[Gtk] Several tests launch g_object_ref/unref assertion messages
[WebKit-https.git] / Tools / ChangeLog
1 2012-10-02  Zan Dobersek  <zandobersek@gmail.com>
2
3         [Gtk] Several tests launch g_object_ref/unref assertion messages
4         https://bugs.webkit.org/show_bug.cgi?id=95062
5
6         Reviewed by Martin Robinson.
7
8         Only increase and decrease reference count of the accessibility
9         UI element if it actually exists.
10
11         * DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
12         (AccessibilityUIElement::AccessibilityUIElement):
13         (AccessibilityUIElement::~AccessibilityUIElement):
14
15 2012-10-02  Douglas Stockwell  <dstockwell@chromium.org>
16
17         [chromium] running layout tests on mountainlion should use chromium TestExpectations
18         https://bugs.webkit.org/show_bug.cgi?id=98119
19
20         Reviewed by Adam Barth.
21
22         Some basic support for mountainlion has already been committed but it failed to refer
23         to the chromium TestExpectations.
24
25         * Scripts/webkitpy/layout_tests/port/chromium.py:
26         (ChromiumPort):
27         * Scripts/webkitpy/layout_tests/port/chromium_port_testcase.py:
28         (ChromiumPortTestCase.test_all_test_configurations):
29
30 2012-10-02  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
31
32         webkitpy: Stop listening on port 8081 when using Apache.
33         https://bugs.webkit.org/show_bug.cgi?id=98155
34
35         Reviewed by Eric Seidel.
36
37         No test currently uses this port, and the lighttpd configuration
38         does not open it.
39
40         This helps the efforts of bringing some consistency to the Apache
41         configuration at the moment by opening the same ports in the
42         webkitpy side and in the httpd.conf side (the idea is to remove
43         the httpd.conf side later), and also makes the behavior consistent
44         across the two servers we currently support.
45
46         * Scripts/webkitpy/layout_tests/servers/apache_http_server.py:
47         (LayoutTestApacheHttpd.__init__):
48
49 2012-10-02  Csaba Osztrogonác  <ossy@webkit.org>
50
51         [WK2] REGRESSION(r128623): It made layout tests extremely slow
52         https://bugs.webkit.org/show_bug.cgi?id=96862
53
54         Reviewed by Simon Hausmann.
55
56         Disable calling the extremely slow WKBundleSetDatabaseQuota() between tests until proper fix.
57
58         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
59         (WTR::InjectedBundle::beginTesting):
60
61 2012-10-02  Philip Rogers  <pdr@google.com>
62
63         Fix PerfTest standard deviation calculation.
64         https://bugs.webkit.org/show_bug.cgi?id=98115
65
66         Reviewed by Ryosuke Niwa.
67
68         Previously our standard deviation calculation was incorrect. This patch
69         updates perftest.py's algorithm to calculate the sample standard deviation
70         (with Bessel's correction) using Knuth's online algorithm:
71         http://en.wikipedia.org/wiki/Algorithms_for_calculating_variance#Online_algorithm
72         An existing test has been modified to prove our new results.
73
74         This patch also updates runner.js to use Bessel's correction in
75         its sample standard deviation calculation, which is more accurate
76         for small sample sizes.
77
78         Additionally, runner.js has been modified to not calculate
79         the 'sum' statistic, which was not very useful.
80
81         * Scripts/webkitpy/performance_tests/perftest.py:
82
83             The unused variable valueSum has also been removed.
84
85         (PageLoadingPerfTest.run):
86         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
87
88             This test calculates the stdev of {2000, 3000, ..., 20000} which
89             was hand-calculated using a spreadsheet.
90
91         (TestPageLoadingPerfTest.test_run):
92
93 2012-10-01  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
94
95         webkitpy should accept a different httpd.conf specified by the user
96         https://bugs.webkit.org/show_bug.cgi?id=98071
97
98         Reviewed by Dirk Pranke.
99
100         The existing httpd.conf variants (or lighttpd.conf, for that
101         matter) we have do not always suit the user's system. This is
102         particularly true on Linux/Unix, where Apache can be installed in
103         a plethora of ways and the LoadModule calls can fail to specify
104         the proper module paths.
105
106         For now, we start accepting the WEBKIT_HTTP_SERVER_CONF_PATH
107         environment variable, which allows the user to specify the
108         absolute path to another http server configuration file that might
109         work on the user's system.
110
111         In the long term, we should try to generate our configuration file
112         and stop requiring all the different httpd.conf files we have as
113         well as this hack.
114
115         * Scripts/webkitpy/layout_tests/port/base.py:
116         (Port._path_to_apache_config_file):
117         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
118         (test_path_to_apache_config_file):
119
120 2012-10-01  Emil A Eklund  <eae@chromium.org>
121
122         Unreviewed, upgrade eae to reviewer. 
123         http://lists.webkit.org/mailman/private/webkit-committers/2012-October/000191.html
124
125         * Scripts/webkitpy/common/config/committers.py:
126
127 2012-10-01  Ming Xie  <mxie@rim.com>
128
129         [BlackBerry] Add a early return in buildCMakeProjectOrExit()
130         https://bugs.webkit.org/show_bug.cgi?id=98052
131
132         Reviewed by Rob Buis.
133
134         This option allows our build to only run the CMake build configuration
135         step when environment variable GENERATE_CMAKE_PROJECT_ONLY is specified.
136
137         * Scripts/webkitdirs.pm:
138         (buildCMakeProjectOrExit):
139
140 2012-10-01  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
141
142         [webkitpy] Detect the multiple names of the wdiff binary in the Port class.
143         https://bugs.webkit.org/show_bug.cgi?id=98039
144
145         Reviewed by Dirk Pranke.
146
147         Unify the duplicate checks for the wdiff binary present in
148         different ports into the base Port class.
149
150         * Scripts/webkitpy/layout_tests/port/base.py:
151         (Port):
152         (Port._path_to_wdiff):
153         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
154         (ChromiumLinuxPort._path_to_wdiff): Removed.
155         * Scripts/webkitpy/layout_tests/port/gtk.py:
156         (GtkPort._path_to_wdiff): Removed.
157
158 2012-10-01  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
159
160         [chromium] Remove custom implementation of _path_to_apache_config_file in chromium_linux.
161         https://bugs.webkit.org/show_bug.cgi?id=98042
162
163         Reviewed by Dirk Pranke.
164
165         The Chromium-based ports inherit from Port these days, so get rid
166         of the custom implementation of _path_to_apache_config_file() in
167         favor of the more portable one in the Port class.
168
169         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
170         (ChromiumLinuxPort._path_to_apache_config_file):
171
172 2012-10-01  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
173
174         [chromium] Remove duplicated _is_redhat_based() implementation from chromium_linux.py.
175         https://bugs.webkit.org/show_bug.cgi?id=98035
176
177         Reviewed by Dirk Pranke.
178
179         Since the Chromium port classes inherit from Port these days, it
180         can rely on the same implementation of _is_redhat_based() present
181         there instead of duplicating it.
182
183         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
184         (ChromiumLinuxPort._is_redhat_based): Removed.
185
186 2012-10-01  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
187
188         [GTK] Rely on the general implementation for detecting the Apache server path.
189         https://bugs.webkit.org/show_bug.cgi?id=98033
190
191         Reviewed by Martin Robinson.
192
193         The implemenentation of _path_to_apache() in base.py already
194         checks for the same paths gtk.py tries; the only difference is
195         that all of the are tried regardless of the current distro.
196
197         Doing so is more portable, and lets us remove some distro-checking
198         from the code.
199
200         * Scripts/webkitpy/layout_tests/port/gtk.py:
201         (GtkPort._path_to_apache): Remove.
202
203 2012-10-01  Csaba Osztrogonác  <ossy@webkit.org>
204
205         Unreviewed, update the URL of the Szeged SVN mirror.
206
207         * BuildSlaveSupport/build.webkit.org-config/config.json:
208
209 2012-10-01  Yuta Kitamura  <yutak@chromium.org>
210
211         Unreviewed. Create WebSocket watch list, and add myself to that list.
212
213         * Scripts/webkitpy/common/config/watchlist:
214
215 2012-09-30  Taiju Tsuiki  <tzik@chromium.org>
216
217         Unreviewed. Adding myself as a committer.
218
219         * Scripts/webkitpy/common/config/committers.py:
220
221 2012-09-30  Sadrul Habib Chowdhury  <sadrul@chromium.org>
222
223         [chromium] Add a test to verify that a plugin does not receive touch-events unless it explicitly requests for them
224         https://bugs.webkit.org/show_bug.cgi?id=97975
225
226         Reviewed by Adam Barth.
227
228         TestWebPlugin can now be configured to accept touch events or not. Use this
229         configurability to verify that a plugin does not receive touch events unless
230         it explicitly requests for them.
231
232         * DumpRenderTree/chromium/TestWebPlugin.cpp:
233         (TestWebPlugin::TestWebPlugin):
234         (TestWebPlugin::initialize):
235         (TestWebPlugin::parseBoolean):
236         * DumpRenderTree/chromium/TestWebPlugin.h:
237         (TestWebPlugin):
238
239 2012-09-28  Mariusz Grzegorczyk  <mariusz.g@samsung.com>
240
241         [WK2][GTK][EFL] Share WebKit2-GTK plugin process implementation with EFL port
242         https://bugs.webkit.org/show_bug.cgi?id=91844
243
244         Reviewed by Simon Hausmann.
245
246         Turn on Netscape Plugin API option by default.
247
248         * Scripts/webkitperl/FeatureList.pm:
249
250 2012-09-28  Dirk Pranke  <dpranke@chromium.org>
251
252         webkit-patch rebaseline-expectations is broken
253         https://bugs.webkit.org/show_bug.cgi?id=97619
254
255         Reviewed by Ojan Vafai.
256
257         The new expectations format is causing two problems for
258         webkit-patch rebaseline expectations
259
260         1) the transformation into and out of the old syntax (which is
261         still used internally) is somewhat lossy, e.g., we're not
262         preserving the case of Bug(x) identifiers. Also, we can't
263         tell if the input was [ WontFix ] or [ Skip WontFix Pass ]
264         2) the new syntax is more lenient, allowing for multiple ways to
265         specify the same result, e.g., "[ Skip ]" may or may not be
266         missing.
267
268         As a result, we end up rewriting a whole bunch of lines for
269         purely syntactic differences.
270
271         Both of these things are bugs that need to be fixed, but
272         we should also not just rewrite lines that shouldn't be
273         modified.
274
275         This patch fixes the last issue (no rewriting).
276
277         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
278         (TestExpectationLine._format_line):
279         (TestExpectations.remove_rebaselined_tests):
280         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
281         (RebaseliningTest.test_remove):
282
283 2012-09-28  Dirk Pranke  <dpranke@chromium.org>
284
285         virtual directories should just rename the beginning of the test path
286         https://bugs.webkit.org/show_bug.cgi?id=97925
287
288         Reviewed by Ojan Vafai.
289
290         We were replacing all occurrences of the base prefix of a
291         virtual test suite with the virtual name, instead of just the
292         first. So, where we had softwarecompositing running the
293         compositing tests, any compositing test that actually had
294         "compositing" in the test base name would get that expanded
295         incorrectly.
296
297         * Scripts/webkitpy/layout_tests/port/base.py:
298         (Port.populated_virtual_test_suites):
299         (Port.lookup_virtual_test_base):
300         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
301         (PortTest.test_tests):
302         * Scripts/webkitpy/layout_tests/port/test.py:
303
304 2012-09-28  Ojan Vafai  <ojan@chromium.org>
305
306         Garden-o-matic sometimes things buildbot warnings are failures
307         https://bugs.webkit.org/show_bug.cgi?id=97926
308
309         Reviewed by Tony Chang.
310
311         It turns out the JSON clearly tells us if it's a failure.
312         We want to skip all warnings even if they don't have "warning" in the text.
313         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/builders.js:
314         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/builders_unittests.js:
315
316 2012-09-28  Harald Tveit Alvestrand  <harald@alvestrand.no>
317
318         Implement the GetStats interface on PeerConnection
319         https://bugs.webkit.org/show_bug.cgi?id=95193
320
321         Reviewed by Adam Barth.
322
323         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
324         (RTCStatsRequestSucceededTask):
325         (RTCStatsRequestSucceededTask::RTCStatsRequestSucceededTask):
326         (MockWebRTCPeerConnectionHandler::getStats):
327         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.h:
328         (MockWebRTCPeerConnectionHandler):
329
330 2012-09-28  Kristóf Kosztyó  <kkristof@inf.u-szeged.hu>
331
332         [NRWT] XvfbDriver should choose the next free display
333         https://bugs.webkit.org/show_bug.cgi?id=88414
334
335         Reviewed by Dirk Pranke.
336
337         This change how the XvfbDriver choose the next display. Before
338         this the choosing are based on the worker number but it caused
339         errors when several nrwt run on the same time. This add process
340         based free diplay search.
341
342         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py:
343         (XvfbDriver.__init__):
344         (XvfbDriver._next_free_display):
345         (XvfbDriver._start):
346         (XvfbDriver.stop):
347         * Scripts/webkitpy/layout_tests/port/xvfbdriver_unittest.py:
348         (XvfbDriverTest.make_driver):
349         (XvfbDriverTest.test_start_no_pixel_tests):
350         (XvfbDriverTest.test_start_pixel_tests):
351         (XvfbDriverTest.test_start_arbitrary_worker_number):
352         (XvfbDriverTest.test_next_free_display):
353         (XvfbDriverTest):
354         (XvfbDriverTest.test_start_next_worker):
355
356 2012-09-28  Huang Dongsung  <luxtella@company100.net>
357
358         [GTK] Enable CSS Shaders layout LayoutTests on GTK+
359         https://bugs.webkit.org/show_bug.cgi?id=97821
360
361         Reviewed by Martin Robinson.
362
363         This feature is disabled via Settings by default, but for testing,
364         we enable it using layoutTestController.overridePreferences. Add the
365         necessary plumbing for DRT.
366
367         WTR already works because support was added for Apple Mac earlier.
368
369         * DumpRenderTree/gtk/DumpRenderTree.cpp:
370         (resetDefaultsToConsistentValues): Feature is off by default.
371         * DumpRenderTree/gtk/TestRunnerGtk.cpp:
372         (TestRunner::overridePreference): Add handling of WebKitCSSCustomFilterEnabled.
373
374 2012-09-28  Csaba Osztrogonác  <ossy@webkit.org>
375
376         [Qt] Make NRWT use wk2/TestExpectations file
377         https://bugs.webkit.org/show_bug.cgi?id=97894
378
379         Reviewed by Simon Hausmann.
380
381         * Scripts/webkitpy/layout_tests/port/qt.py: Remove unnecessary _skipped_file_search_paths function.
382         (QtPort.expectations_files): Add wk2 if --webkit-test-runner option added.
383         * Scripts/webkitpy/layout_tests/port/qt_unittest.py: Remove unnecessary test__skipped_file_search_paths function.
384         (QtPortTest._assert_search_path): Removed.
385         (QtPortTest.test_expectations_files): Add wk2 if --webkit-test-runner option added, and code cleanup.
386
387 2012-09-28  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
388
389         [EFL][DRT] EFL's layoutTestController.keepWebHistory should enable visited links tracking
390         https://bugs.webkit.org/show_bug.cgi?id=97901
391
392         Reviewed by Kenneth Rohde Christiansen.
393
394         EFL's DRT testRunner.keepWebHistory should enable visited links tracking in order
395         to be consistent with WTR implementation and also several tests rely on that (fast/history).
396
397         * DumpRenderTree/efl/TestRunnerEfl.cpp:
398         (TestRunner::keepWebHistory):
399
400 2012-09-28  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
401
402         [EFL][DRT] EFL DRT should disable visited links tracking by default
403         https://bugs.webkit.org/show_bug.cgi?id=97881
404
405         Reviewed by Kenneth Rohde Christiansen.
406
407         Visited links tracking is disabled by default for each test.
408
409         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
410         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
411
412 2012-09-28  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
413
414         Unreviewed, rolling out r129863.
415         http://trac.webkit.org/changeset/129863
416         https://bugs.webkit.org/show_bug.cgi?id=97173
417
418         Broken debug WK2 layout test
419
420         * MiniBrowser/efl/main.c:
421         (quit):
422         (main):
423         * WebKitTestRunner/efl/PlatformWebViewEfl.cpp:
424         (WTR::initEcoreEvas):
425         (WTR::PlatformWebView::~PlatformWebView):
426         * WebKitTestRunner/efl/main.cpp:
427         (main):
428
429 2012-09-28  Kai Koehne  <kai.koehne@digia.com>
430
431         [Qt] Use qInstallMessageHandler()
432         https://bugs.webkit.org/show_bug.cgi?id=96648
433
434         Reviewed by Jocelyn Turcotte.
435
436         qInstallMsgHandler() got deprecated in Qt 5.
437
438         * DumpRenderTree/qt/DumpRenderTreeMain.cpp:
439         (messageHandler):
440         (main):
441         * WebKitTestRunner/qt/main.cpp:
442         (messageHandler):
443         (main):
444
445 2012-09-28  Eunmi Lee  <eunmi15.lee@samsung.com>
446
447         [EFL][WK2] Refactoring initialization and shutdown codes of EFL libraries.
448         https://bugs.webkit.org/show_bug.cgi?id=97173
449
450         Reviewed by Gyuyoung Kim.
451
452         The initialization and shutdown are done inside the webkit, so we don't
453         have to call them out of the webkit.
454
455         * MiniBrowser/efl/main.c:
456         (quit):
457         (main):
458         * WebKitTestRunner/efl/PlatformWebViewEfl.cpp:
459         (WTR::initEcoreEvas):
460         (WTR::PlatformWebView::~PlatformWebView):
461         * WebKitTestRunner/efl/main.cpp:
462         (main):
463
464 2012-09-27  Alexandre Elias  <aelias@chromium.org>
465
466         [chromium] DumpRenderTree support for software compositing
467         https://bugs.webkit.org/show_bug.cgi?id=96853
468
469         Reviewed by James Robinson.
470
471         This adds software compositor support to DumpRenderTree.  The only
472         thing needed is a minimal OutputSurfaceSoftware that holds a
473         viewport-sized bitmap. Then, I added virtual test suite
474         platform/chromium/virtual/softwarecompositing -- a mirror of the
475         compositing/ directory with --enable-software-compositing, and
476         expectations for not yet supported tests.
477
478         * DumpRenderTree/DumpRenderTree.gypi:
479         * DumpRenderTree/chromium/DumpRenderTree.cpp:
480         (main):
481         * DumpRenderTree/chromium/TestShell.cpp:
482         (TestShell::TestShell):
483         * DumpRenderTree/chromium/TestShell.h:
484         (TestShell::softwareCompositingEnabled):
485         (TestShell::setSoftwareCompositingEnabled):
486         (TestShell):
487         * DumpRenderTree/chromium/WebViewHost.cpp:
488         (WebViewHost::createOutputSurface):
489         * DumpRenderTree/chromium/WebViewHostOutputSurface.cpp:
490         (WebKit::WebViewHostOutputSurface::create3d):
491         (WebKit):
492         (WebKit::WebViewHostOutputSurface::createSoftware):
493         (WebKit::WebViewHostOutputSurface::WebViewHostOutputSurface):
494         (WebKit::WebViewHostOutputSurface::bindToClient):
495         (WebKit::WebViewHostOutputSurface::surfaceSoftware):
496         * DumpRenderTree/chromium/WebViewHostOutputSurface.h:
497         (WebKit):
498         (WebViewHostOutputSurface):
499         * DumpRenderTree/chromium/WebViewHostSoftwareOutputDevice.cpp: Copied from Tools/DumpRenderTree/chromium/WebViewHostOutputSurface.cpp.
500         (WebKit):
501         (WebKit::WebViewHostSoftwareOutputDevice::lockForWrite):
502         (WebKit::WebViewHostSoftwareOutputDevice::unlockForWrite):
503         (WebKit::WebViewHostSoftwareOutputDevice::lockForRead):
504         (WebKit::WebViewHostSoftwareOutputDevice::unlockForRead):
505         (WebKit::WebViewHostSoftwareOutputDevice::viewportChanged):
506         * DumpRenderTree/chromium/WebViewHostSoftwareOutputDevice.h: Copied from Tools/DumpRenderTree/chromium/WebViewHostOutputSurface.cpp.
507         (WebKit):
508         (WebViewHostSoftwareOutputDevice):
509         * Scripts/webkitpy/layout_tests/port/chromium.py:
510         (ChromiumPort.virtual_test_suites):
511
512 2012-09-27  Nasko Oskov  <nasko@chromium.org>
513
514         Move DumpRenderTree to use newer version of the name method of
515         WebFrame and the willCheckAndDispatchMessageEvent method of WebFrameClient
516         https://bugs.webkit.org/show_bug.cgi?id=95254
517
518         Reviewed by Adam Barth.
519
520         The WebFrame name() method is removed and replaced by better named
521         uniqueName() method, so DRT needs to be updated to use this new version.
522         Also the willCheckAndDispatchMessageEvent in WebFrameClient gained
523         one more parameter, which had to be added in DRT.
524
525         * DumpRenderTree/chromium/TestShell.cpp:
526         (dumpFramesAsText):
527         (dumpFramesAsPrintedText):
528         (dumpFrameScrollPosition):
529         * DumpRenderTree/chromium/WebViewHost.cpp:
530         (WebViewHost::unableToImplementPolicyWithError):
531         (WebViewHost::willCheckAndDispatchMessageEvent):
532         (WebViewHost::printFrameDescription):
533         * DumpRenderTree/chromium/WebViewHost.h:
534         (WebViewHost):
535
536 2012-09-27  Julie Parent  <jparent@chromium.org>
537
538         Add webkit_unit_tests to the flakiness dashboard
539         https://bugs.webkit.org/show_bug.cgi?id=97814
540
541         Reviewed by Dirk Pranke.
542
543         webkit_unit_tests run on the same bots as test_shell_tests.
544
545         * TestResultServer/static-dashboards/builders.js:
546         (loadBuildersList):
547         * TestResultServer/static-dashboards/dashboard_base.js:
548         (currentBuilderGroupCategory):
549
550 2012-09-27  Florin Malita  <fmalita@chromium.org>
551
552         [Perftest] PerfTestRunner exception when using nested git checkouts
553         https://bugs.webkit.org/show_bug.cgi?id=97796
554
555         Reviewed by Ryosuke Niwa.
556
557         In order to support separate Chromium vs. WebKit git trees, PerfTestsRunner needs to
558         instantiate per-repository SCM objects when collecting revision information.
559
560         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
561         (PerfTestsRunner._generate_results_dict):
562
563 2012-09-27  Sudarsana Nagineni  <sudarsana.nagineni@intel.com>
564
565         [Mac][DRT] Add support to allow non-array modifier arguments in keyDown()
566         https://bugs.webkit.org/show_bug.cgi?id=97805
567
568         Reviewed by Alexey Proskuryakov.
569
570         Add support for handling a string as modifier argument to keyDown().
571         This support has already been added to other ports, so implementing
572         this on Mac too, to be consistent with the other ports.
573
574         * DumpRenderTree/mac/EventSendingController.mm:
575         (modifierFlags):
576         (buildModifierFlags):
577
578 2012-09-27  Dirk Pranke  <dpranke@chromium.org>
579
580         efl wk2 bots weren't looking at platform/wk2/TestExpectations
581         https://bugs.webkit.org/show_bug.cgi?id=97815
582
583         Unreviewed, build fix.
584
585         Should clean up this logic later so that we're just using the
586         implementation in base ...
587
588         * Scripts/webkitpy/layout_tests/port/efl.py:
589         (EflPort._search_paths):
590         (EflPort.expectations_files):
591
592 2012-09-27  Dirk Pranke  <dpranke@chromium.org>
593
594         webkitpy: make warning about missing BUG identifiers per-port configurable
595         https://bugs.webkit.org/show_bug.cgi?id=97751
596
597         Reviewed by Ojan Vafai.
598
599         Prior to this change linting TestExpectations files (checking their style)
600         would raise warnings if lines were missing a BUG modifier, even
601         though such a modifier is actually optional in the syntax. As we
602         migrate the entries over from the Skipped files, these warnings
603         will generate a lot of noise, and it's not yet clear that we
604         actually want to require the modifiers on all ports.
605
606         So, this change makes this per-port configurable so that they
607         will generate warnings in the Chromium TestExpectations and not
608         anywhere else.
609
610         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
611         (TestExpectationParser._parse_modifiers):
612         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
613         (MiscTests.test_parse_warning):
614         (SemanticTests.test_missing_bugid):
615         * Scripts/webkitpy/layout_tests/port/base.py:
616         (Port.warn_if_bug_missing_in_test_expectations):
617         * Scripts/webkitpy/layout_tests/port/chromium.py:
618         (ChromiumPort.warn_if_bug_missing_in_test_expectations):
619
620 2012-09-27  Xianzhu Wang  <wangxianzhu@chromium.org>
621
622         [Chromium-Android] Push more fonts for layout tests to match chromium-linux
623         https://bugs.webkit.org/show_bug.cgi?id=97746
624
625         Reviewed by Dirk Pranke.
626
627         * DumpRenderTree/chromium/android_fallback_fonts.xml: Added the new fonts into fallback list. Removed the Android system fallback font DroidSansFallback.ttf.
628         * Scripts/webkitpy/layout_tests/port/builders.py: Add chromium-android to pass style check.
629         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
630         (ChromiumAndroidPort.check_build): Added check for md5sum and forwarder utilities to give better error message when they are missing.
631         (ChromiumAndroidPort.check_sys_deps): Allow host font files to come from one of multiple directories.
632         (ChromiumAndroidPort._path_to_md5sum):
633         (ChromiumAndroidDriver._setup_md5sum_and_push_data_if_needed):
634         (ChromiumAndroidDriver._push_fonts):
635         (ChromiumAndroidDriver._push_test_resources):
636
637 2012-09-27  Tommy Widenflycht  <tommyw@google.com>
638
639         MediaStream API: Enhance MediaConstraints to make it easier to get the constraint data
640         https://bugs.webkit.org/show_bug.cgi?id=97559
641
642         Reviewed by Adam Barth.
643
644         Refactoring for the new WebMediaConstraints api.
645
646         * DumpRenderTree/chromium/MockConstraints.cpp:
647         (MockConstraints::verify):
648
649 2012-09-27  Tommy Widenflycht  <tommyw@google.com>
650
651         MediaStream API: Rename MockConstraints::verify to verifyConstraints
652         https://bugs.webkit.org/show_bug.cgi?id=97779
653
654         Reviewed by Csaba Osztrogonác.
655
656         The Mac OS X sdk has a macro called verify :/
657
658         * DumpRenderTree/chromium/MockConstraints.cpp:
659         (MockConstraints::verifyConstraints):
660         * DumpRenderTree/chromium/MockConstraints.h:
661         (MockConstraints):
662         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
663         (MockWebRTCPeerConnectionHandler::initialize):
664         * DumpRenderTree/chromium/WebUserMediaClientMock.cpp:
665         (WebUserMediaClientMock::requestUserMedia):
666
667 2012-09-27  Tommy Widenflycht  <tommyw@google.com>
668
669         MediaStream API: Update getUserMedia to match the latest specification
670         https://bugs.webkit.org/show_bug.cgi?id=97540
671
672         Reviewed by Adam Barth.
673
674         http://dev.w3.org/2011/webrtc/editor/getusermedia.html#navigatorusermedia
675         Navigator.getUserMedia is changed so that the audio and video members can either
676         be a bool or a constraints object.
677
678         * public/WebUserMediaRequest.h:
679         (WebKit):
680         (WebUserMediaRequest):
681         * src/WebUserMediaRequest.cpp:
682         (WebKit::WebUserMediaRequest::audioConstraints):
683         (WebKit):
684         (WebKit::WebUserMediaRequest::videoConstraints):
685
686 2012-09-26  Laszlo Gombos  <l.gombos@samsung.com>
687
688         Unreviewed. Update my e-mail address.
689
690         * Scripts/webkitpy/common/config/committers.py:
691
692 2012-09-26  Dirk Pranke  <dpranke@chromium.org>
693
694         modify old-run-webkit-tests to support TestExpectations files a little
695         https://bugs.webkit.org/show_bug.cgi?id=97276
696
697         Reviewed by Daniel Bates.
698
699         This patch adds minimal support to old-run-webkit-tests so that
700         it can read a TestExpectations file and skip the tests listed
701         there. It will skip *all* tests, regardless of the expectation
702         (even if the test is listed as [ Pass ]), and it will only look
703         at the TestExpectations file if there isn't a Skipped file next
704         to it.
705         
706         This will allow us to switch to using just using
707         TestExpectations files without making old-run-webkit-tests
708         totally useless.
709
710         * Scripts/old-run-webkit-tests:
711         (readSkippedFiles):
712         (processSkippedFileEntry): Added.
713         (startsWith): Added.
714
715 2012-09-26  Dirk Pranke  <dpranke@chromium.org>
716
717         The style bot spams about skia_test_expectations.txt
718         https://bugs.webkit.org/show_bug.cgi?id=97699
719
720         Reviewed by Adam Barth.
721
722         Remove the warning about a missing skia_test_expectations.txt
723         file for now. There are some situations where it would be
724         helpful to get this warning as it might indicate that the user
725         is going to get results different than what they're expecting,
726         but at the moment it looks difficult if not impossible to get
727         those warnings and not also get warnings that we don't care
728         about, so the noise is outweighing the value of the signal.
729
730         This at least will stop the false negatives we're getting on the
731         style bots. If it turns out people do need the warning we can
732         add it back in in some other more conditional way.
733
734         * Scripts/webkitpy/layout_tests/port/chromium.py:
735         (ChromiumPort.expectations_files):
736         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
737         (TestRebaseline.test_rebaseline_updates_expectations_file_noop):
738         (test_rebaseline_updates_expectations_file):
739         (test_rebaseline_does_not_include_overrides):
740         (test_rebaseline_expectations):
741         (_assert_command):
742
743 2012-09-26  David Michael Barr  <davidbarr@chromium.org>
744
745         Unreviewed. Upgrade myself from contributor to committer.
746
747         * Scripts/webkitpy/common/config/committers.py:
748
749 2012-09-26  Ryosuke Niwa  <rniwa@webkit.org>
750
751         Remove deprecated code from perftestrunner.py
752         https://bugs.webkit.org/show_bug.cgi?id=97724
753
754         Reviewed by Adam Barth.
755
756         Remove the code deprecated in r129580 now that we have landed r129597,
757         which updated master.cfg to use new options, and restarted the buildbot master.
758
759         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
760         (PerfTestsRunner._parse_args):
761         (PerfTestsRunner._generate_and_show_results):
762         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
763         (test_run_with_slave_config_json):
764         (test_parse_args):
765
766 2012-09-26  Beth Dakin  <bdakin@apple.com>
767
768         https://bugs.webkit.org/show_bug.cgi?id=97629
769         http/tests/security/cross-frame-access-put.html failing on Mac WK2 bots after 
770         being unskipped
771
772         Reviewed by Tim Horton.
773
774         The test was failing because we were failing to use the fake window origin. We 
775         should really always use the fake origin. This patch eliminates 
776         _shouldUseFakeOrigin in favor of always using it.
777         * WebKitTestRunner/mac/PlatformWebViewMac.mm:
778         (-[WebKitTestRunnerWindow setFrameOrigin:]):
779         (-[WebKitTestRunnerWindow setFrame:display:animate:]):
780         (-[WebKitTestRunnerWindow setFrame:display:]):
781         (-[WebKitTestRunnerWindow frameRespectingFakeOrigin]):
782
783 2012-09-26  Tony Chang  <tony@chromium.org>
784
785         C++ style checker should warn when the indentation is wrong
786         https://bugs.webkit.org/show_bug.cgi?id=97602
787
788         Reviewed by Ojan Vafai.
789
790         Rewrite the indentation checker to ensure that indentation is always a factor of 4
791         and that we only indent 4 spaces from the previous line.
792
793         I deleted some of the old indentation checking code since it was filtered out
794         and didn't match the webkit style.
795
796         * Scripts/webkitpy/style/checker.py:
797         * Scripts/webkitpy/style/checkers/cpp.py:
798         (check_indentation_amount): New function that checks indentation. Intentionally runs after
799         other indentation checks.
800         (check_style): Delete the old indentation code since we were already filtering out the whitespace/labels warnings.
801         (CppChecker): Remove the whitespace/labels filter.
802         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
803         (CppStyleTest.test_multi_line_comments): Force 4 space indent.
804         (CppStyleTest.test_explicit_single_argument_constructors):
805         (CppStyleTest.test_explicit_single_argument_constructors.Foo):
806         (CppStyleTest.test_explicit_single_argument_constructors.Qualifier):
807         (CppStyleTest.test_slash_star_comment_on_single_line):
808         (CppStyleTest.test_braces):
809         (CppStyleTest.test_spacing_before_last_semicolon):
810         (CppStyleTest.test_static_or_global_stlstrings):
811         (CppStyleTest.test_indent): Added macro test case.
812         (CppStyleTest.test_build_class):
813         (NoNonVirtualDestructorsTest.test_no_error):
814         (NoNonVirtualDestructorsTest.test_no_error.Foo):
815         (NoNonVirtualDestructorsTest.test_no_error.Qualified):
816         (NoNonVirtualDestructorsTest.test_no_destructor_when_virtual_needed):
817         (NoNonVirtualDestructorsTest.test_no_destructor_when_virtual_needed.Foo):
818         (NoNonVirtualDestructorsTest.test_destructor_non_virtual_when_virtual_needed):
819         (NoNonVirtualDestructorsTest.test_destructor_non_virtual_when_virtual_needed.Foo):
820         (NoNonVirtualDestructorsTest.test_no_warn_when_derived):
821         (NoNonVirtualDestructorsTest.test_no_warn_when_derived.Foo):
822         (NoNonVirtualDestructorsTest.test_internal_braces):
823         (NoNonVirtualDestructorsTest.test_internal_braces.Foo):
824         (NoNonVirtualDestructorsTest.test_inner_class_needs_virtual_destructor):
825         (NoNonVirtualDestructorsTest.test_inner_class_needs_virtual_destructor.Foo):
826         (NoNonVirtualDestructorsTest.test_inner_class_needs_virtual_destructor.Foo.Goo):
827         (NoNonVirtualDestructorsTest.test_outer_class_needs_virtual_destructor):
828         (NoNonVirtualDestructorsTest.test_outer_class_needs_virtual_destructor.Foo):
829         (NoNonVirtualDestructorsTest.test_outer_class_needs_virtual_destructor.Foo.Goo):
830         (NoNonVirtualDestructorsTest.test_qualified_class_needs_virtual_destructor):
831         (NoNonVirtualDestructorsTest.test_qualified_class_needs_virtual_destructor.Qualified):
832         (NoNonVirtualDestructorsTest.test_multi_line_declaration_no_error):
833         (NoNonVirtualDestructorsTest.test_multi_line_declaration_no_error.Foo):
834         (NoNonVirtualDestructorsTest.test_multi_line_declaration_with_error):
835         (NoNonVirtualDestructorsTest.test_multi_line_declaration_with_error.Foo):
836         (WebKitStyleTest.test_indentation):
837         (WebKitStyleTest.test_braces): Add test for call params spanning lines.
838
839 2012-09-26  Adam Barth  <abarth@webkit.org>
840
841         Rollout messages should be clickable in bugzilla
842         https://bugs.webkit.org/show_bug.cgi?id=97711
843
844         Reviewed by Simon Fraser.
845
846         If we say Bug NNNN instead of just NNNN, bugzilla will autolink to the
847         actual bug.
848
849         * Scripts/webkitpy/tool/commands/download_unittest.py:
850         * Scripts/webkitpy/tool/steps/createbug.py:
851         (CreateBug.run):
852
853 2012-09-26  Marcelo Lira  <marcelo.lira@openbossa.org>
854
855         run-perf-tests must expand environment variables in user provided paths
856         https://bugs.webkit.org/show_bug.cgi?id=97686
857
858         Reviewed by Ryosuke Niwa.
859
860         The run-perf-tests command line options that receive paths from the
861         user now expand any possible environment variables, because Python's
862         file handling methods do not handle those.
863
864         Examples:
865         run-perf-tests --platform=qt --release --output-json-path=~/perf-results
866         run-perf-tests --platform=qt --release --output-json-path=$HOME/perf-results
867
868         Also removed unused imports.
869
870         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
871         (PerfTestsRunner._parse_args._expand_path): expand environment variables in a path passed via command line
872         (PerfTestsRunner._parse_args):
873
874 2012-09-26  Sheriff Bot  <webkit.review.bot@gmail.com>
875
876         Unreviewed, rolling out r129654.
877         http://trac.webkit.org/changeset/129654
878         https://bugs.webkit.org/show_bug.cgi?id=97702
879
880         breaks chromium windows build (Requested by schenney on
881         #webkit).
882
883         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
884         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.h:
885         (MockWebRTCPeerConnectionHandler):
886
887 2012-09-26  Michael Saboff  <msaboff@apple.com>
888
889         Unreviewed build change.
890
891         Reverting inadvertent change to project file.
892
893         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
894
895 2012-09-26  Harald Tveit Alvestrand  <harald@alvestrand.no>
896
897         Implement the GetStats interface on PeerConnection
898         https://bugs.webkit.org/show_bug.cgi?id=95193
899
900         Reviewed by Adam Barth.
901
902         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
903         (RTCStatsRequestSucceededTask):
904         (RTCStatsRequestSucceededTask::RTCStatsRequestSucceededTask):
905         (MockWebRTCPeerConnectionHandler::getStats):
906         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.h:
907         (MockWebRTCPeerConnectionHandler):
908
909 2012-09-26  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
910
911         [WK2][WTR] Policy client: dumping from decidePolicyForResponse callback
912         https://bugs.webkit.org/show_bug.cgi?id=97034
913
914         Reviewed by Kenneth Rohde Christiansen.
915
916         Provided dumping from WTR Policy client 'decidePolicyForResponse' callback.
917
918         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
919         (WTR::InjectedBundlePage::decidePolicyForResponse):
920
921 2012-09-26  Zoltan Arvai  <zarvai@inf.u-szeged.hu>
922
923         [Qt][WTR] Build fix for MSVC. 
924         https://bugs.webkit.org/show_bug.cgi?id=97667
925
926         Reviewed by Csaba Osztrogonác.
927
928         WTR misses InjectedBundle headers. Adding $$PWD to Target.pri is required for MSVC. Clean build is necessary on Windows.
929
930         * WebKitTestRunner/InjectedBundle/Target.pri:
931
932 2012-09-26  Xiaobo Wang  <xbwang@torchmobile.com.cn>
933
934         [BlackBerry] [DRT] eventSender.keyDown() doesn't support Ctrl key modifier
935         https://bugs.webkit.org/show_bug.cgi?id=97665
936
937         Reviewed by Rob Buis.
938
939         This patch fixes test fast/forms/input-text-option-delete.html.
940
941         * DumpRenderTree/blackberry/EventSender.cpp:
942         (keyDownCallback):
943
944 2012-09-26  Dinu Jacob  <dinu.s.jacob@intel.com>
945
946         Unreviewed. Adding myself as a committer.
947
948         * Scripts/webkitpy/common/config/committers.py:
949
950 2012-09-26  Carlos Garcia Campos  <cgarcia@igalia.com>
951
952         [GTK] Remove color codes characters from run-gtk-tests output when not writing to a TTY
953         https://bugs.webkit.org/show_bug.cgi?id=97660
954
955         Reviewed by Philippe Normand.
956
957         Remove also the '\r' characters returned by TTY output to try to fix
958         the double line endings of the script output in the bots.
959
960         * Scripts/run-gtk-tests:
961         (TestRunner.__init__): Compile a regexp for color codes if stdout
962         it not a TTY.
963         (TestRunner._run_test_command.parse_line): Remove color codes and
964         '\r' characters from the line returned by the pseudo-terminal if
965         stdout is not a TTY.
966         * gtk/common.py:
967         (parse_output_lines): Fix the last chunk check to make sure we
968         always send entire lines to the callback, otherwise color codes
969         regexp might fail if the line is not complete.
970
971 2012-09-25  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
972
973         [DRT][WTR] Support overriding the 'WebKitDisplayImagesKey' preference
974         https://bugs.webkit.org/show_bug.cgi?id=96883
975
976         Reviewed by Kenneth Rohde Christiansen.
977
978         * DumpRenderTree/efl/TestRunnerEfl.cpp:
979         (TestRunner::overridePreference): Support the
980         'WebKitDisplayImagesKey' preference.
981
982 2012-09-26  Zan Dobersek  <zandobersek@gmail.com>
983
984         [GTK] Enable some of the unstable CSS features
985         https://bugs.webkit.org/show_bug.cgi?id=97572
986
987         Reviewed by Martin Robinson.
988
989         Enable the <style scoped> support via DumpRenderTreeSupportGtk method
990         whenever resetting defualts to consistent values.
991
992         * DumpRenderTree/gtk/DumpRenderTree.cpp:
993         (resetDefaultsToConsistentValues):
994
995 2012-09-26  Zan Dobersek  <zandobersek@gmail.com>
996
997         REGRESSION (r126189): No more mismatch-did-not-occur failures in reftests
998         https://bugs.webkit.org/show_bug.cgi?id=97112
999
1000         Reviewed by Dirk Pranke.
1001
1002         Don't use the reftest's image hash when gathering output of its reference.
1003         This indirectly forces the driver to return the reference's image output
1004         as well. This specifically solves the problem when the reftest and its
1005         mismatch reference are equal but the unexpected match failure is not
1006         reported due to missing image output of the reference file.
1007
1008         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py:
1009         (SingleTestRunner._run_reftest):
1010         * Scripts/webkitpy/layout_tests/port/test.py:
1011         (TestPort.diff_image):
1012         (TestDriver.run_test):
1013
1014 2012-09-25  Ryosuke Niwa  <rniwa@webkit.org>
1015
1016         Update master.cfg to invoke run-perf-tests with --no-show-results and --reset-results
1017         https://bugs.webkit.org/show_bug.cgi?id=97613
1018
1019         Reviewed by Dirk Pranke.
1020
1021         Use options added in r129580.
1022
1023         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1024         (RunAndUploadPerfTests):
1025
1026 2012-09-25  Peter Rybin  <peter.rybin@gmail.com>
1027
1028         Web Inspector: Fix logging in pure protocol test harness
1029         https://bugs.webkit.org/show_bug.cgi?id=97579
1030
1031         Reviewed by Yury Semikhatsky.
1032
1033         JavaScript alert implementation in DumpRenderTree gets immediate flush to make sure that
1034         the message survives crash or kill by timeout.
1035
1036         * DumpRenderTree/blackberry/DumpRenderTree.cpp:
1037         (BlackBerry::WebKit::DumpRenderTree::runJavaScriptAlert):
1038         * DumpRenderTree/chromium/WebViewHost.cpp:
1039         (WebViewHost::runModalAlertDialog):
1040         * DumpRenderTree/efl/DumpRenderTreeView.cpp:
1041         (onJavaScriptAlert):
1042         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1043         (webViewScriptAlert):
1044         * DumpRenderTree/mac/UIDelegate.mm:
1045         (-[UIDelegate webView:runJavaScriptAlertPanelWithMessage:initiatedByFrame:]):
1046         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1047         (WebCore::WebPage::javaScriptAlert):
1048         * DumpRenderTree/win/UIDelegate.cpp:
1049         (UIDelegate::runJavaScriptAlertPanelWithMessage):
1050
1051 2012-09-25  Dan Bernstein  <mitz@apple.com>
1052
1053         Tools part of <rdar://problem/11455228> [mac] Stop using screen fonts
1054         https://bugs.webkit.org/show_bug.cgi?id=97620
1055
1056         Reviewed by John Sullivan.
1057
1058         * DumpRenderTree/mac/DumpRenderTree.mm:
1059         (resetDefaultsToConsistentValues): Enable screen font substitution when
1060         running the tests.
1061         * WebKitTestRunner/TestController.cpp:
1062         (WTR::TestController::resetStateToConsistentValues): Ditto.
1063
1064 2012-09-25  Ryosuke Niwa  <rniwa@webkit.org>
1065
1066         Build fix. "values" doesn't necessarily contain multiple values but it needs to be always parsed as an array.
1067
1068         * Scripts/webkitpy/performance_tests/perftest.py:
1069         (PerfTest.parse_output):
1070
1071 2012-09-25  Ryosuke Niwa  <rniwa@webkit.org>
1072
1073         run-perf-tests: cleanup options and results generation code
1074         https://bugs.webkit.org/show_bug.cgi?id=97611
1075
1076         Reviewed by Dirk Pranke.
1077
1078         Previously, --test-results-server triggered old JSOn format where the outermost structure was a dictionary
1079         instead of an array, and also implicitly triggered --no-show-results, caused the old outputs not to merge,
1080         and prevented the generation of results page. Also, it was not obvious that --source-json-path is an option
1081         used only on buildbot slaves.
1082
1083         This patch will:
1084         - Remove the old format since perf-o-matic supports new format now.
1085         - Add --reset-results option so that we can explicitly clear existing outputs.
1086         - Add --slave-config-json-path option to replace --source-json-path option.
1087
1088         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
1089         (PerfTestsRunner._parse_args): Added --reset-results and --slave-config-json-path options.
1090         (PerfTestsRunner._generate_and_show_results): Refactored. Also removed the code to strip "values" from
1091         results since perf-o-matic can parse and store these values now.
1092         (PerfTestsRunner._merge_outputs_if_needed): Renamed from _merge_outputs_if_needed.
1093         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
1094         (test_run_with_json_output): Test a harmless behavioral change to generate resuls page even when
1095         --test-results-server is present.
1096         (test_run_with_description):
1097         (test_run_generates_json_by_default):
1098         (test_run_merges_output_by_default): Added.
1099         (test_run_respects_reset_results): Added.
1100         (test_run_with_slave_config_json):
1101         (test_run_with_bad_slave_config_json): Use --slave-config-json-path instead of --source-json-path to make
1102         sure this optioon works as well.
1103         (test_run_with_multiple_repositories):
1104         (test_run_with_upload_json):
1105
1106 2012-09-25  Simon Fraser  <simon.fraser@apple.com>
1107
1108         Fix SnowLeopard build by adding #ifdefs.
1109         
1110         Fix crashes on some tests where the CFURLRef appears to be null.
1111
1112         * WebKitTestRunner/InjectedBundle/mac/InjectedBundlePageMac.mm:
1113         (WTR::testPathFromURL):
1114         (WTR::InjectedBundlePage::platformDidStartProvisionalLoadForFrame):
1115
1116 2012-09-25  Simon Fraser  <simon.fraser@apple.com>
1117
1118         Speculative build fix for Production builds.
1119
1120         * DumpRenderTree/mac/Configurations/Base.xcconfig:
1121         * WebKitTestRunner/Configurations/Base.xcconfig:
1122
1123 2012-09-25  Simon Fraser  <simon.fraser@apple.com>
1124
1125         Fix the build after r129566.
1126
1127         * DumpRenderTree/mac/FrameLoadDelegate.mm:
1128
1129 2012-09-25  Dominic Mazzoni  <dmazzoni@google.com>
1130
1131         AX: AccessibilityUIElementChromium.cpp won't build on Windows without wtf/StringExtras.h
1132         https://bugs.webkit.org/show_bug.cgi?id=97614
1133
1134         Unreviewed, fixing windows breakage.
1135
1136         Added wtf/StringExtras.h to the includes to fix the Chromium/Windows build.
1137
1138         * DumpRenderTree/chromium/TestRunner/AccessibilityUIElementChromium.cpp:
1139
1140 2012-09-24  Simon Fraser  <simon.fraser@apple.com>
1141
1142         <rdar://problem/12351906> Have DumpRenderTree and WebKitTestRunner crash logs show which test crashed
1143
1144         Reviewed by Mark Rowe.
1145
1146         Use a new WebKitSytemInterface function to add data to crash logs about which 
1147         test was running when the crash happened.
1148
1149         * DumpRenderTree/mac/Configurations/DebugRelease.xcconfig: Make sure WEBKIT_SYSTEM_INTERFACE_LIBRARY is defined.
1150         * DumpRenderTree/mac/Configurations/DumpRenderTree.xcconfig: link with the WebKitSystemInterface library.
1151         * DumpRenderTree/mac/FrameLoadDelegate.mm:
1152         (testPathFromURL): Given a (probably) file URL, return the part after "/LayoutTests/"
1153         (-[FrameLoadDelegate webView:didStartProvisionalLoadForFrame:]): Tell Crash Reporter which test was running,
1154         if we're the main frame and the test is active.
1155         * WebKitTestRunner/Configurations/DebugRelease.xcconfig: Make sure WEBKIT_SYSTEM_INTERFACE_LIBRARY is defined.
1156         * WebKitTestRunner/Configurations/InjectedBundle.xcconfig: link the injected bundle with the WebKitSystemInterface library.
1157         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1158         (WTR::InjectedBundlePage::didStartProvisionalLoadForFrame): Call platformDidStartProvisionalLoadForFrame().
1159         (WTR::InjectedBundlePage::platformDidStartProvisionalLoadForFrame): Stub for non-Mac platforms.
1160         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
1161         Add a platformDidStartProvisionalLoadForFrame() for platform-specific work.
1162         * WebKitTestRunner/InjectedBundle/mac/InjectedBundlePageMac.mm: Added.
1163         (WTR::testPathFromURL): Given a (probably) file URL, return the part after "/LayoutTests/"
1164         (WTR::InjectedBundlePage::platformDidStartProvisionalLoadForFrame): Tell Crash Reporter which test was running,
1165         if we're the main frame.
1166         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj: Add InjectedBundlePageMac.mm.
1167
1168 2012-09-25  Julie Parent  <jparent@chromium.org>
1169
1170         Make test_shell_tests show up in flakiness dashboard.
1171         https://bugs.webkit.org/show_bug.cgi?id=97585
1172
1173         Reviewed by Dirk Pranke.
1174
1175         test_shell_tests run only on chromium.org, and on the same bots as
1176         the layout tests, so do not use the generic fallbacks.
1177
1178         * TestResultServer/static-dashboards/builders.js:
1179         (loadBuildersList):
1180         * TestResultServer/static-dashboards/dashboard_base.js:
1181         (currentBuilderGroupCategory):
1182
1183 2012-09-25  Lucas Forschler  <lforschler@apple.com>
1184
1185         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.
1186         https://bugs.webkit.org/show_bug.cgi?id=97465
1187         
1188         Reviewed by Tim Horton.
1189
1190         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1191         (RunPythonTests.start):
1192
1193 2012-09-25  Mike West  <mkwst@chromium.org>
1194
1195         Unreviewed. Adding myself as a committer.
1196
1197         Yay. :)
1198
1199         * Scripts/webkitpy/common/config/committers.py:
1200
1201 2012-09-25  Beth Dakin  <bdakin@apple.com>
1202
1203         https://bugs.webkit.org/show_bug.cgi?id=95397
1204         Need to merge didFirstVisuallyNonEmptyLayout and 
1205         didNewFirstVisuallyNonEmptyLayout
1206         -and corresponding-
1207         <rdar://problem/10791680>
1208
1209         Reviewed by Sam Weinig.
1210
1211         Account for the new member of the PageLoadClient.
1212         * MiniBrowser/mac/BrowserWindowController.m:
1213         (-[BrowserWindowController awakeFromNib]):
1214         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1215         (WTR::InjectedBundlePage::InjectedBundlePage):
1216         * WebKitTestRunner/TestController.cpp:
1217         (WTR::TestController::initialize):
1218
1219         The NewFirstVisuallyNonEmptyLayoutForImages has existed for a while, 
1220         but it did not have the right target membership.
1221         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
1222
1223         Use new didLayout API instead of old 
1224         didNewFirstVisuallyNonEmptyLayout API.
1225         * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayout.cpp:
1226         (TestWebKitAPI):
1227         (TestWebKitAPI::didLayout):
1228         (TestWebKitAPI::setPageLoaderClient):
1229         (TestWebKitAPI::TEST):
1230         * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayoutFails.cpp:
1231         (TestWebKitAPI):
1232         (TestWebKitAPI::didLayout):
1233         (TestWebKitAPI::setPageLoaderClient):
1234         (TestWebKitAPI::TEST):
1235         * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayoutForImages.cpp:
1236         (TestWebKitAPI):
1237         (TestWebKitAPI::didLayout):
1238         (TestWebKitAPI::setPageLoaderClient):
1239         (TestWebKitAPI::TEST):
1240         * TestWebKitAPI/Tests/WebKit2/NewFirstVisuallyNonEmptyLayoutFrames.cpp:
1241         (TestWebKitAPI):
1242         (TestWebKitAPI::didLayout):
1243         (TestWebKitAPI::setPageLoaderClient):
1244         (TestWebKitAPI::TEST):
1245
1246 2012-09-25  Benjamin Poulain  <bpoulain@apple.com>
1247
1248         Add missing support for Geolocation tests on WebKit2
1249         https://bugs.webkit.org/show_bug.cgi?id=97518
1250
1251         Reviewed by Sam Weinig.
1252
1253         There are two main features added to the tooling to make the test run:
1254         -Add support for error reporting for position errors through setMockGeolocationPositionUnavailableError.
1255         -Support the 3 states permission: undefined-allowed-denied.
1256
1257         * WebKitTestRunner/GeolocationProviderMock.cpp:
1258         (WTR::GeolocationProviderMock::GeolocationProviderMock):
1259         (WTR::GeolocationProviderMock::setPosition):
1260         (WTR::GeolocationProviderMock::setPositionUnavailableError):
1261         (WTR::GeolocationProviderMock::startUpdating):
1262         (WTR::GeolocationProviderMock::sendPositionIfNeeded):
1263         (WTR::GeolocationProviderMock::sendErrorIfNeeded):
1264         * WebKitTestRunner/GeolocationProviderMock.h:
1265         (GeolocationProviderMock):
1266         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
1267         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
1268         (WTR::InjectedBundle::setMockGeolocationPositionUnavailableError):
1269         * WebKitTestRunner/InjectedBundle/InjectedBundle.h:
1270         (InjectedBundle):
1271         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
1272         (WTR::TestRunner::setMockGeolocationPositionUnavailableError):
1273         * WebKitTestRunner/InjectedBundle/TestRunner.h:
1274         (TestRunner):
1275         * WebKitTestRunner/TestController.cpp:
1276         (WTR::TestController::TestController):
1277         (WTR::decidePolicyForGeolocationPermissionRequest):
1278         (WTR::TestController::resetStateToConsistentValues):
1279         (WTR::TestController::setGeolocationPermission):
1280         (WTR::TestController::setMockGeolocationPosition):
1281         (WTR::TestController::setMockGeolocationPositionUnavailableError):
1282         (WTR::TestController::handleGeolocationPermissionRequest):
1283         (WTR::TestController::decidePolicyForGeolocationPermissionRequestIfPossible):
1284         * WebKitTestRunner/TestController.h:
1285         (TestController):
1286         * WebKitTestRunner/TestInvocation.cpp:
1287         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
1288
1289 2012-09-25  Sheriff Bot  <webkit.review.bot@gmail.com>
1290
1291         Unreviewed, rolling out r129517.
1292         http://trac.webkit.org/changeset/129517
1293         https://bugs.webkit.org/show_bug.cgi?id=97582
1294
1295         Link errors in chromium (Requested by alecf on #webkit).
1296
1297         * DumpRenderTree/DumpRenderTree.gypi:
1298         * DumpRenderTree/chromium/MockConstraints.cpp: Removed.
1299         * DumpRenderTree/chromium/MockConstraints.h: Removed.
1300         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
1301         (RTCSessionDescriptionRequestSuccededTask::RTCSessionDescriptionRequestSuccededTask):
1302         (RTCSessionDescriptionRequestFailedTask::RTCSessionDescriptionRequestFailedTask):
1303         (RTCVoidRequestTask::RTCVoidRequestTask):
1304         (isSupportedConstraint):
1305         (isValidConstraint):
1306         (MockWebRTCPeerConnectionHandler::initialize):
1307         * DumpRenderTree/chromium/WebUserMediaClientMock.cpp:
1308         (WebKit::WebUserMediaClientMock::requestUserMedia):
1309         * DumpRenderTree/chromium/WebUserMediaClientMock.h:
1310         (WebKit):
1311         (WebUserMediaClientMock):
1312         * DumpRenderTree/chromium/WebViewHost.h:
1313         (WebKit):
1314         (WebViewHost):
1315
1316 2012-09-25  Dominic Mazzoni  <dmazzoni@google.com>
1317
1318         AX: Chromium doesn't pass accessibility text range and line number tests
1319         https://bugs.webkit.org/show_bug.cgi?id=97246
1320
1321         Reviewed by Chris Fleizach.
1322
1323         Provide a chromium implementation of some AccessibilityUIElement
1324         methods needed by accessibility tests.
1325
1326         Enables four existing tests to pass on Chromium:
1327         Tests: accessibility/textarea-insertion-point-line-number.html
1328                accessibility/textarea-line-for-index.html
1329                accessibility/textarea-selected-text-range.html
1330                accessibility/textbox-role-reports-selection.html
1331
1332         * DumpRenderTree/chromium/TestRunner/AccessibilityUIElementChromium.cpp:
1333         (AccessibilityUIElement::insertionPointLineNumberGetterCallback):
1334         (AccessibilityUIElement::selectedTextRangeGetterCallback):
1335         (AccessibilityUIElement::lineForIndexCallback):
1336         (AccessibilityUIElement::setSelectedTextRangeCallback):
1337
1338 2012-09-25  Peter Beverloo  <peter@chromium.org>
1339
1340         [Chromium] Android's build archives have grown from 38M to 102M in 11 days
1341         https://bugs.webkit.org/show_bug.cgi?id=97336
1342
1343         Reviewed by Adam Barth.
1344
1345         Each WebKit target results in four APKs in the out/{Debug,Release}/ directory,
1346         while we really only need one. Ignoring the unsigned APKs removes six from
1347         the archive will make sure that we only archive a single APK per target,
1348         bringing its size back to rougly 40 megabytes.
1349
1350         * BuildSlaveSupport/built-product-archive:
1351         (archiveBuiltProduct):
1352
1353 2012-09-25  Tommy Widenflycht  <tommyw@google.com>
1354
1355         MediaStream API: Update getUserMedia to match the latest specification
1356         https://bugs.webkit.org/show_bug.cgi?id=97540
1357
1358         Reviewed by Adam Barth.
1359
1360         Adding mock support for the enhanced getUserMedia call.
1361
1362         * DumpRenderTree/DumpRenderTree.gypi:
1363         * DumpRenderTree/chromium/MockConstraints.cpp: Copied from Source/WebCore/platform/chromium/support/WebMediaConstraints.cpp.
1364         (MockConstraints):
1365         (MockConstraints::isSupported):
1366         (MockConstraints::isValid):
1367         (MockConstraints::verify):
1368         * DumpRenderTree/chromium/MockConstraints.h: Copied from Tools/DumpRenderTree/chromium/WebUserMediaClientMock.h.
1369         (WebKit):
1370         (MockConstraints):
1371         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
1372         (RTCSessionDescriptionRequestSuccededTask::RTCSessionDescriptionRequestSuccededTask):
1373         (RTCSessionDescriptionRequestFailedTask::RTCSessionDescriptionRequestFailedTask):
1374         (RTCVoidRequestTask::RTCVoidRequestTask):
1375         (MockWebRTCPeerConnectionHandler::initialize):
1376         * DumpRenderTree/chromium/WebUserMediaClientMock.cpp:
1377         (UserMediaRequestTask):
1378         (UserMediaRequestTask::UserMediaRequestTask):
1379         (WebUserMediaClientMock::requestUserMedia):
1380         * DumpRenderTree/chromium/WebUserMediaClientMock.h:
1381         (WebUserMediaClientMock):
1382         (WebUserMediaClientMock::taskList):
1383         * DumpRenderTree/chromium/WebViewHost.h:
1384         (WebKit):
1385         (WebViewHost):
1386
1387 2012-09-25  Raphael Kubo da Costa  <raphael.kubo.da.costa@intel.com>
1388
1389         [EFL] Do not dump inspector output in DRT
1390         https://bugs.webkit.org/show_bug.cgi?id=97018
1391
1392         Reviewed by Kenneth Rohde Christiansen.
1393
1394         Some inspector-related layout tests expect console output from the
1395         inspector view to be suppressed. Do that by setting a property in
1396         the inspector view that makes the onConsoleMessage callback return
1397         early and not output anything.
1398
1399         Using evas_object_data_{get,set} is not very beautiful, but it is
1400         still better than properly subclassing Ewk_View and having to
1401         implement something similar to drt_view_console_message_enable_set().
1402
1403         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
1404         (DumpRenderTreeChrome::createWebInspectorView):
1405         * DumpRenderTree/efl/DumpRenderTreeView.cpp:
1406         (onConsoleMessage):
1407
1408 2012-09-25  Christophe Dumez  <christophe.dumez@intel.com>
1409
1410         [WK2][WKTR] TestRunner needs support for setWillSendRequestReturnsNullOnRedirect
1411         https://bugs.webkit.org/show_bug.cgi?id=97542
1412
1413         Reviewed by Kenneth Rohde Christiansen.
1414
1415         Implement support for setWillSendRequestReturnsNullOnRedirect
1416         in WebKitTestRunner.
1417
1418         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
1419         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1420         (WTR::InjectedBundlePage::willSendRequestForFrame):
1421         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
1422         (WTR::TestRunner::TestRunner):
1423         * WebKitTestRunner/InjectedBundle/TestRunner.h:
1424         (WTR::TestRunner::willSendRequestReturnsNullOnRedirect):
1425         (WTR::TestRunner::setWillSendRequestReturnsNullOnRedirect):
1426         (TestRunner):
1427
1428 2012-09-25  Xiaobo Wang  <xbwang@torchmobile.com.cn>
1429
1430         [BlackBerry] [DRT] fast/forms/textarea-live-pseudo-selectors.html failed
1431         https://bugs.webkit.org/show_bug.cgi?id=97189
1432
1433         Reviewed by Rob Buis.
1434
1435         Translate ASCII key code to BlackBerry key code for BACKSPACE and DELETE key.
1436
1437         * DumpRenderTree/blackberry/EventSender.cpp:
1438         (keyDownCallback):
1439
1440 2012-09-25  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
1441
1442         [WK2][WTR] WTR bundle client loads binary data as text
1443         https://bugs.webkit.org/show_bug.cgi?id=97532
1444
1445         Reviewed by Kenneth Rohde Christiansen.
1446
1447         WTR::InjectedBundlePage::decidePolicyForResponse now checks response MIME type
1448         and returns WKBundlePagePolicyActionPassThrough if it cannot be shown.
1449
1450         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1451         (WTR::InjectedBundlePage::decidePolicyForResponse):
1452
1453 2012-09-25  Zan Dobersek  <zandobersek@gmail.com>
1454
1455         Unreviewed, another test case adjustment after r129459. Python 2.7 doesn't
1456         seem to be a requirement yet so the assertDictEqual method is not yet available
1457         on all builders. Assert that the empty dictionary length equals 0 instead.
1458
1459         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
1460         (ResultSummaryTest.test_interpret_test_failures):
1461
1462 2012-09-24  Simon Hausmann  <simon.hausmann@digia.com>
1463
1464         [Qt] Fix build with latest Qt 5
1465         https://bugs.webkit.org/show_bug.cgi?id=97479
1466
1467         Reviewed by Tor Arne Vestbø.
1468
1469         The variables for syncqt handling are now set in api.pri, but let's add a comment here that we can
1470         remove them from default_pre.prf once we're past Qt 5 beta2.
1471
1472         * qmake/mkspecs/features/default_pre.prf:
1473
1474 2012-09-25  Zan Dobersek  <zandobersek@gmail.com>
1475
1476         Unreviewed, adjusting the test cases after r129459. The JSON data
1477         for reftests now contains the 'reftest_type' key and no more 'is_reftest'
1478         or 'is_mismatch_reftest' keys.
1479
1480         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1481         (EndToEndTest.test_reftest_with_two_notrefs):
1482
1483 2012-09-24  Zan Dobersek  <zandobersek@gmail.com>
1484
1485         Unexpected reftest passes are only reported when pixel testing is enabled in results.html as well
1486         https://bugs.webkit.org/show_bug.cgi?id=97426
1487
1488         Reviewed by Ojan Vafai.
1489
1490         The 'is_reftest' member of the TestResult class has been replaced by the reftest_type
1491         member which is a set containing either '!=', '==', both or none if the test represented
1492         by the object is not a reftest.
1493
1494         When summarizing results, the test dictionary is updated to contain the 'reftest_type' key
1495         if the test is a reftest. The value of this key is a list of all the reftest types of this test.
1496
1497         The test failure interpretation method has been refactored, it now reports only
1498         missing results of various formats and image diff percent in case of image or
1499         reftest mismatch.
1500
1501         Unit tests have been modified accordingly.
1502
1503         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner.py:
1504         (LayoutTestRunner._update_summary_with_result):
1505         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner_unittest.py:
1506         (LayoutTestRunnerTests.test_update_summary_with_result):
1507         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1508         (interpret_test_failures):
1509         (summarize_results):
1510         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
1511         (ResultSummaryTest.test_interpret_test_failures):
1512         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py:
1513         (SingleTestRunner.run):
1514         (SingleTestRunner._run_reftest):
1515         * Scripts/webkitpy/layout_tests/models/test_results.py:
1516         (TestResult.__init__):
1517         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1518         (EndToEndTest.test_reftest_with_two_notrefs):
1519
1520 2012-09-24  Sam Weinig  <sam@webkit.org>
1521
1522         Use NSUserDefaults rather than an environment variable to control whether to use an XPC Service for the WebProcess
1523         https://bugs.webkit.org/show_bug.cgi?id=97514
1524
1525         Reviewed by Anders Carlsson.
1526
1527         * Scripts/webkitdirs.pm:
1528         (argumentsForRunAndDebugMacWebKitApp):
1529         Push "-WebKit2UseXPCServiceForWebProcess YES" as arguments when --use-web-process-xpc-service
1530         is provided to set the user default. Move adding of @ARGV to after custom additions as calling
1531         shouldUseXPCServiceForWebProcess() can change @ARGV.
1532
1533         (runMacWebKitApp):
1534         (execMacWebKitAppForDebugging):
1535         Stop setting the WEBKIT_USE_XPC_SERVICE_FOR_WEB_PROCESS environment variable.
1536
1537 2012-09-24  Laszlo Gombos  <l.gombos@samsung.com>
1538
1539         [GTK][EFL] Make sure cairo is built with fontconfig support
1540         https://bugs.webkit.org/show_bug.cgi?id=97513
1541
1542         Reviewed by Martin Robinson.
1543
1544         The GTK and EFL port of WebKit requires a version of cairo that is built with fontconfig support. 
1545
1546         * efl/jhbuild.modules:
1547         * gtk/jhbuild.modules:
1548
1549 2012-09-24  Benjamin Poulain  <bpoulain@apple.com>
1550
1551         Fix Geolocation error reporting in the test support
1552         https://bugs.webkit.org/show_bug.cgi?id=97386
1553
1554         Reviewed by Sam Weinig.
1555         The TestRunner was exposing internal implementation details of
1556         the Geolocation object through setMockGeolocationError().
1557
1558         In practice, only the error code PositionUnavailable is supported
1559         or even implemented.
1560
1561         Exposing setMockGeolocationError() does not make sense for WebKit2 as
1562         passing generic errors from the UIProcess is a bad idea. This is the
1563         opportunity to clean TestRunner and prepare and interface suitable for WebKit2.
1564
1565         * DumpRenderTree/TestRunner.cpp:
1566         (setMockGeolocationPositionUnavailableErrorCallback):
1567         (TestRunner::staticFunctions):
1568         * DumpRenderTree/TestRunner.h:
1569         (TestRunner):
1570         * DumpRenderTree/blackberry/TestRunnerBlackBerry.cpp:
1571         (TestRunner::setMockGeolocationPositionUnavailableError):
1572         * DumpRenderTree/chromium/DRTTestRunner.cpp:
1573         (DRTTestRunner::DRTTestRunner):
1574         (DRTTestRunner::setMockGeolocationPositionUnavailableError):
1575         * DumpRenderTree/chromium/DRTTestRunner.h:
1576         (DRTTestRunner):
1577         * DumpRenderTree/efl/TestRunnerEfl.cpp:
1578         (TestRunner::setMockGeolocationPositionUnavailableError):
1579         * DumpRenderTree/gtk/TestRunnerGtk.cpp:
1580         (TestRunner::setMockGeolocationPositionUnavailableError):
1581         * DumpRenderTree/mac/MockGeolocationProvider.h:
1582         * DumpRenderTree/mac/MockGeolocationProvider.mm:
1583         (-[MockGeolocationProvider dealloc]):
1584         (-[MockGeolocationProvider resetError]):
1585         (-[MockGeolocationProvider setPosition:]):
1586         (-[MockGeolocationProvider setPositionUnavailableErrorWithMessage:]):
1587         (-[MockGeolocationProvider lastPosition]):
1588         (-[MockGeolocationProvider timerFired]):
1589         * DumpRenderTree/mac/TestRunnerMac.mm:
1590         (TestRunner::setMockGeolocationPositionUnavailableError):
1591         * DumpRenderTree/qt/TestRunnerQt.cpp:
1592         (TestRunner::setMockGeolocationPositionUnavailableError):
1593         * DumpRenderTree/qt/TestRunnerQt.h:
1594         (TestRunner):
1595         * DumpRenderTree/win/TestRunnerWin.cpp:
1596         (TestRunner::setMockGeolocationPositionUnavailableError):
1597         * DumpRenderTree/wx/TestRunnerWx.cpp:
1598         (TestRunner::setMockGeolocationPositionUnavailableError):
1599
1600 2012-09-24  Nikhil Bansal  <n.bansal@samsung.com>
1601
1602         [EFL][WK2] TestRunner needs touch events support.
1603         https://bugs.webkit.org/show_bug.cgi?id=96465
1604
1605         Reviewed by Kenneth Rohde Christiansen.
1606
1607         Simulate touch events in WebkitTestRunner for touch related
1608         layout tests. This patch uses ewk_view_feed_touch_event() to
1609         generate touch events.
1610
1611         * WebKitTestRunner/EventSenderProxy.h:
1612         (EventSenderProxy):
1613         * WebKitTestRunner/efl/EventSenderProxyEfl.cpp:
1614         (WTR):
1615         (WTR::setEvasModifiers):
1616         (WTR::EventSenderProxy::EventSenderProxy):
1617         (WTR::EventSenderProxy::~EventSenderProxy):
1618         (WTR::EventSenderProxy::sendTouchEvent):
1619         (WTR::EventSenderProxy::addTouchPoint):
1620         (WTR::EventSenderProxy::updateTouchPoint):
1621         (WTR::EventSenderProxy::setTouchModifier):
1622         (WTR::EventSenderProxy::touchStart):
1623         (WTR::EventSenderProxy::touchMove):
1624         (WTR::EventSenderProxy::touchEnd):
1625         (WTR::EventSenderProxy::touchCancel):
1626         (WTR::EventSenderProxy::clearTouchPoints):
1627         (WTR::EventSenderProxy::releaseTouchPoint):
1628         (WTR::EventSenderProxy::cancelTouchPoint):
1629
1630 2012-09-24  Christophe Dumez  <christophe.dumez@intel.com>
1631
1632         [WK2] TestRunner does not support overridePreference with value "0"
1633         https://bugs.webkit.org/show_bug.cgi?id=97457
1634
1635         Reviewed by Alexey Proskuryakov.
1636
1637         testRunner.overridePreference() was marked as taking a boolean
1638         in argument in testRunner.idl even though the test cases are
1639         passing strings to it. The current code works if the test
1640         case wants to enable a setting and therefore passes "1" string.
1641         However, if the test wants to disable a setting and passes "0"
1642         string, the current code also evaluates this to true causing
1643         certain test cases to fail.
1644
1645         The argument to testRunner.overridePreference() is now of string
1646         type and a function has been added to convert the string passed
1647         by the test to a boolean value.
1648
1649         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
1650         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
1651         (WTR::toBool):
1652         (WTR):
1653         (WTR::TestRunner::overridePreference):
1654         * WebKitTestRunner/InjectedBundle/TestRunner.h:
1655         (TestRunner):
1656
1657 2012-09-24  Sheriff Bot  <webkit.review.bot@gmail.com>
1658
1659         Unreviewed, rolling out r129343.
1660         http://trac.webkit.org/changeset/129343
1661         https://bugs.webkit.org/show_bug.cgi?id=97479
1662
1663         It broke the build (Requested by Ossy_NIGHT on #webkit).
1664
1665         * qmake/mkspecs/features/default_pre.prf:
1666
1667 2012-09-24  Zan Dobersek  <zandobersek@gmail.com>
1668
1669         [Gtk] Unskip the Shadow DOM layout tests
1670         https://bugs.webkit.org/show_bug.cgi?id=90776
1671
1672         Reviewed by Ryosuke Niwa.
1673
1674         Enable Shadow DOM through DumpRenderTreeSupport when resetting defaults in between tests.
1675
1676         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1677         (resetDefaultsToConsistentValues):
1678
1679 2012-09-24  Zan Dobersek  <zandobersek@gmail.com>
1680
1681         REGRESSION (r129211-r129218): http/tests/loading/redirect-with-no-location-crash.html failing on GTK Linux 64-bit Release WK2 (Tests)
1682         https://bugs.webkit.org/show_bug.cgi?id=97404
1683
1684         Reviewed by Philippe Normand.
1685
1686         Print out the 'didReceiveTitle' information when dumping frame load callbacks.
1687
1688         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1689         (webViewTitleChanged):
1690
1691 2012-09-24  Simon Hausmann  <simon.hausmann@digia.com>
1692
1693         [Qt] Fix build with latest Qt 5
1694
1695         Reviewed by Tor Arne Vestbø.
1696
1697         The variables for syncqt handling are now set in api.pri, but let's add a comment here that we can
1698         remove them from default_pre.prf once we're past Qt 5 beta2.
1699
1700         * qmake/mkspecs/features/default_pre.prf:
1701
1702 2012-09-22  Dominic Mazzoni  <dmazzoni@google.com>
1703
1704         AX: Layout tests would be easier to write if AccessibilityController could find an element by id
1705         https://bugs.webkit.org/show_bug.cgi?id=97301
1706
1707         Reviewed by Chris Fleizach.
1708
1709         Adds a new accessibleElementById method to AccessibilityController, to make it
1710         easier to write cross-platform layout tests.
1711
1712         Implemented for Chromium and for Mac (DRT and WKTR).
1713
1714         * DumpRenderTree/AccessibilityController.cpp:
1715         (getAccessibleElementByIdCallback):
1716         (AccessibilityController::getJSClass):
1717         * DumpRenderTree/AccessibilityController.h:
1718         (AccessibilityController):
1719         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
1720         * DumpRenderTree/chromium/TestRunner/AccessibilityControllerChromium.cpp:
1721         (AccessibilityController::AccessibilityController):
1722         (AccessibilityController::findAccessibleElementByIdRecursive):
1723         (AccessibilityController::getAccessibleElementById):
1724         (AccessibilityController::accessibleElementByIdGetterCallback):
1725         * DumpRenderTree/chromium/TestRunner/AccessibilityControllerChromium.h:
1726         (AccessibilityController):
1727         * DumpRenderTree/gtk/AccessibilityControllerGtk.cpp:
1728         (AccessibilityController::accessibleElementById):
1729         * DumpRenderTree/mac/AccessibilityCommonMac.h: added
1730         * DumpRenderTree/mac/AccessibilityCommonMac.mm: added
1731         (+[NSString stringWithJSStringRef:]):
1732         (-[NSString createJSStringRef]):
1733         * DumpRenderTree/mac/AccessibilityControllerMac.mm:
1734         (findAccessibleObjectById):
1735         (AccessibilityController::accessibleElementById):
1736         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
1737         * DumpRenderTree/win/AccessibilityControllerWin.cpp:
1738         (AccessibilityController::accessibleElementById):
1739         * WebKitTestRunner/InjectedBundle/AccessibilityController.cpp:
1740         (WTR::AccessibilityController::accessibleElementById):
1741         * WebKitTestRunner/InjectedBundle/AccessibilityController.h:
1742         (AccessibilityController):
1743         * WebKitTestRunner/InjectedBundle/Bindings/AccessibilityController.idl:
1744         * WebKitTestRunner/InjectedBundle/mac/AccessibilityCommonMac.h: added
1745         * WebKitTestRunner/InjectedBundle/mac/AccessibilityCommonMac.mm: added
1746         (+[NSString stringWithJSStringRef:]):
1747         (-[NSString createJSStringRef]):
1748         * WebKitTestRunner/InjectedBundle/mac/AccessibilityControllerMac.mm:
1749         (WTR::findAccessibleObjectById):
1750         (WTR):
1751         (WTR::AccessibilityController::accessibleElementById):
1752         * WebKitTestRunner/InjectedBundle/mac/AccessibilityUIElementMac.mm:
1753         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
1754
1755 2012-09-22  Zan Dobersek  <zandobersek@gmail.com>
1756
1757         Unexpected reftest passes are only reported when pixel testing is enabled
1758         https://bugs.webkit.org/show_bug.cgi?id=97242
1759
1760         Reviewed by Dirk Pranke.
1761
1762         LayoutTestRunner should know whether the finished test it's handling is
1763         a reftest. This is necessary when updating result summary for reftests
1764         when pixel testing is not enabled - in these circumstances an unexpectedly
1765         passing reftest was not reported as such due to pixel testing then being
1766         determined by searching for reftest failures in failures list, which were
1767         not found in an empty failure list of a passing test. Now the TestResult
1768         objects have a member variable indicating the test is a reftest whenever
1769         that's the case.
1770
1771         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner.py:
1772         (LayoutTestRunner._update_summary_with_result):
1773         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner_unittest.py:
1774         (LayoutTestRunnerTests.test_update_summary_with_result):
1775         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py:
1776         (SingleTestRunner.run):
1777         (SingleTestRunner._run_reftest):
1778         * Scripts/webkitpy/layout_tests/models/test_results.py:
1779         (TestResult.__init__):
1780
1781 2012-09-21  Sam Weinig  <sam@webkit.org>
1782
1783         WebProcess XPC services need have their environment set without disrupting all other XPC services
1784         https://bugs.webkit.org/show_bug.cgi?id=95161
1785
1786         Reviewed by Anders Carlsson.
1787
1788         Remove setting the __XPC_* environment variables.
1789
1790         * Scripts/webkitdirs.pm:
1791         (setUpGuardMallocIfNeeded):
1792         (runMacWebKitApp):
1793         (execMacWebKitAppForDebugging):
1794
1795 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
1796
1797         Fix typo in additional-platform-directory patch just landed
1798         https://bugs.webkit.org/show_bug.cgi?id=97380
1799         
1800         Unreviewed, build fix.
1801
1802         * Scripts/webkitpy/layout_tests/port/base.py:
1803         (Port.relative_test_filename):
1804         (Port.relative_perf_test_filename):
1805
1806 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
1807
1808         nrwt: don't require additional-platform-directory to be an abspath or live under LayoutTests
1809         https://bugs.webkit.org/show_bug.cgi?id=97380
1810
1811         Reviewed by Ojan Vafai.
1812
1813         There doesn't seem to be a good reason for this restriction and
1814         it's useful to be able to point to directories outside the
1815         checkout for results (e.g., for local failures due to a 10.7.4
1816         install ;).
1817
1818         * Scripts/webkitpy/layout_tests/port/base.py:
1819         (Port.relative_test_filename):
1820         (Port.relative_perf_test_filename):
1821         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
1822         (ChromiumAndroidDriver._command_from_driver_input):
1823         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1824         (_set_up_derived_options):
1825         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1826         (MainTest.test_additional_platform_directory):
1827
1828 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
1829
1830         webkitpy: drop support for old TestExpectations syntax
1831         https://bugs.webkit.org/show_bug.cgi?id=97364
1832
1833         Reviewed by Ryosuke Niwa.
1834
1835         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
1836         (TestExpectationParser._collect_matching_tests):
1837         (TestExpectationParser):
1838         (TestExpectationParser._tokenize_line):
1839         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
1840         (SkippedTests.test_skipped_entry_dont_exist):
1841         (ExpectationSyntaxTests.assert_tokenize_exp):
1842         (ExpectationSyntaxTests.test_bare_name):
1843         (ExpectationSyntaxTests.test_bare_name_and_bugs):
1844         (ExpectationSyntaxTests.test_comments):
1845         (ExpectationSyntaxTests.test_config_modifiers):
1846         (ExpectationSyntaxTests.test_unknown_config):
1847         (ExpectationSyntaxTests.test_unknown_expectation):
1848         (ExpectationSyntaxTests.test_skip):
1849         (ExpectationSyntaxTests.test_slow):
1850         (ExpectationSyntaxTests.test_wontfix):
1851         (ExpectationSyntaxTests.test_blank_line):
1852         (ExpectationSyntaxTests.test_warnings):
1853         (RebaseliningTest.test_no_get_rebaselining_failures):
1854
1855 2012-09-21  Simon Fraser  <simon.fraser@apple.com>
1856
1857         Improve WTR unresponsiveness output a little
1858         https://bugs.webkit.org/show_bug.cgi?id=97370
1859
1860         Reviewed by Timothy Horton.
1861
1862         Distinguish between conditions that already set the errorMessage,
1863         and unresponsiveness due to slow about:blank loads when WTR
1864         reports unresponsiveness.
1865
1866         * WebKitTestRunner/TestInvocation.cpp:
1867         (WTR::TestInvocation::invoke):
1868
1869 2012-09-21  Chris Rogers  <crogers@google.com>
1870
1871         Add Web Audio support for deprecated/legacy APIs
1872         https://bugs.webkit.org/show_bug.cgi?id=97050
1873
1874         Reviewed by Eric Carlson.
1875
1876         * Scripts/webkitperl/FeatureList.pm:
1877
1878 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
1879
1880         webkitpy: update remaining tests to use the new expectation syntax
1881         https://bugs.webkit.org/show_bug.cgi?id=97362
1882
1883         Reviewed by Ojan Vafai.
1884
1885         This patch updates all the unit tests that were still using the
1886         old TestExpectations syntax to use the new syntax *except* for
1887         the tests that were specifically testing that we parsed the old
1888         syntax correctly.
1889
1890         Also, a block of tests for the new syntax were duplicated, so
1891         I've deleted the duplicate.
1892
1893         Note that the old syntax is still supported so this change should
1894         produce no visible changes.
1895
1896         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
1897         (MiscTests.test_multiple_results):
1898         (MiscTests.test_category_expectations):
1899         (MiscTests.test_error_on_different_platform):
1900         (MiscTests.test_error_on_different_build_type):
1901         (MiscTests.test_overrides):
1902         (MiscTests.test_overrides__directory):
1903         (MiscTests.test_overrides__duplicate):
1904         (MiscTests.test_more_specific_override_resets_skip):
1905         (SkippedTests.check):
1906         (SkippedTests.test_duplicate_skipped_test_fails_lint):
1907         (SkippedTests.test_skipped_file_overrides_expectations):
1908         (SkippedTests.test_skipped_dir_overrides_expectations):
1909         (SkippedTests.test_skipped_file_overrides_overrides):
1910         (SkippedTests.test_skipped_dir_overrides_overrides):
1911         (ExpectationSyntaxTests.disabled_test_missing_expectation):
1912         (ExpectationSyntaxTests.disabled_test_missing_colon):
1913         (ExpectationSyntaxTests.disabled_test_too_many_colons):
1914         (ExpectationSyntaxTests.disabled_test_too_many_equals_signs):
1915         (ExpectationSyntaxTests):
1916         (ExpectationSyntaxTests.test_unrecognized_expectation):
1917         (ExpectationSyntaxTests.test_macro):
1918         (SemanticTests.test_bug_format):
1919         (SemanticTests.test_bad_bugid):
1920         (SemanticTests.test_missing_bugid):
1921         (SemanticTests.test_slow_and_timeout):
1922         (SemanticTests.test_rebaseline):
1923         (test_missing_file):
1924         (test_ambiguous):
1925         (test_more_modifiers):
1926         (test_order_in_file):
1927         (test_macro_overrides):
1928         (OldExpectationParserTests):
1929         (OldExpectationParserTests._tokenize):
1930         (OldExpectationParserTests.test_tokenize_extra_colon):
1931         (OldExpectationParserTests.test_tokenize_missing_equal):
1932         (OldExpectationParserTests.test_tokenize_extra_equal):
1933         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1934         (MainTest.test_skip_failing_tests):
1935         (MainTest.test_additional_expectations):
1936         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
1937         (TestExpectationsTestCase.test_valid_expectations):
1938         (TestExpectationsTestCase.test_invalid_expectations):
1939         (TestExpectationsTestCase.test_tab):
1940
1941 2012-09-21  Benjamin Poulain  <bpoulain@apple.com>
1942
1943         [WK2] Add basic testing support for Geolocation
1944         https://bugs.webkit.org/show_bug.cgi?id=97278
1945
1946         Reviewed by Kenneth Rohde Christiansen.
1947
1948         Update the Geolocation testing to use the proper API in the UIProcess.
1949
1950         * WebKitTestRunner/CMakeLists.txt:
1951         * WebKitTestRunner/GNUmakefile.am:
1952
1953         * WebKitTestRunner/GeolocationProviderMock.cpp: Added.
1954         (WTR::startUpdatingCallback):
1955         (WTR::stopUpdatingCallback):
1956         (WTR::GeolocationProviderMock::GeolocationProvierMock):
1957         (WTR::GeolocationProviderMock::setMockGeolocationPosition):
1958         (WTR::GeolocationProviderMock::startUpdating):
1959         (WTR::GeolocationProviderMock::stopUpdating):
1960         (GeolocationProviderMock):
1961         The GeolocationProvider store the location update and deliver them as needed.
1962
1963         WebCore GeolocationController do not support asynchronous update on start/stop. This is not
1964         a problem in this case because all the messages between the WebProcess and the UIProcess are
1965         asynchronous. Because of this, unlike GeolocationClientMock, we do not use a timer for event
1966         delivery.
1967
1968         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
1969         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
1970         (WTR::InjectedBundle::setGeolocationPermission):
1971         (WTR::InjectedBundle::setMockGeolocationPosition):
1972         * WebKitTestRunner/InjectedBundle/InjectedBundle.h:
1973         (InjectedBundle):
1974         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
1975         (WTR::TestRunner::setGeolocationPermission):
1976         (WTR::TestRunner::setMockGeolocationPosition):
1977         From the InjectedBundle, we now pass the information to the UIProcess so that
1978         GeolocationProvider and the TestController can respond appropriately.
1979
1980         * WebKitTestRunner/InjectedBundle/TestRunner.h:
1981         (TestRunner):
1982         * WebKitTestRunner/Target.pri:
1983         * WebKitTestRunner/TestController.cpp:
1984         (WTR::TestController::TestController):
1985         (WTR::decidePolicyForGeolocationPermissionRequest):
1986         (WTR::TestController::createOtherPage):
1987         (WTR::TestController::initialize):
1988         (WTR::TestController::setMockGeolocationPosition):
1989         * WebKitTestRunner/TestController.h:
1990         (TestController):
1991         (WTR::TestController::setGeolocationPermission):
1992         (WTR::TestController::isGeolocationPermissionAllowed):
1993         * WebKitTestRunner/TestInvocation.cpp:
1994         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
1995         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
1996         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
1997
1998 2012-09-21  Joanmarie Diggs  <jdiggs@igalia.com>
1999
2000         REGRESSION (r127882): accessibility/spinbutton-value.html failing on GTK
2001         https://bugs.webkit.org/show_bug.cgi?id=96196
2002
2003         Reviewed by Martin Robinson.
2004
2005         The "regression" is that a new test was added but the support was missing
2006         in the Gtk port for spin buttons.
2007
2008         * DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
2009         (AccessibilityUIElement::valueDescription): Updated the FIXME comment to
2010         indicate that this cannot be implemented until it is implemented in ATK.
2011         URL of the newly-filed ATK bug included for reference.
2012
2013 2012-09-21  Dirk Pranke  <dpranke@chromium.org>
2014
2015         Fix test_skip_and_wontfix failure
2016         https://bugs.webkit.org/show_bug.cgi?id=97225
2017
2018         Unreviewed, build fix.
2019
2020         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2021         (SemanticTests.test_skip_and_wontfix):
2022
2023 2012-09-21  Carlos Garcia Campos  <cgarcia@igalia.com>
2024
2025         [GTK] Implement ViewState methods in PageClientImpl in WebKit2
2026         https://bugs.webkit.org/show_bug.cgi?id=97202
2027
2028         Reviewed by Martin Robinson.
2029
2030         * WebKitTestRunner/gtk/PlatformWebViewGtk.cpp:
2031         (WTR::PlatformWebView::focus): Focus the view.
2032
2033 2012-09-21  Sudarsana Nagineni  <sudarsana.nagineni@intel.com>
2034
2035         [WTR] Memory leaks in InjectedBundleController::initialize()
2036         https://bugs.webkit.org/show_bug.cgi?id=97329
2037
2038         Reviewed by Alexey Proskuryakov.
2039
2040         Fix leaks in InjectedBundleController::initialize() by adopting
2041         strings created with WKStringCreateWithUTF8CString().
2042
2043         * TestWebKitAPI/InjectedBundleController.cpp:
2044         (TestWebKitAPI::InjectedBundleController::initialize):
2045
2046 2012-09-21  Marcelo Lira  <marcelo.lira@openbossa.org>
2047
2048         [Qt][WK2] Removed duplicated code from EventSenderProxy::keyDown
2049         https://bugs.webkit.org/show_bug.cgi?id=97235
2050
2051         Reviewed by Luiz Agostini.
2052
2053         * WebKitTestRunner/qt/EventSenderProxyQt.cpp:
2054         (WTR::EventSenderProxy::keyDown):
2055
2056 2012-09-21  Gergely Kis  <gergely@homejinni.com>
2057
2058         Add MIPS build slave to build.webkit.org
2059         https://bugs.webkit.org/show_bug.cgi?id=96713
2060
2061         Reviewed by Csaba Osztrogonác.
2062
2063         Added a build slave for MIPS, and enabled a builder for 
2064         Qt Linux MIPS32R2 little-endian release build.
2065
2066         * BuildSlaveSupport/build.webkit.org-config/config.json:
2067
2068 2012-09-21  Peter Beverloo  <peter@chromium.org>
2069
2070         [Chromium] Switch back to a fixed fifo path for Android
2071         https://bugs.webkit.org/show_bug.cgi?id=97230
2072
2073         Reviewed by Tony Chang.
2074
2075         Because not all external storage cards will be formated using a file
2076         system that supports named pipes, Chromium has been changed to creating
2077         the pipes in a temporary folder on the internal storage. Adapt this in
2078         WebKit so we can continue to run layout tests.
2079
2080         This also includes an *unreviewed* fix for a breakage in the webkitpy
2081         tests I made in r129221. Two lines and related to this code, so I decided
2082         to include it in this change.
2083
2084         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
2085         (ChromiumAndroidDriver.__init__):
2086         (ChromiumAndroidDriver._setup_test):
2087         (ChromiumAndroidDriver._update_version):
2088
2089 2012-09-20  Peter Beverloo  <peter@chromium.org>
2090
2091         Leverage Chromium's code to set up FIFOs for Chromium Android layout tests
2092         https://bugs.webkit.org/show_bug.cgi?id=97227
2093
2094         Reviewed by Tony Chang.
2095
2096         We switched Chromium to using FIFOs in order to achieve better consistency,
2097         which was done by Marcus in r157541. Remove all custom WebKit code in
2098         favor of Chromium's implementation.
2099
2100         Remove more FIFO code in the test runner itself, including the code in
2101         DumpRenderTree that invoked it. We can now switch to Chromium's brand
2102         new FIFO-creating code, which is being set-up for all test targets build
2103         for Android, including DumpRenderTree, TestWebKitAPI and webkit_unit_tests.
2104
2105         This also changes the ChromiumAndroidDriver._remove_all_pipes method to
2106         delete the files individually. "rm" would fail if one of the earlier files
2107         does not exist, and the "-f" argument doesn't seem to be reliable.
2108
2109         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
2110         * DumpRenderTree/chromium/TestShellAndroid.cpp:
2111         (platformInit):
2112         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
2113         (ChromiumAndroidDriver.__init__):
2114         (ChromiumAndroidDriver._setup_test):
2115         (ChromiumAndroidDriver._get_external_storage):
2116         (ChromiumAndroidDriver._drt_cmd_line):
2117         (ChromiumAndroidDriver._remove_all_pipes):
2118         (ChromiumAndroidDriver.stop):
2119         * TestWebKitAPI/TestWebKitAPI.gyp/TestWebKitAPI.gyp:
2120
2121 2012-09-21  Christophe Dumez  <christophe.dumez@intel.com>
2122
2123         [WK2][WKTR] EventSender needs to implement scheduleAsynchronousClick
2124         https://bugs.webkit.org/show_bug.cgi?id=97326
2125
2126         Reviewed by Kenneth Rohde Christiansen.
2127
2128         Implement scheduleAsynchronousClick() in WebKitTestRunner's
2129         EventSender by sending a "MouseDown" and a "MouseUp" message
2130         asynchronously to the WebProcess.
2131
2132         * WebKitTestRunner/InjectedBundle/Bindings/EventSendingController.idl:
2133         * WebKitTestRunner/InjectedBundle/EventSendingController.cpp:
2134         (WTR::createMouseMessageBody):
2135         (WTR):
2136         (WTR::EventSendingController::mouseDown):
2137         (WTR::EventSendingController::mouseUp):
2138         (WTR::EventSendingController::scheduleAsynchronousClick):
2139         * WebKitTestRunner/InjectedBundle/EventSendingController.h:
2140         (EventSendingController):
2141         * WebKitTestRunner/TestController.cpp:
2142         (WTR::TestController::didReceiveMessageFromInjectedBundle):
2143
2144 2012-09-21  Simon Hausmann  <simon.hausmann@digia.com>
2145
2146         [Qt] Error out early if we don't have ICU available
2147
2148         Reviewed by Tor Arne Vestbø.
2149
2150         * qmake/mkspecs/features/configure.prf:
2151
2152 2012-09-21  Simon Hausmann  <simon.hausmann@digia.com>
2153
2154         [Qt] Bail out when trying to build WebKit with Qt != 5
2155
2156         Reviewed by Tor Arne Vestbø.
2157
2158         Moved check for Qt version out of default_pre into top-level WebKit.pro,
2159         because we never reach default_pre.prf due to the lack of .qmake.conf support
2160         in older versions of Qt/QMake.
2161
2162         * qmake/mkspecs/features/default_pre.prf:
2163
2164 2012-09-21  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
2165
2166         WebKitTestRunner needs to print frame load delegate information
2167         https://bugs.webkit.org/show_bug.cgi?id=42705
2168
2169         Reviewed by Kenneth Rohde Christiansen.
2170
2171         Added missing dumping from WTR::InjectedBundlePage::didFailLoadWithErrorForFrame.
2172
2173         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2174         (WTR::InjectedBundlePage::didFailLoadWithErrorForFrame):
2175
2176 2012-09-21  Christophe Dumez  <christophe.dumez@intel.com>
2177
2178         [EFL] EventSender should mimic CTRL+o emacs shortcut
2179         https://bugs.webkit.org/show_bug.cgi?id=97224
2180
2181         Reviewed by Kenneth Rohde Christiansen.
2182
2183         Add support for mimicking CTRL+o emacs shortcut in EFL's
2184         EventSender in DumpRenderTree and WebKitTestRunner.
2185
2186         * DumpRenderTree/efl/EventSender.cpp:
2187         (sendKeyDown):
2188         * WebKitTestRunner/efl/EventSenderProxyEfl.cpp:
2189         (WTR::EventSenderProxy::keyDown):
2190
2191 2012-09-21  Simon Hausmann  <simon.hausmann@digia.com>
2192
2193         [Qt] Re-fix clean builds
2194
2195         Reviewed by Tor Arne Vestbø.
2196
2197         Re-introduce the sanitization for LIBS when creating a module to use
2198         LIBS_PRIVATE and otherwise do _not_ use LIBS_PRIVATE. We decided to
2199         continue to use QT, LIBS and PKGCONFIG instead of their _PRIVATE variants
2200         throughout the code base, so just using LIBS_PRIVATE in linkAgainstLibrary()
2201         causes build issues when depending system libraries end up in LIBS before
2202         LIBS_PRIVATE.
2203
2204         * qmake/mkspecs/features/default_post.prf:
2205         * qmake/mkspecs/features/functions.prf:
2206
2207 2012-09-20  Simon Fraser  <simon.fraser@apple.com>
2208
2209         Comment out a failing webkitpy unit test until Dirk can fix it.
2210
2211         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2212         (SemanticTests.test_skip_and_wontfix):
2213
2214 2012-09-20  Dirk Pranke  <dpranke@chromium.org>
2215
2216         make Skip, WontFix be the only expectations on a line
2217         https://bugs.webkit.org/show_bug.cgi?id=97225
2218
2219         Reviewed by Ojan Vafai.
2220
2221         It is now incorrect in the new syntax to have a line like:
2222
2223             foo.html [ WontFix Crash ]
2224
2225         This will generate a lint warning and be treated as an invalid
2226         line. Fixing this caused a whole bunch of unit tests to need updating
2227         to no longer be marked as WontFix :). Also, this patch adjusts
2228         the warnings so that missing Bug() identifiers will cause lint
2229         warnings but will *not* cause the line to be treated as invalid.
2230         Fixing these issues also revealed that test_hung_thread was no
2231         longer testing the right logic, so I adjusted the timeouts in
2232         test.py to make that test work again.
2233
2234         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner.py:
2235         (Worker._run_test_in_another_thread):
2236         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2237         (ResultSummaryTest.test_summarized_results_wontfix):
2238         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2239         (TestExpectationParser._parse_modifiers):
2240         (TestExpectationParser._tokenize_line_using_new_format):
2241         (TestExpectationLine.is_invalid):
2242         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2243         (BasicTests.test_basic):
2244         (test_get_test_set):
2245         (test_parse_warning):
2246         (test_pixel_tests_flag):
2247         (SemanticTests.test_missing_bugid):
2248         (SemanticTests):
2249         (SemanticTests.test_skip_and_wontfix):
2250         * Scripts/webkitpy/layout_tests/port/test.py:
2251         (TestDriver.run_test):
2252         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2253         (MainTest.test_hung_thread):
2254         * Scripts/webkitpy/tool/commands/queries_unittest.py:
2255         (PrintExpectationsTest.test_basic):
2256         (PrintExpectationsTest.test_multiple):
2257         (PrintExpectationsTest.test_full):
2258         (PrintExpectationsTest.test_exclude):
2259         (PrintExpectationsTest.test_csv):
2260
2261 2012-09-20  Lucas Forschler  <lforschler@apple.com>
2262
2263         Unreviewed.  Start running tests on the mac-ews.
2264         
2265         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
2266         (MacEWS):
2267
2268 2012-09-20  Tony Chang  <tony@chromium.org>
2269
2270         [chromium] Enable more clang warnings
2271         https://bugs.webkit.org/show_bug.cgi?id=97257
2272
2273         Reviewed by James Robinson.
2274
2275         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp: Add -Wunused-parameter. Don't add -Wglobal-constructors because
2276         there are lots of violations in these files.
2277
2278 2012-09-19  Ryosuke Niwa  <rniwa@webkit.org>
2279
2280         run-perf-tests should record individual value instead of statistics
2281         https://bugs.webkit.org/show_bug.cgi?id=97155
2282
2283         Reviewed by Hajime Morita.
2284
2285         Parse the list of individual value reported by tests and include them as "values".
2286         We strip "values" from the output JSON when uploading it to the perf-o-matic
2287         since it doesn't know how to parse "values" or ignore it.
2288
2289         * Scripts/webkitpy/performance_tests/perftest.py:
2290         (PerfTest):
2291         (PerfTest.parse_output): Parse and report "values".
2292         (PageLoadingPerfTest.run): Report indivisual page loading time in "values".
2293         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
2294         (MainTest.test_parse_output):
2295         (MainTest.test_parse_output_with_failing_line):
2296         (TestPageLoadingPerfTest.test_run):
2297         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
2298         (PerfTestsRunner._generate_and_show_results): Strip "values" from each result
2299         until we update perf-o-matic.
2300         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
2301         (test_run_memory_test):
2302         (test_run_with_json_output):
2303         (test_run_with_description):
2304         (test_run_with_slave_config_json):
2305         (test_run_with_multiple_repositories):
2306
2307 2012-09-20  Simon Hausmann  <simon.hausmann@digia.com>
2308
2309         [Qt] Fix production builds
2310
2311         Unreviewed trivial fix: Follow up r129119 and avoid clobbering LIBS through the use of LIBS_PRIVATE.
2312
2313         Otherwise the libQtWebKit.prl file for example contains -lWebKit1 -lWebCore, etc.
2314
2315         * qmake/mkspecs/features/functions.prf:
2316
2317 2012-09-20  Dirk Pranke  <dpranke@chromium.org>
2318
2319         REGRESSION: layout test results doesn't show diffs
2320         https://bugs.webkit.org/show_bug.cgi?id=97182
2321
2322         Reviewed by Ojan Vafai.
2323
2324         Go back to storing TEXT, AUDIO, and IMAGE+TEXT in results.json
2325         so that results.html (and hopefully garden-o-matic) can
2326         determine which things actually failed. However, we keep mapping
2327         these results to Failure so that we still only have a single
2328         expectation type for them.
2329
2330         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
2331         (JSONLayoutResultsGenerator):
2332         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2333         (TestExpectationParser):
2334         (TestExpectations):
2335         (TestExpectations.result_was_expected):
2336         * Scripts/webkitpy/layout_tests/models/test_failures.py:
2337         (determine_result_type):
2338         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2339         (MainTest.test_missing_and_unexpected_results):
2340         (MainTest.test_retrying_and_flaky_tests):
2341
2342 2012-09-20  Tommy Widenflycht  <tommyw@google.com>
2343
2344         MediaStream API: Extend UserMediaRequest with a ownerDocument method
2345         https://bugs.webkit.org/show_bug.cgi?id=97095
2346
2347         Reviewed by Adam Barth.
2348
2349         Extending WebUserMediaClientMock to check that the owning document is valid,
2350         and that the document has a frame.
2351
2352         * DumpRenderTree/chromium/WebUserMediaClientMock.cpp:
2353         (WebKit::WebUserMediaClientMock::requestUserMedia):
2354
2355 2012-09-20  Stephen Chenney  <schenney@chromium.org>
2356
2357         [Chromium] DRT does not support --dump-all-pixels flag
2358         https://bugs.webkit.org/show_bug.cgi?id=95098
2359
2360         Reviewed by Dirk Pranke.
2361
2362         Add support for the --pixel-tests and shorthand -p option in Chromium DumpRenderTree. Use
2363         of this flag causes pixel results to be created for all tests, regardless of
2364         individual test options. If an individual test provides a pixel hash it will be used,
2365         otherwise the hash will be empty. This replaces a previously defined but unused option
2366         --dump-all-pixels, and is useful primarily when debugging DRT instances.
2367
2368         * DumpRenderTree/chromium/DumpRenderTree.cpp:
2369         (runTest): Add a parameter and code to force pixel results for the test.
2370         (main): Add parameter handling for --pixels-test and -p, and remove --dump-all-pixels.
2371
2372 2012-09-20  Carlos Garcia Campos  <cgarcia@igalia.com>
2373
2374         [GTK] run-api-tests should not buffer test stdout
2375         https://bugs.webkit.org/show_bug.cgi?id=88474
2376
2377         Reviewed by Philippe Normand.
2378
2379         * Scripts/run-gtk-tests:
2380         (TestRunner._run_test_command): Use os.forkpty() instead of
2381         subprocess.Popen() so that gtest sends the output with colors to
2382         stdout. Use common.parse_output_lines() to parse the output and
2383         write it to stdout while it's read.
2384         (TestRunner._run_test_command.parse_line): Parse the line to get
2385         the test pid and write the line to stdout.
2386         (TestRunner._run_test_command.waitpid): Helper function to call
2387         waitpid handling EINTR.
2388         (TestRunner._run_test_command.return_code_from_exit_status):
2389         Helper function to convert exit status of test commands to a
2390         return code.
2391         * gtk/common.py:
2392         (parse_output_lines): Helper function that uses select to read
2393         the given file descriptor and call the given callback for every
2394         line read.
2395
2396 2012-09-20  Simon Hausmann  <simon.hausmann@digia.com>
2397
2398         [Qt] Fix initial build
2399
2400         Reviewed by Tor Arne Vestbø.
2401
2402         When building QtWebKit the first time there is no qt_webkit.pri module pri file, and therefore
2403         $$QT.webkit.name isn't set and so creating_module isn't set. That has all sorts of implications
2404         causing incorrect linking for Makefile.api, etc.
2405
2406         Fix the determination by simply checking if MODULE is set, which only happens in api.pri.
2407
2408         * qmake/mkspecs/features/webkit_modules.prf:
2409
2410 2012-09-20  Sheriff Bot  <webkit.review.bot@gmail.com>
2411
2412         Unreviewed, rolling out r129091.
2413         http://trac.webkit.org/changeset/129091
2414         https://bugs.webkit.org/show_bug.cgi?id=97205
2415
2416         It broke perf tests everywhere (Requested by Ossy on #webkit).
2417
2418         * Scripts/webkitpy/performance_tests/perftest.py:
2419         (PerfTest):
2420         (PerfTest.parse_output):
2421         (PageLoadingPerfTest.run):
2422         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
2423         (MainTest.test_parse_output):
2424         (MainTest.test_parse_output_with_failing_line):
2425         (TestPageLoadingPerfTest.test_run):
2426         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
2427         (PerfTestsRunner._generate_and_show_results):
2428         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
2429         (test_run_memory_test):
2430         (test_run_with_json_output):
2431         (test_run_with_description):
2432         (test_run_with_slave_config_json):
2433         (test_run_with_multiple_repositories):
2434
2435 2012-09-20  Simon Hausmann  <simon.hausmann@digia.com>
2436
2437         [Qt] QtWebKit module header includes private dependencies
2438
2439         Reviewed by Tor Arne Vestbø.
2440
2441         * qmake/mkspecs/features/default_post.prf: Don't try to sanitize LIBS, because we can make sure
2442         that LIBS_PRIVATE is set from the beginning. Moved the creating_module and PKGCONFIG/QT(_PRIVATE)
2443         sanitization into webkit_modules.prf. creating_module determination requires TARGET to be set, so
2444         we can't do it in default_pre.
2445         * qmake/mkspecs/features/webkit_modules.prf:
2446
2447 2012-09-20  Simon Hausmann  <simon.hausmann@digia.com>
2448
2449         [Qt] Fix build with MingW
2450
2451         Reviewed by Tor Arne Vestbø.
2452
2453         Don't try to compile WebCore, etc. with debug symbols for production
2454         builds, it's just too big.
2455
2456         * qmake/mkspecs/features/production_build.prf:
2457
2458 2012-09-20  Michał Pakuła vel Rutka  <m.pakula@samsung.com>
2459
2460         [EFL][DRT]DumpRenderTree needs to reset focus state when test starts.
2461         https://bugs.webkit.org/show_bug.cgi?id=97087
2462
2463         Reviewed by Gyuyoung Kim.
2464
2465         Add focusing a main frame on settings reset.
2466         After editing/undo/undo-iframe-location-change was executed a frame
2467         was left in unfocused state. This caused flakiness in two tests results.
2468
2469         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
2470         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
2471
2472 2012-09-20  Allan Sandfeld Jensen  <allan.jensen@digia.com>
2473
2474         [Qt] Add eventSender.gestureTap
2475         https://bugs.webkit.org/show_bug.cgi?id=66173
2476
2477         Reviewed by Kenneth Rohde Christiansen.
2478
2479         Implement support for sending raw QGesture events.
2480
2481         * DumpRenderTree/qt/EventSenderQt.cpp:
2482         (EventSender::EventSender):
2483         (EventSender::gestureTap):
2484         * DumpRenderTree/qt/EventSenderQt.h:
2485         (EventSender):
2486
2487 2012-09-19  Sheriff Bot  <webkit.review.bot@gmail.com>
2488
2489         Unreviewed, rolling out r129007.
2490         http://trac.webkit.org/changeset/129007
2491         https://bugs.webkit.org/show_bug.cgi?id=97172
2492
2493         It broke the build on the Qt bots (Requested by Ossy on
2494         #webkit).
2495
2496         * qmake/config.tests/gccdepends/empty.cpp:
2497         * qmake/config.tests/gccdepends/gccdepends.pro:
2498         * qmake/mkspecs/features/default_pre.prf:
2499
2500 2012-09-19  Ryosuke Niwa  <rniwa@webkit.org>
2501
2502         run-perf-tests should record indivisual value instead of statistics
2503         https://bugs.webkit.org/show_bug.cgi?id=97155
2504
2505         Reviewed by Hajime Morita.
2506
2507         Parse the list of indivisual value reported by tests and include them as "values".
2508         We strip "values" from the output JSON when uploading it to the perf-o-matic
2509         since it doesn't know how to parse "values" or ignore it.
2510
2511         * Scripts/webkitpy/performance_tests/perftest.py:
2512         (PerfTest):
2513         (PerfTest.parse_output): Parse and report "values".
2514         (PageLoadingPerfTest.run): Report indivisual page loading time in "values".
2515         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
2516         (MainTest.test_parse_output):
2517         (MainTest.test_parse_output_with_failing_line):
2518         (TestPageLoadingPerfTest.test_run):
2519         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
2520         (PerfTestsRunner._generate_and_show_results): Strip "values" from each result
2521         until we update perf-o-matic.
2522         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
2523         (test_run_memory_test):
2524         (test_run_with_json_output):
2525         (test_run_with_description):
2526         (test_run_with_slave_config_json):
2527         (test_run_with_multiple_repositories):
2528
2529 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2530
2531         Fix regex groups for bug matching in flakiness dashboard.
2532         https://bugs.webkit.org/show_bug.cgi?id=97152
2533
2534         Unreviewed, build fix.
2535
2536         * TestResultServer/static-dashboards/flakiness_dashboard.js:
2537         (htmlForBugs):
2538
2539 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2540
2541         nrwt: print unexpected results using new TestExpectations syntax
2542         https://bugs.webkit.org/show_bug.cgi?id=97159
2543
2544         Unreviewed, build fix.
2545
2546         Change new-run-webkit-tests to print out failures using the new
2547         syntax when there are unexpected results, e.g.:
2548
2549         52 tests ran as expected, 19 didn't:
2550
2551
2552         Regressions: Unexpected failures : (2)
2553           failures/flaky/text.html [ Failure ]
2554           failures/unexpected/text-image-checksum.html [ Failure ]
2555
2556         and so forth
2557
2558         * Scripts/webkitpy/layout_tests/views/printing.py:
2559         (Printer._print_unexpected_results):
2560
2561 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2562
2563         update flakiness dashboard after cutover to new test expectations syntax
2564         https://bugs.webkit.org/show_bug.cgi?id=97152
2565
2566         Unreviewed, build fix.
2567
2568         Handle (??) the new Bug notations as well. Hopefully we don't
2569         still need the old ones.
2570
2571         * TestResultServer/static-dashboards/flakiness_dashboard.js:
2572         (filterBugs):
2573         (htmlForBugs):
2574         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
2575
2576 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2577
2578         update flakiness dashboard after cutover to new test expectations syntax
2579         https://bugs.webkit.org/show_bug.cgi?id=97152
2580
2581         Reviewed by Ryosuke Niwa.
2582
2583         This change clones the TestExpectation parsing state machine
2584         from python into javascript.
2585
2586         * TestResultServer/static-dashboards/flakiness_dashboard.js:
2587         (parsedExpectations.lines.forEach):
2588         (parsedExpectations):
2589         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
2590
2591 2012-09-19  Dana Jansens  <danakj@chromium.org>
2592
2593         Add backer@chromium.org as contributor
2594         https://bugs.webkit.org/show_bug.cgi?id=97150
2595
2596         * Scripts/webkitpy/common/config/committers.py:
2597
2598 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2599
2600         fix MISSING after TestExpectations conversion
2601         https://bugs.webkit.org/show_bug.cgi?id=97148
2602
2603         Unreviewed, expectations change / build fix.
2604
2605         * Scripts/convert-test-expectations:
2606         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2607         (TestExpectationParser):
2608
2609 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2610
2611         nrwt: convert chromium TestExpectations to the new syntax
2612         https://bugs.webkit.org/show_bug.cgi?id=97139
2613
2614         Reviewed by Ryosuke Niwa.
2615     
2616         Add in a file temporarily that does the conversion of
2617         test expectations formats
2618
2619         * Tools/Scripts/convert-test-expectations: Added.
2620
2621 2012-09-19  Ryosuke Niwa  <rniwa@webkit.org>
2622
2623         REGRESSION: run-perf-tests no longer reports the total test time
2624         https://bugs.webkit.org/show_bug.cgi?id=97138
2625
2626         Reviewed by Tony Chang.
2627
2628         Report the finished time as a info-level log as opposed to a debug level log.
2629
2630         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
2631         (PerfTestsRunner._run_single_test):
2632         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
2633         (MainTest.normalizeFinishedTime): Added. It replaces all finished times by 0.1 seconds.
2634         (test_run_test_pause_before_testing):
2635         (test_run_test_set_for_parser_tests):
2636         (test_run_memory_test):
2637         (_test_run_with_json_output):
2638
2639 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2640
2641         Support new TestExpectations format alongside old one
2642         https://bugs.webkit.org/show_bug.cgi?id=96588
2643
2644         Reviewed by Ojan Vafai.
2645
2646         This patch adds support for actually reading in lines formatted
2647         in the new style, and when we re-serialize/write out the file,
2648         all lines will be written in the new style. Note that reading in
2649         the old style is still supported, and no updates are being made
2650         to the actual TestExpectations files as part of this change.
2651
2652         This change updates most but not all of the unit tests to use
2653         the new syntax. I will update the rest when (or before where
2654         possible) I drop support for the old format.
2655
2656         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2657         (TestExpectationParser):
2658         (TestExpectationParser._tokenize_line):
2659         (TestExpectationParser._tokenize_line_using_new_format):
2660         (TestExpectationLine.to_string):
2661         (TestExpectationLine._format_line):
2662         (TestExpectations):
2663         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2664         (test_parse_warning):
2665         (SkippedTests.test_skipped_entry_dont_exist):
2666         (NewExpectationSyntaxTests.test_warnings):
2667         (RemoveConfigurationsTest.test_remove):
2668         (test_remove_line):
2669         (RebaseliningTest.test_remove):
2670         (RebaseliningTest.test_no_get_rebaselining_failures):
2671         (TestExpectationParserTests.test_tokenize_blank):
2672         (TestExpectationParserTests.test_tokenize_extra_colon):
2673         (TestExpectationParserTests.test_tokenize_missing_equal):
2674         (TestExpectationParserTests.test_tokenize_extra_equal):
2675         (TestExpectationSerializationTests.test_unparsed_to_string):
2676         (TestExpectationSerializationTests.test_unparsed_list_to_string):
2677         (TestExpectationSerializationTests.test_parsed_to_string):
2678         (TestExpectationSerializationTests.test_format_line):
2679         (TestExpectationSerializationTests.test_string_roundtrip):
2680         (TestExpectationSerializationTests.test_list_roundtrip):
2681         (TestExpectationSerializationTests.test_reconstitute_only_these):
2682         (TestExpectationSerializationTests.test_string_whitespace_stripping):
2683         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2684         (LintTest.test_lint_test_files__errors):
2685         * Scripts/webkitpy/tool/commands/queries_unittest.py:
2686         (PrintExpectationsTest.test_basic):
2687         (PrintExpectationsTest.test_multiple):
2688         (PrintExpectationsTest.test_full):
2689         (PrintExpectationsTest.test_exclude):
2690         * Scripts/webkitpy/tool/commands/rebaseline.py:
2691         (RebaselineExpectations.execute):
2692         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
2693         (TestRebaseline.test_rebaseline_updates_expectations_file_noop):
2694         (test_rebaseline_updates_expectations_file):
2695         (test_rebaseline_does_not_include_overrides):
2696         (test_rebaseline_expectations_noop):
2697         (test_overrides_are_included_correctly):
2698
2699 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2700
2701         implement first part of support for the new TestExpectations syntax
2702         https://bugs.webkit.org/show_bug.cgi?id=96569
2703
2704         Reviewed by Ryosuke Niwa.
2705
2706         This patch implements support for parsing a line of the new
2707         format for the TestExpectations file and converting it back into
2708         the old format for compatibility. This routine is not yet used
2709         by anything.
2710
2711         The new format is documented at:
2712           http://trac.webkit.org/wiki/TestExpectations
2713
2714         but, in short:
2715
2716           [bugs] [ "[" modifiers "]" ] test_name [ "[" expectations "]" ]
2717
2718         - Comments are indicated with "#" instead of "//"
2719         - If no expectations are specified we default to Skip for
2720           compatibility with the Skipped files (these two changes make
2721           Skipped files a subset of TestExpectations files)
2722
2723         - All of the tokens are now CamelCase instead of ALLCAPS.
2724         - FAIL -> Failure
2725         - IMAGE -> ImageOnlyFailure
2726         - WONTFIX -> WontFix
2727         - modifiers refer to just the platforms and configurations
2728           (release/debug) that the line applies to.
2729         - WontFix, Rebaseline, Slow, and Skip move to the right-hand side as
2730           expectations
2731         - expectations will typically be written out in lexicographic order
2732         - We use webkit.org/b/12345, crbug.com/12345, and Bug(dpranke)
2733           instead of BUGWK12345, BUGCR12345, and BUGDPRANKE.
2734
2735         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2736         (TestExpectationParser):
2737         (TestExpectationParser._tokenize_line_using_new_format):
2738         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2739         (NewExpectationSyntaxTests):
2740         (NewExpectationSyntaxTests.assert_exp):
2741         (NewExpectationSyntaxTests.test_bare_name):
2742         (NewExpectationSyntaxTests.test_bare_name_and_bugs):
2743         (NewExpectationSyntaxTests.test_comments):
2744         (NewExpectationSyntaxTests.test_config_modifiers):
2745         (NewExpectationSyntaxTests.test_unknown_config):
2746         (NewExpectationSyntaxTests.test_unknown_expectation):
2747         (NewExpectationSyntaxTests.test_skip):
2748         (NewExpectationSyntaxTests.test_slow):
2749         (NewExpectationSyntaxTests.test_wontfix):
2750         (NewExpectationSyntaxTests.test_blank_line):
2751         (NewExpectationSyntaxTests.test_warnings):
2752
2753 2012-09-19  Dirk Pranke  <dpranke@chromium.org>
2754
2755         nrwt: replace TEXT, AUDIO, and IMAGE+TEXT with FAIL
2756         https://bugs.webkit.org/show_bug.cgi?id=96845
2757
2758         Reviewed by Ojan Vafai.
2759
2760         In preparation for the new TestExpectations syntax, we replace
2761         all TEXT, IMAGE+TEXT, and AUDIO failures with FAIL. This will
2762         make switching to the new syntax lossless (i.e., we lose
2763         information now, not then).
2764
2765         Note that we can still parse in results.json files that have the
2766         old data for backwards compatibility.
2767         
2768
2769         * Scripts/webkitpy/common/net/resultsjsonparser.py:
2770         (JSONTestResult._failure_types_from_actual_result):
2771         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
2772         (ResultsJSONParserTest):
2773         (test_basic):
2774         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2775         (ResultSummaryTest.test_summarized_results_wontfix):
2776         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
2777         (JSONLayoutResultsGenerator):
2778         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2779         (TestExpectations):
2780         (TestExpectations.remove_pixel_failures):
2781         (TestExpectations.has_pixel_failures):
2782         (TestExpectations.suffixes_for_expectations):
2783         (TestExpectations.get_rebaselining_failures):
2784         (TestExpectations.remove_configuration_from_test):
2785         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
2786         (assert_bad_expectations):
2787         (BasicTests):
2788         (BasicTests.test_basic):
2789         (MiscTests):
2790         (MiscTests.test_multiple_results):
2791         (MiscTests.test_result_was_expected):
2792         (MiscTests.test_remove_pixel_failures):
2793         (MiscTests.test_suffixes_for_expectations):
2794         (test_get_expectations_string):
2795         (test_parse_warning):
2796         (test_error_on_different_platform):
2797         (test_error_on_different_build_type):
2798         (test_overrides):
2799         (test_overrides__directory):
2800         (test_overrides__duplicate):
2801         (test_pixel_tests_flag):
2802         (test_more_specific_override_resets_skip):
2803         (SkippedTests.test_skipped_file_overrides_expectations):
2804         (SkippedTests.test_skipped_dir_overrides_expectations):
2805         (SkippedTests.test_skipped_file_overrides_overrides):
2806         (SkippedTests.test_skipped_dir_overrides_overrides):
2807         (ExpectationSyntaxTests.test_missing_colon):
2808         (ExpectationSyntaxTests.test_too_many_equals_signs):
2809         (SemanticTests.test_bug_format):
2810         (SemanticTests.test_bad_bugid):
2811         (SemanticTests.test_missing_bugid):
2812         (SemanticTests.test_rebaseline):
2813         (test_missing_file):
2814         (test_more_modifiers):
2815         (test_order_in_file):
2816         (test_macro_overrides):
2817         (RemoveConfigurationsTest.test_remove):
2818         (test_remove_line):
2819         (RebaseliningTest.test_remove):
2820         (TestExpectationSerializationTests.test_serialize_parsed_expectations):
2821         * Scripts/webkitpy/layout_tests/models/test_failures.py:
2822         (determine_result_type):
2823         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
2824         (test_test_expectations):
2825         * Scripts/webkitpy/layout_tests/port/test.py:
2826         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2827         (MainTest.test_missing_and_unexpected_results):
2828         (MainTest.test_retrying_and_flaky_tests):
2829         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2830         (TestExpectationsTestCase.test_valid_expectations):
2831         * Scripts/webkitpy/tool/commands/queries_unittest.py:
2832         (PrintExpectationsTest.test_basic):
2833         (PrintExpectationsTest.test_multiple):
2834         (PrintExpectationsTest.test_full):
2835         (PrintExpectationsTest.test_exclude):
2836         (PrintExpectationsTest.test_csv):
2837         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
2838         (test_rebaseline_does_not_include_overrides):
2839         (test_overrides_are_included_correctly):
2840
2841 2012-09-19  Dominic Mazzoni  <dmazzoni@google.com>
2842
2843         AX: A few control types are returning the wrong answer for isReadOnly
2844         https://bugs.webkit.org/show_bug.cgi?id=96735
2845
2846         Reviewed by Chris Fleizach.
2847
2848         Exposing isReadOnly in an AccessibilityObject to DumpRenderTree.
2849
2850         * DumpRenderTree/chromium/TestRunner/AccessibilityUIElementChromium.cpp:
2851         (AccessibilityUIElement::AccessibilityUIElement):
2852         (AccessibilityUIElement::isReadOnlyGetterCallback):
2853         * DumpRenderTree/chromium/TestRunner/AccessibilityUIElementChromium.h:
2854         (AccessibilityUIElement):
2855
2856 2012-09-19  Sudarsana Nagineni  <sudarsana.nagineni@intel.com>
2857
2858         [WTR] Memory leaks in TestRunner::deliverWebIntent()
2859         https://bugs.webkit.org/show_bug.cgi?id=97111
2860
2861         Reviewed by Kenneth Rohde Christiansen.
2862
2863         Fix memory leaks in deliverWebIntent() by adopting strings
2864         created with WKStringCreateWithUTF8CString().
2865
2866         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
2867         (WTR::TestRunner::deliverWebIntent):
2868
2869 2012-09-19  Simon Hausmann  <simon.hausmann@digia.com>
2870
2871         [Qt] Fix incremental builds with all-in-one-files and gccdepends
2872
2873         Reviewed by Tor Arne Vestbø.
2874
2875         Pass -MP to gcc when we use the gcc depends feature, to ensure that implicit rules
2876         are not only created for header files but also for .cpp files. AllInOne.cpp files
2877         include other .cpp files, and when those are removed we need those dummy rules to
2878         avoid a "No rule to make Foo.cpp required by AllInOne.o" error.
2879
2880         * qmake/config.tests/gccdepends/empty.cpp:
2881         (main):
2882         * qmake/config.tests/gccdepends/gccdepends.pro:
2883         * qmake/mkspecs/features/default_pre.prf:
2884
2885 2012-09-19  Rick Byers  <rbyers@chromium.org>
2886
2887         Do touch adjustment on GestureTapDown
2888         https://bugs.webkit.org/show_bug.cgi?id=96677
2889
2890         Reviewed by Antonio Gomes.
2891
2892         Allow radius to be set for GestureTapDown events.
2893         * DumpRenderTree/chromium/TestRunner/EventSender.cpp:
2894         (EventSender::gestureEvent):
2895
2896 2012-09-19  Christophe Dumez  <christophe.dumez@intel.com>
2897
2898         [EFL] EFL's DRT does not support overriding 'WebKitCSSRegionsEnabled' preference
2899         https://bugs.webkit.org/show_bug.cgi?id=97100
2900
2901         Reviewed by Gyuyoung Kim.
2902
2903         EFL's DRT now supports overriding the 'WebKitCSSRegionsEnabled'
2904         preference, in order to match WebKitTestRunner functionality.
2905
2906         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
2907         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
2908         * DumpRenderTree/efl/TestRunnerEfl.cpp:
2909         (TestRunner::overridePreference):
2910
2911 2012-09-18  Kangil Han  <kangil.han@samsung.com>
2912
2913         [WK2][WTR] CodeGeneratorTestRunner could keep original copyright.
2914         https://bugs.webkit.org/show_bug.cgi?id=96181
2915
2916         Reviewed by Daniel Bates.
2917
2918         This patch enabled derived files, in DerivedSources/InjectedBundle, to keep original copyright.
2919
2920         * WebKitTestRunner/InjectedBundle/Bindings/CodeGeneratorTestRunner.pm:
2921         (new):
2922         (_parseLicenseBlock):
2923         (_parseLicenseBlockFromFile):
2924         (_defaultLicenseBlock):
2925         (_licenseBlock):
2926         (_generateHeaderFile):
2927         (_generateImplementationFile):
2928
2929 2012-09-18  Byungwoo Lee  <bw80.lee@samsung.com>
2930
2931         Title string should be changed when document.title is set to ''.
2932         https://bugs.webkit.org/show_bug.cgi?id=96793
2933
2934         Reviewed by Kenneth Rohde Christiansen.
2935
2936         Change dump format of dumpTitleChanges more understandable.
2937         Uses single quotation marks for the title string.
2938
2939         * DumpRenderTree/blackberry/DumpRenderTree.cpp:
2940         (BlackBerry::WebKit::DumpRenderTree::didReceiveTitleForFrame):
2941         * DumpRenderTree/chromium/WebViewHost.cpp:
2942         (WebViewHost::didReceiveTitle):
2943         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
2944         (DumpRenderTreeChrome::onFrameTitleChanged):
2945         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2946         (webViewTitleChanged):
2947         * DumpRenderTree/mac/FrameLoadDelegate.mm:
2948         (-[FrameLoadDelegate webView:didReceiveTitle:forFrame:]):
2949         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
2950         (WebCore::DumpRenderTree::titleChanged):
2951         * DumpRenderTree/win/FrameLoadDelegate.cpp:
2952         (FrameLoadDelegate::didReceiveTitle):
2953         * DumpRenderTree/wx/DumpRenderTreeWx.cpp:
2954         (LayoutWebViewEventHandler::OnReceivedTitleEvent):
2955         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2956         (WTR::InjectedBundlePage::didReceiveTitleForFrame):
2957
2958 2012-09-18  Szilard Ledan  <szledan@inf.u-szeged.hu>
2959
2960         EWS shouldn't sleep if there are new patches in its queue
2961         https://bugs.webkit.org/show_bug.cgi?id=83038
2962
2963         Reviewed by Eric Seidel.
2964
2965         EWS tries to process a security patch. Of course it can't, because the EWS isn't
2966         the member of the security group. But the problem is that after it can't process
2967         the attachment, it says that queue is empty (but it isn't!) and it sleeps 2 minutes
2968         and push the security patch to the end of the queue.
2969         Now it stays in the loop until it finds a patch or the queue gets empty.
2970
2971         * Scripts/webkitpy/tool/commands/queues.py:
2972         (AbstractPatchQueue._next_patch):
2973         * Scripts/webkitpy/tool/commands/queues_unittest.py:
2974         (AbstractPatchQueueTest.test_next_patch):
2975
2976 2012-09-18  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
2977
2978         [WK2][WTR] InjectedBundle::booleanForKey() should handle literals effectively
2979         https://bugs.webkit.org/show_bug.cgi?id=97014
2980
2981         Reviewed by Kenneth Rohde Christiansen.
2982
2983         According to http://trac.webkit.org/wiki/EfficientStrings WTF::StringBuilder::appendLiteral() shall
2984         be used for literals rather than WTF::StringBuilder::append().
2985
2986         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
2987         (WTR::InjectedBundle::booleanForKey):
2988
2989 2012-09-18  Simon Hausmann  <simon.hausmann@digia.com>
2990
2991         Update my e-mail address.
2992
2993         * Scripts/webkitpy/common/config/committers.py:
2994
2995 2012-09-18  Andras Becsi  <andras.becsi@digia.com>
2996
2997         Update my e-mail address.
2998
2999         Unreviewed.
3000
3001         * Scripts/webkitpy/common/config/committers.py:
3002
3003 2012-09-18  Andras Becsi  <andras.becsi@digia.com>
3004
3005         [Qt] qt_webkit.pri should not be listed in Tools.pro
3006
3007         Reviewed and rubber-stamped by Simon Hausmann and Tor Arne Vestbø.
3008
3009         Since r128751 the module pri file is auto-generated
3010         but it was still listed in OTHER_FILES.
3011
3012         * Tools.pro: Remove unneeded line.
3013
3014 2012-09-18  Simon Hausmann  <simon.hausmann@digia.com>
3015
3016         [Qt] Fix build with some versions of the gold linker
3017
3018         Reviewed by Tor Arne Vestbø.
3019
3020         Don't unconditionally pass --no-keep-memory to the linker, some versions might not support it.
3021         Instead run a compile/link test first to see if it works.
3022
3023         * qmake/config.tests/gnuld/gnuld.pro: Added.
3024         * qmake/config.tests/gnuld/main.cpp: Added.
3025         (main):
3026         * qmake/mkspecs/features/unix/default_post.prf:
3027
3028 2012-09-18  Simon Hausmann  <simon.hausmann@digia.com>
3029
3030         [Qt] Remove forced use of gold.
3031
3032         Reviewed by Tor Arne Vestbø.
3033
3034         The choice of what linker to use with WebKit should be taken by Qt's build system and ideally the same for all
3035         modules of Qt. Then in turn it is usually up to the administrator of the machine. Recent Debian based systems
3036         often offer the automatic use of gold through a symlink and a dpkg-diversion when installing the gold package.
3037
3038         * qmake/mkspecs/features/unix/default_post.prf:
3039
3040 2012-09-18  Simon Hausmann  <simon.hausmann@digia.com>
3041
3042         [Qt] Prospective Qt/Windows cross-compiling fix
3043
3044         Reviewed by Tor Arne Vestbø.
3045
3046         The win32 scope is not set when cross-compiling from Linux to Windows.
3047
3048         * qmake/mkspecs/features/functions.prf:
3049
3050 2012-09-18  Seokju Kwon  <seokju.kwon@samsung.com>
3051
3052         [EFL] Remove background view on EWebLauncher and MiniBrowser
3053         https://bugs.webkit.org/show_bug.cgi?id=96905
3054
3055         Reviewed by Gyuyoung Kim.
3056
3057         The size of webview was changed after adding url bar.
3058         And the background view is not necessary anymore, since it was used for debugging back in the day.
3059
3060         * EWebLauncher/main.c:
3061         (_ELauncher):
3062         (on_ecore_evas_resize):
3063         (browserCreate):
3064         * MiniBrowser/efl/main.c:
3065         (_MiniBrowser):
3066         (on_ecore_evas_resize):
3067         (browserCreate):
3068
3069 2012-09-18  Sergio Villar Senin  <svillar@igalia.com>
3070
3071         [GTK] run-webkit-tests unable to find TestExpectations for WK2
3072         https://bugs.webkit.org/show_bug.cgi?id=96998
3073
3074         Reviewed by Philippe Normand.
3075
3076         We should look for TestExpectations files in all the locations where
3077         we currently look for Skipped files. This will allow
3078         run-webkit-tests to look for TestExpectations files in
3079         platform/gtk-wk2 and platform/wk2 if the "-2" flag is used.
3080
3081         * Scripts/webkitpy/layout_tests/port/gtk.py:
3082         (GtkPort.expectations_files):
3083
3084 2012-09-18  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3085
3086         [Qt] Fix build without the QtQuick module
3087
3088         Reviewed by Simon Hausmann.
3089
3090         * qmake/mkspecs/features/features.prf:
3091
3092 2012-09-18  Stephanie Lewis  <slewis@apple.com>
3093
3094         Build fix after http://trac.webkit.org/projects/webkit/changeset/128852.
3095
3096         Unreviewed.
3097
3098         * DumpRenderTree/mac/DumpRenderTree.mm:
3099         (poseAsClass):
3100
3101 2012-09-18  Allan Sandfeld Jensen  <allan.jensen@nokia.com>
3102
3103         Unreviewed update of email addresses for Berlin QtWebKit office.
3104
3105         * Scripts/webkitpy/common/config/committers.py:
3106
3107 2012-09-17  Zan Dobersek  <zandobersek@gmail.com>
3108
3109         [GTK] fast/loader/display-image-unset-can-block-image-and-can-reload-in-place.html failing after r128645
3110         https://bugs.webkit.org/show_bug.cgi?id=96899
3111
3112         Reviewed by Martin Robinson.
3113
3114         When overriding the 'WebKitDisplayImageKey' preference, use the
3115         'auto-load-images' property of WebKitWebSettings as the property which
3116         should be updated with the corresponding preference value.
3117
3118         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3119         (resetDefaultsToConsistentValues):
3120         * DumpRenderTree/gtk/TestRunnerGtk.cpp:
3121         (TestRunner::overridePreference):
3122
3123 2012-09-17  Pratik Solanki  <psolanki@apple.com>
3124
3125         DumpRenderTree and WebKitTestRunner should compile with -Wundef on Mac
3126         https://bugs.webkit.org/show_bug.cgi?id=96973
3127
3128         Reviewed by Dan Bernstein.
3129
3130         * DumpRenderTree/mac/Configurations/Base.xcconfig:
3131         * WebKitTestRunner/Configurations/Base.xcconfig:
3132         * WebKitTestRunner/PlatformWebView.h: Use #ifdef __OBJC__ and not #if.
3133
3134 2012-09-17  Dirk Pranke  <dpranke@chromium.org>
3135
3136         nrwt: remove "unexpected EOF" warnings
3137         https://bugs.webkit.org/show_bug.cgi?id=96970
3138
3139         Reviewed by Ojan Vafai.
3140
3141         After debugging this a bit, it looks like there aren't any cases
3142         that I can reproduce where a read() of zero indicates something
3143         actually wrong; either it is a prelude to a crash, or a false
3144         negative. So, I'm removing these warnings and adding a comment.
3145
3146         * Scripts/webkitpy/layout_tests/port/server_process.py:
3147         (ServerProcess._wait_for_data_and_update_buffers_using_select):
3148
3149 2012-09-17  Dirk Pranke  <dpranke@chromium.org>
3150
3151         [chromium] ASAN bot is crashing at the end of the run
3152         https://bugs.webkit.org/show_bug.cgi?id=96967
3153
3154         Reviewed by Abhishek Arya.
3155
3156         The ASAN bot is crashing attempting to decode some output into
3157         UTF-8; there's no reason to do this, so let's not do this and
3158         see if something else is going on as well.
3159
3160         * Scripts/webkitpy/layout_tests/port/chromium.py:
3161         (ChromiumPort._get_crash_log):
3162
3163 2012-09-17  Dirk Pranke  <dpranke@chromium.org>
3164
3165         nrwt: --results-directory isn't getting printed properly
3166         https://bugs.webkit.org/show_bug.cgi?id=96965
3167
3168         Reviewed by Ojan Vafai.
3169
3170         options.results_directory isn't actually initialized with the
3171         default values until after we call print_config(), so this
3172         changes things to print the value directly.
3173
3174         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3175         (run):
3176         * Scripts/webkitpy/layout_tests/views/printing.py:
3177         (Printer.print_config):
3178         * Scripts/webkitpy/layout_tests/views/printing_unittest.py:
3179         (Testprinter.test_print_config):
3180
3181 2012-09-17  Rob Buis  <rbuis@rim.com>
3182
3183         [BlackBerry] Enable VIDEO_TRACK
3184         https://bugs.webkit.org/show_bug.cgi?id=96949
3185
3186         Reviewed by Antonio Gomes.
3187
3188         * Scripts/webkitperl/FeatureList.pm:
3189
3190 2012-09-17  Rick Byers  <rbyers@chromium.org>
3191
3192         Add handling of new GestureTapCancel in DRT
3193
3194         https://bugs.webkit.org/show_bug.cgi?id=96183
3195
3196         Reviewed by Antonio Gomes.
3197
3198         * DumpRenderTree/chromium/TestWebPlugin.cpp:
3199         (TestWebPlugin::handleInputEvent):
3200         * DumpRenderTree/chromium/EventSender.cpp:
3201         (EventSender::gestureTapCancel):
3202
3203 2012-09-17  Philip Rogers  <pdr@google.com>
3204
3205         Teach style checker about preprocessor directive indentation rules
3206         https://bugs.webkit.org/show_bug.cgi?id=96874
3207
3208         Reviewed by Adam Barth.
3209
3210         Preprocessor directives (#ifdef, #include, #define, etc.) should not be indented.
3211         This is not explicit in our style guide but is generally followed in our code.
3212         Searching for violations in our codebase shows these are rarely indented:
3213             #include - indented in 6 files
3214             #ifdef - indented in 0 files
3215             #ifndef - indented in 1 file
3216             #define - indented in 11 files
3217             #if - indented in 7 files
3218
3219         * Scripts/webkitpy/style/checkers/cpp.py:
3220         (check_directive_indentation):
3221
3222             This is the simple test where we look for spaces followed by a #.
3223
3224         (check_style):
3225         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
3226
3227             A few tests needed to be modified because they had unintentionally indented
3228             preprocessor directives.
3229
3230         (CppStyleTest.test_build_class.Foo):
3231         (CppStyleTest.test_build_class):
3232         (CppStyleTest.test_build_class.DERIVE_FROM_GOO):
3233         (WebKitStyleTest.test_line_breaking):
3234         (WebKitStyleTest.test_directive_indentation):
3235
3236 2012-09-17  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3237
3238         [Qt] Auto-generate the module pri file for QtWebKit
3239
3240         Reviewed by Simon Hausmann.
3241
3242         * qmake/qt_webkit.pri: Removed.
3243
3244 2012-09-14  Dirk Pranke  <dpranke@chromium.org>
3245
3246         nrwt: --additional-platform-dir is broken on chromium ports
3247         https://bugs.webkit.org/show_bug.cgi?id=96840
3248
3249         Reviewed by Ojan Vafai.
3250
3251         Looks like I broke this when cleaning things up as part of
3252         adding ML support. Fixing and adding a test.
3253
3254         * Scripts/webkitpy/layout_tests/port/chromium.py:
3255         (ChromiumPort.default_baseline_search_path):
3256         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
3257         (ChromiumAndroidPort.default_baseline_search_path):
3258         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
3259         (ChromiumLinuxPort.default_baseline_search_path):
3260         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
3261         (ChromiumMacPort.__init__):
3262         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
3263         (ChromiumWinPort.setup_environ_for_server):
3264         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
3265         (test_check_build):
3266         (test_additional_platform_directory):
3267
3268 2012-09-13  Stephanie Lewis  <slewis@apple.com>
3269
3270         mac-future ignores previous platform TestExpectations.
3271         https://bugs.webkit.org/show_bug.cgi?id=96718.
3272
3273         Reviewed by Dirk Pranke.
3274
3275         TestExpectations match an expectation's configuraton against a list of configurations that are 
3276         valid for that port.  That list does not contain mac-future so none of the expectations can be
3277         applied.
3278
3279         * Scripts/webkitpy/layout_tests/port/apple.py:
3280         (ApplePort._generate_all_test_configurations):
3281
3282 2012-09-14  Adam Barth  <abarth@webkit.org>
3283
3284         Remove webkitPostMessage
3285         https://bugs.webkit.org/show_bug.cgi?id=96577
3286
3287         Reviewed by Ojan Vafai.
3288
3289         Add ENABLE_LEGACY_VENDOR_PREFIXES flag.
3290
3291         * Scripts/webkitperl/FeatureList.pm:
3292         * qmake/mkspecs/features/features.pri:
3293
3294 2012-09-14  Alexey Proskuryakov  <ap@apple.com>
3295
3296         [WK2] webarchive/loading/javascript-url-iframe-crash.html fails
3297         https://bugs.webkit.org/show_bug.cgi?id=96817
3298
3299         Reviewed by Kenneth Rohde Christiansen.
3300
3301         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3302         (WTR::InjectedBundlePage::didFailProvisionalLoadWithErrorForFrame): This callback
3303         got overlooked when adding logging elsewhere.
3304
3305 2012-09-14  Peter Beverloo  <peter@chromium.org>
3306
3307         [Chromium] Build fix for DumpRenderTree following r128628
3308         https://bugs.webkit.org/show_bug.cgi?id=96808
3309
3310         Reviewed by Tony Chang.
3311
3312         When building DumpRenderTree from the Chromium project, a compile error
3313         shows up because including the header ForwardIOStreamsAndroid.h cannot
3314         be located. TestWebKitAPI does this correctly.
3315
3316         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
3317
3318 2012-09-14  Dana Jansens  <danakj@chromium.org>
3319
3320         Change cc-bugs@google.com watchlist to cc-bugs@chromium.org
3321         https://bugs.webkit.org/show_bug.cgi?id=96805
3322
3323         Reviewed by Dirk Pranke.
3324
3325         * Scripts/webkitpy/common/config/committers.py:
3326         * Scripts/webkitpy/common/config/watchlist:
3327
3328 2012-09-14  Peter Beverloo  <peter@chromium.org>
3329
3330         [Chromium] Support the --{in,out,err}-fifo arguments on TestWebKitAPI and webkit_unit_tests
3331         https://bugs.webkit.org/show_bug.cgi?id=96687
3332
3333         Reviewed by Tony Chang.
3334
3335         Android's DumpRenderTree currently supports these arguments, implemented
3336         as part of TestShellAndroid:
3337         http://trac.webkit.org/browser/trunk/Tools/DumpRenderTree/chromium/TestShellAndroid.cpp?rev=128496
3338
3339         They're used by the layout test runner to get the STDOUT and STDERR while
3340         a layout test run is in process, which is a safer alternative to parsing
3341         all the logcat output manually. The implementation can be seen here:
3342         http://trac.webkit.org/browser/trunk/Tools/Scripts/webkitpy/layout_tests/port/chromium_android.py?rev=128496#L590
3343
3344         This patch generalizes parsing of and applying the effects of these arguments
3345         so that they can be used for TestWebKitAPI and webkit_unit_tests as well.
3346         After this patch, this will make it possible to pull out the output-reading
3347         code from Android's layout test port and generalize it so it can be re-used
3348         in the new test-runner for the other two test suites.
3349
3350         This has no effect when compiling and running these tests as part of Chromium
3351         code, which has a much more advanced test-runner that does parse complete log
3352         output, but also directly depends on code licensed under Apache 2.
3353
3354         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
3355         * DumpRenderTree/chromium/TestShellAndroid.cpp:
3356         (platformInit):
3357         * TestWebKitAPI/TestWebKitAPI.gyp/TestWebKitAPI.gyp:
3358
3359 2012-09-14  Zoltan Horvath  <zoltan@webkit.org>
3360
3361         check-webkit-style should not warn in case of NONCOPYABLE and FAST_ALLOCATED macros
3362         https://bugs.webkit.org/show_bug.cgi?id=96656
3363
3364         Reviewed by Adam Barth.
3365
3366         We should allow to use WTF_MAKE_NONCOPYABLE(ClassName) and WTF_MAKE_FAST_ALLOCATED
3367         macros in 1 line.
3368
3369         * Scripts/webkitpy/style/checkers/cpp.py:
3370         (check_style.definitions):
3371         (check_style): Add rule.
3372         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
3373         (WebKitStyleTest.test_line_breaking): Add unittest.
3374
3375 2012-09-14  Christophe Dumez  <christophe.dumez@intel.com>
3376
3377         WebKitTestRunner needs layoutTestController.dumpDatabaseCallbacks
3378         https://bugs.webkit.org/show_bug.cgi?id=57570
3379
3380         Reviewed by Kenneth Rohde Christiansen.
3381
3382         Implement testRunner.dumpDatabaseCallbacks in WebKitTestRunner
3383         and dump the information expected by the test cases.
3384
3385         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
3386         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
3387         (WTR::InjectedBundle::beginTesting): Explicitly set the default
3388         database quota.
3389         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3390         (WTR::securityOriginToStr): New convenience function to convert
3391         a security origin to the string expected in test results.
3392         (WTR):
3393         (WTR::InjectedBundlePage::InjectedBundlePage):
3394         (WTR::InjectedBundlePage::didExceedDatabaseQuota):
3395         (WTR::InjectedBundlePage::didReachApplicationCacheOriginQuota): Minor
3396         refactoring to share code with didExceedDatabaseQuota.
3397         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
3398         (InjectedBundlePage):
3399         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
3400         (WTR::TestRunner::TestRunner):
3401         * WebKitTestRunner/InjectedBundle/TestRunner.h:
3402         (WTR::TestRunner::dumpDatabaseCallbacks):
3403         (WTR::TestRunner::shouldDumpDatabaseCallbacks):
3404         (TestRunner):
3405
3406 2012-09-14  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3407
3408         [Qt] Set force_static_libs_as_shared early enough to be caught by configure
3409
3410         Rubber-stamped by Simon Hausmann.
3411
3412         * qmake/mkspecs/features/default_post.prf:
3413
3414 2012-09-14  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3415
3416         [Qt] Remove old cruft from the qmake build system
3417
3418         Rubber-stamped by Simon Hausmann.
3419
3420         * DumpRenderTree/qt/DumpRenderTree.pro:
3421         * qmake/mkspecs/features/default_post.prf:
3422
3423 2012-09-14  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3424
3425         [Qt] Enable force_static_libs_as_shared for development
3426
3427         Not enabled for buildbots or production-builds (part of Qt5)
3428
3429         Reviewed by Simon Hausmann.
3430
3431         * qmake/mkspecs/features/default_post.prf:
3432
3433 2012-09-14  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3434
3435         [Qt] Make force_static_libs_as_shared work on Mac OS
3436
3437         We had to move a few LIBS += around that were in the wrong place,
3438         and not caught when everything was just linked into the final
3439         QtWebKit library.
3440
3441         Reviewed by Simon Hausmann.
3442
3443         * qmake/mkspecs/features/default_post.prf:
3444         * qmake/mkspecs/features/force_static_libs_as_shared.prf:
3445
3446 2012-09-14  Simon Hausmann  <simon.hausmann@nokia.com>
3447
3448         [Qt] Fix build of WTR on Windows
3449
3450         Reviewed by Tor Arne Vestbø.
3451
3452         WTR/qt/main.cpp includes TestController.h, which lives in just WTR/
3453         With qmake's unix makefile generator as well as with the MingW generator,
3454         the directory where the pro file lives (WTR/) is automatically added to
3455         the include search path, but not so with MSVC. Instead of relying on qmake's
3456         implicit feature here, add WTR/ explicitly to the include search path.
3457
3458         * WebKitTestRunner/Target.pri:
3459
3460 2012-09-14  Christophe Dumez  <christophe.dumez@intel.com>
3461
3462         WebKitTestRunner needs layoutTestController.setStopProvisionalFrameLoads
3463         https://bugs.webkit.org/show_bug.cgi?id=42691
3464
3465         Reviewed by Kenneth Rohde Christiansen.
3466
3467         Add implementation for testrunner.setStopProvisionalFrameLoads()
3468         in WebKitTestRunner.
3469
3470         Original patch by Alexey Proskuryakov.
3471
3472         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
3473         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3474         (WTR::InjectedBundlePage::didStartProvisionalLoadForFrame):
3475         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
3476         (WTR::TestRunner::TestRunner):
3477         * WebKitTestRunner/InjectedBundle/TestRunner.h:
3478         (TestRunner):
3479         (WTR::TestRunner::setStopProvisionalFrameLoads):
3480         (WTR::TestRunner::shouldStopProvisionalFrameLoads):
3481
3482 2012-09-14  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3483
3484         [Qt] Fix handling of debug/release/debug_and_release/build_all
3485
3486         We now pick up the defaults from Qt, in default_pre (CONFIG already
3487         contains the appropriate values for debug and release). We then let
3488         the command line arguments to qmake, or the project files themselves,
3489         override the configuration, and finally we sanitize the configuration
3490         in default_post.
3491
3492         Reviewed by Simon Hausmann.
3493
3494         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
3495         * Scripts/webkitdirs.pm:
3496         (buildQMakeProjects):
3497         * qmake/mkspecs/features/default_post.prf:
3498         * qmake/mkspecs/features/default_pre.prf:
3499         * qmake/mkspecs/features/mac/default_pre.prf:
3500
3501 2012-09-14  Sudarsana Nagineni  <sudarsana.nagineni@linux.intel.com>
3502
3503         WebKitTestRunner needs layoutTestController.setTabKeyCyclesThroughElements
3504         https://bugs.webkit.org/show_bug.cgi?id=42687
3505
3506         Reviewed by Kenneth Rohde Christiansen.
3507
3508         Implement testRunner.setTabKeyCyclesThroughElements for
3509         WebKitTestRunner.
3510
3511         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
3512         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
3513         (WTR::InjectedBundle::beginTesting):
3514         * WebKitTestRunner/InjectedBundle/TestRunner.cpp:
3515         (WTR::TestRunner::setTabKeyCyclesThroughElements):
3516         (WTR):
3517         * WebKitTestRunner/InjectedBundle/TestRunner.h:
3518         (TestRunner):
3519
3520 2012-09-14  Sudarsana Nagineni  <sudarsana.nagineni@linux.intel.com>
3521
3522         EventSendingController::keyDown does not support non-array modifier arguments
3523         https://bugs.webkit.org/show_bug.cgi?id=96727
3524
3525         Reviewed by Kenneth Rohde Christiansen.
3526
3527         Add support for handling a string as modifier argument to
3528         EventSendingController::keyDown.
3529
3530         * WebKitTestRunner/InjectedBundle/EventSendingController.cpp:
3531         (WTR::parseModifierArray):
3532
3533 2012-09-14  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
3534
3535         WebKitTestRunner needs layoutTestController.setCustomPolicyDelegate
3536         https://bugs.webkit.org/show_bug.cgi?id=42546
3537
3538         Reviewed by Kenneth Rohde Christiansen.
3539
3540         Exported TestRunner::setCustomPolicyDelegate() method.
3541         Allowed Policy Delegate dumping even if we're not in 'waitUntilDone' mode.
3542
3543         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
3544         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3545         (WTR::InjectedBundlePage::decidePolicyForNavigationAction):
3546
3547 2012-09-14  Zan Dobersek  <zandobersek@gmail.com>
3548
3549         [GTK] Clear application cache between tests in DumpRenderTree
3550         https://bugs.webkit.org/show_bug.cgi?id=96543
3551
3552         Reviewed by Philippe Normand.
3553
3554         Call the new DumpRenderTreeSupportGtk method to clear application cache
3555         after each test in DumpRenderTree.
3556
3557         Set the XDG_CACHE_HOME environment variable when running DumpRenderTree
3558         as well. It's now being set in XvfbDriver._start so that each driver is assigned
3559         a cache directory that's constructed out of the driver's name and its worker
3560         number. These directories are located in the layout tests results directory.
3561         At least on the builders this ensures a driver-specific empty cache directory
3562         with each test run.
3563
3564         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3565         (runTest):
3566         * Scripts/webkitpy/layout_tests/port/gtk.py:
3567         (GtkPort.setup_environ_for_server):
3568         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py:
3569         (XvfbDriver._start):
3570
3571 2012-09-14  Simon Hausmann  <simon.hausmann@nokia.com>
3572
3573         [Qt] Fix linking on Windows
3574
3575         Reviewed by Tor Arne Vestbø.
3576
3577         When compiling code _after_ QtWebKit.dll, we need to make sure that the QWEBKIT_EXPORT
3578         macro turns into an import macro, which means QT_MAKEDLL should only be set _up until_
3579         the QtWebKit.dll creation in the build process and be off afterwards. This is done in
3580         win32/default_post.prf using a !contains(QT, webkit): DEFINES += QT_MAKEDLL.
3581
3582         r128450 sanitized the values of QT and PKGCONFIG to only contain the dependencies we
3583         want for QtWebKit.dll, but unfortunately that code was applied to any binary built also
3584         after the module. Therefore QT did never contain webkit anymore and we always passed
3585         QT_MAKEDLL to DEFINES, resulting in QWEBKIT_EXPORT to _not_ turn into an import macro.
3586
3587         This patch introduces a creating_module configuration that is used to
3588         scope the qt/pkgconfig sanitization.
3589
3590         * qmake/mkspecs/features/default_post.prf:
3591
3592 2012-09-14  Christophe Dumez  <christophe.dumez@intel.com>
3593
3594         [WK2][WKTR] TestRunner is not dumping some frame load callbacks information
3595         https://bugs.webkit.org/show_bug.cgi?id=96740
3596
3597         Reviewed by Kenneth Rohde Christiansen.
3598
3599         Dump information expected by the test cases about the following
3600         frame load callbacks:
3601         "didDisplayInsecureContentForFrame"
3602         "didRunInsecureContentForFrame"
3603         "didDetectXSSForFrame"
3604
3605         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3606         (WTR::InjectedBundlePage::willPerformClientRedirectForFrame): Print a URL
3607         as expected by test results by calling pathSuitableForTestResult() instead
3608         of printing raw URL.
3609         (WTR::InjectedBundlePage::didDisplayInsecureContentForFrame):
3610         (WTR::InjectedBundlePage::didRunInsecureContentForFrame):
3611         (WTR::InjectedBundlePage::didDetectXSSForFrame):
3612
3613 2012-09-14  Peter Beverloo  <peter@chromium.org>
3614
3615         The runtime/unsigned category should be valid in the cpp style-checker.
3616         https://bugs.webkit.org/show_bug.cgi?id=96748
3617
3618         Reviewed by Csaba Osztrogonác.
3619
3620         This is causing the webkitpy tests to fail because it's not listed in the
3621         category array. Furthermore, change an "int" to "long" because it's
3622         throwing a double error in one of the earlier asserts.
3623
3624         * Scripts/webkitpy/style/checkers/cpp.py:
3625         (CppChecker):
3626         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
3627         (WebKitStyleTest.test_names):
3628
3629 2012-09-13  Kenneth Rohde Christiansen  <kenneth@webkit.org>
3630
3631         Evas_Object* is a ref'ed structure, so tread it as such
3632         https://bugs.webkit.org/show_bug.cgi?id=96659
3633
3634         Reviewed by Adam Barth.
3635
3636         Replace OwnPtr<Evas_Object> with RefPtr.