51c216a11a8e446c2a173dd2dbf20cc153efc86a
[WebKit-https.git] / Tools / ChangeLog
1 2011-07-14  Dimitri Glazkov  <dglazkov@chromium.org>
2
3         Introduce TestExpectationsModel, split out of TestExpectations.
4         https://bugs.webkit.org/show_bug.cgi?id=64531
5
6         This is a simple split-and-make-work refactoring, a first step among many.
7
8         Reviewed by Adam Barth.
9
10         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Moved all model-related members
11             out of TestExpectations and into TestExpectationsModel.
12
13 2011-07-14  Eric Seidel  <eric@webkit.org>
14
15         Move webkitpy off of loose mocks
16         https://bugs.webkit.org/show_bug.cgi?id=64508
17
18         Unreviewed.  Fixing a test which fails under test-webkitpy --all (but not in a normal run).
19
20         * Scripts/webkitpy/common/checkout/scm/git.py:
21
22 2011-07-14  Adam Barth  <abarth@webkit.org>
23
24         garden-o-matic should have a "rebaseline" button
25         https://bugs.webkit.org/show_bug.cgi?id=64446
26
27         Reviewed by Ojan Vafai.
28
29         This patch adds a basic Rebaseline button that copies the baselines
30         displayed in the results pane into the appropriate directory in your
31         working copy.
32
33         There are two main limitations:
34
35         1) There is no UI for actually committing the baselines.
36
37         2) The baselines are not optimized for redundancy (meaning you can have
38            identical baselines in both chromium-mac and chromium-win).
39
40         * Scripts/webkitpy/tool/commands/rebaseline.py:
41             - Turns out we need to create the directory for the baseline if it doesn't exist yet.
42         * Scripts/webkitpy/tool/servers/data/gardeningserver/checkout.js:
43             - Add an programatic API to call the server.
44         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
45             - Add the rebaseline button itself.
46         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
47             - Change the CSS so that the Rebaseline and Close buttons can
48               appear in the normal order in the DOM.
49         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
50             - Bind the event and translate the parameters.
51             - Hide/show the rebaseline button, as appropriate.
52         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
53             - Add some helpful utility functions for manipulating failure types.
54         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
55
56 2011-07-14  Ryosuke Niwa  <rniwa@webkit.org>
57
58         REGRESSION: webkit-patch roll-chromium-deps is broken
59         https://bugs.webkit.org/show_bug.cgi?id=64568
60
61         Reviewed by Dirk Pranke
62
63         Use _filesystem.join.
64
65         * Scripts/webkitpy/common/checkout/checkout.py:
66         * Scripts/webkitpy/common/checkout/checkout_unittests.py:
67
68 2011-07-14  Noam Rosenthal  <noam.rosenthal@nokia.com>
69
70         Adding myself to the reviewers list; No review needed.
71
72         * Scripts/webkitpy/common/config/committers.py:
73
74 2011-07-14  Adam Roben  <aroben@apple.com>
75
76         Don't use Element.prototype.classList in TestFailures
77
78         Safari 5 doesn't support it.
79
80         Fixes <http://webkit.org/b/64550> Can't expand flaky tests on TestFailures page in Safari 5
81
82         Reviewed by Daniel Bates.
83
84         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
85         (Element.prototype.hasStyleClass):
86         (Element.prototype.addStyleClass):
87         (Element.prototype.removeStyleClass):
88         (Element.prototype.toggleStyleClass):
89         Added these helper functions which simulate classList functionality.
90
91         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities_unittests.js:
92         Added. Tests for the above.
93
94         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
95         (ViewController.prototype._domForPossiblyFlakyTests): Changed to use
96         toggleStyleClass/hasStyleClass instead of classList.
97
98         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/run-unittests.html:
99         Added Utilities_unittests.js, and reordered the tested files to be in
100         roughly dependency order (i.e., the lowest-level files are imported
101         and tested first).
102
103 2011-07-14  Eric Seidel  <eric@webkit.org>
104
105         NRWT doesn't store the svn revision in full_results.json on chromium-win
106         https://bugs.webkit.org/show_bug.cgi?id=64492
107
108         Unreviewed.  Just fixing my test-webkitpy regression.
109
110         Fix the unit tests.  Unfortunately scm does not use a filesystem
111         object so we can't control the result of detect_scm_system.
112         When detect_scm_system would fail, we would log, which would
113         cause all passing_run integration tests to fail.
114
115         * Scripts/webkitpy/common/checkout/scm/scm.py:
116         * Scripts/webkitpy/layout_tests/controllers/manager.py:
117         * Scripts/webkitpy/layout_tests/port/base.py:
118         * Scripts/webkitpy/layout_tests/port/test.py:
119         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
120
121 2011-07-14  Eric Seidel  <eric@webkit.org>
122
123         NRWT doesn't store the svn revision in full_results.json on chromium-win
124         https://bugs.webkit.org/show_bug.cgi?id=64492
125
126         Reviewed by Ojan Vafai.
127
128         This should fix the bug.  Unfortunately this code is currently
129         impossible to test since it's impossible to mock detect_scm_system at the
130         moment.  I started re-writing scm.detection.py to be mockable, but decided that
131         was best left for another day.
132
133         * Scripts/webkitpy/layout_tests/controllers/manager.py:
134         * Scripts/webkitpy/layout_tests/port/base.py:
135
136 2011-07-14  Ojan Vafai  <ojan@chromium.org>
137
138         fix coding style of dashboard_base.js
139         https://bugs.webkit.org/show_bug.cgi?id=64545
140
141         Reviewed by Adam Roben.
142
143         The code changes to the files other than dashboard_base.js
144         are just fallout from renaming globals in dashboard_base.js.
145
146         * TestResultServer/static-dashboards/aggregate_results.html:
147         * TestResultServer/static-dashboards/builders.js:
148         * TestResultServer/static-dashboards/dashboard_base.js:
149         * TestResultServer/static-dashboards/flakiness_dashboard.html:
150         * TestResultServer/static-dashboards/flakiness_dashboard_tests.js:
151         * TestResultServer/static-dashboards/timeline_explorer.html:
152         * TestResultServer/static-dashboards/treemap.html:
153
154 2011-07-14  Carlos Garcia Campos  <cgarcia@igalia.com>
155
156         Unreviewed. Fix WebKit2 GTK build after r90953.
157
158         * WebKitTestRunner/GNUmakefile.am:
159
160 2011-07-14  Adam Barth  <abarth@webkit.org>
161
162         garden-o-matic results pane should be more discoverable
163         https://bugs.webkit.org/show_bug.cgi?id=64513
164
165         Reviewed by Eric Seidel.
166
167         This patch causes us to trigger the results pane on mouse clicks
168         instead of mousein.  Also, handle the case of no results details more
169         elegantly.
170
171         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
172         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
173             - Change the binding to "click".
174         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
175             - Handle the "no results URLs" case explicitly instead of never
176               calling the callback.
177         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
178         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
179             - Show a message when there are no result details to show.
180         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
181
182 2011-07-14  Adam Barth  <abarth@webkit.org>
183
184         garden-o-matic should badge TIMEOUT failures
185         https://bugs.webkit.org/show_bug.cgi?id=64435
186
187         Reviewed by Eric Seidel.
188
189         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
190         (.regression .what a[draggable].TIMEOUT::before):
191             - Add TIMEOUT.
192
193 2011-07-14  Adam Barth  <abarth@webkit.org>
194
195         Fix Python exception blocking the commit-queue.
196
197         * Scripts/webkitpy/common/checkout/scm/git.py:
198
199 2011-07-13  Eric Seidel  <eric@webkit.org>
200
201         Move webkitpy off of loose mocks
202         https://bugs.webkit.org/show_bug.cgi?id=64508
203
204         Reviewed by Adam Barth.
205
206         Using Mock has caused us more pain than help.
207         It's possible that there was a cleaner way to use it
208         (maybe Mock(class) instead of inheriting from it?).
209         But for now, I've removed all uses of Mock from mocktool.py.
210
211         I also moved run_command into the only 3 files which call it
212         instead of leaving the deprecated method in executive.py.
213
214         I changed various direct calls to os.* to use filesystem instead.
215
216         * Scripts/webkitpy/common/checkout/checkout.py:
217         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
218         * Scripts/webkitpy/common/checkout/scm/git.py:
219         * Scripts/webkitpy/common/checkout/scm/scm.py:
220         * Scripts/webkitpy/common/checkout/scm/scm_unittest.py:
221         * Scripts/webkitpy/common/checkout/scm/svn.py:
222         * Scripts/webkitpy/common/system/executive.py:
223         * Scripts/webkitpy/common/system/executive_unittest.py:
224         * Scripts/webkitpy/tool/mocktool.py:
225         * Scripts/webkitpy/tool/steps/cleanworkingdirectory_unittest.py:
226         * Scripts/webkitpy/tool/steps/ensurelocalcommitifneeded.py:
227
228 2011-07-13  Eric Seidel  <eric@webkit.org>
229
230         NRWT doesn't store the svn revision in full_results.json on chromium-win
231         https://bugs.webkit.org/show_bug.cgi?id=64492
232
233         Reviewed by Adam Barth.
234
235         Add a warning message in the revision='' case to help us diagnose
236         what's going wrong on the bot.
237
238         * Scripts/webkitpy/layout_tests/controllers/manager.py:
239
240 2011-07-13  Dirk Pranke  <dpranke@chromium.org>
241
242         new-run-webkit-tests: print baseline search path as part of config output
243         https://bugs.webkit.org/show_bug.cgi?id=64499
244
245         Reviewed by Eric Seidel.
246
247         * Scripts/webkitpy/layout_tests/controllers/manager.py:
248         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
249
250 2011-07-13  Adam Barth  <abarth@webkit.org>
251
252         gardening server should have an API for parsing changelogs
253         https://bugs.webkit.org/show_bug.cgi?id=64495
254
255         Reviewed by Eric Seidel.
256
257         This patch exposes much of the same information from CommitInfo in a
258         dictionary form, which is easier to send over-the-wire as JSON to the
259         frontend.
260
261         * Scripts/webkitpy/common/checkout/checkout.py:
262         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
263         * Scripts/webkitpy/tool/mocktool.py:
264         * Scripts/webkitpy/tool/servers/gardeningserver.py:
265         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py:
266         * Scripts/webkitpy/tool/servers/reflectionhandler.py:
267
268 2011-07-13  Eric Seidel  <eric@webkit.org>
269
270         REGRESSION: GitTestWithMock.test_create_patch fails
271         https://bugs.webkit.org/show_bug.cgi?id=62945
272
273         Reviewed by Daniel Bates.
274
275         I was not able to reproduce the exact failure seen in the bug,
276         however this test was failing on my machine for other reasons.
277
278         I went through and did an audit of our run_command usage, it's
279         entirely in scm classes after this change.  (Not surprising given
280         that scm.py was the second file ever created in webkit.py.)
281
282         The real bug I'm fixing here is that we were setting executive.should_log
283         when the value had changed to executive._should_log.  Now we set the right one
284         and the test works again.
285
286         * Scripts/webkitpy/common/checkout/checkout.py:
287         * Scripts/webkitpy/common/checkout/scm/git.py:
288         * Scripts/webkitpy/common/checkout/scm/scm.py:
289         * Scripts/webkitpy/common/checkout/scm/scm_unittest.py:
290         * Scripts/webkitpy/common/checkout/scm/svn.py:
291
292 2011-07-13  Ilya Sherman  <isherman@chromium.org>
293
294         Fix WTF header guard style check
295         https://bugs.webkit.org/show_bug.cgi?id=64488
296
297         Reviewed by David Levin.
298
299         * Scripts/webkitpy/style/checkers/cpp.py:
300         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Verify that we don't *always* suggest "WTF_" as a prefix
301
302 2011-07-13  Ojan Vafai  <ojan@chromium.org>
303
304         bring flakiness_dashboard.html closer to webkit style
305         https://bugs.webkit.org/show_bug.cgi?id=64477
306
307         Reviewed by Adam Barth.
308
309         * TestResultServer/static-dashboards/dashboard_base.js:
310         Fix bad variable name that would hit only when running the tests.
311         * TestResultServer/static-dashboards/flakiness_dashboard.html:
312         No code/logic changes. All moving/removing brackets, indents and moving things
313         to one line.
314
315 2011-07-13  Xan Lopez  <xlopez@igalia.com>
316
317         [GTK] Fix distcheck
318
319         Reviewed by Martin Robinson.
320
321         * WebKitTestRunner/GNUmakefile.am: mark built sources as nodist.
322
323 2011-07-13  Joseph Pecoraro  <joepeck@webkit.org>
324
325         Unreviewed fix. Add a "\n" to a printf that somehow got lost.
326
327         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
328         (WebCore::DumpRenderTree::dumpApplicationCacheQuota):
329
330 2011-07-13  Joseph Pecoraro  <joepeck@webkit.org>
331
332         ApplicationCache Quota Output is Flakey
333         https://bugs.webkit.org/show_bug.cgi?id=64410
334
335         Reviewed by Alexey Proskuryakov.
336
337         Unify the delegate logging for reaching application cache
338         quotas. Truncate the space needed to the nearest 10000
339         for less flakey test output. This also better supports
340         printing NSUInteger on non-64 bit systems by casting to
341         an unsigned long.
342
343         * DumpRenderTree/mac/UIDelegate.mm:
344         (-[UIDelegate webView:exceededApplicationCacheOriginQuotaForSecurityOrigin:totalSpaceNeeded:]):
345         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
346         (WebCore::DumpRenderTree::dumpApplicationCacheQuota):
347
348 2011-07-13  Dimitri Glazkov  <dglazkov@chromium.org>
349
350         Remove unused TestExpectations._overrides.
351         https://bugs.webkit.org/show_bug.cgi?id=64470
352
353         Reviewed by Adam Barth.
354
355         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Removed unused member.
356
357 2011-07-13  Dimitri Glazkov  <dglazkov@chromium.org>
358
359         Introduce TestExpectationSerializer.list_to_string.
360         https://bugs.webkit.org/show_bug.cgi?id=64462
361
362         Reviewed by Adam Barth.
363
364         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Added list_to_string and change the relevant callsite to use it.
365         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: Added tests for it.
366
367 2011-07-13  Dimitri Glazkov  <dglazkov@chromium.org>
368
369         Consolidate expectations parsing code.
370         https://bugs.webkit.org/show_bug.cgi?id=64460
371
372         Reviewed by Adam Barth.
373
374         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Folded TestExpectationParser._split_expectation_string into TestExpectationParser.parse.
375
376 2011-07-12  Brent Fulgham  <bfulgham@webkit.org>
377
378         Standardize WinCairo conditionalized code under PLATFORM macro.
379         https://bugs.webkit.org/show_bug.cgi?id=64377
380
381         Reviewed by Maciej Stachowiak.
382
383         Update compile-time conditions that had previously used #ifdef
384         of WIN_CAIRO to consistently use PLATFORM(WIN_CAIRO).
385
386         * DumpRenderTree/config.h: Switch to PLATFORM(WIN_CAIRO)
387         * DumpRenderTree/win/DumpRenderTreeCairo.vsprops: Remove redundant
388           define of WIN_CAIRO=1.
389         * DumpRenderTree/win/DumpRenderTreeDebugCairoCFLite.vsprops: Add
390           WinCairo.vsprops to set (to match Release build).
391
392 2011-07-13  Dimitri Glazkov  <dglazkov@chromium.org>
393
394         Eliminate TestExpectationsFile.
395         https://bugs.webkit.org/show_bug.cgi?id=64458
396
397         Turns out, we can just use a Python list.
398
399         Reviewed by Adam Barth.
400
401         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Folded TestExpectationsFile.append into TestExpectationParser.parse_list,
402             removed TestExpectationsFile.
403         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: Moved tests to better reflect new names, removed iterator test,
404             since there's no more custom iterator machinery.
405
406 2011-07-13  Philippe Normand  <pnormand@igalia.com>
407
408         Unreviewed, added my other email addresses.
409
410         * Scripts/webkitpy/common/config/committers.py:
411
412 2011-07-13  Adam Barth  <abarth@webkit.org>
413
414         Folks should only be listed once.
415
416         * Scripts/webkitpy/common/config/committers.py:
417
418 2011-07-13  Philippe Normand  <pnormand@igalia.com>
419
420         Unreviewed, adding myself as Reviewer.
421
422         * Scripts/webkitpy/common/config/committers.py:
423
424 2011-07-13  Adam Barth  <abarth@webkit.org>
425
426         Fix flickering bug introduced by my previous patch.  I forgot to change
427         the name of the class everywhere.
428
429         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
430
431 2011-07-12  Philippe Normand  <pnormand@igalia.com>
432
433         [GTK] media/media-blocked-by-willsendrequest.html fails
434         https://bugs.webkit.org/show_bug.cgi?id=63699
435
436         Reviewed by Martin Robinson.
437
438         * DumpRenderTree/gtk/DumpRenderTree.cpp:
439         (willSendRequestCallback): Abort the request if explicitely asked
440         by the LayoutTestController's willSendRequestReturnsNull() function.
441
442 2011-07-13  Adam Barth  <abarth@webkit.org>
443
444         Remove "Dismiss" button from garden-o-matic butterbar
445         https://bugs.webkit.org/show_bug.cgi?id=64443
446
447         Reviewed by Dimitri Glazkov.
448
449         We don't have any persistent butter bar messages yet, so the dismiss
450         button is premature (and fairly heavy, visually).
451
452         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
453         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
454
455 2011-07-13  Adam Barth  <abarth@webkit.org>
456
457         Refine garden-o-matic status pane
458         https://bugs.webkit.org/show_bug.cgi?id=64442
459
460         Reviewed by Dimitri Glazkov.
461
462         This patch makes two improvements to the status pane:
463
464         1) We only query the server for failure types that we're expecting.
465            This dramatically reduces the number of HTTP requests, making
466            loading the status pane faster.
467
468         2) The status pane now displays which test and which builder it is
469            showing results for.
470
471         * Scripts/webkitpy/tool/servers/data/gardeningserver/config.js:
472         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
473         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
474         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
475         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
476         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
477         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
478
479 2011-07-13  Adam Barth  <abarth@webkit.org>
480
481         garden-o-matic should show test results
482         https://bugs.webkit.org/show_bug.cgi?id=64440
483
484         Reviewed by Adam Roben.
485
486         This patch introduces the results detail pane, which appears at the
487         bottom of the window and contains test results from the bots.
488         Currently, you can activate the pane by mousing over one of the builder
489         names associated with a failing test.
490
491         This is just a first iteration of the UI.  There's no way to resize or
492         zoom in on elements of the details pane, and images likely aren't sized
493         correctly, but it's a place to start.
494
495         * Scripts/webkitpy/tool/servers/data/gardeningserver/config.js:
496             - Add constants for our data attributes so we don't typo them!
497         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
498             - Add DOM structure for the details pane.
499         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
500         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
501             - Bind events for showing and hiding the details pane.
502         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
503             - Reduce the number of result types that we query for to avoid
504               overfilling the details pane.  We'll probably need another
505               solution here in the long-term.  For example, we could use tabs
506               to pack more results into the pane.
507         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
508         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html:
509             - Now that we're using the config package during testing, we need
510               to include it in the testing HTML.
511         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
512         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
513             - Change the DOM structure of the results detail to make it fit
514               nicely in the details pane (rather than flowing freely in the
515               body, as it did before).
516
517 2011-07-13  Adam Roben  <aroben@apple.com>
518
519         Make TestFailures show every time a possibly-flaky test failed, but hide it by default
520
521         It's useful to be able to see every time a flaky test failed to see whether it failed the
522         same way every time. But doing so takes a lot of space, so the list of failures is now
523         collapsed by default and can be revealed using a disclosure triangle.
524
525         Fixes <http://webkit.org/b/64455> TestFailures page doesn't show as much information for
526         flaky tests as I would like, even though the page is already so long
527
528         Reviewed by Dimitri Glazkov.
529
530         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/FlakyLayoutTestDetector.js:
531         (FlakyLayoutTestDetector.prototype.allFailures): Replaced flakinessExamples with this
532         function. Now returns all failures for the given test.
533
534         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/FlakyLayoutTestDetector_unittests.js:
535         Added. This just contains some simple tests of the FlakyLayoutTestDetector class. We'll add
536         more over time.
537
538         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
539         (LayoutTestHistoryAnalyzer.prototype.start): Updated the documentation comment to reflect
540         that we no longer return passing builds for possibly-flaky tests.
541
542         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
543         (.existing-bugs-list, .suspect-revisions-list, .flakiness-examples-list): Make the list of
544         flakiness examples small, too, since it can get quite long.
545
546         (.disclosure-triangle):
547         (.expanded > .disclosure-triangle):
548         Simple styles for the disclosure triangle.
549
550         (.flakiness-examples-list): Collapse the list by default.
551
552         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
553         (ViewController.prototype._displayBuilder): Pass the total number of builds analyzed to
554         _domForPossiblyFlakyTests.
555         (ViewController.prototype._domForPossiblyFlakyTests): Put a disclosure triangle to the left
556         of each test name, and the number of failures to the right. When the disclosure triangle is
557         clicked for the first time, we build up the list of failures and expand the element. After
558         that we just collapse or expand the element on subsequent clicks.
559
560         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/run-unittests.html:
561         Pulled in new tests.
562
563 2011-07-13  Adam Roben  <aroben@apple.com>
564         
565         Teach TestFailures to understand NRWT's output when it exits early due to too many failures
566
567         Fixes <http://webkit.org/b/64456> TestFailures page reports way too many failures when NRWT
568         exits early
569
570         Reviewed by Dimitri Glazkov.
571
572         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
573         (Builder.prototype.getNumberOfFailingTests): Relaxed the "Exiting early" test not to require
574         it to be at the beginning of the line, since NRWT prints a bunch of junk earlier in the
575         line. Tightened up the regex that's used to parse the number of failing tests to require the
576         leading number to be followed by whitespace so that we won't parse the "2011" in
577         "2011-07-13" as a number of failures.
578
579         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder_unittests.js:
580         Added a test for the above.
581
582 2011-07-12  Dimitri Glazkov  <dglazkov@chromium.org>
583
584         Extract model-like TestExpectationLine and TestExpectationFile from TestExpectations.
585         https://bugs.webkit.org/show_bug.cgi?id=64386
586
587         This is the first step in converting TestExpectations to a real model.
588         * TestExpectationsLine represents a line in the test_expectations.txt file, and
589         * TestExpectationsFile represents the file, which is a collection of lines.
590
591         Reviewed by Adam Barth.
592
593         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
594         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
595         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
596
597 2011-07-13  Xan Lopez  <xlopez@igalia.com>
598
599         [GTK] Do not grab focus too early in DRT.
600
601         Reviewed by Gustavo Noronha.
602
603         It causes a layout to happen and a progress signal to be emitted
604         since r90900, but at this point we don't have a
605         LayoutTestController object and we'll eventually crash. Since we
606         already grab focus at the beginning of runTest() this is
607         redundant, so get rid of it to fix the crash.
608
609         * DumpRenderTree/gtk/DumpRenderTree.cpp:
610         (main): remove call to grab_focus
611
612 2011-07-13  Sheriff Bot  <webkit.review.bot@gmail.com>
613
614         Unreviewed, rolling out r90893 and r90894.
615         http://trac.webkit.org/changeset/90893
616         http://trac.webkit.org/changeset/90894
617         https://bugs.webkit.org/show_bug.cgi?id=64441
618
619         NRWT still doesn't work on qt-mac platform (Requested by Ossy
620         on #webkit).
621
622         * Scripts/run-webkit-tests:
623         (useNewRunWebKitTests):
624
625 2011-07-13  Csaba Osztrogonác  <ossy@webkit.org>
626
627         [Qt] NRWT should pick up the right httpd config file
628         https://bugs.webkit.org/show_bug.cgi?id=64086
629
630         * Scripts/run-webkit-tests: Enable NRWT on qt-mac platform after r90810.
631         (useNewRunWebKitTests):
632
633 2011-07-12  Mark Rowe  <mrowe@apple.com>
634
635         Fix the 32-bit build.
636
637         * DumpRenderTree/mac/UIDelegate.mm:
638         (-[UIDelegate webView:exceededApplicationCacheOriginQuotaForSecurityOrigin:totalSpaceNeeded:]):
639         Cast the NSUInteger value to unsigned long to match the format specifier.
640
641 2011-07-12  Adam Barth  <abarth@webkit.org>
642
643         Improve garden-o-matic UI when the bots fail to report revision numbers
644         https://bugs.webkit.org/show_bug.cgi?id=64427
645
646         Reviewed by Dimitri Glazkov.
647
648         At least Win (dbg)(2) seems to fail to report the SVN revision number
649         in full_results.json.  This patch makes garden-o-matic more robust to
650         missing revision numbers.
651
652         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
653         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
654         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
655         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
656         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
657
658 2011-07-12  Adam Barth  <abarth@webkit.org>
659
660         Adjust garden-o-matic layout to use fewer lines
661         https://bugs.webkit.org/show_bug.cgi?id=64422
662
663         Reviewed by Ojan Vafai.
664
665         This patch effectively merges the test name line with the list of
666         builders on which the test fails.  Each test failure now occupies two
667         lines instead of three.
668
669         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
670         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
671         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
672
673 2011-07-12  Adam Barth  <abarth@webkit.org>
674
675         garden-o-matic should label tests failures of new tests as such
676         https://bugs.webkit.org/show_bug.cgi?id=64421
677
678         Reviewed by Ojan Vafai.
679
680         Also, we shouldn't dim failures of new tests, even if we've only seen
681         them once because they're likely to be real problems that need
682         attention.
683
684         * Scripts/webkitpy/tool/servers/data/gardeningserver/base_unittests.js:
685         * Scripts/webkitpy/tool/servers/data/gardeningserver/checkout.js: Added.
686         * Scripts/webkitpy/tool/servers/data/gardeningserver/checkout_unittests.js: Added.
687         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
688         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
689         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
690         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
691         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html:
692         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
693         * Scripts/webkitpy/tool/servers/gardeningserver.py:
694
695 2011-07-12  Adam Barth  <abarth@webkit.org>
696
697         garden-o-matic should display how many times we've seen a failure
698         https://bugs.webkit.org/show_bug.cgi?id=64417
699
700         Reviewed by Ojan Vafai.
701
702         This patch adds some UI to display how many times we've seen a given
703         failure, which can be helpful for determining whether that failure is a
704         real failure or a flaky test.
705
706         When a failure has only been seen once (i.e., only a single run on a
707         single bot), we set the opacity of to 50% to avoid distracting the
708         gardener.
709
710         This patch also refactors the failure walker to have a simpler API
711         internally by moving from an object-oriented paradigm to a functional
712         paradigm.
713
714         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
715         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
716         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
717         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
718         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
719         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
720
721 2011-07-12  Chris Rogers  <crogers@google.com>
722
723         Enable Web Audio for chromium DRT
724         https://bugs.webkit.org/show_bug.cgi?id=64409
725
726         Reviewed by James Robinson.
727
728         * DumpRenderTree/chromium/TestShell.cpp:
729         (TestShell::TestShell):
730
731 2011-07-12  Adam Barth  <abarth@webkit.org>
732
733         Tweak some UI in garden-o-matic now that the tree actually has a
734         failure and I can see what this all looks like.
735
736         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
737         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
738         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
739
740 2011-07-12  Joseph Pecoraro  <joepeck@webkit.org>
741
742         Unreviewed. Skipping a few tests which fail due to differing output
743         in recent ApplicationCache quota tests. Reenabling the tests is
744         tracked by <http://webkit.org/b/64410>.
745
746         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
747         (WebCore::DumpRenderTree::dumpApplicationCacheQuota): This should
748         have been a space, not a newline.
749
750 2011-07-12  Adam Barth  <abarth@webkit.org>
751
752         garden-o-matic should display regression ranges
753         https://bugs.webkit.org/show_bug.cgi?id=64407
754
755         Reviewed by Dimitri Glazkov.
756
757         This patch computes a regression range for a failure by intersecting
758         the regression ranges seen by the various bots.  We make the underlying
759         assumption that a test is only failing due to one revision at any given
760         moment.  If that's not true, this code probably explodes.
761
762         The regression ranges appear asynchronously, which might be a jarring
763         UI.  We'll have to experiment to see.
764
765         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
766         * Scripts/webkitpy/tool/servers/data/gardeningserver/base_unittests.js:
767         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
768         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
769         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
770         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
771         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
772
773 2011-07-12  Adam Barth  <abarth@webkit.org>
774
775         cr-linux-ews should run pixel tests
776         https://bugs.webkit.org/show_bug.cgi?id=64394
777
778         Reviewed by Ojan Vafai.
779
780         This patch enabled pixel test on the cr-linux-ews.  Running a quick
781         test on the bots, it looks like we have five pixel failures, related to
782         some missing fonts.  I've also made the bots skip failing tests to
783         avoid uploading huge zip files to bugs.webkit.org (with all the exected
784         IMAGE failures).
785
786         * Scripts/webkitpy/common/config/ports.py:
787         * Scripts/webkitpy/common/config/ports_unittest.py:
788
789 2011-07-12  Joseph Pecoraro  <joepeck@webkit.org>
790
791         ApplicationCache update should not immediately fail when reaching per-origin quota
792         https://bugs.webkit.org/show_bug.cgi?id=64177
793
794         Reviewed by Alexey Proskuryakov.
795
796         Add layoutTestController.disallowIncreaseForApplicationCacheQuota()
797         to disable the default behavior of raising the default per-origin
798         quota to 5MB when reached. This allows us to test what happens
799         when the quota is reached and not increased.
800
801         * DumpRenderTree/LayoutTestController.cpp:
802         (LayoutTestController::LayoutTestController):
803         (disallowIncreaseForApplicationCacheQuotaCallback):
804         (LayoutTestController::staticFunctions):
805         * DumpRenderTree/LayoutTestController.h:
806         (LayoutTestController::disallowIncreaseForApplicationCacheQuota):
807         (LayoutTestController::setDisallowIncreaseForApplicationCacheQuota):
808         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
809         (LayoutTestController::reset):
810         * DumpRenderTree/qt/LayoutTestControllerQt.h:
811         (LayoutTestController::disallowIncreaseForApplicationCacheQuota):
812         (LayoutTestController::dumpApplicationCacheDelegateCallbacks):
813         Expose a function to prevent automatically increasing the per-origin
814         quota when the quota is reached. This allows us to test what happens
815         when a user would "disallow" a quota increase.
816
817         * DumpRenderTree/mac/UIDelegate.mm:
818         (-[UIDelegate webView:exceededApplicationCacheOriginQuotaForSecurityOrigin:spaceNeeded:]):
819         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
820         (WebCore::DumpRenderTree::dumpApplicationCacheQuota):
821         Respect the new disallowIncrease flag and dump the spaceNeeded values
822         for ports that implement application cache quotas.
823
824 2011-07-12  Ojan Vafai  <ojan@chromium.org>
825
826         fix flakiness dashboard to work with new crash log filenames
827         https://bugs.webkit.org/show_bug.cgi?id=64393
828
829         Reviewed by Adam Barth.
830
831         * TestResultServer/static-dashboards/flakiness_dashboard.html:
832
833 2011-07-12  Adam Barth  <abarth@webkit.org>
834
835         Add a happy animation to garden-o-matic when there are no failures
836         https://bugs.webkit.org/show_bug.cgi?id=64382
837
838         Reviewed by Ojan Vafai.
839
840         As requested by Ojan.
841
842         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
843         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
844         * Scripts/webkitpy/tool/servers/data/gardeningserver/partytime.gif: Added.
845         * Scripts/webkitpy/tool/servers/gardeningserver.py:
846
847 2011-07-12  Adam Barth  <abarth@webkit.org>
848
849         sheriffbot can't perform rollouts
850         https://bugs.webkit.org/show_bug.cgi?id=64370
851
852         Reviewed by Eric Seidel.
853
854         Now that we're not updating the working copy before processing every
855         IRC message, the rollout command needs to update it manually.
856
857         * Scripts/webkitpy/tool/bot/irc_command.py:
858         * Scripts/webkitpy/tool/bot/irc_command_unittest.py:
859
860 2011-07-12  Adam Barth  <abarth@webkit.org>
861
862         Fix the grammar in the FIXME comment.
863
864         * Scripts/webkitpy/common/net/resultsjsonparser.py:
865
866 2011-07-12  Eric Seidel  <eric@webkit.org>
867
868         NRWT should open test results page with Safari trunk, not the system provided one on Mac
869         https://bugs.webkit.org/show_bug.cgi?id=64346
870
871         Reviewed by Adam Barth.
872
873         To fix this I implemented Port.show_results_html_file in Mac, Gtk and Qt ports with
874         implementations (mostly) matching those found in old-run-webkit-tests.
875         There are still some minor differences for Qt which Qt hackers may wish to tweak.
876
877         I had to add a WebKitPort._port_flag_for_scripts method (similar to flag() in common.config.ports.py)
878         for the Qt/Gtk ports which always require a flag passed to scripts.
879
880         While trying to test this, I found that FactoryTest.test_chromium_gpu_linux
881         was using a real filesystem (due to assert_platform) which due to
882         global variables in config.py was causing set-webkit-configuration to have
883         an affect on unit test results!  So I fixed this by making FactoryTest.assert_port
884         pass a mock file system whenever calling factory.get.
885
886         Unfortunately TestPort was depending on always being passed a None filesystem
887         and asserting filesystem._tests (only true for unit_test_filesystem()).
888         So I just removed the FactoryTest.test_test and FactoryTest.test_dryrun tests
889         deciding that they were pretty much useless anyway.  If others feel strongly
890         I'm happy to fix this in a different way.
891
892         * Scripts/webkitpy/common/system/executive.py:
893          - default arguments in python are screwy.  They use a single shared
894            instance, so it's better to use argument=None and then argument = argument or Default()
895            if you have any chance of mutating (or returning) the default argument.
896         * Scripts/webkitpy/layout_tests/port/config.py:
897          - This code is wrong.  We don't need to use a global variable here (as far as I can tell).
898            I'm not fixing it in this patch, but I've marked it with a FIXME and we can convert to
899            storing the results of the read on the Config object (which should only be created once during normal operation).
900            Unit tests shouldn't be hitting the disk anyway.  It's possible Config should move off of Port and onto Tool/Host directly.
901         * Scripts/webkitpy/layout_tests/port/factory.py:
902         * Scripts/webkitpy/layout_tests/port/factory_unittest.py:
903         * Scripts/webkitpy/layout_tests/port/gtk.py:
904         * Scripts/webkitpy/layout_tests/port/gtk_unittest.py: Added.
905         * Scripts/webkitpy/layout_tests/port/mac.py:
906         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
907         * Scripts/webkitpy/layout_tests/port/qt.py:
908         * Scripts/webkitpy/layout_tests/port/qt_unittest.py:
909         * Scripts/webkitpy/layout_tests/port/webkit.py:
910
911 2011-07-12  Adam Roben  <aroben@apple.com>
912
913         Teach TestFailures to recognize when run-webkit-tests gets killed by buildbot
914
915         Fixes <http://webkit.org/b/64358> TestFailures page thinks all tests passed in
916         http://build.webkit.org/builders/Windows%207%20Release%20(Tests)/builds/14672
917
918         Reviewed by Daniel Bates.
919
920         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
921         (Builder.prototype.getNumberOfFailingTests): If run-webkit-tests exited with a non-zero
922         exit status but we didn't find any failure counts, assume that there was some error that
923         caused run-webkit-tests to die early (like being killed by buildbot due to a timeout).
924
925         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder_unittests.js:
926         Added a new test that shows that we get a failingTestCount of -1 when run-webkit-tests dies
927         early.
928         (runGetNumberOfFailingTestsTest): Moved most code here from the only pre-existing test in this
929         file.
930
931         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
932         (LayoutTestResultsLoader.prototype.start): Bump the cache number so old cached data that was
933         tainted by the bug fixed in this patch will be evicted.
934
935 2011-07-12  Adam Barth  <abarth@webkit.org>
936
937         cr-linux-ews complains about tests that aren't actually failing
938         https://bugs.webkit.org/show_bug.cgi?id=64367
939
940         The underlying problem here is that full_results.json doesn't have
941         enough information to determine whether a given test result was
942         expected because whether an actual result is expected depends on
943         whether full_results.json was generated duing a test run that included
944         pixel tests.
945
946         The right long-term solution is to make full_results.json a complete
947         description of what happened durning a test run.  In this patch
948         however, to stop the spam, this patch makes the jsonresultsparser
949         ignore pixel failures.
950
951         (I'm landing this patch unreviewed in order to stop the bots from
952         spamming.  I'll happily address any review comments in a follow-up
953         patch.)
954
955         * Scripts/webkitpy/common/net/resultsjsonparser.py:
956         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
957
958 2011-07-12  Adam Barth  <abarth@webkit.org>
959
960         Re-work garden-o-matic UI to begin to look like the mocks
961         https://bugs.webkit.org/show_bug.cgi?id=64334
962
963         Reviewed by Dimitri Glazkov.
964
965         This patch changes how we display failures to be more like the mock the
966         dglazkov made. We now can display the same information much more
967         compactly.
968
969         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
970         * Scripts/webkitpy/tool/servers/data/gardeningserver/base_unittests.js:
971         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
972         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css: Added.
973         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
974         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
975         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
976         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
977         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
978         * Scripts/webkitpy/tool/servers/gardeningserver.py:
979
980 2011-07-11  Adam Roben  <aroben@apple.com>
981
982         Extract TestFailures's bug-filing code into two new classes
983
984         Fixes <http://webkit.org/b/64300> TestFailures page's new-bug-filing code is a mess!
985
986         Reviewed by Darin Adler and Adam Barth.
987
988         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Buildbot.js:
989         (Buildbot.prototype.resultsDirectoryURL): Changed to return a URI-encoded URL. Otherwise the
990         URL isn't valid (and it's harder to mock this function).
991
992         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Buildbot_unittests.js:
993         Added. Just tests the above change (for now).
994
995         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder_unittests.js:
996         Added a license header, enclosed everything in a closure to avoid polluting the global
997         namespace, changed the test name to actually describe the passing condition, and replaced
998         uses of equals() with equal(). (The latter is the actual name of the function, and matches
999         deepEqual, etc.)
1000
1001         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/NewBugForm.js: Added.
1002         (NewBugForm): This class knows how to construct a <form> used for filing new bugs in
1003         Bugzilla based on some parameters.
1004         (NewBugForm.prototype.domElement): Creates and returns the <form> element. Code came from
1005         ViewController.prototype._domForNewAndExistingBugs.
1006
1007         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/NewBugForm_unittests.js:
1008         Added. Tests the above code.
1009
1010         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailureBugForm.js: Added.
1011         (TestFailureBugForm): This class knows how to construct a <form> element used for filing new
1012         bugs specifically about test failures. Code came from
1013         ViewController.prototype._domForNewAndExistingBugs.
1014         (TestFailureBugForm.prototype.domElement): Slightly customizes the <form> element returned
1015         by NewBugForm.
1016         (TestFailureBugForm.prototype._computeOperatingSystem):
1017         (TestFailureBugForm.prototype._computePlatform):
1018         (TestFailureBugForm.prototype._createBugTitle):
1019         (TestFailureBugForm.prototype._failingResultsHTMLURL):
1020         (TestFailureBugForm.prototype._failingRevision):
1021         (TestFailureBugForm.prototype._passingRevision):
1022         (TestFailureBugForm.prototype._regressionRangeString):
1023         Code came from ViewController.prototype._domForNewAndExistingBugs. I broke it out into
1024         separate functions to break up the rat's nest a bit.
1025
1026         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailureBugForm_unittests.js:
1027         Added. Tests the above code.
1028
1029         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1030         (ViewController.prototype._domForNewAndExistingBugs): Moved code from here to
1031         TestFailureBugForm/NewBugForm, and changed this code to use a TestFailureBugForm.
1032
1033         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/index.html: Pull in
1034         NewBugForm/TestFailureBugForm.js.
1035
1036         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/run-unittests.html:
1037         Added new tests and required files.
1038
1039 2011-07-12  Adam Roben  <aroben@apple.com>
1040
1041         Test that no intermediate WTF::Strings are created when concatenating with string literals
1042
1043         Test for <http://webkit.org/b/63330> Concatenating string literals and WTF::Strings using
1044         operator+ is suboptimal
1045
1046         Reviewed by Darin Adler.
1047
1048         * TestWebKitAPI/Tests/WTF/StringOperators.cpp: Added.
1049         (TestWebKitAPI::TEST): Test that a bunch of different string concatenation expressions don't
1050         create any intermediate WTF::Strings while they're being evaluated.
1051
1052         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
1053         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
1054         Added new file.
1055
1056 2011-07-12  Eric Seidel  <eric@webkit.org>
1057
1058         [Qt] NRWT should pick up the right httpd config file
1059         https://bugs.webkit.org/show_bug.cgi?id=64086
1060
1061         Reviewed by Adam Barth.
1062
1063         This is more fall-out from the Port class inappropriately encapsulating
1064         both platform and port knowledge.  We need to split out some of this
1065         platform knowledge into a separate class which can be better shared
1066         between ports.
1067
1068         The fix was to move all the _path_to_apache_config_file logic down
1069         into the WebKitPort baseclass so that all of the ports can find the
1070         right apache config file, regardless of what platform they may be running on.
1071
1072         I did not move this down into base.Port (even though I considered it).
1073         Chromium duplicates some of this logic, but since they have separate
1074         subclasses for each port-OS combination (e.g. ChromiumMac, ChromiumWin)
1075         they wouldn't notice the change.  Eventually we'll move this logic
1076         out of Port entirely, and then it will be shared by all ports.
1077
1078         I also cleaned up the http-config logic for ORWT while I was there,
1079         although since we're killing that code, I'm happy to revert that part
1080         of the change if changing it is viewed as needlessly risky.
1081
1082         During this cleanup, I noticed that no ports use the "default" httpd.conf
1083         which is left over from Mac OS X Tiger (and old linux distros) which used
1084         Apache 1.3.  I've removed httpd.conf (and associated support in ORWT) since
1085         we no longer support any ports which use this httpd.conf.
1086
1087         * Scripts/webkitperl/httpd.pm:
1088         (getHTTPDConfigPathForTestDirectory):
1089         * Scripts/webkitpy/layout_tests/port/base.py:
1090         * Scripts/webkitpy/layout_tests/port/gtk.py:
1091         * Scripts/webkitpy/layout_tests/port/mac.py:
1092         * Scripts/webkitpy/layout_tests/port/qt.py:
1093         * Scripts/webkitpy/layout_tests/port/webkit.py:
1094         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
1095         * Scripts/webkitpy/layout_tests/port/win.py:
1096         * Scripts/webkitpy/layout_tests/servers/apache_http_server.py:
1097
1098 2011-07-11  Hironori Bono  <hbono@chromium.org>
1099
1100         Reviewed by Adam Roben.
1101
1102         Implement layoutTestController.setTextDirection for WebKit2 and Windows.
1103         https://bugs.webkit.org/show_bug.cgi?id=61931
1104
1105         This change implements layoutTextController.setTextDirection for WebKit2
1106         and Windows so we can run a layout test added by r87770 on them. Each
1107         implementation adds a wrapper function for Editor::setBaseWritingDirection()
1108         so LayoutTestController can call it.
1109
1110         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1111         (LayoutTestController::setTextDirection): Implemented the binding function
1112         for layoutTestController.setTextDirection (Windows).
1113         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
1114         Added a binding function for layoutTestController.setTextDirection (WebKit2).
1115         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: ditto.
1116         (WTR::LayoutTestController::setTextDirection):
1117         * WebKitTestRunner/InjectedBundle/LayoutTestController.h: ditto.
1118
1119 2011-07-11  Raphael Kubo da Costa  <kubo@profusion.mobi>
1120
1121         [EFL] ImageDiff: Make sure gEcoreEvas is destroyed at the right time.
1122         https://bugs.webkit.org/show_bug.cgi?id=64293
1123
1124         Reviewed by Kent Tamura.
1125
1126         gEcoreEvas was being destroyed after shutdownEfl() was called, so the
1127         call to ecore_evas_free() failed.
1128
1129         We now shut it down manually before shutdownEfl() is called.
1130
1131         * DumpRenderTree/efl/ImageDiff.cpp:
1132         (main): Destroy gEcoreEvas before shutting down the EFL.
1133
1134 2011-07-11  Ryosuke Niwa  <rniwa@webkit.org>
1135
1136         Make Chromium Mac and Linux tests bot core builders
1137         https://bugs.webkit.org/show_bug.cgi?id=63196
1138
1139         Reviewed by Adam Barth.
1140
1141         Added Chromium Mac and Linux Release Tests bots core builders.
1142
1143         Also rearranged the order of bots so that test bots show up right next to
1144         respective builders.
1145
1146         * BuildSlaveSupport/build.webkit.org-config/config.json:
1147         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
1148         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1149
1150 2011-07-11  Raphael Kubo da Costa  <kubo@profusion.mobi>
1151
1152         [EFL] DRT: Add code related to font management.
1153         https://bugs.webkit.org/show_bug.cgi?id=63989
1154
1155         Reviewed by Kent Tamura.
1156
1157         The added files are responsible for managing fontconfig and adding the
1158         fonts required by some layout tests to fontconfig's path.
1159
1160         * DumpRenderTree/efl/FontManagement.cpp: Added.
1161         (getFontDirectories):
1162         (getFontFiles):
1163         (addFontDirectories):
1164         (addFontFiles):
1165         (addFontsToEnvironment):
1166         * DumpRenderTree/efl/FontManagement.h: Added.
1167
1168 2011-07-08  Dirk Pranke  <dpranke@chromium.org>
1169
1170         nrwt: linting fixes
1171         https://bugs.webkit.org/show_bug.cgi?id=64225
1172
1173         Reviewed by Eric Siedel.
1174
1175         Miscellaneous linting fixes. The most notable change is that
1176         we add public attributes for user, executive, filesystem, and
1177         options on the Port object, so we don't have to refer to the
1178         "protected" versions all over the place".
1179
1180         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1181         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
1182         * Scripts/webkitpy/layout_tests/controllers/message_broker.py:
1183         * Scripts/webkitpy/layout_tests/controllers/message_broker_unittest.py:
1184         * Scripts/webkitpy/layout_tests/controllers/worker.py:
1185         * Scripts/webkitpy/layout_tests/port/base.py:
1186         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1187
1188 2011-07-11  Ryosuke Niwa  <rniwa@webkit.org>
1189
1190         webkit-patch roll-chromium-deps no longer works
1191         https://bugs.webkit.org/show_bug.cgi?id=64324
1192
1193         Reviewed by James Robinson.
1194
1195         Lower the minimum commit message length to 10.
1196
1197         * Scripts/webkitpy/tool/steps/commit.py:
1198
1199 2011-07-11  Alice Boxhall  <aboxhall@chromium.org>
1200
1201         Convert json_results_generator.py to output version 4 JSON.
1202         https://bugs.webkit.org/show_bug.cgi?id=60869
1203
1204         Reviewed by Ojan Vafai.
1205
1206         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
1207         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1208         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
1209
1210 2011-07-11  Martin Robinson  <mrobinson@igalia.com>
1211
1212         [GTK] [NRWT] Pixel tests do not work
1213         https://bugs.webkit.org/show_bug.cgi?id=64091
1214
1215         Reviewed by Eric Seidel.
1216
1217         Fix pixel test runs for GTK+ new-run-webkit-tests by exposing the
1218         path to the ImageDiff binary.
1219
1220         * Scripts/webkitpy/layout_tests/port/gtk.py: Implement _path_to_image_diff.
1221
1222 2011-07-11  Dirk Pranke  <dpranke@chromium.org>
1223
1224         Fix hang on win32 in a multiprocessing test that shouldn't have been running.
1225
1226         Unreviewed, build fix.
1227
1228         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1229
1230 2011-07-11  Adam Roben  <aroben@apple.com>
1231
1232         Don't count leaks as test failures on TestFailures's front page
1233
1234         As a bonus, this patch adds our first unit test.
1235
1236         Fixes <http://webkit.org/b/64303> REGRESSION (r90489): TestFailures page says far more tests
1237         are failing on the Leaks bot than actually are
1238
1239         Reviewed by Adam Barth.
1240
1241         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
1242         (Builder.prototype.getNumberOfFailingTests): Exclude lines containing "leak" when summing
1243         failure counts.
1244
1245         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder_unittests.js:
1246         Added. Contains a simple test to show that leaks aren't counted as test failures.
1247
1248         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/run-unittests.html:
1249         Added. This is the test harness.
1250
1251 2011-06-14  Mike Stegeman  <mrstegeman@gmail.com>
1252
1253         [GTK] Add API to allow setting local storage database path
1254         https://bugs.webkit.org/show_bug.cgi?id=62091
1255
1256         Reviewed by Martin Robinson and Gustavo Noronha.
1257
1258         Expose the path of the localStorage databases through a setting
1259         to allow HTML5 localStorage to be persistent. New setting is:
1260         html5-local-storage-database-path
1261
1262         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1263         (resetDefaultsToConsistentValues):
1264
1265 2011-07-11  Adam Barth  <abarth@webkit.org>
1266
1267         garden-o-matic should be able to rebaseline tests
1268         https://bugs.webkit.org/show_bug.cgi?id=64186
1269
1270         Reviewed by Eric Seidel.
1271
1272         This patch contains a simple HTTP server binding for the
1273         rebaseline-test command.  The frontend will appear in another patch.
1274
1275         * Scripts/webkitpy/tool/servers/gardeningserver.py:
1276         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py:
1277         * Scripts/webkitpy/tool/servers/reflectionhandler.py:
1278
1279 2011-07-11  Adam Barth  <abarth@webkit.org>
1280
1281         Add some basic UI for showing regression ranges
1282         https://bugs.webkit.org/show_bug.cgi?id=64243
1283
1284         Reviewed by Eric Seidel.
1285
1286         This UI is all temporary.  It exists mostly just to have some buttons
1287         to click to exercise the code.
1288
1289         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
1290         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
1291         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
1292         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
1293         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
1294
1295 2011-07-11  Adam Barth  <abarth@webkit.org>
1296
1297         garden-o-matic should be able to determine which revisions caused a given failure
1298         https://bugs.webkit.org/show_bug.cgi?id=64189
1299
1300         Reviewed by Adam Roben.
1301
1302         Walking the failure history looking for failures turns out to be
1303         slightly tricky because the network requests are asynchronous.
1304         Currently we do all the fetches serially and our cache is unbounded.
1305         We'll probably optimize both those parameters eventually.
1306
1307         This patch also generalizes some functionality in the unit testing
1308         framework to make testing this sort of code easier.
1309
1310         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
1311         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
1312         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
1313
1314 2011-07-11  Adam Barth  <abarth@webkit.org>
1315
1316         Add a webkit-patch command for rebaselining an individual test
1317         https://bugs.webkit.org/show_bug.cgi?id=64246
1318
1319         Reviewed by Eric Seidel.
1320
1321         This patch introduces a command that's able to rebaseline a single
1322         test.  Currently, the command works only with the build.chromium.org
1323         buildbots, but extending it to work with the build.webkit.org bots
1324         shouldn't be that hard.
1325
1326         A complete rebaseling tool should also include an "optimize baselines"
1327         command (which moves/deletes baselines in order to reduce the number of
1328         expected results files), but that will come in a future patch.
1329
1330         Really BuilderToPort should be merged into builders.py, but I'm going
1331         to save that for a future patch as well.  (We need to stop shaving yaks
1332         at some point.)
1333
1334         * Scripts/webkitpy/tool/commands/rebaseline.py:
1335
1336 2011-07-11  Adam Barth  <abarth@webkit.org>
1337
1338         garden-o-matic should be able to roll out patches
1339         https://bugs.webkit.org/show_bug.cgi?id=64185
1340
1341         Reviewed by Eric Seidel.
1342
1343         This gardening server API is simply a binding to the rollout machinery
1344         already present in webkit-patch.
1345
1346         * Scripts/webkitpy/tool/commands/gardenomatic.py:
1347         * Scripts/webkitpy/tool/servers/gardeningserver.py:
1348         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py: Added.
1349
1350 2011-07-11  Csaba Osztrogonác  <ossy@webkit.org>
1351
1352         [Qt][Mac] Unreviewed fix after r90746.
1353
1354         * Scripts/run-webkit-tests:
1355         (useNewRunWebKitTests): Really disable NRWT for qt-mac platform.
1356
1357 2011-07-11  Adam Roben  <aroben@apple.com>
1358
1359         Teach TestFailures that ORWT's results.html file might be missing due to all tests passing
1360
1361         There are three reasons why we might fail to fetch ORWT's results.html:
1362           - All tests passed, so no results.html was generated
1363           - Some error during the test run caused results.html not to be generated (e.g., ORWT
1364             timed out)
1365           - Some network error occurred when fetching results.html
1366
1367         We were failing to account for the first possibility in some cases. For test runs before
1368         r89610, we first check build.webkit.org/json to determine how many tests failed and whether
1369         ORWT exited early due to too many failures; if all tests passed then we don't fetch
1370         results.html at all. r89610 changed ORWT to put information in results.html about exiting
1371         early due to too many failures, so we no longer needed to check build.webkit.org/json to get
1372         that information, and in r89619 I changed TestFailures to do just that. But I forgot that we
1373         still needed to check build.webkit.org/json to find out if all tests passed!
1374
1375         Now, for test runs after r89610, we check results.html first, and then check
1376         build.webkit.org/json if we fail to fetch results.html. This lets us distinguish between all
1377         tests passing and the error cases.
1378
1379         Fixes <http://webkit.org/b/64280> TestFailures page can't pinpoint that r90699 caused 13
1380         tests to fail on Windows 7 Release (WebKit2 Tests)
1381
1382         Reviewed by David Kilzer.
1383
1384         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
1385         (LayoutTestResultsLoader.prototype.start): Bumped the cache version so that old, buggy
1386         cached data will get evicted. We were marking builds where all tests passed as errors!
1387         (LayoutTestResultsLoader.prototype._fetchAndParseORWTResults): Added success/error callback
1388         parameters to the fetchAndParseResultsHTML helper function, and added a similar
1389         fetchNumberOfFailingTests function that fetches data from build.webkit.org/json (code came
1390         from later in the function). For test runs before r89610, we first check
1391         build.webkit.org/json then check results.html. For builds after r89610, we first check
1392         results.html and then check build.webkit.org/json if we couldn't fetch results.html.
1393
1394 2011-07-11  Csaba Osztrogonác  <ossy@webkit.org>
1395
1396         new-run-webkit-tests does not support qt-arm or qt-4.8 results
1397         https://bugs.webkit.org/show_bug.cgi?id=64071
1398
1399         [Qt] NRWT should pick up the right httpd config file
1400         https://bugs.webkit.org/show_bug.cgi?id=64086
1401
1402         Reviewed by Zoltán Herczeg.
1403
1404         * Scripts/run-webkit-tests: Disable NRWT for qt-mac, qt-arm and qt-4.8 platforms until fix.
1405         (useNewRunWebKitTests):
1406
1407 2011-07-11  Kenichi Ishibashi  <bashi@chromium.org>
1408
1409         Add TestNetscapePlugIn/Tests/FormValue.cpp to DRT build files
1410         https://bugs.webkit.org/show_bug.cgi?id=64248
1411
1412         Reviewed by Kent Tamura.
1413
1414         * DumpRenderTree/DumpRenderTree.gypi: Added FormValue.cpp.
1415         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Ditto.
1416         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro: Ditto.
1417         * GNUmakefile.am: Ditto.
1418
1419 2011-07-10  Kenichi Ishibashi  <bashi@chromium.org>
1420
1421         Let plugins participate in form submission
1422         https://bugs.webkit.org/show_bug.cgi?id=13061
1423
1424         Adds NPPVformValue support to the testing plugin.
1425
1426         Reviewed by Anders Carlsson.
1427
1428         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Added FormValue.cpp
1429         * DumpRenderTree/TestNetScapePlugIn/Tests/FormValue.cpp: Added.
1430         (FormValue::FormValue): Ditto.
1431         (FormValue::NPP_GetValue): Ditto.
1432
1433 2011-07-11  Roland Steiner  <rolandsteiner@chromium.org>
1434
1435         Reviewed by Tony Chang.
1436
1437         Complete functions in filesystem.py
1438         https://bugs.webkit.org/show_bug.cgi?id=63528
1439
1440         - missing text file functions added
1441         - functions sorted
1442         - removed 'append' optional parameters (were unused)
1443         - adapted filesystem_mock in the same way
1444
1445         * Scripts/webkitpy/common/system/filesystem.py:
1446         * Scripts/webkitpy/common/system/filesystem_mock.py:
1447         * Scripts/webkitpy/common/system/filesystem_unittest.py:
1448
1449 2011-07-10  Adam Barth  <abarth@webkit.org>
1450
1451         Clean up style in fallback path calculation
1452         https://bugs.webkit.org/show_bug.cgi?id=64239
1453
1454         Reviewed by Daniel Bates.
1455
1456         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1457         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1458         * Scripts/webkitpy/layout_tests/port/mac.py:
1459         * Scripts/webkitpy/layout_tests/port/win.py:
1460
1461 2011-07-10  Balazs Kelemen  <kbalazs@webkit.org>
1462
1463         [Qt][WK2] Implement the rest of WTR::PlatformWebView
1464         https://bugs.webkit.org/show_bug.cgi?id=63630
1465
1466         Reviewed by Kenneth Rohde Christiansen.
1467
1468         * WebKitTestRunner/qt/PlatformWebViewQt.cpp:
1469         (WTR::PlatformWebView::windowFrame):
1470         (WTR::PlatformWebView::setWindowFrame):
1471
1472 2011-07-10  Mark Rowe  <mrowe@apple.com>
1473
1474         Teach build-webkit how to find the latest WebKitSystemInterface binary.
1475
1476         * Scripts/build-webkit:
1477
1478 2011-07-09  Adam Roben  <aroben@apple.com>
1479
1480         Teach TestFailures to abbreviate the examples of test flakiness
1481
1482         These lists can get quite long, and it's not really helpful in most cases to have soooooo
1483         many examples of flakiness.
1484
1485         Fixes <http://webkit.org/b/64203> Lists of flaky revisions on TestFailures page can get so
1486         long they're hard to navigate
1487
1488         Reviewed by Dan Bates.
1489
1490         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/FlakyLayoutTestDetector.js:
1491         (FlakyLayoutTestDetector.prototype.flakinessExamples): If we have more than a certain number
1492         of examples, replace the middle items with a separator.
1493
1494         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
1495         (.flakiness-example-separator): Added styles for the separator.
1496
1497         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1498         (ViewController.prototype._domForPossiblyFlakyTests): Use a vertical ellipsis to represent
1499         the separator.
1500
1501 2011-07-09  Dirk Pranke  <dpranke@chromium.org>
1502
1503         nrwt: stack traces from worker-side exceptions aren't very useful inside test-webkitpy
1504         https://bugs.webkit.org/show_bug.cgi?id=64218
1505
1506         Reviewed by Eric Seidel.
1507
1508         Exceptions aren't picklable and can't be sent across the
1509         manager/worker message queue without losing information. NRWT
1510         handles this by turning the stack trace into a set of strings,
1511         and logging the strings when we receive an exception from the
1512         worker. However, when you are running tests and something
1513         crashes on the worker side, test-webkitpy prints the
1514         manager-side stack trace, which is just confusing and useless.
1515
1516         This patch changes the logic so that exceptions are passed
1517         through as-is when the worker and manager are in the same
1518         process (the --worker-model=inline option). This increases the
1519         code paths slightly but makes crashes much more useful.
1520
1521         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1522         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
1523         * Scripts/webkitpy/layout_tests/controllers/message_broker.py:
1524         * Scripts/webkitpy/layout_tests/controllers/worker.py:
1525         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1526
1527 2011-07-09  Darin Fisher  <darin@chromium.org>
1528
1529         Eliminate bad dependency on gfx::Point.
1530         https://bugs.webkit.org/show_bug.cgi?id=64228
1531
1532         Reviewed by Kent Tamura.
1533
1534         * DumpRenderTree/chromium/EventSender.cpp:
1535         (initMouseEvent):
1536
1537 2011-07-08  Ojan Vafai  <ojan@chromium.org>
1538
1539         pull static dashboard files into the appengine server from the chromium repository
1540         https://bugs.webkit.org/show_bug.cgi?id=64208
1541
1542         Reviewed by Dirk Pranke.
1543
1544         These files belong in the WebKit repo since they are tied primarily to the webkit tests.
1545         They have some extra bits to support chromium's gtests, but that seems fine.
1546         Mainly, this will allow other WebKit hackers to hack on the dashboards.
1547
1548         As a nice side-effect, we can now push the dashboard files when we do appengine pushes
1549         instead of the weird thing we used to do of pulling them from the Chromium repository
1550         and storing them in the appengine datastore. This allows for cleaning up a lot of code
1551         and will likely make the dashboards load a bit faster.
1552
1553         The new JS files don't fully match WebKit style, but I'd like to clean that up in a
1554         followup patch if possible to maintain my sanity with this patch.
1555
1556         * TestResultServer/app.yaml:
1557         * TestResultServer/handlers/dashboardhandler.py: Removed.
1558         * TestResultServer/handlers/menu.py:
1559         * TestResultServer/main.py:
1560         * TestResultServer/model/dashboardfile.py: Removed.
1561         * TestResultServer/static-dashboards/LICENSE.dygraph.txt: Added.
1562         * TestResultServer/static-dashboards/README.dygraph.txt: Added.
1563         * TestResultServer/static-dashboards/README.webtreemap.txt: Added.
1564         * TestResultServer/static-dashboards/aggregate_results.html: Added.
1565         * TestResultServer/static-dashboards/builders.js: Added.
1566         * TestResultServer/static-dashboards/dashboard_base.js: Added.
1567         * TestResultServer/static-dashboards/dygraph-combined.js: Added.
1568         * TestResultServer/static-dashboards/flakiness_dashboard.html: Added.
1569         * TestResultServer/static-dashboards/flakiness_dashboard_tests.js: Added.
1570         * TestResultServer/static-dashboards/timeline_explorer.html: Added.
1571         * TestResultServer/static-dashboards/treemap.html: Added.
1572         * TestResultServer/static-dashboards/webtreemap.css: Added.
1573         * TestResultServer/static-dashboards/webtreemap.js: Added.
1574         * TestResultServer/stylesheets/dashboardfile.css: Removed.
1575         * TestResultServer/stylesheets/menu.css:
1576         * TestResultServer/templates/dashboardfilelist.html: Removed.
1577         * TestResultServer/templates/menu.html:
1578
1579 2011-07-08  Adam Roben  <aroben@apple.com>
1580
1581         Make TestFailures's list of flaky tests look more like the list of non-flaky tests
1582
1583         Fixes <http://webkit.org/b/64204> TestFailures page's flaky tests list is ugly!
1584
1585         Reviewed by Daniel Bates.
1586
1587         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
1588         (#failure-history, #possibly-flaky-tests): Expanded this rule to apply to the list of flaky
1589         tests.
1590         (#failure-history > li, #possibly-flaky-tests > li): Ditto, but moved the 50px left padding
1591         from here...
1592         (#failure-history > li): ...to here.
1593
1594         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1595         (ViewController.prototype._domForPossiblyFlakyTests): Give the list an id attribute for
1596         styling purposes.
1597
1598 2011-07-08  Jeffrey Pfau  <jpfau@apple.com>
1599
1600         Unreviewed, add myself as committer.
1601
1602         * Scripts/webkitpy/common/config/committers.py:
1603
1604 2011-07-08  Adam Barth  <abarth@webkit.org>
1605
1606         TestResultsServer should keep old test results
1607         https://bugs.webkit.org/show_bug.cgi?id=64199
1608
1609         Reviewed by Ojan Vafai.
1610
1611         Having historical data will help us do failure archeology.
1612
1613         * TestResultServer/handlers/testfilehandler.py:
1614         * TestResultServer/model/testfile.py:
1615         * TestResultServer/templates/showfilelist.jsonp: Added.
1616
1617 2011-07-08  Adam Roben  <aroben@apple.com>
1618
1619         Remove commit-log-editor's dependency on Module::Load::Conditional
1620
1621         This module isn't available in Perl 5.8.8 (the version used on Leopard).
1622
1623         Fixes <http://webkit.org/b/64198> REGRESSION (r90583):
1624         webkitpy.common.checkout.checkout_unittest failing on Leopard
1625
1626         Reviewed by Daniel Bates.
1627
1628         * Scripts/commit-log-editor: Use the new loadTermReadKey() function instead of
1629         Module::Load::Conditional::can_load.
1630         (loadTermReadKey): Added. Tries to load Term::ReadKey and returns true if it's successful.
1631
1632 2011-07-08  Adam Roben  <aroben@apple.com>
1633
1634         Make checkout_unittest more robust against files moving around
1635
1636         Fixes <http://webkit.org/b/64197> checkout_unittest contains ugly, fragile code to find the
1637         Scripts directory
1638
1639         Reviewed by Adam Barth.
1640
1641         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
1642         (CommitMessageForThisCommitTest.test_commit_message_for_this_commit): Instantiate a real SCM
1643         object and use it to get the path to the Scripts directory, rather than hard-coding the
1644         relative path from this file to Scripts.
1645
1646 2011-07-08  Adam Roben  <aroben@apple.com>
1647
1648         Make TestFailures remember that run-webkit-tests timed out, even across reloads
1649
1650         Fixes <http://webkit.org/b/64193> TestFailures page incorrectly thinks all tests passed in
1651         http://build.webkit.org/builders/Windows%207%20Release%20(Tests)/builds/14589 after a reload
1652
1653         Reviewed by Daniel Bates.
1654
1655         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
1656         (LayoutTestResultsLoader.prototype.start): Store an "error" attribute in the cached data.
1657         When true, it indicates that there was an error retrieving the results for this build and
1658         that the errorCallback should be called.
1659
1660 2011-07-08  Adam Barth  <abarth@webkit.org>
1661
1662         Teach garden-o-matic how to display test results
1663         https://bugs.webkit.org/show_bug.cgi?id=64141
1664
1665         Reviewed by Ojan Vafai.
1666
1667         This patch includes basic infrastructure for probing build.chromium.org
1668         for test results.  We only handle text and image tests, not anything
1669         complicated like reftests.  Also, we're using the revision/build
1670         independent results store on the server, so we're avoiding that
1671         complication for now.
1672
1673         It's slightly hacky that we need to probe the server to see what kinds
1674         of results exist.  A better solution would be to add CORS support to
1675         the server or to use the local server to help.
1676
1677         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
1678         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
1679         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
1680         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
1681         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
1682         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
1683
1684 2011-07-08  Dirk Pranke  <dpranke@chromium.org>
1685
1686         REGRESSION(90419) NRWT's httpd locking is broken for --child-processes=1
1687         https://bugs.webkit.org/show_bug.cgi?id=64092
1688
1689         Reviewed by Tony Chang.
1690
1691         The code was incorrectly creating empty shards in the
1692         shard_in_two case.
1693
1694         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1695         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
1696
1697 2011-07-08  Adam Roben  <aroben@apple.com>
1698
1699         Teach buildbot to figure out how many webkitpy/webkitperl tests failed
1700
1701         Fixes <http://webkit.org/b/64192> It's hard to tell how many test-webkitpy/test-webkitperl
1702         tests failed when looking at build.webkit.org
1703
1704         Reviewed by Eric Seidel.
1705
1706         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1707         (TestWithFailureCount): New class that represents a test build step which has an associated
1708         failure count. Eventually we should move more of our test classes to deriving from this.
1709         (TestWithFailureCount.countFailures): Method for subclasses to override to say how many
1710         failures occurred.
1711
1712         (TestWithFailureCount.commandComplete):
1713         (TestWithFailureCount.evaluateCommand):
1714         (TestWithFailureCount.getText):
1715         (TestWithFailureCount.getText2):
1716         These were all based on RunGtkAPITests.
1717
1718         (RunPythonTests): Changed to inherit from TestWithFailureCount.
1719         (RunPythonTests.countFailures): Parses the test-webkitpy output looking for the count of
1720         failures.
1721         (RunPerlTests): Changed to inherit from TestWithFailureCount.
1722         (RunPerlTests.countFailures): Parses the test-webkitperl output looking for the count of
1723         failures.
1724
1725 2011-07-08  Adam Roben  <aroben@apple.com>
1726
1727         Ensure $CHANGE_LOG_EMAIL_ADDRESS is set when testing webkitpy's commit-log-editor integration
1728
1729         Fixes <http://webkit.org/b/64180> REGRESSION (r90564): test-webkitpy failing on multiple
1730         bots due to commit-log-editor errors
1731
1732         Reviewed by Adam Barth.
1733
1734         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
1735         (CommitMessageForThisCommitTest.test_commit_message_for_this_commit): Set
1736         $CHANGE_LOG_EMAIL_ADDRESS to the patch author's email address. This ensures that
1737         commit-log-editor can find a value for the committer's email, and that the committer and
1738         author email addresses match, which will prevent commit-log-editor from inserting a "Patch
1739         by" line in the commit message.
1740
1741 2011-07-08  Andreas Kling  <kling@webkit.org>
1742
1743         [Qt][WK2] Views should expose QActions for basic navigation.
1744         https://bugs.webkit.org/show_bug.cgi?id=64174
1745
1746         Reviewed by Benjamin Poulain.
1747
1748         Bring the toolbar in Qt's MiniBrowser back to life.
1749
1750         * MiniBrowser/qt/BrowserView.cpp:
1751         (BrowserView::navigationAction):
1752         * MiniBrowser/qt/BrowserView.h:
1753         * MiniBrowser/qt/BrowserWindow.cpp:
1754         (BrowserWindow::BrowserWindow):
1755
1756 2011-07-08  Adam Roben  <aroben@apple.com>
1757
1758         Teach TestFailures how to find test names in commit-log-editor-style commit messages
1759
1760         TestFailures was relying on Trac turning the list of modified files in our commit messages
1761         into an HTML list. But Trac only does this when the list of modified files is indented.
1762         commit-log-editor doesn't indent the file list, so the list wasn't being turned into an HTML
1763         list, which was confusing TestFailures.
1764
1765         TestFailures now does much simpler parsing of the commit message (i.e., just a substring
1766         search) without relying at all on its structure.
1767
1768         Fixes <http://webkit.org/b/64173> TestFailures page fails to blame r90608 for breaking
1769         fast/dom/HTMLProgressElement/progress-element-markup.html on Windows
1770
1771         Reviewed by David Kilzer.
1772
1773         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Trac.js:
1774         (Trac.prototype.getCommitDataForRevisionRange): Instead of trying to parse the commit
1775         message, just return its text.
1776
1777         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1778         (ViewController.prototype._domForRegressionRange): Instead of searching for test names in
1779         each commit's list of modified files, just search for test names anywhere in the commit's
1780         message.
1781
1782 2011-07-08  Adam Barth  <abarth@webkit.org>
1783
1784         sheriffbot is too spammy in IRC
1785         https://bugs.webkit.org/show_bug.cgi?id=64153
1786
1787         Reviewed by Eric Seidel.
1788
1789         Reporting failures in IRC worked well when the tree was greener than it
1790         is today.  Nowadays, this feature mostly just results in spam about
1791         false positives.  If we reach a greener state, we can bring this code
1792         back.
1793
1794         (Another possibility is to restrict this feature to builder bots,
1795         rather than including testers, as we did before this patch.)
1796
1797         * Scripts/webkitpy/tool/commands/sheriffbot.py:
1798         * Scripts/webkitpy/tool/commands/sheriffbot_unittest.py:
1799
1800 2011-07-08  Adam Roben  <aroben@apple.com>
1801
1802         Ensure commit-log-editor adds a "Patch by" line when the author and committer are different
1803
1804         Previously we were only adding a "Patch by" line when the ChangeLog contained a "Reviewed
1805         by" line. But some patches (like rollout patches) don't contain that line. Now we always add
1806         "Patch by" to the commit log regardless of the ChangeLog's contents.
1807
1808         Fixes <http://webkit.org/b/64127> Committer for r90588 is commit-queue@webkit.org, but
1809         should have been abarth@webkit.org
1810
1811         Reviewed by Anders Carlsson.
1812
1813         * Scripts/commit-log-editor:
1814         (createCommitMessage): Try to put the "Patch by" line just above the "Reviewed by" line, as
1815         before. If there is no "Reviewed by" line, try to put it just above the first modified file.
1816         If all else fails, put it at the end of the commit message.
1817         (patchAuthorshipString): Added. Code came from createCommitMessage.
1818
1819 2011-07-08  Herczeg Zoltan  <zherczeg@webkit.org>
1820
1821         Reviewed by Andreas Kling.
1822
1823         Adding myself as a reviewer.
1824
1825         * Scripts/webkitpy/common/config/committers.py:
1826
1827 2011-07-07  Kevin Ollivier  <kevino@theolliviers.com>
1828
1829         [wx] Unreviewed build fix, remove old files from the tree and update paths to new ones.
1830
1831         * wx/build: Removed.
1832         * wx/build/build_utils.py: Removed.
1833         * wx/build/settings.py: Removed.
1834         * wx/build/waf_extensions.py: Removed.
1835         * wx/build/wxpresets.py: Removed.
1836         * wx/packaging/build-mac-installer.py:
1837
1838 2011-07-07  Kevin Ollivier  <kevino@theolliviers.com>
1839
1840         [wx] Unreviewed build fix, more fixes for install name issues.
1841
1842         * wx/packaging/build-mac-installer.py:
1843
1844 2011-07-07  Dirk Pranke  <dpranke@chromium.org>
1845
1846         test-webkitpy fails on chromium win
1847         https://bugs.webkit.org/show_bug.cgi?id=64137
1848
1849         Reviewed by Eric Seidel.
1850
1851         test-webkitpy doesn't play nicely with the multiprocessing
1852         module on win32. An earlier change actually reenabled the tests
1853         on win32 by mistake.
1854
1855         This patch also fixes a few cases where path names will trip
1856         things up on win32 (e.g., by testing unix paths on windows).
1857         We do not lose any real coverage here with those fixes.
1858
1859         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
1860         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
1861         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
1862         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1863
1864 2011-07-07  Leandro Pereira  <leandro@profusion.mobi>
1865
1866         [ImageDiff] Calculate/print difference right after reading baseline image.
1867         https://bugs.webkit.org/show_bug.cgi?id=64117
1868
1869         Reviewed by Kent Tamura.
1870
1871         Otherwise, ImageDiff will block on fgets() until the universe ends or you ^C
1872         it.  Whichever happens first.
1873
1874         * DumpRenderTree/efl/ImageDiff.cpp:
1875         (main):
1876
1877 2011-07-07  Raphael Kubo da Costa  <kubo@profusion.mobi>
1878
1879         [EFL] DRT: Leak cairo_t to make the pixel tests stop crashing
1880         https://bugs.webkit.org/show_bug.cgi?id=64107
1881
1882         Reviewed by Kent Tamura.
1883
1884         The pointer is later managed by BitmapContext, but it was being
1885         de-refed and deleted earlier by our RefPtr, causing crashes in all
1886         pixel tests.
1887
1888         We now call leakRef() to make sure it is not removed when our
1889         createBitmapContextFromWebView goes out of context.
1890
1891         * DumpRenderTree/efl/PixelDumpSupportEfl.cpp:
1892         (createBitmapContextFromWebView):
1893
1894 2011-07-07  Adam Roben  <aroben@apple.com>
1895
1896         Update TestFailures's title and header immediately upon navigation
1897
1898         Fixes <http://webkit.org/b/64125> TestFailures page seems slow to react on link clicks
1899
1900         Reviewed by Daniel Bates.
1901
1902         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1903         (ViewController.prototype.loaded): Create and store an h1 element for displaying the page's
1904         title.
1905         (ViewController.prototype._displayBuilder): Set the page's title and clear out any old
1906         content immediately instead of waiting for the first set of results for the history
1907         analyzer. As we receive new results from the analyzer we'll just clear out the main content
1908         area and replace it with the new info.
1909         (ViewController.prototype._displayTesters): Set the page's title and clear out any old
1910         content immediately instead of waiting for the list of testers from the buildbot. Sprinkled
1911         in a little use of our removeAllChildren helper, too.
1912         (ViewController.prototype._setTitle): Added. Set's the page's title and the header text.
1913
1914 2011-07-07  Adam Roben  <aroben@apple.com>
1915
1916         Make Checkout use SCM's Executive instead of conjuring up its own
1917
1918         This will improve integration with the rest of webkitpy, particularly when invoked via
1919         webkit-patch.
1920
1921         Fixes <http://webkit.org/b/64115> REGRESSION (r90564): webkitpy's Checkout class uses
1922         Executive inappropriately
1923
1924         Reviewed by Adam Barth.
1925
1926         * Scripts/webkitpy/common/checkout/checkout.py:
1927         (Checkout.commit_message_for_this_commit): Use SCM.run instead of creating an Executive for
1928         our own use. SCM might have some super-special Executive that it uses under the covers, and
1929         we want to use it, too!
1930
1931         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
1932         (CommitMessageForThisCommitTest.test_commit_message_for_this_commit): Mock the SCM.run
1933         method to call through to Executive.run_command.
1934
1935 2011-07-07  Adam Roben  <aroben@apple.com>
1936
1937         Make Term/ReadKey.pm an optional dependency of commit-log-editor
1938
1939         When Term/ReadKey.pm isn't present, the --regenerate-log switch will be non-functional.
1940
1941         Fixes <http://webkit.org/b/64113> REGRESSION (r90564):
1942         webkitpy.common.checkout.checkout_unittest failing on GTK bots due to missing
1943         Term/ReadKey.pm module
1944
1945         Reviewed by Xan Lopez.
1946
1947         * Scripts/commit-log-editor: Use Module::Load::Conditional::can_load to only load
1948         Term/ReadKey.pm if it's present. If it isn't present, ignore the --regenerate-log switch so
1949         that existing commit messages will be preserved. (Users can manually remove existing commit
1950         messages using git-reset.)
1951
1952 2011-07-07  Adam Roben  <aroben@apple.com>
1953
1954         Keep testing that commit messages containing Unicode are handled properly by webkitpy
1955
1956         Fixes <http://webkit.org/b/64109> REGRESSION (r90571): test-webkitpy no longer tests that we
1957         correctly handle Unicode in commit messages
1958
1959         Reviewed by Eric Seidel.
1960
1961         * Scripts/webkitpy/common/checkout/checkout_unittest.py: Put some Unicode characters in
1962         places in the ChangeLog that will end up being included in the commit message.
1963
1964 2011-07-07  Andreas Kling  <kling@webkit.org>
1965
1966         [Qt][WK2] Remove Symbian code.
1967         https://bugs.webkit.org/show_bug.cgi?id=64101
1968
1969         Reviewed by Benjamin Poulain.
1970
1971         * MiniBrowser/qt/BrowserWindow.cpp:
1972         (BrowserWindow::BrowserWindow):
1973         (BrowserWindow::screenshot):
1974         * MiniBrowser/qt/MiniBrowser.pro:
1975         * MiniBrowser/qt/MiniBrowserApplication.h:
1976         (WindowOptions::WindowOptions):
1977
1978 2011-07-07  Andreas Kling  <kling@webkit.org>
1979
1980         [Qt][WK2] Don't support ridiculous matrix of QT_NO_FEATURE combinations.
1981         https://bugs.webkit.org/show_bug.cgi?id=64099
1982
1983         Reviewed by Benjamin Poulain.
1984
1985         * MiniBrowser/qt/BrowserWindow.cpp:
1986         (BrowserWindow::openFile):
1987         (BrowserWindow::screenshot):
1988         (BrowserWindow::loadURLListFromFile):
1989
1990 2011-07-07  Adam Roben  <aroben@apple.com>
1991
1992         Completely ignore too-many-failures builds in TestFailures in most circumstances
1993
1994         Because a semi-arbitrary subset of tests are run in a too-many-failures build, we can't
1995         really use them to perform regression analysis. The only time we want to pay attention to
1996         too-many-failures builds is when we're trying to explain when the current bout of
1997         too-many-failures started.
1998
1999         Fixes <http://webkit.org/b/64106> TestFailures page sometimes claims a test started failing
2000         in a build that didn't even run it (because it exited early due to too many failues)
2001
2002         Reviewed by David Kilzer.
2003
2004         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
2005         (LayoutTestHistoryAnalyzer.prototype._incorporateBuildHistory): Removed old, broken
2006         too-many-failures handling that would cause us to blame builds that didn't even run a given
2007         test for breaking it. Instead, skip over all too-many-failures builds unless the most recent
2008         build was itself a too-many-failures build.
2009
2010 2011-07-07  Adam Roben  <aroben@apple.com>
2011
2012         Teach webkitpy's Checkout class to use commit-log-editor to create commit messages
2013
2014         Fixes <http://webkit.org/b/26755> webkit-patch's commit messages are less readable than
2015         commit-log-editor's
2016
2017         Reviewed by David Kilzer.
2018
2019         * Scripts/webkitpy/common/checkout/checkout.py:
2020         (Checkout.commit_message_for_this_commit): Run commit-log-editor, passing it the paths of
2021         the modified ChangeLogs, to generate the commit message, rather than trying to generate one
2022         ourselves.
2023
2024         * Scripts/webkitpy/common/checkout/checkout_unittest.py: Updated the expected commit message
2025         to match commit-log-editor's format.
2026         (CommitMessageForThisCommitTest.setUp): Write the ChangeLogs into Tools and LayoutTests
2027         directories so we can see how the various entries get labeled in the commit message.
2028         (CommitMessageForThisCommitTest.test_commit_message_for_this_commit): Create a mock SCM
2029         instance that knows how to find commit-log-editor and pass it to our Checkout instance.
2030         Don't bother capturing output, since there shouldn't be any.
2031
2032 2011-07-07  Adam Roben  <aroben@apple.com>
2033
2034         Add a --print-log option to commit-log-editor
2035
2036         When specified, commit-log-editor takes a set of ChangeLog paths on the command line,
2037         generates a commit message from those ChangeLogs, prints it to stdout, and exits.
2038
2039         Prep work for fixing <http://webkit.org/b/26755> webkit-patch's commit messages are less
2040         readable than commit-log-editor's
2041
2042         Reviewed by David Kilzer.
2043
2044         * Scripts/commit-log-editor: Changed to use Getopt::Long to parse options. Added --print-log
2045         option, which calls through to createCommitMessage, prints the result, and exits.
2046         (printUsageAndExit): Renamed from "usage" for clarity. Beefed up the usage statement to
2047         explain commit-log-editor's 2.5 (normal, --print-log, and --help) modes of operation.
2048
2049 2011-07-07  Adam Roben  <aroben@apple.com>
2050
2051         Extract commit-log-editor's commit-message-generation code into a separate function
2052
2053         As a bonus, we now skip a bunch of work in the case where we are generating a commit message
2054         for staged git changes.
2055
2056         Prep work for fixing <http://webkit.org/b/26755> webkit-patch's commit messages are less
2057         readable than commit-log-editor's
2058
2059         Reviewed by David Kilzer.
2060
2061         * Scripts/commit-log-editor: Moved the default definition of $endl up toward the top of the
2062         file so that it will be defined even if we aren't operating on an existing log message.
2063         Moved a bunch of top-level code to generate the commit message from here...
2064         (createCommitMessage): ...to here.
2065
2066 2011-07-07  Carlos Garcia Campos  <cgarcia@igalia.com>
2067
2068         [GTK] Open links in a new window when clicking with the middle button in MiniBrowser
2069         https://bugs.webkit.org/show_bug.cgi?id=63988
2070
2071         Reviewed by Martin Robinson.
2072
2073         * MiniBrowser/gtk/BrowserWindow.c:
2074         (browserWindowConstructed):
2075         (decidePolicyForNavigationAction): Ignore the action if a link is
2076         clicked with the middle buttonm and open the link in a new window.
2077         (browserWindowPolicyClientInit):
2078
2079 2011-07-07  Adam Barth  <abarth@webkit.org>
2080
2081         Remove style_references.py
2082         https://bugs.webkit.org/show_bug.cgi?id=64070
2083
2084         Reviewed by Eric Seidel.
2085
2086         We never really adopted the _references design, and this file is a bit
2087         of an orphan at this point.
2088
2089         * Scripts/check-webkit-style:
2090         * Scripts/webkitpy/style/checker.py:
2091         * Scripts/webkitpy/style/checker_unittest.py:
2092         * Scripts/webkitpy/style/checkers/python.py:
2093         * Scripts/webkitpy/style/checkers/test_expectations.py:
2094         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2095         * Scripts/webkitpy/style/main_unittest.py:
2096         * Scripts/webkitpy/style_references.py: Removed.
2097
2098 2011-07-07  Eric Seidel  <eric@webkit.org>
2099
2100         Fix WebKit2 expected results search paths for Mac and Qt under new-run-webkit-tests
2101         https://bugs.webkit.org/show_bug.cgi?id=64056
2102
2103         Unreviewed.
2104
2105         I had written this unittest prior to landing my change, but
2106         had named it qt_unitest.py, so it wasn't running (nor was
2107         it added to my git repo).
2108
2109         Once I fixe the name of the unittest file, it was very easy
2110         to fix the error in webkit.py (which already had a FIXME).
2111
2112         * Scripts/webkitpy/layout_tests/port/qt_unittest.py: Added.
2113         * Scripts/webkitpy/layout_tests/port/webkit.py:
2114
2115 2011-07-07  Eric Seidel  <eric@webkit.org>
2116
2117         Unreviewed.  Add more debug logging about which test expectations we're using.
2118
2119         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2120         * Scripts/webkitpy/layout_tests/port/webkit.py:
2121
2122 2011-07-06  Adam Barth  <abarth@webkit.org>
2123
2124         Remove unexpected_results.json
2125         https://bugs.webkit.org/show_bug.cgi?id=64065
2126
2127         Reviewed by Eric Seidel.
2128
2129         No one should be using this file any more.  This patch removes it from
2130         the face of the earth.  This patch also removes two NRWT-specific
2131         command line options that were implemented using
2132         unexpected_results.json.
2133
2134         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
2135         * Scripts/webkitpy/layout_tests/controllers/manager.py:
2136         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2137         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2138         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2139         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
2140         * Scripts/webkitpy/tool/servers/rebaselineserver_unittest.py:
2141
2142 2011-07-07  Adam Barth  <abarth@webkit.org>
2143
2144         REGRESSION(90520): chromium GPU tests produce output in the LayoutTests directory
2145         https://bugs.webkit.org/show_bug.cgi?id=64067
2146
2147         Reviewed by Eric Seidel.
2148
2149         The author of r90520 forgot to change the chromium_gpu.py implemenation
2150         of this method as well.
2151
2152         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2153
2154 2011-07-06  Eric Seidel  <eric@webkit.org>
2155
2156         Fix WebKit2 expected results search paths for Mac and Qt under new-run-webkit-tests
2157         https://bugs.webkit.org/show_bug.cgi?id=64056
2158
2159         Reviewed by Adam Barth.
2160
2161         This fixes the MacPort's version of baseline_search_path to include
2162         wk2 fallback like WebKitPort's does.  This also re-works the QtPort
2163         to set self._name (as other ports do) so that WebKitPort's default
2164         implementations will understand Qt operating system flavors correctly
2165         (including when running wk2 tests).
2166
2167         In trying to test this, I realized that nearly all of our port-tests
2168         use the default FileSystem(), User() and Executive() objects, which
2169         is really bad, because they expose details of your system in the unittest results!
2170
2171         Once I fixed many of the Port() calls to pass MockFileSystem(), then
2172         MacPort.test_tests_for_other_platforms started failing, and I had to
2173         teach MockFileSystem.glob how to handle directories to make it work again.
2174
2175         I removed the useless PortTestCase.test_baseline_search_path and replaced
2176         it with more useful test_baseline_search_path tests in MacPortTest and QtPortTest.
2177
2178         * Scripts/webkitpy/common/config/ports.py:
2179         * Scripts/webkitpy/common/system/filesystem_mock.py:
2180         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2181         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2182         * Scripts/webkitpy/layout_tests/port/mac.py:
2183         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
2184         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
2185         * Scripts/webkitpy/layout_tests/port/qt.py:
2186         * Scripts/webkitpy/layout_tests/port/test_files.py:
2187         * Scripts/webkitpy/layout_tests/port/win.py:
2188         * Scripts/webkitpy/to_be_moved/deduplicate_tests.py:
2189         * Scripts/webkitpy/tool/mocktool.py:
2190
2191 2011-07-06  Adam Barth  <abarth@webkit.org>
2192
2193         Wean rebaseline-server off unexpected_results.json
2194         https://bugs.webkit.org/show_bug.cgi?id=64063
2195
2196         Reviewed by Eric Seidel.
2197
2198         In the process of changing this code to use full_results.json, I
2199         noticed that the code was broken (because it wasn't tested).  This
2200         patch also adds test coverage for the broken code.
2201
2202         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
2203         * Scripts/webkitpy/tool/servers/rebaselineserver.py:
2204         * Scripts/webkitpy/tool/servers/rebaselineserver_unittest.py:
2205
2206 2011-07-06  Adam Barth  <abarth@webkit.org>
2207
2208         Wean resultsjsonparser off unexpected_results.json
2209         https://bugs.webkit.org/show_bug.cgi?id=64061
2210
2211         Reviewed by Eric Seidel.
2212
2213         We're about to delete unexpected_results.json, so we need to remove all
2214         the users.
2215
2216         * Scripts/webkitpy/common/config/ports.py:
2217         * Scripts/webkitpy/common/net/resultsjsonparser.py:
2218         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
2219         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2220
2221 2011-07-06  Adam Barth  <abarth@webkit.org>
2222
2223         Cleanup result_summary.py
2224         https://bugs.webkit.org/show_bug.cgi?id=64057
2225
2226         Reviewed by Eric Seidel.
2227
2228         This class had a bunch of out-of-date docstrings that no longer make
2229         any sense.
2230
2231         * Scripts/webkitpy/layout_tests/models/result_summary.py:
2232
2233 2011-07-06  Adam Barth  <abarth@webkit.org>
2234
2235         Tweak the ChangeLog for DEPS rolls to say "Unreviewed" so that the
2236         patches can be landed by the commit-queue.
2237
2238         * Scripts/webkitpy/tool/steps/preparechangelogfordepsroll.py:
2239
2240 2011-07-06  Adam Barth  <abarth@webkit.org>
2241
2242         Normalize callers of logging.getLogger to use __name__
2243         https://bugs.webkit.org/show_bug.cgi?id=64054
2244
2245         Reviewed by Eric Seidel.
2246
2247         Most callers were already using __name__.  This patch makes everyone
2248         consistent.
2249
2250         * Scripts/webkitpy/common/checkout/diff_parser.py:
2251         * Scripts/webkitpy/common/checkout/scm/svn.py:
2252         * Scripts/webkitpy/common/net/statusserver.py:
2253         * Scripts/webkitpy/common/system/executive.py:
2254         * Scripts/webkitpy/common/system/file_lock.py:
2255         * Scripts/webkitpy/common/system/user.py:
2256         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
2257         * Scripts/webkitpy/layout_tests/models/result_summary.py:
2258         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2259         * Scripts/webkitpy/layout_tests/port/chromium.py:
2260         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2261         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
2262         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2263         * Scripts/webkitpy/layout_tests/port/gtk.py:
2264         * Scripts/webkitpy/layout_tests/port/http_lock.py:
2265         * Scripts/webkitpy/layout_tests/port/mac.py:
2266         * Scripts/webkitpy/layout_tests/port/qt.py:
2267         * Scripts/webkitpy/layout_tests/port/server_process.py:
2268         * Scripts/webkitpy/layout_tests/port/webkit.py:
2269         * Scripts/webkitpy/layout_tests/port/win.py:
2270         * Scripts/webkitpy/layout_tests/servers/apache_http_server.py:
2271         * Scripts/webkitpy/layout_tests/servers/http_server.py:
2272         * Scripts/webkitpy/layout_tests/servers/http_server_base.py:
2273         * Scripts/webkitpy/layout_tests/servers/websocket_server.py:
2274         * Scripts/webkitpy/layout_tests/views/printing.py:
2275         * Scripts/webkitpy/python24/versioning.py:
2276         * Scripts/webkitpy/style/checker.py:
2277         * Scripts/webkitpy/style/checkers/test_expectations.py:
2278         * Scripts/webkitpy/style/patchreader.py:
2279         * Scripts/webkitpy/to_be_moved/update_webgl_conformance_tests.py:
2280
2281 2011-07-06  Adam Barth  <abarth@webkit.org>
2282
2283         Move model classes out of layout_package and into models
2284         https://bugs.webkit.org/show_bug.cgi?id=64053
2285
2286         Reviewed by Eric Seidel.
2287
2288         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2289         * Scripts/webkitpy/common/net/layouttestresults.py:
2290         * Scripts/webkitpy/common/net/layouttestresults_unittest.py:
2291         * Scripts/webkitpy/common/net/resultsjsonparser.py:
2292         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
2293         * Scripts/webkitpy/layout_tests/controllers/manager.py:
2294         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2295         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py:
2296         * Scripts/webkitpy/layout_tests/controllers/worker.py:
2297         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
2298         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
2299         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
2300         * Scripts/webkitpy/layout_tests/layout_package/result_summary.py: Removed.
2301         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py: Removed.
2302         * Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py: Removed.
2303         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py: Removed.
2304         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py: Removed.
2305         * Scripts/webkitpy/layout_tests/layout_package/test_input.py: Removed.
2306         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py:
2307         * Scripts/webkitpy/layout_tests/layout_package/test_results.py: Removed.
2308         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py: Removed.
2309         * Scripts/webkitpy/layout_tests/models: Added.
2310         * Scripts/webkitpy/layout_tests/models/__init__.py: Added.
2311         * Scripts/webkitpy/layout_tests/models/result_summary.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/result_summary.py.
2312         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_expectations.py.
2313         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py.
2314         * Scripts/webkitpy/layout_tests/models/test_failures.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_failures.py.
2315         * Scripts/webkitpy/layout_tests/models/test_failures_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py.
2316         * Scripts/webkitpy/layout_tests/models/test_input.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_input.py.
2317         * Scripts/webkitpy/layout_tests/models/test_results.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_results.py.
2318         * Scripts/webkitpy/layout_tests/models/test_results_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py.
2319         * Scripts/webkitpy/layout_tests/port/chromium.py:
2320         * Scripts/webkitpy/layout_tests/views/printing.py:
2321         * Scripts/webkitpy/layout_tests/views/printing_unittest.py:
2322         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2323         * Scripts/webkitpy/style_references.py:
2324         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests.py:
2325         * Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
2326         * Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py:
2327         * Scripts/webkitpy/tool/commands/queues_unittest.py:
2328         * Scripts/webkitpy/tool/commands/rebaseline.py:
2329
2330 2011-07-06  Dirk Pranke  <dpranke@chromium.org>
2331
2332         nrwt: fix typo preventing http startup on windows
2333         https://bugs.webkit.org/show_bug.cgi?id=64050
2334
2335         Reviewed by Eric Siedel.
2336
2337         NRWT checks whether it needs to start the servers by looking
2338         for "/http/" in the test names. After r90520, the leading slash
2339         is not present, so the check isn't quite right.
2340
2341         * Scripts/webkitpy/layout_tests/controllers/manager.py:
2342         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2343
2344 2011-07-06  Dirk Pranke  <dpranke@chromium.org>
2345
2346         nrwt: remove --use-apache from the command line
2347         https://bugs.webkit.org/show_bug.cgi?id=63358
2348
2349         Reviewed by Adam Barth.
2350
2351         Re-land the fix in bug 63358 - the prior fix was flawed in that
2352         we were trying to start the http server during check_sys_deps()
2353         but not actually setting up the environment properly. This
2354         broken prior to this change, but exposed by it, since the patch
2355         changes the chromium port to call the base class's checks.
2356
2357         * Scripts/webkitpy/common/system/executive.py:
2358         * Scripts/webkitpy/layout_tests/port/base.py:
2359         * Scripts/webkitpy/layout_tests/port/chromium.py:
2360         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2361         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2362         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2363         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
2364         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
2365         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2366
2367 2011-07-06  Adam Barth  <abarth@webkit.org>
2368
2369         Move view-related code out of layout_package and into views
2370         https://bugs.webkit.org/show_bug.cgi?id=64051
2371
2372         Reviewed by Eric Seidel.
2373
2374         * Scripts/webkitpy/layout_tests/controllers/manager.py:
2375         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2376         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
2377         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py:
2378         * Scripts/webkitpy/layout_tests/layout_package/metered_stream.py: Removed.
2379         * Scripts/webkitpy/layout_tests/layout_package/metered_stream_unittest.py: Removed.
2380         * Scripts/webkitpy/layout_tests/layout_package/printing.py: Removed.
2381         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py: Removed.
2382         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2383         * Scripts/webkitpy/layout_tests/views: Added.
2384         * Scripts/webkitpy/layout_tests/views/__init__.py: Added.
2385         * Scripts/webkitpy/layout_tests/views/metered_stream.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/metered_stream.py.
2386         * Scripts/webkitpy/layout_tests/views/metered_stream_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/metered_stream_unittest.py.
2387         * Scripts/webkitpy/layout_tests/views/printing.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/printing.py.
2388         * Scripts/webkitpy/layout_tests/views/printing_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py.
2389
2390 2011-07-06  Adam Barth  <abarth@webkit.org>
2391
2392         Add goofy comments to explain why we have these files.
2393
2394         * Scripts/webkitpy/layout_tests/controllers/__init__.py:
2395         * Scripts/webkitpy/layout_tests/layout_package/__init__.py:
2396
2397 2011-07-06  Adam Barth  <abarth@webkit.org>
2398
2399         Move a bunch of classes out of layout_tests.layout_package and into layout_tests.controllers
2400         https://bugs.webkit.org/show_bug.cgi?id=64048
2401
2402         Reviewed by Eric Seidel.
2403
2404         * Scripts/webkitpy/layout_tests/controllers: Added.
2405         * Scripts/webkitpy/layout_tests/controllers/__init__.py: Added.
2406         * Scripts/webkitpy/layout_tests/controllers/manager.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/manager.py.
2407         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py.
2408         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py.
2409         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py.
2410         * Scripts/webkitpy/layout_tests/controllers/message_broker.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/message_broker.py.
2411         * Scripts/webkitpy/layout_tests/controllers/message_broker_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py.
2412         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py.
2413         * Scripts/webkitpy/layout_tests/controllers/worker.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/worker.py.
2414         * Scripts/webkitpy/layout_tests/layout_package/manager.py: Removed.
2415         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py: Removed.
2416         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py: Removed.
2417         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py: Removed.
2418         * Scripts/webkitpy/layout_tests/layout_package/message_broker.py: Removed.
2419         * Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py: Removed.
2420         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
2421         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py: Removed.
2422         * Scripts/webkitpy/layout_tests/layout_package/worker.py: Removed.
2423         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2424
2425 2011-07-06  Adam Barth  <abarth@webkit.org>
2426
2427         Attempt to fix unittest on Leopard.  Let's enumerate the keys in sorted
2428         order so the test results are predictable.
2429
2430         * Scripts/webkitpy/tool/bot/irc_command.py:
2431         * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:
2432
2433 2011-07-06  Dirk Pranke  <dpranke@chromium.org>
2434
2435         Reviewed by Eric Siedel.
2436
2437         nrwt: make the code be consistent about using test names instead of filenames or paths
2438
2439         https://bugs.webkit.org/show_bug.cgi?id=63597
2440
2441         It can be hard to figure out where in the code we're assuming
2442         test names are listed as unix-style relative filenames and where
2443         they are either absolute paths or relative paths following the
2444         host filesystem convention.
2445
2446         This patch changes things so that everything outside of the
2447         Port object uses (and must assume) unix-style relative
2448         filenames (with one exception, which is specifying host-local
2449         filenames as a list of test arguments on the command line).
2450
2451         This should make things clearer and more consistent, and also
2452         removes a lot of path conversion calls.
2453
2454         The changes in this patch outside of port/base.py are largely
2455         cosmetic changes of "filename" to "test". We add a few routines
2456         to the base Port implementation to manipulate test names to
2457         split them, figure out if we're referring to a directory of
2458         tests, etc.
2459
2460         * Scripts/webkitpy/common/net/layouttestresults.py:
2461         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
2462         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
2463         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
2464         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
2465         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
2466         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
2467         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
2468         * Scripts/webkitpy/layout_tests/layout_package/result_summary.py:
2469         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
2470         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
2471         * Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
2472         * Scripts/webkitpy/layout_tests/layout_package/test_input.py:
2473         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py:
2474         * Scripts/webkitpy/layout_tests/layout_package/test_results.py:
2475         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:
2476         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
2477         * Scripts/webkitpy/layout_tests/port/base.py:
2478         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2479         * Scripts/webkitpy/layout_tests/port/chromium.py:
2480         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2481         * Scripts/webkitpy/layout_tests/port/dryrun.py:
2482         * Scripts/webkitpy/layout_tests/port/test.py:
2483         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2484         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2485         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests.py:
2486         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests_unittest.py:
2487         * Scripts/webkitpy/tool/bot/flakytestreporter.py:
2488         * Scripts/webkitpy/tool/commands/rebaseline.py:
2489         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
2490
2491 2011-07-06  Chris Guillory  <ctguil@chromium.org>
2492
2493         No review necessary.
2494
2495         Adding myself to committers list in committers.py.
2496         https://bugs.webkit.org/show_bug.cgi?id=63478
2497
2498         * Scripts/webkitpy/common/config/committers.py:
2499
2500 2011-07-06  Johnny Ding  <jnd@chromium.org>
2501
2502         https://bugs.webkit.org/show_bug.cgi?id=63857
2503         Implement createPopupMenu for Chromium DRT.
2504
2505         Reviewed by Tony Chang.
2506
2507         * DumpRenderTree/chromium/WebViewHost.cpp:
2508         (WebViewHost::createPopupMenu):
2509         (WebViewHost::~WebViewHost):
2510         * DumpRenderTree/chromium/WebViewHost.h:
2511
2512 2011-07-06  Mihai Parparita  <mihaip@chromium.org>
2513
2514         Unreviewed, rolling out r90503.
2515         http://trac.webkit.org/changeset/90503
2516         https://bugs.webkit.org/show_bug.cgi?id=63358
2517
2518         Makes Chromium Windows canary bots hang.
2519
2520         * Scripts/webkitpy/layout_tests/port/base.py:
2521         * Scripts/webkitpy/layout_tests/port/chromium.py:
2522         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2523         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2524         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2525         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
2526         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
2527         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2528
2529 2011-07-06  Adam Barth  <abarth@webkit.org>
2530
2531         Move buildbot URL in webkitpy to config.urls
2532         https://bugs.webkit.org/show_bug.cgi?id=64041
2533
2534         Reviewed by Eric Seidel.
2535
2536         This patch prepares for using the buildbot class to talk to the
2537         Chromium buildbots.
2538
2539         * Scripts/webkitpy/common/config/urls.py:
2540         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
2541         * Scripts/webkitpy/tool/commands/download_unittest.py:
2542         * Scripts/webkitpy/tool/commands/queries.py:
2543         * Scripts/webkitpy/tool/mocktool.py:
2544         * Scripts/webkitpy/tool/steps/ensurebuildersaregreen.py:
2545
2546 2011-07-06  Kevin Ollivier  <kevino@theolliviers.com>
2547
2548         [wx] Unreviewed build fix, add the correct path to the waf build settings.
2549         
2550         * wx/packaging/build-win-installer.py:
2551
2552 2011-07-06  Adam Barth  <abarth@webkit.org>
2553
2554         Add missing include.
2555
2556         * Scripts/webkitpy/tool/bot/irc_command.py:
2557
2558 2011-07-06  Adam Barth  <abarth@webkit.org>
2559
2560         Add roll-chromium-deps command to sheriff-bot
2561         https://bugs.webkit.org/show_bug.cgi?id=64037
2562
2563         Reviewed by Eric Seidel.
2564
2565         Lo, and the DEPS were rolled on command.
2566
2567         * Scripts/webkitpy/tool/bot/irc_command.py:
2568         * Scripts/webkitpy/tool/bot/irc_command_unittest.py:
2569         * Scripts/webkitpy/tool/bot/sheriff.py:
2570         * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:
2571         * Scripts/webkitpy/tool/commands/roll.py:
2572         * Scripts/webkitpy/tool/commands/roll_unittest.py:
2573         * Scripts/webkitpy/tool/steps/updatechromiumdeps.py:
2574
2575 2011-07-06  Eric Seidel  <eric@webkit.org>
2576
2577         Split Driver, DriverOutput and DriverInput out into a new driver.py file
2578         https://bugs.webkit.org/show_bug.cgi?id=64031
2579
2580         Reviewed by Adam Barth.
2581
2582         This ended up kinda large and I'm happy to break it up if desired.
2583
2584         As part of cleaning up the Port module, I split Driver, DriverInput and DriverOutput
2585         out into its own driver.py file.
2586
2587         I also cleaned up how we import Port objects, by making port/__init__.py
2588         expose Port, Driver, DriverInput, and DriverOutput.  I removed exposure of
2589         some various test objects as they were causing circular dependencies.
2590         base.py is now an implementation detail of port.  No longer is "base" referenced
2591         anywhere in the code.
2592
2593         I also made Driver.__init__ non-virtual so that sub-classers could all share
2594         the data storage on Driver.  This more closely matches how other subclasses
2595         work throughout our python code.
2596
2597         Standard line-unwrapping and doc-string removal as I went through the code.
2598
2599         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2600         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
2601         * Scripts/webkitpy/layout_tests/port/__init__.py:
2602         * Scripts/webkitpy/layout_tests/port/base.py:
2603         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2604         * Scripts/webkitpy/layout_tests/port/chromium.py:
2605         * Scripts/webkitpy/layout_tests/port/driver.py: Added.
2606         * Scripts/webkitpy/layout_tests/port/dryrun.py:
2607         * Scripts/webkitpy/layout_tests/port/mock_drt.py:
2608         * Scripts/webkitpy/layout_tests/port/test.py:
2609         * Scripts/webkitpy/layout_tests/port/webkit.py:
2610         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2611         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests_unittest.py:
2612         * Scripts/webkitpy/tool/servers/rebaselineserver_unittest.py:
2613
2614 2011-07-06  Adam Roben  <aroben@apple.com>
2615
2616         Clear TestFailures's getNumberOfFailingTests cache
2617
2618         Now that we support NRWT, we need to clear out old zero values for the NRWT bots that might
2619         have been cached by old versions of TestFailures.
2620
2621         Fixes <http://webkit.org/b/64029> TestFailures page still doesn't show NRWT bots if you've
2622         used it recently
2623
2624         Reviewed by Adam Barth.
2625
2626         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
2627         (Builder.prototype.getNumberOfFailingTests): Added a .version property to the cache data
2628         that must match the current version. Otherwise we'll ignore the cached data.
2629
2630 2011-06-24  Dirk Pranke  <dpranke@chromium.org>
2631
2632         Reviewed by Adam Barth.
2633
2634         nrwt: remove --use-apache from the command line
2635         https://bugs.webkit.org/show_bug.cgi?id=63358
2636
2637         This change removes the --use-apache command line argument. It
2638         was initially put in when we were trying to get the cygwin
2639         apache instance to work with Chromium win, but that code has
2640         bitrotted and doesn't work at all now.
2641
2642         Arguably we should remove all of the code to allow for a choice
2643         of web servers, but since we may still want to switch off of
2644         LigHTTPd at some point on Windows, I'll leave the rest in for
2645         now (or at least yank it in a different change.)
2646
2647         * Scripts/webkitpy/layout_tests/port/base.py:
2648         * Scripts/webkitpy/layout_tests/port/chromium.py:
2649         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2650         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2651         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
2652         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2653
2654 2011-07-06  Dimitri Glazkov  <dglazkov@chromium.org>
2655
2656         garden-o-matic should use a favicon to indicate current state.
2657         https://bugs.webkit.org/show_bug.cgi?id=64027
2658
2659         Reviewed by Adam Barth.
2660
2661         * Scripts/webkitpy/tool/servers/data/gardeningserver/favicon-green.png: Added.
2662         * Scripts/webkitpy/tool/servers/data/gardeningserver/favicon-red.png: Added.
2663         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html: Added blurb about icons.
2664         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js: Added setting icon state.
2665         * Scripts/webkitpy/tool/servers/gardeningserver.py: Allowed the use of favicons.
2666
2667 2011-07-06  Daniel Bates  <dbates@rim.com>
2668
2669         Fix webkitpy error after changeset 90483 <http://trac.webkit.org/changeset/90483>; Add missing
2670         parameter to _fetch_file_from_results lamdba function.
2671
2672         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2673
2674 2011-07-06  Kevin Ollivier  <kevino@theolliviers.com>
2675
2676         [wx] Unreviewed build fix. Generate the proper install_name for dylibs on Mac.
2677
2678         * wx/packaging/build-mac-installer.py:
2679
2680 2011-07-06  Xan Lopez  <xlopez@igalia.com>
2681
2682         [GTK] Only set env. variables in start when strictly needed
2683         https://bugs.webkit.org/show_bug.cgi?id=64026
2684
2685         Reviewed by Eric Seidel.
2686
2687         * Scripts/webkitpy/layout_tests/port/gtk.py: add a
2688         setup_environ_for_server method in GtkPort and set as many
2689         env. variables as possible there.
2690
2691 2011-07-06  Adam Roben  <aroben@apple.com>
2692
2693         Teach TestFailures how to load, parse, and interpret NRWT test results
2694
2695         Fixes <http://webkit.org/b/61877> TestFailures page doesn't show testers that use
2696         new-run-webkit-tests
2697
2698         Reviewed by Adam Barth.
2699
2700         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
2701         (Builder.prototype.failureDiagnosisTextAndURL): Added support for the new 'flaky' failure
2702         type. For now we don't account for image-only flakes (but TestFailures doesn't deal with
2703         pixel tests at all currently).
2704         (Builder.prototype.getNumberOfFailingTests): Relaxed the regex used for parsing the number
2705         of failing tests from the buildbot output. Make sure not to count "new passes" (a new
2706         category introduced by NRWT) as failures.
2707
2708         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
2709         (LayoutTestResultsLoader.prototype.start): Moved a bunch of code from here to
2710         _fetchAndParseORWTResults. This function now attempts to load NRWT results, then falls back
2711         to loading ORWT results.
2712         (LayoutTestResultsLoader.prototype._fetchAndParseNRWTResults): Added. Tries to load and
2713         parse the full_results.json file for the given build.
2714         (LayoutTestResultsLoader.prototype._fetchAndParseORWTResults): Added. Code came from start.
2715         Fixed a bug along the way where we were sometimes calling the error callback instead of the
2716         success callback when all tests passed.
2717
2718         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/NRWTResultsParser.js: Added.
2719         (NRWTResultsParser): Do-nothing constructor.
2720         (NRWTResultsParser.prototype.parse): Uses eval() (eww!) to get the test results out of the
2721         JS string, then iterates over all the tests in the results data and builds up a data
2722         structure matching what ORWTResultsParser returns.
2723
2724         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
2725         (Array.prototype.contains):
2726         (String.prototype.contains):
2727         Added these simple helper functions.
2728
2729         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/index.html: Pull in
2730         NRWTResultsParser.js.
2731
2732 2011-07-06  Adam Roben  <aroben@apple.com>
2733
2734         Extract code to parse ORWT's results.html file into its own class
2735
2736         Prep work for fixing <http://webkit.org/b/61877> TestFailures page doesn't show testers that
2737         use new-run-webkit-tests
2738
2739         Reviewed by Adam Barth.
2740
2741         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
2742         (LayoutTestResultsLoader.prototype.start): Moved a bunch of code from here...
2743
2744         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ORWTResultsParser.js: Added.
2745         (ORWTResultsParser):
2746         (ORWTResultsParser.prototype.parse):
2747         ...to here.
2748
2749         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/index.html: Pull in
2750         ORWTResultsParser.js.
2751
2752 2011-07-06  Eric Seidel  <eric@webkit.org>
2753
2754         Unreviewed.  Fix a unittest from hitting the network. :)
2755
2756         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2757
2758 2011-07-06  Eric Seidel  <eric@webkit.org>
2759
2760         Unreviewed.  Fixing an exception in test-webkitpy, caused by bug 64006.
2761
2762         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2763
2764 2011-07-06  Gustavo Noronha Silva  <gustavo.noronha@collabora.com>
2765
2766         Run an Xvfb instance to run the API tests on, like we do for the
2767         layout tests.
2768
2769         Reviewed by Xan Lopez.
2770
2771         * Scripts/run-gtk-tests:
2772
2773 2011-07-06  Daniel Bates  <dbates@webkit.org>
2774
2775         Make SCM unit tests faster
2776         https://bugs.webkit.org/show_bug.cgi?id=63883
2777
2778         Speeds up the SCM unit tests by a factor of 4. Currently, we create a
2779         mock SVN repo for each test_ method in SVNTest and GitTest and creating
2780         this repo is expensive.
2781
2782         Instead, it is significantly faster to create the mock SVN repo once
2783         and then perform a filesystem copy of it for each test_ method.
2784
2785         Note, Python 2.7's unittest module implements support for per class and
2786         per module setup and tear down methods which could be used to implement
2787         similar functionality. At the time of writing, test-webkitpy is designed
2788         to support Python 2.5. So, we can't take advantage of these Python 2.7
2789         features :(
2790
2791         Reviewed by Eric Seidel.
2792
2793         * Scripts/webkitpy/common/checkout/scm/scm_unittest.py:
2794
2795 2011-07-06  Adam Barth  <abarth@webkit.org>
2796
2797         Remove unused parameter from _add_test in test_expectations.py
2798         https://bugs.webkit.org/show_bug.cgi?id=64008
2799
2800         Reviewed by Eric Seidel.
2801
2802         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
2803
2804 2011-07-06  Adam Barth  <abarth@webkit.org>
2805
2806         Use old-run-webkit-tests when --leaks is present
2807         https://bugs.webkit.org/show_bug.cgi?id=64012
2808
2809         Reviewed by Adam Roben.
2810
2811         We'll remove this branch once support of --leaks is added to
2812         new-run-webkit-tests.
2813
2814         * Scripts/run-webkit-tests:
2815
2816 2011-07-06  Benjamin Poulain  <benjamin@webkit.org>
2817
2818         [Qt] [WK2] Add touch mocking to Qt's MiniBrowser
2819         https://bugs.webkit.org/show_bug.cgi?id=63995
2820
2821         Reviewed by Andreas Kling.
2822
2823         The mouse events are intercepted through QCoreApplication::notify(),
2824         and fake touch events are generated for mouse events when necessary.
2825
2826         If touch events are received from the platform, we assume the current
2827         hardware has a touch screen and we stop generating fake events.
2828
2829         * MiniBrowser/qt/MiniBrowserApplication.cpp:
2830         (isTouchEvent):
2831         (isMouseEvent):
2832         (MiniBrowserApplication::MiniBrowserApplication):
2833         (MiniBrowserApplication::notify):
2834         * MiniBrowser/qt/MiniBrowserApplication.h:
2835
2836 2011-07-06  Xan Lopez  <xlopez@igalia.com>
2837
2838         Reviewed by Gustavo Noronha.
2839
2840         * Scripts/webkitpy/layout_tests/port/gtk.py: use Popen correctly.
2841
2842 2011-07-06  Eric Seidel  <eric@webkit.org>
2843
2844         webkit-patch failure-reason does not understand NRWT results
2845         https://bugs.webkit.org/show_bug.cgi?id=64006
2846
2847         Reviewed by Adam Barth.
2848
2849         Adam Barth tells me failure-reason should be deleted soon,
2850         but aroben's fancy new TestFailures/ page doesn't work for
2851         NRWT yet, and I needed to know when
2852         third-party-cookie-relaxing started failing.
2853
2854         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
2855         * Scripts/webkitpy/tool/commands/queries.py:
2856
2857 2011-07-06  Xan Lopez  <xlopez@igalia.com>
2858
2859         [GTK] Add missing environment variables in NWRT
2860         https://bugs.webkit.org/show_bug.cgi?id=64004
2861
2862         Reviewed by Gustavo Noronha Silva.
2863
2864         Add missing environment variables needed by the GTK+ port.
2865
2866         * Scripts/webkitpy/layout_tests/port/gtk.py: ditto.
2867
2868 2011-07-06  Adam Roben  <aroben@apple.com>
2869
2870         Make run-api-tests output intelligible on the bots
2871
2872         run-api-tests's output on the bots was very hard to read for two reasons:
2873           1) It was mixing its own output with gtest's
2874           2) It was using ANSI escape sequences to print colored text, but the bots don't support
2875              that
2876
2877         Now, the --verbose flag turns off almost all of run-api-tests's own output so that gtest's
2878         output will not be obscured. We still print "Timeout" messages even in verbose mode, since
2879         gtest doesn't have any native support for timeouts. Also, when our output is being
2880         redirected to a file, we don't print ANSI escape sequences.
2881
2882         Fixes <http://webkit.org/b/63996> It's very hard to read run-api-tests output on the bots
2883
2884         Reviewed by Sam Weinig.
2885
2886         * Scripts/run-api-tests: Don't use the :constants interface of Term::ANSIColor. We want to
2887         use the colored function instead.
2888         (runAllTestsInSuite): Don't print out suite names in verbose mode; gtest will give us enough
2889         context that they aren't needed.
2890         (runTest): Don't print out test names or pass/fail messages in verbose mode; gtest will do
2891         that for us. Also, use the new possiblyColored function instead of always coloring output.
2892         (possiblyColored): Added. When printing to a tty, returns the string with the appropriate
2893         ANSI color escape sequences added. Otherwise just returns the string unmodified.
2894
2895 2011-07-06  Xan Lopez  <xlopez@igalia.com>
2896
2897         Reviewed by Gustavo Noronha.
2898
2899         Launch Xvfb (mostly) the same way we were launching it in the
2900         bots, for maximum compatibility.
2901
2902         * Scripts/webkitpy/layout_tests/port/gtk.py:
2903
2904 2011-07-06  Andreas Kling  <kling@webkit.org>
2905
2906         [Qt][WK2] Split Qt API into two different web views (touch and desktop)
2907         https://bugs.webkit.org/show_bug.cgi?id=63950
2908
2909         Co-authored with Benjamin Poulain <benjamin@webkit.org>.
2910
2911         Reviewed by Kenneth Rohde Christiansen.
2912
2913         Port WTR/Qt to use QDesktopWebView and force a 800x600 viewport.
2914
2915         The Qt MiniBrowser now defaults to QDesktopWebView but can be started
2916         in QTouchWebView mode by passing -touch on the command line.
2917
2918         A lot of functionality has been removed and/or disabled from
2919         MiniBrowser until we reintroduce the necessary API's.
2920
2921         * MiniBrowser/qt/BrowserView.cpp:
2922         (BrowserView::BrowserView):
2923         (BrowserView::~BrowserView):
2924         (BrowserView::load):
2925         (BrowserView::view):
2926         (BrowserView::touchWebView):
2927         (BrowserView::desktopWebView):
2928         * MiniBrowser/qt/BrowserView.h:
2929         * MiniBrowser/qt/BrowserWindow.cpp:
2930         (newPageFunction):
2931         (BrowserWindow::BrowserWindow):
2932         (BrowserWindow::webView):
2933         (BrowserWindow::newWindow):
2934         (BrowserWindow::showUserAgentDialog):
2935         (BrowserWindow::updateUserAgentList):
2936         * MiniBrowser/qt/BrowserWindow.h:
2937         * MiniBrowser/qt/MiniBrowserApplication.cpp:
2938         (MiniBrowserApplication::handleUserOptions):
2939         * MiniBrowser/qt/MiniBrowserApplication.h:
2940         (WindowOptions::WindowOptions):
2941         * MiniBrowser/qt/UrlLoader.cpp:
2942         (UrlLoader::UrlLoader):
2943         * MiniBrowser/qt/main.cpp:
2944         (main):
2945         * WebKitTestRunner/qt/PlatformWebViewQt.cpp:
2946         (WTR::WebView::wkView):
2947         (WTR::WebView::pageRef):
2948         (WTR::WebView::WebView):
2949         (WTR::PlatformWebView::page):
2950         (WTR::PlatformWebView::windowFrame):
2951         (WTR::PlatformWebView::setWindowFrame):
2952
2953 2011-07-06  Eric Seidel  <eric@webkit.org>
2954
2955         Unreviewed..
2956
2957         * Scripts/run-webkit-tests:
2958         (useNewRunWebKitTests):
2959          - Fix the pass-thru of --gtk and --qt for old-run-webkit-tests,
2960            it was previously broken by the existence of isGtk/isQt checks which
2961            removed the --gtk/--qt args from @ARGV.
2962
2963 2011-07-06  Eric Seidel  <eric@webkit.org>
2964
2965         new-run-webkit-tests uses a 35s timeout for Mac and 6s for all other ports
2966         https://bugs.webkit.org/show_bug.cgi?id=63983
2967
2968         Reviewed by Xan Lopez.
2969
2970         All WebKit ports want to use a 35s timeout.  Only chromium currently
2971         uses a 6s timeout.  Eventually we'll move all WebKit ports to 6s too
2972         to further speed up new-run-webkit-tests runs, but that's going to be a while.
2973
2974         * Scripts/webkitpy/layout_tests/port/mac.py:
2975         * Scripts/webkitpy/layout_tests/port/webkit.py:
2976         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
2977
2978 2011-07-06  Xan Lopez  <xlopez@igalia.com>
2979
2980         Unreviewed.
2981
2982         NWRT actually seems to work locally, so it should work in the bots
2983         as long as we take care of the TIMEOUT issue (which is being fixed
2984         as I write this). Let's give this another shot and fix whatever
2985         breaks.
2986
2987         * Scripts/run-webkit-tests:
2988         (useNewRunWebKitTests):
2989
2990 2011-07-06  Xan Lopez  <xlopez@igalia.com>
2991
2992         Unreviewed.
2993
2994         Disable NWRT on GTK. At the very least it needs an updated exected
2995         results file since NWRT is more sensitive to timeouts, not sure
2996         what the exact problem is.
2997
2998         * Scripts/run-webkit-tests:
2999         (useNewRunWebKitTests):
3000
3001 2011-07-06  Adam Barth  <abarth@webkit.org>
3002
3003         Fix global variable leak detected by noglobals.
3004
3005         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
3006
3007 2011-07-06  Adam Barth  <abarth@webkit.org>
3008
3009         garden-o-matic shouldn't show flaky tests by default
3010         https://bugs.webkit.org/show_bug.cgi?id=63979
3011
3012         Reviewed by Ojan Vafai.
3013
3014         If any of the actual test results are PASS, the test is flaky and we
3015         can ignore it (at least for now).  If we get awesome at dealing with
3016         failing tests, we might want to tackle flaky tests by refering the user
3017         to the flakiness dashboard, which is way more awesome at dealing with
3018         flaky tests that this tool ever will be.
3019
3020         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
3021         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
3022
3023 2011-07-05  Adam Barth  <abarth@webkit.org>
3024
3025         garden-o-matic should list the failing tests
3026         https://bugs.webkit.org/show_bug.cgi?id=63978
3027
3028         Reviewed by Eric Seidel.
3029
3030         * Scripts/webkitpy/tool/servers/data/gardeningserver/config.js: Added.
3031         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
3032         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
3033         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
3034         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
3035         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html:
3036         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js: Added.
3037         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js: Added.
3038         * Scripts/webkitpy/tool/servers/gardeningserver.py:
3039
3040
3041 2011-07-05  Eric Seidel  <eric@webkit.org>
3042
3043         Re-enable NRWT on Leopard now that bug 63973 is fixed.
3044
3045         Reviewed by Adam Barth.
3046
3047         I also made the should-enable expression a black-list instead
3048         of a white-list since every bot on build.webkit.org is
3049         now using NRWT except windows and WebKit2 bots.
3050
3051         * Scripts/run-webkit-tests:
3052         (useNewRunWebKitTests):
3053
3054 2011-07-05  Eric Seidel  <eric@webkit.org>
3055
3056         Perl string concat failure on Leopard (causes NRWT to fail)
3057         https://bugs.webkit.org/show_bug.cgi?id=63973
3058
3059         Reviewed by Daniel Bates.
3060
3061         Leopard's "xcodebuild -version" has a different format from Snow Leopard/Lion
3062         thus the version parsing logic in webkitdirs.pm seems to fail.
3063
3064         Default to "3.0" instead of undef when we can't parse the version number.
3065         This should send us down the correct path for Leopard.
3066
3067         * Scripts/webkitdirs.pm:
3068         (determineBaseProductDir):
3069
3070 2011-07-05  Kevin Ollivier  <kevino@theolliviers.com>
3071
3072         [wx] Unreviewed build fix after addition of Source/WebCore/xml/parser dir.
3073
3074         * waf/build/settings.py:
3075
3076 2011-07-05  Eric Seidel  <eric@webkit.org>
3077
3078         Disable NRWT for Leopard until https://bugs.webkit.org/show_bug.cgi?id=63973 can be fixed.
3079
3080         Reviewed by Adam Barth.
3081
3082         * Scripts/run-webkit-tests:
3083         (useNewRunWebKitTests):
3084
3085 2011-07-05  Eric Seidel  <eric@webkit.org>
3086
3087         Make run-webkit-tests pass --verbose to new-run-webkit-tests on qt/gtk bots.
3088
3089         Reviewed by Adam Barth.
3090
3091         * Scripts/run-webkit-tests:
3092         (runningOnBuildBot):
3093
3094 2011-07-05  Adam Barth  <abarth@webkit.org>
3095
3096         Add results_unittests.js to garden-o-matic
3097         https://bugs.webkit.org/show_bug.cgi?id=63971
3098
3099         Reviewed by Eric Seidel.
3100
3101         Adding these unit tests forced me to commit to an API for the results
3102         module.  I'm not sure if this is the API we'll end up with, but it's
3103         something not entirely unreasonable.
3104
3105         Also, I've changed results.js over to WebKit style, as discussed previously.
3106
3107         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
3108         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
3109         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js: Added.
3110         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html:
3111
3112 2011-07-05  Eric Seidel  <eric@webkit.org>
3113
3114         Teach run-webkit-tests how to translate --qt and --gtk for new-run-webkit-tests.
3115
3116         Reviewed by Adam Barth.
3117
3118         * Scripts/run-webkit-tests:
3119
3120 2011-07-05  Adam Barth  <abarth@webkit.org>
3121
3122         Add first unit tests for garden-o-matic
3123         https://bugs.webkit.org/show_bug.cgi?id=63969
3124
3125         Reviewed by Eric Seidel.
3126
3127         I've also reformated base.js to match WebKit style, as discussed in the
3128         previous bug.
3129
3130         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
3131         * Scripts/webkitpy/tool/servers/data/gardeningserver/base_unittests.js: Added.
3132         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html: Added.
3133
3134 2011-07-05  Eric Seidel  <eric@webkit.org>
3135
3136         Move Qt, Gtk and Leopard to NRWT.
3137
3138         We'll roll out any of these if we see problems.
3139
3140         Reviewed by Adam Barth.
3141
3142         * Scripts/run-webkit-tests:
3143         (useNewRunWebKitTests):
3144
3145 2011-07-05  Adam Barth  <abarth@webkit.org>
3146
3147         garden-o-matic should understand which tests have unexpected failures
3148         https://bugs.webkit.org/show_bug.cgi?id=63965
3149
3150         Reviewed by Eric Seidel.
3151
3152         Currently, this code just logs the list of unexpected failures to the
3153         console.  A future patch will do something useful with this
3154         information.  A testing harness will also come in the next patch.
3155
3156         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js: Added.
3157         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
3158         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
3159         * Scripts/webkitpy/tool/servers/gardeningserver.py:
3160
3161 2011-07-05  Adam Barth  <abarth@webkit.org>
3162
3163         garden-o-matic should know how to fetch test results from the (NRWT) bots
3164         https://bugs.webkit.org/show_bug.cgi?id=63959
3165
3166         Reviewed by Eric Seidel.
3167
3168         Currently, this code just fetches the results and expectations JSON
3169         blob for Linux from the server and logs it to the console.
3170
3171         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
3172         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js: Added.
3173         * Scripts/webkitpy/tool/servers/gardeningserver.py:
3174
3175 2011-07-05  Dirk Pranke  <dpranke@chromium.org>
3176
3177         Reviewed by Eric Seidel.
3178
3179         Python tests are failing on leopard
3180         https://bugs.webkit.org/show_bug.cgi?id=63842
3181
3182         Fix the missing flush() call that appears to be needed by the
3183         logging package in python 2.5. Also fix the regressions
3184         introduced in the run_webkit_test tests when we switched from
3185         thread to processes (since processes aren't available on 2.5).
3186
3187         Also fix a missing "from __future__ import with_statement' in
3188         testfilehandler (unreviewed).
3189
3190         * Scripts/webkitpy/layout_tests/layout_package/metered_stream.py:
3191         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
3192         * TestResultServer/handlers/testfilehandler.py:
3193
3194 2011-07-05  Eric Seidel  <eric@webkit.org>
3195
3196         Reviewed by Adam Barth.
3197
3198         re-enabling new-run-webkit-tests on the Snow Leopard WebKit1 bots
3199         now that we believe the stray http-server issue to be solved.
3200
3201         * Scripts/run-webkit-tests:
3202         (useNewRunWebKitTests):
3203
3204 2011-07-05  Adam Barth  <abarth@webkit.org>
3205
3206         Turns out we need to support numerals too because that's what jQuery
3207         uses by default.
3208
3209         * TestResultServer/handlers/testfilehandler.py:
3210
3211 2011-07-05  Dirk Pranke  <dpranke@chromium.org>
3212
3213         nrwt: allow for multiple http shards
3214         https://bugs.webkit.org/show_bug.cgi?id=63116
3215
3216         Reviewed by Tony Chang.
3217
3218         This modifies the sharding logic to support multiple http
3219         shards, but for now we clamp to one shard until we can test
3220         perf impact and flakiness impact.
3221
3222         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3223         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
3224
3225 2011-07-05  Dirk Pranke  <dpranke@chromium.org>
3226
3227         Re-land nrwt: make sharding tests needing locks less hard-coded
3228         https://bugs.webkit.org/show_bug.cgi?id=63112
3229
3230         Reviewed by Ojan Vafai.
3231
3232         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3233         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
3234
3235 2011-07-05  Adam Barth  <abarth@webkit.org>
3236
3237         TestResultServer should support callback parameter for JSON
3238         https://bugs.webkit.org/show_bug.cgi?id=63961
3239
3240         Reviewed by Ojan Vafai.
3241
3242         This makes using jQuery.ajax so much more pleasant.
3243
3244         * TestResultServer/handlers/testfilehandler.py:
3245
3246 2011-07-05  Eric Seidel  <eric@webkit.org>
3247
3248         buildbot needs to understand whether NRWT exited early after having too many failures
3249         https://bugs.webkit.org/show_bug.cgi?id=63839
3250
3251         Reviewed by Adam Barth.
3252
3253         Fix-up the exited early messages printed by NRWT so that
3254         the buildbot can parse them as expected.
3255         It looks for lines using "if line.find('Exiting early') >= 0:"
3256
3257         I also plumbed the "early exit" status through to results.json
3258         in the form of an "interrupted" bool.  It was unclear to me
3259         if results.json already had enough information to compute this bool
3260         itself.  It's possible Ojan could come up with a better fix.
3261
3262         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3263         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
3264         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
3265         * Scripts/webkitpy/layout_tests/layout_package/result_summary.py:
3266
3267 2011-07-05  Eric Seidel  <eric@webkit.org>
3268
3269         new-run-webkit-tests fails to start http server if one is already running
3270         https://bugs.webkit.org/show_bug.cgi?id=63956
3271
3272         Reviewed by Adam Barth.
3273
3274         * BuildSlaveSupport/kill-old-processes:
3275          - Removed ^M line endings and sorted the process names.
3276          - There are no functional changes to this file.
3277            (except that now that its using unix line endings it can be executed directly!)
3278         * Scripts/webkitpy/layout_tests/servers/apache_http_server.py:
3279          - Change the name to 'httpd' to match old-run-webkit-test paths.
3280
3281 2011-07-05  Adam Barth  <abarth@webkit.org>
3282
3283         Add basic ajax support to garden-o-matic
3284         https://bugs.webkit.org/show_bug.cgi?id=63874
3285
3286         Reviewed by Eric Seidel.
3287
3288         This patch adds some basic infrastructure to garden-o-matic.
3289         Currently, the infrastructure is just used to make the "quit" command
3290         use Ajax, but in the future, this infrastructure will be used to do
3291         more sophistocated remote proceedure calls.
3292
3293         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
3294         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js: Added.
3295         * Scripts/webkitpy/tool/servers/gardeningserver.py:
3296         * Scripts/webkitpy/tool/servers/reflectionhandler.py:
3297
3298 2011-07-05  Adam Barth  <abarth@webkit.org>
3299
3300         Add trivial garden-o-matic command (with server)
3301         https://bugs.webkit.org/show_bug.cgi?id=63872
3302
3303         Reviewed by Eric Seidel.
3304
3305         This patch adds a trivial "Hello, world" garden-o-matic command,
3306         complete with an HTTP server.  This command re-uses a bunch of code
3307         from the existing rebaseline-server command.  Over time, this command
3308         will grow to be a tool useful for gardening the WebKit tree.
3309
3310         * Scripts/webkitpy/tool/commands/__init__.py:
3311         * Scripts/webkitpy/tool/commands/abstractservercommand.py: Added.
3312         * Scripts/webkitpy/tool/commands/gardenomatic.py: Added.
3313         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
3314         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html: Added.
3315         * Scripts/webkitpy/tool/servers/gardeningserver.py: Added.
3316         * Scripts/webkitpy/tool/servers/rebaselineserver.py:
3317
3318 2011-07-05  Adam Roben  <aroben@apple.com>
3319
3320         Make prepare-ChangeLog include modified Perl functions in its ChangeLog template
3321
3322         This is a very simple first cut. Functions must start with a line that starts with "sub "
3323         and end with a line that starts with a closing brace. No leading whitespace is allowed.
3324         Package names aren't parsed at all.
3325
3326         Fixes <http://webkit.org/b/21591> prepare-ChangeLog should know how to find functions in
3327         Perl files
3328
3329         Reviewed by David Kilzer.
3330
3331         * Scripts/prepare-ChangeLog: Removed redundant code that ignored certain files when
3332         searching for function line ranges. This is already done inside the get_function_line_ranges
3333         function.
3334         (get_function_line_ranges): Cleaned up coding style a little bit. Call
3335         get_function_line_ranges_for_perl for files with .pl and .pm extensions. For files with an
3336         unknown extension or no extension, read the shebang line to try to determine the script
3337         interpreter. Call get_function_line_ranges_for_perl if the interpreter seems to be Perl.
3338         (get_function_line_ranges_for_perl): Added. Does extremely basic parsing of the file to find
3339         lines starting with "sub " or "}".
3340
3341 2011-07-05  Adam Roben  <aroben@apple.com>
3342
3343         Clean up run-api-tests output on Windows
3344
3345         We were mixing run-api-tests output with gtest output, and the result was a mess.
3346
3347         Fixes <http://webkit.org/b/63954> run-api-tests output is very confusing on Windows
3348
3349         Reviewed by Dan Bates.
3350
3351         * Scripts/run-api-tests:
3352         (runTest):
3353         (populateTests):
3354         Made Windows use the formerly-Mac-only codepaths which correctly suppress gtest output
3355         except in verbose mode. Most changes are due to indentation. Use of the arch utility is now
3356         the only Mac-specific thing in these functions. Some other platform-specific code was moved
3357         from here...
3358
3359         (prepareEnvironmentForRunningTestTool):
3360         (testToolPath):
3361         ...to here.
3362
3363 2011-07-04  Gavin Barraclough  <barraclough@apple.com>
3364
3365         Reviewed by Sam Weinig.
3366
3367         https://bugs.webkit.org/show_bug.cgi?id=16652
3368         Firefox and JavaScriptCore differ in Number.toString(integer)
3369
3370         Added forwarding header.
3371
3372         * DumpRenderTree/ForwardingHeaders/wtf/StdLibExtras.h: Added.
3373
3374 2011-07-04  Anders Carlsson  <andersca@apple.com>
3375
3376         Try to fix Qt build.
3377
3378         * DumpRenderTree/TestNetscapePlugIn/Tests/NPRuntimeRemoveProperty.cpp:
3379
3380 2011-07-04  Anders Carlsson  <andersca@apple.com>
3381
3382         NP_RemoveProperty is not called back by Safari when delete npObject.prop is encountered in JavaScript
3383         https://bugs.webkit.org/show_bug.cgi?id=63915
3384         <rdar://problem/7124300>
3385
3386         Reviewed by Sam Weinig.
3387
3388         Extend the NPRuntimeRemoveProperty to handle delete object.property from JavaScript.
3389
3390         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
3391         (PluginTest::NPN_IdentifierIsString):
3392         (PluginTest::NPN_UTF8FromIdentifier):
3393         (PluginTest::NPN_IntFromIdentifier):
3394         (PluginTest::NPN_RetainObject):
3395         (PluginTest::NPN_ReleaseObject):
3396         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
3397         (PluginTest::Object::removeProperty):
3398         (PluginTest::Object::identifierIs):
3399         (PluginTest::Object::NP_RemoveProperty):
3400         (PluginTest::Object::npClass):
3401         * DumpRenderTree/TestNetscapePlugIn/Tests/NPRuntimeRemoveProperty.cpp:
3402         (NPRuntimeRemoveProperty::TestObject::TestObject):
3403         (NPRuntimeRemoveProperty::TestObject::hasProperty):
3404         (NPRuntimeRemoveProperty::TestObject::getProperty):
3405         (NPRuntimeRemoveProperty::TestObject::removeProperty):
3406         (NPRuntimeRemoveProperty::PluginObject::PluginObject):
3407         (NPRuntimeRemoveProperty::PluginObject::~PluginObject):
3408         (NPRuntimeRemoveProperty::PluginObject::hasMethod):
3409         (NPRuntimeRemoveProperty::PluginObject::invoke):
3410         (NPRuntimeRemoveProperty::PluginObject::hasProperty):
3411         (NPRuntimeRemoveProperty::PluginObject::getProperty):
3412         (NPRuntimeRemoveProperty::NPP_GetValue):
3413
3414 2011-07-04  Leandro Pereira  <leandro@profusion.mobi>
3415
3416         Reviewed by Kent Tamura.
3417
3418         [EFL] Add LayoutTestControllerEfl.
3419         https://bugs.webkit.org/show_bug.cgi?id=61974
3420
3421         Only the core features are implemented. Most functions are actually
3422         stubs and will be implemented as soon as its features are needed.
3423         Build system changes will be made as soon as DRT is completely
3424         upstreamed.
3425
3426         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: Added.
3427         (LayoutTestController::~LayoutTestController):
3428         (LayoutTestController::addDisallowedURL):
3429         (LayoutTestController::clearBackForwardList):
3430         (LayoutTestController::copyDecodedHostName):
3431         (LayoutTestController::copyEncodedHostName):
3432         (LayoutTestController::dispatchPendingLoadRequests):
3433         (LayoutTestController::display):
3434         (LayoutTestController::counterValueForElementById):
3435         (LayoutTestController::keepWebHistory):
3436         (LayoutTestController::computedStyleIncludingVisitedInfo):
3437         (LayoutTestController::nodesFromRect):
3438         (LayoutTestController::layerTreeAsText):
3439         (LayoutTestController::pageNumberForElementById):
3440         (LayoutTestController::numberOfPages):
3441         (LayoutTestController::pageProperty):
3442         (LayoutTestController::isPageBoxVisible):
3443         (LayoutTestController::pageSizeAndMarginsInPixels):
3444         (LayoutTestController::webHistoryItemCount):
3445         (LayoutTestController::workerThreadCount):
3446         (LayoutTestController::notifyDone):
3447         (LayoutTestController::pathToLocalResource):
3448         (LayoutTestController::queueLoad):
3449         (LayoutTestController::setAcceptsEditing):
3450         (LayoutTestController::setAlwaysAcceptCookies):
3451         (LayoutTestController::setCustomPolicyDelegate):
3452         (LayoutTestController::waitForPolicyDelegate):
3453         (LayoutTestController::setScrollbarPolicy):
3454         (LayoutTestController::addOriginAccessWhitelistEntry):
3455         (LayoutTestController::removeOriginAccessWhitelistEntry):
3456         (LayoutTestController::setMainFrameIsFirstResponder):
3457         (LayoutTestController::setTabKeyCyclesThroughElements):
3458         (LayoutTestController::setUseDashboardCompatibilityMode):
3459         (LayoutTestController::setUserStyleSheetEnabled):
3460         (LayoutTestController::setUserStyleSheetLocation):
3461         (LayoutTestController::setValueForUser):
3462         (LayoutTestController::setViewModeMediaFeature):
3463         (LayoutTestController::setWindowIsKey):
3464         (LayoutTestController::setSmartInsertDeleteEnabled):
3465         (waitToDumpWatchdogFired):
3466         (LayoutTestController::setWaitToDump):
3467         (LayoutTestController::windowCount):
3468         (LayoutTestController::setPrivateBrowsingEnabled):
3469         (LayoutTestController::setJavaScriptCanAccessClipboard):
3470         (LayoutTestController::setXSSAuditorEnabled):
3471         (LayoutTestController::setFrameFlatteningEnabled):
3472         (LayoutTestController::setSpatialNavigationEnabled):
3473         (LayoutTestController::setAllowUniversalAccessFromFileURLs):
3474         (LayoutTestController::setAllowFileAccessFromFileURLs):
3475         (LayoutTestController::setAuthorAndUserStylesEnabled):
3476         (LayoutTestController::setAutofilled):
3477         (LayoutTestController::disableImageLoading):
3478         (LayoutTestController::setMockDeviceOrientation):
3479         (LayoutTestController::setMockGeolocationPosition):
3480         (LayoutTestController::setMockGeolocationError):
3481         (LayoutTestController::setGeolocationPermission):
3482         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
3483         (LayoutTestController::addMockSpeechInputResult):
3484         (LayoutTestController::setIconDatabaseEnabled):
3485         (LayoutTestController::setJavaScriptProfilingEnabled):
3486         (LayoutTestController::setSelectTrailingWhitespaceEnabled):
3487         (LayoutTestController::setPopupBlockingEnabled):
3488         (LayoutTestController::setPluginsEnabled):
3489         (LayoutTestController::elementDoesAutoCompleteForElementWithId):
3490         (LayoutTestController::execCommand):
3491         (LayoutTestController::findString):
3492         (LayoutTestController::isCommandEnabled):
3493         (LayoutTestController::setCacheModel):
3494         (LayoutTestController::setPersistentUserStyleSheetLocation):
3495         (LayoutTestController::clearPersistentUserStyleSheet):
3496         (LayoutTestController::clearAllApplicationCaches):
3497         (LayoutTestController::setApplicationCacheOriginQuota):
3498         (LayoutTestController::clearApplicationCacheForOrigin):
3499         (LayoutTestController::localStorageDiskUsageForOrigin):
3500         (LayoutTestController::originsWithApplicationCache):
3501         (LayoutTestController::applicationCacheDiskUsageForOrigin):
3502         (LayoutTestController::clearAllDatabases):
3503         (LayoutTestController::setDatabaseQuota):
3504         (LayoutTestController::originsWithLocalStorage):
3505         (LayoutTestController::deleteAllLocalStorage):
3506         (LayoutTestController::deleteLocalStorageForOrigin):
3507         (LayoutTestController::observeStorageTrackerNotifications):
3508         (LayoutTestController::syncLocalStorage):
3509         (LayoutTestController::setDomainRelaxationForbiddenForURLScheme):
3510         (LayoutTestController::goBack):
3511         (LayoutTestController::setDefersLoading):
3512         (LayoutTestController::setAppCacheMaximumSize):
3513         (LayoutTestController::pauseAnimationAtTimeOnElementWithId):
3514         (LayoutTestController::pauseTransitionAtTimeOnElementWithId):
3515         (LayoutTestController::sampleSVGAnimationForElementAtTime):
3516         (LayoutTestController::numberOfActiveAnimations):
3517         (LayoutTestController::suspendAnimations):
3518         (LayoutTestController::resumeAnimations):
3519         (LayoutTestController::overridePreference):
3520         (LayoutTestController::addUserScript):
3521         (LayoutTestController::addUserStyleSheet):
3522         (LayoutTestController::setDeveloperExtrasEnabled):
3523         (LayoutTestController::setAsynchronousSpellCheckingEnabled):
3524         (LayoutTestController::showWebInspector):
3525         (LayoutTestController::closeWebInspector):
3526         (LayoutTestController::evaluateInWebInspector):
3527         (LayoutTestController::evaluateScriptInIsolatedWorld):
3528         (LayoutTestController::removeAllVisitedLinks):
3529         (LayoutTestController::callShouldCloseOnWebView):
3530         (LayoutTestController::apiTestNewWindowDataLoadBaseURL):
3531         (LayoutTestController::apiTestGoToCurrentBackForwardItem):
3532         (LayoutTestController::setWebViewEditable):
3533         (LayoutTestController::markerTextForListItem):
3534         (LayoutTestController::authenticateSession):
3535         (LayoutTestController::setEditingBehavior):
3536         (LayoutTestController::abortModal):
3537         (LayoutTestController::hasSpellingMarker):
3538         (LayoutTestController::hasGrammarMarker):
3539         (LayoutTestController::dumpConfigurationForViewport):
3540         (LayoutTestController::setSerializeHTTPLoads):
3541         (LayoutTestController::setMinimumTimerInterval):
3542         (LayoutTestController::setTextDirection):
3543         (LayoutTestController::allowRoundingHacks):
3544
3545 2011-06-24  Vsevolod Vlasov  <vsevik@chromium.org>
3546
3547         Unreviewed, add myself as committer.
3548
3549         * Scripts/webkitpy/common/config/committers.py:
3550
3551 2011-07-03  Sheriff Bot  <webkit.review.bot@gmail.com>
3552
3553         Unreviewed, rolling out r90347.
3554         http://trac.webkit.org/changeset/90347
3555         https://bugs.webkit.org/show_bug.cgi?id=63886
3556
3557         Build breaks on Leopard, Chromium-win, WinCairo, and WinCE.
3558         (Requested by tkent on #webkit).
3559
3560         * DumpRenderTree/ForwardingHeaders/wtf/StdLibExtras.h: Removed.
3561
3562 2011-07-03  Daniel Bates  <dbates@webkit.org>
3563
3564         Import Checkout in scm_unittests.py and logging in svn.py to fix
3565         Python missing global variable errors
3566
3567         Rubber-stamped by Adam Barth.
3568
3569         * Scripts/webkitpy/common/checkout/scm/scm_unittest.py: Import Checkout.
3570         * Scripts/webkitpy/common/checkout/scm/svn.py: Import logging.
3571
3572 2011-07-02  Gavin Barraclough  <barraclough@apple.com>
3573
3574         Reviewed by Sam Weinig.
3575
3576         https://bugs.webkit.org/show_bug.cgi?id=16652
3577         Firefox and JavaScriptCore differ in Number.toString(integer)
3578
3579         Added forwarding header.
3580
3581         * DumpRenderTree/ForwardingHeaders/wtf/StdLibExtras.h: Added.
3582
3583 2011-04-02  Robert Hogan  <robert@webkit.org>
3584
3585         Reviewed by Benjamin Poulain.
3586
3587         [Qt] Support third-party cookie policy for Qt clients
3588         https://bugs.webkit.org/show_bug.cgi?id=45455
3589
3590         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3591         (LayoutTestController::reset):
3592         (LayoutTestController::setAlwaysAcceptCookies):
3593         (LayoutTestController::setAlwaysBlockCookies):
3594         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3595
3596 2011-07-03  Adam Barth  <abarth@webkit.org>
3597
3598         Factor ReflectionHandler out of the rebaseline server.  Just code motion.
3599
3600         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
3601         * Scripts/webkitpy/tool/servers/rebaselineserver.py:
3602         * Scripts/webkitpy/tool/servers/reflectionhandler.py: Added.
3603
3604 2011-07-03  Adam Barth  <abarth@webkit.org>
3605
3606         Move the rebaseline server out of the commands package and into the
3607         (new) servers package.  This patch prepares for refactoring this code
3608         to share server infrastructure with the (forthcoming) garden-o-matic
3609         command.
3610
3611         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
3612         * Scripts/webkitpy/tool/servers/__init__.py: Added.
3613         * Scripts/webkitpy/tool/servers/data/rebaselineserver/loupe.js:
3614         * Scripts/webkitpy/tool/servers/data/rebaselineserver/main.css:
3615         * Scripts/webkitpy/tool/servers/data/rebaselineserver/queue.js:
3616         * Scripts/webkitpy/tool/servers/rebaselineserver.py: Added.
3617
3618 2011-07-01  Adam Barth  <abarth@webkit.org>
3619
3620         Fix typo in master.cfg, hoping to trigger a master restart.
3621
3622         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3623
3624 2011-07-01  Adam Barth  <abarth@webkit.org>
3625
3626         Turn off new-run-webkit-tests and remove some debug logging.
3627
3628         * Scripts/run-webkit-tests:
3629         * Scripts/webkitpy/layout_tests/port/http_lock.py:
3630
3631 2011-07-01  Adam Barth  <abarth@webkit.org>
3632
3633         Add some more debug logging to try to understand why we're getting
3634         stuck acquring the HTTP lock.
3635
3636         * Scripts/webkitpy/layout_tests/port/http_lock.py:
3637
3638 2011-07-01  Adam Barth  <abarth@webkit.org>
3639
3640         Refactor NRWT parsing logic in master.cfg to understand that run-webkit-tests might run NRWT
3641         https://bugs.webkit.org/show_bug.cgi?id=63854
3642
3643         Reviewed by Dirk Pranke.
3644
3645         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3646
3647 2011-07-01  Eric Seidel  <eric@webkit.org>
3648
3649         new-run-webkit-tests results does not understand that mac uses test_expectations files
3650         https://bugs.webkit.org/show_bug.cgi?id=63838
3651
3652         Reviewed by Adam Barth.
3653
3654         We've decided to "enable" test_expectations for the Mac port for now.
3655         It makes the results.html results much more confusing to read, but at least they're
3656         no longer lying to us.
3657
3658         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3659         * Scripts/webkitpy/layout_tests/port/base.py:
3660         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
3661
3662 2011-07-01  Adam Barth  <abarth@webkit.org>
3663
3664         Temporarily disable waiting for ReportCrash to see how this code
3665         affects performance on the bots.
3666
3667         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3668         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
3669
3670 2011-07-01  Dmitry Lomov  <dslomov@google.com>
3671
3672         Reviewed by Adam Roben.
3673
3674         https://bugs.webkit.org/show_bug.cgi?id=63818
3675         Test bots shouldn't attempt to build TestWebKitAPI on Windows.
3676         TestWebKitAPI builds as a part of general webkit build and is a
3677         part of build products archive.
3678
3679         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3680         * Scripts/run-api-tests:
3681
3682 2011-07-01  Adam Barth  <abarth@webkit.org>
3683
3684         We shouldn't wait for ReportCrash while the timeout is running
3685         https://bugs.webkit.org/show_bug.cgi?id=63837
3686
3687         Reviewed by Eric Seidel.
3688
3689         Instead, we should move this check to worker.py.
3690
3691         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
3692         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
3693
3694 2011-07-01  Adam Roben  <aroben@apple.com>
3695
3696         Stop setting unsafe headers on POSTs via XHR from TestFailures page
3697
3698         Fixes <http://webkit.org/b/63833> TestFailures page causes a bunch of spew in the
3699         Inspector's console about unsafe headers
3700
3701         Reviewed by Dan Bates.
3702
3703         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
3704         (fetchResource): Removed unnecessary code to set Content-Length and Connection headers.
3705
3706 2011-07-01  Eric Seidel  <eric@webkit.org>
3707
3708         Reviewed by Adam Barth.
3709
3710         Switch Snow Leopard over to new-run-webkit-tests (except for wk2)
3711         https://bugs.webkit.org/show_bug.cgi?id=63786
3712
3713         We're down to one bug blocking the "move to NRWT" bug:
3714         https://bugs.webkit.org/show_bug.cgi?id=34984
3715         so we're trying transitioning Snow Leopard to NRWT as an experiment.
3716
3717         If we run into any blocking issues we'll of course roll this back out
3718         (or anyone reading this should feel encouraged to).
3719
3720         * Scripts/run-webkit-tests:
3721
3722 2011-07-01  Adam Barth  <abarth@webkit.org>
3723
3724         Add a compat shim for --leaks.  Support for --leaks will be added in
3725         https://bugs.webkit.org/show_bug.cgi?id=63832.
3726
3727         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3728
3729 2011-07-01  Adam Barth  <abarth@webkit.org>
3730
3731         Reviewed by Darin Adler.
3732
3733         ReportCrash destabilizes new-run-webkit-tests
3734         https://bugs.webkit.org/show_bug.cgi?id=63795
3735
3736         ReportCrash chews up too many system resources that it destablizes
3737         tests run concurrently.  This patch causes us to spin down all the
3738         worker processes when ReportCrash is running.
3739
3740         Also, this patch causes the master process to tell the user we're
3741         waiting for the crash reporter rather than having that be the worker's
3742         job.  This stops the user from getting spammed with too many "waiting
3743         for crash reporter" messages and also cleans up some minor UI issues
3744         with the meter.
3745
3746         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3747         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
3748         * Scripts/webkitpy/layout_tests/port/base.py:
3749         * Scripts/webkitpy/layout_tests/port/server_process.py:
3750
3751 2011-07-01  Adam Roben  <aroben@apple.com>
3752
3753         Teach webkitpy about the new format of our ChangeLog template
3754
3755         r90229 moved the "Need a short description and bug URL" line to the top of the ChangeLog
3756         template. But webkitpy didn't know this, so the various rollout-related commands were
3757         leaving that line in the ChangeLog.
3758
3759         Fixes <http://webkit.org/b/63815> REGRESSION (r90229): webkit-patch rollout and sheriffbot
3760         rollout are broken
3761
3762         Reviewed by Dan Bates.
3763
3764         * Scripts/webkitpy/common/checkout/changelog.py:
3765         (ChangeLog.update_with_unreviewed_message): Changed to look for the "Need a short
3766         description and bug URL" line instead of the "Reviewed by" line as the first line of the
3767         ChangeLog template.
3768
3769         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
3770         (ChangeLogTest): Updated _new_entry_boilerplate to match what prepare-ChangeLog now
3771         generates.
3772
3773 2011-07-01  Adam Roben  <aroben@apple.com>
3774
3775         Make TestFailures stop loading old builds once we've explained all current failures
3776
3777         We used to do this, but some refactoring in r90165 broke this logic.
3778
3779         Fixes <http://webkit.org/b/63808> REGRESSION (r90165): TestFailures page always loads all
3780         builds back to the beginning of time
3781
3782         Reviewed by Dan Bates.
3783
3784         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
3785         (LayoutTestHistoryAnalyzer.prototype.start): Stop fetching old builds once we're told that
3786         all current failures have been explained.
3787
3788 2011-07-01  Adam Roben  <aroben@apple.com>
3789
3790         commit-log-editor reorders ChangeLog entries in unexpected ways
3791         https://bugs.webkit.org/show_bug.cgi?id=63804
3792
3793         Reviewed by Darin Adler.
3794
3795         commit-log-editor was changed in r46899 to move the bug title and URL to the top of the
3796         commit message in order to make git-based tools, which expect the first line of the commit
3797         message to be a summary of the change, to work better. But this reordering of the ChangeLog
3798         entry was unpredictable and unexpected.
3799
3800         I changed prepare-ChangeLog to put the bug title and URL at the top of the ChangeLog entry
3801         so that commit-log-editor doesn't have to move them. Then I changed commit-log-editor not to
3802         try to move them.
3803
3804         * Scripts/commit-log-editor: Partially reverted r46899. We no longer save the "Reviewed by"
3805         line and try to move it down below the bug title and URL. The order of the text in the
3806         ChangeLog is now preserved, except that we insert a "Patch by" line just above the "Reviewed
3807         by" line in cases where the committer and author are not the same person.
3808
3809         * Scripts/prepare-ChangeLog: Moved the "Reviewed by" line down below the bug title and URL.
3810
3811 2011-07-01  Mihnea Ovidenie  <mihnea@adobe.com>
3812
3813         Reviewed by Adam Roben.
3814
3815         [CSSRegions] Add build slave
3816         https://bugs.webkit.org/show_bug.cgi?id=62372
3817
3818         * BuildSlaveSupport/build.webkit.org-config/config.json:
3819         Add an entry for a release build of Mac.
3820
3821 2011-07-01  Sheriff Bot  <webkit.review.bot@gmail.com>
3822
3823         Unreviewed, rolling out r90213.
3824         http://trac.webkit.org/changeset/90213
3825         https://bugs.webkit.org/show_bug.cgi?id=63792
3826
3827         The SL bot has an errant HTTP server holding port 8080
3828         (Requested by abarth on #webkit).
3829
3830         * Scripts/run-webkit-tests:
3831
3832 2011-07-01  Carlos Garcia Campos  <cgarcia@igalia.com>
3833
3834         Unreviewed. Fix WebKit2 GTK build after r90163.
3835
3836         * MiniBrowser/gtk/WebBundle/WebBundleMain.c:
3837         (didCreatePage):
3838
3839 2011-07-01  Adam Barth  <abarth@webkit.org>
3840
3841         Exit faster when we get a keyboard interrupt.  Previously, we'd open
3842         Safari if you ^C NRWT after it found a failure.
3843
3844         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3845
3846 2011-07-01  Eric Seidel  <eric@webkit.org>
3847
3848         Reviewed by Adam Barth.
3849
3850         Switch Snow Leopard over to new-run-webkit-tests (except for wk2)
3851         https://bugs.webkit.org/show_bug.cgi?id=63786
3852
3853         We're down to one bug blocking the "move to NRWT" bug:
3854         https://bugs.webkit.org/show_bug.cgi?id=34984
3855         so we're trying transitioning Snow Leopard to NRWT as an experiment.
3856
3857         If we run into any blocking issues we'll of course roll this back out
3858         (or anyone reading this should feel encouraged to).
3859
3860         * Scripts/run-webkit-tests:
3861
3862 2011-07-01  Sheriff Bot  <webkit.review.bot@gmail.com>
3863
3864         Unreviewed, rolling out r90192.
3865         http://trac.webkit.org/changeset/90192
3866         https://bugs.webkit.org/show_bug.cgi?id=63788
3867
3868         Appears to have caused NRWT on Chromium WebKit Vista to hang
3869         (Requested by abarth on #webkit).
3870
3871         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3872         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
3873
3874 2011-06-30  Adam Barth  <abarth@webkit.org>
3875
3876         Reviewed by Eric Seidel.
3877
3878         Remove the concept of "being wedged" from new-run-webkit-tests
3879         https://bugs.webkit.org/show_bug.cgi?id=63767
3880
3881         Worker processes shouldn't ever become wedged.  My understanding is
3882         that this code was originally motivated by the old threading-based
3883         design but no longer servers any purpose.
3884
3885         Note: If we actually have a problem with the test harness getting
3886         stuck, buildbot will kill us.
3887
3888         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3889         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py:
3890         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
3891
3892 2011-06-30  Adam Barth  <abarth@webkit.org>
3893
3894         The win32 implementation of fork is sad about forking test-webkitpy.
3895         Hopefully this will heal the bot.
3896
3897         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
3898
3899 2011-06-30  Eric Seidel  <eric@webkit.org>
3900
3901         Reviewed by Adam Barth.
3902
3903         Update new-run-webkit-tests now that the wk2 shared skipped list has moved
3904         https://bugs.webkit.org/show_bug.cgi?id=63780
3905
3906         The shared list was moved from mac-wk2 to wk2 as part of
3907         http://trac.webkit.org/changeset/90125.
3908
3909         * Scripts/webkitpy/layout_tests/port/webkit.py:
3910         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
3911
3912 2011-06-30  Kenichi Ishibashi  <bashi@chromium.org>
3913
3914         Reviewed by Darin Adler.
3915
3916         Style Checker should flag removal of "developmentRegion = English;" from project.pbxproj.
3917         https://bugs.webkit.org/show_bug.cgi?id=62022
3918
3919         Adds XcodeProjectFileChecker to check removal of "developmentRegion".
3920
3921         * Scripts/webkitpy/style/checker.py: Added XcodeProjectFileChecker as a checker for *.pbxproj.
3922         * Scripts/webkitpy/style/checkers/xcodeproj.py: Added.
3923         * Scripts/webkitpy/style/checkers/xcodeproj_unittest.py: Added.
3924
3925 2011-06-30  Dirk Pranke  <dpranke@chromium.org>
3926
3927         Reviewed by Ojan Vafai.
3928
3929         nrwt: make sharding tests needing locks less hard-coded
3930         https://bugs.webkit.org/show_bug.cgi?id=63112
3931
3932         This change also changes the manager logic so that it will
3933         drop the server lock as soon as all of the shards requiring
3934         the lock have completed.
3935
3936         This change includes some minor namespace/import changes in the
3937         unit tests, and also makes the Manager a new-style object, which
3938         it should've been all along.
3939
3940         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3941         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
3942
3943 2011-06-30  Adam Barth  <abarth@webkit.org>
3944
3945         Reviewed by Eric Seidel.
3946
3947         Remove threaded mode from new-run-webkit-tests
3948         https://bugs.webkit.org/show_bug.cgi?id=63771
3949
3950         This mode is not used and is buggy.  Rather than carry around a bunch
3951         of unused buggy code, we should rip it out and focus on the
3952         multiprocess implementation.
3953
3954         * Scripts/webkitpy/common/system/executive.py:
3955         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py:
3956         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
3957         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
3958         * Scripts/webkitpy/layout_tests/port/base.py:
3959         * Scripts/webkitpy/layout_tests/port/chromium.py:
3960         * Scripts/webkitpy/layout_tests/port/mac.py:
3961         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
3962         * Scripts/webkitpy/layout_tests/port/server_process.py:
3963         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3964         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
3965
3966 2011-06-30  Adam Barth  <abarth@webkit.org>
3967
3968         Reviewed by Darin Adler.
3969
3970         Clean up output from new-run-webkit-tests
3971         https://bugs.webkit.org/show_bug.cgi?id=63759
3972
3973         Printing messages from the child process looks super ugly because of
3974         the way the pretty-printer works.  Printing a blank line first is a
3975         hack, but it makes things at least partially sane.
3976
3977         Also, handle the case where calling sample throws an exception.
3978
3979         * Scripts/webkitpy/layout_tests/port/server_process.py:
3980         * Scripts/webkitpy/layout_tests/port/server_process_unittest.py:
3981
3982 2011-06-27  Diego Gonzalez  <diegohcg@webkit.org>
3983
3984         Reviewed by Antonio Gomes.
3985
3986         [Qt] DRT: Provide LayoutTestController::setDefersLoading and ::goBack support
3987         https://bugs.webkit.org/show_bug.cgi?id=63489
3988
3989         Unskipped:
3990             loader/navigation-while-deferring-loads.html
3991
3992         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3993         (LayoutTestController::goBack):
3994         (LayoutTestController::setDefersLoading):
3995         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3996
3997 2011-06-30  Adam Barth  <abarth@webkit.org>
3998
3999         Reviewed by Dirk Pranke.
4000
4001         new-run-webkit-tests should stop when I ^C it
4002         https://bugs.webkit.org/show_bug.cgi?id=63763
4003
4004         It still doesn't stop fast enough, but at least with this patch it
4005         doesn't run more tests.
4006
4007         Also, I tweaked the UI that happens when there's a keyboard interrupt
4008         to not be amazingly ugly.  There's still a garbage "Te" left on the
4009         console, but that's a problem for another day.
4010
4011         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
4012
4013 2011-06-30  Adam Roben  <aroben@apple.com>
4014
4015         Make TestFailures never treat a too-many-failures build as a last-passing build for a test
4016
4017         Fixes <http://webkit.org/b/63758> TestFailures sometimes incorrectly says that a build was
4018         the last passing build for a test, when in fact the test wasn't run due to too many failures
4019         causing old-run-webkit-tests to exit early
4020
4021         Reviewed by Anders Carlsson.
4022
4023         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
4024         (LayoutTestHistoryAnalyzer.prototype._incorporateBuildHistory): If there were too many
4025         failures in this build, assume that all the tests failed in the last build would have failed
4026         in this build, too, had they been run. Introduced some new local variables to reduce typing,
4027         and used them throughout.
4028
4029         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
4030         (LayoutTestResultsLoader.prototype.start): Fix a regression from r89841 where we wouldn't
4031         correctly remember whether old-run-webkit-tests exited early due to too many failures for
4032         builds older than r89610. Also removed an unused parameter while I was at it.
4033
4034 2011-06-30  Adam Roben  <aroben@apple.com>
4035
4036         Make TestFaiulres only load old-results directories as needed
4037
4038         Fixes <http://webkit.org/b/63752> Tester pages on TestFailures page load very slowly
4039
4040         Reviewed by Anders Carlsson.
4041
4042         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
4043         (Builder.prototype.getBuildNames): Moved up into the API section of the class. Now just
4044         calls through to _getBuildNamesFromResultsDirectory.
4045         (Builder.prototype.getOldBuildNames): Added. Just calls through to
4046         _getBuildNamesFromResultsDirectory.
4047         (Builder.prototype._getBuildNamesFromResultsDirectory): Renamed from getBuildNames. Now
4048         takes the directory URL as an argument and only fetches that single URL.
4049
4050         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
4051         (LayoutTestHistoryAnalyzer.prototype.start): Moved most logic to _analyzeBuilds. First
4052         analyzes builds from Builder.getBuildNames, then from Builder.getOldBuildNames if needed.
4053         (LayoutTestHistoryAnalyzer.prototype._analyzeBuilds): Moved logic here from start. (Most
4054         changes are just indentation.) Now takes a callback to call when we've finished analyzing
4055         all builds in buildNames so that we can try to fetch more build names if needed.
4056
4057 2011-06-30  Mark Rowe  <mrowe@apple.com>
4058
4059         Reviewed by Anders Carlsson.
4060
4061         Update MiniBrowser and WebKitTestRunner to use the client version constants.
4062
4063         * MiniBrowser/gtk/BrowserWindow.c:
4064         (browserWindowLoaderClientInit):
4065         (browserWindowUIClientInit):
4066         * MiniBrowser/gtk/WebBundle/WebBundleMain.c:
4067         (didCreatePage):
4068         (WKBundleInitialize):
4069         * MiniBrowser/gtk/main.c:
4070         (createWKContextWithInjectedBundle):
4071         * MiniBrowser/mac/AppDelegate.m:
4072         (-[BrowserAppDelegate init]):
4073         * MiniBrowser/mac/BrowserWindowController.m:
4074         (-[BrowserWindowController awakeFromNib]):
4075         * MiniBrowser/mac/WebBundle/WebBundleMain.m:
4076         (WKBundleInitialize):
4077         * MiniBrowser/win/BrowserView.cpp:
4078         (BrowserView::create):
4079         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
4080         (WTR::InjectedBundle::initialize):
4081         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
4082         (WTR::InjectedBundlePage::InjectedBundlePage):
4083         * WebKitTestRunner/TestController.cpp:
4084         (WTR::TestController::createOtherPage):
4085         (WTR::TestController::initialize):
4086
4087 2011-06-30  Adam Roben  <aroben@apple.com>
4088
4089         Identify revisions that modified failing tests on TestFailures page
4090
4091         Fixes <http://webkit.org/b/63716> TestFailures page should try to infer which commit caused
4092         a failure by looking at commit logs
4093
4094         Reviewed by David Kilzer.
4095
4096         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
4097         (.existing-bugs-list, .suspect-revisions-list): Made this rule apply to the suspect
4098         revisions list, too.
4099
4100         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Trac.js:
4101         (Trac): Initialize our memory cache.
4102         (Trac.prototype.getCommitDataForRevisionRange): Added. Fetches the commit log in RSS form,
4103         then parses out the revision number, Trac's idea of the commit title, and the files modified
4104         by the commit. (To get the modified files, we rely on the commit message including a
4105         prepare-ChangeLog-style file list.)
4106         (Trac.prototype.logURL): Added a new formatAsRSS parameter.
4107
4108         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
4109         (removePathExtension): Added. Returns a new string with the last dot and everything after it
4110         removed.
4111         (sorted): Added sortFunction parameter.
4112         (Node.prototype.removeAllChildren): Added. Does what it says.
4113
4114         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
4115         (ViewController.prototype._displayBuilder): Updated for change to _domForRegressionRange.
4116         (ViewController.prototype._domForRegressionRange): Changed argument order to match
4117         _domForNewAndExistingBugs. Now takes the list of failing tests and uses it to figure out
4118         which revisions modified the failing tests and displays the suspect revisions in a list.
4119
4120 2011-06-30  Eric Seidel  <eric@webkit.org>
4121
4122         Reviewed by Adam Barth.
4123
4124         Rename should_kill_dump_render_tree to driver_needs_restart in preparation for supporting WebKitTestRunner
4125         https://bugs.webkit.org/show_bug.cgi?id=63743
4126
4127         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
4128         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
4129
4130 2011-06-30  Martin Robinson  <mrobinson@igalia.com>
4131
4132         Reviewed by Anders Carlsson.
4133
4134         [GTK] Crash observed with nspluginwrapper and flash
4135         https://bugs.webkit.org/show_bug.cgi?id=62249
4136
4137         Added a TestNetscapePlugin test which verifies that WebKit properly
4138         handles situations where InvalidateRect is called with a null instance.
4139
4140         * DumpRenderTree/TestNetscapePlugIn/Tests/x11/CallInvalidateRectWithNullNPPArgument.cpp: Added.
4141         (CallInvalidateRectWithNullNPPArgument::CallInvalidateRectWithNullNPPArgument):
4142         (CallInvalidateRectWithNullNPPArgument::NPP_New):
4143         * GNUmakefile.am: Add the new file to sources list.
4144
4145 2011-06-30  Eric Seidel  <eric@webkit.org>
4146
4147         Reviewed by Adam Barth.
4148
4149         Remove redundant tests in TestFailuresTest
4150         https://bugs.webkit.org/show_bug.cgi?id=63742
4151
4152         These tests are all already covered by TestFailuresTest.test_loads.
4153         I also added the missing "T" from TestFailuresest. :)
4154
4155         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
4156
4157 2011-06-30  Adam Barth  <abarth@webkit.org>
4158
4159         Reviewed by Eric Seidel.
4160
4161         REGRESSION(r84294): new-run-webkit-tests results.html generate links to diffs.html or diff.png that don't exist
4162         https://bugs.webkit.org/show_bug.cgi?id=63494
4163
4164         This patch does two things:
4165
4166         1) Refactor diff_image to return the image diff instead of writing the
4167         diff to a file.  This fixes a bunch of hacks and disentangles a bunch
4168         of code.
4169
4170         2) When there's a checksum mismatch but not image diff, we no longer
4171         report an IMAGE failure to results.html.  That fixes the bug in
4172         question because results.html won't try to link to a non-existent image
4173         diff.
4174
4175         * Scripts/webkitpy/common/system/executive_mock.py:
4176         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
4177         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py:
4178         * Scripts/webkitpy/layout_tests/port/base.py:
4179         * Scripts/webkitpy/layout_tests/port/chromium.py:
4180         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
4181         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
4182         * Scripts/webkitpy/layout_tests/port/test.py:
4183         * Scripts/webkitpy/layout_tests/port/webkit.py:
4184         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
4185         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests.py:
4186
4187 2011-06-28  Hans Wennborg  <hans@chromium.org>
4188
4189         Reviewed by Tony Chang.
4190
4191         IndexedDB: Prepare for running layout tests with LevelDB
4192         https://bugs.webkit.org/show_bug.cgi?id=63593
4193
4194         The TestShell must always provide a temporary folder for LevelDB,
4195         not just when the backing store type is overridden.
4196
4197         * DumpRenderTree/chromium/LayoutTestController.cpp:
4198         (LayoutTestController::setOverrideIndexedDBBackingStore):
4199         (LayoutTestController::clearAllDatabases):
4200         * DumpRenderTree/chromium/LayoutTestController.h:
4201         * DumpRenderTree/chromium/TestShell.cpp:
4202         (TestShell::TestShell):
4203         * DumpRenderTree/chromium/TestShell.h:
4204
4205 2011-06-30  Zsolt Fehér  <h490267@stud.u-szeged.hu>
4206
4207         Reviewed by Csaba Osztrogonác.
4208
4209         [WK2] Create a general wk2 specific LayoutTests/platform/wk2 platform