5d989e742cefc147dc6583660d9849034367ce8d
[WebKit-https.git] / Tools / ChangeLog
1 2012-07-16  Ryuan Choi  <ryuan.choi@samsung.com>
2
3         [EFL][WK2] Add APIs to support theme.
4         https://bugs.webkit.org/show_bug.cgi?id=90107
5
6         Reviewed by Hajime Morita.
7
8         Override default theme path for MiniBrowser and WebKitTestRunner.
9
10         * MiniBrowser/efl/CMakeLists.txt:
11         * MiniBrowser/efl/main.c:
12         (browserCreate):
13         * WebKitTestRunner/PlatformEfl.cmake:
14         * WebKitTestRunner/efl/PlatformWebViewEfl.cpp:
15         (WTR::PlatformWebView::PlatformWebView):
16
17 2012-07-16  Ryuan Choi  <ryuan.choi@samsung.com>
18
19         [EFL][DRT] Implement dumpFrameScrollPosition
20         https://bugs.webkit.org/show_bug.cgi?id=87638
21
22         Reviewed by Hajime Morita.
23
24         * DumpRenderTree/efl/DumpRenderTree.cpp:
25         (dumpFrameScrollPosition): Implemented.
26         (shouldDumpFrameScrollPosition): Fixed wrong condition like other ports.
27
28 2012-07-16  Don Olmstead  <don.olmstead@am.sony.com>
29
30         NRWRT Should provide a VS project to work on
31         https://bugs.webkit.org/show_bug.cgi?id=91436
32
33         Reviewed by Dirk Pranke.
34
35         Adding VS2010 project for modifying webkitpy. Uses pytools
36         <http://pytools.codeplex.com> which integrates python into 
37         Visual Studio.
38
39         * Scripts/webkitpy/webkitpy.pyproj: Added.
40         * Scripts/webkitpy/webkitpy.sln: Added.
41
42 2012-07-16  Adam Barth  <abarth@webkit.org>
43
44         [Chromium] Move GamepadController into TestRunner.a
45         https://bugs.webkit.org/show_bug.cgi?id=91311
46
47         Reviewed by Tony Chang.
48
49         GamepadController looks like a simple class to move into TestRunner.a
50         because it has almost zero dependence on TestShell.
51
52         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
53         * DumpRenderTree/DumpRenderTree.gypi:
54         * DumpRenderTree/chromium/TestRunner/CppBoundClass.cpp: Renamed from Tools/DumpRenderTree/chromium/CppBoundClass.cpp.
55         * DumpRenderTree/chromium/TestRunner/CppBoundClass.h: Renamed from Tools/DumpRenderTree/chromium/CppBoundClass.h.
56         * DumpRenderTree/chromium/TestRunner/CppVariant.cpp: Renamed from Tools/DumpRenderTree/chromium/CppVariant.cpp.
57         * DumpRenderTree/chromium/TestRunner/CppVariant.h: Renamed from Tools/DumpRenderTree/chromium/CppVariant.h.
58             - GamepadController depends on these bindings helpers.
59         * DumpRenderTree/chromium/TestRunner/GamepadController.cpp: Renamed from Tools/DumpRenderTree/chromium/GamepadController.cpp.
60         * DumpRenderTree/chromium/TestRunner/GamepadController.h: Renamed from Tools/DumpRenderTree/chromium/GamepadController.h.
61             - Just moved these files and removed the unused TestShell
62               references.
63         * DumpRenderTree/chromium/TestRunner/Stub.cpp: Removed.
64             - No longer needed.
65         * DumpRenderTree/chromium/TestShell.cpp:
66             - Update call to constructor to avoid passing in this.
67
68 2012-07-16  Xianzhu Wang  <wangxianzhu@chromium.org>
69
70         [Chromium-android] Don't use test_shell mode of DRT
71         https://bugs.webkit.org/show_bug.cgi?id=88542
72
73         Reviewed by Dirk Pranke.
74
75         Test shell mode is about to be removed.
76         Switch to use DRT mode for chromium-android.
77
78         Summary of changes:
79         1. ChromiumAndroidDriver now inherits from WebKitDriver instead of ChromiumDriver (to be deprecated).
80         2. Conforms to the DRT mode protocol for input/output of DumpRenderTree.
81         3. Added support for Android 'adb shell' input/output (base64, newline mode, etc.)
82
83         * DumpRenderTree/chromium/TestEventPrinter.cpp:
84         (DRTPrinter::handleImage): Outputs base64 on Android.
85         * DumpRenderTree/chromium/TestShellAndroid.cpp:
86         (platformInit): Changed err_file to err_fifo, required by python ServerProcess.
87         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
88         (ChromiumAndroidPort.__init__):
89         (ChromiumAndroidPort.create_driver): Override to create driver without DriverProxy to ensure 1 Driver per run.
90         (ChromiumAndroidDriver):
91         (ChromiumAndroidDriver.__init__):
92         (ChromiumAndroidDriver.cmd_line):
93         (ChromiumAndroidDriver._deadlock_detector):
94         (ChromiumAndroidDriver._drt_cmd_line):
95         (ChromiumAndroidDriver.start):
96         (ChromiumAndroidDriver._start):
97         (ChromiumAndroidDriver._start_once):
98         (ChromiumAndroidDriver.run_test):
99         (ChromiumAndroidDriver.stop):
100         (ChromiumAndroidDriver._command_from_driver_input):
101         (ChromiumAndroidDriver._read_prompt):
102         * Scripts/webkitpy/layout_tests/port/chromium_android_unittest.py:
103         (ChromiumAndroidPortTest.test_driver_cmd_line):
104         (ChromiumAndroidDriverTest.test_cmd_line):
105         (ChromiumAndroidDriverTest):
106         (ChromiumAndroidDriverTest.test_drt_cmd_line):
107         (ChromiumAndroidDriverTest.test_read_prompt):
108         (ChromiumAndroidDriverTest.test_command_from_driver_input):
109         (ChromiumAndroidDriverTest.test_write_command_and_read_line):
110         * Scripts/webkitpy/layout_tests/port/server_process.py:
111         (ServerProcess.__init__): Added universal_newlines to handle Android 'adb shell' line ends.
112         (ServerProcess._start):
113         (ServerProcess._wait_for_data_and_update_buffers_using_select): Handles unexpected EOF which indicates crash on Android.
114         (ServerProcess.stop): Added kill_directly parameter to kill the process without waiting it (which always timeouts for Android).
115         (ServerProcess.replace_outputs): Added to combine different input/output pipes into one ServerProcess.
116         * Scripts/webkitpy/layout_tests/port/webkit.py:
117         (WebKitDriver._command_from_driver_input):
118         (WebKitDriver.run_test): Changed timeout origin so that slow start() on Android won't cause timeout of layout test case.
119         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
120         (MockServerProcess.read_stdout):
121         (MockServerProcess.start):
122         (MockServerProcess):
123         (MockServerProcess.stop):
124         (MockServerProcess.kill):
125         (WebKitDriverTest.test_read_block):
126         (WebKitDriverTest.test_read_binary_block):
127         (WebKitDriverTest.test_read_base64_block):
128
129 2012-07-16  Dirk Pranke  <dpranke@chromium.org>
130
131         nrwt: move a bunch of printing code from manager.py to printing.py
132         https://bugs.webkit.org/show_bug.cgi?id=91439
133
134         Reviewed by Ojan Vafai.
135
136         All of the logic that handles what gets logged should eventually
137         live in printing.py; this patch moves a large chunk of code that
138         prints all of the stuff after the run completes.
139
140         There are no functional changes, this is just moving code
141         around. Covered by existing tests (updated as necessary).
142
143         This code is all pretty messy and most of it should be deleted
144         (or extracted after the fact from results.json), but I'm saving
145         that for further patches. My goal is just to manager.py to a
146         more manageable state.
147
148         * Scripts/webkitpy/layout_tests/controllers/manager.py:
149         (Manager.prepare_lists_and_print_output):
150         (Manager.run):
151         (Manager._mark_interrupted_tests_as_skipped):
152         (Manager._update_summary_with_result):
153         (Manager._num_digits):
154         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
155         (ManagerTest.test_interrupt_if_at_failure_limits):
156         (ResultSummaryTest.summarized_results):
157         * Scripts/webkitpy/layout_tests/models/result_summary.py:
158         (ResultSummary.__init__):
159         (ResultSummary.add):
160         Here we had to add a list of the slow tests to the result
161         summary so that we didn't need to call back into the manager and
162         the test expectations object to figure out if a test is SLOW.
163         * Scripts/webkitpy/layout_tests/views/printing.py:
164         (Printer.print_results):
165         (Printer):
166         (Printer._print_timing_statistics):
167         (Printer._print_aggregate_test_statistics):
168         (Printer._print_individual_test_times):
169         (Printer._print_test_list_timing):
170         (Printer._print_directory_timings):
171         (Printer._print_statistics_for_test_timings):
172         (Printer._print_result_summary):
173         (Printer._print_result_summary_entry):
174         * Scripts/webkitpy/layout_tests/views/printing_unittest.py:
175         (Testprinter.test_print_unexpected_results.get_unexpected_results):
176         (Testprinter):
177
178 2012-07-16  Adam Barth  <abarth@webkit.org>
179
180         [Chromium] Create a stub TestRunner.a target
181         https://bugs.webkit.org/show_bug.cgi?id=91309
182
183         Reviewed by Ryosuke Niwa.
184
185         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
186         * DumpRenderTree/DumpRenderTree.gypi:
187         * DumpRenderTree/chromium/TestRunner/Stub.cpp: Added.
188
189 2012-07-16  Dirk Pranke  <dpranke@chromium.org>
190
191         test-webkitpy: clean up logging to make it ninja-esque
192         https://bugs.webkit.org/show_bug.cgi?id=91297
193
194         Reviewed by Adam Barth.
195
196         This patch changes the output of test-webkitpy to be closer to
197         what ninja produces. Namely:
198
199         If you are running with stderr writing to a tty (and not -v), we
200         will no longer print '...' as tests complete. Instead, we use a
201         metered stream and print entries of the form:
202            '[X/Y] test_name passed/failed/erred'
203         where X is the current test # and Y is the total number of tests.
204
205         If you are running with stderr piped to a file or other non-tty
206         object (or with -v), you get the same output, one per line for
207         every test.
208
209         In addition, if tests fail or err out, you get the stack trace
210         immediately; you don't have to wait until the end of the run.
211
212         Lastly, this change cleans up the unit tests for test-webkitpy
213         itself to not confuse the logger and to work cleanly w/ multiple
214         processes. It looks like between this and all the clean up in
215         MessagePool that happened as it landed, we can now run
216         multiprocessing tests in parallel.
217
218         * Scripts/webkitpy/test/finder.py:
219         (Finder._default_names):
220         * Scripts/webkitpy/test/finder_unittest.py:
221         (FinderTest.setUp):
222         (FinderTest.tearDown):
223         * Scripts/webkitpy/test/printer.py:
224         (Printer.__init__):
225         (Printer.configure):
226         (Printer.print_started_test):
227         (Printer.print_finished_test):
228         (Printer._test_line):
229         (Printer.print_result):
230         * Scripts/webkitpy/test/runner.py:
231         (Runner.run):
232         * Scripts/webkitpy/test/runner_unittest.py:
233         (RunnerTest.setUp):
234         (RunnerTest):
235         (RunnerTest.tearDown):
236         (RunnerTest.assert_run):
237         (RunnerTest.test_regular):
238         (RunnerTest.test_verbose):
239         (RunnerTest.test_timing):
240
241 2012-07-16  Dirk Pranke  <dpranke@chromium.org>
242
243         test-webkitpy: handle failures properly when running in parallel
244         https://bugs.webkit.org/show_bug.cgi?id=91416
245
246         Reviewed by Adam Barth.
247
248         It turns out that unittest.TestResults contain a handle to the
249         test method itself, which isn't picklable; it's sufficient to just
250         store the test name instead of the actual method. By doing so
251         we can move the test_name() method from the printer to the
252         runner where it belongs (so the printer is less dependent on the
253         unittest framework's data structures).
254
255         This change should really have a test but I don't know how to
256         write one that properly captures the behavior and won't cause
257         test-webkitpy itself to fail. I've verified the fix by hand, at
258         least, in the meantime.
259
260         * Scripts/webkitpy/test/printer.py:
261         (Printer.__init__):
262         (Printer.print_result):
263         * Scripts/webkitpy/test/runner.py:
264         (_test_name):
265         (Runner.all_test_names):
266         (_Worker.handle):
267
268 2012-07-16  Dirk Pranke  <dpranke@chromium.org>
269
270         test-webkitpy: handle failures properly when running in parallel
271         https://bugs.webkit.org/show_bug.cgi?id=91416
272
273         Reviewed by Tony Chang.
274
275         It turns out that unittest.TestResults contain a handle to the
276         test method itself, which isn't picklable; it's sufficient to just
277         store the test name instead of the actual method. By doing so
278         we can move the test_name() method from the printer to the
279         runner where it belongs (so the printer is less dependent on the
280         unittest framework's data structures).
281
282         This change should really have a test but I don't know how to
283         write one that properly captures the behavior and won't cause
284         test-webkitpy itself to fail. I've verified the fix by hand, at
285         least, in the meantime.
286
287         * Scripts/webkitpy/test/printer.py:
288         (Printer.__init__):
289         (Printer.print_result):
290         * Scripts/webkitpy/test/runner.py:
291         (_test_name):
292         (Runner.all_test_names):
293         (_Worker.handle):
294
295 2012-07-16  Dirk Pranke  <dpranke@chromium.org>
296
297         test-webkitpy: run tests in parallel
298         https://bugs.webkit.org/show_bug.cgi?id=91294
299
300         Reviewed by Ojan Vafai.
301
302         This change adds support for running tests in parallel. This is
303         not yet on by default, since the logging isn't very pretty w/
304         parallel tests.
305
306         Also, there are some (multiprocessing-related) tests that can't be
307         run in parallel and so we skip them in that situation; I need to
308         come up with a mechanism for dealing with this, since you
309         apparently can't use multiprocessing as both a parent and a
310         child process.
311
312         * Scripts/webkitpy/test/finder.py:
313         (Finder.find_names):
314         (Finder._default_names):
315         * Scripts/webkitpy/test/main.py:
316         (Tester._parse_args):
317         (Tester.run):
318         * Scripts/webkitpy/test/main_unittest.py:
319         (TesterTest.test_no_tests_found):
320         * Scripts/webkitpy/test/runner.py:
321         (Runner.run):
322         * Scripts/webkitpy/test/runner_unittest.py:
323         (RunnerTest.test_regular):
324         (RunnerTest.test_verbose):
325         (RunnerTest.test_timing):
326
327 2012-07-16  Dirk Pranke  <dpranke@chromium.org>
328
329         test-webkitpy: use message pools
330         https://bugs.webkit.org/show_bug.cgi?id=91292
331
332         Reviewed by Ojan Vafai.
333
334         Restructure the test-running code to be message-driven and
335         use a MessagePool; note that this does not yet actually run the
336         tests in parallel.
337
338         Also clean up the unit tests so that the fake loader is passed
339         to the _Worker properly, and reduce a lot of the cut&pasted code
340         in the tests.
341
342         No functional changes; covered by existing tests.
343
344         * Scripts/webkitpy/test/printer.py:
345         (Printer.print_started_test):
346         (Printer.print_finished_test):
347         * Scripts/webkitpy/test/runner.py:
348         (Runner.__init__):
349         (Runner.run):
350         (Runner.handle):
351         (_Worker):
352         (_Worker.__init__):
353         (_Worker.handle):
354         * Scripts/webkitpy/test/runner_unittest.py:
355         (RunnerTest.assert_run):
356         (RunnerTest.test_regular):
357         (RunnerTest.test_verbose):
358         (RunnerTest.test_timing):
359
360 2012-07-16  Simon Hausmann  <simon.hausmann@nokia.com>
361
362         [Qt] Fix Qt5 module dependencies
363         https://bugs.webkit.org/show_bug.cgi?id=91388
364
365         Reviewed by Laszlo Gombos.
366
367         Removed ancient qtscript and qtxmlpatterns module dependencies.
368
369         * qmake/qt_webkit.pri:
370
371 2012-07-16  Thiago Marcos P. Santos  <thiago.santos@intel.com>
372
373         run-webkit-test returns 254 at exit even when it succeeds
374         https://bugs.webkit.org/show_bug.cgi?id=91370
375
376         Reviewed by Simon Hausmann.
377
378         sys.exit() is implemented by raising SystemExit which inherits from
379         BaseException. Since we are catching BaseException, run_webkit_tests.py
380         was always returning an error code.
381
382         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
383
384 2012-07-15  Carlos Garcia Campos  <cgarcia@igalia.com>
385
386         Unreviewed. Fix make distcheck.
387
388         * GNUmakefile.am: Add missing file to compilation.
389
390 2012-07-15  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
391
392         Unreviewed. Add EFLWebKit2PublicAPI, EFLWebKit2PlatformSpecific to watch list.
393
394         * Scripts/webkitpy/common/config/watchlist:
395
396 2012-07-15  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
397
398         Unreviewed. Add gyuyoung.kim@samsung.com to watch list.
399
400         * Scripts/webkitpy/common/config/watchlist:
401
402 2012-07-15  Joseph Pecoraro  <pecoraro@apple.com>
403
404         Windowless WebView not firing JavaScript load event if there is a media element
405         https://bugs.webkit.org/show_bug.cgi?id=91331
406
407         Reviewed by Eric Carlson.
408
409         Test a windowless WebView loading a page with a media element. The load
410         event should happen on the next spin of the run loop, but we spin check
411         the page, with a timeout of 250ms.
412
413         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
414         * TestWebKitAPI/Tests/mac/WindowlessWebViewWithMedia.html: Added.
415         * TestWebKitAPI/Tests/mac/WindowlessWebViewWithMedia.mm: Added.
416         (-[WindowlessWebViewWithMediaFrameLoadDelegate webView:didFinishLoadForFrame:]):
417         (TestWebKitAPI::spinLoop): Spin check with timeout. Runs a block each spin to early bail.
418         (TestWebKitAPI::TEST):
419
420 2012-07-14  Benjamin Poulain  <bpoulain@apple.com>
421
422         [Mac] Do not try to update the cache model for every WebPreferences change
423         https://bugs.webkit.org/show_bug.cgi?id=91302
424
425         Reviewed by Joseph Pecoraro.
426
427         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
428         * TestWebKitAPI/Tests/mac/SetAndUpdateCacheModel.mm: Added.
429         (TestWebKitAPI):
430         (TestWebKitAPI::TEST):
431
432 2012-07-14  Sheriff Bot  <webkit.review.bot@gmail.com>
433
434         Unreviewed, rolling out r122665.
435         http://trac.webkit.org/changeset/122665
436         https://bugs.webkit.org/show_bug.cgi?id=91321
437
438         Broke Mac builds (Requested by rniwa on #webkit).
439
440         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
441         * TestWebKitAPI/Tests/mac/SetAndUpdateCacheModel.mm: Removed.
442
443 2012-07-13  Benjamin Poulain  <bpoulain@apple.com>
444
445         [Mac] Do not try to update the cache model for every WebPreferences change
446         https://bugs.webkit.org/show_bug.cgi?id=91302
447
448         Reviewed by Joseph Pecoraro.
449
450         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
451         * TestWebKitAPI/Tests/mac/SetAndUpdateCacheModel.mm: Added.
452         (TestWebKitAPI):
453         (TestWebKitAPI::TEST):
454
455 2012-07-13  Dirk Pranke  <dpranke@chromium.org>
456
457         run-webkit-test outputs the wrong number of tests executed when some are skipped.
458         https://bugs.webkit.org/show_bug.cgi?id=89894
459
460         Reviewed by Ojan Vafai.
461
462         Fix the logging of the actual number of tests run so that tests
463         that are skipped aren't included.
464
465         Also revamp the 'expected' output so we distinguish the number
466         of tests found from the number of tests run (to account for
467         --repeat-each and --iterations).
468
469         Covered by existing tests.
470
471         * Scripts/webkitpy/layout_tests/controllers/manager.py:
472         (Manager.prepare_lists_and_print_output):
473         (Manager._log_num_workers):
474         (Manager.run):
475         (Manager._print_result_summary):
476         * Scripts/webkitpy/layout_tests/models/result_summary.py:
477         (ResultSummary.__init__):
478         (ResultSummary.add):
479         * Scripts/webkitpy/layout_tests/views/printing.py:
480         (Printer.print_one_line_summary):
481         * Scripts/webkitpy/layout_tests/views/printing_unittest.py:
482         (Testprinter.test_print_one_line_summary):
483
484 2012-07-13  Dirk Pranke  <dpranke@chromium.org>
485
486         nrwt: actually print the exception name and message for otherwise unhandled exceptions
487         https://bugs.webkit.org/show_bug.cgi?id=91305
488
489         Reviewed by Adam Barth.
490
491         Two more places where I was printing the stack trace but not the
492         exception itself :(. These two spots can't easily be
493         unit-tested, but I tested them by hand.
494
495         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
496         (run):
497         (main):
498
499 2012-07-13  Josh Hawn  <jhawn@apple.com>
500
501         Fix for WebContext::getWebCoreStatistics() causes crash if no m_process
502         https://bugs.webkit.org/show_bug.cgi?id=91116
503
504         Reviewed by Simon Fraser.
505
506         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
507             Added new test file.
508         * TestWebKitAPI/Tests/WebKit2/WebCoreStatisticsWithNoWebProcess.cpp: Added.
509         (TestWebKitAPI::wkContextGetStatisticsCallback):
510             Tests that callback function receives an error.
511         (TestWebKitAPI::TEST):
512             Creates a dummy web context object (no web process).
513             Calls WKContextGetStatistics with the web context and test callback.
514             The test callback should get an expected error.
515
516 2012-07-13  Dirk Pranke  <dpranke@chromium.org>
517
518         test-webkitpy: move printing-related code out of the runner
519         https://bugs.webkit.org/show_bug.cgi?id=91289
520
521         Reviewed by Ryosuke Niwa.
522
523         More refactoring ... this moves all printing-related stuff out
524         of runner.py and into printer.py.
525
526         No functional changes; covered by existing tests.
527
528         * Scripts/webkitpy/test/main.py:
529         (Tester._run_tests):
530         * Scripts/webkitpy/test/printer.py:
531         (Printer.__init__):
532         (Printer):
533         (Printer.test_name):
534         (Printer.print_started_test):
535         (Printer.print_finished_test):
536         (Printer.print_result):
537         * Scripts/webkitpy/test/runner.py:
538         (Runner.__init__):
539         (Runner.all_test_names):
540         (Runner.run):
541         * Scripts/webkitpy/test/runner_unittest.py:
542         (RunnerTest.test_regular):
543         (RunnerTest.test_verbose):
544         (RunnerTest.test_timing):
545
546 2012-07-13  Dirk Pranke  <dpranke@chromium.org>
547
548         webkitpy: split printing/logging code for test-webkitpy out into a new class
549         https://bugs.webkit.org/show_bug.cgi?id=91282
550
551         Reviewed by Ojan Vafai.
552
553         This patch is the first step at splitting all of the
554         printing/logging code out separately from the actual
555         test-running code.
556
557         This is just moving stuff around; no new functionality and no
558         new tests needed.
559
560         * Scripts/webkitpy/test/finder_unittest.py:
561         (FinderTest.setUp):
562         * Scripts/webkitpy/test/main.py:
563         (Tester.__init__):
564         (Tester._parse_args):
565         (Tester.run):
566         (Tester._run_tests):
567         (Tester._log_exception):
568         * Scripts/webkitpy/test/main_unittest.py:
569         (TesterTest.test_no_tests_found):
570         * Scripts/webkitpy/test/printer.py: Added.
571         (Printer):
572         (Printer.__init__):
573         (Printer.configure):
574         (Printer.configure.filter):
575         (_CaptureAndPassThroughStream):
576         (_CaptureAndPassThroughStream.__init__):
577         (_CaptureAndPassThroughStream.write):
578         (_CaptureAndPassThroughStream._message_is_from_pdb):
579         (_CaptureAndPassThroughStream.flush):
580         (_CaptureAndPassThroughStream.getvalue):
581
582 2012-07-13  James Simonsen  <simonjam@chromium.org>
583
584         [Navigation Timing] Imported W3C tests contain duplicates and are DOS formatted
585         https://bugs.webkit.org/show_bug.cgi?id=91184
586
587         Reviewed by Adam Barth.
588
589         The upstream 'html5' tests are just duplicates of the 'html' tests.
590
591         * Scripts/import-w3c-performance-wg-tests:
592
593 2012-07-13  Dirk Pranke  <dpranke@chromium.org>
594
595         webkitpy: hide yield_to_caller from callers in MessagePool :)
596         https://bugs.webkit.org/show_bug.cgi?id=91269
597
598         Reviewed by Adam Barth.
599
600         yield_to_caller() was an optimization/hack to allow us to run
601         both manager and worker in a single process/loop without
602         starving the manager while the worker is running tests. The
603         worker was required to call yield_to_caller() periodically. It
604         turns out that I can get equivalent responsiveness by yielding
605         inside the MessagePool every time the worker posts a message, and this
606         allows me to no longer need the worker to call the routine. Thus
607         I rename yield_to_caller() to _yield_to_manager() to be a little
608         clearer about its purpose.
609
610         Tested by existing tests.
611
612         * Scripts/webkitpy/common/message_pool.py:
613         (_Worker.run):
614         (_Worker.post):
615         (_Worker._yield_to_manager):
616         * Scripts/webkitpy/layout_tests/controllers/worker.py:
617         (Worker.handle):
618
619 2012-07-13  Adam Barth  <abarth@webkit.org>
620
621         EWSTools should be able to build a commit-queue instance from scratch
622         https://bugs.webkit.org/show_bug.cgi?id=91264
623
624         Reviewed by Eric Seidel.
625
626         I've been using this script to build commit-queue instances on Google
627         Compute Engine and it seems to work.
628
629         * EWSTools/GoogleComputeEngine/build-commit-queue.sh: Added.
630
631 2012-07-13  Dirk Pranke  <dpranke@chromium.org>
632
633         webkitpy: make worker.start() and worker.stop() optional in the messagepool
634         https://bugs.webkit.org/show_bug.cgi?id=91170
635
636         Reviewed by Ojan Vafai.
637
638         test-webkitpy will use messagepool workers that don't actually
639         have any per-worker state, so they don't need start() and stop()
640         methods. Now we will only call the methods if they exist; this
641         means that workers only need to expose a handle() method.
642
643         * Scripts/webkitpy/common/message_pool.py:
644         (_Worker.terminate):
645         (_Worker.run):
646
647 2012-07-13  Dirk Pranke  <dpranke@chromium.org>
648
649         NRWT doesn't print exceptions
650         https://bugs.webkit.org/show_bug.cgi?id=91129
651
652         Reviewed by Ojan Vafai.
653
654         Although we printed exceptions in most cases, if an unexpected
655         exception (like a runtime error) was raised when creating a
656         port, we wouldn't. This patch fixes that, and also cleans up
657         how we were logging exceptions from the workers to be less
658         verbose.
659
660         Because of the corner cases where these errors are occurring,
661         it's difficult to write automated unit tests for them. I've
662         tested it quite a bit by hand, though.
663
664         * Scripts/webkitpy/common/message_pool.py:
665         (_MessagePool._close):
666         (_MessagePool._handle_worker_exception):
667         (_Worker.run):
668         (_Worker._raise):
669         * Scripts/webkitpy/layout_tests/controllers/manager.py:
670         (Manager._run_tests):
671         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
672         (main):
673
674 2012-07-13  Wei James  <james.wei@intel.com>
675
676         enable TestWebKitAPI/webkit_unit_tests apk on x86 android platform by adding abi support
677         https://bugs.webkit.org/show_bug.cgi?id=91194
678
679         Reviewed by Adam Barth.
680
681         * TestWebKitAPI/TestWebKitAPI.gyp/TestWebKitAPI.gyp:
682
683 2012-07-13  Simon Pena  <spena@igalia.com>
684
685         [GTK] Gardening: update API tests skipped list
686         https://bugs.webkit.org/show_bug.cgi?id=91224
687
688         Unreviewed gardening.
689
690         Skip "next" and "previous" tests of FindController until bug #91083
691         is fixed.
692
693         * gtk/run-api-tests:
694         (TestRunner):
695
696 2012-07-13  Zeno Albisser  <zeno@webkit.org>
697
698         [Qt][WK2] Implement GraphicsSurface for Linux/GLX.
699         https://bugs.webkit.org/show_bug.cgi?id=90881
700
701         Enable GraphicsSurface for Linux based platforms
702         whenever the Xcomposite extension is available.
703
704         Reviewed by Noam Rosenthal.
705
706         * qmake/config.tests/libXcomposite/libXcomposite.cpp: Added.
707         (main):
708         * qmake/config.tests/libXcomposite/libXcomposite.pro: Added.
709           Add a configure test to detect Xcomposite extension and
710           activate GraphicsSurface on linux in case the extension is available.
711         * qmake/configure.pri:
712         * qmake/mkspecs/features/features.prf:
713
714 2012-07-13  David Grogan  <dgrogan@chromium.org>
715
716         nrwt: don't choke when printing invalid utf-8 to stderr
717         https://bugs.webkit.org/show_bug.cgi?id=91181
718
719         Reviewed by Dirk Pranke.
720
721         * Scripts/webkitpy/layout_tests/controllers/test_result_writer.py:
722         (TestResultWriter.write_stderr):
723
724 2012-07-13  Dirk Pranke  <dpranke@chromium.org>
725
726         test-webkitpy: more class renaming cleanup
727         https://bugs.webkit.org/show_bug.cgi?id=91182
728
729         Reviewed by Adam Barth.
730
731         More removing of the unnecessary "Test" prefix.
732
733         * Scripts/webkitpy/test/finder.py:
734         (_DirectoryTree):
735         (Finder.add_tree):
736         * Scripts/webkitpy/test/main.py:
737         (Tester._run_tests):
738         * Scripts/webkitpy/test/runner.py:
739         (Runner):
740         * Scripts/webkitpy/test/runner_unittest.py:
741         (RunnerTest.test_regular):
742         (RunnerTest.test_verbose):
743         (RunnerTest.test_timing):
744
745 2012-07-12  Christophe Dumez  <christophe.dumez@intel.com>
746
747         [WK2][EFL] Facilitate debugging of the Web Process
748         https://bugs.webkit.org/show_bug.cgi?id=90768
749
750         Reviewed by Kenneth Rohde Christiansen.
751
752         Add a new --webprocess-cmd-prefix argument to
753         run-webkit-tests script for EFL port. If provided,
754         the prefix will be prepended to the command used
755         to spawn the Web process. This can be used for
756         debugging purposes with prefixes such as:
757         "xterm -title renderer -e gdb --args".
758
759         * Scripts/webkitpy/layout_tests/port/efl.py:
760         (EflPort.__init__):
761         (EflPort.setup_environ_for_server):
762         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
763         (parse_args):
764         * WebKitTestRunner/efl/TestControllerEfl.cpp:
765         (WTR::TestController::platformRunUntil): Implement support for
766         m_noTimeout timeout value.
767
768 2012-07-12  Adam Barth  <abarth@webkit.org>
769
770         Fix crash in the commit-queue.  We need to initialize self.port during __init__.
771
772         * Scripts/webkitpy/tool/commands/queues.py:
773         (CommitQueue.__init__):
774         (CommitQueue.begin_work_queue):
775
776 2012-07-12  Dirk Pranke  <dpranke@chromium.org>
777
778         test-webkitpy: rename test_finder to finder
779         https://bugs.webkit.org/show_bug.cgi?id=91175
780
781         Reviewed by Adam Barth.
782
783         Rename test_finder -> finder, TestFinder -> Finder to remove
784         some of the stutter in the names.
785
786         * Scripts/webkitpy/test/finder.py: Renamed from Tools/Scripts/webkitpy/test/test_finder.py.
787         * Scripts/webkitpy/test/finder_unittest.py: Renamed from Tools/Scripts/webkitpy/test/test_finder_unittest.py.
788         * Scripts/webkitpy/test/main.py:
789         (Tester.__init__):
790
791 2012-07-12  Adam Barth  <abarth@webkit.org>
792
793         CommitQueue is confused about what port it is using
794         https://bugs.webkit.org/show_bug.cgi?id=91040
795
796         Reviewed by Dirk Pranke.
797
798         On EC2, we explicitly pass --port to the commit-queue, but that
799         requires editing the start-queue.sh script locally on each bot. In
800         moving to Google Compute Engine, we're try to avoid any local edits to
801         the EWSTools.
802
803         Rather than passing --port to the commit-queue, this patch teaches the
804         CommitQueue which port its running, which is the approach we use for
805         the EWS bots.
806
807         Mutating tool._deprecated_port is a bit ugly, but it's what we're doing
808         currently for the EWS bots.
809
810         * Scripts/webkitpy/tool/commands/queues.py:
811         (CommitQueue):
812         (CommitQueue.begin_work_queue):
813         (CommitQueue.run_command):
814         * Scripts/webkitpy/tool/commands/queues_unittest.py:
815         (CommitQueueTest.test_commit_queue):
816         (mock_run_webkit_patch):
817         (test_rollout):
818         (test_rollout_lands):
819         (test_manual_reject_during_processing):
820
821 2012-07-12  James Simonsen  <simonjam@chromium.org>
822
823         [Navigation Timing] Import the W3C Navigation Timing test suite
824         https://bugs.webkit.org/show_bug.cgi?id=84887
825
826         Reviewed by Tony Gentilcore.
827
828         * Scripts/import-w3c-performance-wg-tests: Added.
829
830 2012-07-12  Adam Barth  <abarth@webkit.org>
831
832         Unreviewed.  Nit: git config files use tabs, not spaces.
833
834         * EWSTools/cold-boot.sh:
835
836 2012-07-12  Kwang Yul Seo  <skyul@company100.net>
837
838         Unreviewed. Add Dongsung Huang to the list of contributors. He
839         has submitted over 30 patches on texture mapper, canvas and image decoders.
840
841         * Scripts/webkitpy/common/config/committers.py:
842
843 2012-07-12  Kwang Yul Seo  <skyul@company100.net>
844
845         Unreviewed. Change my irc nickname to kseo.
846
847         * Scripts/webkitpy/common/config/committers.py:
848
849 2012-07-12  Dirk Pranke  <dpranke@chromium.org>
850
851         webkitpy: clean up logging handlers, lint common.message_pool
852         https://bugs.webkit.org/show_bug.cgi?id=91152
853
854         Reviewed by Ojan Vafai.
855
856         The unix implementation of multiprocessing clones any logging
857         handlers from the parent process into the child; we currently
858         don't want this behavior in our code, so I was hand-removing the
859         installed handlers in the child process I knew about. After thinking
860         about it further, I think it was simpler and safe enough to just
861         remove all handlers in the child, since the message pool
862         propagates any message from the child back into the parent.
863         
864         We can always change this in the future if it turns out to be an issue.
865
866         I'm also fixing a couple of other lint warnings while I'm at it.
867
868         * Scripts/webkitpy/common/message_pool.py:
869         (_MessagePool.__exit__):
870         (_MessagePool._handle_worker_exception):
871         (_Worker._set_up_logging):
872
873 2012-07-12  Dirk Pranke  <dpranke@chromium.org>
874
875         webkitpy: rename manager_worker_broker to message_pool
876         https://bugs.webkit.org/show_bug.cgi?id=91145
877
878         Reviewed by Ojan Vafai.
879
880         Since the MessagePool interface is more generic (and simpler)
881         now and will be reused by test-webkitpy, I'm renaming it and
882         moving it to webkitpy.common.
883
884         * Scripts/webkitpy/common/message_pool.py: Renamed from Tools/Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py.
885         * Scripts/webkitpy/layout_tests/controllers/manager.py:
886         (TestRunInterruptedException.__reduce__):
887         (Manager._run_tests.worker_factory):
888         (Manager._run_tests):
889         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
890
891 2012-07-12  Dirk Pranke  <dpranke@chromium.org>
892
893         nrwt crashes saving the output for a platform-specific expected test reference
894         https://bugs.webkit.org/show_bug.cgi?id=90872
895
896         Reviewed by Ojan Vafai.
897
898         The expected output for a test is copied alongside the test
899         itself in the layout-test-results directory; in other words, for
900         foo/bar-expected.txt sits alongside foo/bar.html even if we're
901         actually using platform/mac/foo/bar-expected.txt.
902
903         Unless the test is a reftest, in which case we would copy the
904         output to platform/mac/foo/bar-expected.html and set a
905         'ref_file' parameter in results.json to indicate the path. This
906         can be useful in the cases where we have multiple references for
907         a single test or when multiple tests share the same reference.
908
909         We found a bug where we weren't creating platform/mac/foo under
910         the results directory, and so this wasn't actually working.
911         However, treating reftests differently seems like a bad thing,
912         so we should probably be consistent. This change puts the
913         -expected.html next to the test, and reworks test_result_writer
914         so that we create directories uniformly and consistently.
915
916         Note that we weren't catching this problem in unit tests because
917         the MockFileSystem creates a directory automatically if it
918         doesn't exist; this was done intentionally for convenience, but
919         is really a bug and should be fixed; see https://bugs.webkit.org/show_bug.cgi?id=91028.
920
921         I have not added additional tests here since fixing that bug
922         should be sufficient.
923
924         * Scripts/webkitpy/layout_tests/controllers/manager.py:
925         (interpret_test_failures):
926         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
927         (ResultSummaryTest.test_interpret_test_failures):
928         * Scripts/webkitpy/layout_tests/controllers/test_result_writer.py:
929         (write_test_result):
930         (TestResultWriter._write_binary_file):
931         (TestResultWriter):
932         (TestResultWriter._write_text_file):
933         (TestResultWriter.write_output_files):
934         (TestResultWriter.write_stderr):
935         (TestResultWriter.write_crash_log):
936         (TestResultWriter.create_text_diff_and_write_result):
937         (TestResultWriter.write_image_diff_files):
938         (write_reftest):
939         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
940         (EndToEndTest.test_reftest_with_two_notrefs):
941
942 2012-07-12  Dirk Pranke  <dpranke@chromium.org>
943
944         nrwt: reimplement manager_worker_broker in a much simpler form
945         https://bugs.webkit.org/show_bug.cgi?id=90513
946
947         Reviewed by Ojan Vafai.
948
949         This is a wholesale replacement of the MessagePool() implementation
950         and the other classes in manager_worker_broker.py. All of the
951         BrokerConnection*, Broker*, etc. classes are gone, and there are now
952         just a MessagePool class and a _Worker class. Happiness ensues.
953  
954         I'm removing manager_worker_broker_unittest.py as well; we get
955         nearly complete coverage from the integration tests, and will
956         get more coverage when test-webkitpy moves to use this as well,
957         so having unit tests seems like unnecessary overhead. (running
958         coverage numbers with test-webkitpy shows that pretty much the only
959         uncovered lines are lines that are only run in the child processes,
960         which coverage doesn't handle at the moment).
961  
962         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
963         (_MessagePool.__init__):
964         (_MessagePool.run):
965         (_MessagePool._start_workers):
966         (_MessagePool):
967         (_MessagePool.wait):
968         (_MessagePool._close):
969         (_MessagePool._handle_done):
970         (_MessagePool._can_pickle):
971         (_MessagePool._loop):
972         (WorkerException):
973         (_Message.__init__):
974         (_Message.__repr__):
975         (_Worker):
976         (_Worker.__init__):
977         (_Worker.terminate):
978         (_Worker._close):
979         (_Worker.run):
980         (_Worker.post):
981         (_Worker.yield_to_caller):
982         (_Worker._post):
983         (_Worker._raise):
984         (_Worker._set_up_logging):
985         (_WorkerLogHandler.__init__):
986         (_WorkerLogHandler.emit):
987         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py: Removed.
988
989 2012-07-12  Tony Chang  <tony@chromium.org>
990
991         [chromium] Remove drag and drop API methods that are no longer used
992         https://bugs.webkit.org/show_bug.cgi?id=90996
993
994         Reviewed by Adam Barth.
995
996         Migrate DRT to use the methods that take modifier keys.
997
998         * DumpRenderTree/chromium/EventSender.cpp:
999         (EventSender::doDragDrop):
1000         (EventSender::doMouseUp):
1001         (EventSender::doMouseMove):
1002         (EventSender::beginDragWithFiles):
1003
1004 2012-07-12  Ojan Vafai  <ojan@chromium.org>
1005
1006         Allow putting ranges in user.py list prompts
1007         https://bugs.webkit.org/show_bug.cgi?id=91115
1008
1009         Reviewed by Adam Barth.
1010
1011         Ranges are inclusive and denoted by a dash. This is useful for rebaselining a whole port
1012         since the items are listed with each port's builders being contiguous.
1013
1014         * Scripts/webkitpy/common/system/user.py:
1015         (User._wait_on_list_response):
1016         * Scripts/webkitpy/common/system/user_unittest.py:
1017         (UserTest.test_prompt_with_multiple_lists.run_prompt_test):
1018         (UserTest.test_prompt_with_multiple_lists):
1019
1020 2012-07-12  Arnaud Renevier  <arno@renevier.net>
1021
1022         [GTK] Implement disableImageLoading in DRT
1023         https://bugs.webkit.org/show_bug.cgi?id=87973
1024
1025         Reviewed by Martin Robinson.
1026
1027         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1028         (resetDefaultsToConsistentValues):
1029         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1030         (LayoutTestController::disableImageLoading):
1031
1032 2012-07-12  Simon Hausmann  <simon.hausmann@nokia.com>
1033
1034         [Qt] Internal symbols are exported on Linux
1035         https://bugs.webkit.org/show_bug.cgi?id=90981
1036
1037         Reviewed by Jocelyn Turcotte.
1038
1039         I should've re-enabled ELF symbol visibility when removing the symbol map in
1040         r106650.
1041
1042         * qmake/mkspecs/features/default_post.prf:
1043
1044 2012-07-11  Mark Rowe  <mrowe@apple.com>
1045
1046         <http://webkit.org/b/91024> Build against the latest SDK when targeting older OS X versions.
1047
1048         Reviewed by Dan Bernstein.
1049
1050         The deployment target is already set to the version that we're targeting, and it's that setting
1051         which determines which functionality from the SDK is available to us.
1052
1053         * DumpRenderTree/mac/Configurations/Base.xcconfig:
1054         * TestWebKitAPI/Configurations/Base.xcconfig:
1055         * WebKitTestRunner/Configurations/Base.xcconfig:
1056
1057 2012-07-11  Mark Rowe  <mrowe@apple.com>
1058
1059         Replace definitions of BUILDING_ON / TARGETING macros with macros that will error when used.
1060
1061         Part of <http://webkit.org/b/91015> Remove BUILDING_ON / TARGETING macros in favor of system availability macros.
1062
1063         Reviewed by Anders Carlsson.
1064
1065         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h: Remove the macros completely from here since
1066         they're completely unused in TestNetscapePlugIn.
1067         * DumpRenderTree/config.h:
1068
1069 2012-07-11  Mark Rowe  <mrowe@apple.com>
1070
1071         <http://webkit.org/b/91015> Remove BUILDING_ON / TARGETING macros in favor of system availability macros
1072
1073         This removal was handled by a script that translates the relevant macros in to the equivalent checks
1074         using the system availability macros.
1075
1076         Reviewed by Filip Pizlo.
1077
1078         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
1079         * DumpRenderTree/mac/CheckedMalloc.cpp:
1080         * DumpRenderTree/mac/DumpRenderTree.mm:
1081         * DumpRenderTree/mac/DumpRenderTreeDraggingInfo.mm:
1082         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1083         * DumpRenderTree/mac/TextInputController.m:
1084         * TestWebKitAPI/mac/InjectedBundleControllerMac.mm:
1085         * WebKitTestRunner/InjectedBundle/mac/ActivateFonts.mm:
1086         * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm:
1087
1088 2012-07-11  Robert Kroeger  <rjkroege@chromium.org>
1089
1090         Suppress horizontal conversion of PlatformWheelEvents when hasPreciseScrollingDeltas is true
1091         https://bugs.webkit.org/show_bug.cgi?id=89580
1092
1093         WebKit GTK and Chromium Linux force vertical wheel events to
1094         scroll horizontally when over horizontal scroll bars.  This is
1095         undesirable for touchpad scrolling with
1096         hasPreciseScrollingDeltas() == true. Modified DumpRenderTree to
1097         let a layout test specify this attribute so that the change's impact
1098         on scrolling can be tested in a layout test.
1099
1100         Reviewed by Adam Barth.
1101
1102         * DumpRenderTree/chromium/EventSender.cpp:
1103         (EventSender::handleMouseWheel):
1104
1105 2012-07-11  Simon Fraser  <simon.fraser@apple.com>
1106
1107         Fix the build by declaring -isPaginated before use.
1108
1109         * MiniBrowser/mac/BrowserWindowController.m:
1110
1111 2012-07-11  Sheriff Bot  <webkit.review.bot@gmail.com>
1112
1113         Unreviewed, rolling out r122358.
1114         http://trac.webkit.org/changeset/122358
1115         https://bugs.webkit.org/show_bug.cgi?id=91037
1116
1117         Build break on WebKit Win (Requested by hayato on #webkit).
1118
1119         * DumpRenderTree/chromium/EventSender.cpp:
1120         (EventSender::doDragDrop):
1121         (EventSender::doMouseUp):
1122         (EventSender::doMouseMove):
1123         (EventSender::beginDragWithFiles):
1124
1125 2012-07-03  Dirk Pranke  <dpranke@chromium.org>
1126
1127         nrwt: add a MessagePool abstraction that the manager will call to replace the broker
1128         https://bugs.webkit.org/show_bug.cgi?id=90511
1129
1130         Reviewed by Ojan Vafai.
1131
1132         This change introduces the new MessagePool abstraction that will
1133         replace the classes in manager_worker_broker. It is a minimal
1134         interface that tries to follow the conventions in
1135         multiprocessing.Pool and concurrency.futures ... it provides a
1136         context manager and a run() method that sends N messages to M
1137         workers processes (starting workers as necessary) and waits for
1138         them all to complete, handling cleanup as necessary. The caller
1139         is responsible for providing a handle() method to handle
1140         messages received from the workers.
1141         
1142         This interface basically hides all of the multiprocessing logic from
1143         the manager class.
1144
1145         The initial implementation of MessagePool is a simple shim
1146         around the existing broker classes; a subsequent change will
1147         replace all the other classes with a much simpler
1148         implementation.
1149
1150         No additional tests are provided for now; existing tests should
1151         provide adequate coverage, and I will add new unit tests for the
1152         MessagePool class when I replace the existing implementation.
1153
1154         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1155         (TestRunInterruptedException.__reduce__):
1156         (Manager.__init__):
1157         (Manager._run_tests):
1158         (Manager._run_tests.instead):
1159         (Manager.handle):
1160         (Manager._handle_started_test):
1161         (Manager._handle_finished_test_list):
1162         (Manager._handle_finished_test):
1163         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
1164         (get):
1165         (_MessagePool):
1166         (_MessagePool.__init__):
1167         (_MessagePool.__enter__):
1168         (_MessagePool.__exit__):
1169         (_MessagePool.run):
1170         (_MessagePool.wait):
1171         (_MessagePool.is_done):
1172         (_MessagePool._worker_is_done):
1173         (_MessagePool._close):
1174         (_MessagePool.handle_done):
1175         (_MessagePool.handle_started_test):
1176         (_MessagePool.handle_finished_test):
1177         (_MessagePool.handle_finished_test_list):
1178         (_MessagePool.handle_exception):
1179         (_MessagePool._log_messages):
1180         (_MessagePool._handle_worker_exception):
1181         (_WorkerState):
1182         (_WorkerState.for):
1183         (_WorkerState.__init__):
1184         (_WorkerState.__repr__):
1185         (_get_broker):
1186         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py:
1187         (make_broker):
1188
1189 2012-07-11  Simon Fraser  <simon.fraser@apple.com>
1190
1191         Add an option to enter paginated mode in MiniBrowser
1192         https://bugs.webkit.org/show_bug.cgi?id=91035
1193
1194         Reviewed by Dan Bernstein.
1195
1196         Add an item to the Debug menu for MiniBrowser that puts
1197         the web view into paginated mode. For now, we only
1198         do left-to-right pagination, with a fixed column width.
1199
1200         * MiniBrowser/mac/BrowserWindowController.h:
1201         * MiniBrowser/mac/BrowserWindowController.m:
1202         (-[BrowserWindowController validateMenuItem:]): Update the checked
1203         state of the menu item.
1204         (-[BrowserWindowController isPaginated]): Return YES if paginated.
1205         (-[BrowserWindowController togglePaginationMode:]): Toggle in
1206         and out of pagination mode.
1207         * MiniBrowser/mac/MainMenu.xib: Add the debug menu item.
1208
1209 2012-07-11  Adam Barth  <abarth@webkit.org>
1210
1211         commit-queue instances on Compute Engine are missing git-svn
1212         https://bugs.webkit.org/show_bug.cgi?id=91034
1213
1214         Reviewed by Eric Seidel.
1215
1216         git-svn is needed to actually commit to SVN from a git working copy.
1217
1218         * EWSTools/cold-boot.sh:
1219
1220 2012-07-11  Dirk Pranke  <dpranke@chromium.org>
1221
1222         webkitpy: lint code in webkitpy.layout_tests.models
1223         https://bugs.webkit.org/show_bug.cgi?id=90416
1224
1225         Reviewed by Ojan Vafai.
1226
1227         Cleaning up errors reported from lint-webkitpy.
1228
1229         Also, suppress the warnings about wildcard imports in pylintrc;
1230         we have nothing particularly against them.
1231
1232         * Scripts/webkitpy/layout_tests/models/test_configuration.py:
1233         (TestConfigurationConverter.combinations):
1234         * Scripts/webkitpy/layout_tests/models/test_configuration_unittest.py:
1235         (TestConfigurationTest.test_hash.query_unknown_key):
1236         (TestConfigurationTest.test_eq):
1237         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
1238         (ParseError.__init__):
1239         (TestExpectationLine.__init__):
1240         (TestExpectationsModel.get_expectations_string):
1241         (TestExpectationsModel):
1242         (TestExpectationsModel.expectation_to_string):
1243         (TestExpectationsModel.add_expectation_line):
1244         (TestExpectationsModel._clear_expectations_for_test):
1245         (TestExpectationsModel._remove_from_sets):
1246         (TestExpectations.get_expectations_string):
1247         (TestExpectations.expectation_to_string):
1248         (TestExpectations._report_warnings):
1249         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
1250         (Base.__init__):
1251         (parse_exp):
1252         (SkippedTests.check):
1253         (TestExpectationParserTests.test_parse_empty_string):
1254         * Scripts/webkitpy/layout_tests/models/test_failures.py:
1255         (FailureTimeout.__init__):
1256         (FailureCrash.__init__):
1257         (FailureImageHashMismatch.__init__):
1258         (FailureReftestMismatch.__init__):
1259         (FailureReftestMismatchDidNotOccur.__init__):
1260         (FailureReftestNoImagesGenerated.__init__):
1261         * Scripts/webkitpy/layout_tests/models/test_failures_unittest.py:
1262         (TestFailuresTest.test_unknown_failure_type.UnknownFailure.message):
1263         (TestFailuresTest.test_unknown_failure_type):
1264         (TestFailuresTest):
1265         (TestFailuresTest.test_message_is_virtual):
1266         * Scripts/webkitpy/layout_tests/models/test_results.py:
1267         (TestResult.loads):
1268         (TestResult.has_failure_matching_types):
1269         * Scripts/webkitpy/pylintrc:
1270
1271 2012-07-11  Dirk Pranke  <dpranke@chromium.org>
1272
1273         nrwt: clean up names in worker.py
1274         https://bugs.webkit.org/show_bug.cgi?id=90510
1275
1276         Reviewed by Ojan Vafai.
1277
1278         This is the last patch in the series of refactoring worker.py;
1279         all this does is change some names of methods, instance
1280         variables, and method parameters to be clearer (it also changes
1281         some code in manager.py and manager_worker.py to be consistent).
1282
1283         There are no functional changes in this patch and the existing
1284         tests should cover everything.
1285
1286         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1287         (Manager.__init__):
1288         (Manager.prepare_lists_and_print_output):
1289         (Manager._run_tests.worker_factory):
1290         (Manager._run_tests):
1291         (Manager._show_results_html_file):
1292         (Manager.handle_finished_test_list):
1293         (_WorkerState.__init__):
1294         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
1295         (_Broker.post_message):
1296         (_Broker._dispatch_message):
1297         (AbstractWorker.__init__):
1298         (AbstractWorker.run):
1299         (AbstractWorker.yield_to_caller):
1300         (AbstractWorker.post):
1301         (_WorkerConnection.__init__):
1302         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py:
1303         (_TestWorker.__init__):
1304         (_TestWorker.start):
1305         (_TestWorker.handle):
1306         (_TestWorker.stop):
1307         (_TestsMixin):
1308         (_TestsMixin.test_name):
1309         * Scripts/webkitpy/layout_tests/controllers/worker.py:
1310         (Worker.__init__):
1311         (Worker.__del__):
1312         (Worker.start):
1313         (Worker.handle):
1314         (Worker._run_test):
1315         (Worker.stop):
1316         (Worker._timeout):
1317         (Worker._kill_driver):
1318         (Worker._run_test_with_timeout):
1319         (Worker._clean_up_after_test):
1320         (Worker._run_test_in_another_thread.SingleTestThread.run):
1321         (Worker._run_test_in_this_thread):
1322         (Worker._run_single_test):
1323
1324 2012-07-11  Adam Barth  <abarth@webkit.org>
1325
1326         The commit-queue needs some extra git config to be able to commit
1327         https://bugs.webkit.org/show_bug.cgi?id=91025
1328
1329         Reviewed by Eric Seidel.
1330
1331         In order for the commit-queue to actually commit, it needs to know the
1332         location of the SVN server.
1333
1334         * EWSTools/cold-boot.sh:
1335
1336 2012-07-11  Adam Barth  <abarth@webkit.org>
1337
1338         Teach EWSTools how to configure SVN auth credentials
1339         https://bugs.webkit.org/show_bug.cgi?id=91021
1340
1341         Reviewed by Eric Seidel.
1342
1343         To move the commit-queue over to Google Compute Engine, we need a way
1344         to configure the commit-queue's credentials during the machine build
1345         process. There doesn't seem to be a nice command line way of
1346         configuring SVN auth credentials. I tried doing a bogus commit and
1347         supplying the credentials that way, but that's super hacky.
1348
1349         The approach in this patch is to write the config file directly. The
1350         format of these configurations files is documented in the SVN book, so
1351         this approach doesn't see too sketchy.
1352
1353         * EWSTools/configure-svn-auth.sh: Added.
1354
1355 2012-07-11  Ojan Vafai  <ojan@chromium.org>
1356
1357         Dedupe suffixes passed to webkit-patch rebaseline
1358         https://bugs.webkit.org/show_bug.cgi?id=91017
1359
1360         Reviewed by Dirk Pranke.
1361
1362         * Scripts/webkitpy/tool/commands/rebaseline.py:
1363         (Rebaseline._suffixes_to_update):
1364         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
1365         (test_rebaseline_multiple_builders_and_tests_command_line):
1366
1367 2012-07-11  Kevin Ollivier  <kevino@theolliviers.com>
1368
1369         [wx] Unreviewed build fix. Add new directories and a new LayoutTestController method.
1370
1371         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
1372         (LayoutTestController::setStorageDatabaseIdleInterval):
1373         * waf/build/settings.py:
1374
1375 2012-07-11  Tony Chang  <tony@chromium.org>
1376
1377         [chromium] Remove drag and drop API methods that are no longer used
1378         https://bugs.webkit.org/show_bug.cgi?id=90996
1379
1380         Reviewed by Adam Barth.
1381
1382         Migrate DRT to use the methods that take modifier keys.
1383
1384         * DumpRenderTree/chromium/EventSender.cpp:
1385         (EventSender::doDragDrop):
1386         (EventSender::doMouseUp):
1387         (EventSender::doMouseMove):
1388         (EventSender::beginDragWithFiles):
1389
1390 2012-07-11  Mark Rowe  <mrowe@apple.com>
1391
1392         Add a Mountain Lion version of libWebKitSystemInterface.a.
1393
1394         Reviewed by John Sullivan.
1395
1396         * Scripts/copy-webkitlibraries-to-product-directory: Include libWebKitSystemInterfaceMountainLion.a in the list of libraries to copy.
1397
1398 2012-07-09  Mark Rowe  <mrowe@apple.com>
1399
1400         <http://webkit.org/b/90835> Teach bisect-builds to work with a Safari.app that has entitlements.
1401
1402         Reviewed by Dan Bernstein.
1403
1404         * Scripts/bisect-builds: Use safariPathFromSafariBundle to determine which binary within the application
1405         should be invoked.
1406
1407 2012-07-11  Zan Dobersek  <zandobersek@gmail.com>
1408
1409         [Gtk] fast/events/keydown-function-keys.html is failing
1410         https://bugs.webkit.org/show_bug.cgi?id=90891
1411
1412         Reviewed by Martin Robinson.
1413
1414         Work around the context menu being shown on F10 key being pressed by
1415         unbiding the key when running tests in DumpRenderTree. The problem
1416         appears when using a recent version of the xkeyboard-config package.
1417
1418         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1419         (setDefaultsToConsistentStateValuesForTesting):
1420
1421 2012-07-11  No'am Rosenthal  <noam.rosenthal@nokia.com>
1422
1423         [Qt] QRawWebView should notify when rendering is done, so that pixel results can be grabbed at the appropriate moment.
1424         https://bugs.webkit.org/show_bug.cgi?id=90641
1425
1426         Reviewed by Jocelyn Turcotte.
1427
1428         * MiniBrowser/qt/raw/View.h: Comment used old class name (WKView).
1429         (View):
1430
1431 2012-07-11  Min Qin  <qinmin@chromium.org>
1432
1433         [Android] sending an extra to the DRT apk so that DRT can be run in a seperate thread
1434         https://bugs.webkit.org/show_bug.cgi?id=90831
1435
1436         Reviewed by Adam Barth.
1437
1438         On android, DRT needs to run in a background thread to avoid ANR.
1439         However, the java tests are running on UI thread by default.
1440         We need to send an intent extra to the apk so that it can run on a sub thread.
1441
1442         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
1443         (ChromiumAndroidDriver._start_once):
1444
1445 2012-07-11  Csaba Osztrogonác  <ossy@webkit.org>
1446
1447         [Qt] REGRESSION(r107171): Fix --timeout option of Qt's DRT
1448         https://bugs.webkit.org/show_bug.cgi?id=90966
1449
1450         Reviewed by Ryosuke Niwa.
1451
1452         * DumpRenderTree/qt/main.cpp:
1453         (main): Don't remove the argument, because takeOptionValue() did it before.
1454
1455 2012-07-11  Allan Sandfeld Jensen  <allan.jensen@nokia.com>
1456
1457         First commit; moving myself to commiters.
1458
1459         Unreviewed.
1460
1461         * Scripts/webkitpy/common/config/committers.py:
1462
1463 2012-07-11  Balazs Ankes  <bank@inf.u-szeged.hu>
1464
1465         webkit-patch land should automatically add svn:mime-type for .png files
1466         https://bugs.webkit.org/show_bug.cgi?id=75825
1467
1468         Reviewed by Dirk Pranke.
1469
1470         Refactor the png.py to avoid code duplication.
1471
1472         * Scripts/webkitpy/common/checksvnconfigfile.py: Added.
1473         (check):
1474         (config_file_path):
1475         (errorstr_autoprop):
1476         (errorstr_png):
1477         * Scripts/webkitpy/style/checkers/png.py:
1478         (PNGChecker.check):
1479         * Scripts/webkitpy/tool/commands/download.py:
1480         (Land):
1481         * Scripts/webkitpy/tool/steps/__init__.py:
1482         * Scripts/webkitpy/tool/steps/addsvnmimetypeforpng.py: Added.
1483         (AddSvnMimetypeForPng):
1484         (AddSvnMimetypeForPng.__init__):
1485         (AddSvnMimetypeForPng.run):
1486         (AddSvnMimetypeForPng._check_pngs):
1487         * Scripts/webkitpy/tool/steps/addsvnmimetypeforpng_unittest.py: Added.
1488         (MockSCMDetector):
1489         (MockSCMDetector.__init__):
1490         (MockSCMDetector.display_name):
1491         (AddSvnMimetypeForPngTest):
1492         (AddSvnMimetypeForPngTest.test_run):
1493
1494 2012-07-11  Hayato Ito  <hayato@chromium.org>
1495
1496         Unreviewed gardening.
1497         One more fix for r122292.
1498
1499         * TestWebKitAPI/TestWebKitAPI.gyp/TestWebKitAPI.gyp:
1500
1501 2012-07-11  Sheriff Bot  <webkit.review.bot@gmail.com>
1502
1503         Unreviewed, rolling out r122301.
1504         http://trac.webkit.org/changeset/122301
1505         https://bugs.webkit.org/show_bug.cgi?id=90947
1506
1507         Android builder started to fail (Requested by hayato on
1508         #webkit).
1509
1510         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
1511
1512 2012-07-10  Adam Barth  <abarth@webkit.org>
1513
1514         [Chromium] Merge final nits to DumpRenderTree.gyp for Android
1515         https://bugs.webkit.org/show_bug.cgi?id=90920
1516
1517         Reviewed by Tony Chang.
1518
1519         This patch contains the last few small changes to DumpRenderTree.gyp
1520         from the chromium-android branch. After this change, this file will be
1521         fully merged.
1522
1523         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
1524
1525 2012-07-10  Adam Barth  <abarth@webkit.org>
1526
1527         [Chromium-Android] Add apk test targets for webkit_unit_tests and TestWebKitAPI
1528         https://bugs.webkit.org/show_bug.cgi?id=90918
1529
1530         Reviewed by Tony Chang.
1531
1532         The rules are similar to what we have added for DumpRenderTree apk.
1533         All references to gtest_target_type can be removed once we enable APK
1534         tests on the all bots.
1535
1536         * TestWebKitAPI/TestWebKitAPI.gyp/TestWebKitAPI.gyp:
1537
1538 2012-07-10  Kenneth Russell  <kbr@google.com>
1539
1540         Add --no-build option to perf test runner
1541         https://bugs.webkit.org/show_bug.cgi?id=90916
1542
1543         Reviewed by Ryosuke Niwa.
1544
1545         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
1546         (PerfTestsRunner._parse_args):
1547             Support --no-build option to perf tests, as in layout tests.
1548
1549 2012-07-10  Adam Barth  <abarth@webkit.org>
1550
1551         EWSTools should be able to build a chromium-ews bot from scratch
1552         https://bugs.webkit.org/show_bug.cgi?id=90912
1553
1554         Reviewed by Eric Seidel.
1555
1556         I've been using this script to kick off the build process for the
1557         chromium-ews bots on Google Compute Engine.
1558
1559         * EWSTools/GoogleComputeEngine: Added.
1560         * EWSTools/GoogleComputeEngine/build-chromium-ews.sh: Added.
1561
1562 2012-07-10  Alec Flett  <alecflett@chromium.org>
1563
1564         Add alecflett to the list of committers
1565         https://bugs.webkit.org/show_bug.cgi?id=90903
1566
1567         Reviewed by Ojan Vafai.
1568
1569         * Scripts/webkitpy/common/config/committers.py:
1570
1571 2012-07-10  Xianzhu Wang  <wangxianzhu@chromium.org>
1572
1573         Unreviewed fix of webkitpy unittest break about chromium_android.py
1574         caused by r122251.
1575
1576         * Scripts/webkitpy/layout_tests/port/chromium_android_unittest.py:
1577         (ChromiumAndroidDriverTest.test_cmd_line):
1578
1579 2012-07-10  Ojan Vafai  <ojan@chromium.org>
1580
1581         webkit-patch rebaseline doesn't work for audio/pixel tests
1582         https://bugs.webkit.org/show_bug.cgi?id=90905
1583
1584         Reviewed by Adam Barth.
1585
1586         We just need to pass the correct suffix list to rebaseline-test-internal
1587         and optimize-baselines. By default, pass all suffixes.
1588
1589         * Scripts/webkitpy/tool/commands/rebaseline.py:
1590         (Rebaseline):
1591         (Rebaseline.__init__):
1592         (Rebaseline._suffixes_to_update):
1593         (Rebaseline.execute):
1594         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
1595         (test_rebaseline_multiple_builders_and_tests_command_line):
1596
1597 2012-07-10  Adam Barth  <abarth@webkit.org>
1598
1599         Remove LayoutTestController and WebKitTestRunner support for Hixie76 WebSockets
1600         https://bugs.webkit.org/show_bug.cgi?id=90853
1601
1602         Reviewed by Eric Seidel.
1603
1604         WebKit no longer implements the Hixie76 version of the WebSocket
1605         protocol. We don't need to support it in our test framework.
1606
1607         * DumpRenderTree/chromium/LayoutTestController.cpp:
1608         (LayoutTestController::overridePreference):
1609         * DumpRenderTree/chromium/WebPreferences.cpp:
1610         (WebPreferences::reset):
1611         (WebPreferences::applyTo):
1612         * DumpRenderTree/chromium/WebPreferences.h:
1613         (WebPreferences):
1614         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1615         (resetDefaultsToConsistentValues):
1616         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1617         (LayoutTestController::overridePreference):
1618         * DumpRenderTree/mac/DumpRenderTree.mm:
1619         (resetDefaultsToConsistentValues):
1620         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1621         (WebCore::WebPage::resetSettings):
1622         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1623         (LayoutTestController::overridePreference):
1624         * DumpRenderTree/win/DumpRenderTree.cpp:
1625         (resetDefaultsToConsistentValues):
1626
1627 2012-07-10  Xianzhu Wang  <wangxianzhu@chromium.org>
1628
1629         [Chromium-Android] Fix typos in chromium_android.py in r151492
1630         https://bugs.webkit.org/show_bug.cgi?id=90904
1631
1632         Reviewed by Ojan Vafai.
1633
1634         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
1635         (ChromiumAndroidPort.setup_test_run):
1636
1637 2012-07-10  Xianzhu Wang  <wangxianzhu@chromium.org>
1638
1639         [Chromium-Android] Use setup_test_runner() instead of start_helper() to setup test environment
1640         https://bugs.webkit.org/show_bug.cgi?id=90894
1641
1642         Reviewed by Adam Barth.
1643
1644         start_helper() is actually start_pixel_test_helper() since r115601 (bug 81729).
1645         Should use setup_test_runner() to setup test environment for chromium-android.
1646
1647         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
1648         (ChromiumAndroidPort.setup_test_run): Renamed from start_helper(). Added cache cleanup code.
1649         (ChromiumAndroidPort.clean_up_test_run): Renamed from stop_helper().
1650         (ChromiumAndroidPort._path_to_helper): Returns None as we don't have a helper now.
1651         (ChromiumAndroidPort):
1652         (ChromiumAndroidPort._path_to_forwarder): Original _path_to_helper().
1653         (ChromiumAndroidPort._push_executable):
1654         (ChromiumAndroidDriver.__init__):
1655         (ChromiumAndroidDriver.cmd_line):
1656
1657 2012-07-09  Ojan Vafai  <ojan@chromium.org>
1658
1659         Improve webkit-patch rebaseline to work for more cases
1660         https://bugs.webkit.org/show_bug.cgi?id=90504
1661
1662         Reviewed by Adam Barth.
1663
1664         -Makes it work for the build.chromium.org bots.
1665         -Lets you rebaseline all builders instead of just one.
1666         -Lets you pass in the builders or tests to rebaseline.
1667
1668         * Scripts/webkitpy/common/host.py:
1669         (Host.buildbot_for_builder_name):
1670         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
1671         (Builder.__init__):
1672         (Builder.latest_layout_test_results):
1673         Provide a way to get to the LayoutTestResults of the latest build.
1674         Most of the time we only need to get the latest one and the Chromium bots
1675         only expose the full_results.json file for the latest build.
1676
1677         (Builder):
1678         (Builder._fetch_file_from_results):
1679         (Builder.fetch_layout_test_results):
1680         Move these functions into Builder so that Builder and Build can both
1681         fetch layout test results.
1682
1683         (Build.results_zip_url):
1684         (Build.layout_test_results):
1685         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1686         (BuilderTest.test_latest_layout_test_results):
1687         (BuildTest.test_layout_test_results):
1688         * Scripts/webkitpy/common/system/user.py:
1689         (User.prompt_with_multiple_lists):
1690         Prompt with multiple sublists, but still have a global numbering.
1691         This lets the build.chromium.org bots be clearly separated from the
1692         build.webkit.org bots, which helps understand the builder names.
1693
1694         (User):
1695         (User._wait_on_list_response):
1696         (User.prompt_with_list):
1697         * Scripts/webkitpy/common/system/user_unittest.py:
1698         (UserTest.test_prompt_with_multiple_lists):
1699         (UserTest.test_prompt_with_multiple_lists.run_prompt_test):
1700         (UserTest.test_prompt_with_multiple_lists.run_prompt_test.mock_raw_input):
1701         * Scripts/webkitpy/layout_tests/port/base.py:
1702         (Port.is_chromium):
1703         Provide a way to tell if a port is a Chromium port that doesn't involve string
1704         manipulation on the port name or builder name.
1705
1706         * Scripts/webkitpy/layout_tests/port/builders.py:
1707         (builder_path_from_name):
1708         (all_builder_names):
1709         Memoizing here is incorrect because the test override _exact_matches,
1710         so these can return different values. In either case, I'm pretty sure these
1711         are not remotely hot codepaths. I manually inspected all call sites and they're
1712         all in high-level calls (e.g. execute calls for webkit-patch commands) and not
1713         called in a loop.
1714
1715         * Scripts/webkitpy/layout_tests/port/chromium.py:
1716         (ChromiumPort.is_chromium):
1717         * Scripts/webkitpy/tool/commands/rebaseline.py:
1718         (RebaselineTest._results_url):
1719         (AbstractParallelRebaselineCommand.__init__):
1720         (Rebaseline.__init__):
1721         (Rebaseline):
1722         (Rebaseline._builders_to_pull_from):
1723         (Rebaseline._tests_to_update):
1724         (Rebaseline.execute):
1725         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
1726         (test_rebaseline.mock_builders_to_pull_from):
1727         (test_rebaseline):
1728         (test_rebaseline_command_line_flags):
1729         (test_rebaseline_multiple_builders):
1730         (test_rebaseline_multiple_builders.mock_builders_to_pull_from):
1731         (test_rebaseline_multiple_builders.mock_tests_to_update):
1732         (test_rebaseline_multiple_builders_and_tests_command_line):
1733         * Scripts/webkitpy/tool/mocktool.py:
1734         (MockTool.irc):
1735         (MockTool):
1736         (MockTool.buildbot_for_builder_name):
1737
1738 2012-07-10  Alice Cheng  <alice_cheng@apple.com>
1739
1740         Editing: Reproducible crasher when pasting a 0x0 image into Mail
1741         https://bugs.webkit.org/show_bug.cgi?id=90640
1742         <rdar://problem/11141920>
1743
1744         Reviewed by Brady Eidson.
1745
1746         Test cases for the patch. 0x0 images don't get a resource representation in the WebArchive, so we need a null check.
1747
1748         * TestWebKitAPI/Tests/mac/0.png: Added.
1749         * TestWebKitAPI/Tests/mac/WebViewCanPasteZeroPng.mm: Added.
1750         (TestWebKitAPI):
1751         (TestWebKitAPI::TEST):
1752
1753 2012-07-10  Leandro Gracia Gil  <leandrogracia@chromium.org>
1754
1755         WebSurroundingText layout tests should use the same code path as the rest of the feature.
1756         https://bugs.webkit.org/show_bug.cgi?id=90807
1757
1758         Reviewed by Adam Barth.
1759
1760         Make the textSurroundingNode method take a pair of point coordinates
1761         instead of a node offset.
1762
1763         * DumpRenderTree/chromium/LayoutTestController.cpp:
1764         (LayoutTestController::textSurroundingNode):
1765         * DumpRenderTree/chromium/LayoutTestController.h:
1766         (LayoutTestController):
1767
1768 2012-07-10  Csaba Osztrogonác  <ossy@webkit.org>
1769
1770         [Qt] REGRESSION(r122175): It broke the Windows build
1771         https://bugs.webkit.org/show_bug.cgi?id=90850
1772
1773         Reviewed by Noam Rosenthal.
1774
1775         * qmake/mkspecs/features/features.prf: Disable USE(3D_GRAPHICS) on Windows temporarily until proper fix.
1776
1777 2012-07-10  János Badics  <jbadics@inf.u-szeged.hu>
1778
1779         [Qt][NRWT] Enable cascaded TestExpectations
1780         https://bugs.webkit.org/show_bug.cgi?id=89108
1781
1782         Reviewed by Dirk Pranke.
1783
1784         * Scripts/webkitpy/layout_tests/port/qt.py:
1785         (QtPort.expectations_files):
1786         * Scripts/webkitpy/layout_tests/port/qt_unittest.py:
1787         (QtPortTest._assert_expectations_files):
1788         (QtPortTest.test_expectations_files):
1789
1790 2012-07-10  Christophe Dumez  <christophe.dumez@intel.com>
1791
1792         [EFL] WebKit DRT and WTR fail to build due to undefined reference to WTF::MD5::*
1793         https://bugs.webkit.org/show_bug.cgi?id=90868
1794
1795         Unreviewed EFL build fix.
1796
1797         Correct CMake configuration to have EFL's DRT and WebKitTestRunner link
1798         WTF library. This is needed to resolve undefined reference to WTF::MD5::*.
1799
1800         * DumpRenderTree/efl/CMakeLists.txt:
1801         * WebKitTestRunner/PlatformEfl.cmake:
1802
1803 2012-07-10  Adam Barth  <abarth@webkit.org>
1804
1805         LayoutTestController.dumpConfigurationForViewport should move to Internals
1806         https://bugs.webkit.org/show_bug.cgi?id=45652
1807
1808         Reviewed by Eric Seidel.
1809
1810         Remove dumpConfigurationForViewport from LayoutTestController. This API
1811         has been replaced by an API on Internals that does the same thing in a
1812         port-agnostic way.
1813
1814         * DumpRenderTree/LayoutTestController.cpp:
1815         (LayoutTestController::staticFunctions):
1816         * DumpRenderTree/LayoutTestController.h:
1817         (LayoutTestController):
1818         * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp:
1819         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp:
1820         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1821         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1822         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1823         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1824         (LayoutTestController):
1825         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1826         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
1827
1828 2012-07-09  Matt Falkenhagen  <falken@chromium.org>
1829
1830         Add ENABLE_DIALOG_ELEMENT and skeleton files
1831         https://bugs.webkit.org/show_bug.cgi?id=90521
1832
1833         Reviewed by Kent Tamura.
1834
1835         * Scripts/webkitperl/FeatureList.pm:
1836
1837 2012-07-09  Kent Tamura  <tkent@chromium.org>
1838
1839         Improve the boilerplate generated by prepare-ChangeLog
1840         https://bugs.webkit.org/show_bug.cgi?id=89560
1841
1842         Reviewed by Ryosuke Niwa.
1843
1844         Produce the following boilerplate:
1845
1846         > 2012-06-20  Kent Tamura  <tkent@chromium.org>
1847         >
1848         >         Need a short description (Oops!).
1849         >         Need the bug URL (Oops!).
1850         >
1851         >         Reviewed by NOBODY (Oops!).
1852         >
1853         >         Additional information of the change such as approach, rationale. Please add per-function descriptions below. (Oops!).
1854         >
1855         >         No new tests (Oops!).
1856
1857         * Scripts/prepare-ChangeLog:
1858         (generateNewChangeLogs): Generate the updated template.
1859         * Scripts/webkitpy/common/checkout/changelog.py:
1860         (ChangeLog.update_with_unreviewed_message):
1861         Update the pattern for the beginning of the boilerplate.
1862         (ChangeLog.set_short_description_and_bug_url):
1863         - Use self._changelog_indent
1864         - Update the pattern for the description line
1865         - Skip the bug boilerplate.
1866         We substitute the specified bug URL to the description boilerplate
1867         because we need to add it even if the bug URL boilerplate is missing.
1868         * Scripts/webkitpy/common/checkout/changelog_unittest.py: Update the boilerplate data.
1869         (test_set_short_description_and_bug_url):
1870         - Update the description boilerplate.
1871         - Add a test for a case with both of the description boilerplace and the
1872           bug URL boilerplate.
1873         * Scripts/webkitpy/tool/steps/preparechangelog_unittest.py:
1874         (PrepareChangeLogTest.test_ensure_bug_url): Update the expected boilerplate.
1875
1876 2012-07-09  No'am Rosenthal  <noam.rosenthal@nokia.com>
1877
1878         Shared code that is guarded with ENABLE(WEBGL) should be guarded with USE()
1879         https://bugs.webkit.org/show_bug.cgi?id=90506
1880
1881         Reviewed by Martin Robinson.
1882
1883         Always enable the WTF_USE_3D_GRAPHICS flag in features.prf if we're compiling with Qt5
1884         or if the OpenGL module is enabled.
1885
1886         * qmake/mkspecs/features/features.prf:
1887
1888 2012-07-09  Yongjun Zhang  <yongjun_zhang@apple.com>
1889
1890         Consider closing unused localStorage database after a timeout.
1891         https://bugs.webkit.org/show_bug.cgi?id=90713
1892
1893         For a localStorage, if there is no active document referencing to it for certain amount of time (300 seconds),
1894         we can close the underlying sqlite database.
1895
1896         Reviewed by Brady Eidson.
1897
1898         * DumpRenderTree/LayoutTestController.cpp: add a JS method setStorageDatabaseIdleInterval to testRunner
1899             to set a different timeout value for closing localStorage database.
1900         (setStorageDatabaseIdleIntervalCallback):
1901         (LayoutTestController::staticFunctions):
1902         * DumpRenderTree/LayoutTestController.h:
1903         (LayoutTestController):
1904         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp:
1905         (LayoutTestController::setStorageDatabaseIdleInterval): add empty method.
1906         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1907         (LayoutTestController::setStorageDatabaseIdleInterval): ditto.
1908         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1909         (LayoutTestController::setStorageDatabaseIdleInterval): ditto.
1910         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1911         (LayoutTestController::setStorageDatabaseIdleInterval): ditto.
1912
1913 2012-07-09  Adam Klein  <adamk@chromium.org>
1914
1915         Rename WebCore::WebKitMutationObserver to WebCore::MutationObserver
1916         https://bugs.webkit.org/show_bug.cgi?id=90810
1917
1918         Reviewed by Ojan Vafai.
1919
1920         * DumpRenderTree/gtk/EventSender.cpp:
1921
1922 2012-07-09  Mihai Parparita  <mihaip@chromium.org>
1923
1924         Handle missing results in TestResultsServer better
1925         https://bugs.webkit.org/show_bug.cgi?id=90816
1926
1927         Reviewed by Ojan Vafai.
1928
1929         If we can't load the results JSON, don't try to wrap it with the JSONP
1930         callback invocation.
1931
1932         * TestResultServer/handlers/testfilehandler.py:
1933         (GetFile.get):
1934
1935 2012-07-09  Sheriff Bot  <webkit.review.bot@gmail.com>
1936
1937         Unreviewed, rolling out r122124.
1938         http://trac.webkit.org/changeset/122124
1939         https://bugs.webkit.org/show_bug.cgi?id=90815
1940
1941         It broke NRWT on Qt (Requested by Ossy_HOME on #webkit).
1942
1943         * Scripts/webkitpy/layout_tests/port/qt.py:
1944         (QtPort._skipped_file_search_paths):
1945         * Scripts/webkitpy/layout_tests/port/qt_unittest.py:
1946         (QtPortTest._assert_skipped_path):
1947         (QtPortTest.test_skipped_file_search_path):
1948
1949 2012-07-09  Zan Dobersek  <zandobersek@gmail.com>
1950
1951         [Gtk] Implement dumpFrameScrollPosition in DumpRenderTree
1952         https://bugs.webkit.org/show_bug.cgi?id=89356
1953
1954         Reviewed by Martin Robinson.
1955
1956         Implement dumpFrameScrollPosition, acquiring the scroll position
1957         through WebKitDOMDOMWindow of the WebKitDOMDocument loaded in the frame
1958         whose scroll position is being dumped.
1959
1960         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1961         (dumpFrameScrollPosition):
1962
1963 2012-07-09  János Badics  <jbadics@inf.u-szeged.hu>
1964
1965         [Qt][NRWT] Enable cascaded TestExpectations
1966         https://bugs.webkit.org/show_bug.cgi?id=89108
1967
1968         Reviewed by Dirk Pranke.
1969
1970         * Scripts/webkitpy/layout_tests/port/qt.py:
1971         (QtPort.expectations_files):
1972         * Scripts/webkitpy/layout_tests/port/qt_unittest.py:
1973         (QtPortTest._assert_expectations_files):
1974         (QtPortTest.test_expectations_files):
1975
1976 2012-07-09  Dan Bernstein  <mitz@apple.com>
1977
1978         Fixed clean builds of the All target in the DumpRenderTree project.
1979
1980         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Let Xcode know that the
1981         DumpRenderTree Perl Support target depends on the DumpRenderTree target.
1982
1983 2012-07-09  Zan Dobersek  <zandobersek@gmail.com>
1984
1985         [Gtk] Unskip the CSS Regions layout tests
1986         https://bugs.webkit.org/show_bug.cgi?id=90771
1987
1988         Reviewed by Martin Robinson.
1989
1990         Use the DumpRenderTreeSupportGtk method for enabling the CSS Regions,
1991         calling it when resetting the defaults in DumpRenderTree or when the
1992         settings value is overriden through the testRunner.
1993
1994         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1995         (resetDefaultsToConsistentValues):
1996         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1997         (LayoutTestController::overridePreference):
1998
1999 2012-07-09  Sheriff Bot  <webkit.review.bot@gmail.com>
2000
2001         Unreviewed, rolling out r122107.
2002         http://trac.webkit.org/changeset/122107
2003         https://bugs.webkit.org/show_bug.cgi?id=90794
2004
2005         Build failure on Mac debug bots (Requested by falken_ on
2006         #webkit).
2007
2008         * Scripts/webkitperl/FeatureList.pm:
2009
2010 2012-07-09  Matt Falkenhagen  <falken@chromium.org>
2011
2012         Add ENABLE_DIALOG_ELEMENT and skeleton files
2013         https://bugs.webkit.org/show_bug.cgi?id=90521
2014
2015         Reviewed by Kent Tamura.
2016
2017         * Scripts/webkitperl/FeatureList.pm:
2018
2019 2012-07-08  Kent Tamura  <tkent@chromium.org>
2020
2021         Unreviewed, rolling out r122076.
2022         http://trac.webkit.org/changeset/122076
2023         https://bugs.webkit.org/show_bug.cgi?id=89560
2024
2025         Broke sheriffbot rollout
2026
2027         * Scripts/prepare-ChangeLog:
2028         (generateNewChangeLogs):
2029
2030 2012-07-08  Leandro Gracia Gil  <leandrogracia@chromium.org>
2031
2032         Update the description of LayoutTestController::textSurroundingElement
2033         https://bugs.webkit.org/show_bug.cgi?id=90620
2034
2035         Reviewed by Adam Barth.
2036
2037         Rename and fix the description of the method as its behaviour changed
2038         during the review of https://bugs.webkit.org/show_bug.cgi?id=82461
2039
2040         * DumpRenderTree/chromium/LayoutTestController.cpp:
2041         (LayoutTestController::LayoutTestController):
2042         (LayoutTestController::textSurroundingNode):
2043         * DumpRenderTree/chromium/LayoutTestController.h:
2044         (LayoutTestController):
2045
2046 2012-07-08  Kent Tamura  <tkent@chromium.org>
2047
2048         Improve the template generated by prepare-ChangeLog
2049         https://bugs.webkit.org/show_bug.cgi?id=89560
2050
2051         Reviewed by Ryosuke Niwa.
2052
2053         Produce the following template:
2054
2055         > 2012-06-20  Kent Tamura  <tkent@chromium.org>
2056         >
2057         >         Need a short description (Oops!).
2058         >         Need the bug URL (Oops!).
2059         >
2060         >         Reviewed by Ryosuke Niwa.
2061         >
2062         >         Additional information of the change such as approach, rationale. Please add per-function descriptions below. (Oops!).
2063         >
2064         >         No new tests (Oops!).
2065
2066         * Scripts/prepare-ChangeLog:
2067         (generateNewChangeLogs):
2068
2069 2012-07-08  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
2070
2071         [wx] Unreviewed. Fix notificaitons directory path after notifications move.
2072
2073         * waf/build/settings.py:
2074
2075 2012-07-07  Nuno Lopes  <nlopes@apple.com>
2076
2077         Fix build with recent clang.
2078         https://bugs.webkit.org/show_bug.cgi?id=90712
2079
2080         Reviewed by Alexey Proskuryakov.
2081
2082         * TestWebKitAPI/Tests/WebKit2/DOMWindowExtensionBasic_Bundle.cpp:
2083         (DOMWindowExtensionBasic):
2084         * TestWebKitAPI/Tests/WebKit2/DOMWindowExtensionNoCache_Bundle.cpp:
2085         (DOMWindowExtensionNoCache):
2086         * TestWebKitAPI/Tests/WebKit2/DocumentStartUserScriptAlertCrash_Bundle.cpp:
2087         (TestWebKitAPI::DocumentStartUserScriptAlertCrashTest::initialize):
2088
2089 2012-07-06  Zan Dobersek  <zandobersek@gmail.com>
2090
2091         [Gtk] Add a configuration option for disabling unstable features in releases
2092         https://bugs.webkit.org/show_bug.cgi?id=87995
2093
2094         Reviewed by Martin Robinson.
2095
2096         Pass the --enable-unstable-features flag when building the Gtk port through build-webkit.
2097
2098         * Scripts/webkitdirs.pm:
2099         (buildAutotoolsProject):
2100
2101 2012-07-06  Jessie Berlin  <jberlin@apple.com>
2102
2103         WKContext should ask for its initialization data when it needs it so the client doesn't have
2104         to keep it up to date.
2105         https://bugs.webkit.org/show_bug.cgi?id=90627
2106
2107         Reviewed by Anders Carlsson.
2108
2109         Add tests and update other WKContextInjectedBundleClients.
2110
2111         * MiniBrowser/mac/AppDelegate.m:
2112         (-[BrowserAppDelegate init]):
2113         Updated for the change to WKContextInjectedBundleClient.
2114         * WebKitTestRunner/TestController.cpp:
2115         (WTR::TestController::initialize):
2116         Ditto.
2117
2118         * TestWebKitAPI/PlatformUtilities.cpp:
2119         (TestWebKitAPI::Util::createInitializationDictionaryForInjectedBundleTest):
2120         Moved the logic to create the initialization dictionary here ...
2121         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
2122         ... from here so that it can be used without automatically using
2123         WKContextSetInitializationUserDataForInjectedBundle.
2124         * TestWebKitAPI/PlatformUtilities.h:
2125
2126         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
2127         Add the new test files.
2128         * TestWebKitAPI/GNUmakefile.am:
2129         Ditto.
2130
2131         * TestWebKitAPI/Tests/WebKit2/GetInjectedBundleInitializationUserDataCallback.cpp: Added.
2132         (TestWebKitAPI::didReceiveMessageFromInjectedBundle):
2133         Check that the message received from the injected bundle matches the user data it was
2134         initialized with.
2135         (TestWebKitAPI::getInjectedBundleInitializationUserData):
2136         Return the user data that the injected bundle should be initialized with.
2137         (TestWebKitAPI::TEST):
2138         Set up WKContextInjectedBundleClient and load a page.
2139         * TestWebKitAPI/Tests/WebKit2/GetInjectedBundleInitializationUserDataCallback_Bundle.cpp: Added.
2140         (TestWebKitAPI::GetInjectedBundleInitializationUserDataCallbackTest::initialize):
2141         Send the initialization user data back up to the UI Process.
2142
2143         * TestWebKitAPI/Tests/WebKit2/InjectedBundleInitializationUserDataCallbackWins.cpp: Added.
2144         (TestWebKitAPI::didReceiveMessageFromInjectedBundle):
2145         Check that the message received from the injected bundle matches the user data it was
2146         initialized with in the callback.
2147         (TestWebKitAPI::getInjectedBundleInitializationUserData):
2148         Return the user data that the injected bundle should be initialized with.
2149         (TestWebKitAPI::TEST):
2150         Set up the context and use WKContextSetInitializationUserDataForInjectedBundle to set the
2151         initialization user data (which should be overridden by the user data returned in
2152         getInjectedBundleInitializationUserData).
2153         * TestWebKitAPI/Tests/WebKit2/InjectedBundleInitializationUserDataCallbackWins_Bundle.cpp: Added.
2154         (TestWebKitAPI::InjectedBundleInitializationUserDataCallbackWinsTest::initialize):
2155         Send the initialization user data back up to the UI Process.
2156
2157 2012-07-06  Dongwoo Im  <dw.im@samsung.com>
2158
2159         [EFL][GTK] jhbuild : Disable pixman demos build depending on GTK+
2160         https://bugs.webkit.org/show_bug.cgi?id=90593
2161
2162         Reviewed by Philippe Normand.
2163
2164         * efl/jhbuild.modules: Give '--enable-gtk=no' option to pixman.
2165         * gtk/jhbuild.modules: Give '--enable-gtk=no' option to pixman.
2166
2167 2012-07-06  Luciano Wolf <luciano.wolf@openbossa.org>
2168
2169         [Qt] DumpRenderTree does not use 'monospace' font when directed
2170         https://bugs.webkit.org/show_bug.cgi?id=85203
2171
2172         The test fonts used for Qt tests were moved to the Liberation font family.
2173
2174         Due to this change we are skipping tons of tests. They will be gradually
2175         unskipped in batches, ASAP.
2176
2177         Also, the now unneeded special font configuration for Qt5 WK1 was removed.
2178
2179         Reviewed by Tor Arne Vestbø.
2180
2181         * DumpRenderTree/qt/QtInitializeTestFonts.cpp:
2182         (WebKit::initializeTestFonts):
2183
2184 2012-07-06  Csaba Osztrogonác  <ossy@webkit.org>
2185
2186         [Qt] Buildfix for newer Qt5
2187         https://bugs.webkit.org/show_bug.cgi?id=90519
2188
2189         Reviewed by Tor Arne Vestbø.
2190
2191         * MiniBrowser/qt/MiniBrowser.pro: We also need to depend on gui-private in the MiniBrowser to get access to these headers.
2192         * MiniBrowser/qt/MiniBrowserApplication.h: Include qpa/qwindowsysteminterface.h instead of deprecated qwindowsysteminterface_qpa.h
2193         * WebKitTestRunner/qt/PlatformWebViewQt.cpp: Include qpa/qwindowsysteminterface.h instead of deprecated qwindowsysteminterface_qpa.h
2194
2195 2012-07-06  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
2196
2197         [Qt] Make use of Qt5 qmake's changed makefile recursion behavior
2198
2199         When not using the -r option, qmake now interleaves qmake and make calls,
2200         so we don't need custom logic for this. There's also an option to supress
2201         the effect of the -r option, which we use to make WebKit.pro the only
2202         project file parsed in a recursive qmake-run.
2203
2204         https://bugs.webkit.org/show_bug.cgi?id=90461
2205
2206         Reviewed by Tor Arne Vestbø.
2207
2208         * qmake/mkspecs/features/default_post.prf:
2209         * qmake/mkspecs/features/functions.prf:
2210
2211 2012-07-06  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
2212
2213         [Qt] Remove custom qmake logic for module creation
2214
2215         Qmake now has the necessary hooks to cleanly override the build locations.
2216
2217         https://bugs.webkit.org/show_bug.cgi?id=90461
2218
2219         Reviewed by Tor Arne Vestbø.
2220
2221         * qmake/mkspecs/features/default_pre.prf:
2222
2223 2012-07-06  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
2224
2225         [Qt] Add top-level .qmake.conf
2226
2227         With Qt5, this makes setting $QMAKEPATH externally unnecessary.
2228
2229         The magic in the perl scripts to set QMAKEPATH is still there, as it doesn't
2230         hurt, and is still required for Qt4.
2231
2232         https://bugs.webkit.org/show_bug.cgi?id=90461
2233
2234         Reviewed by Tor Arne Vestbø.
2235
2236         * qmake/mkspecs/features/default_pre.prf:
2237
2238 2012-07-06  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
2239
2240         [Qt] Unify qtFeatureDefaults code paths
2241
2242         There's no reason to run qmake on features.prf directly anymore, as we
2243         can selectivly run configure tests in the project file now.
2244
2245         https://bugs.webkit.org/show_bug.cgi?id=90461
2246
2247         Reviewed by Tor Arne Vestbø.
2248
2249         * Scripts/webkitdirs.pm:
2250         (qtFeatureDefaults):
2251         * qmake/configure.pri:
2252         * qmake/mkspecs/features/features.prf:
2253
2254 2012-07-06  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
2255
2256         [Qt] Don't let qt_webkit.pri proclaim its own location
2257
2258         This won't work any more with recent Qt5 versions, as the forwarding
2259         pri is created by qt_module_config, which needs MODULE_PRI to be set
2260         up already.
2261
2262         We also need to load build_config, not qt_module.
2263
2264         https://bugs.webkit.org/show_bug.cgi?id=90461
2265
2266         Reviewed by Tor Arne Vestbø.
2267
2268         * qmake/qt_webkit.pri:
2269
2270 2012-07-06  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
2271
2272         [Qt] Remove redundant CONFIG+=module
2273
2274         qt_module_config takes care of that.
2275
2276         In api.pri we are actually testing the flag ourselves, so now we need to
2277         test a related flag qt_module_config sets instead.
2278
2279         Reviewed by Tor Arne Vestbø.
2280
2281         * qmake/mkspecs/features/default_post.prf:
2282
2283 2012-07-06  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
2284
2285         [Qt] Switch to new-style Qt 5 configure tests
2286
2287         Use explicit project file action instead of syncqt magic.
2288
2289         https://bugs.webkit.org/show_bug.cgi?id=90461
2290
2291         Reviewed by Tor Arne Vestbø.
2292
2293         * Scripts/webkitdirs.pm:
2294         (buildQMakeProjects):
2295         * qmake/configure.pri: Copied from Tools/qmake/configure.pro.
2296         * qmake/configure.pro:
2297         * qmake/mkspecs/features/default_pre.prf:
2298         * qmake/mkspecs/features/features.prf:
2299         * qmake/sync.profile: Removed.
2300
2301 2012-07-05  Jason Liu  <jason.liu@torchmobile.com.cn>
2302
2303         Add Jason Liu to committers.py.
2304         https://bugs.webkit.org/show_bug.cgi?id=90654
2305
2306         Unreviewed.
2307
2308         * Scripts/webkitpy/common/config/committers.py:
2309
2310 2012-07-05  Benjamin Poulain  <bpoulain@apple.com>
2311
2312         Double release of resources if the load is canceled in a callback of ResourceLoader::didFinishLoading
2313         https://bugs.webkit.org/show_bug.cgi?id=90431
2314
2315         Reviewed by Anders Carlsson.
2316
2317         Add a Mac API test.
2318
2319         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
2320         * TestWebKitAPI/Tests/mac/CancelLoadFromResourceLoadDelegate.html: Added.
2321         * TestWebKitAPI/Tests/mac/CancelLoadFromResourceLoadDelegate.mm: Added.
2322         (-[CancelLoadFromResourceLoadDelegate webView:resource:didFinishLoadingFromDataSource:]):
2323         (-[CancelLoadFromResourceLoadDelegateFrameLoadDelegate webView:didFinishLoadForFrame:]):
2324         (TestWebKitAPI):
2325         (TestWebKitAPI::TEST):
2326
2327 2012-07-05  Dave Tharp  <dtharp@codeaurora.org>
2328
2329         Adding myself as committer to committers.py
2330
2331         Unreviewed.
2332
2333         * Scripts/webkitpy/common/config/committers.py:
2334
2335 2012-07-05  Alexey Proskuryakov  <ap@apple.com>
2336
2337         [Mac][WK2] Enable HTTPS tests
2338         https://bugs.webkit.org/show_bug.cgi?id=90600
2339
2340         Reviewed by Brady Eidson.
2341
2342         * DumpRenderTree/mac/DumpRenderTree.mm: (dumpRenderTree): Removed a misleading comment.
2343         It explained why we were doing this here instead of a frame load delegate, but that's
2344         just history, not an explanation of why we need this code.
2345
2346         * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm: (WTR::InjectedBundle::platformInitialize):
2347         Ignore certificate errors for localhost and 127.0.0.1. Since this is only needed
2348         for tests, it seems appropriate to have platform specific solutions in every WTR,
2349         and not add a WebKit2 API.
2350
2351 2012-07-05  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2352
2353         [Qt] Remove rpath-hacks in qt_webkit.pri
2354
2355         Having it in qt_webkit.pri meant that all client apps would inherit the rpath,
2356         even if they were not using QtWebKit.
2357
2358         This is a regression for people who build client apps against a non-installed
2359         QtWebKit, as the client app will be missing a rpath to the standalone WebKit
2360         build dir. This can be solved by LD_LIBRARY_PATH or custom QMAKE_RPATHDIR
2361         logic, until the real issue is fixed in Qt.
2362
2363         Reviewed by Csaba Osztrogonác.
2364
2365         * qmake/mkspecs/features/rpath.prf:
2366         * qmake/qt_webkit.pri:
2367
2368 2012-07-05  Dongwoo Im  <dw.im@samsung.com>
2369
2370         [EFL] Enable the CUSTOM_SCHEME_HANDLER feature as default.
2371         https://bugs.webkit.org/show_bug.cgi?id=88608
2372
2373         Reviewed by Hajime Morita.
2374
2375         * Scripts/webkitperl/FeatureList.pm: Enable the CUSTOM_SCHEME_HANDLER feature as default.
2376
2377 2012-07-05  Ryuan Choi  <ryuan.choi@samsung.com>
2378
2379         [Wk2][EFL] EFL needs a WebKitTestRunner
2380         https://bugs.webkit.org/show_bug.cgi?id=87659
2381
2382         Reviewed by Kenneth Rohde Christiansen.
2383
2384         Add an implementation of WebKitTestRunner for Efl port.
2385
2386         * CMakeLists.txt: Includes WebKitTestRunner build script.
2387         * Scripts/build-webkittestrunner: Added knowledge of WebKitTestRunner/Efl.
2388         * Scripts/webkitpy/layout_tests/port/efl.py: Ditto.
2389         (EflPort.setup_environ_for_server):
2390         * WebKitTestRunner/CMakeLists.txt: Added.
2391         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
2392         * WebKitTestRunner/InjectedBundle/efl/ActivateFontsEfl.cpp: Added.
2393         * WebKitTestRunner/InjectedBundle/efl/InjectedBundleEfl.cpp: Added.
2394         * WebKitTestRunner/InjectedBundle/efl/LayoutTestControllerEfl.cpp: Added.
2395         * WebKitTestRunner/PlatformEfl.cmake: Added.
2396         * WebKitTestRunner/PlatformWebView.h:
2397         * WebKitTestRunner/TestController.cpp:
2398         * WebKitTestRunner/config.h: Added knowledge of WebKitTestRunner/Efl.
2399         * WebKitTestRunner/efl/PlatformWebViewEfl.cpp: Added.
2400         * WebKitTestRunner/efl/TestControllerEfl.cpp: Added.
2401         * WebKitTestRunner/efl/main.cpp: Added.
2402         (main):
2403
2404 2012-07-05  Christophe Dumez  <christophe.dumez@intel.com>
2405
2406         [WK2][EFL] Ewk_View needs API to load HTML data
2407         https://bugs.webkit.org/show_bug.cgi?id=90540
2408
2409         Reviewed by Kenneth Rohde Christiansen.
2410
2411         Update EFL MiniBrowser to catch the "load,error" signal
2412         on the view and display an error page.
2413
2414         * MiniBrowser/efl/main.c:
2415         (on_error):
2416         (browserCreate):
2417
2418 2012-07-05  Sergio Villar Senin  <svillar@igalia.com>
2419
2420         Hidden dirs are not copied when creating the built product archive
2421         https://bugs.webkit.org/show_bug.cgi?id=90559
2422
2423         Reviewed by Carlos Garcia Campos.
2424
2425         Hidden dirs are not copied inside the built product archive
2426         because the recursive copy command was ignoring them. We need those
2427         hidden directories for the GTK WebKit2 testing bot to work fine.
2428
2429         * BuildSlaveSupport/built-product-archive:
2430         (archiveBuiltProduct): replaced "*" by "." to include hidden
2431         dirs/files.
2432
2433 2012-07-04  Yoshifumi Inoue  <yosin@chromium.org>
2434
2435         Unreviewed Chromium gardening, Disable always failed python tests on Windows.
2436         https://bugs.webkit.org/show_bug.cgi?id=90587
2437
2438         PortTest.qt_linux and mac_lion are always failed on Chromium Windows at
2439         creating child process.
2440
2441         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2442         (PortTest.disabled_test_qt_linux):
2443         (PortTest.disabled_test_mac_lion):
2444
2445 2012-07-04  Yoshifumi Inoue  <yosin@chromium.org>
2446
2447         [Tools] webkit_unittest.py got assertion
2448         https://bugs.webkit.org/show_bug.cgi?id=90579
2449
2450         Reviewed by Hajime Morita.
2451
2452         This patch updates expectation of assertion in test_skipped_directories_for_symbols().
2453         r121874 added new (exclude) directory (inspector/styles/variables), however it didn't
2454         update one of two assertions.
2455
2456         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
2457
2458 2012-07-04  Balazs Ankes  <bank@inf.u-szeged.hu>
2459
2460         webkit-patch should add reviewer if "Reviewed by NOBODY ..." is missing
2461         https://bugs.webkit.org/show_bug.cgi?id=67935
2462
2463         Reviewed by Ryosuke Niwa.
2464
2465         * Scripts/webkitpy/common/checkout/changelog.py:
2466         (ChangeLog.set_reviewer):
2467         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
2468         (test_set_reviewer):
2469         (test_set_short_description_and_bug_url):
2470
2471 2012-07-04  Sheriff Bot  <webkit.review.bot@gmail.com>
2472
2473         Unreviewed, rolling out r121862.
2474         http://trac.webkit.org/changeset/121862
2475         https://bugs.webkit.org/show_bug.cgi?id=90569
2476
2477         It broke built-product-archive for GTK, Qt and EFL (Requested
2478         by svillar on #webkit).
2479
2480         * BuildSlaveSupport/built-product-archive:
2481         (archiveBuiltProduct):
2482
2483 2012-07-04  Sergio Villar Senin  <svillar@igalia.com>
2484
2485         Hidden dirs are not copied when creating the built product archive
2486         https://bugs.webkit.org/show_bug.cgi?id=90559
2487
2488         Reviewed by Carlos Garcia Campos.
2489
2490         Hidden dirs are not copied inside the built product archive because
2491         the recursive copy command was ignoring them. We need those hidden
2492         directories for the GTK WebKit2 testing bot to work fine.
2493
2494         * BuildSlaveSupport/built-product-archive:
2495         (archiveBuiltProduct): removed the trailing "*" from the copy
2496         command.
2497
2498 2012-07-04  John Mellor  <johnme@chromium.org>
2499
2500         Text Autosizing: Add compile flag and runtime setting
2501         https://bugs.webkit.org/show_bug.cgi?id=87394
2502
2503         This patch renames Font Boosting to Text Autosizing.
2504
2505         Reviewed by Adam Barth.
2506
2507         * Scripts/webkitperl/FeatureList.pm:
2508         * qmake/mkspecs/features/features.pri:
2509
2510 2012-07-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2511
2512         [Qt] Replace deprecated IN_PWD with PWD
2513
2514         Reviewed by Csaba Osztrogonác.
2515
2516         * qmake/mkspecs/features/default_pre.prf:
2517
2518 2012-07-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2519
2520         [Qt] Don't try to build WTR twice
2521
2522         Reviewed by Csaba Osztrogonác.
2523
2524         * Tools.pro:
2525
2526 2012-07-04  Sheriff Bot  <webkit.review.bot@gmail.com>
2527
2528         Unreviewed, rolling out r121821.
2529         http://trac.webkit.org/changeset/121821
2530         https://bugs.webkit.org/show_bug.cgi?id=90551
2531
2532         This patch did not receive a high-quality review and has a
2533         number of errors (Requested by abarth on #webkit).
2534
2535         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
2536         (Build.results_url):
2537         * Scripts/webkitpy/common/net/buildbot/chromiumbuildbot.py:
2538         (ChromiumBuilder.accumulated_results_url):
2539         * Scripts/webkitpy/layout_tests/port/builders.py:
2540         * Scripts/webkitpy/tool/commands/rebaseline.py:
2541         (AbstractParallelRebaselineCommand.__init__):
2542         (Rebaseline):
2543         (Rebaseline._builder_to_pull_from):
2544         (Rebaseline._tests_to_update):
2545         (Rebaseline.execute):
2546         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
2547         (test_rebaseline.mock_builder_to_pull_from):
2548         (test_rebaseline):
2549
2550 2012-07-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2551
2552         [Qt] Get rid of un-needed QT += declarative for Qt 5
2553
2554         The declarative module has been renamed to quick1 in Qt 5, and the
2555         engine-only module for Qt 5 is named 'qml'. For many of the instances
2556         we could just remove 'declarative', since the project file was only
2557         used for Qt5/WebKit2 builds. In the other cases the module was wrapped
2558         in a haveQt(4) scope.
2559
2560         Reviewed by Csaba Osztrogonác.
2561
2562         * MiniBrowser/qt/MiniBrowser.pro:
2563         * WebKitTestRunner/InjectedBundle/Target.pri:
2564         * WebKitTestRunner/Target.pri:
2565
2566 2012-07-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2567
2568         [Qt] Remove a few un-needed load(features) after r121777
2569
2570         The features are computed by configure.pro and cached in .qmake.cache.
2571
2572         Reviewed by Jocelyn Turcotte..
2573
2574         * MiniBrowser/qt/raw/Target.pri:
2575
2576 2012-07-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2577
2578         [Qt] Use haveQtModule() in project files instead of manual checks
2579
2580         Reviewed by Jocelyn Turcotte.
2581
2582         * qmake/mkspecs/features/default_pre.prf:
2583         * qmake/mkspecs/features/features.prf:
2584
2585 2012-07-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2586
2587         [Qt] Fix haveQtModule() check
2588
2589         Without {} brackets we were pulling out the value of the module.name variable.
2590
2591         Reviewed by Jocelyn Turcotte.
2592
2593         * qmake/mkspecs/features/functions.prf:
2594
2595 2012-07-04  Sergio Villar Senin  <svillar@igalia.com>
2596
2597         [WK2] [GTK] WK2 testing bot unable to find the InjectedBundle library
2598         https://bugs.webkit.org/show_bug.cgi?id=90541
2599
2600         Reviewed by Carlos Garcia Campos.
2601
2602         We must add the contents of WebKitBuild/*/Libraries/ directory to
2603         the file to be uploaded to b.w.o to successfully run WebKit2 tests
2604         in the WebKit2 testing bot.
2605
2606         * BuildSlaveSupport/built-product-archive:
2607         (archiveBuiltProduct): appended Libraries to the list of
2608         neededDirectories.
2609
2610 2012-07-04  Sergio Villar Senin  <svillar@igalia.com>
2611
2612         [GTK] WebKit2 testing bot unable to find the build directory
2613         https://bugs.webkit.org/show_bug.cgi?id=90523
2614
2615         Reviewed by Carlos Garcia Campos.
2616
2617         The presence of GNUmakefile was used to determine whether a given
2618         directory was a valid build directory or not. That's not correct for
2619         testing bots as we export only the minimum required stuff to run the
2620         tests.
2621
2622         * gtk/common.py:
2623         (get_build_path.is_valid_build_directory): added an extra check.:
2624
2625 2012-07-04  Joe Thomas  <joethomas@motorola.com>
2626
2627         Adding myself to committers list.
2628
2629         Unreviewed.
2630
2631         * Scripts/webkitpy/common/config/committers.py:
2632
2633 2012-07-03  Christophe Dumez  <christophe.dumez@intel.com>
2634
2635         [WK2][EFL] Ewk_View should report the load progress
2636         https://bugs.webkit.org/show_bug.cgi?id=90457
2637
2638         Reviewed by Kenneth Rohde Christiansen.
2639
2640         Update EFL MiniBrowser so that it listens for the
2641         "load,progress" on the Ewk_View and updates its
2642         window title accordingly.
2643
2644         * MiniBrowser/efl/main.c:
2645         (title_set):
2646         (on_title_changed):
2647         (on_progress):
2648         (browserCreate):
2649
2650 2012-07-03  Dirk Pranke  <dpranke@chromium.org>
2651
2652         disable failing mock-chromium-mac python unit tests
2653
2654         Unreviewed, build fix.
2655
2656         Disable these tests until I can figure out why they're failing
2657         on the bots.
2658
2659         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2660         (PortTest.disabled_test_chromium_mac_lion):
2661         (PortTest.disabled_test_chromium_mac_lion_in_test_shell_mode):
2662
2663 2012-07-03  Ian Vollick  <vollick@chromium.org>
2664
2665         Add vollick to the list of committers
2666         https://bugs.webkit.org/show_bug.cgi?id=90512
2667
2668         Reviewed by Adrienne Walker.
2669
2670         * Scripts/webkitpy/common/config/committers.py:
2671
2672 2012-07-03  Ojan Vafai  <ojan@chromium.org>
2673
2674         Improve webkit-patch rebaseline to work for more cases
2675         https://bugs.webkit.org/show_bug.cgi?id=90504
2676
2677         Reviewed by Dirk Pranke.
2678
2679         -Makes it work for the build.chromium.org bots.
2680         -Lets you rebaseline all builders instead of just one.
2681         -Lets you pass in the builder or test to rebaseline.
2682
2683         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
2684         (Build.results_url):
2685         Make this work for build.chromium.org builders as well.
2686
2687         * Scripts/webkitpy/common/net/buildbot/chromiumbuildbot.py:
2688         (ChromiumBuilder):
2689         (ChromiumBuilder.latest_cached_build):
2690         ChromiumBuilder doesn't support large parts of the Builder interface.
2691         This provides the bare minimum for this patch to work. Eventually,
2692         we should create a single interface that can be supported for both
2693         sets of buildbots.
2694
2695         * Scripts/webkitpy/layout_tests/port/builders.py:
2696         (builder_path_from_name):
2697         (all_builder_names):
2698         memoizing here is incorrect because the test override _exact_matches,
2699         so these can return different values. In either case, I'm pretty sure these
2700         are not remotely hot codepaths.
2701
2702         * Scripts/webkitpy/tool/commands/rebaseline.py:
2703         (AbstractParallelRebaselineCommand.__init__):
2704         (Rebaseline):
2705         (Rebaseline.__init__):
2706         (Rebaseline._builders_to_pull_from):
2707         (Rebaseline._builder_with_name):
2708         (Rebaseline._tests_to_update):
2709         (Rebaseline.execute):
2710         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
2711         (test_rebaseline.mock_builders_to_pull_from):
2712         (test_rebaseline):
2713         (test_rebaseline_command_line_flags):
2714         (test_rebaseline_multiple_builders):
2715         (test_rebaseline_multiple_builders.mock_builders_to_pull_from):
2716         (test_rebaseline_multiple_builders.mock_tests_to_update):
2717
2718 2012-07-03  Christophe Dumez  <christophe.dumez@intel.com>
2719
2720         [EFL] Enable CSS variables support at compile time
2721         https://bugs.webkit.org/show_bug.cgi?id=90448
2722
2723         Reviewed by Kenneth Rohde Christiansen.
2724
2725         Turn on CSS_VARIABLES flag by default on EFL port.
2726
2727         * Scripts/webkitperl/FeatureList.pm:
2728
2729 2012-07-03  Dirk Pranke  <dpranke@chromium.org>
2730
2731         nrwt: clean up exception handling and make sure we log some more failures
2732         https://bugs.webkit.org/show_bug.cgi?id=90503
2733
2734         Reviewed by Ojan Vafai.
2735
2736         There were several places where exceptions weren't getting
2737         logged, most notably if you passed a bad value to --platform.
2738         This change tests that and cleans things up a bit; more cleanup
2739         will be possible when we rework the manager_worker_broker code.
2740
2741         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
2742         (_BrokerConnection.raise_exception):
2743         (_InlineWorkerConnection.raise_exception):
2744         * Scripts/webkitpy/layout_tests/controllers/worker.py:
2745         (Worker.run):
2746         (Worker.kill_driver):
2747         * Scripts/webkitpy/layout_tests/port/factory.py:
2748         (PortFactory.get):
2749         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2750         (run):
2751         (main):
2752         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2753         (MainTest.test_unsupported_platfrom):
2754
2755 2012-07-03  Dirk Pranke  <dpranke@chromium.org>
2756
2757         nrwt: fix mock port
2758         https://bugs.webkit.org/show_bug.cgi?id=90500
2759
2760         Reviewed by Ojan Vafai.
2761
2762         The MockDRT code was never updated when we switched the chromium
2763         ports to using "drt mode" by default. This change updates that
2764         code, fixes a typo in the chromium port that went undetected
2765         (default_test_timeout_ms -> default_timeout_ms), and adds tests
2766         that actually exercise some of the mock ports. These tests are
2767         useful in that they will exercise the port-specific code in an
2768         end-to-end-manner, but they are a bit slow for some reason (>1s
2769         each) that I need to look into.
2770
2771         * Scripts/webkitpy/layout_tests/port/chromium.py:
2772         (ChromiumDriver.stop):
2773         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
2774         (ChromiumAndroidPort.default_timeout_ms):
2775         * Scripts/webkitpy/layout_tests/port/chromium_android_unittest.py:
2776         (TestChromiumAndroidPort.test_default_timeout_ms):
2777         * Scripts/webkitpy/layout_tests/port/mock_drt.py:
2778         (main):
2779         (parse_options):
2780         (MockTestShell):
2781         (MockTestShell.output_for_test):
2782         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
2783         (MockDRTTest.assertTest):
2784         (MockTestShellTest):
2785         (MockTestShellTest.make_drt):
2786         (MockTestShellTest.test_test_shell_parse_options):
2787         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2788         (PortTest):
2789         (PortTest.assert_mock_port_works):
2790         (PortTest.test_chromium_mac_lion):
2791         (PortTest.test_chromium_mac_lion_in_test_shell_mode):
2792         (PortTest.test_qt_linux):
2793         (PortTest.test_mac_lion):
2794
2795 2012-07-03  Dirk Pranke  <dpranke@chromium.org>
2796
2797         nrwt: make the worker class stand alone with a cleaner interface
2798         https://bugs.webkit.org/show_bug.cgi?id=90409
2799
2800         Reviewed by Ojan Vafai.
2801
2802         Currently the Worker class derives from AbstractWorker, which is
2803         kind of crufty and awkward; it would be better if we did not
2804         rely on shared state.
2805
2806         This change changes that so that Worker derives from object, and
2807         exposes the following interface:
2808           __init__() - called in the manager process
2809           safe_init() - called in the worker process to initialize
2810             unpicklable state
2811           handle() - a single routine to handle all messages
2812           cleanup() - called so the worker can clean up
2813
2814         Also, all of the "administrative" messages that are handled by
2815         the worker (notification of start/stop/etc.) move into
2816         manager_worker_broker - this reduces worker.py to just handling
2817         the mechanics of actually running each test.
2818
2819         For the moment, we do this by creating Yet Another wrapper/proxy
2820         class in manager_worker_broker, but this will get simpler
2821         shortly when the rest of m_w_b is cleaned up.
2822
2823         With this change worker is now in its new form but there will be
2824         a follow-on change that cleans up some names and other minor
2825         things.
2826
2827         This change is again mostly just moving things around and should
2828         be covered by the (updated) existing tests.
2829
2830         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
2831         (get):
2832         (AbstractWorker.__init__):
2833         (AbstractWorker.run):
2834         (AbstractWorker):
2835         (AbstractWorker.handle_stop):
2836         (AbstractWorker.handle_test_list):
2837         (AbstractWorker.yield_to_broker):
2838         (AbstractWorker.post_message):
2839         (_WorkerConnection.__init__):
2840         (_Process.run):
2841         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py:
2842         (_TestWorker):
2843         (_TestWorker.__init__):
2844         (_TestWorker.name):
2845         (_TestWorker.cleanup):
2846         (_TestWorker.handle):
2847         (_TestWorker.safe_init):
2848         (_TestWorker.stop):
2849         (_TestsMixin.handle_finished_test):
2850         (_TestsMixin.setUp):
2851         (_TestsMixin.test_cancel):
2852         (_TestsMixin.test_done):
2853         * Scripts/webkitpy/layout_tests/controllers/worker.py:
2854         (Worker):
2855         (Worker.__init__):
2856         (Worker.safe_init):
2857         (Worker.handle):
2858
2859 2012-07-03  Dirk Pranke  <dpranke@chromium.org>
2860
2861         nrwt: moving child process logging code into manager_worker_broker
2862         https://bugs.webkit.org/show_bug.cgi?id=90408
2863
2864         Reviewed by Ojan Vafai.
2865
2866         Users of manager_worker_broker should not have to be aware of
2867         whether they're in the same process or different processes and
2868         configure logging themselves; mwb should hide this complexity.
2869         We can't quite do this completely/correctly yet, since the
2870         manager expects to get a list of messages to log, but this
2871         change fixes the worker side of it, at least.
2872
2873         This is just moving code around, there is no new functionality
2874         and this should be covered by existing tests.
2875
2876         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
2877         (AbstractWorker.__init__):
2878         (_WorkerConnection.__init__):
2879         (_WorkerConnection.post_message):
2880         (_WorkerConnection):
2881         (_WorkerConnection.set_up_logging):
2882         (_WorkerConnection.clean_up_logging):
2883         (_InlineWorkerConnection.run):
2884         (_MultiProcessWorkerConnection.run):
2885         (_WorkerLogHandler):
2886         (_WorkerLogHandler.__init__):
2887         (_WorkerLogHandler.emit):
2888         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py:
2889         (_TestWorker.run):
2890         (_TestsMixin.handle_done):
2891         * Scripts/webkitpy/layout_tests/controllers/worker.py:
2892         (Worker.__init__):
2893         (Worker.run):
2894         (Worker._run_test):
2895         (Worker.cleanup):
2896         (Worker.run_single_test):
2897
2898 2012-07-03  Tony Chang  <tony@chromium.org>
2899
2900         [chromium] Don't archive build files generated by VS2010
2901         https://bugs.webkit.org/show_bug.cgi?id=90493
2902
2903         Reviewed by Ojan Vafai.
2904
2905         In particular, the pch (pre compiled header) files are gigantic, making
2906         release.zip almost a 1gb in size.
2907
2908         * BuildSlaveSupport/built-product-archive:
2909         (archiveBuiltProduct):
2910
2911 2012-07-03  Tony Chang  <tony@chromium.org>
2912
2913         [chromium] Unreviewed, update .gitignore to handle VS2010 files.
2914
2915         * .gitignore:
2916         * DumpRenderTree/DumpRenderTree.gyp/.gitignore:
2917         * TestWebKitAPI/TestWebKitAPI.gyp/.gitignore:
2918
2919 2012-07-03  Ojan Vafai  <ojan@chromium.org>
2920
2921         Have webkit-patch rebaseline use rebaseline-test-internal
2922         https://bugs.webkit.org/show_bug.cgi?id=90491
2923
2924         Reviewed by Dirk Pranke.
2925
2926         This lets it handle new files, reduces duplicate code and lays the
2927         groundwork for making rebaseline have a richer interface.
2928
2929         * Scripts/webkitpy/common/net/buildbot/buildbot_mock.py:
2930         (MockBuild):
2931         (MockBuild.__init__):
2932         (MockBuilder.build):
2933         * Scripts/webkitpy/tool/commands/rebaseline.py:
2934         (AbstractParallelRebaselineCommand._files_to_add):
2935         (Rebaseline):
2936         (Rebaseline._builder_to_pull_from):
2937         (Rebaseline._tests_to_update):
2938         (Rebaseline.execute):
2939         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
2940         (test_overrides_are_included_correctly):
2941         (test_rebaseline):
2942         (test_rebaseline.mock_builder_to_pull_from):
2943         (test_rebaseline.mock_tests_to_update):
2944
2945 2012-07-03  Ojan Vafai  <ojan@chromium.org>
2946
2947         Rename rebaseline-test to rebaseline-test-internal
2948         https://bugs.webkit.org/show_bug.cgi?id=90485
2949
2950         Reviewed by Adam Barth.
2951
2952         It's now only used by other rebaseline commands. It's still useful to leave it
2953         as it's own command to aid in debugging when something goes wrong. In a followup
2954         patch, I'll make webkit-patch rebaseline cover any use-cases that rebaseline-test
2955         might have covered.
2956
2957         We no longer need the --print-scm-changes option since the only caller always passes
2958         that option in.
2959
2960         Also, make all the arguments command-line flags instead. Simplifies the code
2961         a bit in my opinion.
2962
2963         * Scripts/webkitpy/tool/commands/rebaseline.py:
2964         (RebaselineTest):
2965         (RebaselineTest.__init__):
2966         (RebaselineTest.execute):
2967         (AbstractParallelRebaselineCommand._rebaseline_commands):
2968         (RebaselineExpectations.execute):
2969         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
2970         (test_rebaseline_all):
2971
2972 2012-07-03  Balazs Kelemen  <kbalazs@webkit.org>
2973
2974         [Qt][WTR] Get rid of using DumpRenderTreeSupportQt
2975         https://bugs.webkit.org/show_bug.cgi?id=90262
2976
2977         Unreviewed build fix.
2978
2979         Reverting r121550 manually because it caused build
2980         break on ARM and Mac. It turned out that the Qt build
2981         system is too broken for this change now, first we have
2982         to address two issues:
2983             * symbol hiding was forbidden accidentally on x86 Linux
2984             * Qt lacks a separate library for test support
2985
2986         * Tools.pro:
2987         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2988         (WTR::InjectedBundlePage::resetAfterTest):
2989         (WTR::InjectedBundlePage::didClearWindowForFrame):
2990         * WebKitTestRunner/InjectedBundle/qt/ActivateFontsQt.cpp:
2991
2992 2012-07-03  Tony Chang  <tony@chromium.org>
2993
2994         [chromium] Handle trailing backslash on %VSINSTALLDIR%
2995         https://bugs.webkit.org/show_bug.cgi?id=90410
2996
2997         Reviewed by Ojan Vafai.
2998
2999         * Scripts/webkitdirs.pm:
3000         (buildChromiumVisualStudioProject): Handle VSINSTALLDIR ending in a backslash, the backslash escapes the closing double quote.
3001
3002 2012-07-02  Ojan Vafai  <ojan@chromium.org>
3003
3004         Make the skia_test_expectations.txt file optional.
3005         https://bugs.webkit.org/show_bug.cgi?id=90400
3006
3007         Reviewed by Dirk Pranke.
3008
3009         It used to be optional. This regressed at some point. It's important that it be
3010         optional so that webkit-patch commands work in a pure-webkit checkout for chromium bots.
3011         Specifically, this was breaking webkit-patch rebaseline-test when it would go to update
3012         TestExpectations.
3013
3014         * Scripts/webkitpy/layout_tests/port/chromium.py:
3015         (ChromiumPort.expectations_files):
3016         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
3017         (ChromiumDriverTest.test_expectations_dict):
3018
3019 2012-07-03  Raphael Kubo da Costa  <rakuco@webkit.org>
3020
3021         [jhbuild][EFL] Bump libffi dependency.
3022         https://bugs.webkit.org/show_bug.cgi?id=85373
3023
3024         Reviewed by Gustavo Noronha Silva.
3025
3026         Update libffi to 3.0.11, which fixes the build on some platforms
3027         where 3.0.10 had problems.
3028
3029         * efl/jhbuild.modules:
3030
3031 2012-07-03  George Staikos  <staikos@webkit.org>
3032
3033         [BlackBerry] Enable microdata support for BlackBerry.
3034         https://bugs.webkit.org/show_bug.cgi?id=90429
3035
3036         Reviewed by Rob Buis.
3037
3038         * Scripts/webkitperl/FeatureList.pm:
3039
3040 2012-07-03  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3041
3042         [Qt] Make use of .qmake.cache for caching features
3043
3044         Instead of loading() features from the files that need them (and re-running
3045         a bunch of checks), we now run feature detection as part of configure.pro,
3046         and have build-webkit write the computed feature-defines and CONFIG to
3047         .qmake.cache, which is then loaded by qmake _before_ even defaults_pre
3048         when building WebKit.pro.
3049
3050         At some point we'll be able to selectivly prevent running of config tests
3051         in configure.pro, which means we don't need a separate code-path for
3052         the build-webkit --help case.
3053
3054         We should also move the code in build-webkit that now uses .webkit.config
3055         to detect clean builds, to use .qmake.cache, since we now store the same
3056         thing there.
3057
3058         Original patch by Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
3059
3060         Reviewed by Tor Arne Vestbø.
3061
3062         * DumpRenderTree/qt/DumpRenderTree.pro:
3063         * QtTestBrowser/QtTestBrowser.pro:
3064         * Scripts/webkitdirs.pm:
3065         (qtFeatureDefaults):
3066         (buildQMakeProjects):
3067         * Tools.pro:
3068         * WebKitTestRunner/InjectedBundle/DerivedSources.pri:
3069         * WebKitTestRunner/InjectedBundle/Target.pri:
3070         * WebKitTestRunner/Target.pri:
3071         * qmake/.qmake.conf: Added.
3072         * qmake/configure.pro:
3073         * qmake/mkspecs/features/default_post.prf:
3074         * qmake/mkspecs/features/default_pre.prf:
3075         * qmake/mkspecs/features/features.prf:
3076
3077 2012-07-03  Nico Weber  <thakis@chromium.org>
3078
3079         -Wunused-private-field cleanup, linux edition
3080         https://bugs.webkit.org/show_bug.cgi?id=90463
3081
3082         Reviewed by Stephen White.
3083
3084         * DumpRenderTree/chromium/DRTDevToolsAgent.h:
3085         * DumpRenderTree/chromium/GamepadController.h:
3086         (GamepadController):
3087         * DumpRenderTree/chromium/NotificationPresenter.h:
3088         (NotificationPresenter::NotificationPresenter):
3089         (NotificationPresenter):
3090
3091 2012-07-03  Sheriff Bot  <webkit.review.bot@gmail.com>
3092
3093         Unreviewed, rolling out r121766.
3094         http://trac.webkit.org/changeset/121766
3095         https://bugs.webkit.org/show_bug.cgi?id=90465
3096
3097         It caused flakey build errors on the bots (Requested by Ossy
3098         on #webkit).
3099
3100         * DumpRenderTree/qt/DumpRenderTree.pro:
3101         * QtTestBrowser/QtTestBrowser.pro:
3102         * Scripts/webkitdirs.pm:
3103         (qtFeatureDefaults):
3104         (buildQMakeProjects):
3105         * Tools.pro:
3106         * WebKitTestRunner/InjectedBundle/DerivedSources.pri:
3107         * WebKitTestRunner/InjectedBundle/Target.pri:
3108         * WebKitTestRunner/Target.pri:
3109         * qmake/.qmake.conf: Removed.
3110         * qmake/configure.pro:
3111         * qmake/mkspecs/features/default_post.prf:
3112         * qmake/mkspecs/features/default_pre.prf:
3113         * qmake/mkspecs/features/features.prf:
3114
3115 2012-07-03  George Staikos  <staikos@webkit.org>
3116
3117         [BlackBerry] Enable Custom Scheme Handlers for BlackBerry.
3118         https://bugs.webkit.org/show_bug.cgi?id=90422
3119
3120         Reviewed by Rob Buis.
3121
3122         * Scripts/webkitperl/FeatureList.pm:
3123
3124 2012-07-03  George Staikos  <staikos@webkit.org>
3125
3126         [BlackBerry] Enable RegisterProtocolHandler for BlackBerry.
3127         https://bugs.webkit.org/show_bug.cgi?id=90422
3128
3129         Reviewed by Rob Buis.
3130
3131         * Scripts/webkitperl/FeatureList.pm:
3132
3133 2012-07-03  Thiago Marcos P. Santos  <thiago.santos@intel.com>
3134
3135         [EFL] Add run API tests as a step on the build bots
3136         https://bugs.webkit.org/show_bug.cgi?id=87252
3137
3138         Reviewed by Daniel Bates.
3139
3140         Run EFL API test suite on the build bots.
3141
3142         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3143         (RunEflAPITests):
3144         (TestFactory.__init__):
3145         (BuildAndTestFactory.__init__):
3146         * Scripts/run-efl-tests: Added.
3147
3148 2012-07-03  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3149
3150         [Qt] Make use of .qmake.cache for caching features
3151
3152         Instead of loading() features from the files that need them (and re-running
3153         a bunch of checks), we now run feature detection as part of configure.pro,
3154         and have build-webkit write the computed feature-defines and CONFIG to
3155         .qmake.cache, which is then loaded by qmake _before_ even defaults_pre
3156         when building WebKit.pro.
3157
3158         At some point we'll be able to selectivly prevent running of config tests
3159         in configure.pro, which means we don't need a separate code-path for
3160         the build-webkit --help case.
3161
3162         We should also move the code in build-webkit that now uses .webkit.config
3163         to detect clean builds, to use .qmake.cache, since we now store the same
3164         thing there.
3165
3166         Original patch by Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
3167
3168         Reviewed by Tor Arne Vestbø.
3169
3170         * DumpRenderTree/qt/DumpRenderTree.pro:
3171         * QtTestBrowser/QtTestBrowser.pro:
3172         * Scripts/webkitdirs.pm:
3173         (qtFeatureDefaults):
3174         (buildQMakeProjects):
3175         * Tools.pro:
3176         * WebKitTestRunner/InjectedBundle/DerivedSources.pri:
3177         * WebKitTestRunner/InjectedBundle/Target.pri:
3178         * WebKitTestRunner/Target.pri:
3179         * qmake/.qmake.conf: Added.
3180         * qmake/configure.pro:
3181         * qmake/mkspecs/features/default_post.prf:
3182         * qmake/mkspecs/features/default_pre.prf:
3183         * qmake/mkspecs/features/features.prf:
3184
3185 2012-07-03  Jocelyn Turcotte  <jocelyn.turcotte@nokia.com>  Joel Dillon <joel.dillon@codethink.co.uk>
3186
3187         [Qt][Win] Fix broken QtWebKit5.lib linking
3188         https://bugs.webkit.org/show_bug.cgi?id=88321
3189
3190         Reviewed by Kenneth Rohde Christiansen.
3191
3192         On windows the import/export definition is part of the symbol's signature.
3193         Define STATICALLY_LINKED_WITH_$$library for each dependend module
3194         that is being linked statically to make sure that they can be linked together
3195         later on.
3196
3197         Also do not compile Assertions.cpp in DumpRenderTree anymore since all the
3198         used symbols are exported and it would cause a duplicate symbols error.
3199
3200         * DumpRenderTree/qt/DumpRenderTree.pro:
3201         * qmake/mkspecs/features/default_post.prf:
3202
3203 2012-07-03  Szilard Ledan  <szledan@inf.u-szeged.hu>
3204
3205         master.cfg cleanup, pass BuildStep instances instead of BuildStep subclasses
3206         https://bugs.webkit.org/show_bug.cgi?id=89001
3207
3208         Reviewed by Csaba Osztrogonác.
3209
3210         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3211         (UploadBuiltProduct.__init__):
3212         (UploadTestResults.__init__):
3213         (ExtractTestResults.__init__):
3214         (Factory.__init__):
3215         (BuildFactory.__init__):
3216         (TestFactory.__init__):
3217         (BuildAndTestFactory.__init__):
3218         (BuildAndPerfTestFactory.__init__):
3219         (BuildAndPerfTestWebKit2Factory.__init__):
3220         (DownloadAndPerfTestFactory.__init__):
3221
3222 2012-07-03  Ádám Kallai  <kadam@inf.u-szeged.hu>
3223
3224         Update unit test after r121724.
3225         https://bugs.webkit.org/show_bug.cgi?id=90444
3226
3227         Reviewed by Csaba Osztrogonác.
3228
3229         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py:
3230         (GardeningServerTest.test_rebaseline_new_port):
3231
3232 2012-07-03  Sheriff Bot  <webkit.review.bot@gmail.com>
3233
3234         Unreviewed, rolling out r121749.
3235         http://trac.webkit.org/changeset/121749
3236         https://bugs.webkit.org/show_bug.cgi?id=90441
3237
3238         Tests crash on the paralel NRWT Qt bot (Requested by Ossy on
3239         #webkit).
3240
3241         * Scripts/webkitpy/layout_tests/port/qt.py:
3242         (QtPort._build_driver):
3243         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py:
3244         (XvfbDriver._start):
3245         (XvfbDriver._start.x_filter):
3246         (XvfbDriver.stop):
3247
3248 2012-07-03  Kristóf Kosztyó  <kkristof@inf.u-szeged.hu>
3249
3250         [NRWT] XvfbDriver should choose the next free display
3251         https://bugs.webkit.org/show_bug.cgi?id=88414
3252
3253         Reviewed by Dirk Pranke.
3254
3255         * Scripts/webkitpy/layout_tests/port/qt.py:
3256         (QtPort._driver_class):
3257         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py:
3258         (XvfbDriver.__init__):
3259         (XvfbDriver._start):
3260         (XvfbDriver._start.next_free_id):
3261         (XvfbDriver.stop):
3262
3263 2012-07-03  Christophe Dumez  <christophe.dumez@intel.com>
3264
3265         [EFL] Enable MICRODATA support
3266         https://bugs.webkit.org/show_bug.cgi?id=90377
3267
3268         Reviewed by Kenneth Rohde Christiansen.
3269
3270         Turn on MICRODATA flag for EFL port by default.
3271
3272         * Scripts/webkitperl/FeatureList.pm:
3273
3274 2012-07-02  George Staikos  <staikos@webkit.org>
3275
3276         [BlackBerry] Enable scoped style for BlackBerry.
3277         https://bugs.webkit.org/show_bug.cgi?id=90418
3278
3279         Reviewed by Rob Buis.
3280
3281         * Scripts/webkitperl/FeatureList.pm:
3282
3283 2012-07-02  Xiaobo Wang  <xbwang@torchmobile.com.cn>
3284
3285         [BlackBerry] Use PUBLIC_BUILD to enable/disable DRT
3286         https://bugs.webkit.org/show_bug.cgi?id=90271
3287
3288         Reviewed by George Staikos.
3289
3290         RIM PR #154707
3291
3292         Currently DRT code will be compiled only if ENABLE_DRT is set, and it's not
3293         defined by default.
3294         We should enable DRT by default unless PUBLIC_BUILD is set. In this way we don't
3295         need to rebuild webkit before running DRT.
3296
3297         * Scripts/webkitdirs.pm:
3298         (blackberryCMakeArguments):
3299
3300 2012-07-02  Ojan Vafai  <ojan@chromium.org>
3301
3302         webkit-patch rebaseline-expectations should share code with rebaseline-all
3303         https://bugs.webkit.org/show_bug.cgi?id=90413
3304
3305         Reviewed by Dirk Pranke.
3306
3307         Make them share code. In addition to reducing code duplication this makes
3308         rebaseline-expectations considerably faster by rebaselining in parallel.
3309
3310         * Scripts/webkitpy/tool/commands/rebaseline.py:
3311         (AbstractParallelRebaselineCommand):
3312         (AbstractParallelRebaselineCommand._run_webkit_patch):
3313         (AbstractParallelRebaselineCommand._rebaseline):
3314         (RebaselineJson):
3315         (RebaselineJson.execute):
3316         (RebaselineExpectations):
3317         (RebaselineExpectations._update_expectations_file):
3318         (RebaselineExpectations._tests_to_rebaseline):
3319         (RebaselineExpectations._add_tests_to_rebaseline_for_port):
3320         (RebaselineExpectations.execute):
3321         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
3322         (test_rebaseline_all):
3323         (test_rebaseline_expectations.run_in_parallel):
3324         (test_rebaseline_expectations):
3325
3326 2012-07-02  Xiaobo Wang  <xbwang@torchmobile.com.cn>
3327
3328         [BlackBerry] Update DumpRenderTree to have it work interactively in parallel
3329         https://bugs.webkit.org/show_bug.cgi?id=88326
3330
3331         Reviewed by Rob Buis.
3332
3333         1. Test name is sent to torch-launcher via PPS message(from host machine).
3334            So we get test list by monitoring and receiving PPS message instead of
3335            reading file index.drt.
3336         2. Torch-launcher create a <test file>.done file when it finished a test.
3337         3. We don't need to search for Ref-tests in DumpRenderTree.cpp any more. NRWT
3338            will get them for us.
3339
3340         * DumpRenderTree/blackberry/DumpRenderTree.cpp:
3341         (BlackBerry::WebKit::DumpRenderTree::DumpRenderTree):
3342         (BlackBerry::WebKit::DumpRenderTree::doneDrt):
3343         (BlackBerry::WebKit::DumpRenderTree::runRemainingTests):
3344         (BlackBerry::WebKit::DumpRenderTree::ensurePPS):
3345         (WebKit):
3346         (BlackBerry::WebKit::DumpRenderTree::handlePPSData):
3347         (BlackBerry::WebKit::DumpRenderTree::waitForTest):
3348         (BlackBerry::WebKit::DumpRenderTree::runTests):
3349         (BlackBerry::WebKit::DumpRenderTree::dump):
3350         * DumpRenderTree/blackberry/DumpRenderTreeBlackBerry.h:
3351         (DumpRenderTree):
3352
3353 2012-07-02  Adam Barth  <abarth@webkit.org>
3354
3355         Remove flashplugin-installer from the EWS image because it causes some tests to crash
3356         https://bugs.webkit.org/show_bug.cgi?id=90403
3357
3358         Reviewed by Tony Chang.
3359
3360         * EWSTools/cold-boot.sh:
3361
3362 2012-07-02  Leandro Gracia Gil  <leandrogracia@chromium.org>
3363
3364         [Chromium] Implement a Layout Test for editing/SurroundingText
3365         https://bugs.webkit.org/show_bug.cgi?id=82461
3366
3367         Reviewed by Ryosuke Niwa.
3368
3369         Add a new method to the layout test controller in order to retrieve the
3370         text surrounding a provided element.
3371
3372         * DumpRenderTree/chromium/LayoutTestController.cpp:
3373         (LayoutTestController::LayoutTestController):
3374         (LayoutTestController::textSurroundingElement):
3375         * DumpRenderTree/chromium/LayoutTestController.h:
3376         (LayoutTestController):
3377
3378 2012-07-02  Ojan Vafai  <ojan@chromium.org>
3379
3380         Delete unused rebaseline method in gardeningserver.py
3381         https://bugs.webkit.org/show_bug.cgi?id=90396
3382
3383         Reviewed by Eric Seidel.
3384
3385         As best I can tell, the only usage was in this unittest.
3386
3387         * Scripts/webkitpy/tool/servers/gardeningserver.py:
3388         (GardeningHTTPRequestHandler.updateexpectations):
3389         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py:
3390         (GardeningServerTest.test_rebaseline_new_port):
3391
3392 2012-07-02  Csaba Osztrogonác  <ossy@webkit.org>
3393
3394         REGRESSION(r121497): It switched off and broke many unittests
3395         https://bugs.webkit.org/show_bug.cgi?id=90371
3396
3397         Reviewed by Adam Barth.
3398
3399         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
3400         (ChromiumMacPortTest): Use snowleopard as os_version instead of leopard, because it isn't supported anymore.
3401         * Scripts/webkitpy/layout_tests/port/chromium_port_testcase.py:
3402         (ChromiumPortTestCase.test_all_test_configurations): Remove leopard testcases, because it isn't supported anymore.
3403         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
3404         (PortTestCase): Inherit class PortTestCase from unittest.TestCase instead of object.
3405
3406 2012-07-02  Ojan Vafai  <ojan@chromium.org>
3407
3408         Move rebaseline-all command from the gardening-server down into webkit-patch
3409         https://bugs.webkit.org/show_bug.cgi?id=90395
3410
3411         Reviewed by Adam Barth.
3412
3413         This is just moving code. It it in preparation for making rebaseline-expectations
3414         use the same code in order to get the parallelism benefits and reduces the amount
3415         of code we have for doing rebaselines.
3416
3417         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
3418         (CheckoutTest.test_apply_patch):
3419         Updated due to the change to executive_mock.
3420         * Scripts/webkitpy/common/system/executive_mock.py:
3421         (MockExecutive.run_command):
3422         Update to print out the input passed to stdin.
3423         * Scripts/webkitpy/tool/commands/download_unittest.py:
3424         Updated due to executive_mock change.
3425         * Scripts/webkitpy/tool/commands/rebaseline.py:
3426         (RebaselineAll):
3427         (RebaselineAll._run_webkit_patch):
3428         (RebaselineAll._builders_to_fetch_from):
3429         (RebaselineAll._rebaseline_commands):
3430         (RebaselineAll._files_to_add):
3431         (RebaselineAll._optimize_baselines):
3432         (RebaselineAll._rebaseline):
3433         (RebaselineAll.execute):
3434         All this code is just copy-pasted except for mechanical changes
3435         (e.g. self.server.tool --> self._tool) and the reading in of the 
3436         JSON from stdin instead of the post body.
3437         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
3438         (test_rebaseline_all):
3439         Copied the test-case out of gardeningserver_unittest.py.
3440         * Scripts/webkitpy/tool/servers/gardeningserver.py:
3441         (GardeningHTTPRequestHandler):
3442         (GardeningHTTPRequestHandler.rebaseline):
3443         (GardeningHTTPRequestHandler.rebaselineall):
3444         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py:
3445         (GardeningServerTest.test_rebaselineall):
3446         (GardeningServerTest.test_rebaselineall.run_command):
3447
3448 2012-07-02  Ojan Vafai  <ojan@chromium.org>
3449
3450         Remove Leopard support from the flakiness dashboard
3451         https://bugs.webkit.org/show_bug.cgi?id=90390
3452
3453         Reviewed by Adam Barth.
3454
3455         * TestResultServer/static-dashboards/flakiness_dashboard.js:
3456         (nonChromiumPlatform):
3457         (chromiumPlatform):
3458         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
3459         (test):
3460         * TestResultServer/static-dashboards/run-embedded-unittests.html:
3461         * TestResultServer/static-dashboards/run-unittests.html:
3462
3463 2012-07-02  Ojan Vafai  <ojan@chromium.org>
3464
3465         Fix posting from garden-o-matic. This broke in moving away from jquery's ajax method
3466         in http://trac.webkit.org/changeset/121617.
3467         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/net.js:
3468
3469 2012-07-02  Carlos Garcia Campos  <cgarcia@igalia.com>
3470
3471         [GTK] Read fonts path when running layout tests from alternative fonts dir when main dir doesn't exist
3472         https://bugs.webkit.org/show_bug.cgi?id=89437
3473
3474         Reviewed by Martin Robinson.
3475
3476         If main fonts directory doesn't exist, try with an alternative
3477         fonts directory at build directory.
3478
3479         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3480         (getOutputDir):
3481         (getFontsPath):
3482         (initializeFonts):
3483         * WebKitTestRunner/InjectedBundle/gtk/ActivateFontsGtk.cpp:
3484         (WTR::getOutputDir):
3485         (WTR):
3486         (WTR::getFontsPath):
3487         (WTR::inititializeFontConfigSetting):
3488
3489 2012-07-02  Carlos Garcia Campos  <cgarcia@igalia.com>
3490
3491         [GTK] Don't run the tests with jhbuild wrapper if it's already running under jhbuild
3492         https://bugs.webkit.org/show_bug.cgi?id=89435
3493
3494         Reviewed by Martin Robinson.
3495
3496         * Scripts/new-run-webkit-tests: Don't run the tests with the
3497         jhbuild wrapper if there isn't a Dependencies directory inside the
3498         build directory.
3499
3500 2012-06-29  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
3501
3502         [Qt] Simplify detection of non-installed module file
3503
3504         Has the additional advantage that we do not rely on additional information.
3505
3506         Reviewed by Tor Arne Vestbø.
3507
3508         * qmake/qt_webkit.pri:
3509
3510 2012-07-01  Christophe Dumez  <christophe.dumez@intel.com>
3511
3512         [EFL] Add Gamepad support
3513         https://bugs.webkit.org/show_bug.cgi?id=90170
3514
3515         Reviewed by Kenneth Rohde Christiansen.
3516
3517         * Scripts/webkitperl/FeatureList.pm: Turn on GAMEPAD flag
3518         by default for EFL port.
3519         * efl/jhbuild.modules: Bump dependency versions of EFL libs
3520         since the latest Eeze is required for gamepad support.
3521
3522 2012-07-01  Adam Barth  <abarth@webkit.org>
3523
3524         Unreviewed.
3525
3526         Using the default start cylinder for fdisk causes a warning about the
3527         partition not starting on physical sector boundary.  The Ubuntu forums
3528         recommend using a number that's divisible by 8, which is what we do in
3529         this patch.
3530
3531         * EWSTools/cold-boot.sh:
3532
3533 2012-07-01  Adam Barth  <abarth@webkit.org>
3534
3535         Unreviewed.
3536
3537         It turns out we need to use Ubuntu 10.04 to get the right image results
3538         for chromium-linux. This patch updates our scripts to be compatible
3539         with Ubuntu 10.04.
3540
3541         * EWSTools/cold-boot.sh:
3542             - fdisk doesn't have p and 1 as default commands in 10.04.
3543         * EWSTools/start-queue.sh:
3544             - git doesn't understanding the -B argument in 10.04. We've been
3545               using this change locally on the EC2 bots for a while.
3546
3547 2012-07-01  Adam Barth  <abarth@webkit.org>
3548
3549         Unreviewed.
3550
3551         * EWSTools/boot.sh:
3552             - We need to start the screen in detached mode so that we can run
3553               it remotely via ssh.
3554
3555 2012-07-01  Adam Barth  <abarth@webkit.org>
3556
3557         Unreviewed.
3558
3559         * EWSTools/cold-boot.sh:
3560             - Turns out we need to sudo this command in order for it to
3561               actually dimiss the EULA screen. :)
3562
3563 2012-06-29  Adam Barth  <abarth@webkit.org>
3564
3565         Turns out we need zip too.
3566
3567         * EWSTools/cold-boot.sh:
3568
3569 2012-06-29  Adam Barth  <abarth@webkit.org>
3570
3571         Add a cold-boot.sh script for the EWS
3572         https://bugs.webkit.org/show_bug.cgi?id=90330
3573
3574         Unreviewed.
3575
3576         * EWSTools/cold-boot.sh: Added.
3577             - This script can take us from a cold GCE instance to a running EWS
3578               bot in one fell swoop.
3579         * EWSTools/start-queue.sh:
3580             - The if-block at the top of this script was causing trouble. I
3581               removed it from the bots a while ago. Now that we're using SVN to
3582               cold-boot the EWS bots, we need to make this change in the repo.
3583
3584 2012-06-29  Luiz Agostini  <luiz.agostini@nokia.com>
3585
3586         [Qt][WK2] Private non-QtQuick API
3587         https://bugs.webkit.org/show_bug.cgi?id=84532
3588
3589         MiniBrowserRaw is an usage example for the QRawWebView API.
3590         It is only an example and is not fully implemented.
3591
3592         Reviewed by Noam Rosenthal.
3593
3594         * MiniBrowser/qt/raw/DerivedSources.pri: Added.
3595         * MiniBrowser/qt/raw/MiniBrowserRaw.pro: Added.
3596         * MiniBrowser/qt/raw/Target.pri: Added.
3597         * MiniBrowser/qt/raw/View.cpp: Added.
3598         * MiniBrowser/qt/raw/View.h: Added.
3599         * Scripts/webkitpy/style/checker.py:
3600         * Tools.pro:
3601
3602 2012-06-29  Ojan Vafai  <ojan@chromium.org>
3603
3604         Remove leopard bots from garden-o-matic
3605         https://bugs.webkit.org/show_bug.cgi?id=90328
3606
3607         Reviewed by Adam Barth.
3608
3609         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/config.js:
3610         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/failures_unittests.js:
3611         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/notifications_unittests.js:
3612
3613 2012-06-29  Ojan Vafai  <ojan@chromium.org>
3614
3615         garden-o-matic broken: TypeError: 'undefined' is not an object (evaluating 'buildLocations[currentIndex].url')
3616         https://bugs.webkit.org/show_bug.cgi?id=90243
3617
3618         Reviewed by Dirk Pranke.
3619
3620         jQuery was trying to be too smart and parsing the jsonp as json because of it's content-type.
3621         Excise jQuery and just use XHR directly since it's easier to maintain something where we control it
3622         all.
3623
3624         Not really sure how to unittest this. I tested it all manually of course.
3625
3626         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/net.js:
3627         Made net.ajax a drop-in replacement for the features of $.ajax that we were using.
3628         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui_unittests.js:
3629         Not related to this patch, but figured I'd update the failing test while I was at it.
3630
3631 2012-06-29  Yaron Friedman  <yfriedman@chromium.org>
3632
3633         Fix layout test runner for Android after https://bugs.webkit.org/show_bug.cgi?id=88134
3634         https://bugs.webkit.org/show_bug.cgi?id=90309
3635
3636         Reviewed by Adam Barth.
3637
3638         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
3639         (ChromiumAndroidPort.start_http_server):
3640
3641 2012-06-29  James Weatherall  <wez@chromium.org>
3642
3643         NPObjectWrapper may not address all window script object lifetime issues
3644         https://bugs.webkit.org/show_bug.cgi?id=85679
3645
3646         TestNetscapePlugin now has a leak-window-scriptable-object test which takes a reference to the window script object, and a second reference to it via the "self" property, and does not release those references. This is used to simulate a leaky plugin in layout tests of the NPAPI scripting interface glue code.
3647
3648         Reviewed by Nate Chapin.
3649
3650         * DumpRenderTree/DumpRenderTree.gypi:
3651         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
3652         (PluginTest::NPN_GetProperty):
3653         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
3654         (PluginTest):
3655         * DumpRenderTree/TestNetscapePlugIn/Tests/LeakWindowScriptableObject.cpp: Added.
3656         (LeakWindowScriptableObject):
3657         (LeakWindowScriptableObject::LeakWindowScriptableObject):
3658         (LeakWindowScriptableObject::NPP_New):
3659
3660 2012-06-29  Dirk Pranke  <dpranke@chromium.org>
3661
3662         webkitpy: add comment about how determine_full_port_name() works for apple ports, fix -wk2 bug
3663         https://bugs.webkit.org/show_bug.cgi?id=90314
3664
3665         Reviewed by Ojan Vafai.
3666
3667         Add comments and fix a bug in how we would handle the mac-wk2
3668         and win-wk2 port names after confusion around in bug 90312 :).
3669
3670         * Scripts/webkitpy/layout_tests/port/apple.py:
3671         (ApplePort.determine_full_port_name):
3672
3673 2012-06-29  Dirk Pranke  <dpranke@chromium.org>
3674
3675         webkitpy: remove support for mac leopard from chromium configurations
3676         https://bugs.webkit.org/show_bug.cgi?id=90313
3677
3678         Reviewed by Tony Chang.
3679
3680         Google has shipped the last version of Chrome that will support
3681         Mac OS 10.5 (Leopard), and we no longer have bots that run this
3682         configuration, so we're removing support for it.
3683
3684         A subsequent change will remove the baselines in platform/chromium-mac-leopard.
3685
3686         * Scripts/webkitpy/common/checkout/baselineoptimizer_unittest.py:
3687         (BaselineOptimizerTest.test_complex_shadowing):
3688         * Scripts/webkitpy/layout_tests/port/builders.py:
3689         * Scripts/webkitpy/layout_tests/port/chromium.py:
3690         (ChromiumPort):
3691         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
3692         (ChromiumMacPort):
3693         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
3694         (ChromiumMacPortTest.test_versions):
3695         * Scripts/webkitpy/layout_tests/port/factory_unittest.py:
3696         (FactoryTest.test_mac):
3697         (FactoryTest.test_chromium_mac):
3698         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
3699         (test_rebaseline_and_copy_test_with_lion_result):
3700
3701 2012-06-29  Ojan Vafai  <ojan@chromium.org>
3702
3703         Fix optimize-baselines to not move baselines from win to win-7sp0
3704         https://bugs.webkit.org/show_bug.cgi?id=90312
3705
3706         Reviewed by Dirk Pranke.
3707
3708         It used to consider win-7sp0 as the common directory for all the Apple
3709         windows ports and incorrectly move results out of win.
3710
3711         * Scripts/webkitpy/common/checkout/baselineoptimizer.py:
3712         * Scripts/webkitpy/common/checkout/baselineoptimizer_unittest.py:
3713         (BaselineOptimizerTest.test_win_does_not_drop_to_win_7sp0):
3714         (BaselineOptimizerTest.test_common_directory_includes_root):
3715
3716 2012-06-28  Ojan Vafai  <ojan@chromium.org>
3717
3718         Fix rebaselining for Qt and Apple ports
3719         https://bugs.webkit.org/show_bug.cgi?id=90204
3720
3721         Reviewed by Dirk Pranke.
3722
3723         -Apporpriately put wk2 results in the -wk2 directories.
3724         -Since Qt and Apple-Win don't have bots that correspond to the
3725         platform/qt and platform/win directories, we need to fudge it
3726         and always put the results in those directories for those ports.
3727
3728         * Scripts/webkitpy/layout_tests/port/builders.py:
3729         (rebaseline_override_dir):
3730         * Scripts/webkitpy/layout_tests/port/factory.py:
3731         (_builder_options):
3732         Identify webkit2 builders by the WK2 in the builder name.
3733         * Scripts/webkitpy/tool/commands/rebaseline.py:
3734         (RebaselineTest._baseline_directory):
3735         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
3736         (TestRebaseline.test_baseline_directory):
3737
3738 2012-06-29  Zoltan Horvath  <zoltan@webkit.org>
3739
3740         Add support for --force parameter to run-performance-tests
3741         https://bugs.webkit.org/show_bug.cgi?id=90279
3742
3743         Reviewed by Dirk Pranke.
3744
3745         It's helpful to be able to run tests from the Skipped list of the performance tests.
3746
3747         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
3748         (PerfTestsRunner._parse_args):
3749         (PerfTestsRunner._collect_tests):
3750         * Scripts/webkitpy/performance_tests/perftestsrunner_unittest.py: Add test.
3751         (test_collect_tests_with_skipped_list):
3752
3753 2012-06-29  Tony Chang  <tony@chromium.org>
3754
3755         [GTK] Enable CSS grid layout LayoutTests on GTK+
3756         https://bugs.webkit.org/show_bug.cgi?id=90226
3757
3758         Reviewed by Martin Robinson.
3759
3760         This feature is disabled via Settings by default, but for testing,
3761         we enable it using layoutTestController.overridePreferences. Add the
3762         necessary plumbing for DRT.
3763
3764         WTR already works because support was added for Apple Mac earlier.
3765
3766         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3767         (resetDefaultsToConsistentValues): Feature is off by default.
3768         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3769         (LayoutTestController::overridePreference): Add handling of WebKitCSSGridLayoutEnabled.
3770
3771 2012-06-29  Ádám Kallai  <kadam@inf.u-szeged.hu>
3772
3773         [Qt] Added Qt port for garden-o-matic.
3774         https://bugs.webkit.org/show_bug.cgi?id=82719
3775
3776         Reviewed by Adam Barth.
3777
3778         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/config.js:
3779         (.):
3780
3781 2012-06-29  Balazs Kelemen  <kbalazs@webkit.org>
3782
3783         [Qt][WTR] Get rid of using DumpRenderTreeSupportQt
3784         https://bugs.webkit.org/show_bug.cgi?id=90262
3785
3786         Reviewed by Alexey Proskuryakov.
3787
3788         Now that we decided to not support v8 in WebKit2
3789         we can get rid of using DumpRenderTreeSupportQt
3790         in WebKitTestRunner.
3791
3792         * Tools.pro:
3793         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3794         (WTR::InjectedBundlePage::resetAfterTest):
3795         (WTR::InjectedBundlePage::didClearWindowForFrame):
3796         * WebKitTestRunner/InjectedBundle/qt/ActivateFontsQt.cpp:
3797
3798 2012-06-29  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3799
3800         Revert r121540, it broke most Qt builds
3801
3802         * qmake/mkspecs/features/default_post.prf:
3803         * qmake/mkspecs/features/functions.prf:
3804
3805 2012-06-29  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
3806
3807         [Qt] Make build-webkit reject uknown configurations, eg. --profile
3808
3809         The qmake-based buildsystem doesn't support it.
3810
3811         Reviewed by Tor Arne Vestbø.
3812
3813         * Scripts/webkitdirs.pm:
3814         (buildQMakeProjects):
3815
3816 2012-06-29  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
3817
3818         [Qt] Don't add Qt module dependencies in features.prf
3819
3820         The required dependencies are already added in WebCore.pri.
3821
3822         Reviewed by Tor Arne Vestbø.
3823
3824         * qmake/mkspecs/features/features.prf:
3825
3826 2012-06-27  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
3827
3828         [Qt] Use LIBS_PRIVATE instead of putting dependencies into LIBS
3829
3830         Reviewed by Tor Arne Vestbø..
3831
3832         * qmake/mkspecs/features/default_post.prf:
3833         * qmake/mkspecs/features/functions.prf:
3834
3835 2012-06-28  Ryosuke Niwa  <rniwa@webkit.org>
3836
3837         DOMHTMLCollection::item may return a wrong element after namedItem is called
3838         https://bugs.webkit.org/show_bug.cgi?id=90240
3839
3840         Reviewed by Antti Koivisto.
3841
3842         Add a WebKit API test since namedItem is not used in the JS/V8 binding code.
3843
3844         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
3845         * TestWebKitAPI/Tests/mac/HTMLCollectionNamedItem.html: Copied from Tools/TestWebKitAPI/Tests/mac/HTMLFormCollectionNamedItem.html.
3846         * TestWebKitAPI/Tests/mac/HTMLCollectionNamedItem.mm: Copied from Tools/TestWebKitAPI/Tests/mac/HTMLFormCollectionNamedItem.mm.
3847         (TestWebKitAPI::TEST):
3848
3849 2012-06-28  Dirk Pranke  <dpranke@chromium.org>
3850
3851         nrwt: remove the 'google-chrome' port code
3852         https://bugs.webkit.org/show_bug.cgi?id=88824
3853
3854         Reviewed by Ojan Vafai.
3855
3856         NRWT now supports passing additional baseline directories
3857         and expectations files on the command line, so there's no need
3858         to support the concept of a 'google-chrome' port directly.
3859
3860         * Scripts/webkitpy/layout_tests/port/base.py:
3861         (Port.path_to_test_expectations_file):
3862         * Scripts/webkitpy/layout_tests/port/builders.py:
3863         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
3864         (ChromiumMacPort.__init__):
3865         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
3866         (ChromiumWinPort.__init__):
3867         * Scripts/webkitpy/layout_tests/port/factory.py:
3868         (PortFactory):
3869         * Scripts/webkitpy/layout_tests/port/factory_unittest.py:
3870         (FactoryTest.test_win):
3871         * Scripts/webkitpy/layout_tests/port/google_chrome.py: Removed.
3872         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py: Removed.
3873
3874 2012-06-28  Dirk Pranke  <dpranke@chromium.org>
3875
3876         nrwt: clean up how arguments are passed to workers
3877         https://bugs.webkit.org/show_bug.cgi?id=90126
3878
3879         Reviewed by Ojan Vafai.
3880
3881         The way arguments are passed to workers has been crufty. It
3882         turns out it can be a lot cleaner via two things:
3883         1) using a factory method instead of instantiating objects
3884         directly in manager_worker_broker removes the need for passing
3885         'worker arguments' to the broker.
3886         2) it turns out that since mock hosts and test ports are purely
3887         in-memory constructions, they can be pickled and passed to child
3888         workers, meaning that the worker no longer needs hacky code to
3889         pass the port in a special case or to guess what to do if we
3890         don't have a host - all of the test-specific logic can move to
3891         the test file, where we can stub out the mock host's
3892         port_factory to return the same already-created port when it
3893         needs to be shared.
3894
3895         This change also moves WorkerException to manager_worker_broker.py
3896         where it belongs, and removes several useless tests that were
3897         just a maintenance burden (and would've needed rewriting when we
3898         change the rest of the broker implementation).
3899
3900         * Scripts/webkitpy/layout_tests/controllers/manager.py:
3901         (Manager._run_tests.worker_factory):
3902         (Manager._run_tests):
3903         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
3904         (get):
3905         (WorkerException):
3906         (AbstractWorker.__init__):
3907         (_ManagerConnection.__init__):
3908         (_ManagerConnection.start_worker):
3909         (_InlineManager.__init__):
3910         (_InlineManager.start_worker):
3911         (_MultiProcessManager._can_pickle_host):
3912         (_MultiProcessManager):
3913         (_MultiProcessManager.start_worker):
3914         (_WorkerConnection.__init__):
3915         (_InlineWorkerConnection.__init__):
3916         (_InlineWorkerConnection.join):
3917         (_InlineWorkerConnection.run):
3918         (_Process.run):
3919         (_MultiProcessWorkerConnection.__init__):
3920         (_MultiProcessWorkerConnection.start):
3921         (_MultiProcessWorkerConnection):
3922         (_MultiProcessWorkerConnection.run):
3923         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py:
3924         (_TestWorker.__init__):
3925         (_TestWorker.run):
3926         (_TestsMixin.test_name):
3927         (_TestsMixin.test_cancel):
3928         (_TestsMixin.test_done):
3929         (_TestsMixin.test_unknown_message):
3930         (InlineBrokerTests.setUp):
3931         (MultiProcessBrokerTests.setUp):
3932         * Scripts/webkitpy/layout_tests/controllers/worker.py:
3933         (Worker.__init__):
3934         (Worker.run):
3935         * Scripts/webkitpy/layout_tests/controllers/worker_unittest.py: Removed.
3936         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
3937         (passing_run):
3938         (logging_run):
3939         (run_and_capture):
3940         (MainTest.test_child_processes_2):
3941         (MainTest.test_child_processes_min):
3942         (MainTest.test_exception_raised):
3943         (MainTest.test_keyboard_interrupt):
3944         (MainTest.test_retrying_and_flaky_tests):
3945         (MainTest.test_run_order__inline):
3946
3947 2012-06-28  Dirk Pranke  <dpranke@chromium.org>
3948
3949         nrwt: don't try to catch worker exceptions in run_webkit_tests.__main__
3950         https://bugs.webkit.org/show_bug.cgi?id=90125
3951
3952         Reviewed by Ojan Vafai.
3953
3954         It turns out run_webkit_tests.py wasn't really using
3955         WorkerException and the catch we had for it was pointless. I've
3956         removed the symbol import and moved it to the integration tests
3957         where it is needed. Eventually the definition of the exception
3958         moves to the broker module, and so minimizing the number of
3959         users of it is a good thing.
3960
3961         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3962         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
3963         (MainTest.test_exception_raised):
3964
3965 2012-06-28  Dirk Pranke  <dpranke@chromium.org>
3966
3967         nrwt: clean up passing of log messages between processes
3968         https://bugs.webkit.org/show_bug.cgi?id=90123
3969
3970         Reviewed by Ojan Vafai.
3971
3972         It turns out log messages are perfectly picklable by themselves
3973         and contain the process id of the process that generated the
3974         message, so if we just pass the record from the worker to the
3975         manager and call logger.handle() in the manager, everything just
3976         works :).
3977
3978         The change is covered by existing tests.
3979
3980         * Scripts/webkitpy/layout_tests/controllers/manager.py:
3981         (Manager._log_messages):
3982         * Scripts/webkitpy/layout_tests/controllers/worker.py:
3983         (_WorkerLogHandler.emit):
3984         * Scripts/webkitpy/layout_tests/views/metered_stream.py:
3985         (MeteredStream.__init__):
3986         (_LogHandler.emit):
3987
3988 2012-06-28  Dirk Pranke  <dpranke@chromium.org>
3989
3990         add a pylint wrapper for linting python code
3991         https://bugs.webkit.org/show_bug.cgi?id=90232
3992
3993         Reviewed by Adam Barth.
3994
3995         Currently we use 'pep8' to check python code in
3996         check-webkit-style. pep8 is fast but simple; pylint is slower
3997         but has much more robust linting capabilities and will catch
3998         variable typos and all sorts of other things. Eventually we
3999         should switch check-webkit-style to use this, but our code is
4000         far from linting now so it needs to be cleaned up first.
4001
4002         This change adds the infrastructure and a wrapper so we can
4003         start doing that.
4004
4005         * Scripts/lint-webkitpy: Added.
4006         * Scripts/webkitpy/pylintrc: Added.
4007         * Scripts/webkitpy/thirdparty/__init__.py:
4008         (AutoinstallImportHook.find_module):
4009         (AutoinstallImportHook._install_pylint):
4010
4011 2012-06-28  Tony Chang  <tony@chromium.org>
4012
4013         [GTK] Use WEBKITOUTPUTDIR to find fonts in DumpRenderTree
4014         https://bugs.webkit.org/show_bug.cgi?id=90215
4015
4016         Reviewed by Martin Robinson.
4017
4018         * DumpRenderTree/gtk/DumpRenderTree.cpp:
4019         (initializeFonts): Check for WEBKITOUTPUTDIR first.
4020         * Scripts/webkitpy/layout_tests/port/gtk.py:
4021         (GtkPort.setup_environ_for_server): Copy the environment variable to the child process.
4022         * WebKitTestRunner/InjectedBundle/gtk/ActivateFontsGtk.cpp:
4023         (WTR::inititializeFontConfigSetting): Check for WEBKITOUTPUTDIR first.
4024
4025 2012-06-28  Dirk Pranke  <dpranke@chromium.org>
4026
4027         derive ChromiumPort from WebKitPort in NRWT in order to support skipping tests if symbols are missing
4028         https://bugs.webkit.org/show_bug.cgi?id=89706
4029
4030         Reviewed by Ojan Vafai.
4031
4032         Try again to land the change first landed in r121363. This patch
4033         adds a bunch more tests and reworks the handling of
4034         port-specific default values for --pixel-tests and --time-out-ms
4035         to be more consistent (adding a default_pixel_tests() method,
4036         pushing the webkit default_timeout_ms() value up into base.py,
4037         and overriding it properly in the chromium and apple mac ports.
4038
4039         Also change the logic in
4040         run_webkit_tests._setup_derived_options() to not second-guess
4041         what the port wants the timeout value to be for debug builds;
4042         computing this in two different places led to several bugs.
4043
4044         This change also changes the Chromium unittest ports to derive
4045         from ChromiumPortTestCase instead of PortTestCase, so that we
4046         ensure that we're running the same tests on all port variants
4047         more easily. There's more cleanup that can be done here, but
4048         this is good enough for now
4049
4050         * Scripts/webkitpy/layout_tests/port/base.py:
4051         (Port.default_pixel_tests):
4052         (Port):
4053         (Port.default_timeout_ms):
4054         * Scripts/webkitpy/layout_tests/port/chromium.py:
4055         (ChromiumPort):
4056         (ChromiumPort.__init__):
4057         (ChromiumPort.default_pixel_tests):
4058         (ChromiumPort.default_timeout_ms):
4059         (ChromiumPort.driver_name):
4060         (ChromiumPort._driver_class):
4061         (ChromiumPort._missing_symbol_to_skipped_tests):
4062         (ChromiumPort.skipped_layout_tests):
4063         (ChromiumPort.setup_test_run):
4064         (ChromiumPort._path_to_image_diff):
4065         (ChromiumPort._convert_path):
4066         * Scripts/webkitpy/layout_tests/port/chromium_android_unittest.py:
4067         (ChromiumAndroidPortTest):
4068         (ChromiumAndroidPortTest.test_expectations_files):
4069         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
4070         (ChromiumLinuxPort._modules_to_search_for_symbols):
4071         * Scripts/webkitpy/layout_tests/port/chromium_linux_unittest.py:
4072         (ChromiumLinuxPortTest):
4073         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
4074         (ChromiumMacPort._modules_to_search_for_symbols):
4075         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
4076         (ChromiumMacPortTest):
4077         * Scripts/webkitpy/layout_tests/port/chromium_port_testcase.py:
4078         split off from chromium_unittest.
4079         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
4080         (ChromiumWinPort._modules_to_search_for_symbols):
4081         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
4082         (ChromiumWinTest):
4083         * Scripts/webkitpy/layout_tests/port/mac.py:
4084         (MacPort.default_timeout_ms):
4085         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
4086         (MacTest.test_default_timeout_ms):
4087         * Scripts/webkitpy/layout_tests/port/mock_drt.py:
4088         (MockDRTPort.start_http_server):
4089         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
4090         (PortTestCase.test_default_timeout_ms):
4091         (PortTestCase):
4092         (PortTestCase.test_default_pixel_tests):
4093         * Scripts/webkitpy/layout_tests/port/test.py:
4094         (TestPort.default_pixel_tests):
4095         * Scripts/webkitpy/layout_tests/port/webkit.py:
4096         (WebKitPort._modules_to_search_for_symbols):
4097         (WebKitPort):
4098         (WebKitPort._symbols_string):
4099         (WebKitPort._skipped_tests_for_unsupported_features):
4100         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
4101         (TestWebKitPort._symbols_string):
4102         (TestWebKitPort._tests_for_disabled_features):
4103         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
4104         (_set_up_derived_options):
4105
4106 2012-06-28  Joshua Bell  <jsbell@chromium.org>
4107
4108         run-bindings-tests should return non-zero exit code on test failure
4109         https://bugs.webkit.org/show_bug.cgi?id=90205
4110
4111         Reviewed by Adam Barth.
4112
4113         * Scripts/run-bindings-tests:
4114         (main):
4115
4116 2012-06-28  Terry Anderson  <tdanderson@chromium.org>
4117
4118         [chromium] Reset the device scale factor to 1 before each test is run
4119         https://bugs.webkit.org/show_bug.cgi?id=90212
4120
4121         Reviewed by Adam Barth.
4122
4123         Some tests change the device scale factor, so this needs to be reset to
4124         1.0 at the start of each test to avoid test flakiness.
4125
4126         * DumpRenderTree/chromium/LayoutTestController.cpp:
4127         (LayoutTestController::reset):
4128
4129 2012-06-28  Ryosuke Niwa  <rniwa@webkit.org>
4130
4131         Cleanup HTMLFormCollection
4132         https://bugs.webkit.org/show_bug.cgi?id=90111
4133
4134         Reviewed by Andreas Kling.
4135
4136         Add a WebKit API test using copy-paste design pattern per kling's request.
4137
4138         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
4139         * TestWebKitAPI/Tests/mac/HTMLFormCollectionNamedItem.html: Added.
4140         * TestWebKitAPI/Tests/mac/HTMLFormCollectionNamedItem.mm: Added.
4141         (-[HTMLFormCollectionNamedItemTest webView:didFinishLoadForFrame:]):
4142         (TestWebKitAPI):
4143         (TestWebKitAPI::TEST):
4144
4145 2012-06-28  Tim Horton  <timothy_horton@apple.com>
4146
4147         [mac] WKTR windows still don't stay off screen sometimes
4148         https://bugs.webkit.org/show_bug.cgi?id=90214
4149         <rdar://problem/11760263>
4150
4151         Reviewed by Simon Fraser.
4152
4153         In some cases, the system itself will consult [WebKitTestRunnerWindow frame], so we should refrain from
4154         overriding it and instead provide a different method to use when we want the web-facing "fake" window origin
4155         (for PlatformWebView::windowFrame()).
4156
4157         * WebKitTestRunner/mac/PlatformWebViewMac.mm:
4158         (-[WebKitTestRunnerWindow frameRespectingFakeOrigin]):
4159         (WTR::PlatformWebView::windowFrame):
4160
4161 2012-06-28  Martin Robinson  <mrobinson@igalia.com>
4162
4163         [GTK] Add unit tests for GtkInputMethodFilter
4164         https://bugs.webkit.org/show_bug.cgi?id=88698
4165
4166         Reviewed by Carlos Garcia Campos.
4167
4168         Add unit tests for GtkInputMethodFilter in the WebCore platform layer.
4169         This change adds the TestGtk test suite which will be used for all non-API
4170         layer GTK unit tests.
4171
4172         * TestWebKitAPI/GNUmakefile.am: Update the build to include the new tests.
4173         * TestWebKitAPI/Tests/gtk/InputMethodFilter.cpp: Added.
4174         (TestWebKitAPI::PlatformWebView::PlatformWebView): Remove the call to gtk_init here
4175         as it's now in main.cpp.
4176         * TestWebKitAPI/gtk/main.cpp: Change the g_type_init call to gtk_init, because now
4177         a majority of all unit tests depend on GTK+ being initialized.
4178
4179 2012-06-28  Christophe Dumez  <christophe.dumez@intel.com>
4180
4181         [EFL] Enable support for HTML5 datalist
4182         https://bugs.webkit.org/show_bug.cgi?id=90157
4183
4184         Reviewed by Martin Robinson.
4185
4186         Turn on DATALIST flag by default on EFL port to
4187         support HTML5 datalist tag.
4188
4189         * Scripts/webkitperl/FeatureList.pm:
4190
4191 2012-06-28  Tony Chang  <tony@chromium.org>
4192
4193         Enable CSS grid layout LayoutTests on platform Mac
4194         https://bugs.webkit.org/show_bug.cgi?id=90113
4195
4196         Reviewed by Ojan Vafai.
4197
4198         * DumpRenderTree/mac/DumpRenderTree.mm:
4199         (resetDefaultsToConsistentValues): Reset the value to NO between tests.
4200
4201 2012-06-28  Mike Fenton  <mifenton@rim.com>
4202
4203         [BlackBerry] Add watchlist options for Blackberry and editing.
4204         https://bugs.webkit.org/show_bug.cgi?id=90193
4205
4206         Unreviewed.
4207
4208         Add BlackBerry and Editing watchlist and monitor them.
4209
4210         * Scripts/webkitpy/common/config/watchlist:
4211
4212 2012-06-28  Tim Horton  <timothy_horton@apple.com>
4213
4214         Unreviewed, upgrade myself to a reviewer!
4215
4216         http://www.webkit.org/blog/2082/tim-horton-is-now-a-webkit-reviewer/
4217
4218         * Scripts/webkitpy/common/config/committers.py:
4219
4220 2012-06-28  Terry Anderson  <tdanderson@chromium.org>
4221
4222         [Chromium] Chromium's LayoutTestController is missing setBackingScaleFactor
4223         https://bugs.webkit.org/show_bug.cgi?id=83635
4224
4225         Reviewed by Adam Barth.
4226
4227         Added InvokeCallbackTask, a new derived class of MethodTask. When
4228         setBackingScaleFactor is called, a call to setDeviceScaleFactor in
4229         WebView is made and then postTask is used to invoke the callback
4230         function specified in testRunner.setBackingScaleFactor. 
4231
4232         * DumpRenderTree/chromium/LayoutTestController.cpp:
4233         (LayoutTestController::LayoutTestController):
4234         (InvokeCallbackTask):
4235         (InvokeCallbackTask::InvokeCallbackTask):
4236         (InvokeCallbackTask::runIfValid):
4237         (LayoutTestController::setBackingScaleFactor):
4238         * DumpRenderTree/chromium/LayoutTestController.h:
4239         (LayoutTestController):
4240
4241 2012-06-28  Ojan Vafai  <ojan@chromium.org>
4242
4243         Make rebaseline-test and rebaseline-expectations work for non-Chromium ports
4244         https://bugs.webkit.org/show_bug.cgi?id=90186
4245
4246         Reviewed by Adam Barth.
4247
4248         This makes rebaselining work for all ports that have a TestExpectations file
4249         in the tree. I didn't test other ports.
4250
4251         This doesn't address 100% of the problem. The rebaseline code puts the expectations
4252         in the most specific directory and relies on optimize-baselines to merge baselines
4253         appropriately. This only works if every platform directory has an equivalent bot
4254         that runs the tests, which is not true for most ports.
4255
4256         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
4257         (Builder._revision_and_build_for_filename):
4258         Some bots have filenames that aren't revision/build number pairs
4259         e.g. they are random junk like aQhxvx. Handle this gracefully.
4260         (Builder._fetch_revision_to_build_map):
4261         (Builder._file_info_list_to_revision_to_build_list):
4262         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
4263         (BuilderTest.test_build_and_revision_for_filename):
4264         (BuilderTest.test_file_info_list_to_revision_to_build_list):
4265         * Scripts/webkitpy/layout_tests/port/builders.py:
4266         Update the list of builders. This list needs to be kept up
4267         to date for the rebaseline tool to work.
4268         * Scripts/webkitpy/tool/commands/rebaseline.py:
4269         (RebaselineTest._results_url):
4270         (RebaselineExpectations._run_webkit_patch):
4271         (RebaselineExpectations._rebaseline_port):
4272         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
4273         Qt port uses qmake to determine the right version. Systems without qmake correctly fallback
4274         to a specific version.
4275
4276 2012-06-28  Csaba Osztrogonác  <ossy@webkit.org>
4277
4278         [Qt][NRWT] Fix baseline and skipped file search path.
4279         https://bugs.webkit.org/show_bug.cgi?id=89882
4280
4281         Unreviewed trivial typo fix after r121430.
4282
4283         * Scripts/webkitpy/layout_tests/port/qt_unittest.py:
4284         (QtPortTest._assert_search_path):
4285         (QtPortTest._assert_skipped_path):
4286
4287 2012-06-28  János Badics  <jbadics@inf.u-szeged.hu>
4288
4289         [Qt][NRWT] Fix baseline and skipped file search path.
4290         https://bugs.webkit.org/show_bug.cgi?id=89882
4291
4292         Reviewed by Csaba Osztrogonác.
4293
4294         * Scripts/webkitpy/layout_tests/port/qt.py:
4295         (QtPort._search_paths):
4296         (QtPort):
4297         (QtPort.baseline_search_path):
4298         (QtPort._skipped_file_search_paths):
4299         * Scripts/webkitpy/layout_tests/port/qt_unittest.py:
4300         (QtPortTest):
4301         (QtPortTest._assert_skipped_path):
4302         (QtPortTest.test_baseline_search_path):
4303         (QtPortTest.test_skipped_file_search_path):
4304
4305 2012-06-28  Alexander Pavlov  <apavlov@chromium.org>
4306
4307         [Qt] inspector/styles/inject-stylesheet.html makes 4 tests flakey (TEXT PASS)
4308         https://bugs.webkit.org/show_bug.cgi?id=90167
4309
4310         Reviewed by Csaba Osztrogonác.
4311
4312         DRT should remove user stylesheets from the page group when resetting before running another test.
4313
4314         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
4315         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
4316
4317 2012-06-28  Csaba Osztrogonác  <ossy@webkit.org>
4318
4319         [Qt][DRT] Reset AcceleratedCompositingEnabled between tests
4320         https://bugs.webkit.org/show_bug.cgi?id=90164
4321
4322         Reviewed by Simon Hausmann.
4323
4324         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
4325         (WebCore::WebPage::resetSettings):
4326
4327 2012-06-28  Csaba Osztrogonác  <ossy@webkit.org>
4328
4329         [Qt] Restore original value of mock scrollbars between tests
4330         https://bugs.webkit.org/show_bug.cgi?id=90155
4331
4332         Reviewed by Simon Hausmann.
4333
4334         * DumpRenderTree/qt/DumpRenderTreeQt.cpp: