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