[Chromium] Introduce TestInterfaces to hold all the JavaScript interfaces needed...
[WebKit-https.git] / Tools / ChangeLog
1 2012-07-16  Adam Barth  <abarth@webkit.org>
2
3         [Chromium] Introduce TestInterfaces to hold all the JavaScript interfaces needed for LayoutTests
4         https://bugs.webkit.org/show_bug.cgi?id=91312
5
6         Reviewed by Ryosuke Niwa.
7
8         Looking forward to moving more objects into TestRunner.a, we're going
9         to need an object to own all the interfaces and to put them through
10         their lifecycle.
11
12         * DumpRenderTree/DumpRenderTree.gypi:
13         * DumpRenderTree/chromium/TestShell.cpp:
14         (TestShell::initialize):
15         (TestShell::resetTestController):
16         (TestShell::bindJSObjectsToWindow):
17         * DumpRenderTree/chromium/TestShell.h:
18         (TestShell):
19         * DumpRenderTree/chromium/TestRunner/TestInterfaces.cpp: Added.
20         * DumpRenderTree/chromium/TestRunner/TestInterfaces.h: Added.
21
22 2012-07-16  Ryuan Choi  <ryuan.choi@samsung.com>
23
24         [EFL][WK2] Add APIs to support theme.
25         https://bugs.webkit.org/show_bug.cgi?id=90107
26
27         Reviewed by Hajime Morita.
28
29         Override default theme path for MiniBrowser and WebKitTestRunner.
30
31         * MiniBrowser/efl/CMakeLists.txt:
32         * MiniBrowser/efl/main.c:
33         (browserCreate):
34         * WebKitTestRunner/PlatformEfl.cmake:
35         * WebKitTestRunner/efl/PlatformWebViewEfl.cpp:
36         (WTR::PlatformWebView::PlatformWebView):
37
38 2012-07-16  Ryuan Choi  <ryuan.choi@samsung.com>
39
40         [EFL][DRT] Implement dumpFrameScrollPosition
41         https://bugs.webkit.org/show_bug.cgi?id=87638
42
43         Reviewed by Hajime Morita.
44
45         * DumpRenderTree/efl/DumpRenderTree.cpp:
46         (dumpFrameScrollPosition): Implemented.
47         (shouldDumpFrameScrollPosition): Fixed wrong condition like other ports.
48
49 2012-07-16  Don Olmstead  <don.olmstead@am.sony.com>
50
51         NRWRT Should provide a VS project to work on
52         https://bugs.webkit.org/show_bug.cgi?id=91436
53
54         Reviewed by Dirk Pranke.
55
56         Adding VS2010 project for modifying webkitpy. Uses pytools
57         <http://pytools.codeplex.com> which integrates python into 
58         Visual Studio.
59
60         * Scripts/webkitpy/webkitpy.pyproj: Added.
61         * Scripts/webkitpy/webkitpy.sln: Added.
62
63 2012-07-16  Adam Barth  <abarth@webkit.org>
64
65         [Chromium] Move GamepadController into TestRunner.a
66         https://bugs.webkit.org/show_bug.cgi?id=91311
67
68         Reviewed by Tony Chang.
69
70         GamepadController looks like a simple class to move into TestRunner.a
71         because it has almost zero dependence on TestShell.
72
73         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
74         * DumpRenderTree/DumpRenderTree.gypi:
75         * DumpRenderTree/chromium/TestRunner/CppBoundClass.cpp: Renamed from Tools/DumpRenderTree/chromium/CppBoundClass.cpp.
76         * DumpRenderTree/chromium/TestRunner/CppBoundClass.h: Renamed from Tools/DumpRenderTree/chromium/CppBoundClass.h.
77         * DumpRenderTree/chromium/TestRunner/CppVariant.cpp: Renamed from Tools/DumpRenderTree/chromium/CppVariant.cpp.
78         * DumpRenderTree/chromium/TestRunner/CppVariant.h: Renamed from Tools/DumpRenderTree/chromium/CppVariant.h.
79             - GamepadController depends on these bindings helpers.
80         * DumpRenderTree/chromium/TestRunner/GamepadController.cpp: Renamed from Tools/DumpRenderTree/chromium/GamepadController.cpp.
81         * DumpRenderTree/chromium/TestRunner/GamepadController.h: Renamed from Tools/DumpRenderTree/chromium/GamepadController.h.
82             - Just moved these files and removed the unused TestShell
83               references.
84         * DumpRenderTree/chromium/TestRunner/Stub.cpp: Removed.
85             - No longer needed.
86         * DumpRenderTree/chromium/TestShell.cpp:
87             - Update call to constructor to avoid passing in this.
88
89 2012-07-16  Xianzhu Wang  <wangxianzhu@chromium.org>
90
91         [Chromium-android] Don't use test_shell mode of DRT
92         https://bugs.webkit.org/show_bug.cgi?id=88542
93
94         Reviewed by Dirk Pranke.
95
96         Test shell mode is about to be removed.
97         Switch to use DRT mode for chromium-android.
98
99         Summary of changes:
100         1. ChromiumAndroidDriver now inherits from WebKitDriver instead of ChromiumDriver (to be deprecated).
101         2. Conforms to the DRT mode protocol for input/output of DumpRenderTree.
102         3. Added support for Android 'adb shell' input/output (base64, newline mode, etc.)
103
104         * DumpRenderTree/chromium/TestEventPrinter.cpp:
105         (DRTPrinter::handleImage): Outputs base64 on Android.
106         * DumpRenderTree/chromium/TestShellAndroid.cpp:
107         (platformInit): Changed err_file to err_fifo, required by python ServerProcess.
108         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
109         (ChromiumAndroidPort.__init__):
110         (ChromiumAndroidPort.create_driver): Override to create driver without DriverProxy to ensure 1 Driver per run.
111         (ChromiumAndroidDriver):
112         (ChromiumAndroidDriver.__init__):
113         (ChromiumAndroidDriver.cmd_line):
114         (ChromiumAndroidDriver._deadlock_detector):
115         (ChromiumAndroidDriver._drt_cmd_line):
116         (ChromiumAndroidDriver.start):
117         (ChromiumAndroidDriver._start):
118         (ChromiumAndroidDriver._start_once):
119         (ChromiumAndroidDriver.run_test):
120         (ChromiumAndroidDriver.stop):
121         (ChromiumAndroidDriver._command_from_driver_input):
122         (ChromiumAndroidDriver._read_prompt):
123         * Scripts/webkitpy/layout_tests/port/chromium_android_unittest.py:
124         (ChromiumAndroidPortTest.test_driver_cmd_line):
125         (ChromiumAndroidDriverTest.test_cmd_line):
126         (ChromiumAndroidDriverTest):
127         (ChromiumAndroidDriverTest.test_drt_cmd_line):
128         (ChromiumAndroidDriverTest.test_read_prompt):
129         (ChromiumAndroidDriverTest.test_command_from_driver_input):
130         (ChromiumAndroidDriverTest.test_write_command_and_read_line):
131         * Scripts/webkitpy/layout_tests/port/server_process.py:
132         (ServerProcess.__init__): Added universal_newlines to handle Android 'adb shell' line ends.
133         (ServerProcess._start):
134         (ServerProcess._wait_for_data_and_update_buffers_using_select): Handles unexpected EOF which indicates crash on Android.
135         (ServerProcess.stop): Added kill_directly parameter to kill the process without waiting it (which always timeouts for Android).
136         (ServerProcess.replace_outputs): Added to combine different input/output pipes into one ServerProcess.
137         * Scripts/webkitpy/layout_tests/port/webkit.py:
138         (WebKitDriver._command_from_driver_input):
139         (WebKitDriver.run_test): Changed timeout origin so that slow start() on Android won't cause timeout of layout test case.
140         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
141         (MockServerProcess.read_stdout):
142         (MockServerProcess.start):
143         (MockServerProcess):
144         (MockServerProcess.stop):
145         (MockServerProcess.kill):
146         (WebKitDriverTest.test_read_block):
147         (WebKitDriverTest.test_read_binary_block):
148         (WebKitDriverTest.test_read_base64_block):
149
150 2012-07-16  Dirk Pranke  <dpranke@chromium.org>
151
152         nrwt: move a bunch of printing code from manager.py to printing.py
153         https://bugs.webkit.org/show_bug.cgi?id=91439
154
155         Reviewed by Ojan Vafai.
156
157         All of the logic that handles what gets logged should eventually
158         live in printing.py; this patch moves a large chunk of code that
159         prints all of the stuff after the run completes.
160
161         There are no functional changes, this is just moving code
162         around. Covered by existing tests (updated as necessary).
163
164         This code is all pretty messy and most of it should be deleted
165         (or extracted after the fact from results.json), but I'm saving
166         that for further patches. My goal is just to manager.py to a
167         more manageable state.
168
169         * Scripts/webkitpy/layout_tests/controllers/manager.py:
170         (Manager.prepare_lists_and_print_output):
171         (Manager.run):
172         (Manager._mark_interrupted_tests_as_skipped):
173         (Manager._update_summary_with_result):
174         (Manager._num_digits):
175         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
176         (ManagerTest.test_interrupt_if_at_failure_limits):
177         (ResultSummaryTest.summarized_results):
178         * Scripts/webkitpy/layout_tests/models/result_summary.py:
179         (ResultSummary.__init__):
180         (ResultSummary.add):
181         Here we had to add a list of the slow tests to the result
182         summary so that we didn't need to call back into the manager and
183         the test expectations object to figure out if a test is SLOW.
184         * Scripts/webkitpy/layout_tests/views/printing.py:
185         (Printer.print_results):
186         (Printer):
187         (Printer._print_timing_statistics):
188         (Printer._print_aggregate_test_statistics):
189         (Printer._print_individual_test_times):
190         (Printer._print_test_list_timing):
191         (Printer._print_directory_timings):
192         (Printer._print_statistics_for_test_timings):
193         (Printer._print_result_summary):
194         (Printer._print_result_summary_entry):
195         * Scripts/webkitpy/layout_tests/views/printing_unittest.py:
196         (Testprinter.test_print_unexpected_results.get_unexpected_results):
197         (Testprinter):
198
199 2012-07-16  Adam Barth  <abarth@webkit.org>
200
201         [Chromium] Create a stub TestRunner.a target
202         https://bugs.webkit.org/show_bug.cgi?id=91309
203
204         Reviewed by Ryosuke Niwa.
205
206         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
207         * DumpRenderTree/DumpRenderTree.gypi:
208         * DumpRenderTree/chromium/TestRunner/Stub.cpp: Added.
209
210 2012-07-16  Dirk Pranke  <dpranke@chromium.org>
211
212         test-webkitpy: clean up logging to make it ninja-esque
213         https://bugs.webkit.org/show_bug.cgi?id=91297
214
215         Reviewed by Adam Barth.
216
217         This patch changes the output of test-webkitpy to be closer to
218         what ninja produces. Namely:
219
220         If you are running with stderr writing to a tty (and not -v), we
221         will no longer print '...' as tests complete. Instead, we use a
222         metered stream and print entries of the form:
223            '[X/Y] test_name passed/failed/erred'
224         where X is the current test # and Y is the total number of tests.
225
226         If you are running with stderr piped to a file or other non-tty
227         object (or with -v), you get the same output, one per line for
228         every test.
229
230         In addition, if tests fail or err out, you get the stack trace
231         immediately; you don't have to wait until the end of the run.
232
233         Lastly, this change cleans up the unit tests for test-webkitpy
234         itself to not confuse the logger and to work cleanly w/ multiple
235         processes. It looks like between this and all the clean up in
236         MessagePool that happened as it landed, we can now run
237         multiprocessing tests in parallel.
238
239         * Scripts/webkitpy/test/finder.py:
240         (Finder._default_names):
241         * Scripts/webkitpy/test/finder_unittest.py:
242         (FinderTest.setUp):
243         (FinderTest.tearDown):
244         * Scripts/webkitpy/test/printer.py:
245         (Printer.__init__):
246         (Printer.configure):
247         (Printer.print_started_test):
248         (Printer.print_finished_test):
249         (Printer._test_line):
250         (Printer.print_result):
251         * Scripts/webkitpy/test/runner.py:
252         (Runner.run):
253         * Scripts/webkitpy/test/runner_unittest.py:
254         (RunnerTest.setUp):
255         (RunnerTest):
256         (RunnerTest.tearDown):
257         (RunnerTest.assert_run):
258         (RunnerTest.test_regular):
259         (RunnerTest.test_verbose):
260         (RunnerTest.test_timing):
261
262 2012-07-16  Dirk Pranke  <dpranke@chromium.org>
263
264         test-webkitpy: handle failures properly when running in parallel
265         https://bugs.webkit.org/show_bug.cgi?id=91416
266
267         Reviewed by Adam Barth.
268
269         It turns out that unittest.TestResults contain a handle to the
270         test method itself, which isn't picklable; it's sufficient to just
271         store the test name instead of the actual method. By doing so
272         we can move the test_name() method from the printer to the
273         runner where it belongs (so the printer is less dependent on the
274         unittest framework's data structures).
275
276         This change should really have a test but I don't know how to
277         write one that properly captures the behavior and won't cause
278         test-webkitpy itself to fail. I've verified the fix by hand, at
279         least, in the meantime.
280
281         * Scripts/webkitpy/test/printer.py:
282         (Printer.__init__):
283         (Printer.print_result):
284         * Scripts/webkitpy/test/runner.py:
285         (_test_name):
286         (Runner.all_test_names):
287         (_Worker.handle):
288
289 2012-07-16  Dirk Pranke  <dpranke@chromium.org>
290
291         test-webkitpy: handle failures properly when running in parallel
292         https://bugs.webkit.org/show_bug.cgi?id=91416
293
294         Reviewed by Tony Chang.
295
296         It turns out that unittest.TestResults contain a handle to the
297         test method itself, which isn't picklable; it's sufficient to just
298         store the test name instead of the actual method. By doing so
299         we can move the test_name() method from the printer to the
300         runner where it belongs (so the printer is less dependent on the
301         unittest framework's data structures).
302
303         This change should really have a test but I don't know how to
304         write one that properly captures the behavior and won't cause
305         test-webkitpy itself to fail. I've verified the fix by hand, at
306         least, in the meantime.
307
308         * Scripts/webkitpy/test/printer.py:
309         (Printer.__init__):
310         (Printer.print_result):
311         * Scripts/webkitpy/test/runner.py:
312         (_test_name):
313         (Runner.all_test_names):
314         (_Worker.handle):
315
316 2012-07-16  Dirk Pranke  <dpranke@chromium.org>
317
318         test-webkitpy: run tests in parallel
319         https://bugs.webkit.org/show_bug.cgi?id=91294
320
321         Reviewed by Ojan Vafai.
322
323         This change adds support for running tests in parallel. This is
324         not yet on by default, since the logging isn't very pretty w/
325         parallel tests.
326
327         Also, there are some (multiprocessing-related) tests that can't be
328         run in parallel and so we skip them in that situation; I need to
329         come up with a mechanism for dealing with this, since you
330         apparently can't use multiprocessing as both a parent and a
331         child process.
332
333         * Scripts/webkitpy/test/finder.py:
334         (Finder.find_names):
335         (Finder._default_names):
336         * Scripts/webkitpy/test/main.py:
337         (Tester._parse_args):
338         (Tester.run):
339         * Scripts/webkitpy/test/main_unittest.py:
340         (TesterTest.test_no_tests_found):
341         * Scripts/webkitpy/test/runner.py:
342         (Runner.run):
343         * Scripts/webkitpy/test/runner_unittest.py:
344         (RunnerTest.test_regular):
345         (RunnerTest.test_verbose):
346         (RunnerTest.test_timing):
347
348 2012-07-16  Dirk Pranke  <dpranke@chromium.org>
349
350         test-webkitpy: use message pools
351         https://bugs.webkit.org/show_bug.cgi?id=91292
352
353         Reviewed by Ojan Vafai.
354
355         Restructure the test-running code to be message-driven and
356         use a MessagePool; note that this does not yet actually run the
357         tests in parallel.
358
359         Also clean up the unit tests so that the fake loader is passed
360         to the _Worker properly, and reduce a lot of the cut&pasted code
361         in the tests.
362
363         No functional changes; covered by existing tests.
364
365         * Scripts/webkitpy/test/printer.py:
366         (Printer.print_started_test):
367         (Printer.print_finished_test):
368         * Scripts/webkitpy/test/runner.py:
369         (Runner.__init__):
370         (Runner.run):
371         (Runner.handle):
372         (_Worker):
373         (_Worker.__init__):
374         (_Worker.handle):
375         * Scripts/webkitpy/test/runner_unittest.py:
376         (RunnerTest.assert_run):
377         (RunnerTest.test_regular):
378         (RunnerTest.test_verbose):
379         (RunnerTest.test_timing):
380
381 2012-07-16  Simon Hausmann  <simon.hausmann@nokia.com>
382
383         [Qt] Fix Qt5 module dependencies
384         https://bugs.webkit.org/show_bug.cgi?id=91388
385
386         Reviewed by Laszlo Gombos.
387
388         Removed ancient qtscript and qtxmlpatterns module dependencies.
389
390         * qmake/qt_webkit.pri:
391
392 2012-07-16  Thiago Marcos P. Santos  <thiago.santos@intel.com>
393
394         run-webkit-test returns 254 at exit even when it succeeds
395         https://bugs.webkit.org/show_bug.cgi?id=91370
396
397         Reviewed by Simon Hausmann.
398
399         sys.exit() is implemented by raising SystemExit which inherits from
400         BaseException. Since we are catching BaseException, run_webkit_tests.py
401         was always returning an error code.
402
403         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
404
405 2012-07-15  Carlos Garcia Campos  <cgarcia@igalia.com>
406
407         Unreviewed. Fix make distcheck.
408
409         * GNUmakefile.am: Add missing file to compilation.
410
411 2012-07-15  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
412
413         Unreviewed. Add EFLWebKit2PublicAPI, EFLWebKit2PlatformSpecific to watch list.
414
415         * Scripts/webkitpy/common/config/watchlist:
416
417 2012-07-15  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
418
419         Unreviewed. Add gyuyoung.kim@samsung.com to watch list.
420
421         * Scripts/webkitpy/common/config/watchlist:
422
423 2012-07-15  Joseph Pecoraro  <pecoraro@apple.com>
424
425         Windowless WebView not firing JavaScript load event if there is a media element
426         https://bugs.webkit.org/show_bug.cgi?id=91331
427
428         Reviewed by Eric Carlson.
429
430         Test a windowless WebView loading a page with a media element. The load
431         event should happen on the next spin of the run loop, but we spin check
432         the page, with a timeout of 250ms.
433
434         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
435         * TestWebKitAPI/Tests/mac/WindowlessWebViewWithMedia.html: Added.
436         * TestWebKitAPI/Tests/mac/WindowlessWebViewWithMedia.mm: Added.
437         (-[WindowlessWebViewWithMediaFrameLoadDelegate webView:didFinishLoadForFrame:]):
438         (TestWebKitAPI::spinLoop): Spin check with timeout. Runs a block each spin to early bail.
439         (TestWebKitAPI::TEST):
440
441 2012-07-14  Benjamin Poulain  <bpoulain@apple.com>
442
443         [Mac] Do not try to update the cache model for every WebPreferences change
444         https://bugs.webkit.org/show_bug.cgi?id=91302
445
446         Reviewed by Joseph Pecoraro.
447
448         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
449         * TestWebKitAPI/Tests/mac/SetAndUpdateCacheModel.mm: Added.
450         (TestWebKitAPI):
451         (TestWebKitAPI::TEST):
452
453 2012-07-14  Sheriff Bot  <webkit.review.bot@gmail.com>
454
455         Unreviewed, rolling out r122665.
456         http://trac.webkit.org/changeset/122665
457         https://bugs.webkit.org/show_bug.cgi?id=91321
458
459         Broke Mac builds (Requested by rniwa on #webkit).
460
461         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
462         * TestWebKitAPI/Tests/mac/SetAndUpdateCacheModel.mm: Removed.
463
464 2012-07-13  Benjamin Poulain  <bpoulain@apple.com>
465
466         [Mac] Do not try to update the cache model for every WebPreferences change
467         https://bugs.webkit.org/show_bug.cgi?id=91302
468
469         Reviewed by Joseph Pecoraro.
470
471         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
472         * TestWebKitAPI/Tests/mac/SetAndUpdateCacheModel.mm: Added.
473         (TestWebKitAPI):
474         (TestWebKitAPI::TEST):
475
476 2012-07-13  Dirk Pranke  <dpranke@chromium.org>
477
478         run-webkit-test outputs the wrong number of tests executed when some are skipped.
479         https://bugs.webkit.org/show_bug.cgi?id=89894
480
481         Reviewed by Ojan Vafai.
482
483         Fix the logging of the actual number of tests run so that tests
484         that are skipped aren't included.
485
486         Also revamp the 'expected' output so we distinguish the number
487         of tests found from the number of tests run (to account for
488         --repeat-each and --iterations).
489
490         Covered by existing tests.
491
492         * Scripts/webkitpy/layout_tests/controllers/manager.py:
493         (Manager.prepare_lists_and_print_output):
494         (Manager._log_num_workers):
495         (Manager.run):
496         (Manager._print_result_summary):
497         * Scripts/webkitpy/layout_tests/models/result_summary.py:
498         (ResultSummary.__init__):
499         (ResultSummary.add):
500         * Scripts/webkitpy/layout_tests/views/printing.py:
501         (Printer.print_one_line_summary):
502         * Scripts/webkitpy/layout_tests/views/printing_unittest.py:
503         (Testprinter.test_print_one_line_summary):
504
505 2012-07-13  Dirk Pranke  <dpranke@chromium.org>
506
507         nrwt: actually print the exception name and message for otherwise unhandled exceptions
508         https://bugs.webkit.org/show_bug.cgi?id=91305
509
510         Reviewed by Adam Barth.
511
512         Two more places where I was printing the stack trace but not the
513         exception itself :(. These two spots can't easily be
514         unit-tested, but I tested them by hand.
515
516         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
517         (run):
518         (main):
519
520 2012-07-13  Josh Hawn  <jhawn@apple.com>
521
522         Fix for WebContext::getWebCoreStatistics() causes crash if no m_process
523         https://bugs.webkit.org/show_bug.cgi?id=91116
524
525         Reviewed by Simon Fraser.
526
527         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
528             Added new test file.
529         * TestWebKitAPI/Tests/WebKit2/WebCoreStatisticsWithNoWebProcess.cpp: Added.
530         (TestWebKitAPI::wkContextGetStatisticsCallback):
531             Tests that callback function receives an error.
532         (TestWebKitAPI::TEST):
533             Creates a dummy web context object (no web process).
534             Calls WKContextGetStatistics with the web context and test callback.
535             The test callback should get an expected error.
536
537 2012-07-13  Dirk Pranke  <dpranke@chromium.org>
538
539         test-webkitpy: move printing-related code out of the runner
540         https://bugs.webkit.org/show_bug.cgi?id=91289
541
542         Reviewed by Ryosuke Niwa.
543
544         More refactoring ... this moves all printing-related stuff out
545         of runner.py and into printer.py.
546
547         No functional changes; covered by existing tests.
548
549         * Scripts/webkitpy/test/main.py:
550         (Tester._run_tests):
551         * Scripts/webkitpy/test/printer.py:
552         (Printer.__init__):
553         (Printer):
554         (Printer.test_name):
555         (Printer.print_started_test):
556         (Printer.print_finished_test):
557         (Printer.print_result):
558         * Scripts/webkitpy/test/runner.py:
559         (Runner.__init__):
560         (Runner.all_test_names):
561         (Runner.run):
562         * Scripts/webkitpy/test/runner_unittest.py:
563         (RunnerTest.test_regular):
564         (RunnerTest.test_verbose):
565         (RunnerTest.test_timing):
566
567 2012-07-13  Dirk Pranke  <dpranke@chromium.org>
568
569         webkitpy: split printing/logging code for test-webkitpy out into a new class
570         https://bugs.webkit.org/show_bug.cgi?id=91282
571
572         Reviewed by Ojan Vafai.
573
574         This patch is the first step at splitting all of the
575         printing/logging code out separately from the actual
576         test-running code.
577
578         This is just moving stuff around; no new functionality and no
579         new tests needed.
580
581         * Scripts/webkitpy/test/finder_unittest.py:
582         (FinderTest.setUp):
583         * Scripts/webkitpy/test/main.py:
584         (Tester.__init__):
585         (Tester._parse_args):
586         (Tester.run):
587         (Tester._run_tests):
588         (Tester._log_exception):
589         * Scripts/webkitpy/test/main_unittest.py:
590         (TesterTest.test_no_tests_found):
591         * Scripts/webkitpy/test/printer.py: Added.
592         (Printer):
593         (Printer.__init__):
594         (Printer.configure):
595         (Printer.configure.filter):
596         (_CaptureAndPassThroughStream):
597         (_CaptureAndPassThroughStream.__init__):
598         (_CaptureAndPassThroughStream.write):
599         (_CaptureAndPassThroughStream._message_is_from_pdb):
600         (_CaptureAndPassThroughStream.flush):
601         (_CaptureAndPassThroughStream.getvalue):
602
603 2012-07-13  James Simonsen  <simonjam@chromium.org>
604
605         [Navigation Timing] Imported W3C tests contain duplicates and are DOS formatted
606         https://bugs.webkit.org/show_bug.cgi?id=91184
607
608         Reviewed by Adam Barth.
609
610         The upstream 'html5' tests are just duplicates of the 'html' tests.
611
612         * Scripts/import-w3c-performance-wg-tests:
613
614 2012-07-13  Dirk Pranke  <dpranke@chromium.org>
615
616         webkitpy: hide yield_to_caller from callers in MessagePool :)
617         https://bugs.webkit.org/show_bug.cgi?id=91269
618
619         Reviewed by Adam Barth.
620
621         yield_to_caller() was an optimization/hack to allow us to run
622         both manager and worker in a single process/loop without
623         starving the manager while the worker is running tests. The
624         worker was required to call yield_to_caller() periodically. It
625         turns out that I can get equivalent responsiveness by yielding
626         inside the MessagePool every time the worker posts a message, and this
627         allows me to no longer need the worker to call the routine. Thus
628         I rename yield_to_caller() to _yield_to_manager() to be a little
629         clearer about its purpose.
630
631         Tested by existing tests.
632
633         * Scripts/webkitpy/common/message_pool.py:
634         (_Worker.run):
635         (_Worker.post):
636         (_Worker._yield_to_manager):
637         * Scripts/webkitpy/layout_tests/controllers/worker.py:
638         (Worker.handle):
639
640 2012-07-13  Adam Barth  <abarth@webkit.org>
641
642         EWSTools should be able to build a commit-queue instance from scratch
643         https://bugs.webkit.org/show_bug.cgi?id=91264
644
645         Reviewed by Eric Seidel.
646
647         I've been using this script to build commit-queue instances on Google
648         Compute Engine and it seems to work.
649
650         * EWSTools/GoogleComputeEngine/build-commit-queue.sh: Added.
651
652 2012-07-13  Dirk Pranke  <dpranke@chromium.org>
653
654         webkitpy: make worker.start() and worker.stop() optional in the messagepool
655         https://bugs.webkit.org/show_bug.cgi?id=91170
656
657         Reviewed by Ojan Vafai.
658
659         test-webkitpy will use messagepool workers that don't actually
660         have any per-worker state, so they don't need start() and stop()
661         methods. Now we will only call the methods if they exist; this
662         means that workers only need to expose a handle() method.
663
664         * Scripts/webkitpy/common/message_pool.py:
665         (_Worker.terminate):
666         (_Worker.run):
667
668 2012-07-13  Dirk Pranke  <dpranke@chromium.org>
669
670         NRWT doesn't print exceptions
671         https://bugs.webkit.org/show_bug.cgi?id=91129
672
673         Reviewed by Ojan Vafai.
674
675         Although we printed exceptions in most cases, if an unexpected
676         exception (like a runtime error) was raised when creating a
677         port, we wouldn't. This patch fixes that, and also cleans up
678         how we were logging exceptions from the workers to be less
679         verbose.
680
681         Because of the corner cases where these errors are occurring,
682         it's difficult to write automated unit tests for them. I've
683         tested it quite a bit by hand, though.
684
685         * Scripts/webkitpy/common/message_pool.py:
686         (_MessagePool._close):
687         (_MessagePool._handle_worker_exception):
688         (_Worker.run):
689         (_Worker._raise):
690         * Scripts/webkitpy/layout_tests/controllers/manager.py:
691         (Manager._run_tests):
692         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
693         (main):
694
695 2012-07-13  Wei James  <james.wei@intel.com>
696
697         enable TestWebKitAPI/webkit_unit_tests apk on x86 android platform by adding abi support
698         https://bugs.webkit.org/show_bug.cgi?id=91194
699
700         Reviewed by Adam Barth.
701
702         * TestWebKitAPI/TestWebKitAPI.gyp/TestWebKitAPI.gyp:
703
704 2012-07-13  Simon Pena  <spena@igalia.com>
705
706         [GTK] Gardening: update API tests skipped list
707         https://bugs.webkit.org/show_bug.cgi?id=91224
708
709         Unreviewed gardening.
710
711         Skip "next" and "previous" tests of FindController until bug #91083
712         is fixed.
713
714         * gtk/run-api-tests:
715         (TestRunner):
716
717 2012-07-13  Zeno Albisser  <zeno@webkit.org>
718
719         [Qt][WK2] Implement GraphicsSurface for Linux/GLX.
720         https://bugs.webkit.org/show_bug.cgi?id=90881
721
722         Enable GraphicsSurface for Linux based platforms
723         whenever the Xcomposite extension is available.
724
725         Reviewed by Noam Rosenthal.
726
727         * qmake/config.tests/libXcomposite/libXcomposite.cpp: Added.
728         (main):
729         * qmake/config.tests/libXcomposite/libXcomposite.pro: Added.
730           Add a configure test to detect Xcomposite extension and
731           activate GraphicsSurface on linux in case the extension is available.
732         * qmake/configure.pri:
733         * qmake/mkspecs/features/features.prf:
734
735 2012-07-13  David Grogan  <dgrogan@chromium.org>
736
737         nrwt: don't choke when printing invalid utf-8 to stderr
738         https://bugs.webkit.org/show_bug.cgi?id=91181
739
740         Reviewed by Dirk Pranke.
741
742         * Scripts/webkitpy/layout_tests/controllers/test_result_writer.py:
743         (TestResultWriter.write_stderr):
744
745 2012-07-13  Dirk Pranke  <dpranke@chromium.org>
746
747         test-webkitpy: more class renaming cleanup
748         https://bugs.webkit.org/show_bug.cgi?id=91182
749
750         Reviewed by Adam Barth.
751
752         More removing of the unnecessary "Test" prefix.
753
754         * Scripts/webkitpy/test/finder.py:
755         (_DirectoryTree):
756         (Finder.add_tree):
757         * Scripts/webkitpy/test/main.py:
758         (Tester._run_tests):
759         * Scripts/webkitpy/test/runner.py:
760         (Runner):
761         * Scripts/webkitpy/test/runner_unittest.py:
762         (RunnerTest.test_regular):
763         (RunnerTest.test_verbose):
764         (RunnerTest.test_timing):
765
766 2012-07-12  Christophe Dumez  <christophe.dumez@intel.com>
767
768         [WK2][EFL] Facilitate debugging of the Web Process
769         https://bugs.webkit.org/show_bug.cgi?id=90768
770
771         Reviewed by Kenneth Rohde Christiansen.
772
773         Add a new --webprocess-cmd-prefix argument to
774         run-webkit-tests script for EFL port. If provided,
775         the prefix will be prepended to the command used
776         to spawn the Web process. This can be used for
777         debugging purposes with prefixes such as:
778         "xterm -title renderer -e gdb --args".
779
780         * Scripts/webkitpy/layout_tests/port/efl.py:
781         (EflPort.__init__):
782         (EflPort.setup_environ_for_server):
783         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
784         (parse_args):
785         * WebKitTestRunner/efl/TestControllerEfl.cpp:
786         (WTR::TestController::platformRunUntil): Implement support for
787         m_noTimeout timeout value.
788
789 2012-07-12  Adam Barth  <abarth@webkit.org>
790
791         Fix crash in the commit-queue.  We need to initialize self.port during __init__.
792
793         * Scripts/webkitpy/tool/commands/queues.py:
794         (CommitQueue.__init__):
795         (CommitQueue.begin_work_queue):
796
797 2012-07-12  Dirk Pranke  <dpranke@chromium.org>
798
799         test-webkitpy: rename test_finder to finder
800         https://bugs.webkit.org/show_bug.cgi?id=91175
801
802         Reviewed by Adam Barth.
803
804         Rename test_finder -> finder, TestFinder -> Finder to remove
805         some of the stutter in the names.
806
807         * Scripts/webkitpy/test/finder.py: Renamed from Tools/Scripts/webkitpy/test/test_finder.py.
808         * Scripts/webkitpy/test/finder_unittest.py: Renamed from Tools/Scripts/webkitpy/test/test_finder_unittest.py.
809         * Scripts/webkitpy/test/main.py:
810         (Tester.__init__):
811
812 2012-07-12  Adam Barth  <abarth@webkit.org>
813
814         CommitQueue is confused about what port it is using
815         https://bugs.webkit.org/show_bug.cgi?id=91040
816
817         Reviewed by Dirk Pranke.
818
819         On EC2, we explicitly pass --port to the commit-queue, but that
820         requires editing the start-queue.sh script locally on each bot. In
821         moving to Google Compute Engine, we're try to avoid any local edits to
822         the EWSTools.
823
824         Rather than passing --port to the commit-queue, this patch teaches the
825         CommitQueue which port its running, which is the approach we use for
826         the EWS bots.
827
828         Mutating tool._deprecated_port is a bit ugly, but it's what we're doing
829         currently for the EWS bots.
830
831         * Scripts/webkitpy/tool/commands/queues.py:
832         (CommitQueue):
833         (CommitQueue.begin_work_queue):
834         (CommitQueue.run_command):
835         * Scripts/webkitpy/tool/commands/queues_unittest.py:
836         (CommitQueueTest.test_commit_queue):
837         (mock_run_webkit_patch):
838         (test_rollout):
839         (test_rollout_lands):
840         (test_manual_reject_during_processing):
841
842 2012-07-12  James Simonsen  <simonjam@chromium.org>
843
844         [Navigation Timing] Import the W3C Navigation Timing test suite
845         https://bugs.webkit.org/show_bug.cgi?id=84887
846
847         Reviewed by Tony Gentilcore.
848
849         * Scripts/import-w3c-performance-wg-tests: Added.
850
851 2012-07-12  Adam Barth  <abarth@webkit.org>
852
853         Unreviewed.  Nit: git config files use tabs, not spaces.
854
855         * EWSTools/cold-boot.sh:
856
857 2012-07-12  Kwang Yul Seo  <skyul@company100.net>
858
859         Unreviewed. Add Dongsung Huang to the list of contributors. He
860         has submitted over 30 patches on texture mapper, canvas and image decoders.
861
862         * Scripts/webkitpy/common/config/committers.py:
863
864 2012-07-12  Kwang Yul Seo  <skyul@company100.net>
865
866         Unreviewed. Change my irc nickname to kseo.
867
868         * Scripts/webkitpy/common/config/committers.py:
869
870 2012-07-12  Dirk Pranke  <dpranke@chromium.org>
871
872         webkitpy: clean up logging handlers, lint common.message_pool
873         https://bugs.webkit.org/show_bug.cgi?id=91152
874
875         Reviewed by Ojan Vafai.
876
877         The unix implementation of multiprocessing clones any logging
878         handlers from the parent process into the child; we currently
879         don't want this behavior in our code, so I was hand-removing the
880         installed handlers in the child process I knew about. After thinking
881         about it further, I think it was simpler and safe enough to just
882         remove all handlers in the child, since the message pool
883         propagates any message from the child back into the parent.
884         
885         We can always change this in the future if it turns out to be an issue.
886
887         I'm also fixing a couple of other lint warnings while I'm at it.
888
889         * Scripts/webkitpy/common/message_pool.py:
890         (_MessagePool.__exit__):
891         (_MessagePool._handle_worker_exception):
892         (_Worker._set_up_logging):
893
894 2012-07-12  Dirk Pranke  <dpranke@chromium.org>
895
896         webkitpy: rename manager_worker_broker to message_pool
897         https://bugs.webkit.org/show_bug.cgi?id=91145
898
899         Reviewed by Ojan Vafai.
900
901         Since the MessagePool interface is more generic (and simpler)
902         now and will be reused by test-webkitpy, I'm renaming it and
903         moving it to webkitpy.common.
904
905         * Scripts/webkitpy/common/message_pool.py: Renamed from Tools/Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py.
906         * Scripts/webkitpy/layout_tests/controllers/manager.py:
907         (TestRunInterruptedException.__reduce__):
908         (Manager._run_tests.worker_factory):
909         (Manager._run_tests):
910         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
911
912 2012-07-12  Dirk Pranke  <dpranke@chromium.org>
913
914         nrwt crashes saving the output for a platform-specific expected test reference
915         https://bugs.webkit.org/show_bug.cgi?id=90872
916
917         Reviewed by Ojan Vafai.
918
919         The expected output for a test is copied alongside the test
920         itself in the layout-test-results directory; in other words, for
921         foo/bar-expected.txt sits alongside foo/bar.html even if we're
922         actually using platform/mac/foo/bar-expected.txt.
923
924         Unless the test is a reftest, in which case we would copy the
925         output to platform/mac/foo/bar-expected.html and set a
926         'ref_file' parameter in results.json to indicate the path. This
927         can be useful in the cases where we have multiple references for
928         a single test or when multiple tests share the same reference.
929
930         We found a bug where we weren't creating platform/mac/foo under
931         the results directory, and so this wasn't actually working.
932         However, treating reftests differently seems like a bad thing,
933         so we should probably be consistent. This change puts the
934         -expected.html next to the test, and reworks test_result_writer
935         so that we create directories uniformly and consistently.
936
937         Note that we weren't catching this problem in unit tests because
938         the MockFileSystem creates a directory automatically if it
939         doesn't exist; this was done intentionally for convenience, but
940         is really a bug and should be fixed; see https://bugs.webkit.org/show_bug.cgi?id=91028.
941
942         I have not added additional tests here since fixing that bug
943         should be sufficient.
944
945         * Scripts/webkitpy/layout_tests/controllers/manager.py:
946         (interpret_test_failures):
947         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
948         (ResultSummaryTest.test_interpret_test_failures):
949         * Scripts/webkitpy/layout_tests/controllers/test_result_writer.py:
950         (write_test_result):
951         (TestResultWriter._write_binary_file):
952         (TestResultWriter):
953         (TestResultWriter._write_text_file):
954         (TestResultWriter.write_output_files):
955         (TestResultWriter.write_stderr):
956         (TestResultWriter.write_crash_log):
957         (TestResultWriter.create_text_diff_and_write_result):
958         (TestResultWriter.write_image_diff_files):
959         (write_reftest):
960         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
961         (EndToEndTest.test_reftest_with_two_notrefs):
962
963 2012-07-12  Dirk Pranke  <dpranke@chromium.org>
964
965         nrwt: reimplement manager_worker_broker in a much simpler form
966         https://bugs.webkit.org/show_bug.cgi?id=90513
967
968         Reviewed by Ojan Vafai.
969
970         This is a wholesale replacement of the MessagePool() implementation
971         and the other classes in manager_worker_broker.py. All of the
972         BrokerConnection*, Broker*, etc. classes are gone, and there are now
973         just a MessagePool class and a _Worker class. Happiness ensues.
974  
975         I'm removing manager_worker_broker_unittest.py as well; we get
976         nearly complete coverage from the integration tests, and will
977         get more coverage when test-webkitpy moves to use this as well,
978         so having unit tests seems like unnecessary overhead. (running
979         coverage numbers with test-webkitpy shows that pretty much the only
980         uncovered lines are lines that are only run in the child processes,
981         which coverage doesn't handle at the moment).
982  
983         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
984         (_MessagePool.__init__):
985         (_MessagePool.run):
986         (_MessagePool._start_workers):
987         (_MessagePool):
988         (_MessagePool.wait):
989         (_MessagePool._close):
990         (_MessagePool._handle_done):
991         (_MessagePool._can_pickle):
992         (_MessagePool._loop):
993         (WorkerException):
994         (_Message.__init__):
995         (_Message.__repr__):
996         (_Worker):
997         (_Worker.__init__):
998         (_Worker.terminate):
999         (_Worker._close):
1000         (_Worker.run):
1001         (_Worker.post):
1002         (_Worker.yield_to_caller):
1003         (_Worker._post):
1004         (_Worker._raise):
1005         (_Worker._set_up_logging):
1006         (_WorkerLogHandler.__init__):
1007         (_WorkerLogHandler.emit):
1008         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py: Removed.
1009
1010 2012-07-12  Tony Chang  <tony@chromium.org>
1011
1012         [chromium] Remove drag and drop API methods that are no longer used
1013         https://bugs.webkit.org/show_bug.cgi?id=90996
1014
1015         Reviewed by Adam Barth.
1016
1017         Migrate DRT to use the methods that take modifier keys.
1018
1019         * DumpRenderTree/chromium/EventSender.cpp:
1020         (EventSender::doDragDrop):
1021         (EventSender::doMouseUp):
1022         (EventSender::doMouseMove):
1023         (EventSender::beginDragWithFiles):
1024
1025 2012-07-12  Ojan Vafai  <ojan@chromium.org>
1026
1027         Allow putting ranges in user.py list prompts
1028         https://bugs.webkit.org/show_bug.cgi?id=91115
1029
1030         Reviewed by Adam Barth.
1031
1032         Ranges are inclusive and denoted by a dash. This is useful for rebaselining a whole port
1033         since the items are listed with each port's builders being contiguous.
1034
1035         * Scripts/webkitpy/common/system/user.py:
1036         (User._wait_on_list_response):
1037         * Scripts/webkitpy/common/system/user_unittest.py:
1038         (UserTest.test_prompt_with_multiple_lists.run_prompt_test):
1039         (UserTest.test_prompt_with_multiple_lists):
1040
1041 2012-07-12  Arnaud Renevier  <arno@renevier.net>
1042
1043         [GTK] Implement disableImageLoading in DRT
1044         https://bugs.webkit.org/show_bug.cgi?id=87973
1045
1046         Reviewed by Martin Robinson.
1047
1048         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1049         (resetDefaultsToConsistentValues):
1050         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1051         (LayoutTestController::disableImageLoading):
1052
1053 2012-07-12  Simon Hausmann  <simon.hausmann@nokia.com>
1054
1055         [Qt] Internal symbols are exported on Linux
1056         https://bugs.webkit.org/show_bug.cgi?id=90981
1057
1058         Reviewed by Jocelyn Turcotte.
1059
1060         I should've re-enabled ELF symbol visibility when removing the symbol map in
1061         r106650.
1062
1063         * qmake/mkspecs/features/default_post.prf:
1064
1065 2012-07-11  Mark Rowe  <mrowe@apple.com>
1066
1067         <http://webkit.org/b/91024> Build against the latest SDK when targeting older OS X versions.
1068
1069         Reviewed by Dan Bernstein.
1070
1071         The deployment target is already set to the version that we're targeting, and it's that setting
1072         which determines which functionality from the SDK is available to us.
1073
1074         * DumpRenderTree/mac/Configurations/Base.xcconfig:
1075         * TestWebKitAPI/Configurations/Base.xcconfig:
1076         * WebKitTestRunner/Configurations/Base.xcconfig:
1077
1078 2012-07-11  Mark Rowe  <mrowe@apple.com>
1079
1080         Replace definitions of BUILDING_ON / TARGETING macros with macros that will error when used.
1081
1082         Part of <http://webkit.org/b/91015> Remove BUILDING_ON / TARGETING macros in favor of system availability macros.
1083
1084         Reviewed by Anders Carlsson.
1085
1086         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h: Remove the macros completely from here since
1087         they're completely unused in TestNetscapePlugIn.
1088         * DumpRenderTree/config.h:
1089
1090 2012-07-11  Mark Rowe  <mrowe@apple.com>
1091
1092         <http://webkit.org/b/91015> Remove BUILDING_ON / TARGETING macros in favor of system availability macros
1093
1094         This removal was handled by a script that translates the relevant macros in to the equivalent checks
1095         using the system availability macros.
1096
1097         Reviewed by Filip Pizlo.
1098
1099         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
1100         * DumpRenderTree/mac/CheckedMalloc.cpp:
1101         * DumpRenderTree/mac/DumpRenderTree.mm:
1102         * DumpRenderTree/mac/DumpRenderTreeDraggingInfo.mm:
1103         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1104         * DumpRenderTree/mac/TextInputController.m:
1105         * TestWebKitAPI/mac/InjectedBundleControllerMac.mm:
1106         * WebKitTestRunner/InjectedBundle/mac/ActivateFonts.mm:
1107         * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm:
1108
1109 2012-07-11  Robert Kroeger  <rjkroege@chromium.org>
1110
1111         Suppress horizontal conversion of PlatformWheelEvents when hasPreciseScrollingDeltas is true
1112         https://bugs.webkit.org/show_bug.cgi?id=89580
1113
1114         WebKit GTK and Chromium Linux force vertical wheel events to
1115         scroll horizontally when over horizontal scroll bars.  This is
1116         undesirable for touchpad scrolling with
1117         hasPreciseScrollingDeltas() == true. Modified DumpRenderTree to
1118         let a layout test specify this attribute so that the change's impact
1119         on scrolling can be tested in a layout test.
1120
1121         Reviewed by Adam Barth.
1122
1123         * DumpRenderTree/chromium/EventSender.cpp:
1124         (EventSender::handleMouseWheel):
1125
1126 2012-07-11  Simon Fraser  <simon.fraser@apple.com>
1127
1128         Fix the build by declaring -isPaginated before use.
1129
1130         * MiniBrowser/mac/BrowserWindowController.m:
1131
1132 2012-07-11  Sheriff Bot  <webkit.review.bot@gmail.com>
1133
1134         Unreviewed, rolling out r122358.
1135         http://trac.webkit.org/changeset/122358
1136         https://bugs.webkit.org/show_bug.cgi?id=91037
1137
1138         Build break on WebKit Win (Requested by hayato on #webkit).
1139
1140         * DumpRenderTree/chromium/EventSender.cpp:
1141         (EventSender::doDragDrop):
1142         (EventSender::doMouseUp):
1143         (EventSender::doMouseMove):
1144         (EventSender::beginDragWithFiles):
1145
1146 2012-07-03  Dirk Pranke  <dpranke@chromium.org>
1147
1148         nrwt: add a MessagePool abstraction that the manager will call to replace the broker
1149         https://bugs.webkit.org/show_bug.cgi?id=90511
1150
1151         Reviewed by Ojan Vafai.
1152
1153         This change introduces the new MessagePool abstraction that will
1154         replace the classes in manager_worker_broker. It is a minimal
1155         interface that tries to follow the conventions in
1156         multiprocessing.Pool and concurrency.futures ... it provides a
1157         context manager and a run() method that sends N messages to M
1158         workers processes (starting workers as necessary) and waits for
1159         them all to complete, handling cleanup as necessary. The caller
1160         is responsible for providing a handle() method to handle
1161         messages received from the workers.
1162         
1163         This interface basically hides all of the multiprocessing logic from
1164         the manager class.
1165
1166         The initial implementation of MessagePool is a simple shim
1167         around the existing broker classes; a subsequent change will
1168         replace all the other classes with a much simpler
1169         implementation.
1170
1171         No additional tests are provided for now; existing tests should
1172         provide adequate coverage, and I will add new unit tests for the
1173         MessagePool class when I replace the existing implementation.
1174
1175         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1176         (TestRunInterruptedException.__reduce__):
1177         (Manager.__init__):
1178         (Manager._run_tests):
1179         (Manager._run_tests.instead):
1180         (Manager.handle):
1181         (Manager._handle_started_test):
1182         (Manager._handle_finished_test_list):
1183         (Manager._handle_finished_test):
1184         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
1185         (get):
1186         (_MessagePool):
1187         (_MessagePool.__init__):
1188         (_MessagePool.__enter__):
1189         (_MessagePool.__exit__):
1190         (_MessagePool.run):
1191         (_MessagePool.wait):
1192         (_MessagePool.is_done):
1193         (_MessagePool._worker_is_done):
1194         (_MessagePool._close):
1195         (_MessagePool.handle_done):
1196         (_MessagePool.handle_started_test):
1197         (_MessagePool.handle_finished_test):
1198         (_MessagePool.handle_finished_test_list):
1199         (_MessagePool.handle_exception):
1200         (_MessagePool._log_messages):
1201         (_MessagePool._handle_worker_exception):
1202         (_WorkerState):
1203         (_WorkerState.for):
1204         (_WorkerState.__init__):
1205         (_WorkerState.__repr__):
1206         (_get_broker):
1207         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py:
1208         (make_broker):
1209
1210 2012-07-11  Simon Fraser  <simon.fraser@apple.com>
1211
1212         Add an option to enter paginated mode in MiniBrowser
1213         https://bugs.webkit.org/show_bug.cgi?id=91035
1214
1215         Reviewed by Dan Bernstein.
1216
1217         Add an item to the Debug menu for MiniBrowser that puts
1218         the web view into paginated mode. For now, we only
1219         do left-to-right pagination, with a fixed column width.
1220
1221         * MiniBrowser/mac/BrowserWindowController.h:
1222         * MiniBrowser/mac/BrowserWindowController.m:
1223         (-[BrowserWindowController validateMenuItem:]): Update the checked
1224         state of the menu item.
1225         (-[BrowserWindowController isPaginated]): Return YES if paginated.
1226         (-[BrowserWindowController togglePaginationMode:]): Toggle in
1227         and out of pagination mode.
1228         * MiniBrowser/mac/MainMenu.xib: Add the debug menu item.
1229
1230 2012-07-11  Adam Barth  <abarth@webkit.org>
1231
1232         commit-queue instances on Compute Engine are missing git-svn
1233         https://bugs.webkit.org/show_bug.cgi?id=91034
1234
1235         Reviewed by Eric Seidel.
1236
1237         git-svn is needed to actually commit to SVN from a git working copy.
1238
1239         * EWSTools/cold-boot.sh:
1240
1241 2012-07-11  Dirk Pranke  <dpranke@chromium.org>
1242
1243         webkitpy: lint code in webkitpy.layout_tests.models
1244         https://bugs.webkit.org/show_bug.cgi?id=90416
1245
1246         Reviewed by Ojan Vafai.
1247
1248         Cleaning up errors reported from lint-webkitpy.
1249
1250         Also, suppress the warnings about wildcard imports in pylintrc;
1251         we have nothing particularly against them.
1252
1253         * Scripts/webkitpy/layout_tests/models/test_configuration.py:
1254         (TestConfigurationConverter.combinations):
1255         * Scripts/webkitpy/layout_tests/models/test_configuration_unittest.py:
1256         (TestConfigurationTest.test_hash.query_unknown_key):
1257         (TestConfigurationTest.test_eq):
1258         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
1259         (ParseError.__init__):
1260         (TestExpectationLine.__init__):
1261         (TestExpectationsModel.get_expectations_string):
1262         (TestExpectationsModel):
1263         (TestExpectationsModel.expectation_to_string):
1264         (TestExpectationsModel.add_expectation_line):
1265         (TestExpectationsModel._clear_expectations_for_test):
1266         (TestExpectationsModel._remove_from_sets):
1267         (TestExpectations.get_expectations_string):
1268         (TestExpectations.expectation_to_string):
1269         (TestExpectations._report_warnings):
1270         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
1271         (Base.__init__):
1272         (parse_exp):
1273         (SkippedTests.check):
1274         (TestExpectationParserTests.test_parse_empty_string):
1275         * Scripts/webkitpy/layout_tests/models/test_failures.py:
1276         (FailureTimeout.__init__):
1277         (FailureCrash.__init__):
1278         (FailureImageHashMismatch.__init__):
1279         (FailureReftestMismatch.__init__):
1280         (FailureReftestMismatchDidNotOccur.__init__):
1281         (FailureReftestNoImagesGenerated.__init__):
1282         * Scripts/webkitpy/layout_tests/models/test_failures_unittest.py:
1283         (TestFailuresTest.test_unknown_failure_type.UnknownFailure.message):
1284         (TestFailuresTest.test_unknown_failure_type):
1285         (TestFailuresTest):
1286         (TestFailuresTest.test_message_is_virtual):
1287         * Scripts/webkitpy/layout_tests/models/test_results.py:
1288         (TestResult.loads):
1289         (TestResult.has_failure_matching_types):
1290         * Scripts/webkitpy/pylintrc:
1291
1292 2012-07-11  Dirk Pranke  <dpranke@chromium.org>
1293
1294         nrwt: clean up names in worker.py
1295         https://bugs.webkit.org/show_bug.cgi?id=90510
1296
1297         Reviewed by Ojan Vafai.
1298
1299         This is the last patch in the series of refactoring worker.py;
1300         all this does is change some names of methods, instance
1301         variables, and method parameters to be clearer (it also changes
1302         some code in manager.py and manager_worker.py to be consistent).
1303
1304         There are no functional changes in this patch and the existing
1305         tests should cover everything.
1306
1307         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1308         (Manager.__init__):
1309         (Manager.prepare_lists_and_print_output):
1310         (Manager._run_tests.worker_factory):
1311         (Manager._run_tests):
1312         (Manager._show_results_html_file):
1313         (Manager.handle_finished_test_list):
1314         (_WorkerState.__init__):
1315         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
1316         (_Broker.post_message):
1317         (_Broker._dispatch_message):
1318         (AbstractWorker.__init__):
1319         (AbstractWorker.run):
1320         (AbstractWorker.yield_to_caller):
1321         (AbstractWorker.post):
1322         (_WorkerConnection.__init__):
1323         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py:
1324         (_TestWorker.__init__):
1325         (_TestWorker.start):
1326         (_TestWorker.handle):
1327         (_TestWorker.stop):
1328         (_TestsMixin):
1329         (_TestsMixin.test_name):
1330         * Scripts/webkitpy/layout_tests/controllers/worker.py:
1331         (Worker.__init__):
1332         (Worker.__del__):
1333         (Worker.start):
1334         (Worker.handle):
1335         (Worker._run_test):
1336         (Worker.stop):
1337         (Worker._timeout):
1338         (Worker._kill_driver):
1339         (Worker._run_test_with_timeout):
1340         (Worker._clean_up_after_test):
1341         (Worker._run_test_in_another_thread.SingleTestThread.run):
1342         (Worker._run_test_in_this_thread):
1343         (Worker._run_single_test):
1344
1345 2012-07-11  Adam Barth  <abarth@webkit.org>
1346
1347         The commit-queue needs some extra git config to be able to commit
1348         https://bugs.webkit.org/show_bug.cgi?id=91025
1349
1350         Reviewed by Eric Seidel.
1351
1352         In order for the commit-queue to actually commit, it needs to know the
1353         location of the SVN server.
1354
1355         * EWSTools/cold-boot.sh:
1356
1357 2012-07-11  Adam Barth  <abarth@webkit.org>
1358
1359         Teach EWSTools how to configure SVN auth credentials
1360         https://bugs.webkit.org/show_bug.cgi?id=91021
1361
1362         Reviewed by Eric Seidel.
1363
1364         To move the commit-queue over to Google Compute Engine, we need a way
1365         to configure the commit-queue's credentials during the machine build
1366         process. There doesn't seem to be a nice command line way of
1367         configuring SVN auth credentials. I tried doing a bogus commit and
1368         supplying the credentials that way, but that's super hacky.
1369
1370         The approach in this patch is to write the config file directly. The
1371         format of these configurations files is documented in the SVN book, so
1372         this approach doesn't see too sketchy.
1373
1374         * EWSTools/configure-svn-auth.sh: Added.
1375
1376 2012-07-11  Ojan Vafai  <ojan@chromium.org>
1377
1378         Dedupe suffixes passed to webkit-patch rebaseline
1379         https://bugs.webkit.org/show_bug.cgi?id=91017
1380
1381         Reviewed by Dirk Pranke.
1382
1383         * Scripts/webkitpy/tool/commands/rebaseline.py:
1384         (Rebaseline._suffixes_to_update):
1385         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
1386         (test_rebaseline_multiple_builders_and_tests_command_line):
1387
1388 2012-07-11  Kevin Ollivier  <kevino@theolliviers.com>
1389
1390         [wx] Unreviewed build fix. Add new directories and a new LayoutTestController method.
1391
1392         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
1393         (LayoutTestController::setStorageDatabaseIdleInterval):
1394         * waf/build/settings.py:
1395
1396 2012-07-11  Tony Chang  <tony@chromium.org>
1397
1398         [chromium] Remove drag and drop API methods that are no longer used
1399         https://bugs.webkit.org/show_bug.cgi?id=90996
1400
1401         Reviewed by Adam Barth.
1402
1403         Migrate DRT to use the methods that take modifier keys.
1404
1405         * DumpRenderTree/chromium/EventSender.cpp:
1406         (EventSender::doDragDrop):
1407         (EventSender::doMouseUp):
1408         (EventSender::doMouseMove):
1409         (EventSender::beginDragWithFiles):
1410
1411 2012-07-11  Mark Rowe  <mrowe@apple.com>
1412
1413         Add a Mountain Lion version of libWebKitSystemInterface.a.
1414
1415         Reviewed by John Sullivan.
1416
1417         * Scripts/copy-webkitlibraries-to-product-directory: Include libWebKitSystemInterfaceMountainLion.a in the list of libraries to copy.
1418
1419 2012-07-09  Mark Rowe  <mrowe@apple.com>
1420
1421         <http://webkit.org/b/90835> Teach bisect-builds to work with a Safari.app that has entitlements.
1422
1423         Reviewed by Dan Bernstein.
1424
1425         * Scripts/bisect-builds: Use safariPathFromSafariBundle to determine which binary within the application
1426         should be invoked.
1427
1428 2012-07-11  Zan Dobersek  <zandobersek@gmail.com>
1429
1430         [Gtk] fast/events/keydown-function-keys.html is failing
1431         https://bugs.webkit.org/show_bug.cgi?id=90891
1432
1433         Reviewed by Martin Robinson.
1434
1435         Work around the context menu being shown on F10 key being pressed by
1436         unbiding the key when running tests in DumpRenderTree. The problem
1437         appears when using a recent version of the xkeyboard-config package.
1438
1439         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1440         (setDefaultsToConsistentStateValuesForTesting):
1441
1442 2012-07-11  No'am Rosenthal  <noam.rosenthal@nokia.com>
1443
1444         [Qt] QRawWebView should notify when rendering is done, so that pixel results can be grabbed at the appropriate moment.
1445         https://bugs.webkit.org/show_bug.cgi?id=90641
1446
1447         Reviewed by Jocelyn Turcotte.
1448
1449         * MiniBrowser/qt/raw/View.h: Comment used old class name (WKView).
1450         (View):
1451
1452 2012-07-11  Min Qin  <qinmin@chromium.org>
1453
1454         [Android] sending an extra to the DRT apk so that DRT can be run in a seperate thread
1455         https://bugs.webkit.org/show_bug.cgi?id=90831
1456
1457         Reviewed by Adam Barth.
1458
1459         On android, DRT needs to run in a background thread to avoid ANR.
1460         However, the java tests are running on UI thread by default.
1461         We need to send an intent extra to the apk so that it can run on a sub thread.
1462
1463         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
1464         (ChromiumAndroidDriver._start_once):
1465
1466 2012-07-11  Csaba Osztrogonác  <ossy@webkit.org>
1467
1468         [Qt] REGRESSION(r107171): Fix --timeout option of Qt's DRT
1469         https://bugs.webkit.org/show_bug.cgi?id=90966
1470
1471         Reviewed by Ryosuke Niwa.
1472
1473         * DumpRenderTree/qt/main.cpp:
1474         (main): Don't remove the argument, because takeOptionValue() did it before.
1475
1476 2012-07-11  Allan Sandfeld Jensen  <allan.jensen@nokia.com>
1477
1478         First commit; moving myself to commiters.
1479
1480         Unreviewed.
1481
1482         * Scripts/webkitpy/common/config/committers.py:
1483
1484 2012-07-11  Balazs Ankes  <bank@inf.u-szeged.hu>
1485
1486         webkit-patch land should automatically add svn:mime-type for .png files
1487         https://bugs.webkit.org/show_bug.cgi?id=75825
1488
1489         Reviewed by Dirk Pranke.
1490
1491         Refactor the png.py to avoid code duplication.
1492
1493         * Scripts/webkitpy/common/checksvnconfigfile.py: Added.
1494         (check):
1495         (config_file_path):
1496         (errorstr_autoprop):
1497         (errorstr_png):
1498         * Scripts/webkitpy/style/checkers/png.py:
1499         (PNGChecker.check):
1500         * Scripts/webkitpy/tool/commands/download.py:
1501         (Land):
1502         * Scripts/webkitpy/tool/steps/__init__.py:
1503         * Scripts/webkitpy/tool/steps/addsvnmimetypeforpng.py: Added.
1504         (AddSvnMimetypeForPng):
1505         (AddSvnMimetypeForPng.__init__):
1506         (AddSvnMimetypeForPng.run):
1507         (AddSvnMimetypeForPng._check_pngs):
1508         * Scripts/webkitpy/tool/steps/addsvnmimetypeforpng_unittest.py: Added.
1509         (MockSCMDetector):
1510         (MockSCMDetector.__init__):
1511         (MockSCMDetector.display_name):
1512         (AddSvnMimetypeForPngTest):
1513         (AddSvnMimetypeForPngTest.test_run):
1514
1515 2012-07-11  Hayato Ito  <hayato@chromium.org>
1516
1517         Unreviewed gardening.
1518         One more fix for r122292.
1519
1520         * TestWebKitAPI/TestWebKitAPI.gyp/TestWebKitAPI.gyp:
1521
1522 2012-07-11  Sheriff Bot  <webkit.review.bot@gmail.com>
1523
1524         Unreviewed, rolling out r122301.
1525         http://trac.webkit.org/changeset/122301
1526         https://bugs.webkit.org/show_bug.cgi?id=90947
1527
1528         Android builder started to fail (Requested by hayato on
1529         #webkit).
1530
1531         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
1532
1533 2012-07-10  Adam Barth  <abarth@webkit.org>
1534
1535         [Chromium] Merge final nits to DumpRenderTree.gyp for Android
1536         https://bugs.webkit.org/show_bug.cgi?id=90920
1537
1538         Reviewed by Tony Chang.
1539
1540         This patch contains the last few small changes to DumpRenderTree.gyp
1541         from the chromium-android branch. After this change, this file will be
1542         fully merged.
1543
1544         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
1545
1546 2012-07-10  Adam Barth  <abarth@webkit.org>
1547
1548         [Chromium-Android] Add apk test targets for webkit_unit_tests and TestWebKitAPI
1549         https://bugs.webkit.org/show_bug.cgi?id=90918
1550
1551         Reviewed by Tony Chang.
1552
1553         The rules are similar to what we have added for DumpRenderTree apk.
1554         All references to gtest_target_type can be removed once we enable APK
1555         tests on the all bots.
1556
1557         * TestWebKitAPI/TestWebKitAPI.gyp/TestWebKitAPI.gyp:
1558
1559 2012-07-10  Kenneth Russell  <kbr@google.com>
1560
1561         Add --no-build option to perf test runner
1562         https://bugs.webkit.org/show_bug.cgi?id=90916
1563
1564         Reviewed by Ryosuke Niwa.
1565
1566         * Scripts/webkitpy/performance_tests/perftestsrunner.py:
1567         (PerfTestsRunner._parse_args):
1568             Support --no-build option to perf tests, as in layout tests.
1569
1570 2012-07-10  Adam Barth  <abarth@webkit.org>
1571
1572         EWSTools should be able to build a chromium-ews bot from scratch
1573         https://bugs.webkit.org/show_bug.cgi?id=90912
1574
1575         Reviewed by Eric Seidel.
1576
1577         I've been using this script to kick off the build process for the
1578         chromium-ews bots on Google Compute Engine.
1579
1580         * EWSTools/GoogleComputeEngine: Added.
1581         * EWSTools/GoogleComputeEngine/build-chromium-ews.sh: Added.
1582
1583 2012-07-10  Alec Flett  <alecflett@chromium.org>
1584
1585         Add alecflett to the list of committers
1586         https://bugs.webkit.org/show_bug.cgi?id=90903
1587
1588         Reviewed by Ojan Vafai.
1589
1590         * Scripts/webkitpy/common/config/committers.py:
1591
1592 2012-07-10  Xianzhu Wang  <wangxianzhu@chromium.org>
1593
1594         Unreviewed fix of webkitpy unittest break about chromium_android.py
1595         caused by r122251.
1596
1597         * Scripts/webkitpy/layout_tests/port/chromium_android_unittest.py:
1598         (ChromiumAndroidDriverTest.test_cmd_line):
1599
1600 2012-07-10  Ojan Vafai  <ojan@chromium.org>
1601
1602         webkit-patch rebaseline doesn't work for audio/pixel tests
1603         https://bugs.webkit.org/show_bug.cgi?id=90905
1604
1605         Reviewed by Adam Barth.
1606
1607         We just need to pass the correct suffix list to rebaseline-test-internal
1608         and optimize-baselines. By default, pass all suffixes.
1609
1610         * Scripts/webkitpy/tool/commands/rebaseline.py:
1611         (Rebaseline):
1612         (Rebaseline.__init__):
1613         (Rebaseline._suffixes_to_update):
1614         (Rebaseline.execute):
1615         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
1616         (test_rebaseline_multiple_builders_and_tests_command_line):
1617
1618 2012-07-10  Adam Barth  <abarth@webkit.org>
1619
1620         Remove LayoutTestController and WebKitTestRunner support for Hixie76 WebSockets
1621         https://bugs.webkit.org/show_bug.cgi?id=90853
1622
1623         Reviewed by Eric Seidel.
1624
1625         WebKit no longer implements the Hixie76 version of the WebSocket
1626         protocol. We don't need to support it in our test framework.
1627
1628         * DumpRenderTree/chromium/LayoutTestController.cpp:
1629         (LayoutTestController::overridePreference):
1630         * DumpRenderTree/chromium/WebPreferences.cpp:
1631         (WebPreferences::reset):
1632         (WebPreferences::applyTo):
1633         * DumpRenderTree/chromium/WebPreferences.h:
1634         (WebPreferences):
1635         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1636         (resetDefaultsToConsistentValues):
1637         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1638         (LayoutTestController::overridePreference):
1639         * DumpRenderTree/mac/DumpRenderTree.mm:
1640         (resetDefaultsToConsistentValues):
1641         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1642         (WebCore::WebPage::resetSettings):
1643         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1644         (LayoutTestController::overridePreference):
1645         * DumpRenderTree/win/DumpRenderTree.cpp:
1646         (resetDefaultsToConsistentValues):
1647
1648 2012-07-10  Xianzhu Wang  <wangxianzhu@chromium.org>
1649
1650         [Chromium-Android] Fix typos in chromium_android.py in r151492
1651         https://bugs.webkit.org/show_bug.cgi?id=90904
1652
1653         Reviewed by Ojan Vafai.
1654
1655         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
1656         (ChromiumAndroidPort.setup_test_run):
1657
1658 2012-07-10  Xianzhu Wang  <wangxianzhu@chromium.org>
1659
1660         [Chromium-Android] Use setup_test_runner() instead of start_helper() to setup test environment
1661         https://bugs.webkit.org/show_bug.cgi?id=90894
1662
1663         Reviewed by Adam Barth.
1664
1665         start_helper() is actually start_pixel_test_helper() since r115601 (bug 81729).
1666         Should use setup_test_runner() to setup test environment for chromium-android.
1667
1668         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
1669         (ChromiumAndroidPort.setup_test_run): Renamed from start_helper(). Added cache cleanup code.
1670         (ChromiumAndroidPort.clean_up_test_run): Renamed from stop_helper().
1671         (ChromiumAndroidPort._path_to_helper): Returns None as we don't have a helper now.
1672         (ChromiumAndroidPort):
1673         (ChromiumAndroidPort._path_to_forwarder): Original _path_to_helper().
1674         (ChromiumAndroidPort._push_executable):
1675         (ChromiumAndroidDriver.__init__):
1676         (ChromiumAndroidDriver.cmd_line):
1677
1678 2012-07-09  Ojan Vafai  <ojan@chromium.org>
1679
1680         Improve webkit-patch rebaseline to work for more cases
1681         https://bugs.webkit.org/show_bug.cgi?id=90504
1682
1683         Reviewed by Adam Barth.
1684
1685         -Makes it work for the build.chromium.org bots.
1686         -Lets you rebaseline all builders instead of just one.
1687         -Lets you pass in the builders or tests to rebaseline.
1688
1689         * Scripts/webkitpy/common/host.py:
1690         (Host.buildbot_for_builder_name):
1691         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
1692         (Builder.__init__):
1693         (Builder.latest_layout_test_results):
1694         Provide a way to get to the LayoutTestResults of the latest build.
1695         Most of the time we only need to get the latest one and the Chromium bots
1696         only expose the full_results.json file for the latest build.
1697
1698         (Builder):
1699         (Builder._fetch_file_from_results):
1700         (Builder.fetch_layout_test_results):
1701         Move these functions into Builder so that Builder and Build can both
1702         fetch layout test results.
1703
1704         (Build.results_zip_url):
1705         (Build.layout_test_results):
1706         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1707         (BuilderTest.test_latest_layout_test_results):
1708         (BuildTest.test_layout_test_results):
1709         * Scripts/webkitpy/common/system/user.py:
1710         (User.prompt_with_multiple_lists):
1711         Prompt with multiple sublists, but still have a global numbering.
1712         This lets the build.chromium.org bots be clearly separated from the
1713         build.webkit.org bots, which helps understand the builder names.
1714
1715         (User):
1716         (User._wait_on_list_response):
1717         (User.prompt_with_list):
1718         * Scripts/webkitpy/common/system/user_unittest.py:
1719         (UserTest.test_prompt_with_multiple_lists):
1720         (UserTest.test_prompt_with_multiple_lists.run_prompt_test):
1721         (UserTest.test_prompt_with_multiple_lists.run_prompt_test.mock_raw_input):
1722         * Scripts/webkitpy/layout_tests/port/base.py:
1723         (Port.is_chromium):
1724         Provide a way to tell if a port is a Chromium port that doesn't involve string
1725         manipulation on the port name or builder name.
1726
1727         * Scripts/webkitpy/layout_tests/port/builders.py:
1728         (builder_path_from_name):
1729         (all_builder_names):
1730         Memoizing here is incorrect because the test override _exact_matches,
1731         so these can return different values. In either case, I'm pretty sure these
1732         are not remotely hot codepaths. I manually inspected all call sites and they're
1733         all in high-level calls (e.g. execute calls for webkit-patch commands) and not
1734         called in a loop.
1735
1736         * Scripts/webkitpy/layout_tests/port/chromium.py:
1737         (ChromiumPort.is_chromium):
1738         * Scripts/webkitpy/tool/commands/rebaseline.py:
1739         (RebaselineTest._results_url):
1740         (AbstractParallelRebaselineCommand.__init__):
1741         (Rebaseline.__init__):
1742         (Rebaseline):
1743         (Rebaseline._builders_to_pull_from):
1744         (Rebaseline._tests_to_update):
1745         (Rebaseline.execute):
1746         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
1747         (test_rebaseline.mock_builders_to_pull_from):
1748         (test_rebaseline):
1749         (test_rebaseline_command_line_flags):
1750         (test_rebaseline_multiple_builders):
1751         (test_rebaseline_multiple_builders.mock_builders_to_pull_from):
1752         (test_rebaseline_multiple_builders.mock_tests_to_update):
1753         (test_rebaseline_multiple_builders_and_tests_command_line):
1754         * Scripts/webkitpy/tool/mocktool.py:
1755         (MockTool.irc):
1756         (MockTool):
1757         (MockTool.buildbot_for_builder_name):
1758
1759 2012-07-10  Alice Cheng  <alice_cheng@apple.com>
1760
1761         Editing: Reproducible crasher when pasting a 0x0 image into Mail
1762         https://bugs.webkit.org/show_bug.cgi?id=90640
1763         <rdar://problem/11141920>
1764
1765         Reviewed by Brady Eidson.
1766
1767         Test cases for the patch. 0x0 images don't get a resource representation in the WebArchive, so we need a null check.
1768
1769         * TestWebKitAPI/Tests/mac/0.png: Added.
1770         * TestWebKitAPI/Tests/mac/WebViewCanPasteZeroPng.mm: Added.
1771         (TestWebKitAPI):
1772         (TestWebKitAPI::TEST):
1773
1774 2012-07-10  Leandro Gracia Gil  <leandrogracia@chromium.org>
1775
1776         WebSurroundingText layout tests should use the same code path as the rest of the feature.
1777         https://bugs.webkit.org/show_bug.cgi?id=90807
1778
1779         Reviewed by Adam Barth.
1780
1781         Make the textSurroundingNode method take a pair of point coordinates
1782         instead of a node offset.
1783
1784         * DumpRenderTree/chromium/LayoutTestController.cpp:
1785         (LayoutTestController::textSurroundingNode):
1786         * DumpRenderTree/chromium/LayoutTestController.h:
1787         (LayoutTestController):
1788
1789 2012-07-10  Csaba Osztrogonác  <ossy@webkit.org>
1790
1791         [Qt] REGRESSION(r122175): It broke the Windows build
1792         https://bugs.webkit.org/show_bug.cgi?id=90850
1793
1794         Reviewed by Noam Rosenthal.
1795
1796         * qmake/mkspecs/features/features.prf: Disable USE(3D_GRAPHICS) on Windows temporarily until proper fix.
1797
1798 2012-07-10  János Badics  <jbadics@inf.u-szeged.hu>
1799
1800         [Qt][NRWT] Enable cascaded TestExpectations
1801         https://bugs.webkit.org/show_bug.cgi?id=89108
1802
1803         Reviewed by Dirk Pranke.
1804
1805         * Scripts/webkitpy/layout_tests/port/qt.py:
1806         (QtPort.expectations_files):
1807         * Scripts/webkitpy/layout_tests/port/qt_unittest.py:
1808         (QtPortTest._assert_expectations_files):
1809         (QtPortTest.test_expectations_files):
1810
1811 2012-07-10  Christophe Dumez  <christophe.dumez@intel.com>
1812
1813         [EFL] WebKit DRT and WTR fail to build due to undefined reference to WTF::MD5::*
1814         https://bugs.webkit.org/show_bug.cgi?id=90868
1815
1816         Unreviewed EFL build fix.
1817
1818         Correct CMake configuration to have EFL's DRT and WebKitTestRunner link
1819         WTF library. This is needed to resolve undefined reference to WTF::MD5::*.
1820
1821         * DumpRenderTree/efl/CMakeLists.txt:
1822         * WebKitTestRunner/PlatformEfl.cmake:
1823
1824 2012-07-10  Adam Barth  <abarth@webkit.org>
1825
1826         LayoutTestController.dumpConfigurationForViewport should move to Internals
1827         https://bugs.webkit.org/show_bug.cgi?id=45652
1828
1829         Reviewed by Eric Seidel.
1830
1831         Remove dumpConfigurationForViewport from LayoutTestController. This API
1832         has been replaced by an API on Internals that does the same thing in a
1833         port-agnostic way.
1834
1835         * DumpRenderTree/LayoutTestController.cpp:
1836         (LayoutTestController::staticFunctions):
1837         * DumpRenderTree/LayoutTestController.h:
1838         (LayoutTestController):
1839         * DumpRenderTree/blackberry/LayoutTestControllerBlackBerry.cpp:
1840         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp:
1841         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1842         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1843         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1844         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1845         (LayoutTestController):
1846         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1847         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
1848
1849 2012-07-09  Matt Falkenhagen  <falken@chromium.org>
1850
1851         Add ENABLE_DIALOG_ELEMENT and skeleton files
1852         https://bugs.webkit.org/show_bug.cgi?id=90521
1853
1854         Reviewed by Kent Tamura.
1855
1856         * Scripts/webkitperl/FeatureList.pm:
1857
1858 2012-07-09  Kent Tamura  <tkent@chromium.org>
1859
1860         Improve the boilerplate generated by prepare-ChangeLog
1861         https://bugs.webkit.org/show_bug.cgi?id=89560
1862
1863         Reviewed by Ryosuke Niwa.
1864
1865         Produce the following boilerplate:
1866
1867         > 2012-06-20  Kent Tamura  <tkent@chromium.org>
1868         >
1869         >         Need a short description (Oops!).
1870         >         Need the bug URL (Oops!).
1871         >
1872         >         Reviewed by NOBODY (Oops!).
1873         >
1874         >         Additional information of the change such as approach, rationale. Please add per-function descriptions below. (Oops!).
1875         >
1876         >         No new tests (Oops!).
1877
1878         * Scripts/prepare-ChangeLog:
1879         (generateNewChangeLogs): Generate the updated template.
1880         * Scripts/webkitpy/common/checkout/changelog.py:
1881         (ChangeLog.update_with_unreviewed_message):
1882         Update the pattern for the beginning of the boilerplate.
1883         (ChangeLog.set_short_description_and_bug_url):
1884         - Use self._changelog_indent
1885         - Update the pattern for the description line
1886         - Skip the bug boilerplate.
1887         We substitute the specified bug URL to the description boilerplate
1888         because we need to add it even if the bug URL boilerplate is missing.
1889         * Scripts/webkitpy/common/checkout/changelog_unittest.py: Update the boilerplate data.
1890         (test_set_short_description_and_bug_url):
1891         - Update the description boilerplate.
1892         - Add a test for a case with both of the description boilerplace and the
1893           bug URL boilerplate.
1894         * Scripts/webkitpy/tool/steps/preparechangelog_unittest.py:
1895         (PrepareChangeLogTest.test_ensure_bug_url): Update the expected boilerplate.
1896
1897 2012-07-09  No'am Rosenthal  <noam.rosenthal@nokia.com>
1898
1899         Shared code that is guarded with ENABLE(WEBGL) should be guarded with USE()
1900         https://bugs.webkit.org/show_bug.cgi?id=90506
1901
1902         Reviewed by Martin Robinson.
1903
1904         Always enable the WTF_USE_3D_GRAPHICS flag in features.prf if we're compiling with Qt5
1905         or if the OpenGL module is enabled.
1906
1907         * qmake/mkspecs/features/features.prf:
1908
1909 2012-07-09  Yongjun Zhang  <yongjun_zhang@apple.com>
1910
1911         Consider closing unused localStorage database after a timeout.
1912         https://bugs.webkit.org/show_bug.cgi?id=90713
1913
1914         For a localStorage, if there is no active document referencing to it for certain amount of time (300 seconds),
1915         we can close the underlying sqlite database.
1916
1917         Reviewed by Brady Eidson.
1918
1919         * DumpRenderTree/LayoutTestController.cpp: add a JS method setStorageDatabaseIdleInterval to testRunner
1920             to set a different timeout value for closing localStorage database.
1921         (setStorageDatabaseIdleIntervalCallback):
1922         (LayoutTestController::staticFunctions):
1923         * DumpRenderTree/LayoutTestController.h:
1924         (LayoutTestController):
1925         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp:
1926         (LayoutTestController::setStorageDatabaseIdleInterval): add empty method.
1927         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1928         (LayoutTestController::setStorageDatabaseIdleInterval): ditto.
1929         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1930         (LayoutTestController::setStorageDatabaseIdleInterval): ditto.
1931         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1932         (LayoutTestController::setStorageDatabaseIdleInterval): ditto.
1933
1934 2012-07-09  Adam Klein  <adamk@chromium.org>
1935
1936         Rename WebCore::WebKitMutationObserver to WebCore::MutationObserver
1937         https://bugs.webkit.org/show_bug.cgi?id=90810
1938
1939         Reviewed by Ojan Vafai.
1940
1941         * DumpRenderTree/gtk/EventSender.cpp:
1942
1943 2012-07-09  Mihai Parparita  <mihaip@chromium.org>
1944
1945         Handle missing results in TestResultsServer better
1946         https://bugs.webkit.org/show_bug.cgi?id=90816
1947
1948         Reviewed by Ojan Vafai.
1949
1950         If we can't load the results JSON, don't try to wrap it with the JSONP
1951         callback invocation.
1952
1953         * TestResultServer/handlers/testfilehandler.py:
1954         (GetFile.get):
1955
1956 2012-07-09  Sheriff Bot  <webkit.review.bot@gmail.com>
1957
1958         Unreviewed, rolling out r122124.
1959         http://trac.webkit.org/changeset/122124
1960         https://bugs.webkit.org/show_bug.cgi?id=90815
1961
1962         It broke NRWT on Qt (Requested by Ossy_HOME on #webkit).
1963
1964         * Scripts/webkitpy/layout_tests/port/qt.py:
1965         (QtPort._skipped_file_search_paths):
1966         * Scripts/webkitpy/layout_tests/port/qt_unittest.py:
1967         (QtPortTest._assert_skipped_path):
1968         (QtPortTest.test_skipped_file_search_path):
1969
1970 2012-07-09  Zan Dobersek  <zandobersek@gmail.com>
1971
1972         [Gtk] Implement dumpFrameScrollPosition in DumpRenderTree
1973         https://bugs.webkit.org/show_bug.cgi?id=89356
1974
1975         Reviewed by Martin Robinson.
1976
1977         Implement dumpFrameScrollPosition, acquiring the scroll position
1978         through WebKitDOMDOMWindow of the WebKitDOMDocument loaded in the frame
1979         whose scroll position is being dumped.
1980
1981         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1982         (dumpFrameScrollPosition):
1983
1984 2012-07-09  János Badics  <jbadics@inf.u-szeged.hu>
1985
1986         [Qt][NRWT] Enable cascaded TestExpectations
1987         https://bugs.webkit.org/show_bug.cgi?id=89108
1988
1989         Reviewed by Dirk Pranke.
1990
1991         * Scripts/webkitpy/layout_tests/port/qt.py:
1992         (QtPort.expectations_files):
1993         * Scripts/webkitpy/layout_tests/port/qt_unittest.py:
1994         (QtPortTest._assert_expectations_files):
1995         (QtPortTest.test_expectations_files):
1996
1997 2012-07-09  Dan Bernstein  <mitz@apple.com>
1998
1999         Fixed clean builds of the All target in the DumpRenderTree project.
2000
2001         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Let Xcode know that the
2002         DumpRenderTree Perl Support target depends on the DumpRenderTree target.
2003
2004 2012-07-09  Zan Dobersek  <zandobersek@gmail.com>
2005
2006         [Gtk] Unskip the CSS Regions layout tests
2007         https://bugs.webkit.org/show_bug.cgi?id=90771
2008
2009         Reviewed by Martin Robinson.
2010
2011         Use the DumpRenderTreeSupportGtk method for enabling the CSS Regions,
2012         calling it when resetting the defaults in DumpRenderTree or when the
2013         settings value is overriden through the testRunner.
2014
2015         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2016         (resetDefaultsToConsistentValues):
2017         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2018         (LayoutTestController::overridePreference):
2019
2020 2012-07-09  Sheriff Bot  <webkit.review.bot@gmail.com>
2021
2022         Unreviewed, rolling out r122107.
2023         http://trac.webkit.org/changeset/122107
2024         https://bugs.webkit.org/show_bug.cgi?id=90794
2025
2026         Build failure on Mac debug bots (Requested by falken_ on
2027         #webkit).
2028
2029         * Scripts/webkitperl/FeatureList.pm:
2030
2031 2012-07-09  Matt Falkenhagen  <falken@chromium.org>
2032
2033         Add ENABLE_DIALOG_ELEMENT and skeleton files
2034         https://bugs.webkit.org/show_bug.cgi?id=90521
2035
2036         Reviewed by Kent Tamura.
2037
2038         * Scripts/webkitperl/FeatureList.pm:
2039
2040 2012-07-08  Kent Tamura  <tkent@chromium.org>
2041
2042         Unreviewed, rolling out r122076.
2043         http://trac.webkit.org/changeset/122076
2044         https://bugs.webkit.org/show_bug.cgi?id=89560
2045
2046         Broke sheriffbot rollout
2047
2048         * Scripts/prepare-ChangeLog:
2049         (generateNewChangeLogs):
2050
2051 2012-07-08  Leandro Gracia Gil  <leandrogracia@chromium.org>
2052
2053         Update the description of LayoutTestController::textSurroundingElement
2054         https://bugs.webkit.org/show_bug.cgi?id=90620
2055
2056         Reviewed by Adam Barth.
2057
2058         Rename and fix the description of the method as its behaviour changed
2059         during the review of https://bugs.webkit.org/show_bug.cgi?id=82461
2060
2061         * DumpRenderTree/chromium/LayoutTestController.cpp:
2062         (LayoutTestController::LayoutTestController):
2063         (LayoutTestController::textSurroundingNode):
2064         * DumpRenderTree/chromium/LayoutTestController.h:
2065         (LayoutTestController):
2066
2067 2012-07-08  Kent Tamura  <tkent@chromium.org>
2068
2069         Improve the template generated by prepare-ChangeLog
2070         https://bugs.webkit.org/show_bug.cgi?id=89560
2071
2072         Reviewed by Ryosuke Niwa.
2073
2074         Produce the following template:
2075
2076         > 2012-06-20  Kent Tamura  <tkent@chromium.org>
2077         >
2078         >         Need a short description (Oops!).
2079         >         Need the bug URL (Oops!).
2080         >
2081         >         Reviewed by Ryosuke Niwa.
2082         >
2083         >         Additional information of the change such as approach, rationale. Please add per-function descriptions below. (Oops!).
2084         >
2085         >         No new tests (Oops!).
2086
2087         * Scripts/prepare-ChangeLog:
2088         (generateNewChangeLogs):
2089
2090 2012-07-08  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
2091
2092         [wx] Unreviewed. Fix notificaitons directory path after notifications move.
2093
2094         * waf/build/settings.py:
2095
2096 2012-07-07  Nuno Lopes  <nlopes@apple.com>
2097
2098         Fix build with recent clang.
2099         https://bugs.webkit.org/show_bug.cgi?id=90712
2100
2101         Reviewed by Alexey Proskuryakov.
2102
2103         * TestWebKitAPI/Tests/WebKit2/DOMWindowExtensionBasic_Bundle.cpp:
2104         (DOMWindowExtensionBasic):
2105         * TestWebKitAPI/Tests/WebKit2/DOMWindowExtensionNoCache_Bundle.cpp:
2106         (DOMWindowExtensionNoCache):
2107         * TestWebKitAPI/Tests/WebKit2/DocumentStartUserScriptAlertCrash_Bundle.cpp:
2108         (TestWebKitAPI::DocumentStartUserScriptAlertCrashTest::initialize):
2109
2110 2012-07-06  Zan Dobersek  <zandobersek@gmail.com>
2111
2112         [Gtk] Add a configuration option for disabling unstable features in releases
2113         https://bugs.webkit.org/show_bug.cgi?id=87995
2114
2115         Reviewed by Martin Robinson.
2116
2117         Pass the --enable-unstable-features flag when building the Gtk port through build-webkit.
2118
2119         * Scripts/webkitdirs.pm:
2120         (buildAutotoolsProject):
2121
2122 2012-07-06  Jessie Berlin  <jberlin@apple.com>
2123
2124         WKContext should ask for its initialization data when it needs it so the client doesn't have
2125         to keep it up to date.
2126         https://bugs.webkit.org/show_bug.cgi?id=90627
2127
2128         Reviewed by Anders Carlsson.
2129
2130         Add tests and update other WKContextInjectedBundleClients.
2131
2132         * MiniBrowser/mac/AppDelegate.m:
2133         (-[BrowserAppDelegate init]):
2134         Updated for the change to WKContextInjectedBundleClient.
2135         * WebKitTestRunner/TestController.cpp:
2136         (WTR::TestController::initialize):
2137         Ditto.
2138
2139         * TestWebKitAPI/PlatformUtilities.cpp:
2140         (TestWebKitAPI::Util::createInitializationDictionaryForInjectedBundleTest):
2141         Moved the logic to create the initialization dictionary here ...
2142         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
2143         ... from here so that it can be used without automatically using
2144         WKContextSetInitializationUserDataForInjectedBundle.
2145         * TestWebKitAPI/PlatformUtilities.h:
2146
2147         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
2148         Add the new test files.
2149         * TestWebKitAPI/GNUmakefile.am:
2150         Ditto.
2151
2152         * TestWebKitAPI/Tests/WebKit2/GetInjectedBundleInitializationUserDataCallback.cpp: Added.
2153         (TestWebKitAPI::didReceiveMessageFromInjectedBundle):
2154         Check that the message received from the injected bundle matches the user data it was
2155         initialized with.
2156         (TestWebKitAPI::getInjectedBundleInitializationUserData):
2157         Return the user data that the injected bundle should be initialized with.
2158         (TestWebKitAPI::TEST):
2159         Set up WKContextInjectedBundleClient and load a page.
2160         * TestWebKitAPI/Tests/WebKit2/GetInjectedBundleInitializationUserDataCallback_Bundle.cpp: Added.
2161         (TestWebKitAPI::GetInjectedBundleInitializationUserDataCallbackTest::initialize):
2162         Send the initialization user data back up to the UI Process.
2163
2164         * TestWebKitAPI/Tests/WebKit2/InjectedBundleInitializationUserDataCallbackWins.cpp: Added.
2165         (TestWebKitAPI::didReceiveMessageFromInjectedBundle):
2166         Check that the message received from the injected bundle matches the user data it was
2167         initialized with in the callback.
2168         (TestWebKitAPI::getInjectedBundleInitializationUserData):
2169         Return the user data that the injected bundle should be initialized with.
2170         (TestWebKitAPI::TEST):
2171         Set up the context and use WKContextSetInitializationUserDataForInjectedBundle to set the
2172         initialization user data (which should be overridden by the user data returned in
2173         getInjectedBundleInitializationUserData).
2174         * TestWebKitAPI/Tests/WebKit2/InjectedBundleInitializationUserDataCallbackWins_Bundle.cpp: Added.
2175         (TestWebKitAPI::InjectedBundleInitializationUserDataCallbackWinsTest::initialize):
2176         Send the initialization user data back up to the UI Process.
2177
2178 2012-07-06  Dongwoo Im  <dw.im@samsung.com>
2179
2180         [EFL][GTK] jhbuild : Disable pixman demos build depending on GTK+
2181         https://bugs.webkit.org/show_bug.cgi?id=90593
2182
2183         Reviewed by Philippe Normand.
2184
2185         * efl/jhbuild.modules: Give '--enable-gtk=no' option to pixman.
2186         * gtk/jhbuild.modules: Give '--enable-gtk=no' option to pixman.
2187
2188 2012-07-06  Luciano Wolf <luciano.wolf@openbossa.org>
2189
2190         [Qt] DumpRenderTree does not use 'monospace' font when directed
2191         https://bugs.webkit.org/show_bug.cgi?id=85203
2192
2193         The test fonts used for Qt tests were moved to the Liberation font family.
2194
2195         Due to this change we are skipping tons of tests. They will be gradually
2196         unskipped in batches, ASAP.
2197
2198         Also, the now unneeded special font configuration for Qt5 WK1 was removed.
2199
2200         Reviewed by Tor Arne Vestbø.
2201
2202         * DumpRenderTree/qt/QtInitializeTestFonts.cpp:
2203         (WebKit::initializeTestFonts):
2204
2205 2012-07-06  Csaba Osztrogonác  <ossy@webkit.org>
2206
2207         [Qt] Buildfix for newer Qt5
2208         https://bugs.webkit.org/show_bug.cgi?id=90519
2209
2210         Reviewed by Tor Arne Vestbø.
2211
2212         * MiniBrowser/qt/MiniBrowser.pro: We also need to depend on gui-private in the MiniBrowser to get access to these headers.
2213         * MiniBrowser/qt/MiniBrowserApplication.h: Include qpa/qwindowsysteminterface.h instead of deprecated qwindowsysteminterface_qpa.h
2214         * WebKitTestRunner/qt/PlatformWebViewQt.cpp: Include qpa/qwindowsysteminterface.h instead of deprecated qwindowsysteminterface_qpa.h
2215
2216 2012-07-06  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
2217
2218         [Qt] Make use of Qt5 qmake's changed makefile recursion behavior
2219
2220         When not using the -r option, qmake now interleaves qmake and make calls,
2221         so we don't need custom logic for this. There's also an option to supress
2222         the effect of the -r option, which we use to make WebKit.pro the only
2223         project file parsed in a recursive qmake-run.
2224
2225         https://bugs.webkit.org/show_bug.cgi?id=90461
2226
2227         Reviewed by Tor Arne Vestbø.
2228
2229         * qmake/mkspecs/features/default_post.prf:
2230         * qmake/mkspecs/features/functions.prf:
2231
2232 2012-07-06  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
2233
2234         [Qt] Remove custom qmake logic for module creation
2235
2236         Qmake now has the necessary hooks to cleanly override the build locations.
2237
2238         https://bugs.webkit.org/show_bug.cgi?id=90461
2239
2240         Reviewed by Tor Arne Vestbø.
2241
2242         * qmake/mkspecs/features/default_pre.prf:
2243
2244 2012-07-06  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
2245
2246         [Qt] Add top-level .qmake.conf
2247
2248         With Qt5, this makes setting $QMAKEPATH externally unnecessary.
2249
2250         The magic in the perl scripts to set QMAKEPATH is still there, as it doesn't
2251         hurt, and is still required for Qt4.
2252
2253         https://bugs.webkit.org/show_bug.cgi?id=90461
2254
2255         Reviewed by Tor Arne Vestbø.
2256
2257         * qmake/mkspecs/features/default_pre.prf:
2258
2259 2012-07-06  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
2260
2261         [Qt] Unify qtFeatureDefaults code paths
2262
2263         There's no reason to run qmake on features.prf directly anymore, as we
2264         can selectivly run configure tests in the project file now.
2265
2266         https://bugs.webkit.org/show_bug.cgi?id=90461
2267
2268         Reviewed by Tor Arne Vestbø.
2269
2270         * Scripts/webkitdirs.pm:
2271         (qtFeatureDefaults):
2272         * qmake/configure.pri:
2273         * qmake/mkspecs/features/features.prf:
2274
2275 2012-07-06  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
2276
2277         [Qt] Don't let qt_webkit.pri proclaim its own location
2278
2279         This won't work any more with recent Qt5 versions, as the forwarding
2280         pri is created by qt_module_config, which needs MODULE_PRI to be set
2281         up already.
2282
2283         We also need to load build_config, not qt_module.
2284
2285         https://bugs.webkit.org/show_bug.cgi?id=90461
2286
2287         Reviewed by Tor Arne Vestbø.
2288
2289         * qmake/qt_webkit.pri:
2290
2291 2012-07-06  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
2292
2293         [Qt] Remove redundant CONFIG+=module
2294
2295         qt_module_config takes care of that.
2296
2297         In api.pri we are actually testing the flag ourselves, so now we need to
2298         test a related flag qt_module_config sets instead.
2299
2300         Reviewed by Tor Arne Vestbø.
2301
2302         * qmake/mkspecs/features/default_post.prf:
2303
2304 2012-07-06  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
2305
2306         [Qt] Switch to new-style Qt 5 configure tests
2307
2308         Use explicit project file action instead of syncqt magic.
2309
2310         https://bugs.webkit.org/show_bug.cgi?id=90461
2311
2312         Reviewed by Tor Arne Vestbø.
2313
2314         * Scripts/webkitdirs.pm:
2315         (buildQMakeProjects):
2316         * qmake/configure.pri: Copied from Tools/qmake/configure.pro.
2317         * qmake/configure.pro:
2318         * qmake/mkspecs/features/default_pre.prf:
2319         * qmake/mkspecs/features/features.prf:
2320         * qmake/sync.profile: Removed.
2321
2322 2012-07-05  Jason Liu  <jason.liu@torchmobile.com.cn>
2323
2324         Add Jason Liu to committers.py.
2325         https://bugs.webkit.org/show_bug.cgi?id=90654
2326
2327         Unreviewed.
2328
2329         * Scripts/webkitpy/common/config/committers.py:
2330
2331 2012-07-05  Benjamin Poulain  <bpoulain@apple.com>
2332
2333         Double release of resources if the load is canceled in a callback of ResourceLoader::didFinishLoading
2334         https://bugs.webkit.org/show_bug.cgi?id=90431
2335
2336         Reviewed by Anders Carlsson.
2337
2338         Add a Mac API test.
2339
2340         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
2341         * TestWebKitAPI/Tests/mac/CancelLoadFromResourceLoadDelegate.html: Added.
2342         * TestWebKitAPI/Tests/mac/CancelLoadFromResourceLoadDelegate.mm: Added.
2343         (-[CancelLoadFromResourceLoadDelegate webView:resource:didFinishLoadingFromDataSource:]):
2344         (-[CancelLoadFromResourceLoadDelegateFrameLoadDelegate webView:didFinishLoadForFrame:]):
2345         (TestWebKitAPI):
2346         (TestWebKitAPI::TEST):
2347
2348 2012-07-05  Dave Tharp  <dtharp@codeaurora.org>
2349
2350         Adding myself as committer to committers.py
2351
2352         Unreviewed.
2353
2354         * Scripts/webkitpy/common/config/committers.py:
2355
2356 2012-07-05  Alexey Proskuryakov  <ap@apple.com>
2357
2358         [Mac][WK2] Enable HTTPS tests
2359         https://bugs.webkit.org/show_bug.cgi?id=90600
2360
2361         Reviewed by Brady Eidson.
2362
2363         * DumpRenderTree/mac/DumpRenderTree.mm: (dumpRenderTree): Removed a misleading comment.
2364         It explained why we were doing this here instead of a frame load delegate, but that's
2365         just history, not an explanation of why we need this code.
2366
2367         * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm: (WTR::InjectedBundle::platformInitialize):
2368         Ignore certificate errors for localhost and 127.0.0.1. Since this is only needed
2369         for tests, it seems appropriate to have platform specific solutions in every WTR,
2370         and not add a WebKit2 API.
2371
2372 2012-07-05  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2373
2374         [Qt] Remove rpath-hacks in qt_webkit.pri
2375
2376         Having it in qt_webkit.pri meant that all client apps would inherit the rpath,
2377         even if they were not using QtWebKit.
2378
2379         This is a regression for people who build client apps against a non-installed
2380         QtWebKit, as the client app will be missing a rpath to the standalone WebKit
2381         build dir. This can be solved by LD_LIBRARY_PATH or custom QMAKE_RPATHDIR
2382         logic, until the real issue is fixed in Qt.
2383
2384         Reviewed by Csaba Osztrogonác.
2385
2386         * qmake/mkspecs/features/rpath.prf:
2387         * qmake/qt_webkit.pri:
2388
2389 2012-07-05  Dongwoo Im  <dw.im@samsung.com>
2390
2391         [EFL] Enable the CUSTOM_SCHEME_HANDLER feature as default.
2392         https://bugs.webkit.org/show_bug.cgi?id=88608
2393
2394         Reviewed by Hajime Morita.
2395
2396         * Scripts/webkitperl/FeatureList.pm: Enable the CUSTOM_SCHEME_HANDLER feature as default.
2397
2398 2012-07-05  Ryuan Choi  <ryuan.choi@samsung.com>
2399
2400         [Wk2][EFL] EFL needs a WebKitTestRunner
2401         https://bugs.webkit.org/show_bug.cgi?id=87659
2402
2403         Reviewed by Kenneth Rohde Christiansen.
2404
2405         Add an implementation of WebKitTestRunner for Efl port.
2406
2407         * CMakeLists.txt: Includes WebKitTestRunner build script.
2408         * Scripts/build-webkittestrunner: Added knowledge of WebKitTestRunner/Efl.
2409         * Scripts/webkitpy/layout_tests/port/efl.py: Ditto.
2410         (EflPort.setup_environ_for_server):
2411         * WebKitTestRunner/CMakeLists.txt: Added.
2412         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
2413         * WebKitTestRunner/InjectedBundle/efl/ActivateFontsEfl.cpp: Added.
2414         * WebKitTestRunner/InjectedBundle/efl/InjectedBundleEfl.cpp: Added.
2415         * WebKitTestRunner/InjectedBundle/efl/LayoutTestControllerEfl.cpp: Added.
2416         * WebKitTestRunner/PlatformEfl.cmake: Added.
2417         * WebKitTestRunner/PlatformWebView.h:
2418         * WebKitTestRunner/TestController.cpp:
2419         * WebKitTestRunner/config.h: Added knowledge of WebKitTestRunner/Efl.
2420         * WebKitTestRunner/efl/PlatformWebViewEfl.cpp: Added.
2421         * WebKitTestRunner/efl/TestControllerEfl.cpp: Added.
2422         * WebKitTestRunner/efl/main.cpp: Added.
2423         (main):
2424
2425 2012-07-05  Christophe Dumez  <christophe.dumez@intel.com>
2426
2427         [WK2][EFL] Ewk_View needs API to load HTML data
2428         https://bugs.webkit.org/show_bug.cgi?id=90540
2429
2430         Reviewed by Kenneth Rohde Christiansen.
2431
2432         Update EFL MiniBrowser to catch the "load,error" signal
2433         on the view and display an error page.
2434
2435         * MiniBrowser/efl/main.c:
2436         (on_error):
2437         (browserCreate):
2438
2439 2012-07-05  Sergio Villar Senin  <svillar@igalia.com>
2440
2441         Hidden dirs are not copied when creating the built product archive
2442         https://bugs.webkit.org/show_bug.cgi?id=90559
2443
2444         Reviewed by Carlos Garcia Campos.
2445
2446         Hidden dirs are not copied inside the built product archive
2447         because the recursive copy command was ignoring them. We need those
2448         hidden directories for the GTK WebKit2 testing bot to work fine.
2449
2450         * BuildSlaveSupport/built-product-archive:
2451         (archiveBuiltProduct): replaced "*" by "." to include hidden
2452         dirs/files.
2453
2454 2012-07-04  Yoshifumi Inoue  <yosin@chromium.org>
2455
2456         Unreviewed Chromium gardening, Disable always failed python tests on Windows.
2457         https://bugs.webkit.org/show_bug.cgi?id=90587
2458
2459         PortTest.qt_linux and mac_lion are always failed on Chromium Windows at
2460         creating child process.
2461
2462         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2463         (PortTest.disabled_test_qt_linux):
2464         (PortTest.disabled_test_mac_lion):
2465
2466 2012-07-04  Yoshifumi Inoue  <yosin@chromium.org>
2467
2468         [Tools] webkit_unittest.py got assertion
2469         https://bugs.webkit.org/show_bug.cgi?id=90579
2470
2471         Reviewed by Hajime Morita.
2472
2473         This patch updates expectation of assertion in test_skipped_directories_for_symbols().
2474         r121874 added new (exclude) directory (inspector/styles/variables), however it didn't
2475         update one of two assertions.
2476
2477         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
2478
2479 2012-07-04  Balazs Ankes  <bank@inf.u-szeged.hu>
2480
2481         webkit-patch should add reviewer if "Reviewed by NOBODY ..." is missing
2482         https://bugs.webkit.org/show_bug.cgi?id=67935
2483
2484         Reviewed by Ryosuke Niwa.
2485
2486         * Scripts/webkitpy/common/checkout/changelog.py:
2487         (ChangeLog.set_reviewer):
2488         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
2489         (test_set_reviewer):
2490         (test_set_short_description_and_bug_url):
2491
2492 2012-07-04  Sheriff Bot  <webkit.review.bot@gmail.com>
2493
2494         Unreviewed, rolling out r121862.
2495         http://trac.webkit.org/changeset/121862
2496         https://bugs.webkit.org/show_bug.cgi?id=90569
2497
2498         It broke built-product-archive for GTK, Qt and EFL (Requested
2499         by svillar on #webkit).
2500
2501         * BuildSlaveSupport/built-product-archive:
2502         (archiveBuiltProduct):
2503
2504 2012-07-04  Sergio Villar Senin  <svillar@igalia.com>
2505
2506         Hidden dirs are not copied when creating the built product archive
2507         https://bugs.webkit.org/show_bug.cgi?id=90559
2508
2509         Reviewed by Carlos Garcia Campos.
2510
2511         Hidden dirs are not copied inside the built product archive because
2512         the recursive copy command was ignoring them. We need those hidden
2513         directories for the GTK WebKit2 testing bot to work fine.
2514
2515         * BuildSlaveSupport/built-product-archive:
2516         (archiveBuiltProduct): removed the trailing "*" from the copy
2517         command.
2518
2519 2012-07-04  John Mellor  <johnme@chromium.org>
2520
2521         Text Autosizing: Add compile flag and runtime setting
2522         https://bugs.webkit.org/show_bug.cgi?id=87394
2523
2524         This patch renames Font Boosting to Text Autosizing.
2525
2526         Reviewed by Adam Barth.
2527
2528         * Scripts/webkitperl/FeatureList.pm:
2529         * qmake/mkspecs/features/features.pri:
2530
2531 2012-07-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2532
2533         [Qt] Replace deprecated IN_PWD with PWD
2534
2535         Reviewed by Csaba Osztrogonác.
2536
2537         * qmake/mkspecs/features/default_pre.prf:
2538
2539 2012-07-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2540
2541         [Qt] Don't try to build WTR twice
2542
2543         Reviewed by Csaba Osztrogonác.
2544
2545         * Tools.pro:
2546
2547 2012-07-04  Sheriff Bot  <webkit.review.bot@gmail.com>
2548
2549         Unreviewed, rolling out r121821.
2550         http://trac.webkit.org/changeset/121821
2551         https://bugs.webkit.org/show_bug.cgi?id=90551
2552
2553         This patch did not receive a high-quality review and has a
2554         number of errors (Requested by abarth on #webkit).
2555
2556         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
2557         (Build.results_url):
2558         * Scripts/webkitpy/common/net/buildbot/chromiumbuildbot.py:
2559         (ChromiumBuilder.accumulated_results_url):
2560         * Scripts/webkitpy/layout_tests/port/builders.py:
2561         * Scripts/webkitpy/tool/commands/rebaseline.py:
2562         (AbstractParallelRebaselineCommand.__init__):
2563         (Rebaseline):
2564         (Rebaseline._builder_to_pull_from):
2565         (Rebaseline._tests_to_update):
2566         (Rebaseline.execute):
2567         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
2568         (test_rebaseline.mock_builder_to_pull_from):
2569         (test_rebaseline):
2570
2571 2012-07-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2572
2573         [Qt] Get rid of un-needed QT += declarative for Qt 5
2574
2575         The declarative module has been renamed to quick1 in Qt 5, and the
2576         engine-only module for Qt 5 is named 'qml'. For many of the instances
2577         we could just remove 'declarative', since the project file was only
2578         used for Qt5/WebKit2 builds. In the other cases the module was wrapped
2579         in a haveQt(4) scope.
2580
2581         Reviewed by Csaba Osztrogonác.
2582
2583         * MiniBrowser/qt/MiniBrowser.pro:
2584         * WebKitTestRunner/InjectedBundle/Target.pri:
2585         * WebKitTestRunner/Target.pri:
2586
2587 2012-07-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2588
2589         [Qt] Remove a few un-needed load(features) after r121777
2590
2591         The features are computed by configure.pro and cached in .qmake.cache.
2592
2593         Reviewed by Jocelyn Turcotte..
2594
2595         * MiniBrowser/qt/raw/Target.pri:
2596
2597 2012-07-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2598
2599         [Qt] Use haveQtModule() in project files instead of manual checks
2600
2601         Reviewed by Jocelyn Turcotte.
2602
2603         * qmake/mkspecs/features/default_pre.prf:
2604         * qmake/mkspecs/features/features.prf:
2605
2606 2012-07-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
2607
2608         [Qt] Fix haveQtModule() check
2609
2610         Without {} brackets we were pulling out the value of the module.name variable.
2611
2612         Reviewed by Jocelyn Turcotte.
2613
2614         * qmake/mkspecs/features/functions.prf:
2615
2616 2012-07-04  Sergio Villar Senin  <svillar@igalia.com>
2617
2618         [WK2] [GTK] WK2 testing bot unable to find the InjectedBundle library
2619         https://bugs.webkit.org/show_bug.cgi?id=90541
2620
2621         Reviewed by Carlos Garcia Campos.
2622
2623         We must add the contents of WebKitBuild/*/Libraries/ directory to
2624         the file to be uploaded to b.w.o to successfully run WebKit2 tests
2625         in the WebKit2 testing bot.
2626
2627         * BuildSlaveSupport/built-product-archive:
2628         (archiveBuiltProduct): appended Libraries to the list of
2629         neededDirectories.
2630
2631 2012-07-04  Sergio Villar Senin  <svillar@igalia.com>
2632
2633         [GTK] WebKit2 testing bot unable to find the build directory
2634         https://bugs.webkit.org/show_bug.cgi?id=90523
2635
2636         Reviewed by Carlos Garcia Campos.
2637
2638         The presence of GNUmakefile was used to determine whether a given
2639         directory was a valid build directory or not. That's not correct for
2640         testing bots as we export only the minimum required stuff to run the
2641         tests.
2642
2643         * gtk/common.py:
2644         (get_build_path.is_valid_build_directory): added an extra check.:
2645
2646 2012-07-04  Joe Thomas  <joethomas@motorola.com>
2647
2648         Adding myself to committers list.
2649
2650         Unreviewed.
2651
2652         * Scripts/webkitpy/common/config/committers.py:
2653
2654 2012-07-03  Christophe Dumez  <christophe.dumez@intel.com>
2655
2656         [WK2][EFL] Ewk_View should report the load progress
2657         https://bugs.webkit.org/show_bug.cgi?id=90457
2658
2659         Reviewed by Kenneth Rohde Christiansen.
2660
2661         Update EFL MiniBrowser so that it listens for the
2662         "load,progress" on the Ewk_View and updates its
2663         window title accordingly.
2664
2665         * MiniBrowser/efl/main.c:
2666         (title_set):
2667         (on_title_changed):
2668         (on_progress):
2669         (browserCreate):
2670
2671 2012-07-03  Dirk Pranke  <dpranke@chromium.org>
2672
2673         disable failing mock-chromium-mac python unit tests
2674
2675         Unreviewed, build fix.
2676
2677         Disable these tests until I can figure out why they're failing
2678         on the bots.
2679
2680         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2681         (PortTest.disabled_test_chromium_mac_lion):
2682         (PortTest.disabled_test_chromium_mac_lion_in_test_shell_mode):
2683
2684 2012-07-03  Ian Vollick  <vollick@chromium.org>
2685
2686         Add vollick to the list of committers
2687         https://bugs.webkit.org/show_bug.cgi?id=90512
2688
2689         Reviewed by Adrienne Walker.
2690
2691         * Scripts/webkitpy/common/config/committers.py:
2692
2693 2012-07-03  Ojan Vafai  <ojan@chromium.org>
2694
2695         Improve webkit-patch rebaseline to work for more cases
2696         https://bugs.webkit.org/show_bug.cgi?id=90504
2697
2698         Reviewed by Dirk Pranke.
2699
2700         -Makes it work for the build.chromium.org bots.
2701         -Lets you rebaseline all builders instead of just one.
2702         -Lets you pass in the builder or test to rebaseline.
2703
2704         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
2705         (Build.results_url):
2706         Make this work for build.chromium.org builders as well.
2707
2708         * Scripts/webkitpy/common/net/buildbot/chromiumbuildbot.py:
2709         (ChromiumBuilder):
2710         (ChromiumBuilder.latest_cached_build):
2711         ChromiumBuilder doesn't support large parts of the Builder interface.
2712         This provides the bare minimum for this patch to work. Eventually,
2713         we should create a single interface that can be supported for both
2714         sets of buildbots.
2715
2716         * Scripts/webkitpy/layout_tests/port/builders.py:
2717         (builder_path_from_name):
2718         (all_builder_names):
2719         memoizing here is incorrect because the test override _exact_matches,
2720         so these can return different values. In either case, I'm pretty sure these
2721         are not remotely hot codepaths.
2722
2723         * Scripts/webkitpy/tool/commands/rebaseline.py:
2724         (AbstractParallelRebaselineCommand.__init__):
2725         (Rebaseline):
2726         (Rebaseline.__init__):
2727         (Rebaseline._builders_to_pull_from):
2728         (Rebaseline._builder_with_name):
2729         (Rebaseline._tests_to_update):
2730         (Rebaseline.execute):
2731         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
2732         (test_rebaseline.mock_builders_to_pull_from):
2733         (test_rebaseline):
2734         (test_rebaseline_command_line_flags):
2735         (test_rebaseline_multiple_builders):
2736         (test_rebaseline_multiple_builders.mock_builders_to_pull_from):
2737         (test_rebaseline_multiple_builders.mock_tests_to_update):
2738
2739 2012-07-03  Christophe Dumez  <christophe.dumez@intel.com>
2740
2741         [EFL] Enable CSS variables support at compile time
2742         https://bugs.webkit.org/show_bug.cgi?id=90448
2743
2744         Reviewed by Kenneth Rohde Christiansen.
2745
2746         Turn on CSS_VARIABLES flag by default on EFL port.
2747
2748         * Scripts/webkitperl/FeatureList.pm:
2749
2750 2012-07-03  Dirk Pranke  <dpranke@chromium.org>
2751
2752         nrwt: clean up exception handling and make sure we log some more failures
2753         https://bugs.webkit.org/show_bug.cgi?id=90503
2754
2755         Reviewed by Ojan Vafai.
2756
2757         There were several places where exceptions weren't getting
2758         logged, most notably if you passed a bad value to --platform.
2759         This change tests that and cleans things up a bit; more cleanup
2760         will be possible when we rework the manager_worker_broker code.
2761
2762         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
2763         (_BrokerConnection.raise_exception):
2764         (_InlineWorkerConnection.raise_exception):
2765         * Scripts/webkitpy/layout_tests/controllers/worker.py:
2766         (Worker.run):
2767         (Worker.kill_driver):
2768         * Scripts/webkitpy/layout_tests/port/factory.py:
2769         (PortFactory.get):
2770         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2771         (run):
2772         (main):
2773         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2774         (MainTest.test_unsupported_platfrom):
2775
2776 2012-07-03  Dirk Pranke  <dpranke@chromium.org>
2777
2778         nrwt: fix mock port
2779         https://bugs.webkit.org/show_bug.cgi?id=90500
2780
2781         Reviewed by Ojan Vafai.
2782
2783         The MockDRT code was never updated when we switched the chromium
2784         ports to using "drt mode" by default. This change updates that
2785         code, fixes a typo in the chromium port that went undetected
2786         (default_test_timeout_ms -> default_timeout_ms), and adds tests
2787         that actually exercise some of the mock ports. These tests are
2788         useful in that they will exercise the port-specific code in an
2789         end-to-end-manner, but they are a bit slow for some reason (>1s
2790         each) that I need to look into.
2791
2792         * Scripts/webkitpy/layout_tests/port/chromium.py:
2793         (ChromiumDriver.stop):
2794         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
2795         (ChromiumAndroidPort.default_timeout_ms):
2796         * Scripts/webkitpy/layout_tests/port/chromium_android_unittest.py:
2797         (TestChromiumAndroidPort.test_default_timeout_ms):
2798         * Scripts/webkitpy/layout_tests/port/mock_drt.py:
2799         (main):
2800         (parse_options):
2801         (MockTestShell):
2802         (MockTestShell.output_for_test):
2803         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
2804         (MockDRTTest.assertTest):
2805         (MockTestShellTest):
2806         (MockTestShellTest.make_drt):
2807         (MockTestShellTest.test_test_shell_parse_options):
2808         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2809         (PortTest):
2810         (PortTest.assert_mock_port_works):
2811         (PortTest.test_chromium_mac_lion):
2812         (PortTest.test_chromium_mac_lion_in_test_shell_mode):
2813         (PortTest.test_qt_linux):
2814         (PortTest.test_mac_lion):
2815
2816 2012-07-03  Dirk Pranke  <dpranke@chromium.org>
2817
2818         nrwt: make the worker class stand alone with a cleaner interface
2819         https://bugs.webkit.org/show_bug.cgi?id=90409
2820
2821         Reviewed by Ojan Vafai.
2822
2823         Currently the Worker class derives from AbstractWorker, which is
2824         kind of crufty and awkward; it would be better if we did not
2825         rely on shared state.
2826
2827         This change changes that so that Worker derives from object, and
2828         exposes the following interface:
2829           __init__() - called in the manager process
2830           safe_init() - called in the worker process to initialize
2831             unpicklable state
2832           handle() - a single routine to handle all messages
2833           cleanup() - called so the worker can clean up
2834
2835         Also, all of the "administrative" messages that are handled by
2836         the worker (notification of start/stop/etc.) move into
2837         manager_worker_broker - this reduces worker.py to just handling
2838         the mechanics of actually running each test.
2839
2840         For the moment, we do this by creating Yet Another wrapper/proxy
2841         class in manager_worker_broker, but this will get simpler
2842         shortly when the rest of m_w_b is cleaned up.
2843
2844         With this change worker is now in its new form but there will be
2845         a follow-on change that cleans up some names and other minor
2846         things.
2847
2848         This change is again mostly just moving things around and should
2849         be covered by the (updated) existing tests.
2850
2851         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
2852         (get):
2853         (AbstractWorker.__init__):
2854         (AbstractWorker.run):
2855         (AbstractWorker):
2856         (AbstractWorker.handle_stop):
2857         (AbstractWorker.handle_test_list):
2858         (AbstractWorker.yield_to_broker):
2859         (AbstractWorker.post_message):
2860         (_WorkerConnection.__init__):
2861         (_Process.run):
2862         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py:
2863         (_TestWorker):
2864         (_TestWorker.__init__):
2865         (_TestWorker.name):
2866         (_TestWorker.cleanup):
2867         (_TestWorker.handle):
2868         (_TestWorker.safe_init):
2869         (_TestWorker.stop):
2870         (_TestsMixin.handle_finished_test):
2871         (_TestsMixin.setUp):
2872         (_TestsMixin.test_cancel):
2873         (_TestsMixin.test_done):
2874         * Scripts/webkitpy/layout_tests/controllers/worker.py:
2875         (Worker):
2876         (Worker.__init__):
2877         (Worker.safe_init):
2878         (Worker.handle):
2879
2880 2012-07-03  Dirk Pranke  <dpranke@chromium.org>
2881
2882         nrwt: moving child process logging code into manager_worker_broker
2883         https://bugs.webkit.org/show_bug.cgi?id=90408
2884
2885         Reviewed by Ojan Vafai.
2886
2887         Users of manager_worker_broker should not have to be aware of
2888         whether they're in the same process or different processes and
2889         configure logging themselves; mwb should hide this complexity.
2890         We can't quite do this completely/correctly yet, since the
2891         manager expects to get a list of messages to log, but this
2892         change fixes the worker side of it, at least.
2893
2894         This is just moving code around, there is no new functionality
2895         and this should be covered by existing tests.
2896
2897         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
2898         (AbstractWorker.__init__):
2899         (_WorkerConnection.__init__):
2900         (_WorkerConnection.post_message):
2901         (_WorkerConnection):
2902         (_WorkerConnection.set_up_logging):
2903         (_WorkerConnection.clean_up_logging):
2904         (_InlineWorkerConnection.run):
2905         (_MultiProcessWorkerConnection.run):
2906         (_WorkerLogHandler):
2907         (_WorkerLogHandler.__init__):
2908         (_WorkerLogHandler.emit):
2909         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py:
2910         (_TestWorker.run):
2911         (_TestsMixin.handle_done):
2912         * Scripts/webkitpy/layout_tests/controllers/worker.py:
2913         (Worker.__init__):
2914         (Worker.run):
2915         (Worker._run_test):
2916         (Worker.cleanup):
2917         (Worker.run_single_test):
2918
2919 2012-07-03  Tony Chang  <tony@chromium.org>
2920
2921         [chromium] Don't archive build files generated by VS2010
2922         https://bugs.webkit.org/show_bug.cgi?id=90493
2923
2924         Reviewed by Ojan Vafai.
2925
2926         In particular, the pch (pre compiled header) files are gigantic, making
2927         release.zip almost a 1gb in size.
2928
2929         * BuildSlaveSupport/built-product-archive:
2930         (archiveBuiltProduct):
2931
2932 2012-07-03  Tony Chang  <tony@chromium.org>
2933
2934         [chromium] Unreviewed, update .gitignore to handle VS2010 files.
2935
2936         * .gitignore:
2937         * DumpRenderTree/DumpRenderTree.gyp/.gitignore:
2938         * TestWebKitAPI/TestWebKitAPI.gyp/.gitignore:
2939
2940 2012-07-03  Ojan Vafai  <ojan@chromium.org>
2941
2942         Have webkit-patch rebaseline use rebaseline-test-internal
2943         https://bugs.webkit.org/show_bug.cgi?id=90491
2944
2945         Reviewed by Dirk Pranke.
2946
2947         This lets it handle new files, reduces duplicate code and lays the
2948         groundwork for making rebaseline have a richer interface.
2949
2950         * Scripts/webkitpy/common/net/buildbot/buildbot_mock.py:
2951         (MockBuild):
2952         (MockBuild.__init__):
2953         (MockBuilder.build):
2954         * Scripts/webkitpy/tool/commands/rebaseline.py:
2955         (AbstractParallelRebaselineCommand._files_to_add):
2956         (Rebaseline):
2957         (Rebaseline._builder_to_pull_from):
2958         (Rebaseline._tests_to_update):
2959         (Rebaseline.execute):
2960         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
2961         (test_overrides_are_included_correctly):
2962         (test_rebaseline):
2963         (test_rebaseline.mock_builder_to_pull_from):
2964         (test_rebaseline.mock_tests_to_update):
2965
2966 2012-07-03  Ojan Vafai  <ojan@chromium.org>
2967
2968         Rename rebaseline-test to rebaseline-test-internal
2969         https://bugs.webkit.org/show_bug.cgi?id=90485
2970
2971         Reviewed by Adam Barth.
2972
2973         It's now only used by other rebaseline commands. It's still useful to leave it
2974         as it's own command to aid in debugging when something goes wrong. In a followup
2975         patch, I'll make webkit-patch rebaseline cover any use-cases that rebaseline-test
2976         might have covered.
2977
2978         We no longer need the --print-scm-changes option since the only caller always passes
2979         that option in.
2980
2981         Also, make all the arguments command-line flags instead. Simplifies the code
2982         a bit in my opinion.
2983
2984         * Scripts/webkitpy/tool/commands/rebaseline.py:
2985         (RebaselineTest):
2986         (RebaselineTest.__init__):
2987         (RebaselineTest.execute):
2988         (AbstractParallelRebaselineCommand._rebaseline_commands):
2989         (RebaselineExpectations.execute):
2990         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
2991         (test_rebaseline_all):
2992
2993 2012-07-03  Balazs Kelemen  <kbalazs@webkit.org>
2994
2995         [Qt][WTR] Get rid of using DumpRenderTreeSupportQt
2996         https://bugs.webkit.org/show_bug.cgi?id=90262
2997
2998         Unreviewed build fix.
2999
3000         Reverting r121550 manually because it caused build
3001         break on ARM and Mac. It turned out that the Qt build
3002         system is too broken for this change now, first we have
3003         to address two issues:
3004             * symbol hiding was forbidden accidentally on x86 Linux
3005             * Qt lacks a separate library for test support
3006
3007         * Tools.pro:
3008         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3009         (WTR::InjectedBundlePage::resetAfterTest):
3010         (WTR::InjectedBundlePage::didClearWindowForFrame):
3011         * WebKitTestRunner/InjectedBundle/qt/ActivateFontsQt.cpp:
3012
3013 2012-07-03  Tony Chang  <tony@chromium.org>
3014
3015         [chromium] Handle trailing backslash on %VSINSTALLDIR%
3016         https://bugs.webkit.org/show_bug.cgi?id=90410
3017
3018         Reviewed by Ojan Vafai.
3019
3020         * Scripts/webkitdirs.pm:
3021         (buildChromiumVisualStudioProject): Handle VSINSTALLDIR ending in a backslash, the backslash escapes the closing double quote.
3022
3023 2012-07-02  Ojan Vafai  <ojan@chromium.org>
3024
3025         Make the skia_test_expectations.txt file optional.
3026         https://bugs.webkit.org/show_bug.cgi?id=90400
3027
3028         Reviewed by Dirk Pranke.
3029
3030         It used to be optional. This regressed at some point. It's important that it be
3031         optional so that webkit-patch commands work in a pure-webkit checkout for chromium bots.
3032         Specifically, this was breaking webkit-patch rebaseline-test when it would go to update
3033         TestExpectations.
3034
3035         * Scripts/webkitpy/layout_tests/port/chromium.py:
3036         (ChromiumPort.expectations_files):
3037         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
3038         (ChromiumDriverTest.test_expectations_dict):
3039
3040 2012-07-03  Raphael Kubo da Costa  <rakuco@webkit.org>
3041
3042         [jhbuild][EFL] Bump libffi dependency.
3043         https://bugs.webkit.org/show_bug.cgi?id=85373
3044
3045         Reviewed by Gustavo Noronha Silva.
3046
3047         Update libffi to 3.0.11, which fixes the build on some platforms
3048         where 3.0.10 had problems.
3049
3050         * efl/jhbuild.modules:
3051
3052 2012-07-03  George Staikos  <staikos@webkit.org>
3053
3054         [BlackBerry] Enable microdata support for BlackBerry.
3055         https://bugs.webkit.org/show_bug.cgi?id=90429
3056
3057         Reviewed by Rob Buis.
3058
3059         * Scripts/webkitperl/FeatureList.pm:
3060
3061 2012-07-03  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3062
3063         [Qt] Make use of .qmake.cache for caching features
3064
3065         Instead of loading() features from the files that need them (and re-running
3066         a bunch of checks), we now run feature detection as part of configure.pro,
3067         and have build-webkit write the computed feature-defines and CONFIG to
3068         .qmake.cache, which is then loaded by qmake _before_ even defaults_pre
3069         when building WebKit.pro.
3070
3071         At some point we'll be able to selectivly prevent running of config tests
3072         in configure.pro, which means we don't need a separate code-path for
3073         the build-webkit --help case.
3074
3075         We should also move the code in build-webkit that now uses .webkit.config
3076         to detect clean builds, to use .qmake.cache, since we now store the same
3077         thing there.
3078
3079         Original patch by Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
3080
3081         Reviewed by Tor Arne Vestbø.
3082
3083         * DumpRenderTree/qt/DumpRenderTree.pro:
3084         * QtTestBrowser/QtTestBrowser.pro:
3085         * Scripts/webkitdirs.pm:
3086         (qtFeatureDefaults):
3087         (buildQMakeProjects):
3088         * Tools.pro:
3089         * WebKitTestRunner/InjectedBundle/DerivedSources.pri:
3090         * WebKitTestRunner/InjectedBundle/Target.pri:
3091         * WebKitTestRunner/Target.pri:
3092         * qmake/.qmake.conf: Added.
3093         * qmake/configure.pro:
3094         * qmake/mkspecs/features/default_post.prf:
3095         * qmake/mkspecs/features/default_pre.prf:
3096         * qmake/mkspecs/features/features.prf:
3097
3098 2012-07-03  Nico Weber  <thakis@chromium.org>
3099
3100         -Wunused-private-field cleanup, linux edition
3101         https://bugs.webkit.org/show_bug.cgi?id=90463
3102
3103         Reviewed by Stephen White.
3104
3105         * DumpRenderTree/chromium/DRTDevToolsAgent.h:
3106         * DumpRenderTree/chromium/GamepadController.h:
3107         (GamepadController):
3108         * DumpRenderTree/chromium/NotificationPresenter.h:
3109         (NotificationPresenter::NotificationPresenter):
3110         (NotificationPresenter):
3111
3112 2012-07-03  Sheriff Bot  <webkit.review.bot@gmail.com>
3113
3114         Unreviewed, rolling out r121766.
3115         http://trac.webkit.org/changeset/121766
3116         https://bugs.webkit.org/show_bug.cgi?id=90465
3117
3118         It caused flakey build errors on the bots (Requested by Ossy
3119         on #webkit).
3120
3121         * DumpRenderTree/qt/DumpRenderTree.pro:
3122         * QtTestBrowser/QtTestBrowser.pro:
3123         * Scripts/webkitdirs.pm:
3124         (qtFeatureDefaults):
3125         (buildQMakeProjects):
3126         * Tools.pro:
3127         * WebKitTestRunner/InjectedBundle/DerivedSources.pri:
3128         * WebKitTestRunner/InjectedBundle/Target.pri:
3129         * WebKitTestRunner/Target.pri:
3130         * qmake/.qmake.conf: Removed.
3131         * qmake/configure.pro:
3132         * qmake/mkspecs/features/default_post.prf:
3133         * qmake/mkspecs/features/default_pre.prf:
3134         * qmake/mkspecs/features/features.prf:
3135
3136 2012-07-03  George Staikos  <staikos@webkit.org>
3137
3138         [BlackBerry] Enable Custom Scheme Handlers for BlackBerry.
3139         https://bugs.webkit.org/show_bug.cgi?id=90422
3140
3141         Reviewed by Rob Buis.
3142
3143         * Scripts/webkitperl/FeatureList.pm:
3144
3145 2012-07-03  George Staikos  <staikos@webkit.org>
3146
3147         [BlackBerry] Enable RegisterProtocolHandler for BlackBerry.
3148         https://bugs.webkit.org/show_bug.cgi?id=90422
3149
3150         Reviewed by Rob Buis.
3151
3152         * Scripts/webkitperl/FeatureList.pm:
3153
3154 2012-07-03  Thiago Marcos P. Santos  <thiago.santos@intel.com>
3155
3156         [EFL] Add run API tests as a step on the build bots
3157         https://bugs.webkit.org/show_bug.cgi?id=87252
3158
3159         Reviewed by Daniel Bates.
3160
3161         Run EFL API test suite on the build bots.
3162
3163         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3164         (RunEflAPITests):
3165         (TestFactory.__init__):
3166         (BuildAndTestFactory.__init__):
3167         * Scripts/run-efl-tests: Added.
3168
3169 2012-07-03  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3170
3171         [Qt] Make use of .qmake.cache for caching features
3172
3173         Instead of loading() features from the files that need them (and re-running
3174         a bunch of checks), we now run feature detection as part of configure.pro,
3175         and have build-webkit write the computed feature-defines and CONFIG to
3176         .qmake.cache, which is then loaded by qmake _before_ even defaults_pre
3177         when building WebKit.pro.
3178
3179         At some point we'll be able to selectivly prevent running of config tests
3180         in configure.pro, which means we don't need a separate code-path for
3181         the build-webkit --help case.
3182
3183         We should also move the code in build-webkit that now uses .webkit.config
3184         to detect clean builds, to use .qmake.cache, since we now store the same
3185         thing there.
3186
3187         Original patch by Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
3188
3189         Reviewed by Tor Arne Vestbø.
3190
3191         * DumpRenderTree/qt/DumpRenderTree.pro:
3192         * QtTestBrowser/QtTestBrowser.pro:
3193         * Scripts/webkitdirs.pm:
3194         (qtFeatureDefaults):
3195         (buildQMakeProjects):
3196         * Tools.pro:
3197         * WebKitTestRunner/InjectedBundle/DerivedSources.pri:
3198         * WebKitTestRunner/InjectedBundle/Target.pri:
3199         * WebKitTestRunner/Target.pri:
3200         * qmake/.qmake.conf: Added.
3201         * qmake/configure.pro:
3202         * qmake/mkspecs/features/default_post.prf:
3203         * qmake/mkspecs/features/default_pre.prf:
3204         * qmake/mkspecs/features/features.prf:
3205
3206 2012-07-03  Jocelyn Turcotte  <jocelyn.turcotte@nokia.com>  Joel Dillon <joel.dillon@codethink.co.uk>
3207
3208         [Qt][Win] Fix broken QtWebKit5.lib linking
3209         https://bugs.webkit.org/show_bug.cgi?id=88321
3210
3211         Reviewed by Kenneth Rohde Christiansen.
3212
3213         On windows the import/export definition is part of the symbol's signature.
3214         Define STATICALLY_LINKED_WITH_$$library for each dependend module
3215         that is being linked statically to make sure that they can be linked together
3216         later on.
3217
3218         Also do not compile Assertions.cpp in DumpRenderTree anymore since all the
3219         used symbols are exported and it would cause a duplicate symbols error.
3220
3221         * DumpRenderTree/qt/DumpRenderTree.pro:
3222         * qmake/mkspecs/features/default_post.prf:
3223
3224 2012-07-03  Szilard Ledan  <szledan@inf.u-szeged.hu>
3225
3226         master.cfg cleanup, pass BuildStep instances instead of BuildStep subclasses
3227         https://bugs.webkit.org/show_bug.cgi?id=89001
3228
3229         Reviewed by Csaba Osztrogonác.
3230
3231         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3232         (UploadBuiltProduct.__init__):
3233         (UploadTestResults.__init__):
3234         (ExtractTestResults.__init__):
3235         (Factory.__init__):
3236         (BuildFactory.__init__):
3237         (TestFactory.__init__):
3238         (BuildAndTestFactory.__init__):
3239         (BuildAndPerfTestFactory.__init__):
3240         (BuildAndPerfTestWebKit2Factory.__init__):
3241         (DownloadAndPerfTestFactory.__init__):
3242
3243 2012-07-03  Ádám Kallai  <kadam@inf.u-szeged.hu>
3244
3245         Update unit test after r121724.
3246         https://bugs.webkit.org/show_bug.cgi?id=90444
3247
3248         Reviewed by Csaba Osztrogonác.
3249
3250         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py:
3251         (GardeningServerTest.test_rebaseline_new_port):
3252
3253 2012-07-03  Sheriff Bot  <webkit.review.bot@gmail.com>
3254
3255         Unreviewed, rolling out r121749.
3256         http://trac.webkit.org/changeset/121749
3257         https://bugs.webkit.org/show_bug.cgi?id=90441
3258
3259         Tests crash on the paralel NRWT Qt bot (Requested by Ossy on
3260         #webkit).
3261
3262         * Scripts/webkitpy/layout_tests/port/qt.py:
3263         (QtPort._build_driver):
3264         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py:
3265         (XvfbDriver._start):
3266         (XvfbDriver._start.x_filter):
3267         (XvfbDriver.stop):
3268
3269 2012-07-03  Kristóf Kosztyó  <kkristof@inf.u-szeged.hu>
3270
3271         [NRWT] XvfbDriver should choose the next free display
3272         https://bugs.webkit.org/show_bug.cgi?id=88414
3273
3274         Reviewed by Dirk Pranke.
3275
3276         * Scripts/webkitpy/layout_tests/port/qt.py:
3277         (QtPort._driver_class):
3278         * Scripts/webkitpy/layout_tests/port/xvfbdriver.py:
3279         (XvfbDriver.__init__):
3280         (XvfbDriver._start):
3281         (XvfbDriver._start.next_free_id):
3282         (XvfbDriver.stop):
3283
3284 2012-07-03  Christophe Dumez  <christophe.dumez@intel.com>
3285
3286         [EFL] Enable MICRODATA support
3287         https://bugs.webkit.org/show_bug.cgi?id=90377
3288
3289         Reviewed by Kenneth Rohde Christiansen.
3290
3291         Turn on MICRODATA flag for EFL port by default.
3292
3293         * Scripts/webkitperl/FeatureList.pm:
3294
3295 2012-07-02  George Staikos  <staikos@webkit.org>
3296
3297         [BlackBerry] Enable scoped style for BlackBerry.
3298         https://bugs.webkit.org/show_bug.cgi?id=90418
3299
3300         Reviewed by Rob Buis.
3301
3302         * Scripts/webkitperl/FeatureList.pm:
3303
3304 2012-07-02  Xiaobo Wang  <xbwang@torchmobile.com.cn>
3305
3306         [BlackBerry] Use PUBLIC_BUILD to enable/disable DRT
3307         https://bugs.webkit.org/show_bug.cgi?id=90271
3308
3309         Reviewed by George Staikos.
3310
3311         RIM PR #154707
3312
3313         Currently DRT code will be compiled only if ENABLE_DRT is set, and it's not
3314         defined by default.
3315         We should enable DRT by default unless PUBLIC_BUILD is set. In this way we don't
3316         need to rebuild webkit before running DRT.
3317
3318         * Scripts/webkitdirs.pm:
3319         (blackberryCMakeArguments):
3320
3321 2012-07-02  Ojan Vafai  <ojan@chromium.org>
3322
3323         webkit-patch rebaseline-expectations should share code with rebaseline-all
3324         https://bugs.webkit.org/show_bug.cgi?id=90413
3325
3326         Reviewed by Dirk Pranke.
3327
3328         Make them share code. In addition to reducing code duplication this makes
3329         rebaseline-expectations considerably faster by rebaselining in parallel.
3330
3331         * Scripts/webkitpy/tool/commands/rebaseline.py:
3332         (AbstractParallelRebaselineCommand):
3333         (AbstractParallelRebaselineCommand._run_webkit_patch):
3334         (AbstractParallelRebaselineCommand._rebaseline):
3335         (RebaselineJson):
3336         (RebaselineJson.execute):
3337         (RebaselineExpectations):
3338         (RebaselineExpectations._update_expectations_file):
3339         (RebaselineExpectations._tests_to_rebaseline):
3340         (RebaselineExpectations._add_tests_to_rebaseline_for_port):
3341         (RebaselineExpectations.execute):
3342         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
3343         (test_rebaseline_all):
3344         (test_rebaseline_expectations.run_in_parallel):
3345         (test_rebaseline_expectations):
3346
3347 2012-07-02  Xiaobo Wang  <xbwang@torchmobile.com.cn>
3348
3349         [BlackBerry] Update DumpRenderTree to have it work interactively in parallel
3350         https://bugs.webkit.org/show_bug.cgi?id=88326
3351
3352         Reviewed by Rob Buis.
3353
3354         1. Test name is sent to torch-launcher via PPS message(from host machine).
3355            So we get test list by monitoring and receiving PPS message instead of
3356            reading file index.drt.
3357         2. Torch-launcher create a <test file>.done file when it finished a test.
3358         3. We don't need to search for Ref-tests in DumpRenderTree.cpp any more. NRWT
3359            will get them for us.
3360
3361         * DumpRenderTree/blackberry/DumpRenderTree.cpp:
3362         (BlackBerry::WebKit::DumpRenderTree::DumpRenderTree):
3363         (BlackBerry::WebKit::DumpRenderTree::doneDrt):
3364         (BlackBerry::WebKit::DumpRenderTree::runRemainingTests):
3365         (BlackBerry::WebKit::DumpRenderTree::ensurePPS):
3366         (WebKit):
3367         (BlackBerry::WebKit::DumpRenderTree::handlePPSData):
3368         (BlackBerry::WebKit::DumpRenderTree::waitForTest):
3369         (BlackBerry::WebKit::DumpRenderTree::runTests):
3370         (BlackBerry::WebKit::DumpRenderTree::dump):
3371         * DumpRenderTree/blackberry/DumpRenderTreeBlackBerry.h:
3372         (DumpRenderTree):
3373
3374 2012-07-02  Adam Barth  <abarth@webkit.org>
3375
3376         Remove flashplugin-installer from the EWS image because it causes some tests to crash
3377         https://bugs.webkit.org/show_bug.cgi?id=90403
3378
3379         Reviewed by Tony Chang.
3380
3381         * EWSTools/cold-boot.sh:
3382
3383 2012-07-02  Leandro Gracia Gil  <leandrogracia@chromium.org>
3384
3385         [Chromium] Implement a Layout Test for editing/SurroundingText
3386         https://bugs.webkit.org/show_bug.cgi?id=82461
3387
3388         Reviewed by Ryosuke Niwa.
3389
3390         Add a new method to the layout test controller in order to retrieve the
3391         text surrounding a provided element.
3392
3393         * DumpRenderTree/chromium/LayoutTestController.cpp:
3394         (LayoutTestController::LayoutTestController):
3395         (LayoutTestController::textSurroundingElement):
3396         * DumpRenderTree/chromium/LayoutTestController.h:
3397         (LayoutTestController):
3398
3399 2012-07-02  Ojan Vafai  <ojan@chromium.org>
3400
3401         Delete unused rebaseline method in gardeningserver.py
3402         https://bugs.webkit.org/show_bug.cgi?id=90396
3403
3404         Reviewed by Eric Seidel.
3405
3406         As best I can tell, the only usage was in this unittest.
3407
3408         * Scripts/webkitpy/tool/servers/gardeningserver.py:
3409         (GardeningHTTPRequestHandler.updateexpectations):
3410         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py:
3411         (GardeningServerTest.test_rebaseline_new_port):
3412
3413 2012-07-02  Csaba Osztrogonác  <ossy@webkit.org>
3414
3415         REGRESSION(r121497): It switched off and broke many unittests
3416         https://bugs.webkit.org/show_bug.cgi?id=90371
3417
3418         Reviewed by Adam Barth.
3419
3420         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
3421         (ChromiumMacPortTest): Use snowleopard as os_version instead of leopard, because it isn't supported anymore.
3422         * Scripts/webkitpy/layout_tests/port/chromium_port_testcase.py:
3423         (ChromiumPortTestCase.test_all_test_configurations): Remove leopard testcases, because it isn't supported anymore.
3424         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
3425         (PortTestCase): Inherit class PortTestCase from unittest.TestCase instead of object.
3426
3427 2012-07-02  Ojan Vafai  <ojan@chromium.org>
3428
3429         Move rebaseline-all command from the gardening-server down into webkit-patch
3430         https://bugs.webkit.org/show_bug.cgi?id=90395
3431
3432         Reviewed by Adam Barth.
3433
3434         This is just moving code. It it in preparation for making rebaseline-expectations
3435         use the same code in order to get the parallelism benefits and reduces the amount
3436         of code we have for doing rebaselines.
3437
3438         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
3439         (CheckoutTest.test_apply_patch):
3440         Updated due to the change to executive_mock.
3441         * Scripts/webkitpy/common/system/executive_mock.py:
3442         (MockExecutive.run_command):
3443         Update to print out the input passed to stdin.
3444         * Scripts/webkitpy/tool/commands/download_unittest.py:
3445         Updated due to executive_mock change.
3446         * Scripts/webkitpy/tool/commands/rebaseline.py:
3447         (RebaselineAll):
3448         (RebaselineAll._run_webkit_patch):
3449         (RebaselineAll._builders_to_fetch_from):
3450         (RebaselineAll._rebaseline_commands):
3451         (RebaselineAll._files_to_add):
3452         (RebaselineAll._optimize_baselines):
3453         (RebaselineAll._rebaseline):
3454         (RebaselineAll.execute):
3455         All this code is just copy-pasted except for mechanical changes
3456         (e.g. self.server.tool --> self._tool) and the reading in of the 
3457         JSON from stdin instead of the post body.
3458         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
3459         (test_rebaseline_all):
3460         Copied the test-case out of gardeningserver_unittest.py.
3461         * Scripts/webkitpy/tool/servers/gardeningserver.py:
3462         (GardeningHTTPRequestHandler):
3463         (GardeningHTTPRequestHandler.rebaseline):
3464         (GardeningHTTPRequestHandler.rebaselineall):
3465         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py:
3466         (GardeningServerTest.test_rebaselineall):
3467         (GardeningServerTest.test_rebaselineall.run_command):
3468
3469 2012-07-02  Ojan Vafai  <ojan@chromium.org>
3470
3471         Remove Leopard support from the flakiness dashboard
3472         https://bugs.webkit.org/show_bug.cgi?id=90390
3473
3474         Reviewed by Adam Barth.
3475
3476         * TestResultServer/static-dashboards/flakiness_dashboard.js:
3477         (nonChromiumPlatform):
3478         (chromiumPlatform):
3479         * TestResultServer/static-dashboards/flakiness_dashboard_unittests.js:
3480         (test):
3481         * TestResultServer/static-dashboards/run-embedded-unittests.html:
3482         * TestResultServer/static-dashboards/run-unittests.html:
3483
3484 2012-07-02  Ojan Vafai  <ojan@chromium.org>
3485
3486         Fix posting from garden-o-matic. This broke in moving away from jquery's ajax method
3487         in http://trac.webkit.org/changeset/121617.
3488         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/net.js:
3489
3490 2012-07-02  Carlos Garcia Campos  <cgarcia@igalia.com>
3491
3492         [GTK] Read fonts path when running layout tests from alternative fonts dir when main dir doesn't exist
3493         https://bugs.webkit.org/show_bug.cgi?id=89437
3494
3495         Reviewed by Martin Robinson.
3496
3497         If main fonts directory doesn't exist, try with an alternative
3498         fonts directory at build directory.
3499
3500         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3501         (getOutputDir):
3502         (getFontsPath):
3503         (initializeFonts):
3504         * WebKitTestRunner/InjectedBundle/gtk/ActivateFontsGtk.cpp:
3505         (WTR::getOutputDir):
3506         (WTR):
3507         (WTR::getFontsPath):
3508         (WTR::inititializeFontConfigSetting):
3509
3510 2012-07-02  Carlos Garcia Campos  <cgarcia@igalia.com>
3511
3512         [GTK] Don't run the tests with jhbuild wrapper if it's already running under jhbuild
3513         https://bugs.webkit.org/show_bug.cgi?id=89435
3514
3515         Reviewed by Martin Robinson.
3516
3517         * Scripts/new-run-webkit-tests: Don't run the tests with the
3518         jhbuild wrapper if there isn't a Dependencies directory inside the
3519         build directory.
3520
3521 2012-06-29  Oswald Buddenhagen  <oswald.buddenhagen@nokia.com>
3522
3523         [Qt] Simplify detection of non-installed module file
3524
3525         Has the additional advantage that we do not rely on additional information.
3526
3527         Reviewed by Tor Arne Vestbø.
3528
3529         * qmake/qt_webkit.pri:
3530
3531 2012-07-01  Christophe Dumez  <christophe.dumez@intel.com>
3532
3533         [EFL] Add Gamepad support
3534         https://bugs.webkit.org/show_bug.cgi?id=90170
3535
3536         Reviewed by Kenneth Rohde Christiansen.
3537
3538         * Scripts/webkitperl/FeatureList.pm: Turn on GAMEPAD flag
3539         by default for EFL port.
3540         * efl/jhbuild.modules: Bump dependency versions of EFL libs
3541         since the latest Eeze is required for gamepad support.
3542
3543 2012-07-01  Adam Barth  <abarth@webkit.org>
3544
3545         Unreviewed.
3546
3547         Using the default start cylinder for fdisk causes a warning about the
3548         partition not starting on physical sector boundary.  The Ubuntu forums
3549         recommend using a number that's divisible by 8, which is what we do in
3550         this patch.
3551
3552         * EWSTools/cold-boot.sh:
3553
3554 2012-07-01  Adam Barth  <abarth@webkit.org>
3555
3556         Unreviewed.
3557
3558         It turns out we need to use Ubuntu 10.04 to get the right image results
3559         for chromium-linux. This patch updates our scripts to be compatible
3560         with Ubuntu 10.04.
3561
3562         * EWSTools/cold-boot.sh:
3563             - fdisk doesn't have p and 1 as default commands in 10.04.
3564         * EWSTools/start-queue.sh:
3565             - git doesn't understanding the -B argument in 10.04. We've been
3566               using this change locally on the EC2 bots for a while.
3567
3568 2012-07-01  Adam Barth  <abarth@webkit.org>
3569
3570         Unreviewed.
3571
3572         * EWSTools/boot.sh:
3573             - We need to start the screen in detached mode so that we can run
3574               it remotely via ssh.
3575
3576 2012-07-01  Adam Barth  <abarth@webkit.org>
3577
3578         Unreviewed.
3579
3580         * EWSTools/cold-boot.sh:
3581             - Turns out we need to sudo this command in order for it to
3582               actually dimiss the EULA screen. :)
3583
3584 2012-06-29  Adam Barth  <abarth@webkit.org>
3585
3586         Turns out we need zip too.
3587
3588         * EWSTools/cold-boot.sh:
3589
3590 2012-06-29  Adam Barth  <abarth@webkit.org>
3591
3592         Add a cold-boot.sh script for the EWS
3593         https://bugs.webkit.org/show_bug.cgi?id=90330
3594
3595         Unreviewed.
3596
3597         * EWSTools/cold-boot.sh: Added.
3598             - This script can take us from a cold GCE instance to a running EWS
3599               bot in one fell swoop.
3600         * EWSTools/start-queue.sh:
3601             - The if-block at the top of this script was causing trouble. I
3602               removed it from the bots a while ago. Now that we're using SVN to
3603               cold-boot the EWS bots, we need to make this change in the repo.
3604
3605 2012-06-29  Luiz Agostini  <luiz.agostini@nokia.com>
3606
3607         [Qt][WK2] Private non-QtQuick API
3608         https://bugs.webkit.org/show_bug.cgi?id=84532
3609
3610         MiniBrowserRaw is an usage example for the QRawWebView API.
3611         It is only an example and is not fully implemented.
3612
3613         Reviewed by Noam Rosenthal.
3614
3615         * MiniBrowser/qt/raw/DerivedSources.pri: Added.
3616         * MiniBrowser/qt/raw/MiniBrowserRaw.pro: Added.
3617         * MiniBrowser/qt/raw/Target.pri: Added.
3618         * MiniBrowser/qt/raw/View.cpp: Added.
3619         * MiniBrowser/qt/raw/View.h: Added.
3620         * Scripts/webkitpy/style/checker.py:
3621         * Tools.pro:
3622
3623 2012-06-29  Ojan Vafai  <ojan@chromium.org>
3624
3625         Remove leopard bots from garden-o-matic
3626         https://bugs.webkit.org/show_bug.cgi?id=90328
3627
3628         Reviewed by Adam Barth.
3629
3630         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/config.js:
3631         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/failures_unittests.js:
3632         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui/notifications_unittests.js:
3633
3634 2012-06-29  Ojan Vafai  <ojan@chromium.org>
3635
3636         garden-o-matic broken: TypeError: 'undefined' is not an object (evaluating 'buildLocations[currentIndex].url')
3637         https://bugs.webkit.org/show_bug.cgi?id=90243
3638
3639         Reviewed by Dirk Pranke.
3640
3641         jQuery was trying to be too smart and parsing the jsonp as json because of it's content-type.
3642         Excise jQuery and just use XHR directly since it's easier to maintain something where we control it
3643         all.
3644
3645         Not really sure how to unittest this. I tested it all manually of course.
3646
3647         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/net.js:
3648         Made net.ajax a drop-in replacement for the features of $.ajax that we were using.
3649         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/scripts/ui_unittests.js:
3650         Not related to this patch, but figured I'd update the failing test while I was at it.
3651
3652 2012-06-29  Yaron Friedman  <yfriedman@chromium.org>
3653
3654         Fix layout test runner for Android after https://bugs.webkit.org/show_bug.cgi?id=88134
3655         https://bugs.webkit.org/show_bug.cgi?id=90309
3656
3657         Reviewed by Adam Barth.
3658
3659         * Scripts/webkitpy/layout_tests/port/chromium_android.py:
3660         (ChromiumAndroidPort.start_http_server):
3661
3662 2012-06-29  James Weatherall  <wez@chromium.org>
3663
3664         NPObjectWrapper may not address all window script object lifetime issues
3665         https://bugs.webkit.org/show_bug.cgi?id=85679
3666
3667         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.