo.lira@openbossa.org>
2165
2166         [Qt][WK2] Removed duplicated code from EventSenderProxy::keyDown
2167         https://bugs.webkit.org/show_bug.cgi?id=97235
2168
2169         Reviewed by Luiz Agostini.
2170
2171         * WebKitTestRunner/qt/EventSenderProxyQt.cpp:
2172         (WTR::EventSenderProxy::keyDown):
2173
2174 2012-09-21  Gergely Kis  <gergely@homejinni.com>
2175
2176         Add MIPS build slave to build.webkit.org
2177         https://bugs.webkit.org/show_bug.cgi?id=96713
2178
2179         Reviewed by Csaba Osztrogonác.
2180
2181         Added a build slave for MIPS, and enabled a builder for 
2182         Qt Linux MIPS32R2 little-endian release build.
2183
2184         * BuildSlaveSupport/build.webkit.org-config/config.json:
2185
2186 2012-09-21  Peter Beverloo  <peter@chromium.org>
2187
2188         [Chromium] Switch back to a fixed fifo path for Android
2189         https://bugs.webkit.org/show_bug.cgi?id=97230
2190
2191         Reviewed by Tony Chang.
2192
2193         Because not all external storage cards will be formated using a file
2194         system that supports named pipes, Chromium has been changed to creating
2195         the pipes in a temporary folder on the internal storage. Adapt this in
2196         WebKit so we can continue to run layout tests.
2197
2198         This also includes an *unreviewed* fix for a breakage in the webkitpy
2199         tests I made in r129221. Two lines and related to this code, so I decided
2200         to include it in this change.
2201
2202         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
2203         (ChromiumAndroidDriver.__init__):
2204         (ChromiumAndroidDriver._setup_test):
2205         (ChromiumAndroidDriver._update_version):
2206
2207 2012-09-20  Peter Beverloo  <peter@chromium.org>
2208
2209         Leverage Chromium's code to set up FIFOs for Chromium Android layout tests
2210         https://bugs.webkit.org/show_bug.cgi?id=97227
2211
2212         Reviewed by Tony Chang.
2213
2214         We switched Chromium to using FIFOs in order to achieve better consistency,
2215         which was done by Marcus in r157541. Remove all custom WebKit code in
2216         favor of Chromium's implementation.
2217
2218         Remove more FIFO code in the test runner itself, including the code in
2219         DumpRenderTree that invoked it. We can now switch to Chromium's brand
2220         new FIFO-creating code, which is being set-up for all test targets build
2221         for Android, including DumpRenderTree, TestWebKitAPI and webkit_unit_tests.
2222
2223         This also changes the ChromiumAndroidDriver._remove_all_pipes method to
2224         delete the files individually. "rm" would fail if one of the earlier files
2225         does not exist, and the "-f" argument doesn't seem to be reliable.
2226
2227         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
2228         * DumpRenderTree/chromium/TestShellAndroid.cpp:
2229         (platformInit):
2230         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
2231         (ChromiumAndroidDriver.__init__):
2232         (ChromiumAndroidDriver._setup_test):
2233         (ChromiumAndroidDriver._get_external_storage):
2234         (ChromiumAndroidDriver._drt_cmd_line):
2235         (ChromiumAndroidDriver._remove_all_pipes):
2236         (ChromiumAndroidDriver.stop):
2237         * TestWebKitAPI/TestWebKitAPI.gyp/TestWebKitAPI.gyp:
2238
2239 2012-09-21  Christophe Dumez  <christophe.dumez@intel.com>
2240
2241         [WK2][WKTR] EventSender needs to implement scheduleAsynchronousClick
2242         https://bugs.webkit.org/show_bug.cgi?id=97326
2243
2244         Reviewed by Kenneth Rohde Christiansen.
2245
2246         Implement scheduleAsynchronousClick() in WebKitTestRunner's
2247         EventSender by sending a "MouseDown" and a "MouseUp" message
2248         asynchronously to the WebProcess.
2249
2250         * WebKitTestRunner/InjectedBundle/Bindings/EventSendingController.idl:
2251         * WebKitTestRunner/InjectedBundle/EventSendingController.cpp:
2252         (WTR::createMouseMessageBody):
2253         (WTR):
2254         (WTR::EventSendingController::mouseDown):
2255         (WTR::EventSendingController::mouseUp):
2256         (WTR::EventSendingController::scheduleAsynchronousClick):
2257         * WebKitTestRunner/InjectedBundle/EventSendingController.h:
2258         (EventSendingController):
2259         * WebKitTestRunner/TestController.cpp:
2260         (WTR::TestController::didReceiveMessageFromInjectedBundle):
2261
2262 2012-09-21  Simon Hausmann  <simon.hausmann@digia.com>
2263
2264         [Qt] Error out early if we don't have ICU available
2265
2266         Reviewed by Tor Arne Vestbø.
2267
2268         * qmake/mkspecs/features/configure.prf:
2269
2270 2012-09-21  Simon Hausmann  <simon.hausmann@digia.com>
2271
2272         [Qt] Bail out when trying to build WebKit with Qt != 5
2273
2274         Reviewed by Tor Arne Vestbø.
2275
2276         Moved check for Qt version out of default_pre into top-level WebKit.pro,
2277         because we never reach default_pre.prf due to the lack of .qmake.conf support
2278         in older versions of Qt/QMake.
2279
2280         * qmake/mkspecs/features/default_pre.prf:
2281
2282 2012-09-21  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
2283
2284         WebKitTestRunner needs to print frame load delegate information
2285         https://bugs.webkit.org/show_bug.cgi?id=42705
2286
2287         Reviewed by Kenneth Rohde Christiansen.
2288
2289         Added missing dumping from WTR::InjectedBundlePage::didFailLoadWithErrorForFrame.
2290
2291         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2292         (WTR::InjectedBundlePage::didFailLoadWithErrorForFrame):
2293
2294 2012-09-21  Christophe Dumez  <christophe.dumez@intel.com>
2295
2296         [EFL] EventSender should mimic CTRL+o emacs shortcut
2297         https://bugs.webkit.org/show_bug.cgi?id=97224
2298
2299         Reviewed by Kenneth Rohde Christiansen.
2300
2301         Add support for mimicking CTRL+o emacs shortcut in EFL's
2302         EventSender in DumpRenderTree and WebKitTestRunner.
2303
2304         * DumpRenderTree/efl/EventSender.cpp:
2305         (sendKeyDown):
2306         * WebKitTestRunner/efl/EventSenderProxyEfl.cpp:
2307         (WTR::EventSenderProxy::keyDown):
2308
2309 2012-09-21  Simon Hausmann  <simon.hausmann@digia.com>
2310
2311         [Qt] Re-fix clean builds
2312
2313         Reviewed by Tor Arne Vestbø.
2314
2315         Re-introduce the sanitization for LIBS when creating a module to use
2316         LIBS_PRIVATE and otherwise do _not_ use LIBS_PRIVATE. We decided to
2317         continue to use QT, LIBS and PKGCONFIG instead of their _PRIVATE variants
2318         throughout the code base, so just using LIBS_PRIVATE in linkAgainstLibrary()
2319         causes build issues when depending system libraries end up in LIBS before
2320         LIBS_PRIVATE.
2321
2322         * qmake/mkspecs/features/default_post.prf:
2323         * qmake/mkspecs/features/functions.prf:
2324
2325 2012-09-20  Simon Fraser  <simon.fraser@apple.com>
2326
2327         Comment out a failing webkitpy unit test until Dirk can fix it.
2328
2329         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2330         (SemanticTests.test_skip_and_wontfix):
2331
2332 2012-09-20  Dirk Pranke  <dpranke@chromium.org>
2333
2334         make Skip, WontFix be the only expectations on a line
2335         https://bugs.webkit.org/show_bug.cgi?id=97225
2336
2337         Reviewed by Ojan Vafai.
2338
2339         It is now incorrect in the new syntax to have a line like:
2340
2341             foo.html [ WontFix Crash ]
2342
2343         This will generate a lint warning and be treated as an invalid
2344         line. Fixing this caused a whole bunch of unit tests to need updating
2345         to no longer be marked as WontFix :). Also, this patch adjusts
2346         the warnings so that missing Bug() identifiers will cause lint
2347         warnings but will *not* cause the line to be treated as invalid.
2348         Fixing these issues also revealed that test_hung_thread was no
2349         longer testing the right logic, so I adjusted the timeouts in
2350         test.py to make that test work again.
2351
2352         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner.py:
2353         (Worker._run_test_in_another_thread):
2354         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2355         (ResultSummaryTest.test_summarized_results_wontfix):
2356         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2357         (TestExpectationParser._parse_modifiers):
2358         (TestExpectationParser._tokenize_line_using_new_format):
2359         (TestExpectationLine.is_invalid):
2360         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2361         (BasicTests.test_basic):
2362         (test_get_test_set):
2363         (test_parse_warning):
2364         (test_pixel_tests_flag):
2365         (SemanticTests.test_missing_bugid):
2366         (SemanticTests):
2367         (SemanticTests.test_skip_and_wontfix):
2368         * Scripts/webkitpy/layout_tests/port/test.py:
2369         (TestDriver.run_test):
2370         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2371         (MainTest.test_hung_thread):
2372         * Scripts/webkitpy/tool/commands/queries_unittest.py:
2373         (PrintExpectationsTest.test_basic):
2374         (PrintExpectationsTest.test_multiple):
2375         (PrintExpectationsTest.test_full):
2376         (PrintExpectationsTest.test_exclude):
2377         (PrintExpectationsTest.test_csv):
2378
2379 2012-09-20  Lucas Forschler  <lforschler@apple.com>
2380
2381         Unreviewed.  Start running tests on the mac-ews.
2382         
2383         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
2384         (MacEWS):
2385
2386 2012-09-20  Tony Chang  <tony@chromium.org>
2387
2388         [chromium] Enable more clang warnings
2389         https://bugs.webkit.org/show_bug.cgi?id=97257
2390
2391         Reviewed by James Robinson.
2392
2393         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp: Add -Wunused-parameter. Don't add -Wglobal-constructors because
2394         there are lots of violations in these files.
2395
2396 2012-09-19  Ryosuke Niwa  <rniwa@webkit.org>
2397
2398         run-perf-tests should record individual value instead of statistics
2399         https://bugs.webkit.org/show_bug.cgi?id=97155
2400
2401         Reviewed by Hajime Morita.
2402
2403         Parse the list of individual value reported by tests and include them as "values".
2404         We strip "values" from the output JSON when uploading it to the perf-o-matic
2405         since it doesn't know how to parse "values" or ignore it.
2406
2407         * Scripts/webkitpy/performance_tests/perftest.py:
2408         (PerfTest):
2409         (PerfTest.parse_output): Parse and report "values".
2410         (PageLoadingPerfTest.run): Report indivisual page loading time in "values".
2411         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
2412         (MainTest.test_parse_output):
2413         (MainTest.test_parse_output_with_failing_line):
2414         (TestPageLoadingPerfTest.test_run):
2415         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
2416         (PerfTestsRunner._generate_and_show_results): Strip "values" from each result
2417         until we update perf-o-matic.
2418         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
2419         (test_run_memory_test):
2420         (test_run_with_json_output):
2421         (test_run_with_description):
2422         (test_run_with_slave_config_json):
2423         (test_run_with_multiple_repositories):
2424
2425 2012-09-20  Simon Hausmann  <simon.hausmann@digia.com>
2426
2427         [Qt] Fix production builds
2428
2429         Unreviewed trivial fix: Follow up r129119 and avoid clobbering LIBS through the use of LIBS_PRIVATE.
2430
2431         Otherwise the libQtWebKit.prl file for example contains -lWebKit1 -lWebCore, etc.
2432
2433         * qmake/mkspecs/features/functions.prf:
2434
2435 2012-09-20  Dirk Pranke  <dpranke@chromium.org>
2436
2437         REGRESSION: layout test results doesn't show diffs
2438         https://bugs.webkit.org/show_bug.cgi?id=97182
2439
2440         Reviewed by Ojan Vafai.
2441
2442         Go back to storing TEXT, AUDIO, and IMAGE+TEXT in results.json
2443         so that results.html (and hopefully garden-o-matic) can
2444         determine which things actually failed. However, we keep mapping
2445         these results to Failure so that we still only have a single
2446         expectation type for them.
2447
2448         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
2449         (JSONLayoutResultsGenerator):
2450         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2451         (TestExpectationParser):
2452         (TestExpectations):
2453         (TestExpectations.result_was_expected):
2454         * Scripts/webkitpy/layout_tests/models/test_failures.py:
2455         (determine_result_type):
2456         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2457         (MainTest.test_missing_and_unexpected_results):
2458         (MainTest.test_retrying_and_flaky_tests):
2459
2460 2012-09-20  Tommy Widenflycht  <tommyw@google.com>
2461
2462         MediaStream API: Extend UserMediaRequest with a ownerDocument method
2463         https://bugs.webkit.org/show_bug.cgi?id=97095
2464
2465         Reviewed by Adam Barth.
2466
2467         Extending WebUserMediaClientMock to check that the owning document is valid,
2468         and that the document has a frame.
2469
2470         * DumpRenderTree/chromium/WebUserMediaClientMock.cpp:
2471         (WebKit::WebUserMediaClientMock::requestUserMedia):
2472
2473 2012-09-20  Stephen Chenney  <schenney@chromium.org>
2474
2475         [Chromium] DRT does not support --dump-all-pixels flag
2476         https://bugs.webkit.org/show_bug.cgi?id=95098
2477
2478         Reviewed by Dirk Pranke.
2479
2480         Add support for the --pixel-tests and shorthand -p option in Chromium DumpRenderTree. Use
2481         of this flag causes pixel results to be created for all tests, regardless of
2482         individual test options. If an individual test provides a pixel hash it will be used,
2483         otherwise the hash will be empty. This replaces a previously defined but unused option
2484         --dump-all-pixels, and is useful primarily when debugging DRT instances.
2485
2486         * DumpRenderTree/chromium/DumpRenderTree.cpp:
2487         (runTest): Add a parameter and code to force pixel results for the test.
2488         (main): Add parameter handling for --pixels-test and -p, and remove --dump-all-pixels.
2489
2490 2012-09-20  Carlos Garcia Campos  <cgarcia@igalia.com>
2491
2492         [GTK] run-api-tests should not buffer test stdout
2493         https://bugs.webkit.org/show_bug.cgi?id=88474
2494
2495         Reviewed by Philippe Normand.
2496
2497         * Scripts/run-gtk-tests:
2498         (TestRunner._run_test_command): Use os.forkpty() instead of
2499         subprocess.Popen() so that gtest sends the output with colors to
2500         stdout. Use common.parse_output_lines() to parse the output and
2501         write it to stdout while it's read.
2502         (TestRunner._run_test_command.parse_line): Parse the line to get
2503         the test pid and write the line to stdout.
2504         (TestRunner._run_test_command.waitpid): Helper function to call
2505         waitpid handling EINTR.
2506         (TestRunner._run_test_command.return_code_from_exit_status):
2507         Helper function to convert exit status of test commands to a
2508         return code.
2509         * gtk/common.py:
2510         (parse_output_lines): Helper function that uses select to read
2511         the given file descriptor and call the given callback for every
2512         line read.
2513
2514 2012-09-20  Simon Hausmann  <simon.hausmann@digia.com>
2515
2516         [Qt] Fix initial build
2517
2518         Reviewed by Tor Arne Vestbø.
2519
2520         When building QtWebKit the first time there is no qt_webkit.pri module pri file, and therefore
2521         $$QT.webkit.name isn't set and so creating_module isn't set. That has all sorts of implications
2522         causing incorrect linking for Makefile.api, etc.
2523
2524         Fix the determination by simply checking if MODULE is set, which only happens in api.pri.
2525
2526         * qmake/mkspecs/features/webkit_modules.prf:
2527
2528 2012-09-20  Sheriff Bot  <webkit.review.bot@gmail.com>
2529
2530         Unreviewed, rolling out r129091.
2531         http://trac.webkit.org/changeset/129091
2532         https://bugs.webkit.org/show_bug.cgi?id=97205
2533
2534         It broke perf tests everywhere (Requested by Ossy on #webkit).
2535
2536         * Scripts/webkitpy/performance_tests/perftest.py:
2537         (PerfTest):
2538         (PerfTest.parse_output):
2539         (PageLoadingPerfTest.run):
2540         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
2541         (MainTest.test_parse_output):
2542         (MainTest.test_parse_output_with_failing_line):
2543         (TestPageLoadingPerfTest.test_run):
2544         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
2545         (PerfTestsRunner._generate_and_show_results):
2546         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
2547         (test_run_memory_test):
2548         (test_run_with_json_output):
2549         (test_run_with_description):
2550         (test_run_with_slave_config_json):
2551         (test_run_with_multiple_repositories):
2552
2553 2012-09-20  Simon Hausmann  <simon.hausmann@digia.com>
2554
2555         [Qt] QtWebKit module header includes private dependencies
2556
2557         Reviewed by Tor Arne Vestbø.
2558
2559         * qmake/mkspecs/features/default_post.prf: Don't try to sanitize LIBS, because we can make sure
2560         that LIBS_PRIVATE is set from the beginning. Moved the creating_module and PKGCONFIG/QT(_PRIVATE)
2561         sanitization into webkit_modules.prf. creating_module determination requires TARGET to be set, so
2562         we can't do it in default_pre.
2563         * qmake/mkspecs/features/webkit_modules.prf:
2564
2565 2012-09-20  Simon Hausmann  <simon.hausmann@digia.com>
2566
2567         [Qt] Fix build with MingW
2568
2569         Reviewed by Tor Arne Vestbø.
2570
2571         Don't try to compile WebCore, etc. with debug symbols for production
2572         builds, it's just too big.
2573
2574         * qmake/mkspecs/features/production_build.prf:
2575
2576 2012-09-20  Michał Pakuła vel Rutka  <m.pakula@samsung.com>
2577
2578         [EFL][DRT]DumpRenderTree needs to reset focus state when test starts.
2579         https://bugs.webkit.org/show_bug.cgi?id=97087
2580
2581         Reviewed by Gyuyoung Kim.
2582
2583         Add focusing a main frame on settings reset.
2584         After editing/undo/undo-iframe-location-change was executed a frame
2585         was left in unfocused state. This caused flakiness in two tests results.
2586
2587         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
2588         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
2589
2590 2012-09-20  Allan Sandfeld Jensen  <allan.jensen@digia.com>
2591
2592         [Qt] Add eventSender.gestureTap
2593         https://bugs.webkit.org/show_bug.cgi?id=66173
2594
2595         Reviewed by Kenneth Rohde Christiansen.
2596
2597         Implement support for sending raw QGesture events.
2598
2599         * DumpRenderTree/qt/EventSenderQt.cpp:
2600         (EventSender::EventSender):
2601         (EventSender::gestureTap):
2602         * DumpRenderTree/qt/EventSenderQt.h:
2603         (EventSender):
2604
2605 2012-09-19  Sheriff Bot  <webkit.review.bot@gmail.com>
2606
2607         Unreviewed, rolling out r129007.
2608         http://trac.webkit.org/changeset/129007
2609         https://bugs.webkit.org/show_bug.cgi?id=97172
2610
2611         It broke the build on the Qt bots (Requested by Ossy on
2612         #webkit).
2613
2614         * qmake/config.tests/gccdepends/empty.cpp:
2615         * qmake/config.tests/gccdepends/gccdepends.pro:
2616         * qmake/mkspecs/features/default_pre.prf:
2617
2618 2012-09-19  Ryosuke Niwa  <rniwa@webkit.org>
2619
2620         run-perf-tests should record indivisual value instead of statistics
2621         https://bugs.webkit.org/show_bug.cgi?id=97155
2622
2623         Reviewed by Hajime Morita.
2624
2625         Parse the list of indivisual value reported by tests and include them as "values".
2626         We strip "values" from the output JSON when uploading it to the perf-o-matic
2627         since it doesn't know how to parse "values" or ignore it.
2628
2629         * Scripts/webkitpy/performance_tests/perftest.py:
2630         (PerfTest):
2631         (PerfTest.parse_output): Parse and report "values".
2632         (PageLoadingPerfTest.run): Report indivisual page loading time in "values".
2633         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
2634         (MainTest.test_parse_output):
2635         (MainTest.test_parse_output_with_failing_line):
2636         (TestPageLoadingPerfTest.test_run):
2637         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
2638         (PerfTestsRunner._generate_and_show_results): Strip "values" from each result
2639         until we update perf-o-matic.
2640         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
2641         (test_run_memory_test):
2642         (test_run_with_json_output):
2643         (test_run_with_description):
2644         (test_run_with_slave_config_json):
2645         (test_run_with_multiple_repositories):
2646
2647 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2648
2649         Fix regex groups for bug matching in flakiness dashboard.
2650         https://bugs.webkit.org/show_bug.cgi?id=97152
2651
2652         Unreviewed, build fix.
2653
2654         * TestResultServer/static-dashboards/flakiness_dashboard.js:
2655         (htmlForBugs):
2656
2657 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2658
2659         nrwt: print unexpected results using new TestExpectations syntax
2660         https://bugs.webkit.org/show_bug.cgi?id=97159
2661
2662         Unreviewed, build fix.
2663
2664         Change new-run-webkit-tests to print out failures using the new
2665         syntax when there are unexpected results, e.g.:
2666
2667         52 tests ran as expected, 19 didn't:
2668
2669
2670         Regressions: Unexpected failures : (2)
2671           failures/flaky/text.html [ Failure ]
2672           failures/unexpected/text-image-checksum.html [ Failure ]
2673
2674         and so forth
2675
2676         * Scripts/webkitpy/layout_tests/views/printing.py:
2677         (Printer._print_unexpected_results):
2678
2679 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2680
2681         update flakiness dashboard after cutover to new test expectations syntax
2682         https://bugs.webkit.org/show_bug.cgi?id=97152
2683
2684         Unreviewed, build fix.
2685
2686         Handle (??) the new Bug notations as well. Hopefully we don't
2687         still need the old ones.
2688
2689         * TestResultServer/static-dashboards/flakiness_dashboard.js:
2690         (filterBugs):
2691         (htmlForBugs):
2692         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
2693
2694 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2695
2696         update flakiness dashboard after cutover to new test expectations syntax
2697         https://bugs.webkit.org/show_bug.cgi?id=97152
2698
2699         Reviewed by Ryosuke Niwa.
2700
2701         This change clones the TestExpectation parsing state machine
2702         from python into javascript.
2703
2704         * TestResultServer/static-dashboards/flakiness_dashboard.js:
2705         (parsedExpectations.lines.forEach):
2706         (parsedExpectations):
2707         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
2708
2709 2012-09-19  Dana Jansens  <danakj@chromium.org>
2710
2711         Add backer@chromium.org as contributor
2712         https://bugs.webkit.org/show_bug.cgi?id=97150
2713
2714         * Scripts/webkitpy/common/config/committers.py:
2715
2716 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2717
2718         fix MISSING after TestExpectations conversion
2719         https://bugs.webkit.org/show_bug.cgi?id=97148
2720
2721         Unreviewed, expectations change / build fix.
2722
2723         * Scripts/convert-test-expectations:
2724         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2725         (TestExpectationParser):
2726
2727 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2728
2729         nrwt: convert chromium TestExpectations to the new syntax
2730         https://bugs.webkit.org/show_bug.cgi?id=97139
2731
2732         Reviewed by Ryosuke Niwa.
2733     
2734         Add in a file temporarily that does the conversion of
2735         test expectations formats
2736
2737         * Tools/Scripts/convert-test-expectations: Added.
2738
2739 2012-09-19  Ryosuke Niwa  <rniwa@webkit.org>
2740
2741         REGRESSION: run-perf-tests no longer reports the total test time
2742         https://bugs.webkit.org/show_bug.cgi?id=97138
2743
2744         Reviewed by Tony Chang.
2745
2746         Report the finished time as a info-level log as opposed to a debug level log.
2747
2748         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
2749         (PerfTestsRunner._run_single_test):
2750         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
2751         (MainTest.normalizeFinishedTime): Added. It replaces all finished times by 0.1 seconds.
2752         (test_run_test_pause_before_testing):
2753         (test_run_test_set_for_parser_tests):
2754         (test_run_memory_test):
2755         (_test_run_with_json_output):
2756
2757 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2758
2759         Support new TestExpectations format alongside old one
2760         https://bugs.webkit.org/show_bug.cgi?id=96588
2761
2762         Reviewed by Ojan Vafai.
2763
2764         This patch adds support for actually reading in lines formatted
2765         in the new style, and when we re-serialize/write out the file,
2766         all lines will be written in the new style. Note that reading in
2767         the old style is still supported, and no updates are being made
2768         to the actual TestExpectations files as part of this change.
2769
2770         This change updates most but not all of the unit tests to use
2771         the new syntax. I will update the rest when (or before where
2772         possible) I drop support for the old format.
2773
2774         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2775         (TestExpectationParser):
2776         (TestExpectationParser._tokenize_line):
2777         (TestExpectationParser._tokenize_line_using_new_format):
2778         (TestExpectationLine.to_string):
2779         (TestExpectationLine._format_line):
2780         (TestExpectations):
2781         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2782         (test_parse_warning):
2783         (SkippedTests.test_skipped_entry_dont_exist):
2784         (NewExpectationSyntaxTests.test_warnings):
2785         (RemoveConfigurationsTest.test_remove):
2786         (test_remove_line):
2787         (RebaseliningTest.test_remove):
2788         (RebaseliningTest.test_no_get_rebaselining_failures):
2789         (TestExpectationParserTests.test_tokenize_blank):
2790         (TestExpectationParserTests.test_tokenize_extra_colon):
2791         (TestExpectationParserTests.test_tokenize_missing_equal):
2792         (TestExpectationParserTests.test_tokenize_extra_equal):
2793         (TestExpectationSerializationTests.test_unparsed_to_string):
2794         (TestExpectationSerializationTests.test_unparsed_list_to_string):
2795         (TestExpectationSerializationTests.test_parsed_to_string):
2796         (TestExpectationSerializationTests.test_format_line):
2797         (TestExpectationSerializationTests.test_string_roundtrip):
2798         (TestExpectationSerializationTests.test_list_roundtrip):
2799         (TestExpectationSerializationTests.test_reconstitute_only_these):
2800         (TestExpectationSerializationTests.test_string_whitespace_stripping):
2801         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2802         (LintTest.test_lint_test_files__errors):
2803         * Scripts/webkitpy/tool/commands/queries_unittest.py:
2804         (PrintExpectationsTest.test_basic):
2805         (PrintExpectationsTest.test_multiple):
2806         (PrintExpectationsTest.test_full):
2807         (PrintExpectationsTest.test_exclude):
2808         * Scripts/webkitpy/tool/commands/rebaseline.py:
2809         (RebaselineExpectations.execute):
2810         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
2811         (TestRebaseline.test_rebaseline_updates_expectations_file_noop):
2812         (test_rebaseline_updates_expectations_file):
2813         (test_rebaseline_does_not_include_overrides):
2814         (test_rebaseline_expectations_noop):
2815         (test_overrides_are_included_correctly):
2816
2817 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2818
2819         implement first part of support for the new TestExpectations syntax
2820         https://bugs.webkit.org/show_bug.cgi?id=96569
2821
2822         Reviewed by Ryosuke Niwa.
2823
2824         This patch implements support for parsing a line of the new
2825         format for the TestExpectations file and converting it back into
2826         the old format for compatibility. This routine is not yet used
2827         by anything.
2828
2829         The new format is documented at:
2830           http://trac.webkit.org/wiki/TestExpectations
2831
2832         but, in short:
2833
2834           [bugs] [ "[" modifiers "]" ] test_name [ "[" expectations "]" ]
2835
2836         - Comments are indicated with "#" instead of "//"
2837         - If no expectations are specified we default to Skip for
2838           compatibility with the Skipped files (these two changes make
2839           Skipped files a subset of TestExpectations files)
2840
2841         - All of the tokens are now CamelCase instead of ALLCAPS.
2842         - FAIL -> Failure
2843         - IMAGE -> ImageOnlyFailure
2844         - WONTFIX -> WontFix
2845         - modifiers refer to just the platforms and configurations
2846           (release/debug) that the line applies to.
2847         - WontFix, Rebaseline, Slow, and Skip move to the right-hand side as
2848           expectations
2849         - expectations will typically be written out in lexicographic order
2850         - We use webkit.org/b/12345, crbug.com/12345, and Bug(dpranke)
2851           instead of BUGWK12345, BUGCR12345, and BUGDPRANKE.
2852
2853         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2854         (TestExpectationParser):
2855         (TestExpectationParser._tokenize_line_using_new_format):
2856         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2857         (NewExpectationSyntaxTests):
2858         (NewExpectationSyntaxTests.assert_exp):
2859         (NewExpectationSyntaxTests.test_bare_name):
2860         (NewExpectationSyntaxTests.test_bare_name_and_bugs):
2861         (NewExpectationSyntaxTests.test_comments):
2862         (NewExpectationSyntaxTests.test_config_modifiers):
2863         (NewExpectationSyntaxTests.test_unknown_config):
2864         (NewExpectationSyntaxTests.test_unknown_expectation):
2865         (NewExpectationSyntaxTests.test_skip):
2866         (NewExpectationSyntaxTests.test_slow):
2867         (NewExpectationSyntaxTests.test_wontfix):
2868         (NewExpectationSyntaxTests.test_blank_line):
2869         (NewExpectationSyntaxTests.test_warnings):
2870
2871 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2872
2873         nrwt: replace TEXT, AUDIO, and IMAGE+TEXT with FAIL
2874         https://bugs.webkit.org/show_bug.cgi?id=96845
2875
2876         Reviewed by Ojan Vafai.
2877
2878         In preparation for the new TestExpectations syntax, we replace
2879         all TEXT, IMAGE+TEXT, and AUDIO failures with FAIL. This will
2880         make switching to the new syntax lossless (i.e., we lose
2881         information now, not then).
2882
2883         Note that we can still parse in results.json files that have the
2884         old data for backwards compatibility.
2885         
2886
2887         * Scripts/webkitpy/common/net/resultsjsonparser.py:
2888         (JSONTestResult._failure_types_from_actual_result):
2889         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
2890         (ResultsJSONParserTest):
2891         (test_basic):
2892         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2893         (ResultSummaryTest.test_summarized_results_wontfix):
2894         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
2895         (JSONLayoutResultsGenerator):
2896         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2897         (TestExpectations):
2898         (TestExpectations.remove_pixel_failures):
2899         (TestExpectations.has_pixel_failures):
2900         (TestExpectations.suffixes_for_expectations):
2901         (TestExpectations.get_rebaselining_failures):
2902         (TestExpectations.remove_configuration_from_test):
2903         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2904         (assert_bad_expectations):
2905         (BasicTests):
2906         (BasicTests.test_basic):
2907         (MiscTests):
2908         (MiscTests.test_multiple_results):
2909         (MiscTests.test_result_was_expected):
2910         (MiscTests.test_remove_pixel_failures):
2911         (MiscTests.test_suffixes_for_expectations):
2912         (test_get_expectations_string):
2913         (test_parse_warning):
2914         (test_error_on_different_platform):
2915         (test_error_on_different_build_type):
2916         (test_overrides):
2917         (test_overrides__directory):
2918         (test_overrides__duplicate):
2919         (test_pixel_tests_flag):
2920         (test_more_specific_override_resets_skip):
2921         (SkippedTests.test_skipped_file_overrides_expectations):
2922         (SkippedTests.test_skipped_dir_overrides_expectations):
2923         (SkippedTests.test_skipped_file_overrides_overrides):
2924         (SkippedTests.test_skipped_dir_overrides_overrides):
2925         (ExpectationSyntaxTests.test_missing_colon):
2926         (ExpectationSyntaxTests.test_too_many_equals_signs):
2927         (SemanticTests.test_bug_format):
2928         (SemanticTests.test_bad_bugid):
2929         (SemanticTests.test_missing_bugid):
2930         (SemanticTests.test_rebaseline):
2931         (test_missing_file):
2932         (test_more_modifiers):
2933         (test_order_in_file):
2934         (test_macro_overrides):
2935         (RemoveConfigurationsTest.test_remove):
2936         (test_remove_line):
2937         (RebaseliningTest.test_remove):
2938         (TestExpectationSerializationTests.test_serialize_parsed_expectations):
2939         * Scripts/webkitpy/layout_tests/models/test_failures.py:
2940         (determine_result_type):
2941         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
2942         (test_test_expectations):
2943         * Scripts/webkitpy/layout_tests/port/test.py:
2944         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2945         (MainTest.test_missing_and_unexpected_results):
2946         (MainTest.test_retrying_and_flaky_tests):
2947         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2948         (TestExpectationsTestCase.test_valid_expectations):
2949         * Scripts/webkitpy/tool/commands/queries_unittest.py:
2950         (PrintExpectationsTest.test_basic):
2951         (PrintExpectationsTest.test_multiple):
2952         (PrintExpectationsTest.test_full):
2953         (PrintExpectationsTest.test_exclude):
2954         (PrintExpectationsTest.test_csv):
2955         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
2956         (test_rebaseline_does_not_include_overrides):
2957         (test_overrides_are_included_correctly):
2958
2959 2012-09-19  Dominic Mazzoni  <dmazzoni@google.com>
2960
2961         AX: A few control types are returning the wrong answer for isReadOnly
2962         https://bugs.webkit.org/show_bug.cgi?id=96735
2963
2964         Reviewed by Chris Fleizach.
2965
2966         Exposing isReadOnly in an AccessibilityObject to DumpRenderTree.
2967
2968         * DumpRenderTree/chromium/TestRunner/AccessibilityUIElementChromium.cpp:
2969         (AccessibilityUIElement::AccessibilityUIElement):
2970         (AccessibilityUIElement::isReadOnlyGetterCallback):
2971         * DumpRenderTree/chromium/TestRunner/AccessibilityUIElementChromium.h:
2972         (AccessibilityUIElement):
2973
2974 2012-09-19  Sudarsana Nagineni  <sudarsana.nagineni@intel.com>
2975
2976         [WTR] Memory leaks in TestRunner::deliverWebIntent()
2977         https://bugs.webkit.org/show_bug.cgi?id=97111
2978
2979         Reviewed by Kenneth Rohde Christiansen.
2980
2981         Fix memory leaks in deliverWebIntent() by adopting strings
2982         created with WKStringCreateWithUTF8CString().
2983
2984         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
2985         (WTR::TestRunner::deliverWebIntent):
2986
2987 2012-09-19  Simon Hausmann  <simon.hausmann@digia.com>
2988
2989         [Qt] Fix incremental builds with all-in-one-files and gccdepends
2990
2991         Reviewed by Tor Arne Vestbø.
2992
2993         Pass -MP to gcc when we use the gcc depends feature, to ensure that implicit rules
2994         are not only created for header files but also for .cpp files. AllInOne.cpp files
2995         include other .cpp files, and when those are removed we need those dummy rules to
2996         avoid a "No rule to make Foo.cpp required by AllInOne.o" error.
2997
2998         * qmake/config.tests/gccdepends/empty.cpp:
2999         (main):
3000         * qmake/config.tests/gccdepends/gccdepends.pro:
3001         * qmake/mkspecs/features/default_pre.prf:
3002
3003 2012-09-19  Rick Byers  <rbyers@chromium.org>
3004
3005         Do touch adjustment on GestureTapDown
3006         https://bugs.webkit.org/show_bug.cgi?id=96677
3007
3008         Reviewed by Antonio Gomes.
3009
3010         Allow radius to be set for GestureTapDown events.
3011         * DumpRenderTree/chromium/TestRunner/EventSender.cpp:
3012         (EventSender::gestureEvent):
3013
3014 2012-09-19  Christophe Dumez  <christophe.dumez@intel.com>
3015
3016         [EFL] EFL's DRT does not support overriding 'WebKitCSSRegionsEnabled' preference
3017         https://bugs.webkit.org/show_bug.cgi?id=97100
3018
3019         Reviewed by Gyuyoung Kim.
3020
3021         EFL's DRT now supports overriding the 'WebKitCSSRegionsEnabled'
3022         preference, in order to match WebKitTestRunner functionality.
3023
3024         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
3025         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
3026         * DumpRenderTree/efl/TestRunnerEfl.cpp:
3027         (TestRunner::overridePreference):
3028
3029 2012-09-18  Kangil Han  <kangil.han@samsung.com>
3030
3031         [WK2][WTR] CodeGeneratorTestRunner could keep original copyright.
3032         https://bugs.webkit.org/show_bug.cgi?id=96181
3033
3034         Reviewed by Daniel Bates.
3035
3036         This patch enabled derived files, in DerivedSources/InjectedBundle, to keep original copyright.
3037
3038         * WebKitTestRunner/InjectedBundle/Bindings/CodeGeneratorTestRunner.pm:
3039         (new):
3040         (_parseLicenseBlock):
3041         (_parseLicenseBlockFromFile):
3042         (_defaultLicenseBlock):
3043         (_licenseBlock):
3044         (_generateHeaderFile):
3045         (_generateImplementationFile):
3046
3047 2012-09-18  Byungwoo Lee  <bw80.lee@samsung.com>
3048
3049         Title string should be changed when document.title is set to ''.
3050         https://bugs.webkit.org/show_bug.cgi?id=96793
3051
3052         Reviewed by Kenneth Rohde Christiansen.
3053
3054         Change dump format of dumpTitleChanges more understandable.
3055         Uses single quotation marks for the title string.
3056
3057         * DumpRenderTree/blackberry/DumpRenderTree.cpp:
3058         (BlackBerry::WebKit::DumpRenderTree::didReceiveTitleForFrame):
3059         * DumpRenderTree/chromium/WebViewHost.cpp:
3060         (WebViewHost::didReceiveTitle):
3061         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
3062         (DumpRenderTreeChrome::onFrameTitleChanged):
3063         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3064         (webViewTitleChanged):
3065         * DumpRenderTree/mac/FrameLoadDelegate.mm:
3066         (-[FrameLoadDelegate webView:didReceiveTitle:forFrame:]):
3067         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
3068         (WebCore::DumpRenderTree::titleChanged):
3069         * DumpRenderTree/win/FrameLoadDelegate.cpp:
3070         (FrameLoadDelegate::didReceiveTitle):
3071         * DumpRenderTree/wx/DumpRenderTreeWx.cpp:
3072         (LayoutWebViewEventHandler::OnReceivedTitleEvent):
3073         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3074         (WTR::InjectedBundlePage::didReceiveTitleForFrame):
3075
3076 2012-09-18  Szilard Ledan  <szledan@inf.u-szeged.hu>
3077
3078         EWS shouldn't sleep if there are new patches in its queue
3079         https://bugs.webkit.org/show_bug.cgi?id=83038
3080
3081         Reviewed by Eric Seidel.
3082
3083         EWS tries to process a security patch. Of course it can't, because the EWS isn't
3084         the member of the security group. But the problem is that after it can't process
3085         the attachment, it says that queue is empty (but it isn't!) and it sleeps 2 minutes
3086         and push the security patch to the end of the queue.
3087         Now it stays in the loop until it finds a patch or the queue gets empty.
3088
3089         * Scripts/webkitpy/tool/commands/queues.py:
3090         (AbstractPatchQueue._next_patch):
3091         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3092         (AbstractPatchQueueTest.test_next_patch):
3093
3094 2012-09-18  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
3095
3096         [WK2][WTR] InjectedBundle::booleanForKey() should handle literals effectively
3097         https://bugs.webkit.org/show_bug.cgi?id=97014
3098
3099         Reviewed by Kenneth Rohde Christiansen.
3100
3101         According to http://trac.webkit.org/wiki/EfficientStrings WTF::StringBuilder::appendLiteral() shall
3102         be used for literals rather than WTF::StringBuilder::append().
3103
3104         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
3105         (WTR::InjectedBundle::booleanForKey):
3106
3107 2012-09-18  Simon Hausmann  <simon.hausmann@digia.com>
3108
3109         Update my e-mail address.
3110
3111         * Scripts/webkitpy/common/config/committers.py:
3112
3113 2012-09-18  Andras Becsi  <andras.becsi@digia.com>
3114
3115         Update my e-mail address.
3116
3117         Unreviewed.
3118
3119         * Scripts/webkitpy/common/config/committers.py:
3120
3121 2012-09-18  Andras Becsi  <andras.becsi@digia.com>
3122
3123         [Qt] qt_webkit.pri should not be listed in Tools.pro
3124
3125         Reviewed and rubber-stamped by Simon Hausmann and Tor Arne Vestbø.
3126
3127         Since r128751 the module pri file is auto-generated
3128         but it was still listed in OTHER_FILES.
3129
3130         * Tools.pro: Remove unneeded line.
3131
3132 2012-09-18  Simon Hausmann  <simon.hausmann@digia.com>
3133
3134         [Qt] Fix build with some versions of the gold linker
3135
3136         Reviewed by Tor Arne Vestbø.
3137
3138         Don't unconditionally pass --no-keep-memory to the linker, some versions might not support it.
3139         Instead run a compile/link test first to see if it works.
3140
3141         * qmake/config.tests/gnuld/gnuld.pro: Added.
3142         * qmake/config.tests/gnuld/main.cpp: Added.
3143         (main):
3144         * qmake/mkspecs/features/unix/default_post.prf:
3145
3146 2012-09-18  Simon Hausmann  <simon.hausmann@digia.com>
3147
3148         [Qt] Remove forced use of gold.
3149
3150         Reviewed by Tor Arne Vestbø.
3151
3152         The choice of what linker to use with WebKit should be taken by Qt's build system and ideally the same for all
3153         modules of Qt. Then in turn it is usually up to the administrator of the machine. Recent Debian based systems
3154         often offer the automatic use of gold through a symlink and a dpkg-diversion when installing the gold package.
3155
3156         * qmake/mkspecs/features/unix/default_post.prf:
3157
3158 2012-09-18  Simon Hausmann  <simon.hausmann@digia.com>
3159
3160         [Qt] Prospective Qt/Windows cross-compiling fix
3161
3162         Reviewed by Tor Arne Vestbø.
3163
3164         The win32 scope is not set when cross-compiling from Linux to Windows.
3165
3166         * qmake/mkspecs/features/functions.prf:
3167
3168 2012-09-18  Seokju Kwon  <seokju.kwon@samsung.com>
3169
3170         [EFL] Remove background view on EWebLauncher and MiniBrowser
3171         https://bugs.webkit.org/show_bug.cgi?id=96905
3172
3173         Reviewed by Gyuyoung Kim.
3174
3175         The size of webview was changed after adding url bar.
3176         And the background view is not necessary anymore, since it was used for debugging back in the day.
3177
3178         * EWebLauncher/main.c:
3179         (_ELauncher):
3180         (on_ecore_evas_resize):
3181         (browserCreate):
3182         * MiniBrowser/efl/main.c:
3183         (_MiniBrowser):
3184         (on_ecore_evas_resize):
3185         (browserCreate):
3186
3187 2012-09-18  Sergio Villar Senin  <svillar@igalia.com>
3188
3189         [GTK] run-webkit-tests unable to find TestExpectations for WK2
3190         https://bugs.webkit.org/show_bug.cgi?id=96998
3191
3192         Reviewed by Philippe Normand.
3193
3194         We should look for TestExpectations files in all the locations where
3195         we currently look for Skipped files. This will allow
3196         run-webkit-tests to look for TestExpectations files in
3197         platform/gtk-wk2 and platform/wk2 if the "-2" flag is used.
3198
3199         * Scripts/webkitpy/layout_tests/port/gtk.py:
3200         (GtkPort.expectations_files):
3201
3202 2012-09-18  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3203
3204         [Qt] Fix build without the QtQuick module
3205
3206         Reviewed by Simon Hausmann.
3207
3208         * qmake/mkspecs/features/features.prf:
3209
3210 2012-09-18  Stephanie Lewis  <slewis@apple.com>
3211
3212         Build fix after http://trac.webkit.org/projects/webkit/changeset/128852.
3213
3214         Unreviewed.
3215
3216         * DumpRenderTree/mac/DumpRenderTree.mm:
3217         (poseAsClass):
3218
3219 2012-09-18  Allan Sandfeld Jensen  <allan.jensen@nokia.com>
3220
3221         Unreviewed update of email addresses for Berlin QtWebKit office.
3222
3223         * Scripts/webkitpy/common/config/committers.py:
3224
3225 2012-09-17  Zan Dobersek  <zandobersek@gmail.com>
3226
3227         [GTK] fast/loader/display-image-unset-can-block-image-and-can-reload-in-place.html failing after r128645
3228         https://bugs.webkit.org/show_bug.cgi?id=96899
3229
3230         Reviewed by Martin Robinson.
3231
3232         When overriding the 'WebKitDisplayImageKey' preference, use the
3233         'auto-load-images' property of WebKitWebSettings as the property which
3234         should be updated with the corresponding preference value.
3235
3236         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3237         (resetDefaultsToConsistentValues):
3238         * DumpRenderTree/gtk/TestRunnerGtk.cpp:
3239         (TestRunner::overridePreference):
3240
3241 2012-09-17  Pratik Solanki  <psolanki@apple.com>
3242
3243         DumpRenderTree and WebKitTestRunner should compile with -Wundef on Mac
3244         https://bugs.webkit.org/show_bug.cgi?id=96973
3245
3246         Reviewed by Dan Bernstein.
3247
3248         * DumpRenderTree/mac/Configurations/Base.xcconfig:
3249         * WebKitTestRunner/Configurations/Base.xcconfig:
3250         * WebKitTestRunner/PlatformWebView.h: Use #ifdef __OBJC__ and not #if.
3251
3252 2012-09-17  Dirk Pranke  <dpranke@chromium.org>
3253
3254         nrwt: remove "unexpected EOF" warnings
3255         https://bugs.webkit.org/show_bug.cgi?id=96970
3256
3257         Reviewed by Ojan Vafai.
3258
3259         After debugging this a bit, it looks like there aren't any cases
3260         that I can reproduce where a read() of zero indicates something
3261         actually wrong; either it is a prelude to a crash, or a false
3262         negative. So, I'm removing these warnings and adding a comment.
3263
3264         * Scripts/webkitpy/layout_tests/port/server_process.py:
3265         (ServerProcess._wait_for_data_and_update_buffers_using_select):
3266
3267 2012-09-17  Dirk Pranke  <dpranke@chromium.org>
3268
3269         [chromium] ASAN bot is crashing at the end of the run
3270         https://bugs.webkit.org/show_bug.cgi?id=96967
3271
3272         Reviewed by Abhishek Arya.
3273
3274         The ASAN bot is crashing attempting to decode some output into
3275         UTF-8; there's no reason to do this, so let's not do this and
3276         see if something else is going on as well.
3277
3278         * Scripts/webkitpy/layout_tests/port/chromium.py:
3279         (ChromiumPort._get_crash_log):
3280
3281 2012-09-17  Dirk Pranke  <dpranke@chromium.org>
3282
3283         nrwt: --results-directory isn't getting printed properly
3284         https://bugs.webkit.org/show_bug.cgi?id=96965
3285
3286         Reviewed by Ojan Vafai.
3287
3288         options.results_directory isn't actually initialized with the
3289         default values until after we call print_config(), so this
3290         changes things to print the value directly.
3291
3292         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3293         (run):
3294         * Scripts/webkitpy/layout_tests/views/printing.py:
3295         (Printer.print_config):
3296         * Scripts/webkitpy/layout_tests/views/printing_unittest.py:
3297         (Testprinter.test_print_config):
3298
3299 2012-09-17  Rob Buis  <rbuis@rim.com>
3300
3301         [BlackBerry] Enable VIDEO_TRACK
3302         https://bugs.webkit.org/show_bug.cgi?id=96949
3303
3304         Reviewed by Antonio Gomes.
3305
3306         * Scripts/webkitperl/FeatureList.pm:
3307
3308 2012-09-17  Rick Byers  <rbyers@chromium.org>
3309
3310         Add handling of new GestureTapCancel in DRT
3311
3312         https://bugs.webkit.org/show_bug.cgi?id=96183
3313
3314         Reviewed by Antonio Gomes.
3315
3316         * DumpRenderTree/chromium/TestWebPlugin.cpp:
3317         (TestWebPlugin::handleInputEvent):
3318         * DumpRenderTree/chromium/EventSender.cpp:
3319         (EventSender::gestureTapCancel):
3320
3321 2012-09-17  Philip Rogers  <pdr@google.com>
3322
3323         Teach style checker about preprocessor directive indentation rules
3324         https://bugs.webkit.org/show_bug.cgi?id=96874
3325
3326         Reviewed by Adam Barth.
3327
3328         Preprocessor directives (#ifdef, #include, #define, etc.) should not be indented.
3329         This is not explicit in our style guide but is generally followed in our code.
3330         Searching for violations in our codebase shows these are rarely indented:
3331             #include - indented in 6 files
3332             #ifdef - indented in 0 files
3333             #ifndef - indented in 1 file
3334             #define - indented in 11 files
3335             #if - indented in 7 files
3336
3337         * Scripts/webkitpy/style/checkers/cpp.py:
3338         (check_directive_indentation):
3339
3340             This is the simple test where we look for spaces followed by a #.
3341
3342         (check_style):
3343         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
3344
3345             A few tests needed to be modified because they had unintentionally indented
3346             preprocessor directives.
3347
3348         (CppStyleTest.test_build_class.Foo):
3349         (CppStyleTest.test_build_class):
3350         (CppStyleTest.test_build_class.DERIVE_FROM_GOO):
3351         (WebKitStyleTest.test_line_breaking):
3352         (WebKitStyleTest.test_directive_indentation):
3353
3354 2012-09-17  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3355
3356         [Qt] Auto-generate the module pri file for QtWebKit
3357
3358         Reviewed by Simon Hausmann.
3359
3360         * qmake/qt_webkit.pri: Removed.
3361
3362 2012-09-14  Dirk Pranke  <dpranke@chromium.org>
3363
3364         nrwt: --additional-platform-dir is broken on chromium ports
3365         https://bugs.webkit.org/show_bug.cgi?id=96840
3366
3367         Reviewed by Ojan Vafai.
3368
3369         Looks like I broke this when cleaning things up as part of
3370         adding ML support. Fixing and adding a test.
3371
3372         * Scripts/webkitpy/layout_tests/port/chromium.py:
3373         (ChromiumPort.default_baseline_search_path):
3374         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
3375         (ChromiumAndroidPort.default_baseline_search_path):
3376         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
3377         (ChromiumLinuxPort.default_baseline_search_path):
3378         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
3379         (ChromiumMacPort.__init__):
3380         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
3381         (ChromiumWinPort.setup_environ_for_server):
3382         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
3383         (test_check_build):
3384         (test_additional_platform_directory):
3385
3386 2012-09-13  Stephanie Lewis  <slewis@apple.com>
3387
3388         mac-future ignores previous platform TestExpectations.
3389         https://bugs.webkit.org/show_bug.cgi?id=96718.
3390
3391         Reviewed by Dirk Pranke.
3392
3393         TestExpectations match an expectation's configuraton against a list of configurations that are 
3394         valid for that port.  That list does not contain mac-future so none of the expectations can be
3395         applied.
3396
3397         * Scripts/webkitpy/layout_tests/port/apple.py:
3398         (ApplePort._generate_all_test_configurations):
3399
3400 2012-09-14  Adam Barth  <abarth@webkit.org>
3401
3402         Remove webkitPostMessage
3403         https://bugs.webkit.org/show_bug.cgi?id=96577
3404
3405         Reviewed by Ojan Vafai.
3406
3407         Add ENABLE_LEGACY_VENDOR_PREFIXES flag.
3408
3409         * Scripts/webkitperl/FeatureList.pm:
3410         * qmake/mkspecs/features/features.pri:
3411
3412 2012-09-14  Alexey Proskuryakov  <ap@apple.com>
3413
3414         [WK2] webarchive/loading/javascript-url-iframe-crash.html fails
3415         https://bugs.webkit.org/show_bug.cgi?id=96817
3416
3417         Reviewed by Kenneth Rohde Christiansen.
3418
3419         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3420         (WTR::InjectedBundlePage::didFailProvisionalLoadWithErrorForFrame): This callback
3421         got overlooked when adding logging elsewhere.
3422
3423 2012-09-14  Peter Beverloo  <peter@chromium.org>
3424
3425         [Chromium] Build fix for DumpRenderTree following r128628
3426         https://bugs.webkit.org/show_bug.cgi?id=96808
3427
3428         Reviewed by Tony Chang.
3429
3430         When building DumpRenderTree from the Chromium project, a compile error
3431         shows up because including the header ForwardIOStreamsAndroid.h cannot
3432         be located. TestWebKitAPI does this correctly.
3433
3434         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
3435
3436 2012-09-14  Dana Jansens  <danakj@chromium.org>
3437
3438         Change cc-bugs@google.com watchlist to cc-bugs@chromium.org
3439         https://bugs.webkit.org/show_bug.cgi?id=96805
3440
3441         Reviewed by Dirk Pranke.
3442
3443         * Scripts/webkitpy/common/config/committers.py:
3444         * Scripts/webkitpy/common/config/watchlist:
3445
3446 2012-09-14  Peter Beverloo  <peter@chromium.org>
3447
3448         [Chromium] Support the --{in,out,err}-fifo arguments on TestWebKitAPI and webkit_unit_tests
3449         https://bugs.webkit.org/show_bug.cgi?id=96687
3450
3451         Reviewed by Tony Chang.
3452
3453         Android's DumpRenderTree currently supports these arguments, implemented
3454         as part of TestShellAndroid:
3455         http://trac.webkit.org/browser/trunk/Tools/DumpRenderTree/chromium/TestShellAndroid.cpp?rev=128496
3456
3457         They're used by the layout test runner to get the STDOUT and STDERR while
3458         a layout test run is in process, which is a safer alternative to parsing
3459         all the logcat output manually. The implementation can be seen here:
3460         http://trac.webkit.org/browser/trunk/Tools/Scripts/webkitpy/layout_tests/port/chromium_android.py?rev=128496#L590
3461
3462         This patch generalizes parsing of and applying the effects of these arguments
3463         so that they can be used for TestWebKitAPI and webkit_unit_tests as well.
3464         After this patch, this will make it possible to pull out the output-reading
3465         code from Android's layout test port and generalize it so it can be re-used
3466         in the new test-runner for the other two test suites.
3467
3468         This has no effect when compiling and running these tests as part of Chromium
3469         code, which has a much more advanced test-runner that does parse complete log
3470         output, but also directly depends on code licensed under Apache 2.
3471
3472         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
3473         * DumpRenderTree/chromium/TestShellAndroid.cpp:
3474         (platformInit):
3475         * TestWebKitAPI/TestWebKitAPI.gyp/TestWebKitAPI.gyp:
3476
3477 2012-09-14  Zoltan Horvath  <zoltan@webkit.org>
3478
3479         check-webkit-style should not warn in case of NONCOPYABLE and FAST_ALLOCATED macros
3480         https://bugs.webkit.org/show_bug.cgi?id=96656
3481
3482         Reviewed by Adam Barth.
3483
3484         We should allow to use WTF_MAKE_NONCOPYABLE(ClassName) and WTF_MAKE_FAST_ALLOCATED
3485         macros in 1 line.
3486
3487         * Scripts/webkitpy/style/checkers/cpp.py:
3488         (check_style.definitions):
3489         (check_style): Add rule.
3490         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
3491         (WebKitStyleTest.test_line_breaking): Add unittest.
3492
3493 2012-09-14  Christophe Dumez  <christophe.dumez@intel.com>
3494
3495         WebKitTestRunner needs layoutTestController.dumpDatabaseCallbacks
3496         https://bugs.webkit.org/show_bug.cgi?id=57570
3497
3498         Reviewed by Kenneth Rohde Christiansen.
3499
3500         Implement testRunner.dumpDatabaseCallbacks in WebKitTestRunner
3501         and dump the information expected by the test cases.
3502
3503         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
3504         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
3505         (WTR::InjectedBundle::beginTesting): Explicitly set the default
3506         database quota.
3507         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3508         (WTR::securityOriginToStr): New convenience function to convert
3509         a security origin to the string expected in test results.
3510         (WTR):
3511         (WTR::InjectedBundlePage::InjectedBundlePage):
3512         (WTR::InjectedBundlePage::didExceedDatabaseQuota):
3513         (WTR::InjectedBundlePage::didReachApplicationCacheOriginQuota): Minor
3514         refactoring to share code with didExceedDatabaseQuota.
3515         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
3516         (InjectedBundlePage):
3517         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
3518         (WTR::TestRunner::TestRunner):
3519         * WebKitTestRunner/InjectedBundle/TestRunner.h:
3520         (WTR::TestRunner::dumpDatabaseCallbacks):
3521         (WTR::TestRunner::shouldDumpDatabaseCallbacks):
3522         (TestRunner):
3523
3524 2012-09-14  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3525
3526         [Qt] Set force_static_libs_as_shared early enough to be caught by configure
3527
3528         Rubber-stamped by Simon Hausmann.
3529
3530         * qmake/mkspecs/features/default_post.prf:
3531
3532 2012-09-14  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3533
3534         [Qt] Remove old cruft from the qmake build system
3535
3536         Rubber-stamped by Simon Hausmann.
3537
3538         * DumpRenderTree/qt/DumpRenderTree.pro:
3539         * qmake/mkspecs/features/default_post.prf:
3540
3541 2012-09-14  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3542
3543         [Qt] Enable force_static_libs_as_shared for development
3544
3545         Not enabled for buildbots or production-builds (part of Qt5)
3546
3547         Reviewed by Simon Hausmann.
3548
3549         * qmake/mkspecs/features/default_post.prf:
3550
3551 2012-09-14  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3552
3553         [Qt] Make force_static_libs_as_shared work on Mac OS
3554
3555         We had to move a few LIBS += around that were in the wrong place,
3556         and not caught when everything was just linked into the final
3557         QtWebKit library.
3558
3559         Reviewed by Simon Hausmann.
3560
3561         * qmake/mkspecs/features/default_post.prf:
3562         * qmake/mkspecs/features/force_static_libs_as_shared.prf:
3563
3564 2012-09-14  Simon Hausmann  <simon.hausmann@nokia.com>
3565
3566         [Qt] Fix build of WTR on Windows
3567
3568         Reviewed by Tor Arne Vestbø.
3569
3570         WTR/qt/main.cpp includes TestController.h, which lives in just WTR/
3571         With qmake's unix makefile generator as well as with the MingW generator,
3572         the directory where the pro file lives (WTR/) is automatically added to
3573         the include search path, but not so with MSVC. Instead of relying on qmake's
3574         implicit feature here, add WTR/ explicitly to the include search path.
3575
3576         * WebKitTestRunner/Target.pri:
3577
3578 2012-09-14  Christophe Dumez  <christophe.dumez@intel.com>
3579
3580         WebKitTestRunner needs layoutTestController.setStopProvisionalFrameLoads
3581         https://bugs.webkit.org/show_bug.cgi?id=42691
3582
3583         Reviewed by Kenneth Rohde Christiansen.
3584
3585         Add implementation for testrunner.setStopProvisionalFrameLoads()
3586         in WebKitTestRunner.
3587
3588         Original patch by Alexey Proskuryakov.
3589
3590         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
3591         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3592         (WTR::InjectedBundlePage::didStartProvisionalLoadForFrame):
3593         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
3594         (WTR::TestRunner::TestRunner):
3595         * WebKitTestRunner/InjectedBundle/TestRunner.h:
3596         (TestRunner):
3597         (WTR::TestRunner::setStopProvisionalFrameLoads):
3598         (WTR::TestRunner::shouldStopProvisionalFrameLoads):
3599
3600 2012-09-14  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3601
3602         [Qt] Fix handling of debug/release/debug_and_release/build_all
3603
3604         We now pick up the defaults from Qt, in default_pre (CONFIG already
3605         contains the appropriate values for debug and release). We then let
3606         the command line arguments to qmake, or the project files themselves,
3607         override the configuration, and finally we sanitize the configuration
3608         in default_post.
3609
3610         Reviewed by Simon Hausmann.
3611
3612         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
3613         * Scripts/webkitdirs.pm:
3614         (buildQMakeProjects):
3615         * qmake/mkspecs/features/default_post.prf:
3616         * qmake/mkspecs/features/default_pre.prf:
3617         * qmake/mkspecs/features/mac/default_pre.prf:
3618
3619 2012-09-14  Sudarsana Nagineni  <sudarsana.nagineni@linux.intel.com>
3620
3621         WebKitTestRunner needs layoutTestController.setTabKeyCyclesThroughElements
3622         https://bugs.webkit.org/show_bug.cgi?id=42687
3623
3624         Reviewed by Kenneth Rohde Christiansen.
3625
3626         Implement testRunner.setTabKeyCyclesThroughElements for
3627         WebKitTestRunner.
3628
3629         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
3630         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
3631         (WTR::InjectedBundle::beginTesting):
3632         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
3633         (WTR::TestRunner::setTabKeyCyclesThroughElements):
3634         (WTR):
3635         * WebKitTestRunner/InjectedBundle/TestRunner.h: