7399aa9a3bd8e80d303bb673bf57eeea6513fa82
[WebKit-https.git] / Tools / ChangeLog
1 2013-01-07  Ojan Vafai  <ojan@chromium.org>
2
3         Add the Apple MountainLion bots to garden-o-matic.
4         Eventually, we really should move away from hardcoding
5         the list of bots.
6
7         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/config.js:
8
9 2013-01-07  Jochen Eisinger  <jochen@chromium.org>
10
11         [chromium] move web audio related methods to testrunner library
12         https://bugs.webkit.org/show_bug.cgi?id=106211
13
14         Reviewed by Adam Barth.
15
16         * DumpRenderTree/chromium/DRTTestRunner.cpp:
17         (DRTTestRunner::DRTTestRunner):
18         (DRTTestRunner::reset):
19         * DumpRenderTree/chromium/DRTTestRunner.h:
20         (DRTTestRunner):
21         * DumpRenderTree/chromium/TestRunner/public/WebTestRunner.h:
22         (WebKit):
23         (WebTestRunner::WebTestRunner::shouldDumpAsAudio):
24         (WebTestRunner::WebTestRunner::audioData):
25         * DumpRenderTree/chromium/TestRunner/src/TestRunner.cpp:
26         (WebTestRunner::TestRunner::TestRunner):
27         (WebTestRunner::TestRunner::reset):
28         (WebTestRunner::TestRunner::shouldDumpAsAudio):
29         (WebTestRunner):
30         (WebTestRunner::TestRunner::audioData):
31         (WebTestRunner::TestRunner::setAudioData):
32         * DumpRenderTree/chromium/TestRunner/src/TestRunner.h:
33         (TestRunner):
34         * DumpRenderTree/chromium/TestShell.cpp:
35         (TestShell::dump):
36
37 2013-01-07  Mike West  <mkwst@chromium.org>
38
39         Make the IFRAME_SEAMLESS flag runtime-enabled.
40         https://bugs.webkit.org/show_bug.cgi?id=106213
41
42         Reviewed by Ojan Vafai.
43
44         * DumpRenderTree/chromium/TestShell.cpp:
45         (TestShell::TestShell):
46         * DumpRenderTree/efl/DumpRenderTreeChrome.cpp:
47         (DumpRenderTreeChrome::resetDefaultsToConsistentValues):
48         * DumpRenderTree/gtk/DumpRenderTree.cpp:
49         (resetDefaultsToConsistentValues):
50         * DumpRenderTree/mac/DumpRenderTree.mm:
51         (resetDefaultsToConsistentValues):
52         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
53         (WebCore::WebPage::resetSettings):
54         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
55         (WTR::InjectedBundle::beginTesting):
56             Ensure that seamless IFrames are enabled for layout tests in each of
57             the ports that support seamless IFrames.
58
59 2013-01-07  Jochen Eisinger  <jochen@chromium.org>
60
61         [chromium] move dumpAsText and friends to the TestRunner library
62         https://bugs.webkit.org/show_bug.cgi?id=106222
63
64         Reviewed by Adam Barth.
65
66         This moves dumpAsText, dumpChildFramesAsText, and dumpChildFrameScrollPositions to the TestRunner library.
67
68         * DumpRenderTree/chromium/DRTTestRunner.cpp:
69         (DRTTestRunner::DRTTestRunner):
70         (DRTTestRunner::reset):
71         * DumpRenderTree/chromium/DRTTestRunner.h:
72         (DRTTestRunner):
73         * DumpRenderTree/chromium/TestRunner/public/WebTestRunner.h:
74         (WebTestRunner::WebTestRunner::shouldDumpAsText):
75         (WebTestRunner::WebTestRunner::setShouldDumpAsText):
76         (WebTestRunner::WebTestRunner::shouldGeneratePixelResults):
77         (WebTestRunner::WebTestRunner::setShouldGeneratePixelResults):
78         (WebTestRunner::WebTestRunner::shouldDumpChildFrameScrollPositions):
79         (WebTestRunner::WebTestRunner::shouldDumpChildFramesAsText):
80         * DumpRenderTree/chromium/TestRunner/src/TestRunner.cpp:
81         (WebTestRunner::TestRunner::TestRunner):
82         (WebTestRunner::TestRunner::reset):
83         (WebTestRunner::TestRunner::shouldDumpAsText):
84         (WebTestRunner):
85         (WebTestRunner::TestRunner::setShouldDumpAsText):
86         (WebTestRunner::TestRunner::shouldGeneratePixelResults):
87         (WebTestRunner::TestRunner::setShouldGeneratePixelResults):
88         (WebTestRunner::TestRunner::shouldDumpChildFrameScrollPositions):
89         (WebTestRunner::TestRunner::shouldDumpChildFramesAsText):
90         (WebTestRunner::TestRunner::dumpAsText):
91         (WebTestRunner::TestRunner::dumpChildFrameScrollPositions):
92         (WebTestRunner::TestRunner::dumpChildFramesAsText):
93         * DumpRenderTree/chromium/TestRunner/src/TestRunner.h:
94         (TestRunner):
95
96 2013-01-07  Ryosuke Niwa  <rniwa@webkit.org>
97
98         Build fix attempt after r138810.
99
100         Apparently python 2.6 is really bad at floating point computation.
101
102         * Scripts/webkitpy/performance_tests/perftestsrunner_integrationtest.py:
103         (MainTest.test_run_memory_test):
104
105 2013-01-07  Ryosuke Niwa  <rniwa@webkit.org>
106
107         Extract a class to represent a test result to simplify PerfTest.parse_output and PageLoadPerfTest._run_with_driver
108         https://bugs.webkit.org/show_bug.cgi?id=106172
109
110         Reviewed by Tony Chang.
111
112         Extracted PerfTestMetric, which represents a test result and knows how to compute statistics and convert itself
113         to a dictionary. A metric is a type of measurement each test makes such as "Time", "Malloc", and "JSHeap".
114         A single test (e.g. html5-full-render.html) may have multiple metrics, each of which is reported as a separate
115         test on perf-o-matic and perf. test's results page.
116
117         Also, use python's compute_statistics instead of reading statistics off of DriverOutput to simplify the code.
118
119         * Scripts/webkitpy/performance_tests/perftest.py:
120         (PerfTestMetric): Added.
121         (PerfTestMetric.__init__): Added.
122         (PerfTestMetric.metric): Added.
123         (PerfTestMetric.has_values): Added.
124         (PerfTestMetric.legacy_chromium_bot_compatible_test_name): Added. Following the legacy convention, "Malloc" and
125         "JSHeap" will use "test_name:Malloc" whereas "Time" uses "test_name" itself. We should eventually move these
126         conversions to be done on the client side (perf-o-matic and results page).
127         (PerfTestMetric.append): Added.
128         (PerfTestMetric.to_dict): Added. Computes the statistics and generates the dictionary object that represents
129         a result for this metric.
130         (PerfTestMetric.metric_to_unit): Added. Each metric maps to exactly one unit (it's not one-to-one).
131         (PerfTestMetric.compute_statistics): Moved from PerfTest.
132
133         (PerfTest._create_driver): Extracted from PerfTest.run.
134         (PerfTest.run): Given an array of PerfTestMetric, generate the results dictionary. Also print out results
135         as needed.
136         (PerfTest._run_with_driver):
137         (PerfTest): Removed _result_classes and _result_class_regex. "result class" is the old name for "metric".
138         (PerfTest.parse_output): Dramatically simplified. We now just return an array of PerfTestMetric's. Also ignore
139         all keys except "values". We rely on PerfTestMetric to compute the statistics for us.
140
141         (ChromiumStylePerfTest.run): Added and merged _run_with_driver. Since chromium style perf tests are so different
142         from other perf tests, there isn't much point in sharing the code here.
143
144         (PageLoadingPerfTest._run_with_driver): Simplified. Just create an array of PerfTestMetric's without computing
145         statistics for the measured values.
146
147         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
148         (TestPerfTestMetric): Added.
149         (TestPerfTestMetric.test_init_set_missing_unit): Added.
150         (TestPerfTestMetric.test_legacy_chromium_bot_compatible_test_name): Added.
151         (TestPerfTestMetric.test_has_values): Added.
152         (TestPerfTestMetric.test_append): Added.
153         (TestPerfTestMetric.test_compute_statistics.compute_statistics):
154         (TestPerfTestMetric.test_compute_statistics): Added a trivial test case.
155
156         (TestPerfTest._assert_results_are_correct): Updated as needed.
157         (TestPageLoadingPerfTest.test_run): Ditto. Notice that we no longer log anything in _run_with_driver. It's done
158         in PerfTest.run now after generating the results dictionary.
159         (TestPageLoadingPerfTest.test_run_with_memory_output): Ditto.
160
161 2013-01-07  Julien Chaffraix  <jchaffraix@webkit.org>
162
163         Support size_t multiplication and division operators on LayoutUnit
164         https://bugs.webkit.org/show_bug.cgi?id=83848
165
166         Reviewed by Emil A Eklund.
167
168         * TestWebKitAPI/Tests/WebCore/LayoutUnit.cpp:
169         Added some simple tests for the new operators using size_t.
170
171 2013-01-07  Csaba Osztrogonác  <ossy@webkit.org>
172
173         [Qt] run-webkit-tests cleanup for parallel NRWT
174         https://bugs.webkit.org/show_bug.cgi?id=106220
175
176         Reviewed by Dirk Pranke.
177
178         * Scripts/run-webkit-tests:
179         (useNewRunWebKitTests):
180         * Scripts/webkitpy/layout_tests/port/qt.py:
181         (QtPort):
182         (QtPort.default_child_processes):
183
184 2013-01-07  Philippe Normand  <pnormand@igalia.com>
185
186         Create a GStreamerAudio watchlist and add myself to it.
187
188         Unreviewed.
189
190         * Scripts/webkitpy/common/config/watchlist:
191
192 2013-01-05  Ryosuke Niwa  <rniwa@webkit.org>
193
194         Add svn:executable to a bunch of scripts in BuildSlaveSupport.
195         Also removed the empty mac directory.
196
197         * BuildSlaveSupport/built-product-archive: Added property svn:executable.
198         * BuildSlaveSupport/download-built-product: Added property svn:executable.
199         * BuildSlaveSupport/mac: Removed.
200         * BuildSlaveSupport/test-result-archive: Added property svn:executable.
201
202 2013-01-04  Ryosuke Niwa  <rniwa@webkit.org>
203
204         Add a script to delete stale zero-byte build files
205         https://bugs.webkit.org/show_bug.cgi?id=106131
206
207         Reviewed by Tony Chang.
208
209         Delete zero-byte build files on Mac port for now. It doesn't do anything on other ports.
210
211         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
212         (DeleteStaleBuildFiles): Added.
213         (Factory.__init__): Add DeleteStaleBuildFiles step.
214         * BuildSlaveSupport/delete-stale-build-files: Added.
215         (main): Delete zero-byte files except ones intentionally generated by our build process.
216         (webkitBuildDirectory): Copied from build-product-archive.
217
218 2013-01-04  Michael Pruett  <michael@68k.org>
219
220         [GTK] Fix build error on GTK due to r138836
221         https://bugs.webkit.org/show_bug.cgi?id=106154
222
223         Reviewed by Kentaro Hara.
224
225         * GNUmakefile.am:
226
227 2013-01-04  Dan Bernstein  <mitz@apple.com>
228
229         [mac] WebKit1 has no equivalent of WKBundleNodeHandleCopyHTMLTableCellElementCellAbove
230         https://bugs.webkit.org/show_bug.cgi?id=106121
231
232         Reviewed by Adele Peterson.
233
234         Added test for -[DOMTHMLTableCellElement _cellAbove].
235
236         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
237         * TestWebKitAPI/Tests/mac/DOMHTMLTableCellCellAbove.mm: Added.
238         (-[HTMLTableCellElementCellAboveTest webView:didFinishLoadForFrame:]):
239         (TestWebKitAPI):
240         (TestWebKitAPI::expectCellAboveCell):
241         (TestWebKitAPI::TEST):
242         * TestWebKitAPI/Tests/mac/DOMHTMLTableCellElementCellAbove.html: Added.
243
244 2013-01-04  Eric Seidel  <eric@webkit.org>
245
246         FlakyTestReporter should be re-enabled and taught how to post patches
247         https://bugs.webkit.org/show_bug.cgi?id=103839
248
249         Reviewed by Adam Barth.
250
251         This moves LayoutTestResultsReader off of the DeprecatedPort object.
252         It also adds a real Port object to the EWS and CommitQueue objects
253         using a new PatchProcessingQueue superclass.
254         PatchProcessingQueue doesn't seem to be the cleanest insertion point
255         but works for this purpose.
256         I also moved _upload_results_archive_for_patch into PatchProcessingQueue
257         to resolve a FIXME.
258
259         This required some hacks/shiming between the non-specific "old"
260         port names which the EWS system is currently using, and the new
261         port objects.  This shimming is done in PatchProcessingQueue._new_port_name_from_old.
262
263         Hopefully this will fix the flaky test reporter and results uploading.
264         Certainly it's another small step towards getting rid of the old ports.py
265         infrastructure and making the NRWT Port class common webkitpy infrastructure.
266
267         Now that DeprecatedPort is no longer used for the results directory
268         I removed all the explicit results-dir code which webkit-patch was using
269         in --non-interactive mode to tell NRWT to use a results-dir it understood. :)
270
271         * Scripts/webkitpy/common/config/ports.py:
272         (DeprecatedPort.run_perl_unittests_command):
273         * Scripts/webkitpy/common/config/ports_mock.py:
274         (MockPort.name):
275         * Scripts/webkitpy/common/system/outputcapture.py:
276         (OutputCapture.assert_outputs):
277         * Scripts/webkitpy/layout_tests/port/apple.py:
278         (ApplePort.determine_full_port_name):
279         * Scripts/webkitpy/tool/bot/layouttestresultsreader.py:
280         (LayoutTestResultsReader.__init__):
281         (LayoutTestResultsReader):
282         (LayoutTestResultsReader._read_file_contents):
283         (LayoutTestResultsReader._create_layout_test_results):
284         (LayoutTestResultsReader._create_unit_test_results):
285         (LayoutTestResultsReader.results):
286         (LayoutTestResultsReader.archive):
287         * Scripts/webkitpy/tool/bot/layouttestresultsreader_unittest.py:
288         (LayoutTestResultsReaderTest.test_missing_layout_test_results):
289         (LayoutTestResultsReaderTest.test_create_unit_test_results):
290         (test_missing_unit_test_results_path):
291         (test_layout_test_results):
292         (test_archive_last_layout_test_results):
293         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
294         (AbstractEarlyWarningSystem):
295         (AbstractEarlyWarningSystem.__init__):
296         (AbstractEarlyWarningSystem.begin_work_queue):
297         (AbstractEarlyWarningSystem.run_command):
298         * Scripts/webkitpy/tool/commands/queues.py:
299         (AbstractPatchQueue.work_item_log_path):
300         (PatchProcessingQueue):
301         (PatchProcessingQueue._new_port_name_from_old):
302         (PatchProcessingQueue.begin_work_queue):
303         (PatchProcessingQueue._upload_results_archive_for_patch):
304         (CommitQueue):
305         (CommitQueue.begin_work_queue):
306         (CommitQueue.run_command):
307         (AbstractReviewQueue):
308         (AbstractReviewQueue.__init__):
309         (AbstractReviewQueue.begin_work_queue):
310         * Scripts/webkitpy/tool/commands/queues_unittest.py:
311         (AbstractPatchQueueTest.test_next_patch):
312         (PatchProcessingQueueTest):
313         (PatchProcessingQueueTest.test_upload_results_archive_for_patch):
314         * Scripts/webkitpy/tool/steps/runtests.py:
315         (RunTests.run):
316         * Scripts/webkitpy/tool/steps/runtests_unittest.py:
317
318 2013-01-04  Dana Jansens  <danakj@chromium.org>
319
320         [chromium] Remove lost-compositor-context layout tests
321         https://bugs.webkit.org/show_bug.cgi?id=106089
322
323         Reviewed by James Robinson.
324
325         Remove the loseCompositorContext() test function.
326
327         * DumpRenderTree/chromium/TestRunner/src/TestRunner.cpp:
328         (WebTestRunner::TestRunner::TestRunner):
329         * DumpRenderTree/chromium/TestRunner/src/TestRunner.h:
330         (TestRunner):
331
332 2013-01-04  Adam Klein  <adamk@chromium.org>
333
334         Remove ENABLE_MUTATION_OBSERVERS #define
335         https://bugs.webkit.org/show_bug.cgi?id=105459
336
337         Reviewed by Ryosuke Niwa.
338
339         * Scripts/webkitperl/FeatureList.pm:
340         * qmake/mkspecs/features/features.pri:
341
342 2013-01-04  Ryosuke Niwa  <rniwa@webkit.org>
343
344         Statistics used in perftest_unittest.py and perftest_integrationtest.py are bogus
345         https://bugs.webkit.org/show_bug.cgi?id=106078
346
347         Reviewed by Tony Chang.
348
349         This patch updates values for individual iterations in dummy performance tests used in
350         perftest_unittest.py and perftest_integrationtest.py so that they match up with
351         statistics such as stdev and median reported in those tests. This change is required
352         to resolve the bug 97510.
353
354         This patch also refactors integration tests by extracting *TestData classes that
355         centralizes the definitions of each dummy performance test's text (DRT output),
356         runner output, and parsed results (JSON) so that we may share more data between
357         python tests.
358
359         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
360         (MainTest._assert_results_are_correct): Extracted from test_parse_output.
361         Use self.assertAlmostEqual when comparing the standard deviation.
362         (MainTest.test_parse_output):
363         (test_parse_output_with_failing_line):
364         (test_parse_output_with_description): Use _assert_results_are_correct.
365         (test_parse_output_with_subtests): Ditto.
366
367         * Scripts/webkitpy/performance_tests/perftestsrunner_integrationtest.py:
368         (InspectorPassTestData): Extracted from tests below.
369         (EventTargetWrapperTestData): Ditto.
370         (SomeParserTestData): Ditto.
371         (MemoryTestData): Ditto.
372         (TestDriver.run_test): Moved test definitions out of the method.
373         (MainTest._normalize_output): Truncate standard deviations at 5th decimal point.
374         (MainTest._load_output_json): Extracted from create_runner.
375         (MainTest.create_runner):
376         (MainTest.test_run_test_set_for_parser_tests):
377         (MainTest.test_run_memory_test):
378         (MainTest._test_run_with_json_output):
379         (MainTest.test_run_with_json_output):
380         (MainTest.test_run_with_description):
381         (MainTest.test_run_generates_json_by_default):
382         (MainTest.test_run_merges_output_by_default):
383         (MainTest.test_run_respects_reset_results):
384         (MainTest.test_run_generates_and_show_results_page):
385         (MainTest.test_run_with_slave_config_json):
386         (MainTest.test_run_with_multiple_repositories):
387
388 2013-01-04  Martin Robinson  <mrobinson@igalia.com>
389
390         [GTK] [WebKit2] Clicking on a word that is selected in an editable area deletes it
391         https://bugs.webkit.org/show_bug.cgi?id=106045
392
393         Reviewed by Carlos Garcia Campos.
394
395         Update test results to confirm that compositions without preedits do not trigger
396         a composition confirmation.
397
398         * TestWebKitAPI/Tests/gtk/InputMethodFilter.cpp:
399         (TestWebKitAPI::verifyCanceledComposition):
400
401 2013-01-04  Sailesh Agrawal  <sail@chromium.org>
402
403         [Mac] [WK2] platform/mac/fast/events/numpad-keycode-mapping.html fails
404         https://bugs.webkit.org/show_bug.cgi?id=105958
405
406         Reviewed by Alexey Proskuryakov.
407
408         Ported code from DumpRenderTree to WebKitTestRunner to map all numpad keys.
409
410         * WebKitTestRunner/mac/EventSenderProxy.mm:
411         (KeyMappingEntry):
412         (WTR):
413         (WTR::EventSenderProxy::keyDown): Added all numpad keys.
414
415 2013-01-03  Chris Fleizach  <cfleizach@apple.com>
416
417         REGRESSION: [Mac] Intermittent crash in WTR::AccessibilityUIElement::isEqual
418         https://bugs.webkit.org/show_bug.cgi?id=106073
419
420         Reviewed by Ryosuke Niwa.
421
422         Check that the element being compared to is not nil. 
423         This crash is happening because one of the accessibility tests does an isEqual comparison within a callback function.
424         It looks like occasionally, one of the elements passed in to the isEqual is nil because the notification handler is on the accessibility
425         controller, and the notification is being delivered after the document objects are cleared. This shouldn't happen but I've seen it 
426         sometimes before on parallel WebKitTestRunner runs.
427
428         * DumpRenderTree/AccessibilityUIElement.cpp:
429         (AccessibilityUIElement::isEqual):
430         * WebKitTestRunner/InjectedBundle/mac/AccessibilityUIElementMac.mm:
431         (WTR::AccessibilityUIElement::isEqual):
432
433 2013-01-03  Tim 'mithro' Ansell  <mithro@mithis.com>
434
435         Downloads can be flaky, so try downloading multiple times.
436         https://bugs.webkit.org/show_bug.cgi?id=105976
437
438         Reviewed by Eric Seidel.
439
440         * Scripts/webkitpy/common/system/autoinstall.py:
441         (AutoInstaller._download_to_stream):
442
443 2013-01-03  Eric Seidel  <eric@webkit.org>
444
445         Rename Tool.port to deprecated_port in preparation for removing more callers
446         https://bugs.webkit.org/show_bug.cgi?id=106076
447
448         Reviewed by Ryosuke Niwa.
449
450         This resolves a long-standing FIXME to rename this method. :)
451
452         * Scripts/webkitpy/tool/bot/botinfo.py:
453         (BotInfo.summary_text):
454         * Scripts/webkitpy/tool/bot/irc_command.py:
455         (Rollout._update_working_copy):
456         * Scripts/webkitpy/tool/bot/layouttestresultsreader.py:
457         (LayoutTestResultsReader._create_layout_test_results):
458         (LayoutTestResultsReader._create_unit_test_results):
459         (LayoutTestResultsReader._results_directory):
460         * Scripts/webkitpy/tool/bot/layouttestresultsreader_unittest.py:
461         (test_missing_unit_test_results_path):
462         * Scripts/webkitpy/tool/main.py:
463         (WebKitPatch.deprecated_port):
464         * Scripts/webkitpy/tool/mocktool.py:
465         (MockTool.deprecated_port):
466         * Scripts/webkitpy/tool/steps/build.py:
467         (Build.build):
468         * Scripts/webkitpy/tool/steps/checkstyle.py:
469         (CheckStyle.run):
470         * Scripts/webkitpy/tool/steps/commit.py:
471         (Commit._check_test_expectations):
472         * Scripts/webkitpy/tool/steps/preparechangelog.py:
473         (PrepareChangeLog.run):
474         * Scripts/webkitpy/tool/steps/preparechangelogfordepsroll.py:
475         (PrepareChangeLogForDEPSRoll.run):
476         * Scripts/webkitpy/tool/steps/preparechangelogforrevert.py:
477         (PrepareChangeLogForRevert.run):
478         * Scripts/webkitpy/tool/steps/runtests.py:
479         (RunTests.run):
480         * Scripts/webkitpy/tool/steps/steps_unittest.py:
481         (StepsTest.test_runtests_args):
482         * Scripts/webkitpy/tool/steps/update.py:
483         (Update._update_command):
484
485 2013-01-03  Eric Seidel  <eric@webkit.org>
486
487         Add myself to the webkitpy watchlist.
488
489         Unreviewed.
490
491         * Scripts/webkitpy/common/config/watchlist:
492
493 2013-01-03  Ryosuke Niwa  <rniwa@webkit.org>
494
495         Cleanup perftest* tests and add a test for computing statistics
496         https://bugs.webkit.org/show_bug.cgi?id=105685
497
498         Reviewed by Eric Seidel.
499
500         Add a test for PerfTest.compute_statistics (moved and renamed from PageLoadingPerfTest.calculate_statistics) and
501         extracted perftestsrunner_itegrationtests.py from perftestsrunner_unittests.py.
502
503         Also fixed a bug in compute_statistics that the mean ('avg') value can have a large rounding errors in some cases.
504
505         * Scripts/webkitpy/performance_tests/perftest.py:
506         (PerfTest.compute_statistics): Moved from PageLoadingPerfTest to prepare for the bug 97510. Also compute the mean
507         directly from sorted_values instead of using the one from Knuth's online algorithm. This approach gives more
508         accurate result for the mean.
509         (PageLoadingPerfTest.run_single):
510         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
511         (MainTest.test_compute_statistics):
512         (MainTest.test_compute_statistics.compute_statistics): Added.
513         (TestPageLoadingPerfTest.test_run): floatify values.
514         (TestPageLoadingPerfTest.test_run_with_memory_output): Ditto. Also got rid of ".0" from mean values now that Python
515         correctly recognizes them as integers.
516
517         * Scripts/webkitpy/performance_tests/perftestsrunner_integrationtest.py: Copied from
518         Tools/Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py.
519         (TestDriver): Moved out of MainTest.
520         (MainTest): Got rid of assertWritten and all unit tests.
521         (MainTest._normalize_output): Renamed from normalizeFinishedTime to match the PEP8 naming convention.
522         (MainTest.test_run_test_set_kills_drt_per_run.TestDriverWithStopCount):
523         (MainTest.test_run_test_set_for_parser_tests):
524         (MainTest.test_run_memory_test):
525         (MainTest._test_run_with_json_output):
526         (MainTest.test_run_generates_json_by_default):
527         (MainTest.test_run_merges_output_by_default):
528         (MainTest.test_run_respects_reset_results):
529         (MainTest.test_run_generates_and_show_results_page): Use runner.load_output_json() instead of manually loading and
530         parsing output JSON files. Just verify that the output path is correct instead.
531         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
532         (MainTest): Removed all integration tests.
533         (MainTest.create_runner): Simplified.
534
535 2013-01-03  Julie Parent  <jparent@chromium.org>
536
537         Dashboard cleanup: remove usage of global g_defaultBuilderName
538         https://bugs.webkit.org/show_bug.cgi?id=106043
539
540         Use now public defaultBuilder() from the BuilderGroup, rather
541         than global variable.
542
543         Also eliminates the setup() and initBuilders() functions since we
544         no longer need to track the global.
545
546         Reviewed by Adam Barth.
547
548         * TestResultServer/static-dashboards/builders.js:
549         (BuilderGroup.prototype.defaultBuilder):
550         (BuilderGroup.prototype.master):
551         * TestResultServer/static-dashboards/dashboard_base.js:
552         (parseParameters):
553         * TestResultServer/static-dashboards/flakiness_dashboard.js:
554         (updateDefaultBuilderState):
555         * TestResultServer/static-dashboards/loader.js:
556         (.):
557         * TestResultServer/static-dashboards/loader_unittests.js:
558
559 2013-01-03  Ryosuke Niwa  <rniwa@webkit.org>
560
561         Disable an intermittently failing unit test for now.
562         The failure is tracked by the bug 106051.
563
564         * TestWebKitAPI/Tests/mac/MemoryCachePruneWithinResourceLoadDelegate.mm:
565         (TestWebKitAPI::TEST):
566
567 2013-01-03  Dan Bernstein  <mitz@apple.com>
568
569         No way to obtain a DOMNode given a JS wrapper for a Node
570         https://bugs.webkit.org/show_bug.cgi?id=106033
571
572         Added a test for +[DOMNode _nodeFromJSWrapper:].
573
574         Reviewed by Sam Weinig.
575
576         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
577         * TestWebKitAPI/Tests/mac/DOMNodeFromJSObject.mm: Added.
578         (TestWebKitAPI):
579         (TestWebKitAPI::TEST):
580
581 2013-01-03  Emil A Eklund  <eae@chromium.org>
582
583         Fix overflow in LayoutUnit::ceil and floor for SATURATED_LAYOUT_ARITHMETIC
584         https://bugs.webkit.org/show_bug.cgi?id=105961
585
586         Reviewed by Levi Weintraub.
587         
588         Add tests for LayoutUnit::ceil and floor.
589
590         * TestWebKitAPI/Tests/WebCore/LayoutUnit.cpp:
591         (TestWebKitAPI::TEST):
592         (TestWebKitAPI):
593
594 2013-01-03  Julie Parent  <jparent@chromium.org>
595
596         Add unit test for default builder for the dashboards.
597         https://bugs.webkit.org/show_bug.cgi?id=105972
598
599         Reviewed by Dirk Pranke.
600
601         * TestResultServer/static-dashboards/loader_unittests.js:
602
603 2013-01-03  Tony Chang  <tony@chromium.org>
604
605         Generate internal.settings from Settings.in
606         https://bugs.webkit.org/show_bug.cgi?id=104740
607
608         Reviewed by Adam Barth.
609
610         * GNUmakefile.am: Compile new generated files on GTK+.
611         * qmake/mkspecs/features/default_post.prf: Qt: Make it so an action
612         with both extra_sources and add_output_to_sources=false will still
613         compile the files in extra_sources. This makes it possible to compile
614         InternalSettingsGenerated.cpp and not be confused by
615         InternalSettingsGenerated.idl.
616
617 2013-01-03  Zan Dobersek  <zandobersek@gmail.com>
618
619         [style] WebIDL-reflecting uppercase enums reported as style violation
620         https://bugs.webkit.org/show_bug.cgi?id=105832
621
622         Reviewed by Tony Chang.
623
624         Recognize the enum as WebIDL-reflecting one if there's a comment present that
625         indicates this. In enums spanning over multiple lines, the comment is expected
626         on the first line. In enums placed in one line, the comment is expected at the
627         end of that line.
628
629         WebIDL-reflecting enum values are required to be written in uppercase and as such
630         violate the common rule of these values being written in InterCaps. Annotating
631         enums this way makes it possible to avoid false style violation reports.
632
633         The comment is of form '// Web(?:Kit)?IDL enum'.
634
635         * Scripts/webkitpy/style/checkers/cpp.py:
636         (_EnumState.__init__):
637         (_EnumState.process_clean_line):
638         (check_enum_casing):
639         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
640         (NoNonVirtualDestructorsTest.test_enum_casing):
641
642 2013-01-03  Tim 'mithro' Ansell  <mithro@mithis.com>
643
644         Adding favicon.ico to the garden-o-matic.
645         https://bugs.webkit.org/show_bug.cgi?id=105982
646
647         Reviewed by Adam Barth.
648
649         * Scripts/webkitpy/tool/servers/gardeningserver.py:
650         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/favicon.ico:
651
652 2013-01-03  Jonathan Liu  <net147@gmail.com>
653
654         [Qt] Fix library linking order
655         https://bugs.webkit.org/show_bug.cgi?id=105818
656
657         Some Qt libraries are linked before the private static libraries
658         that depend on them which results in QtWebKit failing to link.
659
660         Copy Qt module dependencies from QT to QT_PRIVATE when creating
661         module to ensure the Qt libraries are linked after the private
662         static libraries that depend on them.
663
664         Reviewed by Tor Arne Vestbø.
665
666         * qmake/mkspecs/features/default_post.prf:
667
668 2013-01-02  Byungwoo Lee  <bw80.lee@samsung.com>
669
670         [EFL][WK2] Fix tooltip bugs in MiniBrowser.
671         https://bugs.webkit.org/show_bug.cgi?id=102209
672
673         Reviewed by Gyuyoung Kim.
674
675         Set tooltip on the elm_window instead of ewk_view, and add delay to
676         show the tooltip window.
677
678         Setting tooltip on the ewk_view makes elementary error message because
679         ewk_view is not an elementary widget. To remove the error message,
680         tooltip should be set on the elm_window.
681
682         Most browsers show the tooltip after a mouse pointer stays on an
683         element object that has tooltip text.
684         But elm_tooltip widget itself doesn't have delay or handle mouse
685         events to show/hide.
686
687         So additional codes are added to show/hide tooltip widget along with
688         mouse events.
689           1) When a mouse pointer stays for a second on an element which has
690              tooltip text, then tooltip widget is displayed.
691           2) When a mouse pointer is moved or out of ewk_view, hide the tooltip
692              widget.
693
694         * MiniBrowser/efl/main.c:
695         (_Tooltip_Information):
696         (_Browser_Window):
697         (on_tooltip_show):
698         (window_tooltip_hide):
699         (window_tooltip_update):
700         (on_mouse_in):
701         (on_mouse_move):
702         (on_mouse_out):
703         (window_free):
704         (on_tooltip_text_set):
705         (on_tooltip_text_unset):
706         (window_create):
707
708 2013-01-02  Tony Chang  <tony@chromium.org>
709
710         Unreviewed, rolling out r138661.
711         http://trac.webkit.org/changeset/138661
712         https://bugs.webkit.org/show_bug.cgi?id=104740
713
714         Compile problems on EFL
715
716         * GNUmakefile.am:
717         * qmake/mkspecs/features/default_post.prf:
718
719 2013-01-02  Tony Chang  <tony@chromium.org>
720
721         Generate internal.settings from Settings.in
722         https://bugs.webkit.org/show_bug.cgi?id=104740
723
724         Reviewed by Adam Barth.
725
726         * GNUmakefile.am: Compile new generated files on GTK+.
727         * qmake/mkspecs/features/default_post.prf: Qt: Make it so an action
728         with both extra_sources and add_output_to_sources=false will still
729         compile the files in extra_sources. This makes it possible to compile
730         InternalSettingsGenerated.cpp and not be confused by
731         InternalSettingsGenerated.idl.
732
733 2013-01-02  Adam Barth  <abarth@webkit.org>
734
735         Remove old-run-webkit-tests support from webkitpy
736         https://bugs.webkit.org/show_bug.cgi?id=105619
737
738         Reviewed by Eric Seidel.
739
740         This patch removes support for webkitpy parsing the
741         old-run-webkit-tests results format. There doesn't seem to be any
742         reason to keep this code around anymore now that most bots have
743         switched to new-run-webkit-tests.
744
745         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
746         (Builder.fetch_layout_test_results):
747         * Scripts/webkitpy/common/net/layouttestresults.py:
748         (path_for_layout_test):
749         (LayoutTestResults.results_from_string):
750         (LayoutTestResults):
751         * Scripts/webkitpy/common/net/layouttestresults_unittest.py:
752         (LayoutTestResultsTest.test_results_from_string):
753         * Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
754         (test_flaky_test_failure):
755         * Scripts/webkitpy/tool/bot/layouttestresultsreader_unittest.py:
756         (LayoutTestResultsReaderTest.test_missing_layout_test_results):
757         * Scripts/webkitpy/tool/bot/patchanalysistask.py:
758         (PatchAnalysisTask._test_patch):
759         * Scripts/webkitpy/tool/commands/queuestest.py:
760         (QueuesTest.assert_queue_outputs):
761
762 2013-01-02  Heikki Paajanen  <heikki.paajanen@palm.com>
763
764         [Qt][WK2] Add simple UI to find text from page
765         https://bugs.webkit.org/show_bug.cgi?id=96481
766
767         Reviewed by Jocelyn Turcotte.
768
769         Find bar that slides between navigation bar and WebView.
770
771         * MiniBrowser/qt/BrowserWindow.cpp:
772         (BrowserWindow::toggleFind):
773         * MiniBrowser/qt/BrowserWindow.h:
774         (BrowserWindow):
775         * MiniBrowser/qt/MiniBrowser.qrc:
776         * MiniBrowser/qt/MiniBrowserApplication.cpp:
777         (MiniBrowserApplication::notify):
778         * MiniBrowser/qt/icons/find.png: Added.
779         * MiniBrowser/qt/qml/BrowserWindow.qml:
780
781 2013-01-02  Christophe Dumez  <christophe.dumez@intel.com>
782
783         Regression(r138603): Caused 2 webkitpy tests to fail
784         https://bugs.webkit.org/show_bug.cgi?id=105913
785
786         Reviewed by Kentaro Hara.
787
788         Fix 2 webkitpy tests that started to fail after
789         r138603.
790
791         * Scripts/webkitpy/tool/bot/irc_command_unittest.py:
792         (IRCCommandTest._sheriff_test_data_url): Make sure the
793         returned test file URL is valid no matter what the
794         current working directory is.
795         * Scripts/webkitpy/tool/bot/ircbot_unittest.py:
796         (IRCBotTest.test_help): Add "sheriffs" command to the
797         list of available commands.
798
799 2013-01-01  Christophe Dumez  <christophe.dumez@intel.com>
800
801         [WK2] plugins/clicking-missing-plugin-fires-delegate.html is failing
802         https://bugs.webkit.org/show_bug.cgi?id=105897
803
804         Reviewed by Alexey Proskuryakov.
805
806         Dump in WebKitTestRunner the message that is expected by layout tests
807         when the missing plugin button is clicked.
808
809         * WebKitTestRunner/TestController.cpp:
810         (WTR::TestController::createWebViewWithOptions):
811         (WTR::TestController::unavailablePluginButtonClicked):
812         (WTR):
813         * WebKitTestRunner/TestController.h:
814         (TestController):
815
816 2013-01-01  KwangYong Choi  <ky0.choi@samsung.com>
817
818         [EFL] Enable MHTML feature
819         https://bugs.webkit.org/show_bug.cgi?id=105815
820
821         Reviewed by Laszlo Gombos.
822
823         Enable MHTML feature for EFL port by default.
824
825         * Scripts/webkitperl/FeatureList.pm:
826
827 2013-01-01  Alan Cutter  <alancutter@chromium.org>
828
829         sheriff-bot should know who the gardeners/sheriffs are
830         https://bugs.webkit.org/show_bug.cgi?id=105698
831
832         Reviewed by Eric Seidel.
833
834         Added a "sheriffs" command to sheriffbot to fetch the current Chromium Webkit sheriffs.
835
836         * Scripts/webkitpy/tool/bot/irc_command.py:
837         (Current):
838         (Current._retrieve_webkit_sheriffs):
839         (Current.execute):
840         * Scripts/webkitpy/tool/bot/irc_command_unittest.py:
841         (IRCCommandTest.test_current):
842         * Scripts/webkitpy/tool/bot/testdata/webkit_sheriff_0.js: Added.
843         * Scripts/webkitpy/tool/bot/testdata/webkit_sheriff_1.js: Added.
844         * Scripts/webkitpy/tool/bot/testdata/webkit_sheriff_2.js: Added.
845         * Scripts/webkitpy/tool/bot/testdata/webkit_sheriff_malformed.js: Added.
846         * Scripts/webkitpy/tool/bot/testdata/webkit_sheriff_zero.js: Added.
847
848 2012-12-31  Carlos Garcia Campos  <cgarcia@igalia.com>
849
850         [GTK] Move ImageDiff to the Tools directory root
851         https://bugs.webkit.org/show_bug.cgi?id=105421
852
853         Reviewed by Kenneth Rohde Christiansen.
854
855         * DumpRenderTree/gtk/GNUmakefile.ImageDiff.am: Removed.
856         * GNUmakefile.am:
857         * ImageDiff/gtk/ImageDiff.cpp: Renamed from Tools/DumpRenderTree/gtk/ImageDiff.cpp.
858
859 2012-12-30  Mike West  <mkwst@chromium.org>
860
861         Add myself to watchlists for CSP and console logs.
862         https://bugs.webkit.org/show_bug.cgi?id=105874
863
864         Reviewed by Dirk Pranke.
865
866         Because I really, really, really need more email.
867
868         * Scripts/webkitpy/common/config/committers.py:
869             Add mkwst+watchlist@chromium.org as a contributor.
870         * Scripts/webkitpy/common/config/watchlist:
871             Create watchlist categories for 1) the various mechanisms used to
872             generate console messages, and 2) CSP usage and files.
873
874 2012-12-30  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
875
876         [EFL] Enable TEMPLATE_ELEMENT feature
877         https://bugs.webkit.org/show_bug.cgi?id=105865
878         
879         Reviewed by Laszlo Gombos.
880
881         * Scripts/webkitperl/FeatureList.pm: Enable template element feature for EFL port.
882
883 2012-12-30  Zan Dobersek  <zandobersek@gmail.com>
884
885         Unreviewed GTK gardening of the unit tests.
886
887         Skip the TestInspectorServer unit test that's timing out completely.
888
889         * Scripts/run-gtk-tests:
890         (TestRunner):
891
892 2012-12-29  Ilya Tikhonovsky  <loislo@chromium.org>
893
894         Web Inspector: Native Memory Instrumentation plugin: do not generate 'not instrumented' warning for instrumented mutable members.
895         https://bugs.webkit.org/show_bug.cgi?id=105855
896
897         Reviewed by Vsevolod Vlasov.
898
899         Extract MemberExpr from ImplicitCastExpr. It happens when we instrument a mutable member because
900         addMember expects const T& and the mutable ref to member implicitly converts into const ref to member.
901
902         * clang/ReportMemoryUsagePlugin/ReportMemoryUsage.cpp:
903         (ReportMemoryUsageVisitor):
904         (clang::AddMemberCallVisitor::VisitCallExpr):
905         (clang::ReportMemoryUsageVisitor::VisitCXXMethodDecl):
906         (clang::ReportMemoryUsageVisitor::checkMembersCoverage):
907
908 2012-12-29  Ilya Tikhonovsky  <loislo@chromium.org>
909
910         Web Inspector: Native Memory Instrumentation plugin: move function bodies out of class declarations.
911         https://bugs.webkit.org/show_bug.cgi?id=105852
912
913         Reviewed by Alexander Pavlov.
914
915         * clang/ReportMemoryUsagePlugin/ReportMemoryUsage.cpp:
916         (clang):
917         (AddMemberCallVisitor):
918         (ReportMemoryUsageVisitor):
919         (clang::ReportMemoryUsageAction::ParseArgs):
920         (clang::AddMemberCallVisitor::VisitCallExpr):
921         (clang::ReportMemoryUsageVisitor::VisitCXXMethodDecl):
922         (clang::ReportMemoryUsageVisitor::emitWarning):
923         (clang::ReportMemoryUsageVisitor::findInstrumentationMethod):
924         (clang::ReportMemoryUsageVisitor::needsToBeInstrumented):
925         (clang::ReportMemoryUsageVisitor::CheckMembersCoverage):
926
927 2012-12-29  Zan Dobersek  <zandobersek@gmail.com>
928
929         [webkitpy] Omit webkitpy/thirdparty/BeautifulSoup.py from code coverage checking
930         https://bugs.webkit.org/show_bug.cgi?id=105835
931
932         Reviewed by Dirk Pranke.
933
934         The BeautifulSoup module is a third-party import and as such we don't have much control
935         over its testing. Because of that it should be omitted from code coverage checking.
936
937         * Scripts/webkitpy/test/main.py:
938         (Tester._run_tests):
939
940 2012-12-28  Martin Robinson  <mrobinson@igalia.com>
941
942         [GTK][WK2] Add support for IME Composition
943         https://bugs.webkit.org/show_bug.cgi?id=65093
944
945         Reviewed by Carlos Garcia Campos.
946
947         Update unit tests for GtkInputMethodFilter to note whether or not an event
948         was a fake event synthesized for composition purposes. This is now used
949         for WebKit2.
950
951         * TestWebKitAPI/Tests/gtk/InputMethodFilter.cpp:
952         (TestWebKitAPI::TestInputMethodFilter::sendSimpleKeyEvent):
953         (TestWebKitAPI::TestInputMethodFilter::sendKeyEventWithCompositionResults):
954         (TestWebKitAPI::TEST):
955         * gtk/generate-gtkdoc:
956         (get_webkit2_options): Skip new files in the WebKit2 source directory for documentation
957         generation.
958
959 2012-12-27  Ilya Tikhonovsky  <loislo@chromium.org>
960
961         Web Inspector: Native Memory Instrumentation. Update clang plugin according to the current state of memory instrumentation code.
962         https://bugs.webkit.org/show_bug.cgi?id=105800
963
964         Reviewed by Yury Semikhatsky.
965
966         * clang/ReportMemoryUsagePlugin/ReportMemoryUsage.cpp:
967         (clang::ReportMemoryUsageConsumer::ReportMemoryUsageConsumer):
968         (clang):
969
970 2012-12-27  Zan Dobersek  <zandobersek@gmail.com>
971
972         Create a GTK build system watchlist and add myself to it
973         https://bugs.webkit.org/show_bug.cgi?id=105769
974
975         Reviewed by Dirk Pranke.
976
977         Add a GTK build system watchlist, watching the configure.ac, GNUmakefile.am files
978         and the input GNUmakefile.features.am.in file. The GNUmakefile.list.am files are
979         omitted on purpose due to relatively high change rate they are exposed to.
980
981         The new watchlist would be of use to me as the GTK build system recently went
982         through a series of changes that included a changed policy on adding configuration
983         flags to configure.ac and subsequent additions to GNUmakefile.am files. A PSA on the
984         mailing list about this change will be published later, but I'd also like to be
985         notified of any pending changes to these files just so I can check the changes are
986         in compliance with the new policy.
987
988         * Scripts/webkitpy/common/config/watchlist:
989
990 2012-12-26  Ilya Tikhonovsky  <loislo@chromium.org>
991
992         Unreviewed. Replace find("Tools") with rfind("Tools") because base dir of WebKit may have Tools word.
993         As example ..../DevTools/src/third_party/WebKit
994
995         * Scripts/webkitpy/common/webkit_finder.py:
996         (WebKitFinder.webkit_base):
997
998 2012-12-25  Christophe Dumez  <christophe.dumez@intel.com>
999
1000         [EFL][WK2] Refactor snapshot taking code
1001         https://bugs.webkit.org/show_bug.cgi?id=105687
1002
1003         Reviewed by Kenneth Rohde Christiansen.
1004
1005         Call WKViewCreateSnapshot() instead of WKViewGetSnapshot()
1006         since the function was renamed to follow naming style.
1007
1008         * WebKitTestRunner/efl/PlatformWebViewEfl.cpp:
1009         (WTR::PlatformWebView::windowSnapshotImage):
1010
1011 2012-12-24  Ilya Tikhonovsky  <loislo@chromium.org>
1012
1013         Web Inspector: Native Memory Instrumentation: propagate member type as edge type to the serialized heap graph.
1014         https://bugs.webkit.org/show_bug.cgi?id=105725
1015
1016         Reviewed by Yury Semikhatsky.
1017
1018         MemberType value names were adjusted according to Style Guide.
1019         Existing tests were extended with link type validation.
1020
1021         * TestWebKitAPI/Tests/WTF/MemoryInstrumentationTest.cpp:
1022
1023 2012-12-24  Xiaobo Wang  <xbwang@torchmobile.com.cn>
1024
1025         [BlackBerry] DRT - Crashed when running pixels tests
1026         https://bugs.webkit.org/show_bug.cgi?id=105711
1027
1028         RIM PR 270003
1029
1030         Reviewed by George Staikos.
1031
1032         Calling window->post() here will result in buffer swapping which is
1033         only allowed in UI thread now, and I found we don't need to call it
1034         for GL renderer.
1035
1036         * DumpRenderTree/blackberry/PixelDumpSupportBlackBerry.cpp:
1037         (createBitmapContextFromWebView):
1038
1039 2012-12-22  Dan Winship  <danw@gnome.org>
1040
1041         Add a patch to fix libxml2 2.9.0 regression
1042         https://bugs.webkit.org/show_bug.cgi?id=104494
1043
1044         Reviewed by Martin Robinson.
1045
1046         * gtk/jhbuild.modules:
1047         * gtk/patches/libxml2-2.9.0-dtd.patch: patch from upstream
1048
1049 2012-12-23  Allan Sandfeld Jensen  <allan.jensen@digia.com>
1050
1051         Unreviewed update of status to reviewer.
1052
1053         * Scripts/webkitpy/common/config/committers.py:
1054
1055 2012-12-23  Philippe Normand  <pnormand@igalia.com>
1056
1057         [GTK][jhbuild] Bump GStreamer to 1.0.4
1058         https://bugs.webkit.org/show_bug.cgi?id=105640
1059
1060         Reviewed by Martin Robinson.
1061
1062         * gtk/jhbuild.modules: Update GStreamer modules to 1.0.4.
1063
1064 2012-12-21  Nico Weber  <thakis@chromium.org>
1065
1066         chromium: webkit-build-directory doesn't handle ninja output dirs
1067         https://bugs.webkit.org/show_bug.cgi?id=105671
1068
1069         Reviewed by Daniel Bates.
1070
1071         Teach webkit-build-directory to detect the ninja build directory.
1072
1073         * Scripts/webkitdirs.pm:
1074         (determineBaseProductDir):
1075         (determineIsChromiumNinja):
1076
1077 2012-12-21  Ryosuke Niwa  <rniwa@webkit.org>
1078
1079         [Mountain Lion] platform/mac/editing/spelling/removing-underline-after-accepting-autocorrection-using-punctuation.html failing
1080         https://bugs.webkit.org/show_bug.cgi?id=95477
1081
1082         Reviewed by Sam Weinig.
1083
1084         Co-authored by Jia Pu.
1085
1086         Specify a fixed autocorrection dictionary in resetDefaultsToConsistentValues() so that
1087         we can have consistent autocorrection behavior when running tests.
1088
1089         * DumpRenderTree/mac/DumpRenderTree.mm:
1090         (resetDefaultsToConsistentValues):
1091         * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm:
1092         (WTR::InjectedBundle::platformInitialize):
1093
1094 2012-12-21  Zan Dobersek  <zandobersek@gmail.com>
1095
1096         [GTK] Add WebKit1-specific TestExpectations file
1097         https://bugs.webkit.org/show_bug.cgi?id=97562
1098
1099         Reviewed by Dirk Pranke.
1100
1101         Implement the expectations_files method inside the GtkPort class that
1102         returns the list of the TestExpectations files the current test configuration
1103         requires. The method is implemented in a similar way as on the EFL port.
1104         A unit test is also added.
1105
1106         * Scripts/webkitpy/layout_tests/port/gtk.py:
1107         (GtkPort._search_paths):
1108         (GtkPort):
1109         (GtkPort.expectations_files):
1110         * Scripts/webkitpy/layout_tests/port/gtk_unittest.py:
1111         (GtkPortTest.test_expectations_files):
1112
1113 2012-12-21  Sudarsana Nagineni  <sudarsana.nagineni@intel.com>
1114
1115         [EFL] Enable feature option for the CSS sticky position
1116         https://bugs.webkit.org/show_bug.cgi?id=105630
1117
1118         Reviewed by Kenneth Rohde Christiansen.
1119
1120         Enable CSS sticky position feature by default for the
1121         EFL port. This feature option was added in r138252.
1122
1123         No new tests. Existing layout tests pass.
1124
1125         * Scripts/webkitperl/FeatureList.pm:
1126
1127 2012-12-21  Jussi Kukkonen  <jussi.kukkonen@intel.com>
1128
1129         [WK2] REGRESSION(r138232): It made 4 tests timeout
1130         https://bugs.webkit.org/show_bug.cgi?id=105623
1131
1132         Reviewed by Kenneth Rohde Christiansen.
1133
1134         r138232 changed functionality so that empty text output
1135         no longer leads to a dump() call, which leads to a timeout.
1136         Instead do a text dump if there is text output or if there
1137         is no audio output.
1138
1139         * WebKitTestRunner/TestInvocation.cpp:
1140         (WTR::TestInvocation::dumpResults):
1141
1142 2012-12-21  Zeno Albisser  <zeno@webkit.org>
1143
1144         [Qt][Mac] run-launcher cannot find Qt5WebKitWidgets.
1145         https://bugs.webkit.org/show_bug.cgi?id=105309
1146
1147         On Windows the library name is Qt5WebKitWidgets as well.
1148         So r138136 should not have adjusted the library name for Windows.
1149
1150         Reviewed by Simon Hausmann.
1151
1152         * Scripts/webkitdirs.pm:
1153         (builtDylibPathForName):
1154
1155 2012-12-21  Zeno Albisser  <zeno@webkit.org>
1156
1157         [Qt] Add new TestExpectations for qt-5.0-mac-wk2.
1158         https://bugs.webkit.org/show_bug.cgi?id=105424
1159
1160         This change will allow us to run a very specific
1161         subset of the available LayoutTests on a mac buildbot.
1162         The new baseline search path for mac will look as follows:
1163         qt-5.0-mac-wk2 -> qt-5.0-wk2 -> qt-5.0 -> qt-mac -> qt
1164
1165         Reviewed by Dirk Pranke.
1166
1167         * Scripts/webkitpy/layout_tests/port/qt.py:
1168         (QtPort._search_paths):
1169         * Scripts/webkitpy/layout_tests/port/qt_unittest.py:
1170         (QtPortTest):
1171
1172 2012-12-20  Ryuan Choi  <ryuan.choi@samsung.com>
1173
1174         [EFL] Build break with latest EFL libraries.
1175         https://bugs.webkit.org/show_bug.cgi?id=104827
1176
1177         Reviewed by Laszlo Gombos.
1178
1179         The eo EFL package is introduced and evas and ecore use it since 1.8.
1180         While introducing Eo, EFL changed several structures of Evas and Ecore
1181         from own specific class to Eo.
1182
1183         So, this patch adds FindEo and changes declaration of Evas, Evas_Object,
1184         Ecore_Timer to build with latest EFL libraries.
1185
1186         * DumpRenderTree/efl/CMakeLists.txt: Includes eo EFL library.
1187         * EWebLauncher/CMakeLists.txt: Ditto.
1188         * MiniBrowser/efl/CMakeLists.txt: Ditto.
1189         * ImageDiff/PlatformEfl.cmake: Ditto.
1190         * TestWebKitAPI/PlatformEfl.cmake: Ditto.
1191         * TestWebKitAPI/PlatformWebView.h: Modified declaration of Evas_Object.
1192         * WebKitTestRunner/InjectedBundle/TestRunner.h: Modified declaration of Ecore_Timer.
1193         * WebKitTestRunner/PlatformEfl.cmake: Includes eo EFL library.
1194         * WebKitTestRunner/PlatformWebView.h: Modified declaration of Evas_Object.
1195
1196 2012-12-20  Dirk Pranke  <dpranke@chromium.org>
1197
1198         TestExpectation fallback is broken
1199         https://bugs.webkit.org/show_bug.cgi?id=105583
1200
1201         Reviewed by Ryosuke Niwa.
1202
1203         When I added support for the -wk2 in port names, it looks
1204         like I broke the way we computed the list of TestExpectations
1205         files so that the implementation-version directory was left out :(.
1206         Fixed and cleaned up the tests.
1207
1208         * Scripts/webkitpy/layout_tests/port/base.py:
1209         (Port.expectations_files):
1210         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
1211         (TestWebKitPort.__init__):
1212         (PortTestCase.test_skipped_directories_for_symbols):
1213         (test_skipped_directories_for_features):
1214         (test_skipped_directories_for_features_no_matching_tests_in_test_list):
1215         (test_skipped_tests_for_unsupported_features_empty_test_list):
1216         (test_skipped_layout_tests):
1217         (test_expectations_files):
1218
1219 2012-12-20  Anders Carlsson  <andersca@apple.com>
1220
1221         Don't include WebKit2 headers when building TestWebKitAPI on windows
1222         https://bugs.webkit.org/show_bug.cgi?id=105582
1223
1224         Reviewed by Tim Horton.
1225
1226         * TestWebKitAPI/config.h:
1227         * TestWebKitAPI/win/main.cpp:
1228
1229 2012-12-20  Ryosuke Niwa  <rniwa@webkit.org>
1230
1231         REGRESSION(r133820?): SimplifyMarkupTest API test asserts
1232         https://bugs.webkit.org/show_bug.cgi?id=105370
1233
1234         Reviewed by Simon Fraser.
1235
1236         Re-enable the test.
1237
1238         * TestWebKitAPI/Tests/mac/SimplifyMarkup.mm:
1239         (TestWebKitAPI::TEST):
1240
1241 2012-12-20  Nico Weber  <thakis@chromium.org>
1242
1243         chromium nrwt: Pick the newest binary found in DEFAULT_BUILD_DIRECTORIES, not the first
1244         https://bugs.webkit.org/show_bug.cgi?id=105498
1245
1246         Reviewed by Dirk Pranke.
1247
1248         Use the newest binary available rather than an than always picking one
1249         build directory over another based on iteration order.
1250
1251         * Scripts/webkitpy/layout_tests/port/chromium.py:
1252         (ChromiumPort._static_build_path):
1253         Check for timestamps.
1254         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
1255         (ChromiumMacPortTest.test_build_path_timestamps):
1256         Test that out / xcodebuild selection happens based on timestamps
1257         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1258         (ChromiumWinPortTest.test_build_path_timestamps):
1259         Test that out / build selection happens based on timestamps
1260
1261 2012-12-19  Simon Fraser  <simon.fraser@apple.com>
1262
1263         Lots of sticky tests failing in WK2
1264         https://bugs.webkit.org/show_bug.cgi?id=105464
1265
1266         Reviewed by Beth Dakin.
1267
1268         WebKitTestRunner had a race between snapshotting in the UI process,
1269         and resettting after the test in the web process. InjectedBundle::done()
1270         was a bad place to call page()->resetAfterTest(), because of this race;
1271         it could reset the scroll position before the UI snapshot had been obtained.
1272         
1273         Fix by moving the call to page()->resetAfterTest() into didReceiveMessage(),
1274         for the "Reset" message which will come in before the next test.
1275
1276         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
1277         (WTR::InjectedBundle::didReceiveMessage):
1278         (WTR::InjectedBundle::done):
1279
1280 2012-12-19  Filip Pizlo  <fpizlo@apple.com>
1281
1282         DFG speculation checks that take JumpList should consolidate OSRExits
1283         https://bugs.webkit.org/show_bug.cgi?id=105401
1284
1285         Reviewed by Oliver Hunt.
1286
1287         Update the profiler to understand that an OSR exit may have multiple
1288         candidate exit sites.
1289
1290         * Scripts/display-profiler-output:
1291
1292 2012-12-20  Eric Seidel  <eric@webkit.org>
1293
1294         EWSes do not need to dump build/test logs when things pass
1295         https://bugs.webkit.org/show_bug.cgi?id=105402
1296
1297         Reviewed by Dirk Pranke.
1298
1299         When we were first writing the commit-queue and EWS system
1300         and running them locally, we thought it was important to see the
1301         build/test output as it happened.  Now these queues have been running
1302         for years, and we don't use this feature anymore.  So this patch
1303         makes us stop tee-ing the output directly to stdout, and instead
1304         only print any output from commands when they fail.
1305
1306         I also added some logging when we execute each command
1307         since we are no longer printing the command output itself.
1308         This should make it much easier to see errors with the EWS
1309         system itself, now that all the passing-build output is removed.
1310
1311         This required updating a zillion unittests, since we have a bunch of
1312         "integration" style testing of the queue-bots.  Many of the unittests for the bots were
1313         already approximating this behavior by using a "logging" Executive.
1314         I disabled logging in those cases to avoid redundant logging in those tests.
1315
1316         * Scripts/webkitpy/common/system/executive.py:
1317         (Executive.command_for_printing):
1318         (Executive.run_command):
1319         * Scripts/webkitpy/common/system/executive_mock.py:
1320         (MockExecutive.command_for_printing):
1321         (MockExecutive.run_command):
1322         * Scripts/webkitpy/common/system/executive_unittest.py:
1323         (ExecutiveTest.test_auto_stringify_args):
1324         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
1325         (EarlyWarningSytemTest._default_expected_logs):
1326         (_test_ews):
1327         (_test_ewses):
1328         * Scripts/webkitpy/tool/commands/queues.py:
1329         (AbstractQueue.run_webkit_patch):
1330         * Scripts/webkitpy/tool/commands/queues_unittest.py:
1331         (AbstractQueueTest._assert_run_webkit_patch):
1332         (CommitQueueTest.test_commit_queue):
1333         (test_rollout):
1334         (test_rollout_lands):
1335         (test_manual_reject_during_processing):
1336         (StyleQueueTest.test_style_queue_with_style_exception):
1337         (test_style_queue_with_watch_list_exception):
1338
1339 2012-12-20  Dominic Mazzoni  <dmazzoni@google.com>
1340
1341         AX: support clickPoint in DRT for chromium
1342         https://bugs.webkit.org/show_bug.cgi?id=97359
1343
1344         Reviewed by Chris Fleizach.
1345
1346         Implement clickPointX, clickPointY, and elementAtPoint for Chromium, so that
1347         several more tests can pass.
1348
1349         * DumpRenderTree/chromium/TestRunner/src/AccessibilityUIElementChromium.cpp:
1350         (WebTestRunner::AccessibilityUIElement::AccessibilityUIElement):
1351         (WebTestRunner::AccessibilityUIElement::clickPointXGetterCallback):
1352         (WebTestRunner):
1353         (WebTestRunner::AccessibilityUIElement::clickPointYGetterCallback):
1354         (WebTestRunner::AccessibilityUIElement::elementAtPointCallback):
1355         * DumpRenderTree/chromium/TestRunner/src/AccessibilityUIElementChromium.h:
1356         (AccessibilityUIElement):
1357
1358 2012-12-20  Dan Bernstein  <mitz@apple.com>
1359
1360         prepare-ChangeLog adds "Additional information of the change such as approach, rationale" where it doesn't make sense
1361         https://bugs.webkit.org/show_bug.cgi?id=99268
1362
1363         Reviewed by Alexey Proskuryakov.
1364
1365         * Scripts/prepare-ChangeLog:
1366         (generateNewChangeLogs): Removed the "Additional information of the change" line.
1367
1368 2012-12-20  Carlos Garcia Campos  <cgarcia@igalia.com>
1369
1370         Unreviewed. Skip unit test /webkit/atk/getTextInParagraphAndBodyModerate
1371
1372         * Scripts/run-gtk-tests:
1373         (TestRunner):
1374
1375 2012-12-20  Thiago Marcos P. Santos  <thiago.santos@intel.com>
1376
1377         [Qt] Move ImageDiff to the Tools directory root
1378         https://bugs.webkit.org/show_bug.cgi?id=105420
1379
1380         Reviewed by Tor Arne Vestbø.
1381
1382         Moved ImageDiff from outside DumpRenderTree because it is also a
1383         dependency when running pixel tests for WebKit2.
1384
1385         * ImageDiff/ImageDiff.pro: Renamed from Tools/DumpRenderTree/qt/ImageDiff.pro.
1386         * ImageDiff/qt/ImageDiff.cpp: Renamed from Tools/DumpRenderTree/qt/ImageDiff.cpp.
1387         * Tools.pro:
1388
1389 2012-12-20  Zan Dobersek  <zandobersek@gmail.com>
1390
1391         [GTK] Remove the --enable-unstable-features configuration option
1392         https://bugs.webkit.org/show_bug.cgi?id=105327
1393
1394         Reviewed by Martin Robinson.
1395
1396         Add a missing feature option for the CSS sticky position feature.
1397
1398         * Scripts/webkitperl/FeatureList.pm:
1399
1400 2012-12-20  Carlos Garcia Campos  <cgarcia@igalia.com>
1401
1402         Unreviewed. Skip unit test WebKit2.ForceRepaint that times out
1403
1404         * Scripts/run-gtk-tests:
1405         (TestRunner):
1406
1407 2012-12-20  Dominik Röttsches  <dominik.rottsches@intel.com>
1408
1409         [EFL] MiniBrowser does not play Infinite Gangnam Style
1410         https://bugs.webkit.org/show_bug.cgi?id=103531
1411
1412         Reviewed by Kenneth Rohde Christiansen.
1413
1414         For EFL we would like this feature to be built by default.
1415
1416         * Scripts/webkitperl/FeatureList.pm:
1417
1418 2012-12-20  Csaba Osztrogonác  <ossy@webkit.org>
1419
1420         [Qt] EWS bots should call qmake always
1421         https://bugs.webkit.org/show_bug.cgi?id=90049
1422
1423         Reviewed by Simon Hausmann.
1424
1425         * Scripts/webkitdirs.pm:
1426         (buildQMakeProjects):
1427
1428 2012-12-20  Jochen Eisinger  <jochen@chromium.org>
1429
1430         [chromium] fix stray printf in TestRunner's WebTestProxy
1431         https://bugs.webkit.org/show_bug.cgi?id=105515
1432
1433         Reviewed by Ryosuke Niwa.
1434
1435         * DumpRenderTree/chromium/TestRunner/src/WebTestProxy.cpp:
1436         (WebTestRunner::WebTestProxyBase::shouldInsertNode):
1437
1438 2012-12-20  Dominik Röttsches  <dominik.rottsches@intel.com>
1439
1440         [WK2] WebAudio WKTR support
1441         https://bugs.webkit.org/show_bug.cgi?id=95084
1442
1443         Reviewed by Kenneth Rohde Christiansen.
1444
1445         Implementing audio data dump support for WebKitTestRunner,
1446         by adding the API to the IDL, enabling IPC for transferring
1447         the audio data as bytebuffer and dumping it according
1448         to run-webkit-tests script's expectations.
1449         This enables audio testing on WebKit2 ports.
1450         Based on initial wip-patch from Philippe Normand - thanks.
1451
1452         * WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl: Adding setAudioData API.
1453         * WebKitTestRunner/InjectedBundle/TestRunner.cpp: Assign audio buffer data to a local WKDataRef after extracting it from WK2 using WKBundleDataFromUint8Array bundle API.
1454         (WTR::TestRunner::setAudioData):
1455         (WTR):
1456         * WebKitTestRunner/InjectedBundle/TestRunner.h: Add a whatToDump state for dumping Audio.
1457         (TestRunner):
1458         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp: Add audio to the message that's transferred to the WKTR process.
1459         (WTR::InjectedBundle::done):
1460         * WebKitTestRunner/InjectedBundle/InjectedBundle.h: Allow setting and audio result handle to be transferred.
1461         (WTR::InjectedBundle::setAudioResult):
1462         (InjectedBundle):
1463         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1464         (WTR::InjectedBundlePage::dump): Dummy handling the Audio case in the switch, build fix for Mac.
1465         * WebKitTestRunner/TestInvocation.cpp: Audio dump support, dump audio if we haven't received a text dump from the web process.
1466         (WTR::TestInvocation::dumpResults):
1467         (WTR):
1468         (WTR::TestInvocation::dumpAudio):
1469         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle): Storing received audio data before dumping.
1470         * WebKitTestRunner/TestInvocation.h:
1471         (TestInvocation):
1472
1473 2012-12-20  Alberto Garcia  <agarcia@igalia.com>
1474
1475         [GTK] Count failed and timed out tests separately
1476         https://bugs.webkit.org/show_bug.cgi?id=105448
1477
1478         Reviewed by Carlos Garcia Campos.
1479
1480         Display the number of tests that fail, time out and are skipped
1481         separately in the buildbot page.
1482
1483         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1484         (RunGtkAPITests.commandComplete):
1485         (RunGtkAPITests.evaluateCommand):
1486         (RunGtkAPITests.getText2):
1487         * Scripts/run-gtk-tests:
1488         (TestRunner.run_tests):
1489
1490 2012-12-19  Ryosuke Niwa  <rniwa@webkit.org>
1491
1492         PerfTest.parse_output does too much
1493         https://bugs.webkit.org/show_bug.cgi?id=105391
1494
1495         Reviewed by Tony Chang.
1496
1497         Moved the code to filter lines into PerfTest._filter_output, which has been renamed from _filter_stderr.
1498         Also moved the code to output test results into PerfTest._run_with_driver.
1499
1500         * Scripts/webkitpy/performance_tests/perftest.py:
1501         (PerfTest.__init__): Added _description.
1502         (PerfTest.description): Added.
1503         (PerfTest._run_with_driver): Moved the code to output test description and test results.
1504         (PerfTest._filter_output): Moved and renamed from PerfTest._filter_stderr.
1505         (PerfTest.parse_output): Removed the code to output test results. Return test results and description
1506         respectively. Also removed the code that allowed some tests to omit values since all tests report each
1507         iteration now since r136492.
1508         (PerfTest.output_statistics): Removed the code to print test description, now done in _run_with_driver.
1509         (ChromiumStylePerfTest._run_with_driver): Added. Chromium style tests are sufficiently different from
1510         regular PerfTest that it doesn't make much sense to share _run_with_driver. But really, we should just
1511         get rid of this type of test altogether in favor of regular performance tests that uses runner.js.
1512         (ChromiumStylePerfTest.parse_and_log_output): Renamed from parse_output.
1513         (PageLoadingPerfTest._run_with_driver): Removed the explicit '' for the test description.
1514         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
1515         (MainTest.test_parse_output): Removed the expected logs since parse_output no longer prints out results.
1516         Also added a call to _filter_output since parse_output doesn't filter the output text anymore.
1517         (MainTest.test_parse_output_with_failing_line): Added a call to _filter_output. Also added ',' after 'Time:'
1518         so that it's not string-concatenated with the next line.
1519         (MainTest.test_parse_output_with_description): Added; a test for PerfTest.description().
1520         (MainTest.test_parse_output_with_subtests): Removed the expected logs, and added a call to _filter_output.
1521
1522 2012-12-19  Alexis Menard  <alexis@webkit.org>
1523
1524         Implement CSS parsing for CSS transitions unprefixed.
1525         https://bugs.webkit.org/show_bug.cgi?id=104804
1526
1527         Reviewed by Dean Jackson.
1528
1529         Add a new flag ENABLE_CSS_TRANSFORMS_ANIMATIONS_TRANSITIONS_UNPREFIXED
1530         to cover the work of unprefixing Transforms, Animations and 
1531         Transitions. It will let the possibility of each ports to turn it off 
1532         in their release branches until we're confident that these CSS 
1533         properties are ready to be unprefixed.
1534
1535         * Scripts/webkitperl/FeatureList.pm:
1536         * qmake/mkspecs/features/features.pri:
1537
1538 2012-12-19  Julien Chaffraix  <jchaffraix@webkit.org>
1539
1540         2 clampTo unit tests are failing
1541         https://bugs.webkit.org/show_bug.cgi?id=105253
1542
1543         Reviewed by Emil A Eklund.
1544
1545         Removed clampToUnsigned as it was actually testing exactly what
1546         clampToUnsignedLong was doing. Also made clampToIntFloat more robust
1547         to precision loss (fix suggested by Emil).
1548
1549         * TestWebKitAPI/Tests/WTF/MathExtras.cpp:
1550
1551 2012-12-19  Thiago Marcos P. Santos  <thiago.santos@intel.com>
1552
1553         [EFL] Move ImageDiff to the Tools directory root
1554         https://bugs.webkit.org/show_bug.cgi?id=105305
1555
1556         Reviewed by Laszlo Gombos.
1557
1558         We previously moved this directory from the DRT directory to
1559         efl/ImageDiff. The major reason for doing so was because ImageDiff does
1560         not depend on DumpRenderTree but for legacy reasons was living inside the
1561         DRT folder, even though it is needed when running pixel tests for WK2.
1562
1563         Now we are moving to Tool directory root as other ports agreed to do the same.
1564
1565         * CMakeLists.txt:
1566         * ImageDiff/CMakeLists.txt: Added.
1567         * ImageDiff/PlatformEfl.cmake: Added.
1568         * ImageDiff/efl/ImageDiff.cpp: Renamed from Tools/efl/ImageDiff/ImageDiff.cpp.
1569         * efl/ImageDiff/CMakeLists.txt: Removed.
1570
1571 2012-12-19  Zeno Albisser  <zeno@webkit.org>
1572
1573         [Qt][Mac] run-launcher cannot find Qt5WebKitWidgets.
1574         https://bugs.webkit.org/show_bug.cgi?id=105309
1575
1576         The version number is only used as a part
1577         of the library name on Linux.
1578
1579         Reviewed by Kenneth Rohde Christiansen.
1580
1581         * Scripts/webkitdirs.pm:
1582         (builtDylibPathForName):
1583
1584 2012-12-19  Alberto Garcia  <agarcia@igalia.com>
1585
1586         [GTK] Tests that time out are not considered as failures.
1587         https://bugs.webkit.org/show_bug.cgi?id=105318
1588
1589         Reviewed by Martin Robinson.
1590
1591         Fix parsing of the GTK API tests log files.
1592
1593         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1594         (RunGtkAPITests.commandComplete):
1595
1596 2012-12-18  Ilya Tikhonovsky  <loislo@chromium.org>
1597
1598         Unreviewed, rolling out r138061.
1599         https://bugs.webkit.org/show_bug.cgi?id=105396
1600
1601         Original patch:
1602         http://trac.webkit.org/changeset/138061
1603         https://bugs.webkit.org/show_bug.cgi?id=97359
1604
1605         It is crashing on Debug bots
1606
1607         * DumpRenderTree/chromium/TestRunner/src/AccessibilityUIElementChromium.cpp:
1608         (WebTestRunner::AccessibilityUIElement::AccessibilityUIElement):
1609         (WebTestRunner::AccessibilityUIElement::elementAtPointCallback):
1610         * DumpRenderTree/chromium/TestRunner/src/AccessibilityUIElementChromium.h:
1611         (AccessibilityUIElement):
1612
1613 2012-12-18  Alice Liu  <alice.liu@apple.com>
1614
1615         Test for https://bugs.webkit.org/show_bug.cgi?id=105106
1616         Add SPI to WebKit1 WebFrame for hit testing
1617
1618         Reviewed by Dan Bernstein.
1619
1620         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Added file Tests/mac/ElementAtPointInWebFrame.mm
1621         * TestWebKitAPI/Tests/mac/ElementAtPointInWebFrame.mm: Added.
1622         (TestWebKitAPI::TEST): Loads html with two divs positioned in the 2nd and 4th quadrants of the webview.
1623         Then hit-tests at three points, expecting to hit the two divs and body element. 
1624
1625 2012-12-18  Alice Liu  <alice.liu@apple.com>
1626
1627         Test for https://bugs.webkit.org/show_bug.cgi?id=105262
1628         Add SPI to WebKit1 WebFrame for node conversion to JSValueRef
1629
1630         Reviewed by Anders Carlsson.
1631
1632         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Added file Tests/mac/JSWrapperForNodeInWebFrame.mm.mm
1633         * TestWebKitAPI/Tests/mac/JSWrapperForNodeInWebFrame.mm.mm: Added.
1634         (TestWebKitAPI::TEST): Tests for the correct JS wrapper for a DOMNode, provided a WebScriptWorld.
1635         Loads html with a single div element. In an isolated world, creates a property on that node.
1636         Also, in the standard world, creates a different property on that node.  Then tests for 4 things:
1637             - Existence of the isolated property in the isolated world.
1638             - Existence of the standard property in the standard world.
1639             - Non-existence of the isolated property in the standard world.
1640             - Non-existence of the standard property in the isolated world.
1641
1642 2012-12-18  Simon Fraser  <simon.fraser@apple.com>
1643
1644         Disable the SimplifyMarkup API test, wich asserts in debug builds.
1645         Tracked by https://bugs.webkit.org/show_bug.cgi?id=105370.
1646         
1647         * TestWebKitAPI/Tests/mac/SimplifyMarkup.mm:
1648         (TestWebKitAPI::TEST):
1649
1650 2012-12-18  Joseph Pecoraro  <pecoraro@apple.com>
1651
1652         [Mac] Add Build Phase to Check Headers for Inappropriate Macros (Platform.h macros)
1653         https://bugs.webkit.org/show_bug.cgi?id=104279
1654
1655         Reviewed by David Kilzer.
1656
1657         Add a build phase script that checks a framework's header files for
1658         Platform.h and Compiler.h macros. Also add some tests for the script.
1659
1660         * Scripts/check-for-inappropriate-macros-in-external-headers: Added.
1661         Script to be used in build phases to check for inappropriate macros in headers.
1662
1663         * Scripts/test-webkitruby: Added.
1664         Script that can be used to run all webkitruby tests.
1665
1666         * Scripts/test-webkit-scripts:
1667         Also run the ruby tests.
1668
1669         * Scripts/webkitruby/check-for-inappropriate-macros-in-external-headers-tests/fake-data-failing-expected.txt: Added.
1670         * Scripts/webkitruby/check-for-inappropriate-macros-in-external-headers-tests/pass-expected.txt: Added.
1671         * Scripts/webkitruby/check-for-inappropriate-macros-in-external-headers-tests/resources/Fake.framework/Headers/Fail.h: Added.
1672         * Scripts/webkitruby/check-for-inappropriate-macros-in-external-headers-tests/resources/Fake.framework/Headers/Pass.h: Added.
1673         * Scripts/webkitruby/check-for-inappropriate-macros-in-external-headers-tests/run-test.rb: Added.
1674         Tests for the new script.
1675
1676 2012-12-18  Dominic Mazzoni  <dmazzoni@google.com>
1677
1678         AX: support clickPoint in DRT for chromium
1679         https://bugs.webkit.org/show_bug.cgi?id=97359
1680
1681         Reviewed by Chris Fleizach.
1682
1683         Implement clickPointX, clickPointY, and elementAtPoint for Chromium, so that
1684         several more tests can pass.
1685
1686         * DumpRenderTree/chromium/TestRunner/src/AccessibilityUIElementChromium.cpp:
1687         (WebTestRunner::AccessibilityUIElement::AccessibilityUIElement):
1688         (WebTestRunner::AccessibilityUIElement::clickPointXGetterCallback):
1689         (WebTestRunner):
1690         (WebTestRunner::AccessibilityUIElement::clickPointYGetterCallback):
1691         (WebTestRunner::AccessibilityUIElement::elementAtPointCallback):
1692         * DumpRenderTree/chromium/TestRunner/src/AccessibilityUIElementChromium.h:
1693         (AccessibilityUIElement):
1694
1695 2012-12-18  Ryosuke Niwa  <rniwa@webkit.org>
1696
1697         Build fix after r138045.
1698
1699         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
1700         (PerfTestsRunner._collect_tests): Use relative_path. We skip only when this path
1701         is not explicitly specified. The condition was backwards but never matched so we
1702         always ran the test regardless of whether it's listed in the skipped list or not.
1703         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
1704         (test_collect_tests_with_ignored_skipped_list): Renamed so that it actually runs.
1705
1706 2012-12-18  Emil A Eklund  <eae@chromium.org>
1707
1708         TestWebKitAPI/Tests/WebCore/LayoutUnit fails on mac
1709         https://bugs.webkit.org/show_bug.cgi?id=105332
1710         
1711         Reviewed by Levi Weintraub.
1712         
1713         Create new test suite for LayoutUnit tests as the KURL tests are
1714         of the incompatible TEST_F type.
1715
1716         * TestWebKitAPI/Tests/WebCore/LayoutUnit.cpp:
1717         (TestWebKitAPI):
1718         (TestWebKitAPI::TEST_F):
1719
1720 2012-12-18  Ryosuke Niwa  <rniwa@webkit.org>
1721
1722         Running a skipped test with run-perf-tests could alert the user
1723         https://bugs.webkit.org/show_bug.cgi?id=105219
1724
1725         Reviewed by Eric Seidel.
1726
1727         Force run skipped performance tests explicitly specified in the command line.
1728         Also warn the user when the specified paths don't exist.
1729
1730         * Scripts/webkitpy/layout_tests/port/base.py:
1731         (Port.relative_test_filename):
1732         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
1733         (PerfTestsRunner._parse_args):
1734         (PerfTestsRunner._collect_tests):
1735         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
1736         (test_collect_tests_with_skipped_list_and_files):
1737
1738 2012-12-18  Yaron Friedman  <yfriedman@chromium.org>
1739
1740         Chrome Android java output is too verbose
1741         https://bugs.webkit.org/show_bug.cgi?id=105247
1742
1743         Reviewed by Tony Chang.
1744
1745         Pass -quiet to ant to suppress extraneous output.
1746
1747         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
1748         * TestWebKitAPI/TestWebKitAPI.gyp/TestWebKitAPI.gyp:
1749
1750 2012-12-18  Ryosuke Niwa  <rniwa@webkit.org>
1751
1752         Consider removing --pause-before-testing option
1753         https://bugs.webkit.org/show_bug.cgi?id=105248
1754
1755         Reviewed by Eric Seidel.
1756
1757         Remove --pause-before-testing option.
1758
1759         * Tools/Scripts/webkitpy/layout_tests/port/driver.py:
1760         (DriverProxy.start): Removed.
1761         * Scripts/webkitpy/performance_tests/perftest.py:
1762         (PerfTest.__init__):
1763         (PerfTest.test_path): Renamed from path_or_url.
1764         (PerfTest.run):
1765         (PerfTest): Removed "driver" argument.
1766         (PerfTest._run_with_driver):
1767         (PerfTest.run_single):
1768         (ChromiumStylePerfTest.__init__):
1769         (PageLoadingPerfTest.__init__):
1770         (PageLoadingPerfTest.run_single):
1771         (PageLoadingPerfTest._run_with_driver): Renamed from run so that PerfTest.run can create a driver for us.
1772         (ReplayPerfTest.__init__):
1773         (ReplayPerfTest.prepare):
1774         * Scripts/webkitpy/performance_tests/perftest_unittest.py:
1775         (TestPageLoadingPerfTest.test_run):
1776         (TestPageLoadingPerfTest.test_run_with_memory_output):
1777         (TestPageLoadingPerfTest.test_run_with_bad_output):
1778         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
1779         (PerfTestsRunner._parse_args):
1780         (PerfTestsRunner._run_tests_set): Removed the code to create a driver and pause before running the test.
1781         (PerfTestsRunner._run_single_test): Removed "driver" argument.
1782         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py:
1783         (run_test):
1784         (test_run_test_pause_before_testing): Removed.
1785
1786 2012-12-18  Alberto Garcia  <agarcia@igalia.com>
1787
1788         [GTK] Tests that time out are not considered as failures.
1789         https://bugs.webkit.org/show_bug.cgi?id=105318
1790
1791         Reviewed by Martin Robinson.
1792
1793         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1794         (RunGtkAPITests.commandComplete):
1795         Consider log lines that end with TIMEOUT as errors.
1796
1797         * Scripts/run-gtk-tests:
1798         (TestRunner.run_tests):
1799         Consider timeouts as failures when returning the number of failed
1800         tests.
1801
1802 2012-12-18  Pravin D  <pravin.d@samsung.com>
1803
1804         Unreviewed, adding secondary email.
1805
1806         * Scripts/webkitpy/common/config/committers.py:
1807
1808 2012-12-18  Pravin D  <pravin.d@samsung.com>
1809
1810         Unreviewed, promoting myself to committer's list.
1811
1812         * Scripts/webkitpy/common/config/committers.py:
1813
1814 2012-12-18  Thiago Marcos P. Santos  <thiago.santos@intel.com>
1815
1816         [EFL] ImageDiff does not build on WebKit2 only builds
1817         https://bugs.webkit.org/show_bug.cgi?id=105288
1818
1819         Reviewed by Kenneth Rohde Christiansen.
1820
1821         Moved ImageDiff to the efl/ folder inside Tools/ (so we don't pollute
1822         the folder with things which are EFL-specific.
1823
1824         ImageDiff will be built by default now regardless of WebKit API flavor
1825         you choose.
1826
1827         * CMakeLists.txt:
1828         * DumpRenderTree/efl/CMakeLists.txt:
1829         * efl/ImageDiff/CMakeLists.txt: Added.
1830         * efl/ImageDiff/ImageDiff.cpp: Renamed from Tools/DumpRenderTree/efl/ImageDiff.cpp.
1831
1832 2012-12-18  Kenneth Rohde Christiansen  <kenneth@webkit.org>
1833
1834         [EFL][WK2] When creating new windows, use the opener's context
1835         https://bugs.webkit.org/show_bug.cgi?id=105283
1836
1837         Reviewed by Laszlo Gombos.
1838
1839         Use the context of the opener if exists.
1840
1841         * MiniBrowser/efl/main.c:
1842         (on_key_down):
1843         (on_window_create):
1844         (window_create):
1845         (elm_main):
1846
1847 2012-12-18  Thiago Marcos P. Santos  <thiago.santos@intel.com>
1848
1849         [EFL][WK2] Gardening of TestWebKitAPI tests
1850         https://bugs.webkit.org/show_bug.cgi?id=105268
1851
1852         Reviewed by Kenneth Rohde Christiansen.
1853
1854         Some tests were marked as failing but are passing and the newly added
1855         LayoutUnit should have it's own binary because it conflicts with the
1856         fixture used by KURL and is failing on the bots because of that.
1857
1858         * TestWebKitAPI/CMakeLists.txt:
1859         * TestWebKitAPI/PlatformEfl.cmake:
1860
1861 2012-12-18  Jochen Eisinger  <jochen@chromium.org>
1862
1863         [chromium] move webintent related methods to TestRunner library
1864         https://bugs.webkit.org/show_bug.cgi?id=105164
1865
1866         Reviewed by Darin Fisher.
1867
1868         Also deleted the unused deliveredIntentFailure and
1869         deliveredIntentResult methods.
1870
1871         * DumpRenderTree/chromium/DRTTestRunner.cpp:
1872         (DRTTestRunner::DRTTestRunner):
1873         * DumpRenderTree/chromium/DRTTestRunner.h:
1874         (DRTTestRunner):
1875         * DumpRenderTree/chromium/TestRunner/public/WebTestDelegate.h:
1876         (WebKit):
1877         (WebTestRunner::WebTestDelegate::setCurrentWebIntentRequest):
1878         (WebTestRunner::WebTestDelegate::currentWebIntentRequest):
1879         * DumpRenderTree/chromium/TestRunner/public/WebTestProxy.h:
1880         (WebKit):
1881         (WebTestProxyBase):
1882         (WebTestRunner::WebTestProxy::registerIntentService):
1883         (WebTestRunner::WebTestProxy::dispatchIntent):
1884         * DumpRenderTree/chromium/TestRunner/src/TestRunner.cpp:
1885         (WebTestRunner::TestRunner::TestRunner):
1886         (WebTestRunner::TestRunner::sendWebIntentResponse):
1887         (WebTestRunner):
1888         (WebTestRunner::TestRunner::deliverWebIntent):
1889         * DumpRenderTree/chromium/TestRunner/src/TestRunner.h:
1890         (TestRunner):
1891         * DumpRenderTree/chromium/TestRunner/src/WebTestProxy.cpp:
1892         (WebTestRunner::WebTestProxyBase::postAccessibilityNotification):
1893         (WebTestRunner::WebTestProxyBase::registerIntentService):
1894         (WebTestRunner):
1895         (WebTestRunner::WebTestProxyBase::dispatchIntent):
1896         * DumpRenderTree/chromium/WebViewHost.cpp:
1897         (WebViewHost::setCurrentWebIntentRequest):
1898         (WebViewHost::currentWebIntentRequest):
1899         * DumpRenderTree/chromium/WebViewHost.h:
1900         (WebKit):
1901         (WebViewHost):
1902
1903 2012-12-18  Jochen Eisinger  <jochen@chromium.org>
1904
1905         [chromium] move editing callback dumping to the WebTestProxy
1906         https://bugs.webkit.org/show_bug.cgi?id=105013
1907
1908         Reviewed by Darin Fisher.
1909
1910         To decide whether or not to dump the editing callbacks, the WebTestProxy
1911         needs to query the TestRunner. For this purpose, I introduce a
1912         WebTestRunner class that just exposes the shouldDumpEditingCallbacks
1913         method. In contrast to e.g. WebEventSender, the WebTestRunner is just
1914         an interface, and it is not owned by the WebTestInterfaces, but the
1915         embedder of the TestRunner library must provide it. This will change
1916         once all TestRunner methods are in the library.
1917
1918         * DumpRenderTree/DumpRenderTree.gypi:
1919         * DumpRenderTree/chromium/DRTTestRunner.cpp:
1920         (DRTTestRunner::DRTTestRunner):
1921         (DRTTestRunner::reset):
1922         * DumpRenderTree/chromium/DRTTestRunner.h:
1923         (DRTTestRunner):
1924         (DRTTestRunner::taskList):
1925         (DRTTestRunner::WorkQueue::taskList):
1926         * DumpRenderTree/chromium/TestRunner/public/WebTestInterfaces.h:
1927         (WebTestRunner):
1928         (WebTestInterfaces):
1929         * DumpRenderTree/chromium/TestRunner/public/WebTestProxy.h:
1930         (WebKit):
1931         (WebTestRunner):
1932         (WebTestProxyBase):
1933         (WebTestRunner::WebTestProxy::shouldBeginEditing):
1934         (WebTestRunner::WebTestProxy::shouldEndEditing):
1935         (WebTestRunner::WebTestProxy::shouldInsertNode):
1936         (WebTestRunner::WebTestProxy::shouldInsertText):
1937         (WebTestRunner::WebTestProxy::shouldChangeSelectedRange):
1938         (WebTestRunner::WebTestProxy::shouldDeleteRange):
1939         (WebTestRunner::WebTestProxy::shouldApplyStyle):
1940         (WebTestRunner::WebTestProxy::didBeginEditing):
1941         (WebTestRunner::WebTestProxy::didChangeSelection):
1942         (WebTestRunner::WebTestProxy::didChangeContents):
1943         (WebTestRunner::WebTestProxy::didEndEditing):
1944         * DumpRenderTree/chromium/TestRunner/public/WebTestRunner.h: Copied from Tools/DumpRenderTree/chromium/TestRunner/public/WebTestInterfaces.h.
1945         (WebTestRunner):
1946         (WebTestRunner::WebTestRunner::shouldDumpEditingCallbacks):
1947         * DumpRenderTree/chromium/TestRunner/src/TestRunner.cpp:
1948         (WebTestRunner::TestRunner::TestRunner):
1949         (WebTestRunner::TestRunner::~TestRunner):
1950         (WebTestRunner):
1951         (WebTestRunner::TestRunner::reset):
1952         (WebTestRunner::TestRunner::shouldDumpEditingCallbacks):
1953         (WebTestRunner::TestRunner::dumpEditingCallbacks):
1954         * DumpRenderTree/chromium/TestRunner/src/TestRunner.h:
1955         (TestRunner):
1956         * DumpRenderTree/chromium/TestRunner/src/WebTestInterfaces.cpp:
1957         (WebTestRunner::WebTestInterfaces::Internal::testRunner):
1958         (WebTestRunner::WebTestInterfaces::Internal::setTestRunner):
1959         (WebTestInterfaces::Internal):
1960         (WebTestRunner::WebTestInterfaces::Internal::Internal):
1961         (WebTestRunner::WebTestInterfaces::testRunner):
1962         (WebTestRunner):
1963         (WebTestRunner::WebTestInterfaces::setTestRunner):
1964         * DumpRenderTree/chromium/TestRunner/src/WebTestProxy.cpp:
1965         (WebTestRunner::WebTestProxyBase::postAccessibilityNotification):
1966         (WebTestRunner):
1967         (WebTestRunner::WebTestProxyBase::shouldBeginEditing):
1968         (WebTestRunner::WebTestProxyBase::shouldEndEditing):
1969         (WebTestRunner::WebTestProxyBase::shouldInsertNode):
1970         (WebTestRunner::WebTestProxyBase::shouldInsertText):
1971         (WebTestRunner::WebTestProxyBase::shouldChangeSelectedRange):
1972         (WebTestRunner::WebTestProxyBase::shouldDeleteRange):
1973         (WebTestRunner::WebTestProxyBase::shouldApplyStyle):
1974         (WebTestRunner::WebTestProxyBase::didBeginEditing):
1975         (WebTestRunner::WebTestProxyBase::didChangeSelection):
1976         (WebTestRunner::WebTestProxyBase::didChangeContents):
1977         (WebTestRunner::WebTestProxyBase::didEndEditing):
1978         * DumpRenderTree/chromium/TestShell.cpp:
1979         (TestShell::initialize):
1980         (TestShell::createNewWindow):
1981         * DumpRenderTree/chromium/WebViewHost.cpp:
1982         (WebViewHost::shouldBeginEditing):
1983         (WebViewHost::shouldEndEditing):
1984         (WebViewHost::shouldInsertNode):
1985         (WebViewHost::shouldInsertText):
1986         (WebViewHost::shouldChangeSelectedRange):
1987         (WebViewHost::shouldDeleteRange):
1988         (WebViewHost::shouldApplyStyle):
1989         * DumpRenderTree/chromium/WebViewHost.h:
1990         (WebViewHost):
1991
1992 2012-12-17  Jochen Eisinger  <jochen@chromium.org>
1993
1994         [chromium] Fix methods in TestRunner that directly modify the settings instead of going through WebPreferences
1995         https://bugs.webkit.org/show_bug.cgi?id=105154
1996
1997         Reviewed by Tony Chang.
1998
1999         If the settings are modified directly, the embedder can't keep track of
2000         them and might accidentially override the changed settings.
2001
2002         * DumpRenderTree/chromium/TestRunner/public/WebPreferences.h:
2003         (WebPreferences):
2004         * DumpRenderTree/chromium/TestRunner/src/TestRunner.cpp:
2005         (WebTestRunner::TestRunner::TestRunner):
2006         (WebTestRunner::TestRunner::setAsynchronousSpellCheckingEnabled):
2007         (WebTestRunner):
2008         (WebTestRunner::TestRunner::setMinimumTimerInterval):
2009         (WebTestRunner::TestRunner::setTouchDragDropEnabled):
2010         * DumpRenderTree/chromium/TestRunner/src/TestRunner.h:
2011         (TestRunner):
2012         * DumpRenderTree/chromium/TestRunner/src/WebPreferences.cpp:
2013         (WebTestRunner::WebPreferences::reset):
2014         (WebTestRunner::WebPreferences::applyTo):
2015
2016 2012-12-17  Julien Chaffraix  <jchaffraix@webkit.org>
2017
2018         Unreviewed unit test skipping.
2019
2020         Filed https://bugs.webkit.org/show_bug.cgi?id=105253 to re-enable them.
2021
2022         * TestWebKitAPI/Tests/WTF/MathExtras.cpp:
2023         (TestWebKitAPI):
2024         (TestWebKitAPI::TEST):
2025
2026 2012-12-17  Dean Jackson  <dino@apple.com>
2027
2028         Unreviewed. Add Silvia Pfeiffer to contributor list.
2029
2030         * Scripts/webkitpy/common/config/committers.py:
2031
2032 2012-12-17  Dirk Pranke  <dpranke@chromium.org>
2033
2034         build.webkit.org - add more google windows bots
2035         https://bugs.webkit.org/show_bug.cgi?id=105225
2036
2037         Reviewed by Eric Seidel.
2038
2039         Adds two more windows bots (google-windows-3, a builder, and
2040         google-windows-4, a tester) for the Chromium configs.
2041
2042         * BuildSlaveSupport/build.webkit.org-config/config.json:
2043
2044 2012-12-17  Eric Seidel  <eric@webkit.org>
2045
2046         Cleanup --profiler=perf code a little and allow up to 1000 report files (instead of 100)
2047         https://bugs.webkit.org/show_bug.cgi?id=105217
2048
2049         Reviewed by Dirk Pranke.
2050
2051         Previously --profile would fail strangely if find_unused_filename returned None
2052         (which would happen as soon as we had 100 report files).
2053         Now we'll search through 1000 filenames before giving up (and assert if we failed to find one).
2054         I also made some minor cleanups to the output of perf, and tried to make the
2055         commands more self-documenting by using long-names for args.
2056
2057         * Scripts/webkitpy/common/system/profiler.py:
2058         (SingleFileOutputProfiler.__init__):
2059         (Perf.attach_to_pid):
2060         (Perf.profile_after_exit):
2061
2062 2012-12-17  Sheriff Bot  <webkit.review.bot@gmail.com>
2063
2064         Unreviewed, rolling out r137198.
2065         http://trac.webkit.org/changeset/137198
2066         https://bugs.webkit.org/show_bug.cgi?id=105212
2067
2068         This patch is causing API behavior compatibility problems
2069         (Requested by zdobersek on #webkit).
2070
2071         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2072         (willSendRequestCallback):
2073
2074 2012-12-17  Elliott Sprehn  <esprehn@chromium.org>
2075
2076         Unreviewed. Adding myself to committers list.
2077
2078         * Scripts/webkitpy/common/config/committers.py:
2079
2080 2012-12-17  Dirk Pranke  <dpranke@chromium.org>
2081
2082         nrwt: shuffle code around for cleanup in run_webkit_tests.py
2083         https://bugs.webkit.org/show_bug.cgi?id=105078
2084
2085         Reviewed by Ojan Vafai.
2086
2087         This patch reorders functions so that run_webkit_tests can be
2088         understood in a top-down matter better, and to make things slightly
2089         more sensible when I merge manager.py into it.
2090
2091         Also, this adds tests for the actual main() routine, which had several
2092         bugs get through testing a week or two ago.
2093
2094         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2095         (main):
2096         (parse_args):
2097         (_set_up_derived_options):
2098         (run):
2099         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2100         (RunTest):
2101         (RunTest.test_no_http_tests):
2102         (PortTest.disabled_test_mac_lion):
2103         (MainTest):
2104         (MainTest.test_exception_handling):
2105         (MainTest.test_exception_handling.interrupting_run):
2106         (MainTest.test_exception_handling.successful_run):
2107         (MainTest.test_exception_handling.successful_run.FakeRunDetails):
2108         (MainTest.test_exception_handling.exception_raising_run):
2109
2110 2012-12-17  Dirk Pranke  <dpranke@chromium.org>
2111
2112         webkitpy: move --lint-test-files code into its own module
2113         https://bugs.webkit.org/show_bug.cgi?id=105077
2114
2115         Reviewed by Eric Seidel.
2116
2117         It really didn't need to be in run_webkit_tests.py anymore.
2118         This patch adds a new lint-test-expectations scripts to Tools/Scripts;
2119         rwt --lint-test-files is still supported for backwards compatibility
2120         for now.
2121
2122         * Scripts/lint-test-expectations: Added.
2123         * Scripts/webkitpy/layout_tests/lint_test_expectations.py: Added.
2124         (lint):
2125         (main):
2126         * Scripts/webkitpy/layout_tests/lint_test_expectations_unittest.py: Added.
2127         (FakePort):
2128         (FakePort.__init__):
2129         (FakePort.test_configuration):
2130         (FakePort.expectations_dict):
2131         (FakePort.skipped_layout_tests):
2132         (FakePort.all_test_configurations):
2133         (FakePort.configuration_specifier_macros):
2134         (FakePort.get_option):
2135         (FakeFactory):
2136         (FakeFactory.__init__):
2137         (FakeFactory.get):
2138         (FakeFactory.all_port_names):
2139         (LintTest):
2140         (LintTest.test_all_configurations):
2141         (LintTest.test_lint_test_files):
2142         (LintTest.test_lint_test_files__errors):
2143         (MainTest):
2144         (MainTest.test_success):
2145         (MainTest.test_success.interrupting_lint):
2146         (MainTest.test_success.successful_lint):
2147         (MainTest.test_success.exception_raising_lint):
2148         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2149         (main):
2150         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2151         (StreamTestingMixin.assertNotEmpty):
2152
2153 2012-12-17  Adam Klein  <adamk@chromium.org>
2154
2155         build-webkit: rename --template-tag to --template-element to match ENABLE #define name
2156         https://bugs.webkit.org/show_bug.cgi?id=105072
2157
2158         Reviewed by Laszlo Gombos.
2159
2160         * Scripts/webkitperl/FeatureList.pm:
2161
2162 2012-12-17  Julien Chaffraix  <jchaffraix@webkit.org>
2163
2164         Add some unit testing for WTF::clampTo* functions
2165         https://bugs.webkit.org/show_bug.cgi?id=105060
2166
2167         Reviewed by Emil A Eklund.
2168
2169         The new tests are targeted at clamping to int / unsigned and float as those are
2170         the one explicitely exposed through more dedicated functions (like clampToInteger,
2171         clampToPositiveInteger and clampToFloat).
2172
2173         * TestWebKitAPI/Tests/WTF/MathExtras.cpp:
2174
2175 2012-12-17  Eric Seidel  <eric@webkit.org>
2176
2177         rpt --profile --chromium-android throws exception
2178         https://bugs.webkit.org/show_bug.cgi?id=105061
2179
2180         Reviewed by Daniel Bates.
2181
2182         Somehow popen(args) got turned into popen(*args) at some point,
2183         which caused my new stringify code to fail when args was passed by name.
2184         Fixed and tested.
2185
2186         * Scripts/webkitpy/common/system/executive.py:
2187         (Executive._stringify_args):
2188         (Executive):
2189         (Executive.popen):
2190         * Scripts/webkitpy/common/system/executive_unittest.py:
2191         (ExecutiveTest.test_popen_args):
2192
2193 2012-12-17  Mark Pilgrim  <pilgrim@chromium.org>
2194
2195         Use Platform::current() instead of webKitPlatformSupport() in DumpRenderTree
2196         https://bugs.webkit.org/show_bug.cgi?id=105054
2197
2198         Reviewed by Jochen Eisinger.
2199
2200         Part of a refactoring series; see tracking bug 82948.
2201
2202         * DumpRenderTree/chromium/NotificationPresenter.cpp:
2203         (NotificationPresenter::show):
2204         * DumpRenderTree/chromium/Task.cpp:
2205         (postTask):
2206         * DumpRenderTree/chromium/TestRunner/src/WebTestPlugin.cpp:
2207
2208 2012-12-17  Emil A Eklund  <eae@chromium.org>
2209
2210         Clamp values in LayoutUnit::operator/ when SATURATED_LAYOUT_ARITHMETIC is enabled
2211         https://bugs.webkit.org/show_bug.cgi?id=104955
2212
2213         Reviewed by Julien Chaffraix.
2214
2215         Add tests for LayoutUnit.
2216
2217         * TestWebKitAPI/CMakeLists.txt:
2218         * TestWebKitAPI/ForwardingHeaders: Added.
2219         * TestWebKitAPI/ForwardingHeaders/WebCore: Added.
2220         * TestWebKitAPI/ForwardingHeaders/WebCore/LayoutUnit.h: Added.
2221         * TestWebKitAPI/TestWebKitAPI.gyp/TestWebKitAPI.gyp:
2222         * TestWebKitAPI/TestWebKitAPI.gypi:
2223         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
2224         * TestWebKitAPI/Tests/WebCore/LayoutUnit.cpp: Added.
2225         (TestWebKitAPI):
2226         (TestWebKitAPI::TEST):
2227         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
2228
2229 2012-12-17  Jochen Eisinger  <jochen@chromium.org>
2230
2231         [chromium] removed unused preferences from testRunner.overridePreference
2232         https://bugs.webkit.org/show_bug.cgi?id=105169
2233
2234         Reviewed by Tony Chang.
2235
2236         Most notably this removes the code related to font family maps which is
2237         now handled via window.internals.settings
2238
2239         * DumpRenderTree/chromium/TestRunner/public/WebPreferences.h:
2240         (WebPreferences):
2241         * DumpRenderTree/chromium/TestRunner/src/TestRunner.cpp:
2242         (WebTestRunner::TestRunner::overridePreference):
2243         * DumpRenderTree/chromium/TestRunner/src/TestRunner.h:
2244         (TestRunner):
2245         * DumpRenderTree/chromium/TestRunner/src/WebPreferences.cpp:
2246         (WebTestRunner::WebPreferences::applyTo):
2247
2248 2012-12-17  James Robinson  <jamesr@chromium.org>
2249
2250         [chromium] Move creation of WebLayerTreeView to WebWidgetClient
2251         https://bugs.webkit.org/show_bug.cgi?id=105071
2252
2253         Reviewed by Adrienne Walker.
2254
2255         Update WebViewHost for the new WebWidgetClient interface.
2256
2257         * DumpRenderTree/chromium/WebViewHost.cpp:
2258         (WebViewHost::initializeLayerTreeView):
2259         (WebViewHost::layerTreeView):
2260         (WebViewHost::~WebViewHost):
2261         * DumpRenderTree/chromium/WebViewHost.h:
2262         (WebViewHost):
2263
2264 2012-12-17  Kenneth Rohde Christiansen  <kenneth@webkit.org>
2265
2266         [EFL][WK2] window_create doesn't receive the url
2267         https://bugs.webkit.org/show_bug.cgi?id=105184
2268
2269         Reviewed by Alexis Menard.
2270
2271         * MiniBrowser/efl/main.c: Use the url supplied to window_create
2272         (on_window_create):
2273
2274 2012-12-17  Ilya Tikhonovsky  <loislo@chromium.org>
2275
2276         Web Inspector: Native Memory Instrumentation: MemoryInstrumentation doesn't detect reportMemoryUsage method defined in a base class.
2277         https://bugs.webkit.org/show_bug.cgi?id=105026
2278
2279         Reviewed by Yury Semikhatsky.
2280
2281         Old SFINAE test was replaced with new one based on this article:
2282         http://stackoverflow.com/questions/1966362/sfinae-to-check-for-inherited-member-functions
2283
2284         New test which covers this problem was added.
2285
2286         * TestWebKitAPI/Tests/WTF/MemoryInstrumentationTest.cpp:
2287
2288 2012-12-17  Anthony Scian  <ascian@rim.com>
2289
2290         [Blackberry] Static code analysis warning fixes
2291         https://bugs.webkit.org/show_bug.cgi?id=105129
2292
2293         Reviewed by Rob Buis.
2294
2295         Added null checks for fopen calls, no recovery attempted; access fault prevention only.
2296
2297         * DumpRenderTree/blackberry/DumpRenderTree.cpp:
2298         (BlackBerry::WebKit::createFile):
2299         (BlackBerry::WebKit::DumpRenderTree::runTest):
2300
2301 2012-12-17  John Mellor  <johnme@chromium.org>
2302
2303         Add Anton Vayvod and Tim Volodine as contributors in committers.py
2304         https://bugs.webkit.org/show_bug.cgi?id=105025
2305
2306         Unreviewed. Adds two contributors to committers.py.
2307         They are initially contributing to the Text Autosizing feature (master bug: 84186).
2308
2309         * Scripts/webkitpy/common/config/committers.py:
2310
2311 2012-12-17  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
2312
2313         Unreviewed, add myself to committers.py.
2314
2315         * Scripts/webkitpy/common/config/committers.py:
2316
2317 2012-12-17  Oswald Buddenhagen  <oswald.buddenhagen@digia.com>
2318
2319         [Qt] remove some unnecessary CONFIG additions
2320
2321         Reviewed by Simon Hausmann.
2322
2323         qt is already added by spec_pre.prf, warn_on and depend_includepath by
2324         default_pre.prf.
2325
2326         * qmake/mkspecs/features/default_post.prf:
2327
2328 2012-12-17  Ilya Tikhonovsky  <loislo@chromium.org>
2329
2330         Unreviewed, rolling out r137892.
2331         http://trac.webkit.org/changeset/137892
2332         https://bugs.webkit.org/show_bug.cgi?id=105026
2333
2334         it broke compilation on windows
2335
2336         * TestWebKitAPI/Tests/WTF/MemoryInstrumentationTest.cpp:
2337
2338 2012-12-14  Ilya Tikhonovsky  <loislo@chromium.org>
2339
2340         Web Inspector: Native Memory Instrumentation: MemoryInstrumentation doesn't detect reportMemoryUsage method defined in a base class.
2341         https://bugs.webkit.org/show_bug.cgi?id=105026
2342
2343         Reviewed by Yury Semikhatsky.
2344
2345         Old SFINAE test was replaced with new one based on this article:
2346         http://stackoverflow.com/questions/1966362/sfinae-to-check-for-inherited-member-functions
2347
2348         New test which covers this problem was added.
2349
2350         * TestWebKitAPI/Tests/WTF/MemoryInstrumentationTest.cpp:
2351
2352 2012-12-17  Oswald Buddenhagen  <oswald.buddenhagen@digia.com>
2353
2354         [Qt] rely on automatic output directory setup
2355
2356         Reviewed by Simon Hausmann.
2357
2358         Now being a proper qt module (by virtue of having load(qt_build_config)
2359         in .qmake.conf), webkit gets the path setup goodies for free.
2360
2361         This also fixes Makefile.api.Debug/Release trying to generate qrc_WebKit.cpp
2362         at the same time, instead of the debug and release version ending up in different
2363         directories as they should.
2364
2365         * qmake/mkspecs/features/default_pre.prf:
2366
2367 2012-12-17  Charles Wei  <charles.wei@torchmobile.com.cn>
2368
2369         [BlackBerry] Need to re-enable video track
2370         https://bugs.webkit.org/show_bug.cgi?id=105155
2371
2372         Reviewed by George Staikos.
2373
2374         Just enable Video Track for BlackBerry.
2375
2376         * Scripts/webkitperl/FeatureList.pm:
2377
2378 2012-12-17  János Badics  <jbadics@inf.u-szeged.hu>
2379
2380         Unreviewed. Added myself to committers.py
2381
2382         * Scripts/webkitpy/common/config/committers.py:
2383
2384 2012-12-17  Jonathan Liu  <net147@gmail.com>
2385
2386         Fix linking of libraries with -Wl,-whole-archive
2387         https://bugs.webkit.org/show_bug.cgi?id=104436
2388
2389         Reviewed by Simon Hausmann.
2390
2391         If there is more than one library in LIBS linked with
2392         -Wl,-whole-archive -l... -Wl,-no-whole-archive, qmake tries to
2393         merge the duplicate flags in LIBS which causes linking to fail.
2394
2395         Add no_smart_library_merge to CONFIG to prevent the duplicate
2396         flags from being removed when using -Wl,-whole-archive.
2397
2398         * qmake/mkspecs/features/functions.prf:
2399
2400 2012-12-16  Arvid Nilsson  <anilsson@rim.com>
2401
2402         [BlackBerry] Adapt to new BlackBerry::Platform::TouchPoint API
2403         https://bugs.webkit.org/show_bug.cgi?id=105143
2404         RIM PR 171941
2405
2406         Reviewed by Rob Buis.
2407         Internally reviewed by George Staikos.
2408
2409         Adapt to new method names and encapsulation of TouchPoint data members.
2410
2411         * DumpRenderTree/blackberry/EventSender.cpp:
2412         (addTouchPointCallback):
2413         (updateTouchPointCallback):
2414         (touchEndCallback):
2415         (releaseTouchPointCallback):
2416         (sendTouchEvent):
2417
2418 2012-12-16  Maciej Stachowiak  <mjs@apple.com>
2419
2420         Add a script to compute the mean and 95% confidence interval (using two-sided t-test) of a sample set
2421         https://bugs.webkit.org/show_bug.cgi?id=105148
2422
2423         Reviewed by Filip Pizlo.
2424
2425         Usage examples:
2426         
2427         $ sampstat --help
2428         Usage: sampstat [options] [FILES]
2429           Compute the mean and 95% confidence interval of a sample set.
2430           Standard input or files must contain two or more decimal numbers, one per line.
2431         
2432         Options:
2433           -h, --help            show this help message and exit
2434           -u UNIT, --unit=UNIT  assume values are in units of UNIT
2435           -v, --verbose         print all values (with units)
2436         
2437         $ sampstat -u MB memresults.txt
2438         2356.90 MB +/- 101.34 MB (4.3%)
2439         
2440         $ sampstat -v -u MB memresults.txt
2441               2318.21 MB
2442               2399.56 MB
2443               2352.93 MB
2444         ----------------
2445         Mean: 2356.90 MB +/- 101.34 MB (4.3%)
2446         
2447
2448         * Scripts/sampstat: Added.
2449         (sum): Helper function to compute the sum of a list.
2450         (arithmeticMean): Compute the meam of a list.
2451         (standardDeviation): Compute the sample standard deviation (unbiased estimator).
2452         (standardError): Compute the sample standard error.
2453         (tDist): Compute t(.025, n-1), the t-value for atwo-sided 95% confidence interval. 
2454         (twoSidedConfidenceInterval): Compute the two-sided confidence interval range about the mean.
2455
2456 2012-12-15  Nima Ghanavatian  <nghanavatian@rim.com>
2457
2458         [BlackBerry] Pass key modifiers with touch and mouse events
2459         https://bugs.webkit.org/show_bug.cgi?id=105108
2460
2461         Reviewed by Rob Buis.
2462
2463         Update DRT to new api.
2464
2465         Internally reviewed by Gen Mak.
2466
2467         * DumpRenderTree/blackberry/EventSender.cpp:
2468         (mouseDownCallback):
2469         (mouseUpCallback):
2470         (mouseMoveToCallback):
2471
2472 2012-12-15  Zan Dobersek  <zandobersek@gmail.com>
2473
2474         Stop the XvfbDriver before proceeding with its starting setup
2475         https://bugs.webkit.org/show_bug.cgi?id=105006
2476
2477         Reviewed by Philippe Normand.
2478
2479         Just as with the base Driver implementation, stop the XvfbDriver
2480         before proceeding with setting up lock files and a new Xvfb instance.
2481         This forces any driver or Xvfb instances that are currently under
2482         the driver's control to be closed/killed and cleaned up after.
2483
2484         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py:
2485         (XvfbDriver._start):
2486
2487 2012-12-14  Dirk Pranke  <dpranke@chromium.org>
2488
2489         webkitpy: make port_name a required parameter to the non-test ports
2490         https://bugs.webkit.org/show_bug.cgi?id=105057
2491
2492         Reviewed by Eric Seidel.
2493
2494         This is just minor code cleanup along the path to making Port
2495         constructors simpler and pushing more logic into the port factory.
2496
2497         * Scripts/webkitpy/layout_tests/port/base.py:
2498         (Port.__init__):
2499         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2500         (PortTest.make_port):
2501         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
2502         (TestWebKitPort.__init__):
2503         * Scripts/webkitpy/layout_tests/port/test.py:
2504         (TestPort):
2505         (TestPort.determine_full_port_name):
2506         (TestPort.__init__):
2507         * Scripts/webkitpy/layout_tests/port/xvfbdriver_unittest.py:
2508         (XvfbDriverTest.make_driver):
2509         (XvfbDriverTest.test_stop):
2510
2511 2012-12-14  Dirk Pranke  <dpranke@chromium.org>
2512
2513         garden-o-matic doesn't know about reftests
2514         https://bugs.webkit.org/show_bug.cgi?id=101976
2515
2516         Reviewed by Eric Seidel.
2517
2518         Attempt to make garden-o-matic properly aware of reftests again.
2519
2520         This is re-landing r137407 with a change to
2521         rebaselineWithStatusUpdates() to actually pass the filtered
2522         failureInofLost to checkout.rebaseline().
2523
2524         Unfortunately, the testing coverage for this whole module is almost
2525         non-existent, and it's not obvious that there's a good way to
2526         stub out checkout.rebaseline() to test that the right thing
2527         happens. I verified the change manually.
2528
2529         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/controllers.js:
2530         (.):
2531         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results.js:
2532         (.):
2533         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results_unittests.js:
2534
2535 2012-12-14  Ryosuke Niwa  <rniwa@webkit.org>
2536
2537         "Running 1 DumpRenderTree over X shards" is not a helpful output
2538         https://bugs.webkit.org/show_bug.cgi?id=104858
2539
2540         Reviewed by Dirk Pranke.
2541
2542         The word "shard" is not typically used in test frameworks. This word is confusing for many WebKit
2543         contributors who typically don't have much experience on distributed computations and databases.
2544         So only print this message out in the debug mode.
2545
2546         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2547         * Scripts/webkitpy/layout_tests/views/printing.py:
2548         (Printer.print_workers_and_shards):
2549
2550 2012-12-13  Eric Seidel  <eric@webkit.org>
2551
2552         Add --profiler=PROFILER option to run-perf-tests to allow specifying which profiler to use on platforms with many
2553         https://bugs.webkit.org/show_bug.cgi?id=104891
2554
2555         Reviewed by Ryosuke Niwa.
2556
2557         I also implemented a very simple "Sample" Profiler using
2558         Mac OS X's /usr/bin/sample command line tool.
2559
2560         The real reason for this abstraction is to make it easy
2561         to support both perf and pprof on linux which seem to
2562         be about equally popular among those I ask in the Chrome team.
2563
2564         * Scripts/webkitpy/common/system/profiler.py:
2565         (ProfilerFactory.create_profiler):
2566         (ProfilerFactory):
2567         (ProfilerFactory.available_profilers_by_name):
2568         (ProfilerFactory.default_profiler_name):
2569         (Sample):
2570         (Sample.__init__):
2571         (Sample.attach_to_pid):
2572         (Sample.profile_after_exit):
2573         (IProfiler.attach_to_pid):
2574         * Scripts/webkitpy/layout_tests/port/driver.py:
2575         (Driver.__init__):
2576         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
2577         (PerfTestsRunner._parse_args):
2578
2579 2012-12-13  Eric Seidel  <eric@webkit.org>
2580
2581         Unreviewed, fixing typo in python unittest.
2582
2583         I changed the default profiler on linux from pprof to perf in
2584         bug 104971.  I failed to update the unittest results at that time.
2585
2586         * Scripts/webkitpy/common/system/profiler_unittest.py:
2587         (ProfilerFactoryTest.test_basic):
2588
2589 2012-12-13  Ian Vollick  <vollick@chromium.org>
2590
2591         [chromium] Add a virtual test suite for enabling opt-in to composited scrolling
2592         https://bugs.webkit.org/show_bug.cgi?id=104911
2593
2594         Reviewed by Dirk Pranke.
2595
2596         Adds two virtual tests suites to run the tests in
2597         compositing/overflow/ and scrollbars/ with the flag
2598         --enable-accelerated-overflow-scroll, which I've also plumbed through
2599         in the usual way.
2600
2601         * DumpRenderTree/chromium/DumpRenderTree.cpp:
2602         (main):
2603         * DumpRenderTree/chromium/TestRunner/public/WebPreferences.h:
2604         (WebPreferences):
2605         * DumpRenderTree/chromium/TestRunner/src/WebPreferences.cpp:
2606         (WebTestRunner::WebPreferences::reset):
2607         (WebTestRunner::WebPreferences::applyTo):
2608         * DumpRenderTree/chromium/TestShell.cpp:
2609         (TestShell::TestShell):
2610         (TestShell::resetWebSettings):
2611         * DumpRenderTree/chromium/TestShell.h:
2612         (TestShell::setAcceleratedCompositingForOverflowScrollEnabled):
2613         (TestShell):
2614         * Scripts/webkitpy/layout_tests/port/chromium.py:
2615         (ChromiumPort.virtual_test_suites):
2616
2617 2012-12-13  Eric Seidel  <eric@webkit.org>
2618
2619         Use 'perf' to profile on linux instead of google-pprof by default
2620         https://bugs.webkit.org/show_bug.cgi?id=104971
2621
2622         Reviewed by Dirk Pranke.
2623
2624         This makes Chromium Linux match Chromium Android and use the perf
2625         tool by default.  Once https://bugs.webkit.org/show_bug.cgi?id=104891
2626         lands it will be possible to use pprof again on Linux.
2627
2628         This is slightly more advanced than the AndroidPerf profiler as
2629         instead of using a timeout on "perf record" I instead watch
2630         for the termination of the target process and then control-C
2631         the 'perf record' process.  This required me to add two new
2632         methods to Executive, one to have a limited-time wait() and
2633         the second to be able to send a control-C.  I chose to add
2634         these to Executive to make them easier to mock/fix-for-win32
2635         at a later time if needed.
2636
2637         * Scripts/webkitpy/common/system/executive.py:
2638         (Executive.wait_limited):
2639         (Executive.interrupt):
2640         * Scripts/webkitpy/common/system/profiler.py:
2641         (ProfilerFactory.create_profiler):
2642         (Perf):
2643         (Perf.__init__):
2644         (Perf._perf_path):
2645         (Perf.attach_to_pid):
2646         (Perf._first_ten_lines_of_profile):
2647         (Perf.profile_after_exit):
2648
2649 2012-12-13  Julie Parent  <jparent@chromium.org>
2650
2651         Dashboard cleanup: remove usage of global g_builders.
2652         https://bugs.webkit.org/show_bug.cgi?id=104941
2653
2654         Reviewed by Dirk Pranke.
2655
2656         The dashboards use a lot of global state, which makes hacking on them
2657         complicated. This change removes the use of one such global: g_builders.
2658         In most cases, we can just use currentBuilderGroup().builders instead,
2659         which is now currentBuilders().
2660         Surprisingly, the most changes were required to the unit tests, since
2661         they were even bigger offenders of bad hygiene, relying on global state
2662         set by other tests, randomly clobbering global variables in ways the
2663         real code doesn't, etc.
2664
2665         * TestResultServer/static-dashboards/builders.js:
2666         (BuilderGroup.prototype.setup):
2667         * TestResultServer/static-dashboards/dashboard_base.js:
2668         (.switch.return):
2669         (htmlForTestTypeSwitcher):
2670         * TestResultServer/static-dashboards/flakiness_dashboard.js:
2671         (generatePage):
2672         (getAllTestsTrie):
2673         (processTestRunsForAllBuilders):
2674         (showPopupForBuild):
2675         (generatePageForExpectationsUpdate):
2676         (loadExpectationsLayoutTests):
2677         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
2678         (resetGlobals):
2679         (stubResultsByBuilder):
2680         (test):
2681         * TestResultServer/static-dashboards/loader.js:
2682         (.):
2683         * TestResultServer/static-dashboards/loader_unittests.js:
2684
2685 2012-12-13  Eric Seidel  <eric@webkit.org>
2686
2687         Callers should not have to stringify args before calling Executive run_command/popen
2688         https://bugs.webkit.org/show_bug.cgi?id=104975
2689
2690         Reviewed by Dirk Pranke.
2691
2692         One could argue that we should match the python call syntax here,
2693         but I think it's a more friendly API if we automagically handle
2694         stringification of args in run_command, etc.
2695         This removes map(unicode, args) from several callsites.
2696
2697         When I first tried to land this change, I didn't realize that
2698         Executive._command_for_printing depended on this behavior
2699         having been applied to args in run_command.  The fix is to
2700         call _stringify_args in both run_command and popen.
2701         This is slightly redundant, but given how short args have to be
2702         (due to shell limits), I don't think the double-encode check
2703         matters in practice.
2704
2705         This is slightly complicated by the fact that apache_http_server.py
2706         is the one caller in our codebase which uses shell=True.
2707         shell=True is a well-documented trail-of-tears:
2708         http://stackoverflow.com/questions/3172470/actual-meaning-of-shell-true-in-subprocess
2709         but to support this legacy (windows-only) code (which I can't easily test)
2710         I've added an if-hack to avoid stringifying the the popen(shell=True) case.
2711
2712         * Scripts/webkitpy/common/system/executive.py:
2713         (Executive.run_command):
2714         (Executive._stringify_args):
2715         (Executive.popen):
2716         * Scripts/webkitpy/common/system/executive_unittest.py:
2717         (ExecutiveTest.test_auto_stringify_args):
2718         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
2719         (attach_to_pid):
2720
2721 2012-12-13  Adrienne Walker  <enne@chromium.org>
2722
2723         Unreviewed, rolling out r137645, r137646, and r137667.
2724         http://trac.webkit.org/changeset/137645
2725         http://trac.webkit.org/changeset/137646
2726         http://trac.webkit.org/changeset/137667
2727         https://bugs.webkit.org/show_bug.cgi?id=104911
2728
2729         Breaks some overflow layout tests
2730
2731         * Scripts/webkitpy/layout_tests/port/chromium.py:
2732         (ChromiumPort.virtual_test_suites):
2733
2734 2012-12-13  Dirk Pranke  <dpranke@chromium.org>
2735
2736         Unreviewed, rolling out r137661.
2737         http://trac.webkit.org/changeset/137661
2738         https://bugs.webkit.org/show_bug.cgi?id=104891
2739
2740         broke unit tests, run-webkit-tests
2741
2742         * Scripts/webkitpy/common/system/executive.py:
2743         (Executive._run_command_with_teed_output):
2744         (Executive.run_command):
2745         (Executive.popen):
2746         * Scripts/webkitpy/common/system/profiler.py:
2747         (ProfilerFactory.create_profiler):
2748         (GooglePProf.profile_after_exit):
2749         (IProfiler.attach_to_pid):
2750         * Scripts/webkitpy/layout_tests/port/driver.py:
2751         (Driver.__init__):
2752         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
2753         (PerfTestsRunner._parse_args):
2754
2755 2012-12-13  Dirk Pranke  <dpranke@chromium.org>
2756
2757         nrwt: move rundetails, summarize_results into test_run_results
2758         https://bugs.webkit.org/show_bug.cgi?id=104963
2759
2760         Reviewed by Eric Seidel.
2761
2762         More cleanup; this the plain-old-data functions out of manager.py
2763         and into test_run_results next to the other aggregate data structures.
2764
2765         Also this cleans up a bunch of test code and reduces duplication.
2766
2767         * Scripts/webkitpy/layout_tests/controllers/manager.py:
2768         (Manager.run):
2769         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2770         (ManagerTest.test_look_for_new_crash_logs):
2771         * Scripts/webkitpy/layout_tests/models/test_run_results.py:
2772         (TestRunResults.__init__):
2773         (TestRunResults.add):
2774         (RunDetails):
2775         (RunDetails.__init__):
2776         (_interpret_test_failures):
2777         (summarize_results):
2778         * Scripts/webkitpy/layout_tests/models/test_run_results_unittest.py: Added.
2779         (get_result):
2780         (run_results):
2781         (summarized_results):
2782         (InterpretTestFailuresTest):
2783         (InterpretTestFailuresTest.setUp):
2784         (InterpretTestFailuresTest.test_interpret_test_failures):
2785         (SummarizedResultsTest):
2786         (SummarizedResultsTest.setUp):
2787         (SummarizedResultsTest.test_no_svn_revision):
2788         (SummarizedResultsTest.test_svn_revision):
2789         (SummarizedResultsTest.test_summarized_results_wontfix):
2790         * Scripts/webkitpy/layout_tests/views/buildbot_results_unittest.py:
2791         (BuildBotPrinterTests.get_printer):
2792         (BuildBotPrinterTests.test_print_unexpected_results):
2793         (BuildBotPrinterTests.test_print_results):
2794
2795 2012-12-13  Eric Seidel  <eric@webkit.org>
2796
2797         Make --profile output easier to understand for pprof on linux
2798         https://bugs.webkit.org/show_bug.cgi?id=104957
2799
2800         Reviewed by Dirk Pranke.
2801
2802         pprof is kinda mysterious, so we print some information
2803         on how to use it, and how to learn how to read the sample dump.
2804         This is similar to what we've done for perf output for Chromium Android.
2805
2806         * Scripts/webkitpy/common/system/profiler.py:
2807         (GooglePProf._pprof_path):
2808         (GooglePProf.profile_after_exit):
2809
2810 2012-12-13  Dirk Pranke  <dpranke@chromium.org>
2811
2812         Revert r137407 - it didn't actually work right.
2813         https://bugs.webkit.org/show_bug.cgi?id=101976
2814
2815         Unreviewed, build fix.
2816
2817         Turns out that rebaselining was completely broken with this change :(
2818         We're passing in a list of test names rather than the list of
2819         failureInfo objects.
2820
2821         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/controllers.js:
2822         (.):
2823         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results.js:
2824         (.):
2825         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results_unittests.js:
2826
2827 2012-12-13  Antoine Labour  <piman@chromium.org>
2828
2829         [chromium] Small fixes for WebPluginContainer::setWebLayer
2830         https://bugs.webkit.org/show_bug.cgi?id=104953
2831
2832         Reviewed by James Robinson.
2833
2834         * DumpRenderTree/chromium/TestRunner/src/WebTestPlugin.cpp: Clean up
2835         dangling pointers, force correct invalidation of the layer.
2836
2837 2012-12-13  Eric Seidel  <eric@webkit.org>
2838
2839         Add --profiler=PROFILER option to run-perf-tests to allow specifying which profiler to use on platforms with many
2840         https://bugs.webkit.org/show_bug.cgi?id=104891
2841
2842         Reviewed by Dirk Pranke.
2843
2844         I also implemented a very simple "Sample" Profiler using
2845         Mac OS X's /usr/bin/sample command line tool.
2846
2847         I also moved the map(unicode, args) call into Executive.popen
2848         so that all callers don't have to do it themselves.
2849
2850         The real reason for this abstraction is to make it easy
2851         to support both perf and pprof on linux which seem to
2852         be about equally popular among those I ask in the Chrome team.
2853
2854         * Scripts/webkitpy/common/system/executive.py:
2855         (Executive.popen):
2856         * Scripts/webkitpy/common/system/profiler.py:
2857         (ProfilerFactory.create_profiler):
2858         (ProfilerFactory):
2859         (ProfilerFactory.available_profilers_by_name):
2860         (ProfilerFactory.default_profiler_name):
2861         (Sample):
2862         (Sample.__init__):
2863         (Sample.attach_to_pid):
2864         (Sample.profile_after_exit):
2865         (IProfiler.attach_to_pid):
2866         * Scripts/webkitpy/layout_tests/port/driver.py:
2867         (Driver.__init__):
2868         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
2869         (PerfTestsRunner._parse_args):
2870
2871 2012-12-13  Dirk Pranke  <dpranke@chromium.org>
2872
2873         support -wk2 port names properly in webkitpy.layout_tests.port
2874         https://bugs.webkit.org/show_bug.cgi?id=104761
2875
2876         Reviewed by Eric Seidel.
2877
2878         Our support for WK2 port names was inconsistent at best; this patch
2879         modifies the code so that <port_name>-wk2 is supported for all ports that
2880         have WK2 implementations and is equivalent to specifying <port_name>
2881         and -2/--webkit-test-runner. In addition, this modifies
2882         builders.all_port_names() to include the wk2 variants.
2883
2884         * Scripts/webkitpy/common/checkout/baselineoptimizer.py:
2885         * Scripts/webkitpy/layout_tests/port/apple.py:
2886         (ApplePort.determine_full_port_name):
2887         * Scripts/webkitpy/layout_tests/port/base.py:
2888         (Port.determine_full_port_name):
2889         (Port.__init__):
2890         * Scripts/webkitpy/layout_tests/port/builders.py:
2891         * Scripts/webkitpy/layout_tests/port/driver_unittest.py:
2892         (DriverTest.make_port):
2893         * Scripts/webkitpy/layout_tests/port/mac.py:
2894         (MacPort.default_baseline_search_path):
2895         * Scripts/webkitpy/layout_tests/port/win.py:
2896         (WinPort.default_baseline_search_path):
2897         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
2898         (TestRebaselineExpectations.test_rebaseline_expectations):
2899
2900 2012-12-13  Ian Vollick  <vollick@chromium.org>
2901
2902         [chromium] Add a virtual test suite for enabling opt-in to composited scrolling
2903         https://bugs.webkit.org/show_bug.cgi?id=104911
2904
2905         Reviewed by Adrienne Walker.
2906
2907         Adds two virtual tests suites to run the tests in
2908         compositing/overflow/ and scrollbars/ with the flag
2909         --enable-accelerated-compositing-for-overflow-scroll
2910
2911         * Scripts/webkitpy/layout_tests/port/chromium.py:
2912         (ChromiumPort.virtual_test_suites):
2913
2914 2012-12-13  Dirk Pranke  <dpranke@chromium.org>
2915
2916         delete extraneous self.fail in mastercfg_unittest.py
2917         https://bugs.webkit.org/show_bug.cgi?id=104940
2918
2919         Reviewed by Eric Seidel.
2920
2921         This was accidentally introduced in r137404.
2922
2923         * BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:
2924
2925 2012-12-13  Dirk Pranke  <dpranke@chromium.org>
2926
2927         Fix a mis-named variable after r137621.
2928         https://bugs.webkit.org/show_bug.cgi?id=104871
2929
2930         Unreviewed, build fix.
2931
2932         * Scripts/webkitpy/layout_tests/views/buildbot_results.py:
2933         (BuildBotPrinter.print_results):
2934         * Scripts/webkitpy/layout_tests/views/buildbot_results_unittest.py:
2935
2936 2012-12-13  Dirk Pranke  <dpranke@chromium.org>
2937
2938         webkitpy: rename ResultSummary
2939         https://bugs.webkit.org/show_bug.cgi?id=104871
2940
2941         Reviewed by Eric Seidel.
2942
2943         This patch changes the names of ResultSummary and related modules,
2944         classes, and variables as a part of the larger clean up in bug 103824.
2945
2946         This change is basically a bunch of search and replaces:
2947         - result_summary.py -> test_run_results.py
2948         - class ResultSummary -> class TestRunResults
2949         - "result_summary" or "rs" as a generic variable name -> "run_results",
2950           when it doesn't matter (or we don't know) if this refers to the
2951           initial pass over the tests or the retry pass.
2952         - "result_summary" as a variable containing the results of the initial,
2953           complete test run -> "initial_results"
2954         - "retry_summary" as a variable containing the results of retried
2955           results -> "retry_results"
2956         - ResultSummary.results (the property) -> TestRunResults.results_by_name
2957         - ResultSummary.failures (the property) -> TestRunResults.failures_by_name
2958         - RunDetails.result_summary -> RunDetails.initial_results
2959         - RunDetails.retry_summary -> RunDetails.retry_results
2960
2961         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner.py:
2962         (LayoutTestRunner.__init__):
2963         (LayoutTestRunner.run_tests):
2964         (LayoutTestRunner._mark_interrupted_tests_as_skipped):
2965         (LayoutTestRunner._interrupt_if_at_failure_limits):
2966         (LayoutTestRunner._interrupt_if_at_failure_limits.interrupt_if_at_failure_limit):
2967         (LayoutTestRunner._update_summary_with_result):
2968         (LayoutTestRunner._handle_finished_test):
2969         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner_unittest.py:
2970         (FakePrinter.print_expected):
2971         (LayoutTestRunnerTests.test_interrupt_if_at_failure_limits):
2972         (LayoutTestRunnerTests.test_update_summary_with_result):
2973         * Scripts/webkitpy/layout_tests/controllers/manager.py:
2974         (RunDetails.__init__):
2975         (summarize_results):
2976         (Manager.run):
2977         (Manager._look_for_new_crash_logs):
2978         (Manager._tests_to_retry):
2979         (Manager._upload_json_files):
2980         (Manager._stats_trie):
2981         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2982         (ManagerTest.test_look_for_new_crash_logs):
2983         (SummarizeResultsTest):
2984         (SummarizeResultsTest.get_run_results):
2985         (SummarizeResultsTest.summarized_results):
2986         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
2987         (JSONLayoutResultsGenerator.__init__):
2988         (JSONLayoutResultsGenerator._get_modifier_char):
2989         (JSONLayoutResultsGenerator._insert_failure_summaries):
2990         (JSONLayoutResultsGenerator._get_failure_summary_entry):
2991         * Scripts/webkitpy/layout_tests/models/test_run_results.py: Renamed from Tools/Scripts/webkitpy/layout_tests/models/result_summary.py.
2992         (TestRunResults):
2993         (TestRunResults.__init__):
2994         (TestRunResults.add):
2995         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2996         (get_test_results):
2997         (MainTest.test_basic):
2998         * Scripts/webkitpy/layout_tests/views/buildbot_results.py:
2999         (BuildBotPrinter.print_results):
3000         (BuildBotPrinter.print_run_results):
3001         (BuildBotPrinter._print_run_results_entry):
3002         * Scripts/webkitpy/layout_tests/views/buildbot_results_unittest.py:
3003         (BuildBotPrinterTests.get_run_results):
3004         (BuildBotPrinterTests.test_print_unexpected_results.get_unexpected_results):
3005         (BuildBotPrinterTests):
3006         * Scripts/webkitpy/layout_tests/views/printing.py:
3007         (Printer.print_expected):
3008         (Printer._print_expected_results_of_type):
3009         (Printer.print_results):
3010         (Printer._print_timing_statistics):
3011         (Printer._print_worker_statistics):
3012         (Printer._print_aggregate_test_statistics):
3013         (Printer._print_individual_test_times):
3014         (Printer._print_directory_timings):
3015         * Scripts/webkitpy/layout_tests/views/printing_unittest.py:
3016
3017 2012-12-13  Thiago Marcos P. Santos  <thiago.santos@intel.com>
3018
3019         [EFL] The WebKit2 bots are building and executing WebKit1 tests
3020         https://bugs.webkit.org/show_bug.cgi?id=104690
3021
3022         Reviewed by Csaba Osztrogonác.
3023
3024         Pass --no-webkit1 option to the EFL WK2 bots when building. This
3025         will disable WK1 API and thus, the WK1 API tests.
3026
3027         * BuildSlaveSupport/build.webkit.org-config/config.json:
3028         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3029         (CompileWebKit2Only):
3030         (BuildAndTestFactory):
3031         (BuildAndTestFactory.__init__):
3032         (BuildAndTestWebKit2OnlyFactory):
3033         Now bots using BuildAndTestWebKit2OnlyFactory (for instance only EFL) will
3034         get --no-webkit1 as a parameter on the build-webkit script.
3035         * Scripts/build-webkit:
3036         Added support for -no-webkit1 for the Qt and EFL port.
3037
3038 2012-12-13  Eric Seidel  <eric@webkit.org>
3039
3040         build-webkit --chromium-android forces Make even when update-webkit used --ninja
3041         https://bugs.webkit.org/show_bug.cgi?id=104839
3042
3043         Reviewed by Tony Chang.
3044
3045         This was recommended by Peter as part of:
3046         https://bugs.webkit.org/show_bug.cgi?id=104434#c4
3047         Fixing this makes build-webkit --chromium-android work for the common (ninja) case.
3048
3049         * Scripts/webkitdirs.pm:
3050         (buildChromium):
3051
3052 2012-12-13  Julie Parent  <jparent@chromium.org>
3053
3054         Loader_unittests should reset state between runs.
3055         https://bugs.webkit.org/show_bug.cgi?id=104862
3056
3057         Reviewed by Dirk Pranke.
3058
3059         The loader_unittests were not resetting state between runs, leaving
3060         a lot of data behind in global objects that the dashboard and tests
3061         rely on.  A handful were resetting just a few objects.  Use the common
3062         resetGlobals() method instead.
3063
3064         * TestResultServer/static-dashboards/loader_unittests.js:
3065
3066 2012-12-13  Andras Becsi  <andras.becsi@digia.com>
3067
3068         [Qt][WK2] Fix painting on Mac with retina display
3069         https://bugs.webkit.org/show_bug.cgi?id=104574
3070
3071         Reviewed by Kenneth Rohde Christiansen.
3072
3073         Remove setting the devicePixelRatio experimental property
3074         since the value is now automatically picked up from Qt.
3075
3076         * MiniBrowser/qt/qml/BrowserWindow.qml:
3077
3078 2012-12-13  Jussi Kukkonen  <jussi.kukkonen@intel.com>
3079
3080         [EFL][GTK] Don't call deprecated g_type_init when glib => 2.35
3081         https://bugs.webkit.org/show_bug.cgi?id=103209
3082
3083         Reviewed by Kenneth Rohde Christiansen.
3084
3085         g_type_init() is deprecated from 2.35.0 onwards. Don't call it
3086         in that case to avoid warnings and/or build failure.
3087
3088         * DumpRenderTree/gtk/ImageDiff.cpp:
3089         (main):
3090
3091 2012-12-13  Claudio Saavedra  <csaavedra@igalia.com>
3092
3093         [Tools] Add exception for gtk_style_context_get()
3094         https://bugs.webkit.org/show_bug.cgi?id=104896
3095
3096         Reviewed by Martin Robinson.
3097
3098         Add new exception for gtk_style_context_get().
3099
3100         * Scripts/webkitpy/style/checkers/cpp.py:
3101         (check_for_null): Add exception for gtk_style_context_get().
3102         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
3103         (WebKitStyleTest.test_null_false_zero): Add a unit
3104         test for the new exception.
3105
3106 2012-12-13  Joaquim Rocha  <jrocha@igalia.com>
3107
3108         REGRESSION (r137432): The /webkit2/WebKitCookieManager/accept-policy unit test is failing
3109         https://bugs.webkit.org/show_bug.cgi?id=104790
3110
3111         Reviewed by Carlos Garcia Campos.
3112
3113         Unskip the GTK accept policy test.
3114
3115         * Scripts/run-gtk-tests:
3116         (TestRunner):
3117
3118 2012-12-13  Xingnan Wang  <xingnan.wang@intel.com>
3119
3120         Unreviewed. Adding myself to committers list.
3121
3122         * Scripts/webkitpy/common/config/committers.py:
3123
3124 2012-12-12  Roger Fong  <roger_fong@apple.com>
3125
3126         Unreviewed. Fix typo in NRWT.
3127
3128         * Scripts/webkitpy/layout_tests/port/base.py:
3129         (Port.is_reference_html_file):
3130
3131 2012-12-12  Roger Fong  <roger_fong@apple.com>
3132
3133         Fix detection of ref test files in old and new run webkit tests.
3134         https://bugs.webkit.org/show_bug.cgi?id=104850.
3135
3136         Reviewed by Timothy Horton.
3137
3138         * Scripts/old-run-webkit-tests:
3139         (isUsedInReftest): 
3140         Add notref and ref prefix and suffix cases to isUsedinReftest.
3141         * Scripts/webkitpy/layout_tests/port/base.py:
3142         (Port):
3143         (Port.is_reference_html_file): 
3144         Typo in NRWT.
3145
3146 2012-12-12  Dirk Pranke  <dpranke@chromium.org>
3147
3148         rework ChromiumAndroidDriver.start() et. al to avoid funky layering
3149         https://bugs.webkit.org/show_bug.cgi?id=104856
3150
3151         Reviewed by Eric Seidel.
3152
3153         r137523 / bug 10357 introduced some back-and-forth hopping during
3154         driver.start() in the chromium android port. This patch cleans it
3155         up a bit.
3156
3157         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
3158         (ChromiumAndroidDriver.__init__):
3159         (ChromiumAndroidDriver._start_once):
3160         * Scripts/webkitpy/layout_tests/port/chromium_android_unittest.py:
3161         * Scripts/webkitpy/layout_tests/port/driver.py:
3162         (Driver.start):
3163         (Driver._start):
3164
3165 2012-12-12  Eric Seidel  <eric@webkit.org>
3166
3167         Add support for run-perf-tests --chromium-android --profile
3168         https://bugs.webkit.org/show_bug.cgi?id=103587
3169
3170         Reviewed by Dirk Pranke.
3171
3172         This makes it possible to profile PerformanceTests on an Android device
3173         using run-perf-tests --chromium-android --profile.
3174         CPU profiling on Android is sadly non-trivial, but this goes a long
3175         ways towards making it more accessible to non-Android developers.
3176
3177         Although this script requires your device to be flashed with a
3178         userdebug build (an annoyingly non-trivial endeavor for external contributors)
3179         it endeavors to detect and explain the common errors encountered
3180         while trying to setup a device for perf-based profiling.
3181
3182         * Scripts/webkitpy/common/system/profiler.py:
3183         (GooglePProf.profile_after_exit): Discovered that pprof will just hang if passed a missing file.
3184         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
3185         (AndroidPerf):
3186         (AndroidPerf.__init__):
3187         (AndroidPerf.check_configuration):
3188         (AndroidPerf.print_setup_instructions):
3189         (_file_exists_on_device):
3190         (_run_adb_command):
3191         (attach_to_pid):
3192         (_perf_version_string):
3193         (_find_perfhost_binary):
3194         (_perfhost_path):
3195         (_first_ten_lines_of_profile):
3196         (profile_after_exit):
3197         (ChromiumAndroidDriver.__init__):
3198         (ChromiumAndroidDriver.__init__.methods):
3199         (ChromiumAndroidDriver._find_or_create_symfs):
3200         (ChromiumAndroidDriver._start_once):
3201         (ChromiumAndroidDriver):
3202         (ChromiumAndroidDriver._run_post_start_tasks):
3203         (ChromiumAndroidDriver._pid_from_android_ps_output):
3204         (ChromiumAndroidDriver._pid_on_target):
3205         * Scripts/webkitpy/layout_tests/port/chromium_android_unittest.py:
3206         (ChromiumAndroidDriverTest.test_pid_from_android_ps_output):
3207         (AndroidPerfTest):
3208         (AndroidPerfTest.test_perf_output_regexp):
3209         * Scripts/webkitpy/layout_tests/port/driver.py:
3210         (Driver._setup_environ_for_driver):
3211         (Driver):
3212         (Driver._start):
3213         (Driver._run_post_start_tasks):
3214         (Driver._pid_on_target):
3215
3216 2012-12-12  Dirk Pranke  <dpranke@chromium.org>
3217
3218         nrwt: remove deprecated ORWT compatibility command line flags
3219         https://bugs.webkit.org/show_bug.cgi?id=104301
3220
3221         Reviewed by Eric Seidel.
3222
3223         The command line flags for --noshow-results, --[no-]launch-safari,
3224         and --use-remote-links-to-tests were for compatibility with
3225         old-run-webkit-tests, but the compatibility is not needed any more
3226         since NRWT is basically default and the scripts that passed the
3227         old flags have been updated.
3228
3229         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3230         (_set_up_derived_options):
3231         (parse_args):
3232
3233 2012-12-12  John Griggs  <jgriggs@rim.com>
3234
3235         2012-12-12  John Griggs  <jgriggs@rim.com>
3236
3237         [BlackBerry] Update Media Controls
3238         https://bugs.webkit.org/show_bug.cgi?id=104443
3239         https://przilla.ott.qnx.com/bugzilla/show_bug.cgi?id=204748
3240
3241         Update Media Controls for BlackBerry platform to allow audio, embedded video and fullscreen video controls to be positioned and styled differently.
3242
3243         Reviewed by Rob Buis.
3244
3245         BlackBerry-specific update to Media Controls - No new tests required.
3246
3247         * Scripts/webkitperl/FeatureList.pm:
3248
3249 2012-12-12  Justin Novosad  <junov@google.com>
3250
3251         Adding myself to SkiaGraphics watchlist
3252
3253         Unreviewed
3254
3255         * Scripts/webkitpy/common/config/watchlist:
3256
3257 2012-12-12  Martin Robinson  <mrobinson@igalia.com>
3258
3259         [GTK] Add authentication support to DRT and fix exposed issues in the libsoup backend
3260         https://bugs.webkit.org/show_bug.cgi?id=104285
3261
3262         Reviewed by Gustavo Noronha Silva.
3263
3264         Add support to DumpRenderTree for running authentication tests. Since the DRT
3265         expects an authentication callback, we add one to DRTSupport to avoid #ifdefs
3266         in platform-independent code for GTK+.
3267
3268         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3269         (resetDefaultsToConsistentValues): Reset the authentication password and username.
3270         (authenticationCallback): Added.
3271         (createWebView): Attach the authentiation callback during startup.
3272         * DumpRenderTree/gtk/TestRunnerGtk.cpp:
3273         (soupURIToKURL): soup_uri_to_string does not preserve passwords embedded
3274         in the URL so we add a somewhat messy method of re-adding them when they exist.
3275         It would be nice to use soupURIToKURL here, but it seems we cannot use KURL without
3276         pulling in lots of WebCore code so we use string search and replace.
3277         (TestRunner::queueLoad): Use the new helper.
3278
3279 2012-12-12  Dirk Pranke  <dpranke@chromium.org>
3280
3281         garden-o-matic should prefer efl/ over efl-wk1/ and efl-wk2/ when rebaselining
3282         https://bugs.webkit.org/show_bug.cgi?id=102389
3283
3284         Reviewed by Tony Chang.
3285
3286         The code wasn't aware of both the efl-wk1 and efl-wk2 variants,
3287         because all_port_names() by itself doesn't tell you about the
3288         variants. Work around this for now, and filed bug 104761 to
3289         address the underlying problem.
3290
3291         * Scripts/webkitpy/common/checkout/baselineoptimizer.py:
3292         * Scripts/webkitpy/common/checkout/baselineoptimizer_unittest.py:
3293         (BaselineOptimizerTest.test_efl):
3294
3295 2012-12-12  Krzysztof Czech  <k.czech@samsung.com>
3296
3297         [EFL] Possibility to turn off accessibility feature for WebKit-EFL.
3298         https://bugs.webkit.org/show_bug.cgi?id=103036
3299
3300         Reviewed by Laszlo Gombos.
3301
3302         Guard dependencies for accessibility (ATK), so that they can be turned off.
3303
3304         * WebKitTestRunner/InjectedBundle/AccessibilityUIElement.h:
3305         (AccessibilityUIElement):
3306         * WebKitTestRunner/PlatformEfl.cmake:
3307
3308 2012-12-12  Simon Fraser  <simon.fraser@apple.com>
3309
3310         Add some entries in the watchlist, and register me for them.
3311
3312         * Scripts/webkitpy/common/config/watchlist:
3313
3314 2012-12-12  Ilya Tikhonovsky  <loislo@chromium.org>
3315
3316         Web Inspector: Native Memory Instrumentation: remove fake root MemoryObjectInfo.
3317         https://bugs.webkit.org/show_bug.cgi?id=104796
3318
3319         Reviewed by Yury Semikhatsky.
3320
3321         It was not a good idea to introduce a fake root MemoryObjectInfo.
3322         It makes a problem when we visit an object without its own MemoryObjectType.
3323
3324         Example: RenderBox has a global pointer to a hash map.
3325         HashMap doesn't have its own object type because it is a generic container.
3326         It will inherit object type from the fake root memory object info.
3327         The same could happen for another container in another class with other MemoryObjectType.
3328
3329         This fact forces me to create custom process method for root objects
3330         because they need to have their own MemoryObjectInfo with customisable memory object type.
3331
3332         * TestWebKitAPI/Tests/WTF/MemoryInstrumentationTest.cpp:
3333
3334         Drive by fix: InstrumentedPointer* was replaced with Wrapper* because actually it is using
3335         for instrumented and not instrumented object classes.
3336
3337 2012-12-12  Zan Dobersek  <zandobersek@gmail.com>
3338
3339         [GTK] Remove the last of unnecessary configuration options in configure.ac
3340         https://bugs.webkit.org/show_bug.cgi?id=104793
3341
3342         Reviewed by Martin Robinson.
3343
3344         Removing the filters and indexed database features from the list of features
3345         being configurable through the configure script. This is not the case anymore
3346         as the corresponding configuration options are being removed.
3347
3348         * Scripts/webkitdirs.pm:
3349         (buildAutotoolsProject):
3350
3351 2012-12-12  Mario Sanchez Prada  <mario@webkit.org>
3352
3353         [GTK][WK2] New API to detect display/execution of insecure content
3354         https://bugs.webkit.org/show_bug.cgi?id=104578
3355
3356         Reviewed by Martin Robinson.
3357
3358         Add coding style exception for 'readability/enum_casing' in GTK.
3359
3360         The GTK+ API use upper case, underscore separated, words in
3361         certain types of enums (e.g. signals, properties).
3362
3363         * Scripts/webkitpy/style/checker.py: Added exception, both for
3364         WebKitGTK+ and WebKit2GTK+.
3365
3366 2012-12-12  Jochen Eisinger  <jochen@chromium.org>
3367
3368         [chromium] Fix build on chromium-win
3369
3370         Unreviewed build-fix.
3371
3372         MSVS indeed doesn't have snprintf, adding wtf/StringExtras.h include
3373         to work around that.
3374
3375         * DumpRenderTree/chromium/TestRunner/src/WebTestPlugin.cpp:
3376
3377 2012-12-12  Simon Hausmann  <simon.hausmann@digia.com>
3378
3379         [Qt] Fix build on Mac
3380
3381         Reviewed by Csaba Osztrogonác.
3382
3383         On Mac OS X we have sqlite as a system library available.
3384
3385         * qmake/mkspecs/features/features.prf:
3386
3387 2012-12-12  Zan Dobersek  <zandobersek@gmail.com>
3388
3389         Unreviewed GTK gardening.
3390
3391         Skipping another unit test that regressed in r137432.
3392
3393         * Scripts/run-gtk-tests:
3394         (TestRunner):
3395
3396 2012-12-12  Simon Hausmann  <simon.hausmann@digia.com>
3397
3398         [Qt] Fix linkage against SQLite
3399         https://bugs.webkit.org/show_bug.cgi?id=104781
3400
3401         Reviewed by Csaba Osztrogonác.
3402
3403         Detect SQLite either through pkg-config, from the SQLITE3SRCDIR
3404         environment variable or from qtbase via a qt5.git build. Bail out early
3405         with an error message if neither option applies.
3406
3407         This removes the usage of $$QT.core.sources and the dependency on Qt's
3408         system-sqlite configure setting. We should always favour a system
3409         library over a copy in Qt.
3410
3411         * qmake/mkspecs/features/features.prf:
3412
3413 2012-12-12  Tommy Widenflycht  <tommyw@google.com>
3414
3415         MediaStream API: Change the data channel descriptor pattern to a handler pattern
3416         https://bugs.webkit.org/show_bug.cgi?id=104543
3417
3418         Reviewed by Kent Tamura.
3419
3420         Adding mocks to be able to test this refactoring.
3421
3422         * DumpRenderTree/DumpRenderTree.gypi:
3423         * DumpRenderTree/chromium/MockWebRTCDataChannelHandler.cpp: Added.
3424         (DataChannelReadyStateTask):
3425         (DataChannelReadyStateTask::DataChannelReadyStateTask):
3426         (MockWebRTCDataChannelHandler::MockWebRTCDataChannelHandler):
3427         (MockWebRTCDataChannelHandler::setClient):
3428         (MockWebRTCDataChannelHandler::bufferedAmount):
3429         (MockWebRTCDataChannelHandler::sendStringData):
3430         (MockWebRTCDataChannelHandler::sendRawData):
3431         (MockWebRTCDataChannelHandler::close):
3432         * DumpRenderTree/chromium/MockWebRTCDataChannelHandler.h: Added.
3433         (MockWebRTCDataChannelHandler):
3434         (MockWebRTCDataChannelHandler::taskList):
3435         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.cpp:
3436         (RTCPeerConnectionReadyStateTask::RTCPeerConnectionReadyStateTask):
3437         (RemoteDataChannelTask::RemoteDataChannelTask):
3438         (MockWebRTCPeerConnectionHandler::createDataChannel):
3439         (MockWebRTCPeerConnectionHandler::stop):
3440         * DumpRenderTree/chromium/MockWebRTCPeerConnectionHandler.h:
3441         (MockWebRTCPeerConnectionHandler):
3442
3443 2012-12-12  Jochen Eisinger  <jochen@chromium.org>
3444
3445         [chromium] WebTestPlugin must not depend on WTF::String
3446         https://bugs.webkit.org/show_bug.cgi?id=104783
3447
3448         Reviewed by Kent Tamura.
3449
3450         In a component build, we can't access the WTF that is part of the
3451         WebKit component. Accessing WTF instead will pull in a separate copy
3452         into the TestRunner library. Instead of WTF::String::number use
3453         snprintf() to print numbers.
3454
3455         * DumpRenderTree/chromium/TestRunner/src/WebTestPlugin.cpp:
3456
3457 2012-12-12  Jocelyn Turcotte  <jocelyn.turcotte@digia.com>
3458
3459         [Qt] Do not automatically enable force_static_libs_as_shared when using debug_and_release
3460         https://bugs.webkit.org/show_bug.cgi?id=104562
3461
3462         Reviewed by Tor Arne Vestbø.
3463
3464         force_static_libs_as_shared creates issues on Mac if QtWebKit is built as a framework.
3465         QtWebProcess, for example, will link to the framework's release library and will rely
3466         on DYLD_IMAGE_SUFFIX to switch to the debug build. Since QtWebProcess will also links
3467         directly to the debug build of WebCore, and that QtWebKit release will link in the
3468         release build of WebCore, both the debug and release version will be loaded.
3469
3470         To fix this properly we should make sure that applications always link to the release
3471         build of dependent libraries by bypassing resolveFinalLibraryName.
3472         Disable it for now as force_static_libs_as_shared is mainly for development while
3473         debug_and_release is mainly for production builds.
3474
3475         * qmake/mkspecs/features/default_post.prf:
3476
3477 2012-12-12  Zan Dobersek  <zandobersek@gmail.com>
3478
3479         Unreviewed GTK gardening of the unit tests.
3480
3481         Skip two unit tests that regressed during the hackfest.
3482
3483         * Scripts/run-gtk-tests:
3484         (TestRunner):
3485
3486 2012-12-11  Ryosuke Niwa  <rniwa@webkit.org>
3487
3488         Add Koji Ishii to the list of contributors.
3489
3490         * Scripts/webkitpy/common/config/committers.py:
3491
3492 2012-12-11  Dominic Mazzoni  <dmazzoni@google.com>
3493
3494         AX: Make isActionSupported cross-platform.
3495         https://bugs.webkit.org/show_bug.cgi?id=104681
3496
3497         Reviewed by Chris Fleizach.
3498
3499         Replace isActionSupported, which takes a platform-specific
3500         argument, with three cross-platform methods.
3501
3502         * DumpRenderTree/AccessibilityUIElement.cpp:
3503         (isPressActionSupportedCallback):
3504         (isIncrementActionSupportedCallback):
3505         (isDecrementActionSupportedCallback):
3506         (AccessibilityUIElement::getJSClass):
3507         * DumpRenderTree/AccessibilityUIElement.h:
3508         (AccessibilityUIElement):
3509         * DumpRenderTree/blackberry/AccessibilityUIElementBlackBerry.cpp:
3510         (AccessibilityUIElement::isPressActionSupported):
3511         (AccessibilityUIElement::isIncrementActionSupported):
3512         (AccessibilityUIElement::isDecrementActionSupported):
3513         * DumpRenderTree/chromium/TestRunner/src/AccessibilityUIElementChromium.cpp:
3514         (WebTestRunner::AccessibilityUIElement::AccessibilityUIElement):
3515         (WebTestRunner::AccessibilityUIElement::isPressActionSupportedCallback):
3516         (WebTestRunner):
3517         (WebTestRunner::AccessibilityUIElement::isIncrementActionSupportedCallback):
3518         (WebTestRunner::AccessibilityUIElement::isDecrementActionSupportedCallback):
3519         (WebTestRunner::AccessibilityUIElement::incrementCallback):
3520         (WebTestRunner::AccessibilityUIElement::decrementCallback):
3521         (WebTestRunner::AccessibilityUIElement::pressCallback):
3522         * DumpRenderTree/chromium/TestRunner/src/AccessibilityUIElementChromium.h:
3523         (AccessibilityUIElement):
3524         * DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
3525         (AccessibilityUIElement::isPressActionSupported):
3526         (AccessibilityUIElement::isIncrementActionSupported):
3527         (AccessibilityUIElement::isDecrementActionSupported):
3528         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
3529         (AccessibilityUIElement::isPressActionSupported):
3530         (AccessibilityUIElement::isIncrementActionSupported):
3531         (AccessibilityUIElement::isDecrementActionSupported):
3532         * DumpRenderTree/win/AccessibilityUIElementWin.cpp:
3533         (AccessibilityUIElement::isPressActionSupported):
3534         (AccessibilityUIElement::isIncrementActionSupported):
3535         (AccessibilityUIElement::isDecrementActionSupported):
3536         * WebKitTestRunner/InjectedBundle/AccessibilityUIElement.cpp:
3537         (WTR::AccessibilityUIElement::isPressActionSupported):
3538         (WTR::AccessibilityUIElement::isIncrementActionSupported):
3539         (WTR::AccessibilityUIElement::isDecrementActionSupported):
3540         * WebKitTestRunner/InjectedBundle/AccessibilityUIElement.h:
3541         (AccessibilityUIElement):
3542         * WebKitTestRunner/InjectedBundle/Bindings/AccessibilityUIElement.idl:
3543         * WebKitTestRunner/InjectedBundle/mac/AccessibilityUIElementMac.mm:
3544         (WTR::AccessibilityUIElement::isPressActionSupported):
3545         (WTR):
3546         (WTR::AccessibilityUIElement::isIncrementActionSupported):
3547         (WTR::AccessibilityUIElement::isDecrementActionSupported):
3548
3549 2012-12-11  Dirk Pranke  <dpranke@chromium.org>
3550
3551         nrwt: remove result_summary.total_tests_by_expectation
3552         https://bugs.webkit.org/show_bug.cgi?id=104756
3553
3554         Reviewed by Eric Seidel.
3555
3556         This field wasn't being used by anything.
3557
3558         * Scripts/webkitpy/layout_tests/models/result_summary.py:
3559         (ResultSummary.__init__):
3560
3561 2012-12-11  Dirk Pranke  <dpranke@chromium.org>
3562
3563         garden-o-matic doesn't know about reftests
3564         https://bugs.webkit.org/show_bug.cgi?id=101976
3565
3566         Reviewed by Ojan Vafai.
3567
3568         This is a first attempt at fixing this, sort of a "stop-the-bleeding"
3569         fix. We will now look at the list of tests we're asked to rebaseline,
3570         and log which ones are reftests into the status area, and rebaseline
3571         the rest.
3572
3573         Longer-term, we should display feedback in the results area based on
3574         which tests we're looking at. If the test list contains no reftests,
3575         we should behave as today; if the list is all reftest, we should
3576         probably disable the button, and if the list is a mix, we should
3577         probably display an alert next to the button or something. It would
3578         also be nice to annotate which tests were reftests, maybe with a
3579         "(reftest)" next to the test name or something.
3580
3581         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/controllers.js:
3582         (.):
3583         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results.js:
3584         (.):
3585         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/results_unittests.js:
3586
3587 2012-12-11  Dirk Pranke  <dpranke@chromium.org>
3588
3589         REGRESSION(r133380): new tests without expected file reported as failing tests on the bots
3590         https://bugs.webkit.org/show_bug.cgi?id=101469
3591
3592         Reviewed by Ryosuke Niwa.
3593
3594         Fix parsing of NRWT's output to identify missing results separately
3595         from other regressions.
3596
3597         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3598         (RunWebKitTests._parseNewRunWebKitTestsOutput):
3599         * BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:
3600         (MasterCfgTest.test_nrwt_leaks_parsing):
3601         (test_nrwt_missing_results):
3602
3603 2012-12-11  Dirk Pranke  <dpranke@chromium.org>
3604
3605         buildbot: change run-webkit-tests --no-launch-safari to --no-show-results
3606         https://bugs.webkit.org/show_bug.cgi?id=104299
3607
3608         Reviewed by Ryosuke Niwa.
3609
3610         Since --no-show-results is the more generically-named flag that NRWT uses.
3611
3612         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3613         (RunWebKitTests):
3614
3615 2012-12-11  Dirk Pranke  <dpranke@chromium.org>
3616
3617         remove --use-remote-links-to-tests from buildbot configs (except apple win)
3618         https://bugs.webkit.org/show_bug.cgi?id=104291
3619
3620         Reviewed by Csaba Osztrogonác.
3621
3622         This flag is ignored by NRWT and was just implemented for compatibility.
3623         ORWT still uses it, though.
3624
3625         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3626         (RunWebKitTests):
3627         (RunWebKitTests.start):
3628
3629 2012-12-11  Anders Carlsson  <andersca@apple.com>
3630
3631         ASSERT running platform/mac/plugins/convert-point.html on the WK2 bots
3632         https://bugs.webkit.org/show_bug.cgi?id=104735
3633
3634         Reviewed by Jon Honeycutt.
3635
3636         Do the testing from within NPP_New instead of NPN_SetWindow. Plug-ins aren't supposed to do any real work inside NPP_SetWindow,
3637         and doing so will more easily trigger the assertion failure. Ideally we should get rid of the initial sync painting pass, but this
3638         gets the test passing again!
3639
3640         * DumpRenderTree/TestNetscapePlugIn/Tests/mac/ConvertPoint.cpp:
3641         (ConvertPoint::NPP_New):
3642
3643 2012-12-11  Julie Parent  <jparent@chromium.org>
3644
3645         Dashboard unit tests should not make xhrs.
3646         https://bugs.webkit.org/show_bug.cgi?id=104598
3647
3648         Reviewed by Dirk Pranke.
3649
3650         The unit test were requesting 25 files that they did not use.
3651         Stub out loader.request to avoid making the xhrs.
3652         Result: order of magnitude speed improvement in test run time (from ~1600ms to 160ms on my machine).
3653         
3654         * TestResultServer/static-dashboards/run-embedded-unittests.html:
3655         * TestResultServer/static-dashboards/run-unittests.html:
3656
3657 2012-12-11  Filip Pizlo  <fpizlo@apple.com>
3658
3659         Profiler should show bytecode dumps as they would have been visible to the JITs, including the profiling data that the JITs would see
3660         https://bugs.webkit.org/show_bug.cgi?id=104647
3661
3662         Reviewed by Oliver Hunt.
3663
3664         Added a "profiling" (or "p") command to show the profiling data that the JITs saw
3665         for each JIT compilation of a code block.
3666         
3667         Also added instruction counts in the "full" display and made the "full" display the
3668         default thing you see.
3669
3670         * Scripts/display-profiler-output:
3671
3672 2012-12-11  Eric Seidel  <eric@webkit.org>
3673
3674         Unreviewed, rolling out r137371.
3675         http://trac.webkit.org/changeset/137371
3676         https://bugs.webkit.org/show_bug.cgi?id=104434
3677
3678         Various scripts are not ready for out/ as the build directory,
3679         this can't work as written.
3680
3681         * Scripts/update-webkit:
3682         * Scripts/webkitdirs.pm:
3683         (buildChromium):
3684
3685 2012-12-11  Zan Dobersek  <zandobersek@gmail.com>
3686
3687         Unreviewed GTK build functionality fix.
3688
3689         Turns out the CSP feature was not enabled, that's why all the tests were failing.
3690
3691         * Scripts/webkitperl/FeatureList.pm:
3692
3693 2012-12-11  Eric Seidel  <eric@webkit.org>
3694
3695         Make Ninja the default build system for build-webkit --chromium
3696         https://bugs.webkit.org/show_bug.cgi?id=104434
3697
3698         Reviewed by Daniel Bates.
3699
3700         As far as I can tell, most Chromium developers are now using Ninja
3701         as their default build system instead of the native Xcode/Visual Studio/Make.
3702         This change makes Ninja the default for developers as well as all Chromium webkit bots.
3703
3704         Developers can pass --no-ninja to get the platform-native buildsystem instead of Ninja.
3705         It's possible we may wish to make this a per-platform default, and have
3706         Ninja only be the default for Mac/Linux.
3707
3708         There was a thread on webkit-dev discussing this change:
3709         http://lists.webkit.org/pipermail/webkit-dev/2012-December/023113.html
3710
3711         * Scripts/update-webkit:
3712
3713 2012-12-11  Dirk Pranke  <dpranke@chromium.org>
3714
3715         nrwt: record the order and run times of all of the tests in a single file
3716         https://bugs.webkit.org/show_bug.cgi?id=104184
3717
3718         Reviewed by Ryosuke Niwa.
3719
3720         Replace the N tests_run*.txt files that we output into
3721         layout-test-results with a single packed JSON trie that contains
3722         the test ordering and run statistics. This can eventually be uploaded
3723         to the TestResultsServer to replace the times.json file we currently
3724         upload as well, and will let us replay the results of a previous
3725         run fairly deterministically.
3726
3727         * Scripts/webkitpy/layout_tests/controllers/layout_test_runner.py:
3728         (Worker.__init__):
3729         (Worker.start):
3730         (Worker._run_test):
3731         (Worker.stop):
3732         (Worker._clean_up_after_test):
3733         * Scripts/webkitpy/layout_tests/controllers/manager.py:
3734         (Manager._upload_json_files):
3735         (Manager._copy_results_html_file):
3736         (Manager):
3737         (Manager._stats_trie):
3738         (Manager._stats_trie._worker_number):
3739         * Scripts/webkitpy/layout_tests/models/test_results.py:
3740         (TestResult.__init__):
3741         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
3742         (MainTest.test_missing_and_unexpected_results):
3743         (MainTest.test_retrying_and_flaky_tests):
3744         (EndToEndTest.test_end_to_end):
3745         (EndToEndTest.test_reftest_with_two_notrefs):
3746         (RebaselineTest.test_reset_results):
3747         (RebaselineTest.test_missing_results):
3748         (RebaselineTest.test_new_baseline):
3749
3750 2012-12-11  Dirk Pranke  <dpranke@chromium.org>
3751
3752         check-webkit-style complaining about TestExpectations lines a patch didn't modify
3753         https://bugs.webkit.org/show_bug.cgi?id=104687
3754
3755         Reviewed by Ojan Vafai.
3756
3757         Way back in bug 60466 / r86165, we changed the style checker
3758         for TestExpectations to ensure that the entire file linted, not
3759         just the lines modified by a patch, in order to ensure that a
3760         new line didn't cause errors elsewhere in the file (e.g., by
3761         introducing duplicate lines). At the time, errors in the
3762         expectations files were fatal and we didn't have a separate step
3763         on the bots that would lint the files to ensure things were still
3764         okay, so this was probably the right thing to do, but it could
3765         cause some patches to fail the style check even though the patches
3766         themselves were fine.
3767
3768         However, now errors aren't always fatal, and we have a separate build
3769         step to lint the files. So, we should probably not punish the
3770         patches and the style checker should filter out "unrelated" errors.
3771
3772         In theory we should be comparing the output pre- and post-patch
3773         to filter out pre-existing errors, but I don't think the style
3774         checker has that ability.
3775
3776         * Scripts/webkitpy/style/checkers/test_expectations.py:
3777         (TestExpectationsChecker.__init__):
3778         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
3779         (TestExpectationsTestCase.assert_lines_lint):
3780
3781 2012-12-11  Dirk Pranke  <dpranke@chromium.org>
3782
3783         nrwt isn't rejecting unrecognized expectations
3784         https://bugs.webkit.org/show_bug.cgi?id=104701
3785
3786         Reviewed by Ojan Vafai.
3787
3788         NRWT is silently ignoring unrecognized expectations like "Text"
3789         and "Audio" (largely because I haven't fully removed support for
3790         the old syntax and keywords and restructured the code to do the
3791         proper checking). Temporarily change the tokenizer to reject
3792         unrecognized expectations until I can make those changes.
3793
3794         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
3795         (TestExpectationParser._tokenize_line):
3796         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
3797         (ExpectationSyntaxTests.test_unknown_expectation):
3798         (TestExpectationSerializationTests.disabled_test_string_whitespace_stripping):
3799
3800 2012-12-11  Martin Robinson  <mrobinson@igalia.com>
3801
3802         Fix a typo in my previous commit that is causing test failures.
3803
3804         Reviewed by Gustavo Noronha Silva.
3805
3806         There was an extra bit of statement appended after the end of a conditional. Somehow
3807         this compiled and caused test failures.
3808
3809         * DumpRenderTree/gtk/EventSender.cpp:
3810         (continuousMouseScrollByCallback):
3811
3812 2012-12-11  Martin Robinson  <mrobinson@igalia.com>
3813
3814         [GTK] Spurious stderr output from the event sender
3815         https://bugs.webkit.org/show_bug.cgi?id=104678
3816
3817         Reviewed by Gustavo Noronha Silva.
3818
3819         Eliminate some spurious stderr output from the EventSender by using an early return
3820         instead of an assertion that prints a warning to the console.
3821
3822         * DumpRenderTree/gtk/EventSender.cpp:
3823         (continuousMouseScrollByCallback): We don't support continuous scroll events that
3824         move by page, but we should just fail silently instead of printing an error.
3825
3826 2012-12-11  Joone Hur  <joone.hur@intel.com>
3827
3828         [GTK] Add GraphicsLayerActor
3829         https://bugs.webkit.org/show_bug.cgi?id=91940
3830
3831         Reviewed by Gustavo Noronha Silva.
3832
3833         Add GraphicsLayerActor to the list of classes that contain GObjects
3834         in the style checker script so that it does not complain about GObject
3835         coding style.
3836
3837         * Scripts/webkitpy/style/checker.py:
3838
3839 2012-12-11  Oswald Buddenhagen  <oswald.buddenhagen@digia.com>
3840
3841         [Qt] remove hack which avoids unused .obj directories all over the place
3842
3843         Reviewed by Simon Hausmann.
3844
3845         The feature is subject to upstreaming in Qt itself in
3846         https://codereview.qt-project.org/#change,42011
3847
3848         * qmake/mkspecs/features/default_post.prf:
3849
3850 2012-12-11  Oswald Buddenhagen  <oswald.buddenhagen@digia.com>
3851
3852         [Qt] move SQLITE3SRCDIR assignment to only use site
3853
3854         Reviewed by Simon Hausmann.
3855
3856         Target.pri includes WebCore.pri the almost first thing, so that is
3857         covered as well.
3858
3859         * qmake/mkspecs/features/default_pre.prf:
3860
3861 2012-12-11  Oswald Buddenhagen  <oswald.buddenhagen@digia.com>
3862
3863         [Qt] delete dead include paths
3864
3865         Reviewed by Simon Hausmann.
3866
3867         followup to https://bugs.webkit.org/show_bug.cgi?id=93446
3868
3869         * DumpRenderTree/qt/DumpRenderTree.pro:
3870
3871 2012-12-11  Martin Robinson  <mrobinson@igalia.com>
3872
3873         [GTK] Implement some missing FrameLoaderClient message in DRT
3874         https://bugs.webkit.org/show_bug.cgi?id=104444
3875
3876         Reviewed by Gustavo Noronha Silva.
3877
3878         Install a FrameLoadEvent callback that prints the necessary output for some
3879         missing FrameLoader events.
3880
3881         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3882         (frameLoadEventCallback): Added.
3883         (createWebView): Install the FrameLoader event callback.
3884
3885 2012-12-11  Oswald Buddenhagen  <oswald.buddenhagen@digia.com>
3886
3887         [Qt] change setup of QMAKE_MOC
3888
3889         Reviewed by Simon Hausmann.
3890
3891         don't load(moc) explicitly - turns out that this breaks
3892         debug_and_release builds (the feature does not like being loaded before
3893         default_post has been loaded).
3894         so instead just do a minimal call to find the moc executable.
3895
3896         as it's in fact only two files which need it, so de-centralize this call
3897         for the sake of saving some cpu cycles spent in qmake.
3898
3899         * WebKitTestRunner/InjectedBundle/DerivedSources.pri:
3900         * qmake/mkspecs/features/default_pre.prf:
3901
3902 2012-12-11  Jocelyn Turcotte  <jocelyn.turcotte@digia.com>
3903
3904         [Qt] Fix the linking of debug builds on Mac with force_static_libs_as_shared
3905         https://bugs.webkit.org/show_bug.cgi?id=104559
3906
3907         Reviewed by Tor Arne Vestbø.
3908
3909         "contains(libraries, $$TARGET): return(true)" fails for intermediate libraries since
3910         at this point TARGET may contain the "_debug" suffix.
3911
3912         Fix it by moving the TARGET adjustment later in the file.
3913
3914         * qmake/mkspecs/features/default_post.prf:
3915
3916 2012-12-11  Martin Robinson  <mrobinson@igalia.com>
3917
3918         [Coverity] [GTK] Remove some redundant null checks in DumpRenderTree
3919         https://bugs.webkit.org/show_bug.cgi?id=104570
3920
3921         Reviewed by Daniel Bates.
3922
3923         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3924         (descriptionSuitableForTestResult): Remove a redundant null check.
3925
3926 2012-12-11  Alberto Garcia  <agarcia@igalia.com>
3927
3928         Remove unused GErrors
3929         https://bugs.webkit.org/show_bug.cgi?id=104657
3930
3931         Reviewed by Martin Robinson.
3932
3933         Replace unused GErrors with null pointers.
3934
3935         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3936         (initializeFonts):
3937         * DumpRenderTree/gtk/EditingCallbacks.cpp:
3938         (dumpRange):
3939         * WebKitTestRunner/InjectedBundle/gtk/ActivateFontsGtk.cpp:
3940         (WTR::inititializeFontConfigSetting):
3941
3942 2012-12-11  Martin Robinson  <mrobinson@igalia.com>
3943
3944         [GTK] Bump the jhbuild libsoup version to fix an authentication issue
3945         https://bugs.webkit.org/show_bug.cgi?id=104448
3946
3947         Reviewed by Gustavo Noronha Silva.
3948
3949         Bump the libsoup version to one from the git repository. This requires updating
3950         GLib as well which in turn requires adjusting the CFLAGS we pass when compiling
3951         the scanobj step of the gtkdoc generation.
3952
3953         * gtk/generate-gtkdoc:
3954         * gtk/jhbuild.modules:
3955
3956 2012-12-11  Zan Dobersek  <zandobersek@gmail.com>
3957
3958         Unreviewed functionality fix for the GTK port.
3959
3960         Enable the MicroData feature support for the GTK port in the FeatureList module.
3961         This feature listing now again affects which features are enabled and which disabled
3962         on the development builds of the GTK port.
3963
3964         * Scripts/webkitperl/FeatureList.pm:
3965
3966 2012-12-11  Jinwoo Song  <jinwoo7.song@samsung.com>
3967
3968         [EFL][WK2] MiniBrowser is not built
3969         https://bugs.webkit.org/show_bug.cgi?id=104629
3970
3971         Reviewed by Gyuyoung Kim.
3972
3973         As we are finding an elementary package in Tools/MiniBrowser/efl/CMakeList.txt since r137203,
3974         we need to check if the package is found in the same place not in the Tools/CMakeList.txt.
3975
3976         * CMakeLists.txt:
3977         * MiniBrowser/efl/CMakeLists.txt:
3978
3979 2012-12-11  Xabier Rodriguez Calvar  <calvaris@igalia.com>
3980
3981         [GTK][jhbuild] Switch to GStreamer 1.0 build
3982         https://bugs.webkit.org/show_bug.cgi?id=91727
3983
3984         Reviewed by Philippe Normand.
3985
3986         Switch build-webkit --gtk to GStreamer 1.0 support and build the
3987         necessary GStreamer git modules from JHBuild.
3988
3989         * gtk/jhbuild.modules: Added GStreamer 1.0.3 build support.
3990
3991 2012-12-11  Zan Dobersek  <zandobersek@gmail.com>
3992
3993         [GTK] Feature enabling/disabling should be possible through build-webkit
3994         https://bugs.webkit.org/show_bug.cgi?id=99271
3995
3996         Reviewed by Gustavo Noronha Silva.
3997
3998         Refactor the code in webkitdirs.pm that builds an autotools project to generate
3999         autogen.sh arguments, make arguments and installation prefix from the passed-in
4000         parameters rather than generating all of that in build-webkit. The autogen.sh
4001         arguments now contain only enable/disable flags for options that are actually
4002         configurable in configure.ac, the flag value actually reflecting whether the feature
4003         is enabled or disabled in the feature list.
4004
4005         Other features are overridable through modifying the GNUmakefile.features.am file.
4006         All these features are now stored in the build directory and upon change trigger
4007         a rerun of the autogen.sh script, pretty much like the autogen.sh arguments do.
4008
4009         The override-feature-defines script is called by autogen.sh, before GNUmakefile.in is
4010         generated when calling autoreconf. Its task is to copy the GNUmakefile.features.am.in
4011         into GNUmakefile.features.am and modify it if there's a text file in the build directory
4012         that contains all the feature defines the build-webkit script has written. If the build
4013         is not done through build-webkit (and there's no file in build directory listing all the
4014         feature defines) the script does not advance further from copying.
4015
4016         This approach is taken to overcome the rigidness of the automake system as it's impossible
4017         to effectively generate and use GNUmakefile.features.am or even GNUmakefile.features file
4018         after the autoreconf command execution in autogen.sh.
4019
4020         * Scripts/build-jsc:
4021         (buildMyProject):
4022         * Scripts/build-webkit:
4023         * Scripts/webkitdirs.pm:
4024         (runAutogenForAutotoolsProjectIfNecessary):
4025         (mustReRunAutogen):
4026         (buildAutotoolsProject):
4027         (buildGtkProject):
4028         * Scripts/webkitperl/FeatureList.pm: Update the features that are currently enabled in
4029         Source/WebCore/GNUmakefile.features.am.in but aren't in the feature list.
4030         * Scripts/webkitpy/style/checker.py: Source/WebCore/GNUmakefile.features.am.in is recognized
4031         as a text file and tabulation errors are reported. Skip the file to suppress them.
4032         * gtk/override-feature-defines: Added.
4033         (copy_feature_defines_makefile):
4034         (adjust_feature_defines_makefile):
4035         (adjust_feature_defines_makefile.override_feature_define):
4036         (override_feature_defines):
4037
4038 2012-12-10  Yury Semikhatsky  <yurys@chromium.org>
4039
4040         Memory instrumentation: make sure each edge is reported only once
4041         https://bugs.webkit.org/show_bug.cgi?id=104630
4042
4043         Reviewed by Pavel Feldman.
4044
4045         Test that outgoing edges are reported only once if we come to the same
4046         object twice: first by an address of a base class and then by a real address
4047         of the object.
4048
4049         * TestWebKitAPI/Tests/WTF/MemoryInstrumentationTest.cpp:
4050
4051 2012-12-10  Dana Jansens  <danakj@chromium.org>
4052
4053         [chromium] Remove WebCompositorOutputSurface implementations from DRT
4054         https://bugs.webkit.org/show_bug.cgi?id=104046
4055
4056         Reviewed by Darin Fisher.
4057
4058         The classes are now provided through WebCompositorSupport.
4059
4060         * DumpRenderTree/DumpRenderTree.gypi:
4061         * DumpRenderTree/chromium/WebViewHost.cpp:
4062         (WebViewHost::createOutputSurface):
4063         * DumpRenderTree/chromium/WebViewHostOutputSurface.cpp: Removed.
4064         * DumpRenderTree/chromium/WebViewHostOutputSurface.h: Removed.
4065         * DumpRenderTree/chromium/WebViewHostSoftwareOutputDevice.cpp: Removed.
4066         * DumpRenderTree/chromium/WebViewHostSoftwareOutputDevice.h: Removed.
4067
4068 2012-12-10  Dirk Pranke  <dpranke@chromium.org>
4069
4070         change --no-launch-safari to --no-show-results in scripts
4071         https://bugs.webkit.org/show_bug.cgi?id=104298
4072
4073         Reviewed by Eric Seidel.
4074
4075         Since Safari only exists on the Apple ports, NRWT calls the
4076         command line flag --[no-]show-results instead. This updates
4077         the scripts to use the new flag (but not the buildbot config yet).
4078
4079         * Scripts/generate-coverage-data:
4080         * Scripts/webkit-tools-completion.sh:
4081         * Scripts/webkitpy/tool/steps/runtests.py:
4082         (RunTests.run):
4083         * Scripts/webkitpy/tool/steps/runtests_unittest.py:
4084
4085 2012-12-10  Dirk Pranke  <dpranke@chromium.org>
4086
4087         nrwt: fix mapping of AUDIO onto FAIL when printing unexpected results
4088         https://bugs.webkit.org/show_bug.cgi?id=104597
4089
4090         Reviewed by Ojan Vafai.
4091
4092         This was broken in the recent printer refactoring.
4093
4094         * Scripts/webkitpy/layout_tests/views/buildbot_results.py:
4095         (BuildBotPrinter.print_unexpected_results.add_result.is_expected):
4096         * Scripts/webkitpy/layout_tests/views/buildbot_results_unittest.py:
4097         (BuildBotPrinterTests.get_result):
4098         (BuildBotPrinterTests.test_print_unexpected_results):
4099         (BuildBotPrinterTests.test_print_unexpected_results.get_unexpected_results):
4100         (BuildBotPrinterTests):
4101
4102 2012-12-10  Dirk Pranke  <dpranke@chromium.org>
4103
4104         old-run-webkit-tests: add --show-results properly
4105         https://bugs.webkit.org/show_bug.cgi?id=104608
4106
4107         Reviewed by Eric Seidel.
4108
4109         Rework r137219 / bug 104601 to use the correct perl idiom.
4110
4111         * Scripts/old-run-webkit-tests:
4112
4113 2012-12-10  Aaron Colwell  <acolwell@chromium.org>
4114
4115         Unreviewed. Adding Aaron Colwell to the contributor list.
4116
4117         * Scripts/webkitpy/common/config/committers.py:
4118
4119 2012-12-10  Anders Carlsson  <andersca@apple.com>
4120
4121         Remove WebKit2 tests on Windows
4122         https://bugs.webkit.org/show_bug.cgi?id=104607
4123
4124         Reviewed by Tim Horton.
4125
4126         Stop building WebKit2 tests and remove associated files.
4127
4128         * TestWebKitAPI/win/InjectedBundleControllerWin.cpp: Removed.
4129         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp: Removed.
4130         * TestWebKitAPI/win/PlatformWebViewWin.cpp: Removed.
4131         * TestWebKitAPI/win/TestWebKitAPI.sln: Removed.
4132         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
4133         * TestWebKitAPI/win/TestWebKitAPIGenerated.vcproj: Removed.
4134         * TestWebKitAPI/win/TestWebKitAPIGeneratedCommon.vsprops: Removed.
4135         * TestWebKitAPI/win/TestWebKitAPIInjectedBundle.vcproj: Removed.
4136         * TestWebKitAPI/win/WindowMessageObserver.h: Removed.
4137         * TestWebKitAPI/win/copy-resources.cmd: Removed.
4138
4139 2012-12-10  Dirk Pranke  <dpranke@chromium.org>
4140
4141         old-run-webkit-tests: support --no-show-results for compatibility w/ NRWT
4142         https://bugs.webkit.org/show_bug.cgi?id=104601
4143
4144         Reviewed by Eric Siedel.
4145
4146         This flag is being added so we can move all of the calling scripts
4147         over to using --no-show-results instead of --no-launch-safari.
4148
4149         * Scripts/old-run-webkit-tests:
4150
4151 2012-12-10  Dirk Pranke  <dpranke@chromium.org>
4152
4153         nrwt: use the default logic for skipping platform-specific tests for chromium
4154         https://bugs.webkit.org/show_bug.cgi?id=104403
4155
4156         Reviewed by Stephen White.
4157
4158         I thought I had fixed this a while ago, but the chromium port was
4159         using custom logic to figure out which platform/* tests to run (or
4160         skip) and as a result we were getting some odd behavior (like running
4161         tests in chromium-linux on chromium-mac, but only when there was
4162         an expectation listed for the test).
4163
4164         Ports are supposed to only run the tests found under their
4165         baseline_search_path() by default.
4166
4167         Not writing a test for this; theoretically there shouldn't been a
4168         test checking that the behavior for Chromium was *different* than
4169         the other ports before.
4170
4171         * Scripts/webkitpy/layout_tests/port/chromium.py:
4172         (ChromiumPort._missing_symbol_to_skipped_tests):
4173
4174 2012-12-10  Dirk Pranke  <dpranke@chromium.org>
4175
4176         fix regression in test-webkitpy after r137189
4177         https://bugs.webkit.org/show_bug.cgi?id=104589
4178
4179         Reviewed by Eric Seidel.
4180
4181         Need to specify full port names.
4182
4183         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
4184         (MainTest.test_build_check):
4185
4186 2012-12-10  Laszlo Gombos  <l.gombos@samsung.com>
4187
4188         [EFL] Change the minimum required EFL version to 1.6
4189         https://bugs.webkit.org/show_bug.cgi?id=104431
4190
4191         Reviewed by Kenneth Rohde Christiansen.
4192
4193         Change the minimum required EFL version to 1.6 from 1.7.
4194         The elementary EFL package is only required to build MiniBrowser,
4195         so I moved the required only to the Minibrowser CMake file.
4196
4197         * MiniBrowser/efl/CMakeLists.txt: In addition removed
4198         ENABLE_GLIB_SUPPORT as it is obsolete and it was always set to be on.
4199
4200 2012-12-10  Dean Jackson  <dino@apple.com>
4201
4202         Unreviewed. Adding Antoine to the contributor list so his name autocompletes in bugzilla.
4203
4204         * Scripts/webkitpy/common/config/committers.py:
4205
4206 2012-12-10  Zan Dobersek  <zandobersek@gmail.com>
4207
4208         REGRESSION (r118735): svg/dom/complex-svgView-specification.html, svg/dom/SVGViewSpec.html, svg/dom/viewspec-parser.html failing on GTK Linux 64-bit Release
4209         https://bugs.webkit.org/show_bug.cgi?id=87734
4210
4211         Reviewed by Martin Robinson.
4212
4213         When willSendRequest should return null in testing, cancel the SoupMessage
4214         instead of setting the request's URI to about:blank.
4215
4216         * DumpRenderTree/gtk/DumpRenderTree.cpp:
4217         (willSendRequestCallback):
4218
4219 2012-12-10  Jochen Eisinger  <jochen@chromium.org>
4220
4221         [chromium] Move the test implementation of WebKit::WebPlugin to the TestRunner library
4222         https://bugs.webkit.org/show_bug.cgi?id=104537
4223
4224         Reviewed by Tony Chang.
4225
4226         In order to hide implementation details from the public TestRunner API,
4227         I moved the plugin to WebTestPluginImpl and added a static create
4228         method to WebTestPlugin. Also, I renamed the plugin from TestWebPlugin
4229         to WebTestPlugin to be more in line with other classes in the
4230         TestRunner API.
4231
4232         * DumpRenderTree/DumpRenderTree.gypi:
4233         * DumpRenderTree/chromium/TestRunner/public/WebTestPlugin.h: Added.
4234         (WebTestRunner):
4235         (WebTestPlugin):
4236         * DumpRenderTree/chromium/TestRunner/src/WebTestPlugin.cpp: Renamed from Tools/DumpRenderTree/chromium/TestWebPlugin.cpp.
4237         (WebTestRunner::WebTestPlugin::create):
4238         (WebTestRunner):
4239         (WebTestRunner::WebTestPlugin::WebTestPlugin):
4240         (WebTestRunner::WebTestPlugin::~WebTestPlugin):
4241         (WebTestRunner::WebTestPlugin::mimeType):
4242         * DumpRenderTree/chromium/TestWebPlugin.h: Removed.
4243         * DumpRenderTree/chromium/WebViewHost.cpp:
4244         (WebViewHost::createPlugin):
4245
4246 2012-12-10  Dirk Pranke  <dpranke@chromium.org>
4247
4248         run_webkit_tests throws exception when DRT is not present
4249         https://bugs.webkit.org/show_bug.cgi?id=104565
4250
4251         Reviewed by Jochen Eisinger.
4252
4253         Don't actually try to print the bot results if there aren't
4254         any because we exited before running any tests.
4255
4256         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
4257         (main):
4258         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
4259         (MainTest.test_build_check):
4260
4261 2012-12-10  Benjamin Poulain  <benjamin@webkit.org>
4262
4263         Add convenience methods to use ListHashSet for a LRU cache
4264         https://bugs.webkit.org/show_bug.cgi?id=104499
4265
4266         Reviewed by Sam Weinig.
4267
4268         Test the new methods added to ListHashSet.
4269
4270         * TestWebKitAPI/CMakeLists.txt:
4271         * TestWebKitAPI/GNUmakefile.am:
4272         * TestWebKitAPI/TestWebKitAPI.gypi:
4273         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
4274         * TestWebKitAPI/Tests/WTF/ListHashSet.cpp: Added.
4275         (TestWebKitAPI):
4276         (TestWebKitAPI::TEST):
4277
4278 2012-12-10  Dirk Pranke  <dpranke@chromium.org>
4279
4280         webkit-patch print-expectations doesn't work right for platforms w/ shared expectations
4281         https://bugs.webkit.org/show_bug.cgi?id=104405
4282
4283         Reviewed by Eric Seidel.
4284
4285         If multiple ports shared a TestExpectations file, webkit-patch
4286         print-expectations would use the expectation for the first port
4287         using that file and print the result for all ports, rather than
4288         re-parse and print each port separately (e.g., we might print
4289         the chromium-linux expectation for a file for both chromium-linux
4290         and chromium-mac). Discovered while testing/debugging the fix
4291         for bug 104403.
4292
4293         * Scripts/webkitpy/layout_tests/port/test.py:
4294         (TestPort._skipped_tests_for_unsupported_features):
4295         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: