3e8ac560c820b7ae535161f89672e7e8e2326e4f
[WebKit-https.git] / Tools / ChangeLog
1 2011-07-14  Ryosuke Niwa  <rniwa@webkit.org>
2
3         REGRESSION: webkit-patch roll-chromium-deps is broken
4         https://bugs.webkit.org/show_bug.cgi?id=64568
5
6         Reviewed by Dirk Pranke
7
8         Use _filesystem.join.
9
10         * Scripts/webkitpy/common/checkout/checkout.py:
11         * Scripts/webkitpy/common/checkout/checkout_unittests.py:
12
13 2011-07-14  Noam Rosenthal  <noam.rosenthal@nokia.com>
14
15         Adding myself to the reviewers list; No review needed.
16
17         * Scripts/webkitpy/common/config/committers.py:
18
19 2011-07-14  Adam Roben  <aroben@apple.com>
20
21         Don't use Element.prototype.classList in TestFailures
22
23         Safari 5 doesn't support it.
24
25         Fixes <http://webkit.org/b/64550> Can't expand flaky tests on TestFailures page in Safari 5
26
27         Reviewed by Daniel Bates.
28
29         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
30         (Element.prototype.hasStyleClass):
31         (Element.prototype.addStyleClass):
32         (Element.prototype.removeStyleClass):
33         (Element.prototype.toggleStyleClass):
34         Added these helper functions which simulate classList functionality.
35
36         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities_unittests.js:
37         Added. Tests for the above.
38
39         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
40         (ViewController.prototype._domForPossiblyFlakyTests): Changed to use
41         toggleStyleClass/hasStyleClass instead of classList.
42
43         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/run-unittests.html:
44         Added Utilities_unittests.js, and reordered the tested files to be in
45         roughly dependency order (i.e., the lowest-level files are imported
46         and tested first).
47
48 2011-07-14  Eric Seidel  <eric@webkit.org>
49
50         NRWT doesn't store the svn revision in full_results.json on chromium-win
51         https://bugs.webkit.org/show_bug.cgi?id=64492
52
53         Unreviewed.  Just fixing my test-webkitpy regression.
54
55         Fix the unit tests.  Unfortunately scm does not use a filesystem
56         object so we can't control the result of detect_scm_system.
57         When detect_scm_system would fail, we would log, which would
58         cause all passing_run integration tests to fail.
59
60         * Scripts/webkitpy/common/checkout/scm/scm.py:
61         * Scripts/webkitpy/layout_tests/controllers/manager.py:
62         * Scripts/webkitpy/layout_tests/port/base.py:
63         * Scripts/webkitpy/layout_tests/port/test.py:
64         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
65
66 2011-07-14  Eric Seidel  <eric@webkit.org>
67
68         NRWT doesn't store the svn revision in full_results.json on chromium-win
69         https://bugs.webkit.org/show_bug.cgi?id=64492
70
71         Reviewed by Ojan Vafai.
72
73         This should fix the bug.  Unfortunately this code is currently
74         impossible to test since it's impossible to mock detect_scm_system at the
75         moment.  I started re-writing scm.detection.py to be mockable, but decided that
76         was best left for another day.
77
78         * Scripts/webkitpy/layout_tests/controllers/manager.py:
79         * Scripts/webkitpy/layout_tests/port/base.py:
80
81 2011-07-14  Ojan Vafai  <ojan@chromium.org>
82
83         fix coding style of dashboard_base.js
84         https://bugs.webkit.org/show_bug.cgi?id=64545
85
86         Reviewed by Adam Roben.
87
88         The code changes to the files other than dashboard_base.js
89         are just fallout from renaming globals in dashboard_base.js.
90
91         * TestResultServer/static-dashboards/aggregate_results.html:
92         * TestResultServer/static-dashboards/builders.js:
93         * TestResultServer/static-dashboards/dashboard_base.js:
94         * TestResultServer/static-dashboards/flakiness_dashboard.html:
95         * TestResultServer/static-dashboards/flakiness_dashboard_tests.js:
96         * TestResultServer/static-dashboards/timeline_explorer.html:
97         * TestResultServer/static-dashboards/treemap.html:
98
99 2011-07-14  Carlos Garcia Campos  <cgarcia@igalia.com>
100
101         Unreviewed. Fix WebKit2 GTK build after r90953.
102
103         * WebKitTestRunner/GNUmakefile.am:
104
105 2011-07-14  Adam Barth  <abarth@webkit.org>
106
107         garden-o-matic results pane should be more discoverable
108         https://bugs.webkit.org/show_bug.cgi?id=64513
109
110         Reviewed by Eric Seidel.
111
112         This patch causes us to trigger the results pane on mouse clicks
113         instead of mousein.  Also, handle the case of no results details more
114         elegantly.
115
116         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
117         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
118             - Change the binding to "click".
119         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
120             - Handle the "no results URLs" case explicitly instead of never
121               calling the callback.
122         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
123         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
124             - Show a message when there are no result details to show.
125         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
126
127 2011-07-14  Adam Barth  <abarth@webkit.org>
128
129         garden-o-matic should badge TIMEOUT failures
130         https://bugs.webkit.org/show_bug.cgi?id=64435
131
132         Reviewed by Eric Seidel.
133
134         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
135         (.regression .what a[draggable].TIMEOUT::before):
136             - Add TIMEOUT.
137
138 2011-07-14  Adam Barth  <abarth@webkit.org>
139
140         Fix Python exception blocking the commit-queue.
141
142         * Scripts/webkitpy/common/checkout/scm/git.py:
143
144 2011-07-13  Eric Seidel  <eric@webkit.org>
145
146         Move webkitpy off of loose mocks
147         https://bugs.webkit.org/show_bug.cgi?id=64508
148
149         Reviewed by Adam Barth.
150
151         Using Mock has caused us more pain than help.
152         It's possible that there was a cleaner way to use it
153         (maybe Mock(class) instead of inheriting from it?).
154         But for now, I've removed all uses of Mock from mocktool.py.
155
156         I also moved run_command into the only 3 files which call it
157         instead of leaving the deprecated method in executive.py.
158
159         I changed various direct calls to os.* to use filesystem instead.
160
161         * Scripts/webkitpy/common/checkout/checkout.py:
162         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
163         * Scripts/webkitpy/common/checkout/scm/git.py:
164         * Scripts/webkitpy/common/checkout/scm/scm.py:
165         * Scripts/webkitpy/common/checkout/scm/scm_unittest.py:
166         * Scripts/webkitpy/common/checkout/scm/svn.py:
167         * Scripts/webkitpy/common/system/executive.py:
168         * Scripts/webkitpy/common/system/executive_unittest.py:
169         * Scripts/webkitpy/tool/mocktool.py:
170         * Scripts/webkitpy/tool/steps/cleanworkingdirectory_unittest.py:
171         * Scripts/webkitpy/tool/steps/ensurelocalcommitifneeded.py:
172
173 2011-07-13  Eric Seidel  <eric@webkit.org>
174
175         NRWT doesn't store the svn revision in full_results.json on chromium-win
176         https://bugs.webkit.org/show_bug.cgi?id=64492
177
178         Reviewed by Adam Barth.
179
180         Add a warning message in the revision='' case to help us diagnose
181         what's going wrong on the bot.
182
183         * Scripts/webkitpy/layout_tests/controllers/manager.py:
184
185 2011-07-13  Dirk Pranke  <dpranke@chromium.org>
186
187         new-run-webkit-tests: print baseline search path as part of config output
188         https://bugs.webkit.org/show_bug.cgi?id=64499
189
190         Reviewed by Eric Seidel.
191
192         * Scripts/webkitpy/layout_tests/controllers/manager.py:
193         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
194
195 2011-07-13  Adam Barth  <abarth@webkit.org>
196
197         gardening server should have an API for parsing changelogs
198         https://bugs.webkit.org/show_bug.cgi?id=64495
199
200         Reviewed by Eric Seidel.
201
202         This patch exposes much of the same information from CommitInfo in a
203         dictionary form, which is easier to send over-the-wire as JSON to the
204         frontend.
205
206         * Scripts/webkitpy/common/checkout/checkout.py:
207         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
208         * Scripts/webkitpy/tool/mocktool.py:
209         * Scripts/webkitpy/tool/servers/gardeningserver.py:
210         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py:
211         * Scripts/webkitpy/tool/servers/reflectionhandler.py:
212
213 2011-07-13  Eric Seidel  <eric@webkit.org>
214
215         REGRESSION: GitTestWithMock.test_create_patch fails
216         https://bugs.webkit.org/show_bug.cgi?id=62945
217
218         Reviewed by Daniel Bates.
219
220         I was not able to reproduce the exact failure seen in the bug,
221         however this test was failing on my machine for other reasons.
222
223         I went through and did an audit of our run_command usage, it's
224         entirely in scm classes after this change.  (Not surprising given
225         that scm.py was the second file ever created in webkit.py.)
226
227         The real bug I'm fixing here is that we were setting executive.should_log
228         when the value had changed to executive._should_log.  Now we set the right one
229         and the test works again.
230
231         * Scripts/webkitpy/common/checkout/checkout.py:
232         * Scripts/webkitpy/common/checkout/scm/git.py:
233         * Scripts/webkitpy/common/checkout/scm/scm.py:
234         * Scripts/webkitpy/common/checkout/scm/scm_unittest.py:
235         * Scripts/webkitpy/common/checkout/scm/svn.py:
236
237 2011-07-13  Ilya Sherman  <isherman@chromium.org>
238
239         Fix WTF header guard style check
240         https://bugs.webkit.org/show_bug.cgi?id=64488
241
242         Reviewed by David Levin.
243
244         * Scripts/webkitpy/style/checkers/cpp.py:
245         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Verify that we don't *always* suggest "WTF_" as a prefix
246
247 2011-07-13  Ojan Vafai  <ojan@chromium.org>
248
249         bring flakiness_dashboard.html closer to webkit style
250         https://bugs.webkit.org/show_bug.cgi?id=64477
251
252         Reviewed by Adam Barth.
253
254         * TestResultServer/static-dashboards/dashboard_base.js:
255         Fix bad variable name that would hit only when running the tests.
256         * TestResultServer/static-dashboards/flakiness_dashboard.html:
257         No code/logic changes. All moving/removing brackets, indents and moving things
258         to one line.
259
260 2011-07-13  Xan Lopez  <xlopez@igalia.com>
261
262         [GTK] Fix distcheck
263
264         Reviewed by Martin Robinson.
265
266         * WebKitTestRunner/GNUmakefile.am: mark built sources as nodist.
267
268 2011-07-13  Joseph Pecoraro  <joepeck@webkit.org>
269
270         Unreviewed fix. Add a "\n" to a printf that somehow got lost.
271
272         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
273         (WebCore::DumpRenderTree::dumpApplicationCacheQuota):
274
275 2011-07-13  Joseph Pecoraro  <joepeck@webkit.org>
276
277         ApplicationCache Quota Output is Flakey
278         https://bugs.webkit.org/show_bug.cgi?id=64410
279
280         Reviewed by Alexey Proskuryakov.
281
282         Unify the delegate logging for reaching application cache
283         quotas. Truncate the space needed to the nearest 10000
284         for less flakey test output. This also better supports
285         printing NSUInteger on non-64 bit systems by casting to
286         an unsigned long.
287
288         * DumpRenderTree/mac/UIDelegate.mm:
289         (-[UIDelegate webView:exceededApplicationCacheOriginQuotaForSecurityOrigin:totalSpaceNeeded:]):
290         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
291         (WebCore::DumpRenderTree::dumpApplicationCacheQuota):
292
293 2011-07-13  Dimitri Glazkov  <dglazkov@chromium.org>
294
295         Remove unused TestExpectations._overrides.
296         https://bugs.webkit.org/show_bug.cgi?id=64470
297
298         Reviewed by Adam Barth.
299
300         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Removed unused member.
301
302 2011-07-13  Dimitri Glazkov  <dglazkov@chromium.org>
303
304         Introduce TestExpectationSerializer.list_to_string.
305         https://bugs.webkit.org/show_bug.cgi?id=64462
306
307         Reviewed by Adam Barth.
308
309         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Added list_to_string and change the relevant callsite to use it.
310         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: Added tests for it.
311
312 2011-07-13  Dimitri Glazkov  <dglazkov@chromium.org>
313
314         Consolidate expectations parsing code.
315         https://bugs.webkit.org/show_bug.cgi?id=64460
316
317         Reviewed by Adam Barth.
318
319         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Folded TestExpectationParser._split_expectation_string into TestExpectationParser.parse.
320
321 2011-07-12  Brent Fulgham  <bfulgham@webkit.org>
322
323         Standardize WinCairo conditionalized code under PLATFORM macro.
324         https://bugs.webkit.org/show_bug.cgi?id=64377
325
326         Reviewed by Maciej Stachowiak.
327
328         Update compile-time conditions that had previously used #ifdef
329         of WIN_CAIRO to consistently use PLATFORM(WIN_CAIRO).
330
331         * DumpRenderTree/config.h: Switch to PLATFORM(WIN_CAIRO)
332         * DumpRenderTree/win/DumpRenderTreeCairo.vsprops: Remove redundant
333           define of WIN_CAIRO=1.
334         * DumpRenderTree/win/DumpRenderTreeDebugCairoCFLite.vsprops: Add
335           WinCairo.vsprops to set (to match Release build).
336
337 2011-07-13  Dimitri Glazkov  <dglazkov@chromium.org>
338
339         Eliminate TestExpectationsFile.
340         https://bugs.webkit.org/show_bug.cgi?id=64458
341
342         Turns out, we can just use a Python list.
343
344         Reviewed by Adam Barth.
345
346         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Folded TestExpectationsFile.append into TestExpectationParser.parse_list,
347             removed TestExpectationsFile.
348         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: Moved tests to better reflect new names, removed iterator test,
349             since there's no more custom iterator machinery.
350
351 2011-07-13  Philippe Normand  <pnormand@igalia.com>
352
353         Unreviewed, added my other email addresses.
354
355         * Scripts/webkitpy/common/config/committers.py:
356
357 2011-07-13  Adam Barth  <abarth@webkit.org>
358
359         Folks should only be listed once.
360
361         * Scripts/webkitpy/common/config/committers.py:
362
363 2011-07-13  Philippe Normand  <pnormand@igalia.com>
364
365         Unreviewed, adding myself as Reviewer.
366
367         * Scripts/webkitpy/common/config/committers.py:
368
369 2011-07-13  Adam Barth  <abarth@webkit.org>
370
371         Fix flickering bug introduced by my previous patch.  I forgot to change
372         the name of the class everywhere.
373
374         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
375
376 2011-07-12  Philippe Normand  <pnormand@igalia.com>
377
378         [GTK] media/media-blocked-by-willsendrequest.html fails
379         https://bugs.webkit.org/show_bug.cgi?id=63699
380
381         Reviewed by Martin Robinson.
382
383         * DumpRenderTree/gtk/DumpRenderTree.cpp:
384         (willSendRequestCallback): Abort the request if explicitely asked
385         by the LayoutTestController's willSendRequestReturnsNull() function.
386
387 2011-07-13  Adam Barth  <abarth@webkit.org>
388
389         Remove "Dismiss" button from garden-o-matic butterbar
390         https://bugs.webkit.org/show_bug.cgi?id=64443
391
392         Reviewed by Dimitri Glazkov.
393
394         We don't have any persistent butter bar messages yet, so the dismiss
395         button is premature (and fairly heavy, visually).
396
397         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
398         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
399
400 2011-07-13  Adam Barth  <abarth@webkit.org>
401
402         Refine garden-o-matic status pane
403         https://bugs.webkit.org/show_bug.cgi?id=64442
404
405         Reviewed by Dimitri Glazkov.
406
407         This patch makes two improvements to the status pane:
408
409         1) We only query the server for failure types that we're expecting.
410            This dramatically reduces the number of HTTP requests, making
411            loading the status pane faster.
412
413         2) The status pane now displays which test and which builder it is
414            showing results for.
415
416         * Scripts/webkitpy/tool/servers/data/gardeningserver/config.js:
417         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
418         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
419         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
420         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
421         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
422         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
423
424 2011-07-13  Adam Barth  <abarth@webkit.org>
425
426         garden-o-matic should show test results
427         https://bugs.webkit.org/show_bug.cgi?id=64440
428
429         Reviewed by Adam Roben.
430
431         This patch introduces the results detail pane, which appears at the
432         bottom of the window and contains test results from the bots.
433         Currently, you can activate the pane by mousing over one of the builder
434         names associated with a failing test.
435
436         This is just a first iteration of the UI.  There's no way to resize or
437         zoom in on elements of the details pane, and images likely aren't sized
438         correctly, but it's a place to start.
439
440         * Scripts/webkitpy/tool/servers/data/gardeningserver/config.js:
441             - Add constants for our data attributes so we don't typo them!
442         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
443             - Add DOM structure for the details pane.
444         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
445         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
446             - Bind events for showing and hiding the details pane.
447         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
448             - Reduce the number of result types that we query for to avoid
449               overfilling the details pane.  We'll probably need another
450               solution here in the long-term.  For example, we could use tabs
451               to pack more results into the pane.
452         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
453         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html:
454             - Now that we're using the config package during testing, we need
455               to include it in the testing HTML.
456         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
457         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
458             - Change the DOM structure of the results detail to make it fit
459               nicely in the details pane (rather than flowing freely in the
460               body, as it did before).
461
462 2011-07-13  Adam Roben  <aroben@apple.com>
463
464         Make TestFailures show every time a possibly-flaky test failed, but hide it by default
465
466         It's useful to be able to see every time a flaky test failed to see whether it failed the
467         same way every time. But doing so takes a lot of space, so the list of failures is now
468         collapsed by default and can be revealed using a disclosure triangle.
469
470         Fixes <http://webkit.org/b/64455> TestFailures page doesn't show as much information for
471         flaky tests as I would like, even though the page is already so long
472
473         Reviewed by Dimitri Glazkov.
474
475         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/FlakyLayoutTestDetector.js:
476         (FlakyLayoutTestDetector.prototype.allFailures): Replaced flakinessExamples with this
477         function. Now returns all failures for the given test.
478
479         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/FlakyLayoutTestDetector_unittests.js:
480         Added. This just contains some simple tests of the FlakyLayoutTestDetector class. We'll add
481         more over time.
482
483         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
484         (LayoutTestHistoryAnalyzer.prototype.start): Updated the documentation comment to reflect
485         that we no longer return passing builds for possibly-flaky tests.
486
487         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
488         (.existing-bugs-list, .suspect-revisions-list, .flakiness-examples-list): Make the list of
489         flakiness examples small, too, since it can get quite long.
490
491         (.disclosure-triangle):
492         (.expanded > .disclosure-triangle):
493         Simple styles for the disclosure triangle.
494
495         (.flakiness-examples-list): Collapse the list by default.
496
497         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
498         (ViewController.prototype._displayBuilder): Pass the total number of builds analyzed to
499         _domForPossiblyFlakyTests.
500         (ViewController.prototype._domForPossiblyFlakyTests): Put a disclosure triangle to the left
501         of each test name, and the number of failures to the right. When the disclosure triangle is
502         clicked for the first time, we build up the list of failures and expand the element. After
503         that we just collapse or expand the element on subsequent clicks.
504
505         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/run-unittests.html:
506         Pulled in new tests.
507
508 2011-07-13  Adam Roben  <aroben@apple.com>
509         
510         Teach TestFailures to understand NRWT's output when it exits early due to too many failures
511
512         Fixes <http://webkit.org/b/64456> TestFailures page reports way too many failures when NRWT
513         exits early
514
515         Reviewed by Dimitri Glazkov.
516
517         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
518         (Builder.prototype.getNumberOfFailingTests): Relaxed the "Exiting early" test not to require
519         it to be at the beginning of the line, since NRWT prints a bunch of junk earlier in the
520         line. Tightened up the regex that's used to parse the number of failing tests to require the
521         leading number to be followed by whitespace so that we won't parse the "2011" in
522         "2011-07-13" as a number of failures.
523
524         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder_unittests.js:
525         Added a test for the above.
526
527 2011-07-12  Dimitri Glazkov  <dglazkov@chromium.org>
528
529         Extract model-like TestExpectationLine and TestExpectationFile from TestExpectations.
530         https://bugs.webkit.org/show_bug.cgi?id=64386
531
532         This is the first step in converting TestExpectations to a real model.
533         * TestExpectationsLine represents a line in the test_expectations.txt file, and
534         * TestExpectationsFile represents the file, which is a collection of lines.
535
536         Reviewed by Adam Barth.
537
538         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
539         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
540         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
541
542 2011-07-13  Xan Lopez  <xlopez@igalia.com>
543
544         [GTK] Do not grab focus too early in DRT.
545
546         Reviewed by Gustavo Noronha.
547
548         It causes a layout to happen and a progress signal to be emitted
549         since r90900, but at this point we don't have a
550         LayoutTestController object and we'll eventually crash. Since we
551         already grab focus at the beginning of runTest() this is
552         redundant, so get rid of it to fix the crash.
553
554         * DumpRenderTree/gtk/DumpRenderTree.cpp:
555         (main): remove call to grab_focus
556
557 2011-07-13  Sheriff Bot  <webkit.review.bot@gmail.com>
558
559         Unreviewed, rolling out r90893 and r90894.
560         http://trac.webkit.org/changeset/90893
561         http://trac.webkit.org/changeset/90894
562         https://bugs.webkit.org/show_bug.cgi?id=64441
563
564         NRWT still doesn't work on qt-mac platform (Requested by Ossy
565         on #webkit).
566
567         * Scripts/run-webkit-tests:
568         (useNewRunWebKitTests):
569
570 2011-07-13  Csaba Osztrogonác  <ossy@webkit.org>
571
572         [Qt] NRWT should pick up the right httpd config file
573         https://bugs.webkit.org/show_bug.cgi?id=64086
574
575         * Scripts/run-webkit-tests: Enable NRWT on qt-mac platform after r90810.
576         (useNewRunWebKitTests):
577
578 2011-07-12  Mark Rowe  <mrowe@apple.com>
579
580         Fix the 32-bit build.
581
582         * DumpRenderTree/mac/UIDelegate.mm:
583         (-[UIDelegate webView:exceededApplicationCacheOriginQuotaForSecurityOrigin:totalSpaceNeeded:]):
584         Cast the NSUInteger value to unsigned long to match the format specifier.
585
586 2011-07-12  Adam Barth  <abarth@webkit.org>
587
588         Improve garden-o-matic UI when the bots fail to report revision numbers
589         https://bugs.webkit.org/show_bug.cgi?id=64427
590
591         Reviewed by Dimitri Glazkov.
592
593         At least Win (dbg)(2) seems to fail to report the SVN revision number
594         in full_results.json.  This patch makes garden-o-matic more robust to
595         missing revision numbers.
596
597         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
598         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
599         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
600         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
601         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
602
603 2011-07-12  Adam Barth  <abarth@webkit.org>
604
605         Adjust garden-o-matic layout to use fewer lines
606         https://bugs.webkit.org/show_bug.cgi?id=64422
607
608         Reviewed by Ojan Vafai.
609
610         This patch effectively merges the test name line with the list of
611         builders on which the test fails.  Each test failure now occupies two
612         lines instead of three.
613
614         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
615         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
616         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
617
618 2011-07-12  Adam Barth  <abarth@webkit.org>
619
620         garden-o-matic should label tests failures of new tests as such
621         https://bugs.webkit.org/show_bug.cgi?id=64421
622
623         Reviewed by Ojan Vafai.
624
625         Also, we shouldn't dim failures of new tests, even if we've only seen
626         them once because they're likely to be real problems that need
627         attention.
628
629         * Scripts/webkitpy/tool/servers/data/gardeningserver/base_unittests.js:
630         * Scripts/webkitpy/tool/servers/data/gardeningserver/checkout.js: Added.
631         * Scripts/webkitpy/tool/servers/data/gardeningserver/checkout_unittests.js: Added.
632         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
633         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
634         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
635         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
636         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html:
637         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
638         * Scripts/webkitpy/tool/servers/gardeningserver.py:
639
640 2011-07-12  Adam Barth  <abarth@webkit.org>
641
642         garden-o-matic should display how many times we've seen a failure
643         https://bugs.webkit.org/show_bug.cgi?id=64417
644
645         Reviewed by Ojan Vafai.
646
647         This patch adds some UI to display how many times we've seen a given
648         failure, which can be helpful for determining whether that failure is a
649         real failure or a flaky test.
650
651         When a failure has only been seen once (i.e., only a single run on a
652         single bot), we set the opacity of to 50% to avoid distracting the
653         gardener.
654
655         This patch also refactors the failure walker to have a simpler API
656         internally by moving from an object-oriented paradigm to a functional
657         paradigm.
658
659         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
660         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
661         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
662         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
663         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
664         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
665
666 2011-07-12  Chris Rogers  <crogers@google.com>
667
668         Enable Web Audio for chromium DRT
669         https://bugs.webkit.org/show_bug.cgi?id=64409
670
671         Reviewed by James Robinson.
672
673         * DumpRenderTree/chromium/TestShell.cpp:
674         (TestShell::TestShell):
675
676 2011-07-12  Adam Barth  <abarth@webkit.org>
677
678         Tweak some UI in garden-o-matic now that the tree actually has a
679         failure and I can see what this all looks like.
680
681         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
682         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
683         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
684
685 2011-07-12  Joseph Pecoraro  <joepeck@webkit.org>
686
687         Unreviewed. Skipping a few tests which fail due to differing output
688         in recent ApplicationCache quota tests. Reenabling the tests is
689         tracked by <http://webkit.org/b/64410>.
690
691         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
692         (WebCore::DumpRenderTree::dumpApplicationCacheQuota): This should
693         have been a space, not a newline.
694
695 2011-07-12  Adam Barth  <abarth@webkit.org>
696
697         garden-o-matic should display regression ranges
698         https://bugs.webkit.org/show_bug.cgi?id=64407
699
700         Reviewed by Dimitri Glazkov.
701
702         This patch computes a regression range for a failure by intersecting
703         the regression ranges seen by the various bots.  We make the underlying
704         assumption that a test is only failing due to one revision at any given
705         moment.  If that's not true, this code probably explodes.
706
707         The regression ranges appear asynchronously, which might be a jarring
708         UI.  We'll have to experiment to see.
709
710         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
711         * Scripts/webkitpy/tool/servers/data/gardeningserver/base_unittests.js:
712         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
713         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
714         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
715         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
716         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
717
718 2011-07-12  Adam Barth  <abarth@webkit.org>
719
720         cr-linux-ews should run pixel tests
721         https://bugs.webkit.org/show_bug.cgi?id=64394
722
723         Reviewed by Ojan Vafai.
724
725         This patch enabled pixel test on the cr-linux-ews.  Running a quick
726         test on the bots, it looks like we have five pixel failures, related to
727         some missing fonts.  I've also made the bots skip failing tests to
728         avoid uploading huge zip files to bugs.webkit.org (with all the exected
729         IMAGE failures).
730
731         * Scripts/webkitpy/common/config/ports.py:
732         * Scripts/webkitpy/common/config/ports_unittest.py:
733
734 2011-07-12  Joseph Pecoraro  <joepeck@webkit.org>
735
736         ApplicationCache update should not immediately fail when reaching per-origin quota
737         https://bugs.webkit.org/show_bug.cgi?id=64177
738
739         Reviewed by Alexey Proskuryakov.
740
741         Add layoutTestController.disallowIncreaseForApplicationCacheQuota()
742         to disable the default behavior of raising the default per-origin
743         quota to 5MB when reached. This allows us to test what happens
744         when the quota is reached and not increased.
745
746         * DumpRenderTree/LayoutTestController.cpp:
747         (LayoutTestController::LayoutTestController):
748         (disallowIncreaseForApplicationCacheQuotaCallback):
749         (LayoutTestController::staticFunctions):
750         * DumpRenderTree/LayoutTestController.h:
751         (LayoutTestController::disallowIncreaseForApplicationCacheQuota):
752         (LayoutTestController::setDisallowIncreaseForApplicationCacheQuota):
753         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
754         (LayoutTestController::reset):
755         * DumpRenderTree/qt/LayoutTestControllerQt.h:
756         (LayoutTestController::disallowIncreaseForApplicationCacheQuota):
757         (LayoutTestController::dumpApplicationCacheDelegateCallbacks):
758         Expose a function to prevent automatically increasing the per-origin
759         quota when the quota is reached. This allows us to test what happens
760         when a user would "disallow" a quota increase.
761
762         * DumpRenderTree/mac/UIDelegate.mm:
763         (-[UIDelegate webView:exceededApplicationCacheOriginQuotaForSecurityOrigin:spaceNeeded:]):
764         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
765         (WebCore::DumpRenderTree::dumpApplicationCacheQuota):
766         Respect the new disallowIncrease flag and dump the spaceNeeded values
767         for ports that implement application cache quotas.
768
769 2011-07-12  Ojan Vafai  <ojan@chromium.org>
770
771         fix flakiness dashboard to work with new crash log filenames
772         https://bugs.webkit.org/show_bug.cgi?id=64393
773
774         Reviewed by Adam Barth.
775
776         * TestResultServer/static-dashboards/flakiness_dashboard.html:
777
778 2011-07-12  Adam Barth  <abarth@webkit.org>
779
780         Add a happy animation to garden-o-matic when there are no failures
781         https://bugs.webkit.org/show_bug.cgi?id=64382
782
783         Reviewed by Ojan Vafai.
784
785         As requested by Ojan.
786
787         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
788         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
789         * Scripts/webkitpy/tool/servers/data/gardeningserver/partytime.gif: Added.
790         * Scripts/webkitpy/tool/servers/gardeningserver.py:
791
792 2011-07-12  Adam Barth  <abarth@webkit.org>
793
794         sheriffbot can't perform rollouts
795         https://bugs.webkit.org/show_bug.cgi?id=64370
796
797         Reviewed by Eric Seidel.
798
799         Now that we're not updating the working copy before processing every
800         IRC message, the rollout command needs to update it manually.
801
802         * Scripts/webkitpy/tool/bot/irc_command.py:
803         * Scripts/webkitpy/tool/bot/irc_command_unittest.py:
804
805 2011-07-12  Adam Barth  <abarth@webkit.org>
806
807         Fix the grammar in the FIXME comment.
808
809         * Scripts/webkitpy/common/net/resultsjsonparser.py:
810
811 2011-07-12  Eric Seidel  <eric@webkit.org>
812
813         NRWT should open test results page with Safari trunk, not the system provided one on Mac
814         https://bugs.webkit.org/show_bug.cgi?id=64346
815
816         Reviewed by Adam Barth.
817
818         To fix this I implemented Port.show_results_html_file in Mac, Gtk and Qt ports with
819         implementations (mostly) matching those found in old-run-webkit-tests.
820         There are still some minor differences for Qt which Qt hackers may wish to tweak.
821
822         I had to add a WebKitPort._port_flag_for_scripts method (similar to flag() in common.config.ports.py)
823         for the Qt/Gtk ports which always require a flag passed to scripts.
824
825         While trying to test this, I found that FactoryTest.test_chromium_gpu_linux
826         was using a real filesystem (due to assert_platform) which due to
827         global variables in config.py was causing set-webkit-configuration to have
828         an affect on unit test results!  So I fixed this by making FactoryTest.assert_port
829         pass a mock file system whenever calling factory.get.
830
831         Unfortunately TestPort was depending on always being passed a None filesystem
832         and asserting filesystem._tests (only true for unit_test_filesystem()).
833         So I just removed the FactoryTest.test_test and FactoryTest.test_dryrun tests
834         deciding that they were pretty much useless anyway.  If others feel strongly
835         I'm happy to fix this in a different way.
836
837         * Scripts/webkitpy/common/system/executive.py:
838          - default arguments in python are screwy.  They use a single shared
839            instance, so it's better to use argument=None and then argument = argument or Default()
840            if you have any chance of mutating (or returning) the default argument.
841         * Scripts/webkitpy/layout_tests/port/config.py:
842          - This code is wrong.  We don't need to use a global variable here (as far as I can tell).
843            I'm not fixing it in this patch, but I've marked it with a FIXME and we can convert to
844            storing the results of the read on the Config object (which should only be created once during normal operation).
845            Unit tests shouldn't be hitting the disk anyway.  It's possible Config should move off of Port and onto Tool/Host directly.
846         * Scripts/webkitpy/layout_tests/port/factory.py:
847         * Scripts/webkitpy/layout_tests/port/factory_unittest.py:
848         * Scripts/webkitpy/layout_tests/port/gtk.py:
849         * Scripts/webkitpy/layout_tests/port/gtk_unittest.py: Added.
850         * Scripts/webkitpy/layout_tests/port/mac.py:
851         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
852         * Scripts/webkitpy/layout_tests/port/qt.py:
853         * Scripts/webkitpy/layout_tests/port/qt_unittest.py:
854         * Scripts/webkitpy/layout_tests/port/webkit.py:
855
856 2011-07-12  Adam Roben  <aroben@apple.com>
857
858         Teach TestFailures to recognize when run-webkit-tests gets killed by buildbot
859
860         Fixes <http://webkit.org/b/64358> TestFailures page thinks all tests passed in
861         http://build.webkit.org/builders/Windows%207%20Release%20(Tests)/builds/14672
862
863         Reviewed by Daniel Bates.
864
865         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
866         (Builder.prototype.getNumberOfFailingTests): If run-webkit-tests exited with a non-zero
867         exit status but we didn't find any failure counts, assume that there was some error that
868         caused run-webkit-tests to die early (like being killed by buildbot due to a timeout).
869
870         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder_unittests.js:
871         Added a new test that shows that we get a failingTestCount of -1 when run-webkit-tests dies
872         early.
873         (runGetNumberOfFailingTestsTest): Moved most code here from the only pre-existing test in this
874         file.
875
876         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
877         (LayoutTestResultsLoader.prototype.start): Bump the cache number so old cached data that was
878         tainted by the bug fixed in this patch will be evicted.
879
880 2011-07-12  Adam Barth  <abarth@webkit.org>
881
882         cr-linux-ews complains about tests that aren't actually failing
883         https://bugs.webkit.org/show_bug.cgi?id=64367
884
885         The underlying problem here is that full_results.json doesn't have
886         enough information to determine whether a given test result was
887         expected because whether an actual result is expected depends on
888         whether full_results.json was generated duing a test run that included
889         pixel tests.
890
891         The right long-term solution is to make full_results.json a complete
892         description of what happened durning a test run.  In this patch
893         however, to stop the spam, this patch makes the jsonresultsparser
894         ignore pixel failures.
895
896         (I'm landing this patch unreviewed in order to stop the bots from
897         spamming.  I'll happily address any review comments in a follow-up
898         patch.)
899
900         * Scripts/webkitpy/common/net/resultsjsonparser.py:
901         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
902
903 2011-07-12  Adam Barth  <abarth@webkit.org>
904
905         Re-work garden-o-matic UI to begin to look like the mocks
906         https://bugs.webkit.org/show_bug.cgi?id=64334
907
908         Reviewed by Dimitri Glazkov.
909
910         This patch changes how we display failures to be more like the mock the
911         dglazkov made. We now can display the same information much more
912         compactly.
913
914         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
915         * Scripts/webkitpy/tool/servers/data/gardeningserver/base_unittests.js:
916         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
917         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css: Added.
918         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
919         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
920         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
921         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
922         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
923         * Scripts/webkitpy/tool/servers/gardeningserver.py:
924
925 2011-07-11  Adam Roben  <aroben@apple.com>
926
927         Extract TestFailures's bug-filing code into two new classes
928
929         Fixes <http://webkit.org/b/64300> TestFailures page's new-bug-filing code is a mess!
930
931         Reviewed by Darin Adler and Adam Barth.
932
933         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Buildbot.js:
934         (Buildbot.prototype.resultsDirectoryURL): Changed to return a URI-encoded URL. Otherwise the
935         URL isn't valid (and it's harder to mock this function).
936
937         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Buildbot_unittests.js:
938         Added. Just tests the above change (for now).
939
940         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder_unittests.js:
941         Added a license header, enclosed everything in a closure to avoid polluting the global
942         namespace, changed the test name to actually describe the passing condition, and replaced
943         uses of equals() with equal(). (The latter is the actual name of the function, and matches
944         deepEqual, etc.)
945
946         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/NewBugForm.js: Added.
947         (NewBugForm): This class knows how to construct a <form> used for filing new bugs in
948         Bugzilla based on some parameters.
949         (NewBugForm.prototype.domElement): Creates and returns the <form> element. Code came from
950         ViewController.prototype._domForNewAndExistingBugs.
951
952         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/NewBugForm_unittests.js:
953         Added. Tests the above code.
954
955         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailureBugForm.js: Added.
956         (TestFailureBugForm): This class knows how to construct a <form> element used for filing new
957         bugs specifically about test failures. Code came from
958         ViewController.prototype._domForNewAndExistingBugs.
959         (TestFailureBugForm.prototype.domElement): Slightly customizes the <form> element returned
960         by NewBugForm.
961         (TestFailureBugForm.prototype._computeOperatingSystem):
962         (TestFailureBugForm.prototype._computePlatform):
963         (TestFailureBugForm.prototype._createBugTitle):
964         (TestFailureBugForm.prototype._failingResultsHTMLURL):
965         (TestFailureBugForm.prototype._failingRevision):
966         (TestFailureBugForm.prototype._passingRevision):
967         (TestFailureBugForm.prototype._regressionRangeString):
968         Code came from ViewController.prototype._domForNewAndExistingBugs. I broke it out into
969         separate functions to break up the rat's nest a bit.
970
971         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailureBugForm_unittests.js:
972         Added. Tests the above code.
973
974         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
975         (ViewController.prototype._domForNewAndExistingBugs): Moved code from here to
976         TestFailureBugForm/NewBugForm, and changed this code to use a TestFailureBugForm.
977
978         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/index.html: Pull in
979         NewBugForm/TestFailureBugForm.js.
980
981         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/run-unittests.html:
982         Added new tests and required files.
983
984 2011-07-12  Adam Roben  <aroben@apple.com>
985
986         Test that no intermediate WTF::Strings are created when concatenating with string literals
987
988         Test for <http://webkit.org/b/63330> Concatenating string literals and WTF::Strings using
989         operator+ is suboptimal
990
991         Reviewed by Darin Adler.
992
993         * TestWebKitAPI/Tests/WTF/StringOperators.cpp: Added.
994         (TestWebKitAPI::TEST): Test that a bunch of different string concatenation expressions don't
995         create any intermediate WTF::Strings while they're being evaluated.
996
997         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
998         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
999         Added new file.
1000
1001 2011-07-12  Eric Seidel  <eric@webkit.org>
1002
1003         [Qt] NRWT should pick up the right httpd config file
1004         https://bugs.webkit.org/show_bug.cgi?id=64086
1005
1006         Reviewed by Adam Barth.
1007
1008         This is more fall-out from the Port class inappropriately encapsulating
1009         both platform and port knowledge.  We need to split out some of this
1010         platform knowledge into a separate class which can be better shared
1011         between ports.
1012
1013         The fix was to move all the _path_to_apache_config_file logic down
1014         into the WebKitPort baseclass so that all of the ports can find the
1015         right apache config file, regardless of what platform they may be running on.
1016
1017         I did not move this down into base.Port (even though I considered it).
1018         Chromium duplicates some of this logic, but since they have separate
1019         subclasses for each port-OS combination (e.g. ChromiumMac, ChromiumWin)
1020         they wouldn't notice the change.  Eventually we'll move this logic
1021         out of Port entirely, and then it will be shared by all ports.
1022
1023         I also cleaned up the http-config logic for ORWT while I was there,
1024         although since we're killing that code, I'm happy to revert that part
1025         of the change if changing it is viewed as needlessly risky.
1026
1027         During this cleanup, I noticed that no ports use the "default" httpd.conf
1028         which is left over from Mac OS X Tiger (and old linux distros) which used
1029         Apache 1.3.  I've removed httpd.conf (and associated support in ORWT) since
1030         we no longer support any ports which use this httpd.conf.
1031
1032         * Scripts/webkitperl/httpd.pm:
1033         (getHTTPDConfigPathForTestDirectory):
1034         * Scripts/webkitpy/layout_tests/port/base.py:
1035         * Scripts/webkitpy/layout_tests/port/gtk.py:
1036         * Scripts/webkitpy/layout_tests/port/mac.py:
1037         * Scripts/webkitpy/layout_tests/port/qt.py:
1038         * Scripts/webkitpy/layout_tests/port/webkit.py:
1039         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
1040         * Scripts/webkitpy/layout_tests/port/win.py:
1041         * Scripts/webkitpy/layout_tests/servers/apache_http_server.py:
1042
1043 2011-07-11  Hironori Bono  <hbono@chromium.org>
1044
1045         Reviewed by Adam Roben.
1046
1047         Implement layoutTestController.setTextDirection for WebKit2 and Windows.
1048         https://bugs.webkit.org/show_bug.cgi?id=61931
1049
1050         This change implements layoutTextController.setTextDirection for WebKit2
1051         and Windows so we can run a layout test added by r87770 on them. Each
1052         implementation adds a wrapper function for Editor::setBaseWritingDirection()
1053         so LayoutTestController can call it.
1054
1055         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1056         (LayoutTestController::setTextDirection): Implemented the binding function
1057         for layoutTestController.setTextDirection (Windows).
1058         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
1059         Added a binding function for layoutTestController.setTextDirection (WebKit2).
1060         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: ditto.
1061         (WTR::LayoutTestController::setTextDirection):
1062         * WebKitTestRunner/InjectedBundle/LayoutTestController.h: ditto.
1063
1064 2011-07-11  Raphael Kubo da Costa  <kubo@profusion.mobi>
1065
1066         [EFL] ImageDiff: Make sure gEcoreEvas is destroyed at the right time.
1067         https://bugs.webkit.org/show_bug.cgi?id=64293
1068
1069         Reviewed by Kent Tamura.
1070
1071         gEcoreEvas was being destroyed after shutdownEfl() was called, so the
1072         call to ecore_evas_free() failed.
1073
1074         We now shut it down manually before shutdownEfl() is called.
1075
1076         * DumpRenderTree/efl/ImageDiff.cpp:
1077         (main): Destroy gEcoreEvas before shutting down the EFL.
1078
1079 2011-07-11  Ryosuke Niwa  <rniwa@webkit.org>
1080
1081         Make Chromium Mac and Linux tests bot core builders
1082         https://bugs.webkit.org/show_bug.cgi?id=63196
1083
1084         Reviewed by Adam Barth.
1085
1086         Added Chromium Mac and Linux Release Tests bots core builders.
1087
1088         Also rearranged the order of bots so that test bots show up right next to
1089         respective builders.
1090
1091         * BuildSlaveSupport/build.webkit.org-config/config.json:
1092         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
1093         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1094
1095 2011-07-11  Raphael Kubo da Costa  <kubo@profusion.mobi>
1096
1097         [EFL] DRT: Add code related to font management.
1098         https://bugs.webkit.org/show_bug.cgi?id=63989
1099
1100         Reviewed by Kent Tamura.
1101
1102         The added files are responsible for managing fontconfig and adding the
1103         fonts required by some layout tests to fontconfig's path.
1104
1105         * DumpRenderTree/efl/FontManagement.cpp: Added.
1106         (getFontDirectories):
1107         (getFontFiles):
1108         (addFontDirectories):
1109         (addFontFiles):
1110         (addFontsToEnvironment):
1111         * DumpRenderTree/efl/FontManagement.h: Added.
1112
1113 2011-07-08  Dirk Pranke  <dpranke@chromium.org>
1114
1115         nrwt: linting fixes
1116         https://bugs.webkit.org/show_bug.cgi?id=64225
1117
1118         Reviewed by Eric Siedel.
1119
1120         Miscellaneous linting fixes. The most notable change is that
1121         we add public attributes for user, executive, filesystem, and
1122         options on the Port object, so we don't have to refer to the
1123         "protected" versions all over the place".
1124
1125         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1126         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
1127         * Scripts/webkitpy/layout_tests/controllers/message_broker.py:
1128         * Scripts/webkitpy/layout_tests/controllers/message_broker_unittest.py:
1129         * Scripts/webkitpy/layout_tests/controllers/worker.py:
1130         * Scripts/webkitpy/layout_tests/port/base.py:
1131         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1132
1133 2011-07-11  Ryosuke Niwa  <rniwa@webkit.org>
1134
1135         webkit-patch roll-chromium-deps no longer works
1136         https://bugs.webkit.org/show_bug.cgi?id=64324
1137
1138         Reviewed by James Robinson.
1139
1140         Lower the minimum commit message length to 10.
1141
1142         * Scripts/webkitpy/tool/steps/commit.py:
1143
1144 2011-07-11  Alice Boxhall  <aboxhall@chromium.org>
1145
1146         Convert json_results_generator.py to output version 4 JSON.
1147         https://bugs.webkit.org/show_bug.cgi?id=60869
1148
1149         Reviewed by Ojan Vafai.
1150
1151         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
1152         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1153         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
1154
1155 2011-07-11  Martin Robinson  <mrobinson@igalia.com>
1156
1157         [GTK] [NRWT] Pixel tests do not work
1158         https://bugs.webkit.org/show_bug.cgi?id=64091
1159
1160         Reviewed by Eric Seidel.
1161
1162         Fix pixel test runs for GTK+ new-run-webkit-tests by exposing the
1163         path to the ImageDiff binary.
1164
1165         * Scripts/webkitpy/layout_tests/port/gtk.py: Implement _path_to_image_diff.
1166
1167 2011-07-11  Dirk Pranke  <dpranke@chromium.org>
1168
1169         Fix hang on win32 in a multiprocessing test that shouldn't have been running.
1170
1171         Unreviewed, build fix.
1172
1173         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1174
1175 2011-07-11  Adam Roben  <aroben@apple.com>
1176
1177         Don't count leaks as test failures on TestFailures's front page
1178
1179         As a bonus, this patch adds our first unit test.
1180
1181         Fixes <http://webkit.org/b/64303> REGRESSION (r90489): TestFailures page says far more tests
1182         are failing on the Leaks bot than actually are
1183
1184         Reviewed by Adam Barth.
1185
1186         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
1187         (Builder.prototype.getNumberOfFailingTests): Exclude lines containing "leak" when summing
1188         failure counts.
1189
1190         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder_unittests.js:
1191         Added. Contains a simple test to show that leaks aren't counted as test failures.
1192
1193         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/run-unittests.html:
1194         Added. This is the test harness.
1195
1196 2011-06-14  Mike Stegeman  <mrstegeman@gmail.com>
1197
1198         [GTK] Add API to allow setting local storage database path
1199         https://bugs.webkit.org/show_bug.cgi?id=62091
1200
1201         Reviewed by Martin Robinson and Gustavo Noronha.
1202
1203         Expose the path of the localStorage databases through a setting
1204         to allow HTML5 localStorage to be persistent. New setting is:
1205         html5-local-storage-database-path
1206
1207         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1208         (resetDefaultsToConsistentValues):
1209
1210 2011-07-11  Adam Barth  <abarth@webkit.org>
1211
1212         garden-o-matic should be able to rebaseline tests
1213         https://bugs.webkit.org/show_bug.cgi?id=64186
1214
1215         Reviewed by Eric Seidel.
1216
1217         This patch contains a simple HTTP server binding for the
1218         rebaseline-test command.  The frontend will appear in another patch.
1219
1220         * Scripts/webkitpy/tool/servers/gardeningserver.py:
1221         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py:
1222         * Scripts/webkitpy/tool/servers/reflectionhandler.py:
1223
1224 2011-07-11  Adam Barth  <abarth@webkit.org>
1225
1226         Add some basic UI for showing regression ranges
1227         https://bugs.webkit.org/show_bug.cgi?id=64243
1228
1229         Reviewed by Eric Seidel.
1230
1231         This UI is all temporary.  It exists mostly just to have some buttons
1232         to click to exercise the code.
1233
1234         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
1235         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
1236         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
1237         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
1238         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
1239
1240 2011-07-11  Adam Barth  <abarth@webkit.org>
1241
1242         garden-o-matic should be able to determine which revisions caused a given failure
1243         https://bugs.webkit.org/show_bug.cgi?id=64189
1244
1245         Reviewed by Adam Roben.
1246
1247         Walking the failure history looking for failures turns out to be
1248         slightly tricky because the network requests are asynchronous.
1249         Currently we do all the fetches serially and our cache is unbounded.
1250         We'll probably optimize both those parameters eventually.
1251
1252         This patch also generalizes some functionality in the unit testing
1253         framework to make testing this sort of code easier.
1254
1255         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
1256         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
1257         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
1258
1259 2011-07-11  Adam Barth  <abarth@webkit.org>
1260
1261         Add a webkit-patch command for rebaselining an individual test
1262         https://bugs.webkit.org/show_bug.cgi?id=64246
1263
1264         Reviewed by Eric Seidel.
1265
1266         This patch introduces a command that's able to rebaseline a single
1267         test.  Currently, the command works only with the build.chromium.org
1268         buildbots, but extending it to work with the build.webkit.org bots
1269         shouldn't be that hard.
1270
1271         A complete rebaseling tool should also include an "optimize baselines"
1272         command (which moves/deletes baselines in order to reduce the number of
1273         expected results files), but that will come in a future patch.
1274
1275         Really BuilderToPort should be merged into builders.py, but I'm going
1276         to save that for a future patch as well.  (We need to stop shaving yaks
1277         at some point.)
1278
1279         * Scripts/webkitpy/tool/commands/rebaseline.py:
1280
1281 2011-07-11  Adam Barth  <abarth@webkit.org>
1282
1283         garden-o-matic should be able to roll out patches
1284         https://bugs.webkit.org/show_bug.cgi?id=64185
1285
1286         Reviewed by Eric Seidel.
1287
1288         This gardening server API is simply a binding to the rollout machinery
1289         already present in webkit-patch.
1290
1291         * Scripts/webkitpy/tool/commands/gardenomatic.py:
1292         * Scripts/webkitpy/tool/servers/gardeningserver.py:
1293         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py: Added.
1294
1295 2011-07-11  Csaba Osztrogonác  <ossy@webkit.org>
1296
1297         [Qt][Mac] Unreviewed fix after r90746.
1298
1299         * Scripts/run-webkit-tests:
1300         (useNewRunWebKitTests): Really disable NRWT for qt-mac platform.
1301
1302 2011-07-11  Adam Roben  <aroben@apple.com>
1303
1304         Teach TestFailures that ORWT's results.html file might be missing due to all tests passing
1305
1306         There are three reasons why we might fail to fetch ORWT's results.html:
1307           - All tests passed, so no results.html was generated
1308           - Some error during the test run caused results.html not to be generated (e.g., ORWT
1309             timed out)
1310           - Some network error occurred when fetching results.html
1311
1312         We were failing to account for the first possibility in some cases. For test runs before
1313         r89610, we first check build.webkit.org/json to determine how many tests failed and whether
1314         ORWT exited early due to too many failures; if all tests passed then we don't fetch
1315         results.html at all. r89610 changed ORWT to put information in results.html about exiting
1316         early due to too many failures, so we no longer needed to check build.webkit.org/json to get
1317         that information, and in r89619 I changed TestFailures to do just that. But I forgot that we
1318         still needed to check build.webkit.org/json to find out if all tests passed!
1319
1320         Now, for test runs after r89610, we check results.html first, and then check
1321         build.webkit.org/json if we fail to fetch results.html. This lets us distinguish between all
1322         tests passing and the error cases.
1323
1324         Fixes <http://webkit.org/b/64280> TestFailures page can't pinpoint that r90699 caused 13
1325         tests to fail on Windows 7 Release (WebKit2 Tests)
1326
1327         Reviewed by David Kilzer.
1328
1329         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
1330         (LayoutTestResultsLoader.prototype.start): Bumped the cache version so that old, buggy
1331         cached data will get evicted. We were marking builds where all tests passed as errors!
1332         (LayoutTestResultsLoader.prototype._fetchAndParseORWTResults): Added success/error callback
1333         parameters to the fetchAndParseResultsHTML helper function, and added a similar
1334         fetchNumberOfFailingTests function that fetches data from build.webkit.org/json (code came
1335         from later in the function). For test runs before r89610, we first check
1336         build.webkit.org/json then check results.html. For builds after r89610, we first check
1337         results.html and then check build.webkit.org/json if we couldn't fetch results.html.
1338
1339 2011-07-11  Csaba Osztrogonác  <ossy@webkit.org>
1340
1341         new-run-webkit-tests does not support qt-arm or qt-4.8 results
1342         https://bugs.webkit.org/show_bug.cgi?id=64071
1343
1344         [Qt] NRWT should pick up the right httpd config file
1345         https://bugs.webkit.org/show_bug.cgi?id=64086
1346
1347         Reviewed by Zoltán Herczeg.
1348
1349         * Scripts/run-webkit-tests: Disable NRWT for qt-mac, qt-arm and qt-4.8 platforms until fix.
1350         (useNewRunWebKitTests):
1351
1352 2011-07-11  Kenichi Ishibashi  <bashi@chromium.org>
1353
1354         Add TestNetscapePlugIn/Tests/FormValue.cpp to DRT build files
1355         https://bugs.webkit.org/show_bug.cgi?id=64248
1356
1357         Reviewed by Kent Tamura.
1358
1359         * DumpRenderTree/DumpRenderTree.gypi: Added FormValue.cpp.
1360         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Ditto.
1361         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro: Ditto.
1362         * GNUmakefile.am: Ditto.
1363
1364 2011-07-10  Kenichi Ishibashi  <bashi@chromium.org>
1365
1366         Let plugins participate in form submission
1367         https://bugs.webkit.org/show_bug.cgi?id=13061
1368
1369         Adds NPPVformValue support to the testing plugin.
1370
1371         Reviewed by Anders Carlsson.
1372
1373         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Added FormValue.cpp
1374         * DumpRenderTree/TestNetScapePlugIn/Tests/FormValue.cpp: Added.
1375         (FormValue::FormValue): Ditto.
1376         (FormValue::NPP_GetValue): Ditto.
1377
1378 2011-07-11  Roland Steiner  <rolandsteiner@chromium.org>
1379
1380         Reviewed by Tony Chang.
1381
1382         Complete functions in filesystem.py
1383         https://bugs.webkit.org/show_bug.cgi?id=63528
1384
1385         - missing text file functions added
1386         - functions sorted
1387         - removed 'append' optional parameters (were unused)
1388         - adapted filesystem_mock in the same way
1389
1390         * Scripts/webkitpy/common/system/filesystem.py:
1391         * Scripts/webkitpy/common/system/filesystem_mock.py:
1392         * Scripts/webkitpy/common/system/filesystem_unittest.py:
1393
1394 2011-07-10  Adam Barth  <abarth@webkit.org>
1395
1396         Clean up style in fallback path calculation
1397         https://bugs.webkit.org/show_bug.cgi?id=64239
1398
1399         Reviewed by Daniel Bates.
1400
1401         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1402         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1403         * Scripts/webkitpy/layout_tests/port/mac.py:
1404         * Scripts/webkitpy/layout_tests/port/win.py:
1405
1406 2011-07-10  Balazs Kelemen  <kbalazs@webkit.org>
1407
1408         [Qt][WK2] Implement the rest of WTR::PlatformWebView
1409         https://bugs.webkit.org/show_bug.cgi?id=63630
1410
1411         Reviewed by Kenneth Rohde Christiansen.
1412
1413         * WebKitTestRunner/qt/PlatformWebViewQt.cpp:
1414         (WTR::PlatformWebView::windowFrame):
1415         (WTR::PlatformWebView::setWindowFrame):
1416
1417 2011-07-10  Mark Rowe  <mrowe@apple.com>
1418
1419         Teach build-webkit how to find the latest WebKitSystemInterface binary.
1420
1421         * Scripts/build-webkit:
1422
1423 2011-07-09  Adam Roben  <aroben@apple.com>
1424
1425         Teach TestFailures to abbreviate the examples of test flakiness
1426
1427         These lists can get quite long, and it's not really helpful in most cases to have soooooo
1428         many examples of flakiness.
1429
1430         Fixes <http://webkit.org/b/64203> Lists of flaky revisions on TestFailures page can get so
1431         long they're hard to navigate
1432
1433         Reviewed by Dan Bates.
1434
1435         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/FlakyLayoutTestDetector.js:
1436         (FlakyLayoutTestDetector.prototype.flakinessExamples): If we have more than a certain number
1437         of examples, replace the middle items with a separator.
1438
1439         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
1440         (.flakiness-example-separator): Added styles for the separator.
1441
1442         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1443         (ViewController.prototype._domForPossiblyFlakyTests): Use a vertical ellipsis to represent
1444         the separator.
1445
1446 2011-07-09  Dirk Pranke  <dpranke@chromium.org>
1447
1448         nrwt: stack traces from worker-side exceptions aren't very useful inside test-webkitpy
1449         https://bugs.webkit.org/show_bug.cgi?id=64218
1450
1451         Reviewed by Eric Seidel.
1452
1453         Exceptions aren't picklable and can't be sent across the
1454         manager/worker message queue without losing information. NRWT
1455         handles this by turning the stack trace into a set of strings,
1456         and logging the strings when we receive an exception from the
1457         worker. However, when you are running tests and something
1458         crashes on the worker side, test-webkitpy prints the
1459         manager-side stack trace, which is just confusing and useless.
1460
1461         This patch changes the logic so that exceptions are passed
1462         through as-is when the worker and manager are in the same
1463         process (the --worker-model=inline option). This increases the
1464         code paths slightly but makes crashes much more useful.
1465
1466         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1467         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
1468         * Scripts/webkitpy/layout_tests/controllers/message_broker.py:
1469         * Scripts/webkitpy/layout_tests/controllers/worker.py:
1470         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1471
1472 2011-07-09  Darin Fisher  <darin@chromium.org>
1473
1474         Eliminate bad dependency on gfx::Point.
1475         https://bugs.webkit.org/show_bug.cgi?id=64228
1476
1477         Reviewed by Kent Tamura.
1478
1479         * DumpRenderTree/chromium/EventSender.cpp:
1480         (initMouseEvent):
1481
1482 2011-07-08  Ojan Vafai  <ojan@chromium.org>
1483
1484         pull static dashboard files into the appengine server from the chromium repository
1485         https://bugs.webkit.org/show_bug.cgi?id=64208
1486
1487         Reviewed by Dirk Pranke.
1488
1489         These files belong in the WebKit repo since they are tied primarily to the webkit tests.
1490         They have some extra bits to support chromium's gtests, but that seems fine.
1491         Mainly, this will allow other WebKit hackers to hack on the dashboards.
1492
1493         As a nice side-effect, we can now push the dashboard files when we do appengine pushes
1494         instead of the weird thing we used to do of pulling them from the Chromium repository
1495         and storing them in the appengine datastore. This allows for cleaning up a lot of code
1496         and will likely make the dashboards load a bit faster.
1497
1498         The new JS files don't fully match WebKit style, but I'd like to clean that up in a
1499         followup patch if possible to maintain my sanity with this patch.
1500
1501         * TestResultServer/app.yaml:
1502         * TestResultServer/handlers/dashboardhandler.py: Removed.
1503         * TestResultServer/handlers/menu.py:
1504         * TestResultServer/main.py:
1505         * TestResultServer/model/dashboardfile.py: Removed.
1506         * TestResultServer/static-dashboards/LICENSE.dygraph.txt: Added.
1507         * TestResultServer/static-dashboards/README.dygraph.txt: Added.
1508         * TestResultServer/static-dashboards/README.webtreemap.txt: Added.
1509         * TestResultServer/static-dashboards/aggregate_results.html: Added.
1510         * TestResultServer/static-dashboards/builders.js: Added.
1511         * TestResultServer/static-dashboards/dashboard_base.js: Added.
1512         * TestResultServer/static-dashboards/dygraph-combined.js: Added.
1513         * TestResultServer/static-dashboards/flakiness_dashboard.html: Added.
1514         * TestResultServer/static-dashboards/flakiness_dashboard_tests.js: Added.
1515         * TestResultServer/static-dashboards/timeline_explorer.html: Added.
1516         * TestResultServer/static-dashboards/treemap.html: Added.
1517         * TestResultServer/static-dashboards/webtreemap.css: Added.
1518         * TestResultServer/static-dashboards/webtreemap.js: Added.
1519         * TestResultServer/stylesheets/dashboardfile.css: Removed.
1520         * TestResultServer/stylesheets/menu.css:
1521         * TestResultServer/templates/dashboardfilelist.html: Removed.
1522         * TestResultServer/templates/menu.html:
1523
1524 2011-07-08  Adam Roben  <aroben@apple.com>
1525
1526         Make TestFailures's list of flaky tests look more like the list of non-flaky tests
1527
1528         Fixes <http://webkit.org/b/64204> TestFailures page's flaky tests list is ugly!
1529
1530         Reviewed by Daniel Bates.
1531
1532         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
1533         (#failure-history, #possibly-flaky-tests): Expanded this rule to apply to the list of flaky
1534         tests.
1535         (#failure-history > li, #possibly-flaky-tests > li): Ditto, but moved the 50px left padding
1536         from here...
1537         (#failure-history > li): ...to here.
1538
1539         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1540         (ViewController.prototype._domForPossiblyFlakyTests): Give the list an id attribute for
1541         styling purposes.
1542
1543 2011-07-08  Jeffrey Pfau  <jpfau@apple.com>
1544
1545         Unreviewed, add myself as committer.
1546
1547         * Scripts/webkitpy/common/config/committers.py:
1548
1549 2011-07-08  Adam Barth  <abarth@webkit.org>
1550
1551         TestResultsServer should keep old test results
1552         https://bugs.webkit.org/show_bug.cgi?id=64199
1553
1554         Reviewed by Ojan Vafai.
1555
1556         Having historical data will help us do failure archeology.
1557
1558         * TestResultServer/handlers/testfilehandler.py:
1559         * TestResultServer/model/testfile.py:
1560         * TestResultServer/templates/showfilelist.jsonp: Added.
1561
1562 2011-07-08  Adam Roben  <aroben@apple.com>
1563
1564         Remove commit-log-editor's dependency on Module::Load::Conditional
1565
1566         This module isn't available in Perl 5.8.8 (the version used on Leopard).
1567
1568         Fixes <http://webkit.org/b/64198> REGRESSION (r90583):
1569         webkitpy.common.checkout.checkout_unittest failing on Leopard
1570
1571         Reviewed by Daniel Bates.
1572
1573         * Scripts/commit-log-editor: Use the new loadTermReadKey() function instead of
1574         Module::Load::Conditional::can_load.
1575         (loadTermReadKey): Added. Tries to load Term::ReadKey and returns true if it's successful.
1576
1577 2011-07-08  Adam Roben  <aroben@apple.com>
1578
1579         Make checkout_unittest more robust against files moving around
1580
1581         Fixes <http://webkit.org/b/64197> checkout_unittest contains ugly, fragile code to find the
1582         Scripts directory
1583
1584         Reviewed by Adam Barth.
1585
1586         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
1587         (CommitMessageForThisCommitTest.test_commit_message_for_this_commit): Instantiate a real SCM
1588         object and use it to get the path to the Scripts directory, rather than hard-coding the
1589         relative path from this file to Scripts.
1590
1591 2011-07-08  Adam Roben  <aroben@apple.com>
1592
1593         Make TestFailures remember that run-webkit-tests timed out, even across reloads
1594
1595         Fixes <http://webkit.org/b/64193> TestFailures page incorrectly thinks all tests passed in
1596         http://build.webkit.org/builders/Windows%207%20Release%20(Tests)/builds/14589 after a reload
1597
1598         Reviewed by Daniel Bates.
1599
1600         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
1601         (LayoutTestResultsLoader.prototype.start): Store an "error" attribute in the cached data.
1602         When true, it indicates that there was an error retrieving the results for this build and
1603         that the errorCallback should be called.
1604
1605 2011-07-08  Adam Barth  <abarth@webkit.org>
1606
1607         Teach garden-o-matic how to display test results
1608         https://bugs.webkit.org/show_bug.cgi?id=64141
1609
1610         Reviewed by Ojan Vafai.
1611
1612         This patch includes basic infrastructure for probing build.chromium.org
1613         for test results.  We only handle text and image tests, not anything
1614         complicated like reftests.  Also, we're using the revision/build
1615         independent results store on the server, so we're avoiding that
1616         complication for now.
1617
1618         It's slightly hacky that we need to probe the server to see what kinds
1619         of results exist.  A better solution would be to add CORS support to
1620         the server or to use the local server to help.
1621
1622         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
1623         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
1624         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
1625         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
1626         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
1627         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
1628
1629 2011-07-08  Dirk Pranke  <dpranke@chromium.org>
1630
1631         REGRESSION(90419) NRWT's httpd locking is broken for --child-processes=1
1632         https://bugs.webkit.org/show_bug.cgi?id=64092
1633
1634         Reviewed by Tony Chang.
1635
1636         The code was incorrectly creating empty shards in the
1637         shard_in_two case.
1638
1639         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1640         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
1641
1642 2011-07-08  Adam Roben  <aroben@apple.com>
1643
1644         Teach buildbot to figure out how many webkitpy/webkitperl tests failed
1645
1646         Fixes <http://webkit.org/b/64192> It's hard to tell how many test-webkitpy/test-webkitperl
1647         tests failed when looking at build.webkit.org
1648
1649         Reviewed by Eric Seidel.
1650
1651         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1652         (TestWithFailureCount): New class that represents a test build step which has an associated
1653         failure count. Eventually we should move more of our test classes to deriving from this.
1654         (TestWithFailureCount.countFailures): Method for subclasses to override to say how many
1655         failures occurred.
1656
1657         (TestWithFailureCount.commandComplete):
1658         (TestWithFailureCount.evaluateCommand):
1659         (TestWithFailureCount.getText):
1660         (TestWithFailureCount.getText2):
1661         These were all based on RunGtkAPITests.
1662
1663         (RunPythonTests): Changed to inherit from TestWithFailureCount.
1664         (RunPythonTests.countFailures): Parses the test-webkitpy output looking for the count of
1665         failures.
1666         (RunPerlTests): Changed to inherit from TestWithFailureCount.
1667         (RunPerlTests.countFailures): Parses the test-webkitperl output looking for the count of
1668         failures.
1669
1670 2011-07-08  Adam Roben  <aroben@apple.com>
1671
1672         Ensure $CHANGE_LOG_EMAIL_ADDRESS is set when testing webkitpy's commit-log-editor integration
1673
1674         Fixes <http://webkit.org/b/64180> REGRESSION (r90564): test-webkitpy failing on multiple
1675         bots due to commit-log-editor errors
1676
1677         Reviewed by Adam Barth.
1678
1679         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
1680         (CommitMessageForThisCommitTest.test_commit_message_for_this_commit): Set
1681         $CHANGE_LOG_EMAIL_ADDRESS to the patch author's email address. This ensures that
1682         commit-log-editor can find a value for the committer's email, and that the committer and
1683         author email addresses match, which will prevent commit-log-editor from inserting a "Patch
1684         by" line in the commit message.
1685
1686 2011-07-08  Andreas Kling  <kling@webkit.org>
1687
1688         [Qt][WK2] Views should expose QActions for basic navigation.
1689         https://bugs.webkit.org/show_bug.cgi?id=64174
1690
1691         Reviewed by Benjamin Poulain.
1692
1693         Bring the toolbar in Qt's MiniBrowser back to life.
1694
1695         * MiniBrowser/qt/BrowserView.cpp:
1696         (BrowserView::navigationAction):
1697         * MiniBrowser/qt/BrowserView.h:
1698         * MiniBrowser/qt/BrowserWindow.cpp:
1699         (BrowserWindow::BrowserWindow):
1700
1701 2011-07-08  Adam Roben  <aroben@apple.com>
1702
1703         Teach TestFailures how to find test names in commit-log-editor-style commit messages
1704
1705         TestFailures was relying on Trac turning the list of modified files in our commit messages
1706         into an HTML list. But Trac only does this when the list of modified files is indented.
1707         commit-log-editor doesn't indent the file list, so the list wasn't being turned into an HTML
1708         list, which was confusing TestFailures.
1709
1710         TestFailures now does much simpler parsing of the commit message (i.e., just a substring
1711         search) without relying at all on its structure.
1712
1713         Fixes <http://webkit.org/b/64173> TestFailures page fails to blame r90608 for breaking
1714         fast/dom/HTMLProgressElement/progress-element-markup.html on Windows
1715
1716         Reviewed by David Kilzer.
1717
1718         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Trac.js:
1719         (Trac.prototype.getCommitDataForRevisionRange): Instead of trying to parse the commit
1720         message, just return its text.
1721
1722         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1723         (ViewController.prototype._domForRegressionRange): Instead of searching for test names in
1724         each commit's list of modified files, just search for test names anywhere in the commit's
1725         message.
1726
1727 2011-07-08  Adam Barth  <abarth@webkit.org>
1728
1729         sheriffbot is too spammy in IRC
1730         https://bugs.webkit.org/show_bug.cgi?id=64153
1731
1732         Reviewed by Eric Seidel.
1733
1734         Reporting failures in IRC worked well when the tree was greener than it
1735         is today.  Nowadays, this feature mostly just results in spam about
1736         false positives.  If we reach a greener state, we can bring this code
1737         back.
1738
1739         (Another possibility is to restrict this feature to builder bots,
1740         rather than including testers, as we did before this patch.)
1741
1742         * Scripts/webkitpy/tool/commands/sheriffbot.py:
1743         * Scripts/webkitpy/tool/commands/sheriffbot_unittest.py:
1744
1745 2011-07-08  Adam Roben  <aroben@apple.com>
1746
1747         Ensure commit-log-editor adds a "Patch by" line when the author and committer are different
1748
1749         Previously we were only adding a "Patch by" line when the ChangeLog contained a "Reviewed
1750         by" line. But some patches (like rollout patches) don't contain that line. Now we always add
1751         "Patch by" to the commit log regardless of the ChangeLog's contents.
1752
1753         Fixes <http://webkit.org/b/64127> Committer for r90588 is commit-queue@webkit.org, but
1754         should have been abarth@webkit.org
1755
1756         Reviewed by Anders Carlsson.
1757
1758         * Scripts/commit-log-editor:
1759         (createCommitMessage): Try to put the "Patch by" line just above the "Reviewed by" line, as
1760         before. If there is no "Reviewed by" line, try to put it just above the first modified file.
1761         If all else fails, put it at the end of the commit message.
1762         (patchAuthorshipString): Added. Code came from createCommitMessage.
1763
1764 2011-07-08  Herczeg Zoltan  <zherczeg@webkit.org>
1765
1766         Reviewed by Andreas Kling.
1767
1768         Adding myself as a reviewer.
1769
1770         * Scripts/webkitpy/common/config/committers.py:
1771
1772 2011-07-07  Kevin Ollivier  <kevino@theolliviers.com>
1773
1774         [wx] Unreviewed build fix, remove old files from the tree and update paths to new ones.
1775
1776         * wx/build: Removed.
1777         * wx/build/build_utils.py: Removed.
1778         * wx/build/settings.py: Removed.
1779         * wx/build/waf_extensions.py: Removed.
1780         * wx/build/wxpresets.py: Removed.
1781         * wx/packaging/build-mac-installer.py:
1782
1783 2011-07-07  Kevin Ollivier  <kevino@theolliviers.com>
1784
1785         [wx] Unreviewed build fix, more fixes for install name issues.
1786
1787         * wx/packaging/build-mac-installer.py:
1788
1789 2011-07-07  Dirk Pranke  <dpranke@chromium.org>
1790
1791         test-webkitpy fails on chromium win
1792         https://bugs.webkit.org/show_bug.cgi?id=64137
1793
1794         Reviewed by Eric Seidel.
1795
1796         test-webkitpy doesn't play nicely with the multiprocessing
1797         module on win32. An earlier change actually reenabled the tests
1798         on win32 by mistake.
1799
1800         This patch also fixes a few cases where path names will trip
1801         things up on win32 (e.g., by testing unix paths on windows).
1802         We do not lose any real coverage here with those fixes.
1803
1804         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
1805         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
1806         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
1807         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1808
1809 2011-07-07  Leandro Pereira  <leandro@profusion.mobi>
1810
1811         [ImageDiff] Calculate/print difference right after reading baseline image.
1812         https://bugs.webkit.org/show_bug.cgi?id=64117
1813
1814         Reviewed by Kent Tamura.
1815
1816         Otherwise, ImageDiff will block on fgets() until the universe ends or you ^C
1817         it.  Whichever happens first.
1818
1819         * DumpRenderTree/efl/ImageDiff.cpp:
1820         (main):
1821
1822 2011-07-07  Raphael Kubo da Costa  <kubo@profusion.mobi>
1823
1824         [EFL] DRT: Leak cairo_t to make the pixel tests stop crashing
1825         https://bugs.webkit.org/show_bug.cgi?id=64107
1826
1827         Reviewed by Kent Tamura.
1828
1829         The pointer is later managed by BitmapContext, but it was being
1830         de-refed and deleted earlier by our RefPtr, causing crashes in all
1831         pixel tests.
1832
1833         We now call leakRef() to make sure it is not removed when our
1834         createBitmapContextFromWebView goes out of context.
1835
1836         * DumpRenderTree/efl/PixelDumpSupportEfl.cpp:
1837         (createBitmapContextFromWebView):
1838
1839 2011-07-07  Adam Roben  <aroben@apple.com>
1840
1841         Update TestFailures's title and header immediately upon navigation
1842
1843         Fixes <http://webkit.org/b/64125> TestFailures page seems slow to react on link clicks
1844
1845         Reviewed by Daniel Bates.
1846
1847         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1848         (ViewController.prototype.loaded): Create and store an h1 element for displaying the page's
1849         title.
1850         (ViewController.prototype._displayBuilder): Set the page's title and clear out any old
1851         content immediately instead of waiting for the first set of results for the history
1852         analyzer. As we receive new results from the analyzer we'll just clear out the main content
1853         area and replace it with the new info.
1854         (ViewController.prototype._displayTesters): Set the page's title and clear out any old
1855         content immediately instead of waiting for the list of testers from the buildbot. Sprinkled
1856         in a little use of our removeAllChildren helper, too.
1857         (ViewController.prototype._setTitle): Added. Set's the page's title and the header text.
1858
1859 2011-07-07  Adam Roben  <aroben@apple.com>
1860
1861         Make Checkout use SCM's Executive instead of conjuring up its own
1862
1863         This will improve integration with the rest of webkitpy, particularly when invoked via
1864         webkit-patch.
1865
1866         Fixes <http://webkit.org/b/64115> REGRESSION (r90564): webkitpy's Checkout class uses
1867         Executive inappropriately
1868
1869         Reviewed by Adam Barth.
1870
1871         * Scripts/webkitpy/common/checkout/checkout.py:
1872         (Checkout.commit_message_for_this_commit): Use SCM.run instead of creating an Executive for
1873         our own use. SCM might have some super-special Executive that it uses under the covers, and
1874         we want to use it, too!
1875
1876         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
1877         (CommitMessageForThisCommitTest.test_commit_message_for_this_commit): Mock the SCM.run
1878         method to call through to Executive.run_command.
1879
1880 2011-07-07  Adam Roben  <aroben@apple.com>
1881
1882         Make Term/ReadKey.pm an optional dependency of commit-log-editor
1883
1884         When Term/ReadKey.pm isn't present, the --regenerate-log switch will be non-functional.
1885
1886         Fixes <http://webkit.org/b/64113> REGRESSION (r90564):
1887         webkitpy.common.checkout.checkout_unittest failing on GTK bots due to missing
1888         Term/ReadKey.pm module
1889
1890         Reviewed by Xan Lopez.
1891
1892         * Scripts/commit-log-editor: Use Module::Load::Conditional::can_load to only load
1893         Term/ReadKey.pm if it's present. If it isn't present, ignore the --regenerate-log switch so
1894         that existing commit messages will be preserved. (Users can manually remove existing commit
1895         messages using git-reset.)
1896
1897 2011-07-07  Adam Roben  <aroben@apple.com>
1898
1899         Keep testing that commit messages containing Unicode are handled properly by webkitpy
1900
1901         Fixes <http://webkit.org/b/64109> REGRESSION (r90571): test-webkitpy no longer tests that we
1902         correctly handle Unicode in commit messages
1903
1904         Reviewed by Eric Seidel.
1905
1906         * Scripts/webkitpy/common/checkout/checkout_unittest.py: Put some Unicode characters in
1907         places in the ChangeLog that will end up being included in the commit message.
1908
1909 2011-07-07  Andreas Kling  <kling@webkit.org>
1910
1911         [Qt][WK2] Remove Symbian code.
1912         https://bugs.webkit.org/show_bug.cgi?id=64101
1913
1914         Reviewed by Benjamin Poulain.
1915
1916         * MiniBrowser/qt/BrowserWindow.cpp:
1917         (BrowserWindow::BrowserWindow):
1918         (BrowserWindow::screenshot):
1919         * MiniBrowser/qt/MiniBrowser.pro:
1920         * MiniBrowser/qt/MiniBrowserApplication.h:
1921         (WindowOptions::WindowOptions):
1922
1923 2011-07-07  Andreas Kling  <kling@webkit.org>
1924
1925         [Qt][WK2] Don't support ridiculous matrix of QT_NO_FEATURE combinations.
1926         https://bugs.webkit.org/show_bug.cgi?id=64099
1927
1928         Reviewed by Benjamin Poulain.
1929
1930         * MiniBrowser/qt/BrowserWindow.cpp:
1931         (BrowserWindow::openFile):
1932         (BrowserWindow::screenshot):
1933         (BrowserWindow::loadURLListFromFile):
1934
1935 2011-07-07  Adam Roben  <aroben@apple.com>
1936
1937         Completely ignore too-many-failures builds in TestFailures in most circumstances
1938
1939         Because a semi-arbitrary subset of tests are run in a too-many-failures build, we can't
1940         really use them to perform regression analysis. The only time we want to pay attention to
1941         too-many-failures builds is when we're trying to explain when the current bout of
1942         too-many-failures started.
1943
1944         Fixes <http://webkit.org/b/64106> TestFailures page sometimes claims a test started failing
1945         in a build that didn't even run it (because it exited early due to too many failues)
1946
1947         Reviewed by David Kilzer.
1948
1949         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
1950         (LayoutTestHistoryAnalyzer.prototype._incorporateBuildHistory): Removed old, broken
1951         too-many-failures handling that would cause us to blame builds that didn't even run a given
1952         test for breaking it. Instead, skip over all too-many-failures builds unless the most recent
1953         build was itself a too-many-failures build.
1954
1955 2011-07-07  Adam Roben  <aroben@apple.com>
1956
1957         Teach webkitpy's Checkout class to use commit-log-editor to create commit messages
1958
1959         Fixes <http://webkit.org/b/26755> webkit-patch's commit messages are less readable than
1960         commit-log-editor's
1961
1962         Reviewed by David Kilzer.
1963
1964         * Scripts/webkitpy/common/checkout/checkout.py:
1965         (Checkout.commit_message_for_this_commit): Run commit-log-editor, passing it the paths of
1966         the modified ChangeLogs, to generate the commit message, rather than trying to generate one
1967         ourselves.
1968
1969         * Scripts/webkitpy/common/checkout/checkout_unittest.py: Updated the expected commit message
1970         to match commit-log-editor's format.
1971         (CommitMessageForThisCommitTest.setUp): Write the ChangeLogs into Tools and LayoutTests
1972         directories so we can see how the various entries get labeled in the commit message.
1973         (CommitMessageForThisCommitTest.test_commit_message_for_this_commit): Create a mock SCM
1974         instance that knows how to find commit-log-editor and pass it to our Checkout instance.
1975         Don't bother capturing output, since there shouldn't be any.
1976
1977 2011-07-07  Adam Roben  <aroben@apple.com>
1978
1979         Add a --print-log option to commit-log-editor
1980
1981         When specified, commit-log-editor takes a set of ChangeLog paths on the command line,
1982         generates a commit message from those ChangeLogs, prints it to stdout, and exits.
1983
1984         Prep work for fixing <http://webkit.org/b/26755> webkit-patch's commit messages are less
1985         readable than commit-log-editor's
1986
1987         Reviewed by David Kilzer.
1988
1989         * Scripts/commit-log-editor: Changed to use Getopt::Long to parse options. Added --print-log
1990         option, which calls through to createCommitMessage, prints the result, and exits.
1991         (printUsageAndExit): Renamed from "usage" for clarity. Beefed up the usage statement to
1992         explain commit-log-editor's 2.5 (normal, --print-log, and --help) modes of operation.
1993
1994 2011-07-07  Adam Roben  <aroben@apple.com>
1995
1996         Extract commit-log-editor's commit-message-generation code into a separate function
1997
1998         As a bonus, we now skip a bunch of work in the case where we are generating a commit message
1999         for staged git changes.
2000
2001         Prep work for fixing <http://webkit.org/b/26755> webkit-patch's commit messages are less
2002         readable than commit-log-editor's
2003
2004         Reviewed by David Kilzer.
2005
2006         * Scripts/commit-log-editor: Moved the default definition of $endl up toward the top of the
2007         file so that it will be defined even if we aren't operating on an existing log message.
2008         Moved a bunch of top-level code to generate the commit message from here...
2009         (createCommitMessage): ...to here.
2010
2011 2011-07-07  Carlos Garcia Campos  <cgarcia@igalia.com>
2012
2013         [GTK] Open links in a new window when clicking with the middle button in MiniBrowser
2014         https://bugs.webkit.org/show_bug.cgi?id=63988
2015
2016         Reviewed by Martin Robinson.
2017
2018         * MiniBrowser/gtk/BrowserWindow.c:
2019         (browserWindowConstructed):
2020         (decidePolicyForNavigationAction): Ignore the action if a link is
2021         clicked with the middle buttonm and open the link in a new window.
2022         (browserWindowPolicyClientInit):
2023
2024 2011-07-07  Adam Barth  <abarth@webkit.org>
2025
2026         Remove style_references.py
2027         https://bugs.webkit.org/show_bug.cgi?id=64070
2028
2029         Reviewed by Eric Seidel.
2030
2031         We never really adopted the _references design, and this file is a bit
2032         of an orphan at this point.
2033
2034         * Scripts/check-webkit-style:
2035         * Scripts/webkitpy/style/checker.py:
2036         * Scripts/webkitpy/style/checker_unittest.py:
2037         * Scripts/webkitpy/style/checkers/python.py:
2038         * Scripts/webkitpy/style/checkers/test_expectations.py:
2039         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2040         * Scripts/webkitpy/style/main_unittest.py:
2041         * Scripts/webkitpy/style_references.py: Removed.
2042
2043 2011-07-07  Eric Seidel  <eric@webkit.org>
2044
2045         Fix WebKit2 expected results search paths for Mac and Qt under new-run-webkit-tests
2046         https://bugs.webkit.org/show_bug.cgi?id=64056
2047
2048         Unreviewed.
2049
2050         I had written this unittest prior to landing my change, but
2051         had named it qt_unitest.py, so it wasn't running (nor was
2052         it added to my git repo).
2053
2054         Once I fixe the name of the unittest file, it was very easy
2055         to fix the error in webkit.py (which already had a FIXME).
2056
2057         * Scripts/webkitpy/layout_tests/port/qt_unittest.py: Added.
2058         * Scripts/webkitpy/layout_tests/port/webkit.py:
2059
2060 2011-07-07  Eric Seidel  <eric@webkit.org>
2061
2062         Unreviewed.  Add more debug logging about which test expectations we're using.
2063
2064         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2065         * Scripts/webkitpy/layout_tests/port/webkit.py:
2066
2067 2011-07-06  Adam Barth  <abarth@webkit.org>
2068
2069         Remove unexpected_results.json
2070         https://bugs.webkit.org/show_bug.cgi?id=64065
2071
2072         Reviewed by Eric Seidel.
2073
2074         No one should be using this file any more.  This patch removes it from
2075         the face of the earth.  This patch also removes two NRWT-specific
2076         command line options that were implemented using
2077         unexpected_results.json.
2078
2079         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
2080         * Scripts/webkitpy/layout_tests/controllers/manager.py:
2081         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2082         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2083         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2084         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
2085         * Scripts/webkitpy/tool/servers/rebaselineserver_unittest.py:
2086
2087 2011-07-07  Adam Barth  <abarth@webkit.org>
2088
2089         REGRESSION(90520): chromium GPU tests produce output in the LayoutTests directory
2090         https://bugs.webkit.org/show_bug.cgi?id=64067
2091
2092         Reviewed by Eric Seidel.
2093
2094         The author of r90520 forgot to change the chromium_gpu.py implemenation
2095         of this method as well.
2096
2097         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2098
2099 2011-07-06  Eric Seidel  <eric@webkit.org>
2100
2101         Fix WebKit2 expected results search paths for Mac and Qt under new-run-webkit-tests
2102         https://bugs.webkit.org/show_bug.cgi?id=64056
2103
2104         Reviewed by Adam Barth.
2105
2106         This fixes the MacPort's version of baseline_search_path to include
2107         wk2 fallback like WebKitPort's does.  This also re-works the QtPort
2108         to set self._name (as other ports do) so that WebKitPort's default
2109         implementations will understand Qt operating system flavors correctly
2110         (including when running wk2 tests).
2111
2112         In trying to test this, I realized that nearly all of our port-tests
2113         use the default FileSystem(), User() and Executive() objects, which
2114         is really bad, because they expose details of your system in the unittest results!
2115
2116         Once I fixed many of the Port() calls to pass MockFileSystem(), then
2117         MacPort.test_tests_for_other_platforms started failing, and I had to
2118         teach MockFileSystem.glob how to handle directories to make it work again.
2119
2120         I removed the useless PortTestCase.test_baseline_search_path and replaced
2121         it with more useful test_baseline_search_path tests in MacPortTest and QtPortTest.
2122
2123         * Scripts/webkitpy/common/config/ports.py:
2124         * Scripts/webkitpy/common/system/filesystem_mock.py:
2125         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2126         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2127         * Scripts/webkitpy/layout_tests/port/mac.py:
2128         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
2129         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
2130         * Scripts/webkitpy/layout_tests/port/qt.py:
2131         * Scripts/webkitpy/layout_tests/port/test_files.py:
2132         * Scripts/webkitpy/layout_tests/port/win.py:
2133         * Scripts/webkitpy/to_be_moved/deduplicate_tests.py:
2134         * Scripts/webkitpy/tool/mocktool.py:
2135
2136 2011-07-06  Adam Barth  <abarth@webkit.org>
2137
2138         Wean rebaseline-server off unexpected_results.json
2139         https://bugs.webkit.org/show_bug.cgi?id=64063
2140
2141         Reviewed by Eric Seidel.
2142
2143         In the process of changing this code to use full_results.json, I
2144         noticed that the code was broken (because it wasn't tested).  This
2145         patch also adds test coverage for the broken code.
2146
2147         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
2148         * Scripts/webkitpy/tool/servers/rebaselineserver.py:
2149         * Scripts/webkitpy/tool/servers/rebaselineserver_unittest.py:
2150
2151 2011-07-06  Adam Barth  <abarth@webkit.org>
2152
2153         Wean resultsjsonparser off unexpected_results.json
2154         https://bugs.webkit.org/show_bug.cgi?id=64061
2155
2156         Reviewed by Eric Seidel.
2157
2158         We're about to delete unexpected_results.json, so we need to remove all
2159         the users.
2160
2161         * Scripts/webkitpy/common/config/ports.py:
2162         * Scripts/webkitpy/common/net/resultsjsonparser.py:
2163         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
2164         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2165
2166 2011-07-06  Adam Barth  <abarth@webkit.org>
2167
2168         Cleanup result_summary.py
2169         https://bugs.webkit.org/show_bug.cgi?id=64057
2170
2171         Reviewed by Eric Seidel.
2172
2173         This class had a bunch of out-of-date docstrings that no longer make
2174         any sense.
2175
2176         * Scripts/webkitpy/layout_tests/models/result_summary.py:
2177
2178 2011-07-06  Adam Barth  <abarth@webkit.org>
2179
2180         Tweak the ChangeLog for DEPS rolls to say "Unreviewed" so that the
2181         patches can be landed by the commit-queue.
2182
2183         * Scripts/webkitpy/tool/steps/preparechangelogfordepsroll.py:
2184
2185 2011-07-06  Adam Barth  <abarth@webkit.org>
2186
2187         Normalize callers of logging.getLogger to use __name__
2188         https://bugs.webkit.org/show_bug.cgi?id=64054
2189
2190         Reviewed by Eric Seidel.
2191
2192         Most callers were already using __name__.  This patch makes everyone
2193         consistent.
2194
2195         * Scripts/webkitpy/common/checkout/diff_parser.py:
2196         * Scripts/webkitpy/common/checkout/scm/svn.py:
2197         * Scripts/webkitpy/common/net/statusserver.py:
2198         * Scripts/webkitpy/common/system/executive.py:
2199         * Scripts/webkitpy/common/system/file_lock.py:
2200         * Scripts/webkitpy/common/system/user.py:
2201         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
2202         * Scripts/webkitpy/layout_tests/models/result_summary.py:
2203         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2204         * Scripts/webkitpy/layout_tests/port/chromium.py:
2205         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2206         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
2207         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2208         * Scripts/webkitpy/layout_tests/port/gtk.py:
2209         * Scripts/webkitpy/layout_tests/port/http_lock.py:
2210         * Scripts/webkitpy/layout_tests/port/mac.py:
2211         * Scripts/webkitpy/layout_tests/port/qt.py:
2212         * Scripts/webkitpy/layout_tests/port/server_process.py:
2213         * Scripts/webkitpy/layout_tests/port/webkit.py:
2214         * Scripts/webkitpy/layout_tests/port/win.py:
2215         * Scripts/webkitpy/layout_tests/servers/apache_http_server.py:
2216         * Scripts/webkitpy/layout_tests/servers/http_server.py:
2217         * Scripts/webkitpy/layout_tests/servers/http_server_base.py:
2218         * Scripts/webkitpy/layout_tests/servers/websocket_server.py:
2219         * Scripts/webkitpy/layout_tests/views/printing.py:
2220         * Scripts/webkitpy/python24/versioning.py:
2221         * Scripts/webkitpy/style/checker.py:
2222         * Scripts/webkitpy/style/checkers/test_expectations.py:
2223         * Scripts/webkitpy/style/patchreader.py:
2224         * Scripts/webkitpy/to_be_moved/update_webgl_conformance_tests.py:
2225
2226 2011-07-06  Adam Barth  <abarth@webkit.org>
2227
2228         Move model classes out of layout_package and into models
2229         https://bugs.webkit.org/show_bug.cgi?id=64053
2230
2231         Reviewed by Eric Seidel.
2232
2233         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2234         * Scripts/webkitpy/common/net/layouttestresults.py:
2235         * Scripts/webkitpy/common/net/layouttestresults_unittest.py:
2236         * Scripts/webkitpy/common/net/resultsjsonparser.py:
2237         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
2238         * Scripts/webkitpy/layout_tests/controllers/manager.py:
2239         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2240         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py:
2241         * Scripts/webkitpy/layout_tests/controllers/worker.py:
2242         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
2243         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
2244         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
2245         * Scripts/webkitpy/layout_tests/layout_package/result_summary.py: Removed.
2246         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py: Removed.
2247         * Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py: Removed.
2248         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py: Removed.
2249         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py: Removed.
2250         * Scripts/webkitpy/layout_tests/layout_package/test_input.py: Removed.
2251         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py:
2252         * Scripts/webkitpy/layout_tests/layout_package/test_results.py: Removed.
2253         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py: Removed.
2254         * Scripts/webkitpy/layout_tests/models: Added.
2255         * Scripts/webkitpy/layout_tests/models/__init__.py: Added.
2256         * Scripts/webkitpy/layout_tests/models/result_summary.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/result_summary.py.
2257         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_expectations.py.
2258         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py.
2259         * Scripts/webkitpy/layout_tests/models/test_failures.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_failures.py.
2260         * Scripts/webkitpy/layout_tests/models/test_failures_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py.
2261         * Scripts/webkitpy/layout_tests/models/test_input.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_input.py.
2262         * Scripts/webkitpy/layout_tests/models/test_results.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_results.py.
2263         * Scripts/webkitpy/layout_tests/models/test_results_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py.
2264         * Scripts/webkitpy/layout_tests/port/chromium.py:
2265         * Scripts/webkitpy/layout_tests/views/printing.py:
2266         * Scripts/webkitpy/layout_tests/views/printing_unittest.py:
2267         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2268         * Scripts/webkitpy/style_references.py:
2269         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests.py:
2270         * Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
2271         * Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py:
2272         * Scripts/webkitpy/tool/commands/queues_unittest.py:
2273         * Scripts/webkitpy/tool/commands/rebaseline.py:
2274
2275 2011-07-06  Dirk Pranke  <dpranke@chromium.org>
2276
2277         nrwt: fix typo preventing http startup on windows
2278         https://bugs.webkit.org/show_bug.cgi?id=64050
2279
2280         Reviewed by Eric Siedel.
2281
2282         NRWT checks whether it needs to start the servers by looking
2283         for "/http/" in the test names. After r90520, the leading slash
2284         is not present, so the check isn't quite right.
2285
2286         * Scripts/webkitpy/layout_tests/controllers/manager.py:
2287         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2288
2289 2011-07-06  Dirk Pranke  <dpranke@chromium.org>
2290
2291         nrwt: remove --use-apache from the command line
2292         https://bugs.webkit.org/show_bug.cgi?id=63358
2293
2294         Reviewed by Adam Barth.
2295
2296         Re-land the fix in bug 63358 - the prior fix was flawed in that
2297         we were trying to start the http server during check_sys_deps()
2298         but not actually setting up the environment properly. This
2299         broken prior to this change, but exposed by it, since the patch
2300         changes the chromium port to call the base class's checks.
2301
2302         * Scripts/webkitpy/common/system/executive.py:
2303         * Scripts/webkitpy/layout_tests/port/base.py:
2304         * Scripts/webkitpy/layout_tests/port/chromium.py:
2305         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2306         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2307         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2308         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
2309         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
2310         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2311
2312 2011-07-06  Adam Barth  <abarth@webkit.org>
2313
2314         Move view-related code out of layout_package and into views
2315         https://bugs.webkit.org/show_bug.cgi?id=64051
2316
2317         Reviewed by Eric Seidel.
2318
2319         * Scripts/webkitpy/layout_tests/controllers/manager.py:
2320         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2321         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
2322         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py:
2323         * Scripts/webkitpy/layout_tests/layout_package/metered_stream.py: Removed.
2324         * Scripts/webkitpy/layout_tests/layout_package/metered_stream_unittest.py: Removed.
2325         * Scripts/webkitpy/layout_tests/layout_package/printing.py: Removed.
2326         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py: Removed.
2327         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2328         * Scripts/webkitpy/layout_tests/views: Added.
2329         * Scripts/webkitpy/layout_tests/views/__init__.py: Added.
2330         * Scripts/webkitpy/layout_tests/views/metered_stream.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/metered_stream.py.
2331         * Scripts/webkitpy/layout_tests/views/metered_stream_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/metered_stream_unittest.py.
2332         * Scripts/webkitpy/layout_tests/views/printing.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/printing.py.
2333         * Scripts/webkitpy/layout_tests/views/printing_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py.
2334
2335 2011-07-06  Adam Barth  <abarth@webkit.org>
2336
2337         Add goofy comments to explain why we have these files.
2338
2339         * Scripts/webkitpy/layout_tests/controllers/__init__.py:
2340         * Scripts/webkitpy/layout_tests/layout_package/__init__.py:
2341
2342 2011-07-06  Adam Barth  <abarth@webkit.org>
2343
2344         Move a bunch of classes out of layout_tests.layout_package and into layout_tests.controllers
2345         https://bugs.webkit.org/show_bug.cgi?id=64048
2346
2347         Reviewed by Eric Seidel.
2348
2349         * Scripts/webkitpy/layout_tests/controllers: Added.
2350         * Scripts/webkitpy/layout_tests/controllers/__init__.py: Added.
2351         * Scripts/webkitpy/layout_tests/controllers/manager.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/manager.py.
2352         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py.
2353         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py.
2354         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py.
2355         * Scripts/webkitpy/layout_tests/controllers/message_broker.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/message_broker.py.
2356         * Scripts/webkitpy/layout_tests/controllers/message_broker_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py.
2357         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py.
2358         * Scripts/webkitpy/layout_tests/controllers/worker.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/worker.py.
2359         * Scripts/webkitpy/layout_tests/layout_package/manager.py: Removed.
2360         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py: Removed.
2361         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py: Removed.
2362         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py: Removed.
2363         * Scripts/webkitpy/layout_tests/layout_package/message_broker.py: Removed.
2364         * Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py: Removed.
2365         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
2366         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py: Removed.
2367         * Scripts/webkitpy/layout_tests/layout_package/worker.py: Removed.
2368         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2369
2370 2011-07-06  Adam Barth  <abarth@webkit.org>
2371
2372         Attempt to fix unittest on Leopard.  Let's enumerate the keys in sorted
2373         order so the test results are predictable.
2374
2375         * Scripts/webkitpy/tool/bot/irc_command.py:
2376         * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:
2377
2378 2011-07-06  Dirk Pranke  <dpranke@chromium.org>
2379
2380         Reviewed by Eric Siedel.
2381
2382         nrwt: make the code be consistent about using test names instead of filenames or paths
2383
2384         https://bugs.webkit.org/show_bug.cgi?id=63597
2385
2386         It can be hard to figure out where in the code we're assuming
2387         test names are listed as unix-style relative filenames and where
2388         they are either absolute paths or relative paths following the
2389         host filesystem convention.
2390
2391         This patch changes things so that everything outside of the
2392         Port object uses (and must assume) unix-style relative
2393         filenames (with one exception, which is specifying host-local
2394         filenames as a list of test arguments on the command line).
2395
2396         This should make things clearer and more consistent, and also
2397         removes a lot of path conversion calls.
2398
2399         The changes in this patch outside of port/base.py are largely
2400         cosmetic changes of "filename" to "test". We add a few routines
2401         to the base Port implementation to manipulate test names to
2402         split them, figure out if we're referring to a directory of
2403         tests, etc.
2404
2405         * Scripts/webkitpy/common/net/layouttestresults.py:
2406         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
2407         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
2408         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
2409         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
2410         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
2411         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
2412         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
2413         * Scripts/webkitpy/layout_tests/layout_package/result_summary.py:
2414         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
2415         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
2416         * Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
2417         * Scripts/webkitpy/layout_tests/layout_package/test_input.py:
2418         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py:
2419         * Scripts/webkitpy/layout_tests/layout_package/test_results.py:
2420         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:
2421         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
2422         * Scripts/webkitpy/layout_tests/port/base.py:
2423         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2424         * Scripts/webkitpy/layout_tests/port/chromium.py:
2425         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2426         * Scripts/webkitpy/layout_tests/port/dryrun.py:
2427         * Scripts/webkitpy/layout_tests/port/test.py:
2428         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2429         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2430         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests.py:
2431         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests_unittest.py:
2432         * Scripts/webkitpy/tool/bot/flakytestreporter.py:
2433         * Scripts/webkitpy/tool/commands/rebaseline.py:
2434         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
2435
2436 2011-07-06  Chris Guillory  <ctguil@chromium.org>
2437
2438         No review necessary.
2439
2440         Adding myself to committers list in committers.py.
2441         https://bugs.webkit.org/show_bug.cgi?id=63478
2442
2443         * Scripts/webkitpy/common/config/committers.py:
2444
2445 2011-07-06  Johnny Ding  <jnd@chromium.org>
2446
2447         https://bugs.webkit.org/show_bug.cgi?id=63857
2448         Implement createPopupMenu for Chromium DRT.
2449
2450         Reviewed by Tony Chang.
2451
2452         * DumpRenderTree/chromium/WebViewHost.cpp:
2453         (WebViewHost::createPopupMenu):
2454         (WebViewHost::~WebViewHost):
2455         * DumpRenderTree/chromium/WebViewHost.h:
2456
2457 2011-07-06  Mihai Parparita  <mihaip@chromium.org>
2458
2459         Unreviewed, rolling out r90503.
2460         http://trac.webkit.org/changeset/90503
2461         https://bugs.webkit.org/show_bug.cgi?id=63358
2462
2463         Makes Chromium Windows canary bots hang.
2464
2465         * Scripts/webkitpy/layout_tests/port/base.py:
2466         * Scripts/webkitpy/layout_tests/port/chromium.py:
2467         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2468         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2469         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2470         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
2471         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
2472         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2473
2474 2011-07-06  Adam Barth  <abarth@webkit.org>
2475
2476         Move buildbot URL in webkitpy to config.urls
2477         https://bugs.webkit.org/show_bug.cgi?id=64041
2478
2479         Reviewed by Eric Seidel.
2480
2481         This patch prepares for using the buildbot class to talk to the
2482         Chromium buildbots.
2483
2484         * Scripts/webkitpy/common/config/urls.py:
2485         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
2486         * Scripts/webkitpy/tool/commands/download_unittest.py:
2487         * Scripts/webkitpy/tool/commands/queries.py:
2488         * Scripts/webkitpy/tool/mocktool.py:
2489         * Scripts/webkitpy/tool/steps/ensurebuildersaregreen.py:
2490
2491 2011-07-06  Kevin Ollivier  <kevino@theolliviers.com>
2492
2493         [wx] Unreviewed build fix, add the correct path to the waf build settings.
2494         
2495         * wx/packaging/build-win-installer.py:
2496
2497 2011-07-06  Adam Barth  <abarth@webkit.org>
2498
2499         Add missing include.
2500
2501         * Scripts/webkitpy/tool/bot/irc_command.py:
2502
2503 2011-07-06  Adam Barth  <abarth@webkit.org>
2504
2505         Add roll-chromium-deps command to sheriff-bot
2506         https://bugs.webkit.org/show_bug.cgi?id=64037
2507
2508         Reviewed by Eric Seidel.
2509
2510         Lo, and the DEPS were rolled on command.
2511
2512         * Scripts/webkitpy/tool/bot/irc_command.py:
2513         * Scripts/webkitpy/tool/bot/irc_command_unittest.py:
2514         * Scripts/webkitpy/tool/bot/sheriff.py:
2515         * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:
2516         * Scripts/webkitpy/tool/commands/roll.py:
2517         * Scripts/webkitpy/tool/commands/roll_unittest.py:
2518         * Scripts/webkitpy/tool/steps/updatechromiumdeps.py:
2519
2520 2011-07-06  Eric Seidel  <eric@webkit.org>
2521
2522         Split Driver, DriverOutput and DriverInput out into a new driver.py file
2523         https://bugs.webkit.org/show_bug.cgi?id=64031
2524
2525         Reviewed by Adam Barth.
2526
2527         This ended up kinda large and I'm happy to break it up if desired.
2528
2529         As part of cleaning up the Port module, I split Driver, DriverInput and DriverOutput
2530         out into its own driver.py file.
2531
2532         I also cleaned up how we import Port objects, by making port/__init__.py
2533         expose Port, Driver, DriverInput, and DriverOutput.  I removed exposure of
2534         some various test objects as they were causing circular dependencies.
2535         base.py is now an implementation detail of port.  No longer is "base" referenced
2536         anywhere in the code.
2537
2538         I also made Driver.__init__ non-virtual so that sub-classers could all share
2539         the data storage on Driver.  This more closely matches how other subclasses
2540         work throughout our python code.
2541
2542         Standard line-unwrapping and doc-string removal as I went through the code.
2543
2544         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2545         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
2546         * Scripts/webkitpy/layout_tests/port/__init__.py:
2547         * Scripts/webkitpy/layout_tests/port/base.py:
2548         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2549         * Scripts/webkitpy/layout_tests/port/chromium.py:
2550         * Scripts/webkitpy/layout_tests/port/driver.py: Added.
2551         * Scripts/webkitpy/layout_tests/port/dryrun.py:
2552         * Scripts/webkitpy/layout_tests/port/mock_drt.py:
2553         * Scripts/webkitpy/layout_tests/port/test.py:
2554         * Scripts/webkitpy/layout_tests/port/webkit.py:
2555         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2556         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests_unittest.py:
2557         * Scripts/webkitpy/tool/servers/rebaselineserver_unittest.py:
2558
2559 2011-07-06  Adam Roben  <aroben@apple.com>
2560
2561         Clear TestFailures's getNumberOfFailingTests cache
2562
2563         Now that we support NRWT, we need to clear out old zero values for the NRWT bots that might
2564         have been cached by old versions of TestFailures.
2565
2566         Fixes <http://webkit.org/b/64029> TestFailures page still doesn't show NRWT bots if you've
2567         used it recently
2568
2569         Reviewed by Adam Barth.
2570
2571         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
2572         (Builder.prototype.getNumberOfFailingTests): Added a .version property to the cache data
2573         that must match the current version. Otherwise we'll ignore the cached data.
2574
2575 2011-06-24  Dirk Pranke  <dpranke@chromium.org>
2576
2577         Reviewed by Adam Barth.
2578
2579         nrwt: remove --use-apache from the command line
2580         https://bugs.webkit.org/show_bug.cgi?id=63358
2581
2582         This change removes the --use-apache command line argument. It
2583         was initially put in when we were trying to get the cygwin
2584         apache instance to work with Chromium win, but that code has
2585         bitrotted and doesn't work at all now.
2586
2587         Arguably we should remove all of the code to allow for a choice
2588         of web servers, but since we may still want to switch off of
2589         LigHTTPd at some point on Windows, I'll leave the rest in for
2590         now (or at least yank it in a different change.)
2591
2592         * Scripts/webkitpy/layout_tests/port/base.py:
2593         * Scripts/webkitpy/layout_tests/port/chromium.py:
2594         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2595         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2596         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
2597         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2598
2599 2011-07-06  Dimitri Glazkov  <dglazkov@chromium.org>
2600
2601         garden-o-matic should use a favicon to indicate current state.
2602         https://bugs.webkit.org/show_bug.cgi?id=64027
2603
2604         Reviewed by Adam Barth.
2605
2606         * Scripts/webkitpy/tool/servers/data/gardeningserver/favicon-green.png: Added.
2607         * Scripts/webkitpy/tool/servers/data/gardeningserver/favicon-red.png: Added.
2608         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html: Added blurb about icons.
2609         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js: Added setting icon state.
2610         * Scripts/webkitpy/tool/servers/gardeningserver.py: Allowed the use of favicons.
2611
2612 2011-07-06  Daniel Bates  <dbates@rim.com>
2613
2614         Fix webkitpy error after changeset 90483 <http://trac.webkit.org/changeset/90483>; Add missing
2615         parameter to _fetch_file_from_results lamdba function.
2616
2617         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2618
2619 2011-07-06  Kevin Ollivier  <kevino@theolliviers.com>
2620
2621         [wx] Unreviewed build fix. Generate the proper install_name for dylibs on Mac.
2622
2623         * wx/packaging/build-mac-installer.py:
2624
2625 2011-07-06  Xan Lopez  <xlopez@igalia.com>
2626
2627         [GTK] Only set env. variables in start when strictly needed
2628         https://bugs.webkit.org/show_bug.cgi?id=64026
2629
2630         Reviewed by Eric Seidel.
2631
2632         * Scripts/webkitpy/layout_tests/port/gtk.py: add a
2633         setup_environ_for_server method in GtkPort and set as many
2634         env. variables as possible there.
2635
2636 2011-07-06  Adam Roben  <aroben@apple.com>
2637
2638         Teach TestFailures how to load, parse, and interpret NRWT test results
2639
2640         Fixes <http://webkit.org/b/61877> TestFailures page doesn't show testers that use
2641         new-run-webkit-tests
2642
2643         Reviewed by Adam Barth.
2644
2645         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
2646         (Builder.prototype.failureDiagnosisTextAndURL): Added support for the new 'flaky' failure
2647         type. For now we don't account for image-only flakes (but TestFailures doesn't deal with
2648         pixel tests at all currently).
2649         (Builder.prototype.getNumberOfFailingTests): Relaxed the regex used for parsing the number
2650         of failing tests from the buildbot output. Make sure not to count "new passes" (a new
2651         category introduced by NRWT) as failures.
2652
2653         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
2654         (LayoutTestResultsLoader.prototype.start): Moved a bunch of code from here to
2655         _fetchAndParseORWTResults. This function now attempts to load NRWT results, then falls back
2656         to loading ORWT results.
2657         (LayoutTestResultsLoader.prototype._fetchAndParseNRWTResults): Added. Tries to load and
2658         parse the full_results.json file for the given build.
2659         (LayoutTestResultsLoader.prototype._fetchAndParseORWTResults): Added. Code came from start.
2660         Fixed a bug along the way where we were sometimes calling the error callback instead of the
2661         success callback when all tests passed.
2662
2663         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/NRWTResultsParser.js: Added.
2664         (NRWTResultsParser): Do-nothing constructor.
2665         (NRWTResultsParser.prototype.parse): Uses eval() (eww!) to get the test results out of the
2666         JS string, then iterates over all the tests in the results data and builds up a data
2667         structure matching what ORWTResultsParser returns.
2668
2669         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
2670         (Array.prototype.contains):
2671         (String.prototype.contains):
2672         Added these simple helper functions.
2673
2674         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/index.html: Pull in
2675         NRWTResultsParser.js.
2676
2677 2011-07-06  Adam Roben  <aroben@apple.com>
2678
2679         Extract code to parse ORWT's results.html file into its own class
2680
2681         Prep work for fixing <http://webkit.org/b/61877> TestFailures page doesn't show testers that
2682         use new-run-webkit-tests
2683
2684         Reviewed by Adam Barth.
2685
2686         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
2687         (LayoutTestResultsLoader.prototype.start): Moved a bunch of code from here...
2688
2689         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ORWTResultsParser.js: Added.
2690         (ORWTResultsParser):
2691         (ORWTResultsParser.prototype.parse):
2692         ...to here.
2693
2694         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/index.html: Pull in
2695         ORWTResultsParser.js.
2696
2697 2011-07-06  Eric Seidel  <eric@webkit.org>
2698
2699         Unreviewed.  Fix a unittest from hitting the network. :)
2700
2701         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2702
2703 2011-07-06  Eric Seidel  <eric@webkit.org>
2704
2705         Unreviewed.  Fixing an exception in test-webkitpy, caused by bug 64006.
2706
2707         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2708
2709 2011-07-06  Gustavo Noronha Silva  <gustavo.noronha@collabora.com>
2710
2711         Run an Xvfb instance to run the API tests on, like we do for the
2712         layout tests.
2713
2714         Reviewed by Xan Lopez.
2715
2716         * Scripts/run-gtk-tests:
2717
2718 2011-07-06  Daniel Bates  <dbates@webkit.org>
2719
2720         Make SCM unit tests faster
2721         https://bugs.webkit.org/show_bug.cgi?id=63883
2722
2723         Speeds up the SCM unit tests by a factor of 4. Currently, we create a
2724         mock SVN repo for each test_ method in SVNTest and GitTest and creating
2725         this repo is expensive.
2726
2727         Instead, it is significantly faster to create the mock SVN repo once
2728         and then perform a filesystem copy of it for each test_ method.
2729
2730         Note, Python 2.7's unittest module implements support for per class and
2731         per module setup and tear down methods which could be used to implement
2732         similar functionality. At the time of writing, test-webkitpy is designed
2733         to support Python 2.5. So, we can't take advantage of these Python 2.7
2734         features :(
2735
2736         Reviewed by Eric Seidel.
2737
2738         * Scripts/webkitpy/common/checkout/scm/scm_unittest.py:
2739
2740 2011-07-06  Adam Barth  <abarth@webkit.org>
2741
2742         Remove unused parameter from _add_test in test_expectations.py
2743         https://bugs.webkit.org/show_bug.cgi?id=64008
2744
2745         Reviewed by Eric Seidel.
2746
2747         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
2748
2749 2011-07-06  Adam Barth  <abarth@webkit.org>
2750
2751         Use old-run-webkit-tests when --leaks is present
2752         https://bugs.webkit.org/show_bug.cgi?id=64012
2753
2754         Reviewed by Adam Roben.
2755
2756         We'll remove this branch once support of --leaks is added to
2757         new-run-webkit-tests.
2758
2759         * Scripts/run-webkit-tests:
2760
2761 2011-07-06  Benjamin Poulain  <benjamin@webkit.org>
2762
2763         [Qt] [WK2] Add touch mocking to Qt's MiniBrowser
2764         https://bugs.webkit.org/show_bug.cgi?id=63995
2765
2766         Reviewed by Andreas Kling.
2767
2768         The mouse events are intercepted through QCoreApplication::notify(),
2769         and fake touch events are generated for mouse events when necessary.
2770
2771         If touch events are received from the platform, we assume the current
2772         hardware has a touch screen and we stop generating fake events.
2773
2774         * MiniBrowser/qt/MiniBrowserApplication.cpp:
2775         (isTouchEvent):
2776         (isMouseEvent):
2777         (MiniBrowserApplication::MiniBrowserApplication):
2778         (MiniBrowserApplication::notify):
2779         * MiniBrowser/qt/MiniBrowserApplication.h:
2780
2781 2011-07-06  Xan Lopez  <xlopez@igalia.com>
2782
2783         Reviewed by Gustavo Noronha.
2784
2785         * Scripts/webkitpy/layout_tests/port/gtk.py: use Popen correctly.
2786
2787 2011-07-06  Eric Seidel  <eric@webkit.org>
2788
2789         webkit-patch failure-reason does not understand NRWT results
2790         https://bugs.webkit.org/show_bug.cgi?id=64006
2791
2792         Reviewed by Adam Barth.
2793
2794         Adam Barth tells me failure-reason should be deleted soon,
2795         but aroben's fancy new TestFailures/ page doesn't work for
2796         NRWT yet, and I needed to know when
2797         third-party-cookie-relaxing started failing.
2798
2799         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
2800         * Scripts/webkitpy/tool/commands/queries.py:
2801
2802 2011-07-06  Xan Lopez  <xlopez@igalia.com>
2803
2804         [GTK] Add missing environment variables in NWRT
2805         https://bugs.webkit.org/show_bug.cgi?id=64004
2806
2807         Reviewed by Gustavo Noronha Silva.
2808
2809         Add missing environment variables needed by the GTK+ port.
2810
2811         * Scripts/webkitpy/layout_tests/port/gtk.py: ditto.
2812
2813 2011-07-06  Adam Roben  <aroben@apple.com>
2814
2815         Make run-api-tests output intelligible on the bots
2816
2817         run-api-tests's output on the bots was very hard to read for two reasons:
2818           1) It was mixing its own output with gtest's
2819           2) It was using ANSI escape sequences to print colored text, but the bots don't support
2820              that
2821
2822         Now, the --verbose flag turns off almost all of run-api-tests's own output so that gtest's
2823         output will not be obscured. We still print "Timeout" messages even in verbose mode, since
2824         gtest doesn't have any native support for timeouts. Also, when our output is being
2825         redirected to a file, we don't print ANSI escape sequences.
2826
2827         Fixes <http://webkit.org/b/63996> It's very hard to read run-api-tests output on the bots
2828
2829         Reviewed by Sam Weinig.
2830
2831         * Scripts/run-api-tests: Don't use the :constants interface of Term::ANSIColor. We want to
2832         use the colored function instead.
2833         (runAllTestsInSuite): Don't print out suite names in verbose mode; gtest will give us enough
2834         context that they aren't needed.
2835         (runTest): Don't print out test names or pass/fail messages in verbose mode; gtest will do
2836         that for us. Also, use the new possiblyColored function instead of always coloring output.
2837         (possiblyColored): Added. When printing to a tty, returns the string with the appropriate
2838         ANSI color escape sequences added. Otherwise just returns the string unmodified.
2839
2840 2011-07-06  Xan Lopez  <xlopez@igalia.com>
2841
2842         Reviewed by Gustavo Noronha.
2843
2844         Launch Xvfb (mostly) the same way we were launching it in the
2845         bots, for maximum compatibility.
2846
2847         * Scripts/webkitpy/layout_tests/port/gtk.py:
2848
2849 2011-07-06  Andreas Kling  <kling@webkit.org>
2850
2851         [Qt][WK2] Split Qt API into two different web views (touch and desktop)
2852         https://bugs.webkit.org/show_bug.cgi?id=63950
2853
2854         Co-authored with Benjamin Poulain <benjamin@webkit.org>.
2855
2856         Reviewed by Kenneth Rohde Christiansen.
2857
2858         Port WTR/Qt to use QDesktopWebView and force a 800x600 viewport.
2859
2860         The Qt MiniBrowser now defaults to QDesktopWebView but can be started
2861         in QTouchWebView mode by passing -touch on the command line.
2862
2863         A lot of functionality has been removed and/or disabled from
2864         MiniBrowser until we reintroduce the necessary API's.
2865
2866         * MiniBrowser/qt/BrowserView.cpp:
2867         (BrowserView::BrowserView):
2868         (BrowserView::~BrowserView):
2869         (BrowserView::load):
2870         (BrowserView::view):
2871         (BrowserView::touchWebView):
2872         (BrowserView::desktopWebView):
2873         * MiniBrowser/qt/BrowserView.h:
2874         * MiniBrowser/qt/BrowserWindow.cpp:
2875         (newPageFunction):
2876         (BrowserWindow::BrowserWindow):
2877         (BrowserWindow::webView):
2878         (BrowserWindow::newWindow):
2879         (BrowserWindow::showUserAgentDialog):
2880         (BrowserWindow::updateUserAgentList):
2881         * MiniBrowser/qt/BrowserWindow.h:
2882         * MiniBrowser/qt/MiniBrowserApplication.cpp:
2883         (MiniBrowserApplication::handleUserOptions):
2884         * MiniBrowser/qt/MiniBrowserApplication.h:
2885         (WindowOptions::WindowOptions):
2886         * MiniBrowser/qt/UrlLoader.cpp:
2887         (UrlLoader::UrlLoader):
2888         * MiniBrowser/qt/main.cpp:
2889         (main):
2890         * WebKitTestRunner/qt/PlatformWebViewQt.cpp:
2891         (WTR::WebView::wkView):
2892         (WTR::WebView::pageRef):
2893         (WTR::WebView::WebView):
2894         (WTR::PlatformWebView::page):
2895         (WTR::PlatformWebView::windowFrame):
2896         (WTR::PlatformWebView::setWindowFrame):
2897
2898 2011-07-06  Eric Seidel  <eric@webkit.org>
2899
2900         Unreviewed..
2901
2902         * Scripts/run-webkit-tests:
2903         (useNewRunWebKitTests):
2904          - Fix the pass-thru of --gtk and --qt for old-run-webkit-tests,
2905            it was previously broken by the existence of isGtk/isQt checks which
2906            removed the --gtk/--qt args from @ARGV.
2907
2908 2011-07-06  Eric Seidel  <eric@webkit.org>
2909
2910         new-run-webkit-tests uses a 35s timeout for Mac and 6s for all other ports
2911         https://bugs.webkit.org/show_bug.cgi?id=63983
2912
2913         Reviewed by Xan Lopez.
2914
2915         All WebKit ports want to use a 35s timeout.  Only chromium currently
2916         uses a 6s timeout.  Eventually we'll move all WebKit ports to 6s too
2917         to further speed up new-run-webkit-tests runs, but that's going to be a while.
2918
2919         * Scripts/webkitpy/layout_tests/port/mac.py:
2920         * Scripts/webkitpy/layout_tests/port/webkit.py:
2921         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
2922
2923 2011-07-06  Xan Lopez  <xlopez@igalia.com>
2924
2925         Unreviewed.
2926
2927         NWRT actually seems to work locally, so it should work in the bots
2928         as long as we take care of the TIMEOUT issue (which is being fixed
2929         as I write this). Let's give this another shot and fix whatever
2930         breaks.
2931
2932         * Scripts/run-webkit-tests:
2933         (useNewRunWebKitTests):
2934
2935 2011-07-06  Xan Lopez  <xlopez@igalia.com>
2936
2937         Unreviewed.
2938
2939         Disable NWRT on GTK. At the very least it needs an updated exected
2940         results file since NWRT is more sensitive to timeouts, not sure
2941         what the exact problem is.
2942
2943         * Scripts/run-webkit-tests:
2944         (useNewRunWebKitTests):
2945
2946 2011-07-06  Adam Barth  <abarth@webkit.org>
2947
2948         Fix global variable leak detected by noglobals.
2949
2950         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
2951
2952 2011-07-06  Adam Barth  <abarth@webkit.org>
2953
2954         garden-o-matic shouldn't show flaky tests by default
2955         https://bugs.webkit.org/show_bug.cgi?id=63979
2956
2957         Reviewed by Ojan Vafai.
2958
2959         If any of the actual test results are PASS, the test is flaky and we
2960         can ignore it (at least for now).  If we get awesome at dealing with
2961         failing tests, we might want to tackle flaky tests by refering the user
2962         to the flakiness dashboard, which is way more awesome at dealing with
2963         flaky tests that this tool ever will be.
2964
2965         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
2966         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
2967
2968 2011-07-05  Adam Barth  <abarth@webkit.org>
2969
2970         garden-o-matic should list the failing tests
2971         https://bugs.webkit.org/show_bug.cgi?id=63978
2972
2973         Reviewed by Eric Seidel.
2974
2975         * Scripts/webkitpy/tool/servers/data/gardeningserver/config.js: Added.
2976         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
2977         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
2978         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
2979         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
2980         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html:
2981         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js: Added.
2982         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js: Added.
2983         * Scripts/webkitpy/tool/servers/gardeningserver.py:
2984
2985
2986 2011-07-05  Eric Seidel  <eric@webkit.org>
2987
2988         Re-enable NRWT on Leopard now that bug 63973 is fixed.
2989
2990         Reviewed by Adam Barth.
2991
2992         I also made the should-enable expression a black-list instead
2993         of a white-list since every bot on build.webkit.org is
2994         now using NRWT except windows and WebKit2 bots.
2995
2996         * Scripts/run-webkit-tests:
2997         (useNewRunWebKitTests):
2998
2999 2011-07-05  Eric Seidel  <eric@webkit.org>
3000
3001         Perl string concat failure on Leopard (causes NRWT to fail)
3002         https://bugs.webkit.org/show_bug.cgi?id=63973
3003
3004         Reviewed by Daniel Bates.
3005
3006         Leopard's "xcodebuild -version" has a different format from Snow Leopard/Lion
3007         thus the version parsing logic in webkitdirs.pm seems to fail.
3008
3009         Default to "3.0" instead of undef when we can't parse the version number.
3010         This should send us down the correct path for Leopard.
3011
3012         * Scripts/webkitdirs.pm:
3013         (determineBaseProductDir):
3014
3015 2011-07-05  Kevin Ollivier  <kevino@theolliviers.com>
3016
3017         [wx] Unreviewed build fix after addition of Source/WebCore/xml/parser dir.
3018
3019         * waf/build/settings.py:
3020
3021 2011-07-05  Eric Seidel  <eric@webkit.org>
3022
3023         Disable NRWT for Leopard until https://bugs.webkit.org/show_bug.cgi?id=63973 can be fixed.
3024
3025         Reviewed by Adam Barth.
3026
3027         * Scripts/run-webkit-tests:
3028         (useNewRunWebKitTests):
3029
3030 2011-07-05  Eric Seidel  <eric@webkit.org>
3031
3032         Make run-webkit-tests pass --verbose to new-run-webkit-tests on qt/gtk bots.
3033
3034         Reviewed by Adam Barth.
3035
3036         * Scripts/run-webkit-tests:
3037         (runningOnBuildBot):
3038
3039 2011-07-05  Adam Barth  <abarth@webkit.org>
3040
3041         Add results_unittests.js to garden-o-matic
3042         https://bugs.webkit.org/show_bug.cgi?id=63971
3043
3044         Reviewed by Eric Seidel.
3045
3046         Adding these unit tests forced me to commit to an API for the results
3047         module.  I'm not sure if this is the API we'll end up with, but it's
3048         something not entirely unreasonable.
3049
3050         Also, I've changed results.js over to WebKit style, as discussed previously.
3051
3052         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
3053         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
3054         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js: Added.
3055         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html:
3056
3057 2011-07-05  Eric Seidel  <eric@webkit.org>
3058
3059         Teach run-webkit-tests how to translate --qt and --gtk for new-run-webkit-tests.
3060
3061         Reviewed by Adam Barth.
3062
3063         * Scripts/run-webkit-tests:
3064
3065 2011-07-05  Adam Barth  <abarth@webkit.org>
3066
3067         Add first unit tests for garden-o-matic
3068         https://bugs.webkit.org/show_bug.cgi?id=63969
3069
3070         Reviewed by Eric Seidel.
3071
3072         I've also reformated base.js to match WebKit style, as discussed in the
3073         previous bug.
3074
3075         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
3076         * Scripts/webkitpy/tool/servers/data/gardeningserver/base_unittests.js: Added.
3077         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html: Added.
3078
3079 2011-07-05  Eric Seidel  <eric@webkit.org>
3080
3081         Move Qt, Gtk and Leopard to NRWT.
3082
3083         We'll roll out any of these if we see problems.
3084
3085         Reviewed by Adam Barth.
3086
3087         * Scripts/run-webkit-tests:
3088         (useNewRunWebKitTests):
3089
3090 2011-07-05  Adam Barth  <abarth@webkit.org>
3091
3092         garden-o-matic should understand which tests have unexpected failures
3093         https://bugs.webkit.org/show_bug.cgi?id=63965
3094
3095         Reviewed by Eric Seidel.
3096
3097         Currently, this code just logs the list of unexpected failures to the
3098         console.  A future patch will do something useful with this
3099         information.  A testing harness will also come in the next patch.
3100
3101         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js: Added.
3102         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
3103         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
3104         * Scripts/webkitpy/tool/servers/gardeningserver.py:
3105
3106 2011-07-05  Adam Barth  <abarth@webkit.org>
3107
3108         garden-o-matic should know how to fetch test results from the (NRWT) bots
3109         https://bugs.webkit.org/show_bug.cgi?id=63959
3110
3111         Reviewed by Eric Seidel.
3112
3113         Currently, this code just fetches the results and expectations JSON
3114         blob for Linux from the server and logs it to the console.
3115
3116         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
3117         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js: Added.
3118         * Scripts/webkitpy/tool/servers/gardeningserver.py:
3119
3120 2011-07-05  Dirk Pranke  <dpranke@chromium.org>
3121
3122         Reviewed by Eric Seidel.
3123
3124         Python tests are failing on leopard
3125         https://bugs.webkit.org/show_bug.cgi?id=63842
3126
3127         Fix the missing flush() call that appears to be needed by the
3128         logging package in python 2.5. Also fix the regressions
3129         introduced in the run_webkit_test tests when we switched from
3130         thread to processes (since processes aren't available on 2.5).
3131
3132         Also fix a missing "from __future__ import with_statement' in
3133         testfilehandler (unreviewed).
3134
3135         * Scripts/webkitpy/layout_tests/layout_package/metered_stream.py:
3136         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
3137         * TestResultServer/handlers/testfilehandler.py:
3138
3139 2011-07-05  Eric Seidel  <eric@webkit.org>
3140
3141         Reviewed by Adam Barth.
3142
3143         re-enabling new-run-webkit-tests on the Snow Leopard WebKit1 bots
3144         now that we believe the stray http-server issue to be solved.
3145
3146         * Scripts/run-webkit-tests:
3147         (useNewRunWebKitTests):
3148
3149 2011-07-05  Adam Barth  <abarth@webkit.org>
3150
3151         Turns out we need to support numerals too because that's what jQuery
3152         uses by default.
3153
3154         * TestResultServer/handlers/testfilehandler.py:
3155
3156 2011-07-05  Dirk Pranke  <dpranke@chromium.org>
3157
3158         nrwt: allow for multiple http shards
3159         https://bugs.webkit.org/show_bug.cgi?id=63116
3160
3161         Reviewed by Tony Chang.
3162
3163         This modifies the sharding logic to support multiple http
3164         shards, but for now we clamp to one shard until we can test
3165         perf impact and flakiness impact.
3166
3167         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3168         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
3169
3170 2011-07-05  Dirk Pranke  <dpranke@chromium.org>
3171
3172         Re-land nrwt: make sharding tests needing locks less hard-coded
3173         https://bugs.webkit.org/show_bug.cgi?id=63112
3174
3175         Reviewed by Ojan Vafai.
3176
3177         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3178         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
3179
3180 2011-07-05  Adam Barth  <abarth@webkit.org>
3181
3182         TestResultServer should support callback parameter for JSON
3183         https://bugs.webkit.org/show_bug.cgi?id=63961
3184
3185         Reviewed by Ojan Vafai.
3186
3187         This makes using jQuery.ajax so much more pleasant.
3188
3189         * TestResultServer/handlers/testfilehandler.py:
3190
3191 2011-07-05  Eric Seidel  <eric@webkit.org>
3192
3193         buildbot needs to understand whether NRWT exited early after having too many failures
3194         https://bugs.webkit.org/show_bug.cgi?id=63839
3195
3196         Reviewed by Adam Barth.
3197
3198         Fix-up the exited early messages printed by NRWT so that
3199         the buildbot can parse them as expected.
3200         It looks for lines using "if line.find('Exiting early') >= 0:"
3201
3202         I also plumbed the "early exit" status through to results.json
3203         in the form of an "interrupted" bool.  It was unclear to me
3204         if results.json already had enough information to compute this bool
3205         itself.  It's possible Ojan could come up with a better fix.
3206
3207         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3208         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
3209         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
3210         * Scripts/webkitpy/layout_tests/layout_package/result_summary.py:
3211
3212 2011-07-05  Eric Seidel  <eric@webkit.org>
3213
3214         new-run-webkit-tests fails to start http server if one is already running
3215         https://bugs.webkit.org/show_bug.cgi?id=63956
3216
3217         Reviewed by Adam Barth.
3218
3219         * BuildSlaveSupport/kill-old-processes:
3220          - Removed ^M line endings and sorted the process names.
3221          - There are no functional changes to this file.
3222            (except that now that its using unix line endings it can be executed directly!)
3223         * Scripts/webkitpy/layout_tests/servers/apache_http_server.py:
3224          - Change the name to 'httpd' to match old-run-webkit-test paths.
3225
3226 2011-07-05  Adam Barth  <abarth@webkit.org>
3227
3228         Add basic ajax support to garden-o-matic
3229         https://bugs.webkit.org/show_bug.cgi?id=63874
3230
3231         Reviewed by Eric Seidel.
3232
3233         This patch adds some basic infrastructure to garden-o-matic.
3234         Currently, the infrastructure is just used to make the "quit" command
3235         use Ajax, but in the future, this infrastructure will be used to do
3236         more sophistocated remote proceedure calls.
3237
3238         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
3239         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js: Added.
3240         * Scripts/webkitpy/tool/servers/gardeningserver.py:
3241         * Scripts/webkitpy/tool/servers/reflectionhandler.py:
3242
3243 2011-07-05  Adam Barth  <abarth@webkit.org>
3244
3245         Add trivial garden-o-matic command (with server)
3246         https://bugs.webkit.org/show_bug.cgi?id=63872
3247
3248         Reviewed by Eric Seidel.
3249
3250         This patch adds a trivial "Hello, world" garden-o-matic command,
3251         complete with an HTTP server.  This command re-uses a bunch of code
3252         from the existing rebaseline-server command.  Over time, this command
3253         will grow to be a tool useful for gardening the WebKit tree.
3254
3255         * Scripts/webkitpy/tool/commands/__init__.py:
3256         * Scripts/webkitpy/tool/commands/abstractservercommand.py: Added.
3257         * Scripts/webkitpy/tool/commands/gardenomatic.py: Added.
3258         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
3259         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html: Added.
3260         * Scripts/webkitpy/tool/servers/gardeningserver.py: Added.
3261         * Scripts/webkitpy/tool/servers/rebaselineserver.py:
3262
3263 2011-07-05  Adam Roben  <aroben@apple.com>
3264
3265         Make prepare-ChangeLog include modified Perl functions in its ChangeLog template
3266
3267         This is a very simple first cut. Functions must start with a line that starts with "sub "
3268         and end with a line that starts with a closing brace. No leading whitespace is allowed.
3269         Package names aren't parsed at all.
3270
3271         Fixes <http://webkit.org/b/21591> prepare-ChangeLog should know how to find functions in
3272         Perl files
3273
3274         Reviewed by David Kilzer.
3275
3276         * Scripts/prepare-ChangeLog: Removed redundant code that ignored certain files when
3277         searching for function line ranges. This is already done inside the get_function_line_ranges
3278         function.
3279         (get_function_line_ranges): Cleaned up coding style a little bit. Call
3280         get_function_line_ranges_for_perl for files with .pl and .pm extensions. For files with an
3281         unknown extension or no extension, read the shebang line to try to determine the script
3282         interpreter. Call get_function_line_ranges_for_perl if the interpreter seems to be Perl.
3283         (get_function_line_ranges_for_perl): Added. Does extremely basic parsing of the file to find
3284         lines starting with "sub " or "}".
3285
3286 2011-07-05  Adam Roben  <aroben@apple.com>
3287
3288         Clean up run-api-tests output on Windows
3289
3290         We were mixing run-api-tests output with gtest output, and the result was a mess.
3291
3292         Fixes <http://webkit.org/b/63954> run-api-tests output is very confusing on Windows
3293
3294         Reviewed by Dan Bates.
3295
3296         * Scripts/run-api-tests:
3297         (runTest):
3298         (populateTests):
3299         Made Windows use the formerly-Mac-only codepaths which correctly suppress gtest output
3300         except in verbose mode. Most changes are due to indentation. Use of the arch utility is now
3301         the only Mac-specific thing in these functions. Some other platform-specific code was moved
3302         from here...
3303
3304         (prepareEnvironmentForRunningTestTool):
3305         (testToolPath):
3306         ...to here.
3307
3308 2011-07-04  Gavin Barraclough  <barraclough@apple.com>
3309
3310         Reviewed by Sam Weinig.
3311
3312         https://bugs.webkit.org/show_bug.cgi?id=16652
3313         Firefox and JavaScriptCore differ in Number.toString(integer)
3314
3315         Added forwarding header.
3316
3317         * DumpRenderTree/ForwardingHeaders/wtf/StdLibExtras.h: Added.
3318
3319 2011-07-04  Anders Carlsson  <andersca@apple.com>
3320
3321         Try to fix Qt build.
3322
3323         * DumpRenderTree/TestNetscapePlugIn/Tests/NPRuntimeRemoveProperty.cpp:
3324
3325 2011-07-04  Anders Carlsson  <andersca@apple.com>
3326
3327         NP_RemoveProperty is not called back by Safari when delete npObject.prop is encountered in JavaScript
3328         https://bugs.webkit.org/show_bug.cgi?id=63915
3329         <rdar://problem/7124300>
3330
3331         Reviewed by Sam Weinig.
3332
3333         Extend the NPRuntimeRemoveProperty to handle delete object.property from JavaScript.
3334
3335         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
3336         (PluginTest::NPN_IdentifierIsString):
3337         (PluginTest::NPN_UTF8FromIdentifier):
3338         (PluginTest::NPN_IntFromIdentifier):
3339         (PluginTest::NPN_RetainObject):
3340         (PluginTest::NPN_ReleaseObject):
3341         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
3342         (PluginTest::Object::removeProperty):
3343         (PluginTest::Object::identifierIs):
3344         (PluginTest::Object::NP_RemoveProperty):
3345         (PluginTest::Object::npClass):
3346         * DumpRenderTree/TestNetscapePlugIn/Tests/NPRuntimeRemoveProperty.cpp:
3347         (NPRuntimeRemoveProperty::TestObject::TestObject):
3348         (NPRuntimeRemoveProperty::TestObject::hasProperty):
3349         (NPRuntimeRemoveProperty::TestObject::getProperty):
3350         (NPRuntimeRemoveProperty::TestObject::removeProperty):
3351         (NPRuntimeRemoveProperty::PluginObject::PluginObject):
3352         (NPRuntimeRemoveProperty::PluginObject::~PluginObject):
3353         (NPRuntimeRemoveProperty::PluginObject::hasMethod):
3354         (NPRuntimeRemoveProperty::PluginObject::invoke):
3355         (NPRuntimeRemoveProperty::PluginObject::hasProperty):
3356         (NPRuntimeRemoveProperty::PluginObject::getProperty):
3357         (NPRuntimeRemoveProperty::NPP_GetValue):
3358
3359 2011-07-04  Leandro Pereira  <leandro@profusion.mobi>
3360
3361         Reviewed by Kent Tamura.
3362
3363         [EFL] Add LayoutTestControllerEfl.
3364         https://bugs.webkit.org/show_bug.cgi?id=61974
3365
3366         Only the core features are implemented. Most functions are actually
3367         stubs and will be implemented as soon as its features are needed.
3368         Build system changes will be made as soon as DRT is completely
3369         upstreamed.
3370
3371         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: Added.
3372         (LayoutTestController::~LayoutTestController):
3373         (LayoutTestController::addDisallowedURL):
3374         (LayoutTestController::clearBackForwardList):
3375         (LayoutTestController::copyDecodedHostName):
3376         (LayoutTestController::copyEncodedHostName):
3377         (LayoutTestController::dispatchPendingLoadRequests):
3378         (LayoutTestController::display):
3379         (LayoutTestController::counterValueForElementById):
3380         (LayoutTestController::keepWebHistory):
3381         (LayoutTestController::computedStyleIncludingVisitedInfo):
3382         (LayoutTestController::nodesFromRect):
3383         (LayoutTestController::layerTreeAsText):
3384         (LayoutTestController::pageNumberForElementById):
3385         (LayoutTestController::numberOfPages):
3386         (LayoutTestController::pageProperty):
3387         (LayoutTestController::isPageBoxVisible):
3388         (LayoutTestController::pageSizeAndMarginsInPixels):
3389         (LayoutTestController::webHistoryItemCount):
3390         (LayoutTestController::workerThreadCount):
3391         (LayoutTestController::notifyDone):
3392         (LayoutTestController::pathToLocalResource):
3393         (LayoutTestController::queueLoad):
3394         (LayoutTestController::setAcceptsEditing):
3395         (LayoutTestController::setAlwaysAcceptCookies):
3396         (LayoutTestController::setCustomPolicyDelegate):
3397         (LayoutTestController::waitForPolicyDelegate):
3398         (LayoutTestController::setScrollbarPolicy):
3399         (LayoutTestController::addOriginAccessWhitelistEntry):
3400         (LayoutTestController::removeOriginAccessWhitelistEntry):
3401         (LayoutTestController::setMainFrameIsFirstResponder):
3402         (LayoutTestController::setTabKeyCyclesThroughElements):
3403         (LayoutTestController::setUseDashboardCompatibilityMode):
3404         (LayoutTestController::setUserStyleSheetEnabled):
3405         (LayoutTestController::setUserStyleSheetLocation):
3406         (LayoutTestController::setValueForUser):
3407         (LayoutTestController::setViewModeMediaFeature):
3408         (LayoutTestController::setWindowIsKey):
3409         (LayoutTestController::setSmartInsertDeleteEnabled):
3410         (waitToDumpWatchdogFired):
3411         (LayoutTestController::setWaitToDump):
3412         (LayoutTestController::windowCount):
3413         (LayoutTestController::setPrivateBrowsingEnabled):
3414         (LayoutTestController::setJavaScriptCanAccessClipboard):
3415         (LayoutTestController::setXSSAuditorEnabled):
3416         (LayoutTestController::setFrameFlatteningEnabled):
3417         (LayoutTestController::setSpatialNavigationEnabled):
3418         (LayoutTestController::setAllowUniversalAccessFromFileURLs):
3419         (LayoutTestController::setAllowFileAccessFromFileURLs):
3420         (LayoutTestController::setAuthorAndUserStylesEnabled):
3421         (LayoutTestController::setAutofilled):
3422         (LayoutTestController::disableImageLoading):
3423         (LayoutTestController::setMockDeviceOrientation):
3424         (LayoutTestController::setMockGeolocationPosition):
3425         (LayoutTestController::setMockGeolocationError):
3426         (LayoutTestController::setGeolocationPermission):
3427         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
3428         (LayoutTestController::addMockSpeechInputResult):
3429         (LayoutTestController::setIconDatabaseEnabled):
3430         (LayoutTestController::setJavaScriptProfilingEnabled):
3431         (LayoutTestController::setSelectTrailingWhitespaceEnabled):
3432         (LayoutTestController::setPopupBlockingEnabled):
3433         (LayoutTestController::setPluginsEnabled):
3434         (LayoutTestController::elementDoesAutoCompleteForElementWithId):
3435         (LayoutTestController::execCommand):
3436         (LayoutTestController::findString):
3437         (LayoutTestController::isCommandEnabled):
3438         (LayoutTestController::setCacheModel):
3439         (LayoutTestController::setPersistentUserStyleSheetLocation):
3440         (LayoutTestController::clearPersistentUserStyleSheet):
3441         (LayoutTestController::clearAllApplicationCaches):
3442         (LayoutTestController::setApplicationCacheOriginQuota):
3443         (LayoutTestController::clearApplicationCacheForOrigin):
3444         (LayoutTestController::localStorageDiskUsageForOrigin):
3445         (LayoutTestController::originsWithApplicationCache):
3446         (LayoutTestController::applicationCacheDiskUsageForOrigin):
3447         (LayoutTestController::clearAllDatabases):
3448         (LayoutTestController::setDatabaseQuota):
3449         (LayoutTestController::originsWithLocalStorage):
3450         (LayoutTestController::deleteAllLocalStorage):
3451         (LayoutTestController::deleteLocalStorageForOrigin):
3452         (LayoutTestController::observeStorageTrackerNotifications):
3453         (LayoutTestController::syncLocalStorage):
3454         (LayoutTestController::setDomainRelaxationForbiddenForURLScheme):
3455         (LayoutTestController::goBack):
3456         (LayoutTestController::setDefersLoading):
3457         (LayoutTestController::setAppCacheMaximumSize):
3458         (LayoutTestController::pauseAnimationAtTimeOnElementWithId):
3459         (LayoutTestController::pauseTransitionAtTimeOnElementWithId):
3460         (LayoutTestController::sampleSVGAnimationForElementAtTime):
3461         (LayoutTestController::numberOfActiveAnimations):
3462         (LayoutTestController::suspendAnimations):
3463         (LayoutTestController::resumeAnimations):
3464         (LayoutTestController::overridePreference):
3465         (LayoutTestController::addUserScript):
3466         (LayoutTestController::addUserStyleSheet):
3467         (LayoutTestController::setDeveloperExtrasEnabled):
3468         (LayoutTestController::setAsynchronousSpellCheckingEnabled):
3469         (LayoutTestController::showWebInspector):
3470         (LayoutTestController::closeWebInspector):
3471         (LayoutTestController::evaluateInWebInspector):
3472         (LayoutTestController::evaluateScriptInIsolatedWorld):
3473         (LayoutTestController::removeAllVisitedLinks):
3474         (LayoutTestController::callShouldCloseOnWebView):
3475         (LayoutTestController::apiTestNewWindowDataLoadBaseURL):
3476         (LayoutTestController::apiTestGoToCurrentBackForwardItem):
3477         (LayoutTestController::setWebViewEditable):
3478         (LayoutTestController::markerTextForListItem):
3479         (LayoutTestController::authenticateSession):
3480         (LayoutTestController::setEditingBehavior):
3481         (LayoutTestController::abortModal):
3482         (LayoutTestController::hasSpellingMarker):
3483         (LayoutTestController::hasGrammarMarker):
3484         (LayoutTestController::dumpConfigurationForViewport):
3485         (LayoutTestController::setSerializeHTTPLoads):
3486         (LayoutTestController::setMinimumTimerInterval):
3487         (LayoutTestController::setTextDirection):
3488         (LayoutTestController::allowRoundingHacks):
3489
3490 2011-06-24  Vsevolod Vlasov  <vsevik@chromium.org>
3491
3492         Unreviewed, add myself as committer.
3493
3494         * Scripts/webkitpy/common/config/committers.py:
3495
3496 2011-07-03  Sheriff Bot  <webkit.review.bot@gmail.com>
3497
3498         Unreviewed, rolling out r90347.
3499         http://trac.webkit.org/changeset/90347
3500         https://bugs.webkit.org/show_bug.cgi?id=63886
3501
3502         Build breaks on Leopard, Chromium-win, WinCairo, and WinCE.
3503         (Requested by tkent on #webkit).
3504
3505         * DumpRenderTree/ForwardingHeaders/wtf/StdLibExtras.h: Removed.
3506
3507 2011-07-03  Daniel Bates  <dbates@webkit.org>
3508
3509         Import Checkout in scm_unittests.py and logging in svn.py to fix
3510         Python missing global variable errors
3511
3512         Rubber-stamped by Adam Barth.
3513
3514         * Scripts/webkitpy/common/checkout/scm/scm_unittest.py: Import Checkout.
3515         * Scripts/webkitpy/common/checkout/scm/svn.py: Import logging.
3516
3517 2011-07-02  Gavin Barraclough  <barraclough@apple.com>
3518
3519         Reviewed by Sam Weinig.
3520
3521         https://bugs.webkit.org/show_bug.cgi?id=16652
3522         Firefox and JavaScriptCore differ in Number.toString(integer)
3523
3524         Added forwarding header.
3525
3526         * DumpRenderTree/ForwardingHeaders/wtf/StdLibExtras.h: Added.
3527
3528 2011-04-02  Robert Hogan  <robert@webkit.org>
3529
3530         Reviewed by Benjamin Poulain.
3531
3532         [Qt] Support third-party cookie policy for Qt clients
3533         https://bugs.webkit.org/show_bug.cgi?id=45455
3534
3535         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3536         (LayoutTestController::reset):
3537         (LayoutTestController::setAlwaysAcceptCookies):
3538         (LayoutTestController::setAlwaysBlockCookies):
3539         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3540
3541 2011-07-03  Adam Barth  <abarth@webkit.org>
3542
3543         Factor ReflectionHandler out of the rebaseline server.  Just code motion.
3544
3545         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
3546         * Scripts/webkitpy/tool/servers/rebaselineserver.py:
3547         * Scripts/webkitpy/tool/servers/reflectionhandler.py: Added.
3548
3549 2011-07-03  Adam Barth  <abarth@webkit.org>
3550
3551         Move the rebaseline server out of the commands package and into the
3552         (new) servers package.  This patch prepares for refactoring this code
3553         to share server infrastructure with the (forthcoming) garden-o-matic
3554         command.
3555
3556         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
3557         * Scripts/webkitpy/tool/servers/__init__.py: Added.
3558         * Scripts/webkitpy/tool/servers/data/rebaselineserver/loupe.js:
3559         * Scripts/webkitpy/tool/servers/data/rebaselineserver/main.css:
3560         * Scripts/webkitpy/tool/servers/data/rebaselineserver/queue.js:
3561         * Scripts/webkitpy/tool/servers/rebaselineserver.py: Added.
3562
3563 2011-07-01  Adam Barth  <abarth@webkit.org>
3564
3565         Fix typo in master.cfg, hoping to trigger a master restart.
3566
3567         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3568
3569 2011-07-01  Adam Barth  <abarth@webkit.org>
3570
3571         Turn off new-run-webkit-tests and remove some debug logging.
3572
3573         * Scripts/run-webkit-tests:
3574         * Scripts/webkitpy/layout_tests/port/http_lock.py:
3575
3576 2011-07-01  Adam Barth  <abarth@webkit.org>
3577
3578         Add some more debug logging to try to understand why we're getting
3579         stuck acquring the HTTP lock.
3580
3581         * Scripts/webkitpy/layout_tests/port/http_lock.py:
3582
3583 2011-07-01  Adam Barth  <abarth@webkit.org>
3584
3585         Refactor NRWT parsing logic in master.cfg to understand that run-webkit-tests might run NRWT
3586         https://bugs.webkit.org/show_bug.cgi?id=63854
3587
3588         Reviewed by Dirk Pranke.
3589
3590         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3591
3592 2011-07-01  Eric Seidel  <eric@webkit.org>
3593
3594         new-run-webkit-tests results does not understand that mac uses test_expectations files
3595         https://bugs.webkit.org/show_bug.cgi?id=63838
3596
3597         Reviewed by Adam Barth.
3598
3599         We've decided to "enable" test_expectations for the Mac port for now.
3600         It makes the results.html results much more confusing to read, but at least they're
3601         no longer lying to us.
3602
3603         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3604         * Scripts/webkitpy/layout_tests/port/base.py:
3605         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
3606
3607 2011-07-01  Adam Barth  <abarth@webkit.org>
3608
3609         Temporarily disable waiting for ReportCrash to see how this code
3610         affects performance on the bots.
3611
3612         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3613         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
3614
3615 2011-07-01  Dmitry Lomov  <dslomov@google.com>
3616
3617         Reviewed by Adam Roben.
3618
3619         https://bugs.webkit.org/show_bug.cgi?id=63818
3620         Test bots shouldn't attempt to build TestWebKitAPI on Windows.
3621         TestWebKitAPI builds as a part of general webkit build and is a
3622         part of build products archive.
3623
3624         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3625         * Scripts/run-api-tests:
3626
3627 2011-07-01  Adam Barth  <abarth@webkit.org>
3628
3629         We shouldn't wait for ReportCrash while the timeout is running
3630         https://bugs.webkit.org/show_bug.cgi?id=63837
3631
3632         Reviewed by Eric Seidel.
3633
3634         Instead, we should move this check to worker.py.
3635
3636         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
3637         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
3638
3639 2011-07-01  Adam Roben  <aroben@apple.com>
3640
3641         Stop setting unsafe headers on POSTs via XHR from TestFailures page
3642
3643         Fixes <http://webkit.org/b/63833> TestFailures page causes a bunch of spew in the
3644         Inspector's console about unsafe headers
3645
3646         Reviewed by Dan Bates.
3647
3648         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
3649         (fetchResource): Removed unnecessary code to set Content-Length and Connection headers.
3650
3651 2011-07-01  Eric Seidel  <eric@webkit.org>
3652
3653         Reviewed by Adam Barth.
3654
3655         Switch Snow Leopard over to new-run-webkit-tests (except for wk2)
3656         https://bugs.webkit.org/show_bug.cgi?id=63786
3657
3658         We're down to one bug blocking the "move to NRWT" bug:
3659         https://bugs.webkit.org/show_bug.cgi?id=34984
3660         so we're trying transitioning Snow Leopard to NRWT as an experiment.
3661
3662         If we run into any blocking issues we'll of course roll this back out
3663         (or anyone reading this should feel encouraged to).
3664
3665         * Scripts/run-webkit-tests:
3666
3667 2011-07-01  Adam Barth  <abarth@webkit.org>
3668
3669         Add a compat shim for --leaks.  Support for --leaks will be added in
3670         https://bugs.webkit.org/show_bug.cgi?id=63832.
3671
3672         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3673
3674 2011-07-01  Adam Barth  <abarth@webkit.org>
3675
3676         Reviewed by Darin Adler.
3677
3678         ReportCrash destabilizes new-run-webkit-tests
3679         https://bugs.webkit.org/show_bug.cgi?id=63795
3680
3681         ReportCrash chews up too many system resources that it destablizes
3682         tests run concurrently.  This patch causes us to spin down all the
3683         worker processes when ReportCrash is running.
3684
3685         Also, this patch causes the master process to tell the user we're
3686         waiting for the crash reporter rather than having that be the worker's
3687         job.  This stops the user from getting spammed with too many "waiting
3688         for crash reporter" messages and also cleans up some minor UI issues
3689         with the meter.
3690
3691         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3692         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
3693         * Scripts/webkitpy/layout_tests/port/base.py:
3694         * Scripts/webkitpy/layout_tests/port/server_process.py:
3695
3696 2011-07-01  Adam Roben  <aroben@apple.com>
3697
3698         Teach webkitpy about the new format of our ChangeLog template
3699
3700         r90229 moved the "Need a short description and bug URL" line to the top of the ChangeLog
3701         template. But webkitpy didn't know this, so the various rollout-related commands were
3702         leaving that line in the ChangeLog.
3703
3704         Fixes <http://webkit.org/b/63815> REGRESSION (r90229): webkit-patch rollout and sheriffbot
3705         rollout are broken
3706
3707         Reviewed by Dan Bates.
3708
3709         * Scripts/webkitpy/common/checkout/changelog.py:
3710         (ChangeLog.update_with_unreviewed_message): Changed to look for the "Need a short
3711         description and bug URL" line instead of the "Reviewed by" line as the first line of the
3712         ChangeLog template.
3713
3714         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
3715         (ChangeLogTest): Updated _new_entry_boilerplate to match what prepare-ChangeLog now
3716         generates.
3717
3718 2011-07-01  Adam Roben  <aroben@apple.com>
3719
3720         Make TestFailures stop loading old builds once we've explained all current failures
3721
3722         We used to do this, but some refactoring in r90165 broke this logic.
3723
3724         Fixes <http://webkit.org/b/63808> REGRESSION (r90165): TestFailures page always loads all
3725         builds back to the beginning of time
3726
3727         Reviewed by Dan Bates.
3728
3729         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
3730         (LayoutTestHistoryAnalyzer.prototype.start): Stop fetching old builds once we're told that
3731         all current failures have been explained.
3732
3733 2011-07-01  Adam Roben  <aroben@apple.com>
3734
3735         commit-log-editor reorders ChangeLog entries in unexpected ways
3736         https://bugs.webkit.org/show_bug.cgi?id=63804
3737
3738         Reviewed by Darin Adler.
3739
3740         commit-log-editor was changed in r46899 to move the bug title and URL to the top of the
3741         commit message in order to make git-based tools, which expect the first line of the commit
3742         message to be a summary of the change, to work better. But this reordering of the ChangeLog
3743         entry was unpredictable and unexpected.
3744
3745         I changed prepare-ChangeLog to put the bug title and URL at the top of the ChangeLog entry
3746         so that commit-log-editor doesn't have to move them. Then I changed commit-log-editor not to
3747         try to move them.
3748
3749         * Scripts/commit-log-editor: Partially reverted r46899. We no longer save the "Reviewed by"
3750         line and try to move it down below the bug title and URL. The order of the text in the
3751         ChangeLog is now preserved, except that we insert a "Patch by" line just above the "Reviewed
3752         by" line in cases where the committer and author are not the same person.
3753
3754         * Scripts/prepare-ChangeLog: Moved the "Reviewed by" line down below the bug title and URL.
3755
3756 2011-07-01  Mihnea Ovidenie  <mihnea@adobe.com>
3757
3758         Reviewed by Adam Roben.
3759
3760         [CSSRegions] Add build slave
3761         https://bugs.webkit.org/show_bug.cgi?id=62372
3762
3763         * BuildSlaveSupport/build.webkit.org-config/config.json:
3764         Add an entry for a release build of Mac.
3765
3766 2011-07-01  Sheriff Bot  <webkit.review.bot@gmail.com>
3767
3768         Unreviewed, rolling out r90213.
3769         http://trac.webkit.org/changeset/90213
3770         https://bugs.webkit.org/show_bug.cgi?id=63792
3771
3772         The SL bot has an errant HTTP server holding port 8080
3773         (Requested by abarth on #webkit).
3774
3775         * Scripts/run-webkit-tests:
3776
3777 2011-07-01  Carlos Garcia Campos  <cgarcia@igalia.com>
3778
3779         Unreviewed. Fix WebKit2 GTK build after r90163.
3780
3781         * MiniBrowser/gtk/WebBundle/WebBundleMain.c:
3782         (didCreatePage):
3783
3784 2011-07-01  Adam Barth  <abarth@webkit.org>
3785
3786         Exit faster when we get a keyboard interrupt.  Previously, we'd open
3787         Safari if you ^C NRWT after it found a failure.
3788
3789         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3790
3791 2011-07-01  Eric Seidel  <eric@webkit.org>
3792
3793         Reviewed by Adam Barth.
3794
3795         Switch Snow Leopard over to new-run-webkit-tests (except for wk2)
3796         https://bugs.webkit.org/show_bug.cgi?id=63786
3797
3798         We're down to one bug blocking the "move to NRWT" bug:
3799         https://bugs.webkit.org/show_bug.cgi?id=34984
3800         so we're trying transitioning Snow Leopard to NRWT as an experiment.
3801
3802         If we run into any blocking issues we'll of course roll this back out
3803         (or anyone reading this should feel encouraged to).
3804
3805         * Scripts/run-webkit-tests:
3806
3807 2011-07-01  Sheriff Bot  <webkit.review.bot@gmail.com>
3808
3809         Unreviewed, rolling out r90192.
3810         http://trac.webkit.org/changeset/90192
3811         https://bugs.webkit.org/show_bug.cgi?id=63788
3812
3813         Appears to have caused NRWT on Chromium WebKit Vista to hang
3814         (Requested by abarth on #webkit).
3815
3816         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3817         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
3818
3819 2011-06-30  Adam Barth  <abarth@webkit.org>
3820
3821         Reviewed by Eric Seidel.
3822
3823         Remove the concept of "being wedged" from new-run-webkit-tests
3824         https://bugs.webkit.org/show_bug.cgi?id=63767
3825
3826         Worker processes shouldn't ever become wedged.  My understanding is
3827         that this code was originally motivated by the old threading-based
3828         design but no longer servers any purpose.
3829
3830         Note: If we actually have a problem with the test harness getting
3831         stuck, buildbot will kill us.
3832
3833         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3834         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py:
3835         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
3836
3837 2011-06-30  Adam Barth  <abarth@webkit.org>
3838
3839         The win32 implementation of fork is sad about forking test-webkitpy.
3840         Hopefully this will heal the bot.
3841
3842         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
3843
3844 2011-06-30  Eric Seidel  <eric@webkit.org>
3845
3846         Reviewed by Adam Barth.
3847
3848         Update new-run-webkit-tests now that the wk2 shared skipped list has moved
3849         https://bugs.webkit.org/show_bug.cgi?id=63780
3850
3851         The shared list was moved from mac-wk2 to wk2 as part of
3852         http://trac.webkit.org/changeset/90125.
3853
3854         * Scripts/webkitpy/layout_tests/port/webkit.py:
3855         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
3856
3857 2011-06-30  Kenichi Ishibashi  <bashi@chromium.org>
3858
3859         Reviewed by Darin Adler.
3860
3861         Style Checker should flag removal of "developmentRegion = English;" from project.pbxproj.
3862         https://bugs.webkit.org/show_bug.cgi?id=62022
3863
3864         Adds XcodeProjectFileChecker to check removal of "developmentRegion".
3865
3866         * Scripts/webkitpy/style/checker.py: Added XcodeProjectFileChecker as a checker for *.pbxproj.
3867         * Scripts/webkitpy/style/checkers/xcodeproj.py: Added.
3868         * Scripts/webkitpy/style/checkers/xcodeproj_unittest.py: Added.
3869
3870 2011-06-30  Dirk Pranke  <dpranke@chromium.org>
3871
3872         Reviewed by Ojan Vafai.
3873
3874         nrwt: make sharding tests needing locks less hard-coded
3875         https://bugs.webkit.org/show_bug.cgi?id=63112
3876
3877         This change also changes the manager logic so that it will
3878         drop the server lock as soon as all of the shards requiring
3879         the lock have completed.
3880
3881         This change includes some minor namespace/import changes in the
3882         unit tests, and also makes the Manager a new-style object, which
3883         it should've been all along.
3884
3885         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3886         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
3887
3888 2011-06-30  Adam Barth  <abarth@webkit.org>
3889
3890         Reviewed by Eric Seidel.
3891
3892         Remove threaded mode from new-run-webkit-tests
3893         https://bugs.webkit.org/show_bug.cgi?id=63771
3894
3895         This mode is not used and is buggy.  Rather than carry around a bunch
3896         of unused buggy code, we should rip it out and focus on the
3897         multiprocess implementation.
3898
3899         * Scripts/webkitpy/common/system/executive.py:
3900         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py:
3901         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
3902         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
3903         * Scripts/webkitpy/layout_tests/port/base.py:
3904         * Scripts/webkitpy/layout_tests/port/chromium.py:
3905         * Scripts/webkitpy/layout_tests/port/mac.py:
3906         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
3907         * Scripts/webkitpy/layout_tests/port/server_process.py:
3908         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3909         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
3910
3911 2011-06-30  Adam Barth  <abarth@webkit.org>
3912
3913         Reviewed by Darin Adler.
3914
3915         Clean up output from new-run-webkit-tests
3916         https://bugs.webkit.org/show_bug.cgi?id=63759
3917
3918         Printing messages from the child process looks super ugly because of
3919         the way the pretty-printer works.  Printing a blank line first is a
3920         hack, but it makes things at least partially sane.
3921
3922         Also, handle the case where calling sample throws an exception.
3923
3924         * Scripts/webkitpy/layout_tests/port/server_process.py:
3925         * Scripts/webkitpy/layout_tests/port/server_process_unittest.py:
3926
3927 2011-06-27  Diego Gonzalez  <diegohcg@webkit.org>
3928
3929         Reviewed by Antonio Gomes.
3930
3931         [Qt] DRT: Provide LayoutTestController::setDefersLoading and ::goBack support
3932         https://bugs.webkit.org/show_bug.cgi?id=63489
3933
3934         Unskipped:
3935             loader/navigation-while-deferring-loads.html
3936
3937         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3938         (LayoutTestController::goBack):
3939         (LayoutTestController::setDefersLoading):
3940         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3941
3942 2011-06-30  Adam Barth  <abarth@webkit.org>
3943
3944         Reviewed by Dirk Pranke.
3945
3946         new-run-webkit-tests should stop when I ^C it
3947         https://bugs.webkit.org/show_bug.cgi?id=63763
3948
3949         It still doesn't stop fast enough, but at least with this patch it
3950         doesn't run more tests.
3951
3952         Also, I tweaked the UI that happens when there's a keyboard interrupt
3953         to not be amazingly ugly.  There's still a garbage "Te" left on the
3954         console, but that's a problem for another day.
3955
3956         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3957
3958 2011-06-30  Adam Roben  <aroben@apple.com>
3959
3960         Make TestFailures never treat a too-many-failures build as a last-passing build for a test
3961
3962         Fixes <http://webkit.org/b/63758> TestFailures sometimes incorrectly says that a build was
3963         the last passing build for a test, when in fact the test wasn't run due to too many failures
3964         causing old-run-webkit-tests to exit early
3965
3966         Reviewed by Anders Carlsson.
3967
3968         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
3969         (LayoutTestHistoryAnalyzer.prototype._incorporateBuildHistory): If there were too many
3970         failures in this build, assume that all the tests failed in the last build would have failed
3971         in this build, too, had they been run. Introduced some new local variables to reduce typing,
3972         and used them throughout.
3973
3974         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
3975         (LayoutTestResultsLoader.prototype.start): Fix a regression from r89841 where we wouldn't
3976         correctly remember whether old-run-webkit-tests exited early due to too many failures for
3977         builds older than r89610. Also removed an unused parameter while I was at it.
3978
3979 2011-06-30  Adam Roben  <aroben@apple.com>
3980
3981         Make TestFaiulres only load old-results directories as needed
3982
3983         Fixes <http://webkit.org/b/63752> Tester pages on TestFailures page load very slowly
3984
3985         Reviewed by Anders Carlsson.
3986
3987         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
3988         (Builder.prototype.getBuildNames): Moved up into the API section of the class. Now just
3989         calls through to _getBuildNamesFromResultsDirectory.
3990         (Builder.prototype.getOldBuildNames): Added. Just calls through to
3991         _getBuildNamesFromResultsDirectory.
3992         (Builder.prototype._getBuildNamesFromResultsDirectory): Renamed from getBuildNames. Now
3993         takes the directory URL as an argument and only fetches that single URL.
3994
3995         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
3996         (LayoutTestHistoryAnalyzer.prototype.start): Moved most logic to _analyzeBuilds. First
3997         analyzes builds from Builder.getBuildNames, then from Builder.getOldBuildNames if needed.
3998         (LayoutTestHistoryAnalyzer.prototype._analyzeBuilds): Moved logic here from start. (Most
3999         changes are just indentation.) Now takes a callback to call when we've finished analyzing
4000         all builds in buildNames so that we can try to fetch more build names if needed.
4001
4002 2011-06-30  Mark Rowe  <mrowe@apple.com>
4003
4004         Reviewed by Anders Carlsson.
4005
4006         Update MiniBrowser and WebKitTestRunner to use the client version constants.
4007
4008         * MiniBrowser/gtk/BrowserWindow.c:
4009         (browserWindowLoaderClientInit):
4010         (browserWindowUIClientInit):
4011         * MiniBrowser/gtk/WebBundle/WebBundleMain.c:
4012         (didCreatePage):
4013         (WKBundleInitialize):
4014         * MiniBrowser/gtk/main.c:
4015         (createWKContextWithInjectedBundle):
4016         * MiniBrowser/mac/AppDelegate.m:
4017         (-[BrowserAppDelegate init]):
4018         * MiniBrowser/mac/BrowserWindowController.m:
4019         (-[BrowserWindowController awakeFromNib]):
4020         * MiniBrowser/mac/WebBundle/WebBundleMain.m:
4021         (WKBundleInitialize):
4022         * MiniBrowser/win/BrowserView.cpp:
4023         (BrowserView::create):
4024         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
4025         (WTR::InjectedBundle::initialize):
4026         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
4027         (WTR::InjectedBundlePage::InjectedBundlePage):
4028         * WebKitTestRunner/TestController.cpp:
4029         (WTR::TestController::createOtherPage):
4030         (WTR::TestController::initialize):
4031
4032 2011-06-30  Adam Roben  <aroben@apple.com>
4033
4034         Identify revisions that modified failing tests on TestFailures page
4035
4036         Fixes <http://webkit.org/b/63716> TestFailures page should try to infer which commit caused
4037         a failure by looking at commit logs
4038
4039         Reviewed by David Kilzer.
4040
4041         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
4042         (.existing-bugs-list, .suspect-revisions-list): Made this rule apply to the suspect
4043         revisions list, too.
4044
4045         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Trac.js:
4046         (Trac): Initialize our memory cache.
4047         (Trac.prototype.getCommitDataForRevisionRange): Added. Fetches the commit log in RSS form,
4048         then parses out the revision number, Trac's idea of the commit title, and the files modified
4049         by the commit. (To get the modified files, we rely on the commit message including a
4050         prepare-ChangeLog-style file list.)
4051         (Trac.prototype.logURL): Added a new formatAsRSS parameter.
4052
4053         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
4054         (removePathExtension): Added. Returns a new string with the last dot and everything after it
4055         removed.
4056         (sorted): Added sortFunction parameter.
4057         (Node.prototype.removeAllChildren): Added. Does what it says.
4058
4059         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
4060         (ViewController.prototype._displayBuilder): Updated for change to _domForRegressionRange.
4061         (ViewController.prototype._domForRegressionRange): Changed argument order to match
4062         _domForNewAndExistingBugs. Now takes the list of failing tests and uses it to figure out
4063         which revisions modified the failing tests and displays the suspect revisions in a list.
4064
4065 2011-06-30  Eric Seidel  <eric@webkit.org>
4066
4067         Reviewed by Adam Barth.
4068
4069         Rename should_kill_dump_render_tree to driver_needs_restart in preparation for supporting WebKitTestRunner
4070         https://bugs.webkit.org/show_bug.cgi?id=63743
4071
4072         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
4073         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
4074
4075 2011-06-30  Martin Robinson  <mrobinson@igalia.com>
4076
4077         Reviewed by Anders Carlsson.
4078
4079         [GTK] Crash observed with nspluginwrapper and flash
4080         https://bugs.webkit.org/show_bug.cgi?id=62249
4081
4082         Added a TestNetscapePlugin test which verifies that WebKit properly
4083         handles situations where InvalidateRect is called with a null instance.
4084
4085         * DumpRenderTree/TestNetscapePlugIn/Tests/x11/CallInvalidateRectWithNullNPPArgument.cpp: Added.
4086         (CallInvalidateRectWithNullNPPArgument::CallInvalidateRectWithNullNPPArgument):
4087         (CallInvalidateRectWithNullNPPArgument::NPP_New):
4088         * GNUmakefile.am: Add the new file to sources list.
4089
4090 2011-06-30  Eric Seidel  <eric@webkit.org>
4091
4092         Reviewed by Adam Barth.
4093
4094         Remove redundant tests in TestFailuresTest
4095         https://bugs.webkit.org/show_bug.cgi?id=63742
4096
4097         These tests are all already covered by TestFailuresTest.test_loads.
4098         I also added the missing "T" from TestFailuresest. :)
4099
4100         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
4101
4102 2011-06-30  Adam Barth  <abarth@webkit.org>
4103
4104         Reviewed by Eric Seidel.
4105
4106         REGRESSION(r84294): new-run-webkit-tests results.html generate links to diffs.html or diff.png that don't exist
4107         https://bugs.webkit.org/show_bug.cgi?id=63494
4108
4109         This patch does two things:
4110
4111         1) Refactor diff_image to return the image diff instead of writing the
4112         diff to a file.  This fixes a bunch of hacks and disentangles a bunch
4113         of code.
4114
4115         2) When there's a checksum mismatch but not image diff, we no longer
4116         report an IMAGE failure to results.html.  That fixes the bug in
4117         question because results.html won't try to link to a non-existent image
4118         diff.
4119
4120         * Scripts/webkitpy/common/system/executive_mock.py:
4121         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
4122         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py:
4123         * Scripts/webkitpy/layout_tests/port/base.py:
4124         * Scripts/webkitpy/layout_tests/port/chromium.py:
4125         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
4126         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
4127         * Scripts/webkitpy/layout_tests/port/test.py:
4128         * Scripts/webkitpy/layout_tests/port/webkit.py:
4129         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
4130         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests.py:
4131
4132 2011-06-28  Hans Wennborg  <hans@chromium.org>
4133
4134         Reviewed by Tony Chang.
4135
4136         IndexedDB: Prepare for running layout tests with LevelDB
4137         https://bugs.webkit.org/show_bug.cgi?id=63593
4138
4139         The TestShell must always provide a temporary folder for LevelDB,
4140         not just when the backing store type is overridden.
4141
4142         * DumpRenderTree/chromium/LayoutTestController.cpp:
4143         (LayoutTestController::setOverrideIndexedDBBackingStore):
4144         (LayoutTestController::clearAllDatabases):
4145         * DumpRenderTree/chromium/LayoutTestController.h:
4146         * DumpRenderTree/chromium/TestShell.cpp:
4147         (TestShell::TestShell):
4148         * DumpRenderTree/chromium/TestShell.h:
4149
4150 2011-06-30  Zsolt Fehér  <h490267@stud.u-szeged.hu>
4151
4152         Reviewed by Csaba Osztrogonác.
4153
4154         [WK2] Create a general wk2 specific LayoutTests/platform/wk2 platform
4155         qt-wk2, win-wk2, mac-wk2 use this common Skipped list too
4156         https://bugs.webkit.org/show_bug.cgi?id=61785 
4157
4158         * Scripts/old-run-webkit-tests:
4159
4160 2011-06-30  Adam Roben  <aroben@apple.com>
4161
4162         Use a colon to separate test names from failure info on TestFailures
4163
4164         This matches how the "Possibly Flaky Tests" list works, and is much prettier when the
4165         failure info itself includes parentheses.
4166
4167         Fixes <http://webkit.org/b/63722> Crashing symbols on TestFailures page look ugly due to
4168         double parentheses
4169
4170         Reviewed by Darin Adler.
4171
4172         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
4173         (ViewController.prototype._domForFailedTest): Use a colon instead of surrounding the failure
4174         info in parentheses.
4175
4176 2011-06-30  Adam Roben  <aroben@apple.com>
4177
4178         Robustify PersistentCache against cached data that contains the string ': '
4179
4180         Fixes <http://webkit.org/b/63721> Exceptions thrown while loading TestFailures page due to
4181         passing a bad string to JSON.parse
4182
4183         Reviewed by Darin Adler.
4184
4185         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/PersistentCache.js:
4186         (PersistentCache._parseDateAndJSONFromString): Manually split the string on the separator
4187         using indexOf and substring. (String.prototype.split always splits on all instances of the
4188         separator.)
4189
4190 2011-06-30  Adam Roben  <aroben@apple.com>
4191
4192         Include the crashing symbol in crash logs links on TestFailures
4193
4194         Fixes <http://webkit.org/b/63465> Links to crash logs on TestFailures page should include
4195         the crashing symbol (like links in results.html do)
4196
4197         Reviewed by David Kilzer.
4198
4199         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
4200         (LayoutTestResultsLoader.prototype.start): Bumped the cache version because we now store
4201         crashing symbols for crashing tests. Renamed testsForResultTable to parseResultTable because
4202         it now returns more than just the test names. Specifically, it now looks for crash log links
4203         and extracts the crashing symbol name from them. Updated callers of parseResultTable to
4204         match its new behavior. Changed to store the crashing symbol along with the failure type in
4205         the data we pass to the callback for tests which crashed.
4206
4207         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
4208         (code): Make <code> elements a little smaller because their contents can be quite long.
4209