Unreviewed; updated chromium test expectations.
[WebKit-https.git] / Tools / ChangeLog
1 2011-07-18  Sheriff Bot  <webkit.review.bot@gmail.com>
2
3         Unreviewed, rolling out r91132 and r91135.
4         http://trac.webkit.org/changeset/91132
5         http://trac.webkit.org/changeset/91135
6         https://bugs.webkit.org/show_bug.cgi?id=64681
7
8         Broke GTK and Chromium debug bots (Requested by rniwa on
9         #webkit).
10
11         * DumpRenderTree/AccessibilityController.h:
12         * DumpRenderTree/win/AccessibilityControllerWin.cpp:
13         (AccessibilityController::AccessibilityController):
14         (AccessibilityController::~AccessibilityController):
15         (logEventProc):
16         (AccessibilityController::setLogAccessibilityEvents):
17         (AccessibilityController::addNotificationListener):
18
19 2011-07-17  Philippe Normand  <pnormand@igalia.com>
20
21         test-webkitpy failing with Python 2.5
22         https://bugs.webkit.org/show_bug.cgi?id=64594
23
24         Reviewed by Eric Seidel.
25
26         Updated simplejson to version 2.1.6 that fixes an issue when
27         dumping slashes. Also use cgi.parse_qs instead of the urlparse
28         version to be compatible with Python2.5. Verified the fix with
29         test-webkitpy on python 2.5 and python 2.6.
30
31         * Scripts/webkitpy/thirdparty/simplejson/README.txt:
32         * Scripts/webkitpy/thirdparty/simplejson/__init__.py:
33         * Scripts/webkitpy/thirdparty/simplejson/_speedups.c:
34         (json_PyOS_string_to_double):
35         (_convertPyInt_AsSsize_t):
36         (_convertPyInt_FromSsize_t):
37         (ascii_escape_char):
38         (ascii_escape_unicode):
39         (ascii_escape_str):
40         (raise_errmsg):
41         (join_list_unicode):
42         (join_list_string):
43         (_build_rval_index_tuple):
44         (scanstring_str):
45         (scanstring_unicode):
46         (py_scanstring):
47         (py_encode_basestring_ascii):
48         (scanner_dealloc):
49         (scanner_traverse):
50         (scanner_clear):
51         (_parse_object_str):
52         (_parse_object_unicode):
53         (_parse_array_str):
54         (_parse_array_unicode):
55         (_parse_constant):
56         (_match_number_str):
57         (_match_number_unicode):
58         (scan_once_str):
59         (scan_once_unicode):
60         (scanner_call):
61         (scanner_new):
62         (scanner_init):
63         (encoder_new):
64         (encoder_init):
65         (encoder_call):
66         (_encoded_const):
67         (encoder_encode_float):
68         (encoder_encode_string):
69         (_steal_list_append):
70         (encoder_listencode_obj):
71         (encoder_listencode_dict):
72         (encoder_listencode_list):
73         (encoder_dealloc):
74         (encoder_traverse):
75         (encoder_clear):
76         (init_speedups):
77         * Scripts/webkitpy/thirdparty/simplejson/decoder.py:
78         * Scripts/webkitpy/thirdparty/simplejson/encoder.py:
79         * Scripts/webkitpy/thirdparty/simplejson/ordered_dict.py: Added.
80         * Scripts/webkitpy/thirdparty/simplejson/scanner.py:
81         * Scripts/webkitpy/thirdparty/simplejson/tool.py: Added.
82         * Scripts/webkitpy/tool/servers/reflectionhandler.py:
83
84 2011-07-17  Dimitri Glazkov  <dglazkov@chromium.org>
85
86         Rename ModifierMatcher to SpecificityCalculator.
87         https://bugs.webkit.org/show_bug.cgi?id=64660
88
89         It's a little longer than before, but it is much clearer.
90
91         Reviewed by Adam Barth.
92
93         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Renamed and adjusted
94             all callsites, also rewrote the comments.
95         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: Adjusted callsites
96             and renamed tests.
97
98 2011-07-15  Adam Barth  <abarth@webkit.org>
99
100         gardening server should proxy buildbot status for garden-o-matic
101         https://bugs.webkit.org/show_bug.cgi?id=64588
102
103         Reviewed by Eric Seidel.
104
105         This information will be used to detect build breaks.  I took this
106         opportunity to make ChromiumBuildBot non-static.
107
108         * Scripts/webkitpy/common/host.py:
109         * Scripts/webkitpy/tool/commands/rebaseline.py:
110         * Scripts/webkitpy/tool/commands/rebaseline_unittest.py:
111         * Scripts/webkitpy/tool/mocktool.py:
112         * Scripts/webkitpy/tool/servers/gardeningserver.py:
113         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py:
114
115 2011-07-15  Julien Chaffraix  <jchaffraix@webkit.org>
116
117         [NRWT] Add support for --no-http
118         https://bugs.webkit.org/show_bug.cgi?id=64564
119
120         Reviewed by Dirk Pranke.
121
122         Added support for --no-http, which disables both HTTP and websockets tests.
123         It also matches the old-run-webkit-tests behavior if --force is used.
124
125         * Scripts/webkitpy/layout_tests/controllers/manager.py:
126         Fixed HTTP_SUBDIR and WEBSOCKET_SUBDIR as tests do not start with a leading separator.
127         We check if --no-http is set and add the HTTP / websockets tests to the skipped list prior to looking
128         at the expectation file. Fixed the  _test_requires_lock function to use the same code path to determine
129         what is worth have an HTTP lock as --no-http to avoid badness.
130
131         * Scripts/webkitpy/layout_tests/port/test.py: Added 2 new tests to our mock filesystem to validate that
132         we do skip properly HTTP / websocket tests inside platform/.
133
134         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
135         Added tests that we properly skip all the tests.
136
137         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
138         Added checks for the command line arguments.
139
140 2011-07-13  Jon Honeycutt  <jhoneycutt@apple.com>
141
142         Focus and selection events are not fired when a <select>'s selection
143         changes
144         https://bugs.webkit.org/show_bug.cgi?id=64504
145         <rdar://problem/9319881>
146
147         Reviewed by Alice Liu.
148
149         * DumpRenderTree/AccessibilityController.h:
150         Added m_notificationsEventHook for addNotificationListener().
151         m_allEventsHook will now be used for setLogAccessibilityEvents().
152
153         * DumpRenderTree/win/AccessibilityControllerWin.cpp:
154         (AccessibilityController::AccessibilityController):
155         Initialize m_notificationsEventHook.
156         (AccessibilityController::~AccessibilityController):
157         Turn off logging of all accessibility events. If
158         m_notificationsEventHook is non-null, unhook it.
159         (logEventProc):
160         Add handling of EVENT_OBJECT_SELECTION.
161         (AccessibilityController::setLogAccessibilityEvents):
162         If the state of logging is not changing, return early. If we're turning
163         off logging, unhook m_allEventsHook, and zero it out. Otherwise, add a
164         hook for all events.
165         (AccessibilityController::addNotificationListener):
166         Use m_notificationsEventHook rather than m_allEventsHook.
167
168 2011-07-15  Dimitri Glazkov  <dglazkov@chromium.org>
169
170         Refactor TestExpectationModel to use TestExpectationLine as data item.
171         https://bugs.webkit.org/show_bug.cgi?id=64635
172
173         This is a bit largish in scope. Does the following things:
174
175         1) Adds "path" member to TestExpectationLine to hold normalized path to test, computed at parsing,
176            and changes code that used Port.normalize_test_name to rely on TestExpectationLine.path. As a result, TestExpectationModel no longer
177            needs to have any port knowledge.
178
179         2) Adds "create_passing_expectation" class method to TestExpectationLine to generate a pristine passing expectation out of a test name,
180            and changes TestExpectations._process_tests_without_expectations to use it, thus eliminating the need for a special API entry point.
181            Now all expectations are added to the model in the same way!
182
183         3) Changes TestExpectationModel's main test index to store a tuple consisting of line number and TestExpectationLine instance.
184
185         Reviewed by Adam Barth.
186
187         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Refactored code.
188
189 2011-07-15  Adam Roben  <aroben@apple.com>
190
191         Teach TestFailures how to detect interrupted build steps
192
193         Fixes <http://webkit.org/b/64619> TestFailures page thinks all tests passed in
194         http://build.webkit.org/builders/SnowLeopard%20Intel%20Release%20(WebKit2%20Tests)/builds/13401
195
196         Reviewed by Daniel Bates.
197
198         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
199         (Builder.prototype.getNumberOfFailingTests): If the build step has a result code of 4, the
200         build step was interrupted. Treat it as an error (by returning a failureCount of -1).
201
202         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder_unittests.js:
203         Added a test for the above.
204
205         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
206         (LayoutTestResultsLoader.prototype.start): Bumped the cache version to evict old, buggy
207         cache data that was afflicted by the above bug.
208
209 2011-07-15  Eric Seidel  <eric@webkit.org>
210
211         new-run-webkit-tests crashes on Apple's Windows port when trying to launch Apache
212         https://bugs.webkit.org/show_bug.cgi?id=64533
213
214         Reviewed by Adam Roben.
215
216         Blind removal of Chromium-specific code (now that Chromium does not use this file).
217
218         * Scripts/webkitpy/layout_tests/servers/apache_http_server.py:
219
220 2011-07-15  Adam Roben  <aroben@apple.com>
221
222         Make TestFailures's list of possibly-flaky tests not so tall
223
224         Fixes <http://webkit.org/b/64618> TestFailures page's list of flaky tests takes up way too
225         much room
226
227         Reviewed by Daniel Bates.
228
229         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
230         Removed a no-longer-used .flakiness-example-separator selector.
231
232         (#failure-history > li):
233         (#possibly-flaky-tests > li):
234         (#failure-history > li, #possibly-flaky-tests > li):
235         Use a lot less padding for the possibly-flaky tests.
236
237 2011-07-15  Tony Chang  <tony@chromium.org>
238
239         [chromium] land linux selection color layout test
240         https://bugs.webkit.org/show_bug.cgi?id=64631
241
242         Reviewed by Ojan Vafai.
243
244         I had removed this code in r76620 because I thought it wasn't used.
245         Turns out there was a layout test in the chromium tree that never
246         got migrated that used it.  Re-add the code and land the layout
247         test (I will remove it from the chromium tree soon).
248
249         * DumpRenderTree/chromium/LayoutTestController.cpp:
250         (LayoutTestController::LayoutTestController):
251         (LayoutTestController::forceRedSelectionColors):
252         * DumpRenderTree/chromium/LayoutTestController.h:
253
254 2011-07-15  Dimitri Glazkov  <dglazkov@chromium.org>
255
256         Revert http://trac.webkit.org/changeset/91091, since Python 2.5
257         does not support enumerate() start argument.
258
259         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Reverted.
260
261 2011-07-15  Dimitri Glazkov  <dglazkov@chromium.org>
262
263         Clean up test_expectations.py after refactorings.
264         https://bugs.webkit.org/show_bug.cgi?id=64620
265
266         * Renamed all variables holding TestExpectationLine to expectation_line to avoid confusion with
267           its sub-part, the actual expectation.
268         * Renamed all references to options to modifier to eliminate dual terminology.
269         * Made a bunch of parser constants and changed all callsites to use them.
270         * Various other minor clean-ups.
271
272         Reviewed by Adam Barth.
273
274         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Cleaned up stuff.
275         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: Changed callsites after cleanup.
276         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py: Ditto.
277
278 2011-07-15  Dimitri Glazkov  <dglazkov@chromium.org>
279
280         Move expectation parsing code to TestExpectationParser.
281         https://bugs.webkit.org/show_bug.cgi?id=64605
282
283         This is a somewhat mechanical move, with two interesting bits:
284         
285         1) TestExpectationParser.parse methods renamed to tokenize, to better
286         reflect what they do
287
288         2) TestExpectationLine now carries all of its info, from tokens to parsed data,
289         and even the list of tests that it matches (a line may refer to more than one test).
290
291         Reviewed by Adam Barth.
292
293         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Moved parsing-related TestExpectaions methods
294             to TestExpectationParser, added more members to TestExpectationLine to carry parsed info, renamed existing parse methods
295             to "tokenize", changed callsites to use new code.
296         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: Changed unit tests to reflect renames.
297
298 2011-07-15  Dimitri Glazkov  <dglazkov@chromium.org>
299
300         Switch to use Python enumerate function to enumerate line numbers.
301         https://bugs.webkit.org/show_bug.cgi?id=64602
302
303         Reviewed by Adam Barth.
304
305         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Changed to use enumerate.
306
307 2011-07-15  Martin Robinson  <mrobinson@igalia.com>
308
309         Build fixes for WebKit2. Ensure that all generated sources are
310         on nodist primaries, that they are on forward declared variables
311         so that BUILT_SOURCES is calculated properly and that zlib is 
312         included during linking (for WOFF support).
313
314         * WebKitTestRunner/GNUmakefile.am:
315
316 2011-07-15  Dimitri Glazkov  <dglazkov@chromium.org>
317
318         Store error and warning information on TestExpectationLine.
319         https://bugs.webkit.org/show_bug.cgi?id=64565
320
321         Keeping errors and warnings on the TestExpectationLine instance allows us to decouple storing errors
322         from various parsing and validation mechanisms and have more flexibility in reporting and understanding the origin of the errors.
323
324         Reviewed by Adam Barth.
325
326         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Added TestExpectationLine.warnings list to keep track of non-fatal errors,
327             converted the code to add errors and warnings to corresponding TestExpectationLine instances, removed the code that used to store
328             this info on TestExpectations. In the process, had to refactor ModifierMatcher a bit to take in a TestExpdectationLine instance.
329         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: Changed ModifierMatcher callsite.
330
331 2011-07-15  Dimitri Glazkov  <dglazkov@chromium.org>
332
333         Plumb the use of TestExpectationLine deeper, clean up.
334         https://bugs.webkit.org/show_bug.cgi?id=64559
335
336         Instead of carrying various bits of TestExpectationLine, plumb it down to its consumers,
337         also cleaning up names and remove an unused TestExpectations._get_options_list member.
338
339         Reviewed by Adam Barth.
340
341         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
342
343 2011-07-15  Dimitri Glazkov  <dglazkov@chromium.org>
344
345         Remove the notion of TestExpectationLine.valid, start storing parsing errors in expectations themselves.
346         https://bugs.webkit.org/show_bug.cgi?id=64554
347
348         This moves us toward the world where errors are collected on the expectations, which allows us to
349         easily enumerate them, keep association with the point of origin, and freely pass TestExpectationLine instances around.
350
351         Also eliminate the validator idea, since validation is a context-dependent concept and has to be decoupled from initial
352         parsing.
353
354         Reviewed by Adam Barth.
355
356         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Removed TestExpectation.valid, validator,
357             changed TestExpectationParser to collect errors in TestExpectationLine, refactored surrounding code.
358         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: Changed tests to accommodate changes.
359         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py: Ditto.
360
361 2011-07-15  Adam Roben  <aroben@apple.com>
362
363         Rename TestFailureBugForm to FailingTestsBugForm
364
365         The new name will match better with a forthcoming FlakyTestBugForm class.
366
367         Fixes <http://webkit.org/b/64598> TestFailures page's TestFailureBugForm class has a bad
368         name
369
370         Reviewed by Daniel Bates.
371
372         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/FailingTestsBugForm.js: Renamed from Tools/BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailureBugForm.js.
373         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/FailingTestsBugForm_unittests.js: Renamed from Tools/BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailureBugForm_unittests.js.
374
375         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
376         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/run-unittests.html:
377         Updated for renames.
378
379         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/index.html: Ditto, and
380         reordered <script>s to put files which others depend on for parsing first.
381
382 2011-07-15  Adam Barth  <abarth@webkit.org>
383
384         NRWT stores the Chromium revision number in full_results.json
385         https://bugs.webkit.org/show_bug.cgi?id=64586
386
387         I have no earthly idea how to test this change.
388
389         * Scripts/webkitpy/layout_tests/controllers/manager.py:
390         * Scripts/webkitpy/layout_tests/port/base.py:
391
392 2011-07-14  Ojan Vafai  <ojan@chromium.org>
393
394         fix remaining style issues in the static-dashboards directory
395         https://bugs.webkit.org/show_bug.cgi?id=64561
396
397         Reviewed by Adam Barth.
398
399         * TestResultServer/static-dashboards/aggregate_results.html:
400         * TestResultServer/static-dashboards/builders.js:
401         * TestResultServer/static-dashboards/dashboard_base.js:
402         * TestResultServer/static-dashboards/flakiness_dashboard_tests.js:
403         * TestResultServer/static-dashboards/timeline_explorer.html:
404         * TestResultServer/static-dashboards/treemap.html:
405
406 2011-07-14  Yuta Kitamura  <yutak@chromium.org>
407
408         WebSocket: Introduce pywebsocket-0.6b2
409         https://bugs.webkit.org/show_bug.cgi?id=64534
410
411         Reviewed by Kent Tamura.
412
413         * Scripts/webkitpy/thirdparty/__init__.py:
414
415 2011-07-14  Dimitri Glazkov  <dglazkov@chromium.org>
416
417         Introduce TestExpectationsModel, split out of TestExpectations.
418         https://bugs.webkit.org/show_bug.cgi?id=64531
419
420         This is a simple split-and-make-work refactoring, a first step among many.
421
422         Reviewed by Adam Barth.
423
424         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Moved all model-related members
425             out of TestExpectations and into TestExpectationsModel.
426
427 2011-07-14  Eric Seidel  <eric@webkit.org>
428
429         Move webkitpy off of loose mocks
430         https://bugs.webkit.org/show_bug.cgi?id=64508
431
432         Unreviewed.  Fixing a test which fails under test-webkitpy --all (but not in a normal run).
433
434         * Scripts/webkitpy/common/checkout/scm/git.py:
435
436 2011-07-14  Adam Barth  <abarth@webkit.org>
437
438         garden-o-matic should have a "rebaseline" button
439         https://bugs.webkit.org/show_bug.cgi?id=64446
440
441         Reviewed by Ojan Vafai.
442
443         This patch adds a basic Rebaseline button that copies the baselines
444         displayed in the results pane into the appropriate directory in your
445         working copy.
446
447         There are two main limitations:
448
449         1) There is no UI for actually committing the baselines.
450
451         2) The baselines are not optimized for redundancy (meaning you can have
452            identical baselines in both chromium-mac and chromium-win).
453
454         * Scripts/webkitpy/tool/commands/rebaseline.py:
455             - Turns out we need to create the directory for the baseline if it doesn't exist yet.
456         * Scripts/webkitpy/tool/servers/data/gardeningserver/checkout.js:
457             - Add an programatic API to call the server.
458         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
459             - Add the rebaseline button itself.
460         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
461             - Change the CSS so that the Rebaseline and Close buttons can
462               appear in the normal order in the DOM.
463         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
464             - Bind the event and translate the parameters.
465             - Hide/show the rebaseline button, as appropriate.
466         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
467             - Add some helpful utility functions for manipulating failure types.
468         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
469
470 2011-07-14  Ryosuke Niwa  <rniwa@webkit.org>
471
472         REGRESSION: webkit-patch roll-chromium-deps is broken
473         https://bugs.webkit.org/show_bug.cgi?id=64568
474
475         Reviewed by Dirk Pranke
476
477         Use _filesystem.join.
478
479         * Scripts/webkitpy/common/checkout/checkout.py:
480         * Scripts/webkitpy/common/checkout/checkout_unittests.py:
481
482 2011-07-14  Noam Rosenthal  <noam.rosenthal@nokia.com>
483
484         Adding myself to the reviewers list; No review needed.
485
486         * Scripts/webkitpy/common/config/committers.py:
487
488 2011-07-14  Adam Roben  <aroben@apple.com>
489
490         Don't use Element.prototype.classList in TestFailures
491
492         Safari 5 doesn't support it.
493
494         Fixes <http://webkit.org/b/64550> Can't expand flaky tests on TestFailures page in Safari 5
495
496         Reviewed by Daniel Bates.
497
498         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
499         (Element.prototype.hasStyleClass):
500         (Element.prototype.addStyleClass):
501         (Element.prototype.removeStyleClass):
502         (Element.prototype.toggleStyleClass):
503         Added these helper functions which simulate classList functionality.
504
505         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities_unittests.js:
506         Added. Tests for the above.
507
508         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
509         (ViewController.prototype._domForPossiblyFlakyTests): Changed to use
510         toggleStyleClass/hasStyleClass instead of classList.
511
512         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/run-unittests.html:
513         Added Utilities_unittests.js, and reordered the tested files to be in
514         roughly dependency order (i.e., the lowest-level files are imported
515         and tested first).
516
517 2011-07-14  Eric Seidel  <eric@webkit.org>
518
519         NRWT doesn't store the svn revision in full_results.json on chromium-win
520         https://bugs.webkit.org/show_bug.cgi?id=64492
521
522         Unreviewed.  Just fixing my test-webkitpy regression.
523
524         Fix the unit tests.  Unfortunately scm does not use a filesystem
525         object so we can't control the result of detect_scm_system.
526         When detect_scm_system would fail, we would log, which would
527         cause all passing_run integration tests to fail.
528
529         * Scripts/webkitpy/common/checkout/scm/scm.py:
530         * Scripts/webkitpy/layout_tests/controllers/manager.py:
531         * Scripts/webkitpy/layout_tests/port/base.py:
532         * Scripts/webkitpy/layout_tests/port/test.py:
533         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
534
535 2011-07-14  Eric Seidel  <eric@webkit.org>
536
537         NRWT doesn't store the svn revision in full_results.json on chromium-win
538         https://bugs.webkit.org/show_bug.cgi?id=64492
539
540         Reviewed by Ojan Vafai.
541
542         This should fix the bug.  Unfortunately this code is currently
543         impossible to test since it's impossible to mock detect_scm_system at the
544         moment.  I started re-writing scm.detection.py to be mockable, but decided that
545         was best left for another day.
546
547         * Scripts/webkitpy/layout_tests/controllers/manager.py:
548         * Scripts/webkitpy/layout_tests/port/base.py:
549
550 2011-07-14  Ojan Vafai  <ojan@chromium.org>
551
552         fix coding style of dashboard_base.js
553         https://bugs.webkit.org/show_bug.cgi?id=64545
554
555         Reviewed by Adam Roben.
556
557         The code changes to the files other than dashboard_base.js
558         are just fallout from renaming globals in dashboard_base.js.
559
560         * TestResultServer/static-dashboards/aggregate_results.html:
561         * TestResultServer/static-dashboards/builders.js:
562         * TestResultServer/static-dashboards/dashboard_base.js:
563         * TestResultServer/static-dashboards/flakiness_dashboard.html:
564         * TestResultServer/static-dashboards/flakiness_dashboard_tests.js:
565         * TestResultServer/static-dashboards/timeline_explorer.html:
566         * TestResultServer/static-dashboards/treemap.html:
567
568 2011-07-14  Carlos Garcia Campos  <cgarcia@igalia.com>
569
570         Unreviewed. Fix WebKit2 GTK build after r90953.
571
572         * WebKitTestRunner/GNUmakefile.am:
573
574 2011-07-14  Adam Barth  <abarth@webkit.org>
575
576         garden-o-matic results pane should be more discoverable
577         https://bugs.webkit.org/show_bug.cgi?id=64513
578
579         Reviewed by Eric Seidel.
580
581         This patch causes us to trigger the results pane on mouse clicks
582         instead of mousein.  Also, handle the case of no results details more
583         elegantly.
584
585         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
586         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
587             - Change the binding to "click".
588         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
589             - Handle the "no results URLs" case explicitly instead of never
590               calling the callback.
591         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
592         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
593             - Show a message when there are no result details to show.
594         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
595
596 2011-07-14  Adam Barth  <abarth@webkit.org>
597
598         garden-o-matic should badge TIMEOUT failures
599         https://bugs.webkit.org/show_bug.cgi?id=64435
600
601         Reviewed by Eric Seidel.
602
603         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
604         (.regression .what a[draggable].TIMEOUT::before):
605             - Add TIMEOUT.
606
607 2011-07-14  Adam Barth  <abarth@webkit.org>
608
609         Fix Python exception blocking the commit-queue.
610
611         * Scripts/webkitpy/common/checkout/scm/git.py:
612
613 2011-07-13  Eric Seidel  <eric@webkit.org>
614
615         Move webkitpy off of loose mocks
616         https://bugs.webkit.org/show_bug.cgi?id=64508
617
618         Reviewed by Adam Barth.
619
620         Using Mock has caused us more pain than help.
621         It's possible that there was a cleaner way to use it
622         (maybe Mock(class) instead of inheriting from it?).
623         But for now, I've removed all uses of Mock from mocktool.py.
624
625         I also moved run_command into the only 3 files which call it
626         instead of leaving the deprecated method in executive.py.
627
628         I changed various direct calls to os.* to use filesystem instead.
629
630         * Scripts/webkitpy/common/checkout/checkout.py:
631         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
632         * Scripts/webkitpy/common/checkout/scm/git.py:
633         * Scripts/webkitpy/common/checkout/scm/scm.py:
634         * Scripts/webkitpy/common/checkout/scm/scm_unittest.py:
635         * Scripts/webkitpy/common/checkout/scm/svn.py:
636         * Scripts/webkitpy/common/system/executive.py:
637         * Scripts/webkitpy/common/system/executive_unittest.py:
638         * Scripts/webkitpy/tool/mocktool.py:
639         * Scripts/webkitpy/tool/steps/cleanworkingdirectory_unittest.py:
640         * Scripts/webkitpy/tool/steps/ensurelocalcommitifneeded.py:
641
642 2011-07-13  Eric Seidel  <eric@webkit.org>
643
644         NRWT doesn't store the svn revision in full_results.json on chromium-win
645         https://bugs.webkit.org/show_bug.cgi?id=64492
646
647         Reviewed by Adam Barth.
648
649         Add a warning message in the revision='' case to help us diagnose
650         what's going wrong on the bot.
651
652         * Scripts/webkitpy/layout_tests/controllers/manager.py:
653
654 2011-07-13  Dirk Pranke  <dpranke@chromium.org>
655
656         new-run-webkit-tests: print baseline search path as part of config output
657         https://bugs.webkit.org/show_bug.cgi?id=64499
658
659         Reviewed by Eric Seidel.
660
661         * Scripts/webkitpy/layout_tests/controllers/manager.py:
662         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
663
664 2011-07-13  Adam Barth  <abarth@webkit.org>
665
666         gardening server should have an API for parsing changelogs
667         https://bugs.webkit.org/show_bug.cgi?id=64495
668
669         Reviewed by Eric Seidel.
670
671         This patch exposes much of the same information from CommitInfo in a
672         dictionary form, which is easier to send over-the-wire as JSON to the
673         frontend.
674
675         * Scripts/webkitpy/common/checkout/checkout.py:
676         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
677         * Scripts/webkitpy/tool/mocktool.py:
678         * Scripts/webkitpy/tool/servers/gardeningserver.py:
679         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py:
680         * Scripts/webkitpy/tool/servers/reflectionhandler.py:
681
682 2011-07-13  Eric Seidel  <eric@webkit.org>
683
684         REGRESSION: GitTestWithMock.test_create_patch fails
685         https://bugs.webkit.org/show_bug.cgi?id=62945
686
687         Reviewed by Daniel Bates.
688
689         I was not able to reproduce the exact failure seen in the bug,
690         however this test was failing on my machine for other reasons.
691
692         I went through and did an audit of our run_command usage, it's
693         entirely in scm classes after this change.  (Not surprising given
694         that scm.py was the second file ever created in webkit.py.)
695
696         The real bug I'm fixing here is that we were setting executive.should_log
697         when the value had changed to executive._should_log.  Now we set the right one
698         and the test works again.
699
700         * Scripts/webkitpy/common/checkout/checkout.py:
701         * Scripts/webkitpy/common/checkout/scm/git.py:
702         * Scripts/webkitpy/common/checkout/scm/scm.py:
703         * Scripts/webkitpy/common/checkout/scm/scm_unittest.py:
704         * Scripts/webkitpy/common/checkout/scm/svn.py:
705
706 2011-07-13  Ilya Sherman  <isherman@chromium.org>
707
708         Fix WTF header guard style check
709         https://bugs.webkit.org/show_bug.cgi?id=64488
710
711         Reviewed by David Levin.
712
713         * Scripts/webkitpy/style/checkers/cpp.py:
714         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Verify that we don't *always* suggest "WTF_" as a prefix
715
716 2011-07-13  Ojan Vafai  <ojan@chromium.org>
717
718         bring flakiness_dashboard.html closer to webkit style
719         https://bugs.webkit.org/show_bug.cgi?id=64477
720
721         Reviewed by Adam Barth.
722
723         * TestResultServer/static-dashboards/dashboard_base.js:
724         Fix bad variable name that would hit only when running the tests.
725         * TestResultServer/static-dashboards/flakiness_dashboard.html:
726         No code/logic changes. All moving/removing brackets, indents and moving things
727         to one line.
728
729 2011-07-13  Xan Lopez  <xlopez@igalia.com>
730
731         [GTK] Fix distcheck
732
733         Reviewed by Martin Robinson.
734
735         * WebKitTestRunner/GNUmakefile.am: mark built sources as nodist.
736
737 2011-07-13  Joseph Pecoraro  <joepeck@webkit.org>
738
739         Unreviewed fix. Add a "\n" to a printf that somehow got lost.
740
741         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
742         (WebCore::DumpRenderTree::dumpApplicationCacheQuota):
743
744 2011-07-13  Joseph Pecoraro  <joepeck@webkit.org>
745
746         ApplicationCache Quota Output is Flakey
747         https://bugs.webkit.org/show_bug.cgi?id=64410
748
749         Reviewed by Alexey Proskuryakov.
750
751         Unify the delegate logging for reaching application cache
752         quotas. Truncate the space needed to the nearest 10000
753         for less flakey test output. This also better supports
754         printing NSUInteger on non-64 bit systems by casting to
755         an unsigned long.
756
757         * DumpRenderTree/mac/UIDelegate.mm:
758         (-[UIDelegate webView:exceededApplicationCacheOriginQuotaForSecurityOrigin:totalSpaceNeeded:]):
759         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
760         (WebCore::DumpRenderTree::dumpApplicationCacheQuota):
761
762 2011-07-13  Dimitri Glazkov  <dglazkov@chromium.org>
763
764         Remove unused TestExpectations._overrides.
765         https://bugs.webkit.org/show_bug.cgi?id=64470
766
767         Reviewed by Adam Barth.
768
769         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Removed unused member.
770
771 2011-07-13  Dimitri Glazkov  <dglazkov@chromium.org>
772
773         Introduce TestExpectationSerializer.list_to_string.
774         https://bugs.webkit.org/show_bug.cgi?id=64462
775
776         Reviewed by Adam Barth.
777
778         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Added list_to_string and change the relevant callsite to use it.
779         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: Added tests for it.
780
781 2011-07-13  Dimitri Glazkov  <dglazkov@chromium.org>
782
783         Consolidate expectations parsing code.
784         https://bugs.webkit.org/show_bug.cgi?id=64460
785
786         Reviewed by Adam Barth.
787
788         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Folded TestExpectationParser._split_expectation_string into TestExpectationParser.parse.
789
790 2011-07-12  Brent Fulgham  <bfulgham@webkit.org>
791
792         Standardize WinCairo conditionalized code under PLATFORM macro.
793         https://bugs.webkit.org/show_bug.cgi?id=64377
794
795         Reviewed by Maciej Stachowiak.
796
797         Update compile-time conditions that had previously used #ifdef
798         of WIN_CAIRO to consistently use PLATFORM(WIN_CAIRO).
799
800         * DumpRenderTree/config.h: Switch to PLATFORM(WIN_CAIRO)
801         * DumpRenderTree/win/DumpRenderTreeCairo.vsprops: Remove redundant
802           define of WIN_CAIRO=1.
803         * DumpRenderTree/win/DumpRenderTreeDebugCairoCFLite.vsprops: Add
804           WinCairo.vsprops to set (to match Release build).
805
806 2011-07-13  Dimitri Glazkov  <dglazkov@chromium.org>
807
808         Eliminate TestExpectationsFile.
809         https://bugs.webkit.org/show_bug.cgi?id=64458
810
811         Turns out, we can just use a Python list.
812
813         Reviewed by Adam Barth.
814
815         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Folded TestExpectationsFile.append into TestExpectationParser.parse_list,
816             removed TestExpectationsFile.
817         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: Moved tests to better reflect new names, removed iterator test,
818             since there's no more custom iterator machinery.
819
820 2011-07-13  Philippe Normand  <pnormand@igalia.com>
821
822         Unreviewed, added my other email addresses.
823
824         * Scripts/webkitpy/common/config/committers.py:
825
826 2011-07-13  Adam Barth  <abarth@webkit.org>
827
828         Folks should only be listed once.
829
830         * Scripts/webkitpy/common/config/committers.py:
831
832 2011-07-13  Philippe Normand  <pnormand@igalia.com>
833
834         Unreviewed, adding myself as Reviewer.
835
836         * Scripts/webkitpy/common/config/committers.py:
837
838 2011-07-13  Adam Barth  <abarth@webkit.org>
839
840         Fix flickering bug introduced by my previous patch.  I forgot to change
841         the name of the class everywhere.
842
843         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
844
845 2011-07-12  Philippe Normand  <pnormand@igalia.com>
846
847         [GTK] media/media-blocked-by-willsendrequest.html fails
848         https://bugs.webkit.org/show_bug.cgi?id=63699
849
850         Reviewed by Martin Robinson.
851
852         * DumpRenderTree/gtk/DumpRenderTree.cpp:
853         (willSendRequestCallback): Abort the request if explicitely asked
854         by the LayoutTestController's willSendRequestReturnsNull() function.
855
856 2011-07-13  Adam Barth  <abarth@webkit.org>
857
858         Remove "Dismiss" button from garden-o-matic butterbar
859         https://bugs.webkit.org/show_bug.cgi?id=64443
860
861         Reviewed by Dimitri Glazkov.
862
863         We don't have any persistent butter bar messages yet, so the dismiss
864         button is premature (and fairly heavy, visually).
865
866         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
867         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
868
869 2011-07-13  Adam Barth  <abarth@webkit.org>
870
871         Refine garden-o-matic status pane
872         https://bugs.webkit.org/show_bug.cgi?id=64442
873
874         Reviewed by Dimitri Glazkov.
875
876         This patch makes two improvements to the status pane:
877
878         1) We only query the server for failure types that we're expecting.
879            This dramatically reduces the number of HTTP requests, making
880            loading the status pane faster.
881
882         2) The status pane now displays which test and which builder it is
883            showing results for.
884
885         * Scripts/webkitpy/tool/servers/data/gardeningserver/config.js:
886         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
887         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
888         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
889         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
890         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
891         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
892
893 2011-07-13  Adam Barth  <abarth@webkit.org>
894
895         garden-o-matic should show test results
896         https://bugs.webkit.org/show_bug.cgi?id=64440
897
898         Reviewed by Adam Roben.
899
900         This patch introduces the results detail pane, which appears at the
901         bottom of the window and contains test results from the bots.
902         Currently, you can activate the pane by mousing over one of the builder
903         names associated with a failing test.
904
905         This is just a first iteration of the UI.  There's no way to resize or
906         zoom in on elements of the details pane, and images likely aren't sized
907         correctly, but it's a place to start.
908
909         * Scripts/webkitpy/tool/servers/data/gardeningserver/config.js:
910             - Add constants for our data attributes so we don't typo them!
911         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
912             - Add DOM structure for the details pane.
913         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
914         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
915             - Bind events for showing and hiding the details pane.
916         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
917             - Reduce the number of result types that we query for to avoid
918               overfilling the details pane.  We'll probably need another
919               solution here in the long-term.  For example, we could use tabs
920               to pack more results into the pane.
921         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
922         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html:
923             - Now that we're using the config package during testing, we need
924               to include it in the testing HTML.
925         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
926         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
927             - Change the DOM structure of the results detail to make it fit
928               nicely in the details pane (rather than flowing freely in the
929               body, as it did before).
930
931 2011-07-13  Adam Roben  <aroben@apple.com>
932
933         Make TestFailures show every time a possibly-flaky test failed, but hide it by default
934
935         It's useful to be able to see every time a flaky test failed to see whether it failed the
936         same way every time. But doing so takes a lot of space, so the list of failures is now
937         collapsed by default and can be revealed using a disclosure triangle.
938
939         Fixes <http://webkit.org/b/64455> TestFailures page doesn't show as much information for
940         flaky tests as I would like, even though the page is already so long
941
942         Reviewed by Dimitri Glazkov.
943
944         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/FlakyLayoutTestDetector.js:
945         (FlakyLayoutTestDetector.prototype.allFailures): Replaced flakinessExamples with this
946         function. Now returns all failures for the given test.
947
948         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/FlakyLayoutTestDetector_unittests.js:
949         Added. This just contains some simple tests of the FlakyLayoutTestDetector class. We'll add
950         more over time.
951
952         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
953         (LayoutTestHistoryAnalyzer.prototype.start): Updated the documentation comment to reflect
954         that we no longer return passing builds for possibly-flaky tests.
955
956         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
957         (.existing-bugs-list, .suspect-revisions-list, .flakiness-examples-list): Make the list of
958         flakiness examples small, too, since it can get quite long.
959
960         (.disclosure-triangle):
961         (.expanded > .disclosure-triangle):
962         Simple styles for the disclosure triangle.
963
964         (.flakiness-examples-list): Collapse the list by default.
965
966         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
967         (ViewController.prototype._displayBuilder): Pass the total number of builds analyzed to
968         _domForPossiblyFlakyTests.
969         (ViewController.prototype._domForPossiblyFlakyTests): Put a disclosure triangle to the left
970         of each test name, and the number of failures to the right. When the disclosure triangle is
971         clicked for the first time, we build up the list of failures and expand the element. After
972         that we just collapse or expand the element on subsequent clicks.
973
974         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/run-unittests.html:
975         Pulled in new tests.
976
977 2011-07-13  Adam Roben  <aroben@apple.com>
978         
979         Teach TestFailures to understand NRWT's output when it exits early due to too many failures
980
981         Fixes <http://webkit.org/b/64456> TestFailures page reports way too many failures when NRWT
982         exits early
983
984         Reviewed by Dimitri Glazkov.
985
986         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
987         (Builder.prototype.getNumberOfFailingTests): Relaxed the "Exiting early" test not to require
988         it to be at the beginning of the line, since NRWT prints a bunch of junk earlier in the
989         line. Tightened up the regex that's used to parse the number of failing tests to require the
990         leading number to be followed by whitespace so that we won't parse the "2011" in
991         "2011-07-13" as a number of failures.
992
993         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder_unittests.js:
994         Added a test for the above.
995
996 2011-07-12  Dimitri Glazkov  <dglazkov@chromium.org>
997
998         Extract model-like TestExpectationLine and TestExpectationFile from TestExpectations.
999         https://bugs.webkit.org/show_bug.cgi?id=64386
1000
1001         This is the first step in converting TestExpectations to a real model.
1002         * TestExpectationsLine represents a line in the test_expectations.txt file, and
1003         * TestExpectationsFile represents the file, which is a collection of lines.
1004
1005         Reviewed by Adam Barth.
1006
1007         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
1008         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
1009         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
1010
1011 2011-07-13  Xan Lopez  <xlopez@igalia.com>
1012
1013         [GTK] Do not grab focus too early in DRT.
1014
1015         Reviewed by Gustavo Noronha.
1016
1017         It causes a layout to happen and a progress signal to be emitted
1018         since r90900, but at this point we don't have a
1019         LayoutTestController object and we'll eventually crash. Since we
1020         already grab focus at the beginning of runTest() this is
1021         redundant, so get rid of it to fix the crash.
1022
1023         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1024         (main): remove call to grab_focus
1025
1026 2011-07-13  Sheriff Bot  <webkit.review.bot@gmail.com>
1027
1028         Unreviewed, rolling out r90893 and r90894.
1029         http://trac.webkit.org/changeset/90893
1030         http://trac.webkit.org/changeset/90894
1031         https://bugs.webkit.org/show_bug.cgi?id=64441
1032
1033         NRWT still doesn't work on qt-mac platform (Requested by Ossy
1034         on #webkit).
1035
1036         * Scripts/run-webkit-tests:
1037         (useNewRunWebKitTests):
1038
1039 2011-07-13  Csaba Osztrogonác  <ossy@webkit.org>
1040
1041         [Qt] NRWT should pick up the right httpd config file
1042         https://bugs.webkit.org/show_bug.cgi?id=64086
1043
1044         * Scripts/run-webkit-tests: Enable NRWT on qt-mac platform after r90810.
1045         (useNewRunWebKitTests):
1046
1047 2011-07-12  Mark Rowe  <mrowe@apple.com>
1048
1049         Fix the 32-bit build.
1050
1051         * DumpRenderTree/mac/UIDelegate.mm:
1052         (-[UIDelegate webView:exceededApplicationCacheOriginQuotaForSecurityOrigin:totalSpaceNeeded:]):
1053         Cast the NSUInteger value to unsigned long to match the format specifier.
1054
1055 2011-07-12  Adam Barth  <abarth@webkit.org>
1056
1057         Improve garden-o-matic UI when the bots fail to report revision numbers
1058         https://bugs.webkit.org/show_bug.cgi?id=64427
1059
1060         Reviewed by Dimitri Glazkov.
1061
1062         At least Win (dbg)(2) seems to fail to report the SVN revision number
1063         in full_results.json.  This patch makes garden-o-matic more robust to
1064         missing revision numbers.
1065
1066         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
1067         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
1068         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
1069         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
1070         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
1071
1072 2011-07-12  Adam Barth  <abarth@webkit.org>
1073
1074         Adjust garden-o-matic layout to use fewer lines
1075         https://bugs.webkit.org/show_bug.cgi?id=64422
1076
1077         Reviewed by Ojan Vafai.
1078
1079         This patch effectively merges the test name line with the list of
1080         builders on which the test fails.  Each test failure now occupies two
1081         lines instead of three.
1082
1083         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
1084         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
1085         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
1086
1087 2011-07-12  Adam Barth  <abarth@webkit.org>
1088
1089         garden-o-matic should label tests failures of new tests as such
1090         https://bugs.webkit.org/show_bug.cgi?id=64421
1091
1092         Reviewed by Ojan Vafai.
1093
1094         Also, we shouldn't dim failures of new tests, even if we've only seen
1095         them once because they're likely to be real problems that need
1096         attention.
1097
1098         * Scripts/webkitpy/tool/servers/data/gardeningserver/base_unittests.js:
1099         * Scripts/webkitpy/tool/servers/data/gardeningserver/checkout.js: Added.
1100         * Scripts/webkitpy/tool/servers/data/gardeningserver/checkout_unittests.js: Added.
1101         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
1102         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
1103         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
1104         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
1105         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html:
1106         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
1107         * Scripts/webkitpy/tool/servers/gardeningserver.py:
1108
1109 2011-07-12  Adam Barth  <abarth@webkit.org>
1110
1111         garden-o-matic should display how many times we've seen a failure
1112         https://bugs.webkit.org/show_bug.cgi?id=64417
1113
1114         Reviewed by Ojan Vafai.
1115
1116         This patch adds some UI to display how many times we've seen a given
1117         failure, which can be helpful for determining whether that failure is a
1118         real failure or a flaky test.
1119
1120         When a failure has only been seen once (i.e., only a single run on a
1121         single bot), we set the opacity of to 50% to avoid distracting the
1122         gardener.
1123
1124         This patch also refactors the failure walker to have a simpler API
1125         internally by moving from an object-oriented paradigm to a functional
1126         paradigm.
1127
1128         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
1129         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
1130         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
1131         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
1132         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
1133         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
1134
1135 2011-07-12  Chris Rogers  <crogers@google.com>
1136
1137         Enable Web Audio for chromium DRT
1138         https://bugs.webkit.org/show_bug.cgi?id=64409
1139
1140         Reviewed by James Robinson.
1141
1142         * DumpRenderTree/chromium/TestShell.cpp:
1143         (TestShell::TestShell):
1144
1145 2011-07-12  Adam Barth  <abarth@webkit.org>
1146
1147         Tweak some UI in garden-o-matic now that the tree actually has a
1148         failure and I can see what this all looks like.
1149
1150         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
1151         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
1152         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
1153
1154 2011-07-12  Joseph Pecoraro  <joepeck@webkit.org>
1155
1156         Unreviewed. Skipping a few tests which fail due to differing output
1157         in recent ApplicationCache quota tests. Reenabling the tests is
1158         tracked by <http://webkit.org/b/64410>.
1159
1160         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1161         (WebCore::DumpRenderTree::dumpApplicationCacheQuota): This should
1162         have been a space, not a newline.
1163
1164 2011-07-12  Adam Barth  <abarth@webkit.org>
1165
1166         garden-o-matic should display regression ranges
1167         https://bugs.webkit.org/show_bug.cgi?id=64407
1168
1169         Reviewed by Dimitri Glazkov.
1170
1171         This patch computes a regression range for a failure by intersecting
1172         the regression ranges seen by the various bots.  We make the underlying
1173         assumption that a test is only failing due to one revision at any given
1174         moment.  If that's not true, this code probably explodes.
1175
1176         The regression ranges appear asynchronously, which might be a jarring
1177         UI.  We'll have to experiment to see.
1178
1179         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
1180         * Scripts/webkitpy/tool/servers/data/gardeningserver/base_unittests.js:
1181         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
1182         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
1183         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
1184         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
1185         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
1186
1187 2011-07-12  Adam Barth  <abarth@webkit.org>
1188
1189         cr-linux-ews should run pixel tests
1190         https://bugs.webkit.org/show_bug.cgi?id=64394
1191
1192         Reviewed by Ojan Vafai.
1193
1194         This patch enabled pixel test on the cr-linux-ews.  Running a quick
1195         test on the bots, it looks like we have five pixel failures, related to
1196         some missing fonts.  I've also made the bots skip failing tests to
1197         avoid uploading huge zip files to bugs.webkit.org (with all the exected
1198         IMAGE failures).
1199
1200         * Scripts/webkitpy/common/config/ports.py:
1201         * Scripts/webkitpy/common/config/ports_unittest.py:
1202
1203 2011-07-12  Joseph Pecoraro  <joepeck@webkit.org>
1204
1205         ApplicationCache update should not immediately fail when reaching per-origin quota
1206         https://bugs.webkit.org/show_bug.cgi?id=64177
1207
1208         Reviewed by Alexey Proskuryakov.
1209
1210         Add layoutTestController.disallowIncreaseForApplicationCacheQuota()
1211         to disable the default behavior of raising the default per-origin
1212         quota to 5MB when reached. This allows us to test what happens
1213         when the quota is reached and not increased.
1214
1215         * DumpRenderTree/LayoutTestController.cpp:
1216         (LayoutTestController::LayoutTestController):
1217         (disallowIncreaseForApplicationCacheQuotaCallback):
1218         (LayoutTestController::staticFunctions):
1219         * DumpRenderTree/LayoutTestController.h:
1220         (LayoutTestController::disallowIncreaseForApplicationCacheQuota):
1221         (LayoutTestController::setDisallowIncreaseForApplicationCacheQuota):
1222         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1223         (LayoutTestController::reset):
1224         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1225         (LayoutTestController::disallowIncreaseForApplicationCacheQuota):
1226         (LayoutTestController::dumpApplicationCacheDelegateCallbacks):
1227         Expose a function to prevent automatically increasing the per-origin
1228         quota when the quota is reached. This allows us to test what happens
1229         when a user would "disallow" a quota increase.
1230
1231         * DumpRenderTree/mac/UIDelegate.mm:
1232         (-[UIDelegate webView:exceededApplicationCacheOriginQuotaForSecurityOrigin:spaceNeeded:]):
1233         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1234         (WebCore::DumpRenderTree::dumpApplicationCacheQuota):
1235         Respect the new disallowIncrease flag and dump the spaceNeeded values
1236         for ports that implement application cache quotas.
1237
1238 2011-07-12  Ojan Vafai  <ojan@chromium.org>
1239
1240         fix flakiness dashboard to work with new crash log filenames
1241         https://bugs.webkit.org/show_bug.cgi?id=64393
1242
1243         Reviewed by Adam Barth.
1244
1245         * TestResultServer/static-dashboards/flakiness_dashboard.html:
1246
1247 2011-07-12  Adam Barth  <abarth@webkit.org>
1248
1249         Add a happy animation to garden-o-matic when there are no failures
1250         https://bugs.webkit.org/show_bug.cgi?id=64382
1251
1252         Reviewed by Ojan Vafai.
1253
1254         As requested by Ojan.
1255
1256         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
1257         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
1258         * Scripts/webkitpy/tool/servers/data/gardeningserver/partytime.gif: Added.
1259         * Scripts/webkitpy/tool/servers/gardeningserver.py:
1260
1261 2011-07-12  Adam Barth  <abarth@webkit.org>
1262
1263         sheriffbot can't perform rollouts
1264         https://bugs.webkit.org/show_bug.cgi?id=64370
1265
1266         Reviewed by Eric Seidel.
1267
1268         Now that we're not updating the working copy before processing every
1269         IRC message, the rollout command needs to update it manually.
1270
1271         * Scripts/webkitpy/tool/bot/irc_command.py:
1272         * Scripts/webkitpy/tool/bot/irc_command_unittest.py:
1273
1274 2011-07-12  Adam Barth  <abarth@webkit.org>
1275
1276         Fix the grammar in the FIXME comment.
1277
1278         * Scripts/webkitpy/common/net/resultsjsonparser.py:
1279
1280 2011-07-12  Eric Seidel  <eric@webkit.org>
1281
1282         NRWT should open test results page with Safari trunk, not the system provided one on Mac
1283         https://bugs.webkit.org/show_bug.cgi?id=64346
1284
1285         Reviewed by Adam Barth.
1286
1287         To fix this I implemented Port.show_results_html_file in Mac, Gtk and Qt ports with
1288         implementations (mostly) matching those found in old-run-webkit-tests.
1289         There are still some minor differences for Qt which Qt hackers may wish to tweak.
1290
1291         I had to add a WebKitPort._port_flag_for_scripts method (similar to flag() in common.config.ports.py)
1292         for the Qt/Gtk ports which always require a flag passed to scripts.
1293
1294         While trying to test this, I found that FactoryTest.test_chromium_gpu_linux
1295         was using a real filesystem (due to assert_platform) which due to
1296         global variables in config.py was causing set-webkit-configuration to have
1297         an affect on unit test results!  So I fixed this by making FactoryTest.assert_port
1298         pass a mock file system whenever calling factory.get.
1299
1300         Unfortunately TestPort was depending on always being passed a None filesystem
1301         and asserting filesystem._tests (only true for unit_test_filesystem()).
1302         So I just removed the FactoryTest.test_test and FactoryTest.test_dryrun tests
1303         deciding that they were pretty much useless anyway.  If others feel strongly
1304         I'm happy to fix this in a different way.
1305
1306         * Scripts/webkitpy/common/system/executive.py:
1307          - default arguments in python are screwy.  They use a single shared
1308            instance, so it's better to use argument=None and then argument = argument or Default()
1309            if you have any chance of mutating (or returning) the default argument.
1310         * Scripts/webkitpy/layout_tests/port/config.py:
1311          - This code is wrong.  We don't need to use a global variable here (as far as I can tell).
1312            I'm not fixing it in this patch, but I've marked it with a FIXME and we can convert to
1313            storing the results of the read on the Config object (which should only be created once during normal operation).
1314            Unit tests shouldn't be hitting the disk anyway.  It's possible Config should move off of Port and onto Tool/Host directly.
1315         * Scripts/webkitpy/layout_tests/port/factory.py:
1316         * Scripts/webkitpy/layout_tests/port/factory_unittest.py:
1317         * Scripts/webkitpy/layout_tests/port/gtk.py:
1318         * Scripts/webkitpy/layout_tests/port/gtk_unittest.py: Added.
1319         * Scripts/webkitpy/layout_tests/port/mac.py:
1320         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1321         * Scripts/webkitpy/layout_tests/port/qt.py:
1322         * Scripts/webkitpy/layout_tests/port/qt_unittest.py:
1323         * Scripts/webkitpy/layout_tests/port/webkit.py:
1324
1325 2011-07-12  Adam Roben  <aroben@apple.com>
1326
1327         Teach TestFailures to recognize when run-webkit-tests gets killed by buildbot
1328
1329         Fixes <http://webkit.org/b/64358> TestFailures page thinks all tests passed in
1330         http://build.webkit.org/builders/Windows%207%20Release%20(Tests)/builds/14672
1331
1332         Reviewed by Daniel Bates.
1333
1334         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
1335         (Builder.prototype.getNumberOfFailingTests): If run-webkit-tests exited with a non-zero
1336         exit status but we didn't find any failure counts, assume that there was some error that
1337         caused run-webkit-tests to die early (like being killed by buildbot due to a timeout).
1338
1339         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder_unittests.js:
1340         Added a new test that shows that we get a failingTestCount of -1 when run-webkit-tests dies
1341         early.
1342         (runGetNumberOfFailingTestsTest): Moved most code here from the only pre-existing test in this
1343         file.
1344
1345         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
1346         (LayoutTestResultsLoader.prototype.start): Bump the cache number so old cached data that was
1347         tainted by the bug fixed in this patch will be evicted.
1348
1349 2011-07-12  Adam Barth  <abarth@webkit.org>
1350
1351         cr-linux-ews complains about tests that aren't actually failing
1352         https://bugs.webkit.org/show_bug.cgi?id=64367
1353
1354         The underlying problem here is that full_results.json doesn't have
1355         enough information to determine whether a given test result was
1356         expected because whether an actual result is expected depends on
1357         whether full_results.json was generated duing a test run that included
1358         pixel tests.
1359
1360         The right long-term solution is to make full_results.json a complete
1361         description of what happened durning a test run.  In this patch
1362         however, to stop the spam, this patch makes the jsonresultsparser
1363         ignore pixel failures.
1364
1365         (I'm landing this patch unreviewed in order to stop the bots from
1366         spamming.  I'll happily address any review comments in a follow-up
1367         patch.)
1368
1369         * Scripts/webkitpy/common/net/resultsjsonparser.py:
1370         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
1371
1372 2011-07-12  Adam Barth  <abarth@webkit.org>
1373
1374         Re-work garden-o-matic UI to begin to look like the mocks
1375         https://bugs.webkit.org/show_bug.cgi?id=64334
1376
1377         Reviewed by Dimitri Glazkov.
1378
1379         This patch changes how we display failures to be more like the mock the
1380         dglazkov made. We now can display the same information much more
1381         compactly.
1382
1383         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
1384         * Scripts/webkitpy/tool/servers/data/gardeningserver/base_unittests.js:
1385         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
1386         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css: Added.
1387         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
1388         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
1389         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
1390         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
1391         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
1392         * Scripts/webkitpy/tool/servers/gardeningserver.py:
1393
1394 2011-07-11  Adam Roben  <aroben@apple.com>
1395
1396         Extract TestFailures's bug-filing code into two new classes
1397
1398         Fixes <http://webkit.org/b/64300> TestFailures page's new-bug-filing code is a mess!
1399
1400         Reviewed by Darin Adler and Adam Barth.
1401
1402         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Buildbot.js:
1403         (Buildbot.prototype.resultsDirectoryURL): Changed to return a URI-encoded URL. Otherwise the
1404         URL isn't valid (and it's harder to mock this function).
1405
1406         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Buildbot_unittests.js:
1407         Added. Just tests the above change (for now).
1408
1409         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder_unittests.js:
1410         Added a license header, enclosed everything in a closure to avoid polluting the global
1411         namespace, changed the test name to actually describe the passing condition, and replaced
1412         uses of equals() with equal(). (The latter is the actual name of the function, and matches
1413         deepEqual, etc.)
1414
1415         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/NewBugForm.js: Added.
1416         (NewBugForm): This class knows how to construct a <form> used for filing new bugs in
1417         Bugzilla based on some parameters.
1418         (NewBugForm.prototype.domElement): Creates and returns the <form> element. Code came from
1419         ViewController.prototype._domForNewAndExistingBugs.
1420
1421         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/NewBugForm_unittests.js:
1422         Added. Tests the above code.
1423
1424         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailureBugForm.js: Added.
1425         (TestFailureBugForm): This class knows how to construct a <form> element used for filing new
1426         bugs specifically about test failures. Code came from
1427         ViewController.prototype._domForNewAndExistingBugs.
1428         (TestFailureBugForm.prototype.domElement): Slightly customizes the <form> element returned
1429         by NewBugForm.
1430         (TestFailureBugForm.prototype._computeOperatingSystem):
1431         (TestFailureBugForm.prototype._computePlatform):
1432         (TestFailureBugForm.prototype._createBugTitle):
1433         (TestFailureBugForm.prototype._failingResultsHTMLURL):
1434         (TestFailureBugForm.prototype._failingRevision):
1435         (TestFailureBugForm.prototype._passingRevision):
1436         (TestFailureBugForm.prototype._regressionRangeString):
1437         Code came from ViewController.prototype._domForNewAndExistingBugs. I broke it out into
1438         separate functions to break up the rat's nest a bit.
1439
1440         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailureBugForm_unittests.js:
1441         Added. Tests the above code.
1442
1443         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1444         (ViewController.prototype._domForNewAndExistingBugs): Moved code from here to
1445         TestFailureBugForm/NewBugForm, and changed this code to use a TestFailureBugForm.
1446
1447         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/index.html: Pull in
1448         NewBugForm/TestFailureBugForm.js.
1449
1450         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/run-unittests.html:
1451         Added new tests and required files.
1452
1453 2011-07-12  Adam Roben  <aroben@apple.com>
1454
1455         Test that no intermediate WTF::Strings are created when concatenating with string literals
1456
1457         Test for <http://webkit.org/b/63330> Concatenating string literals and WTF::Strings using
1458         operator+ is suboptimal
1459
1460         Reviewed by Darin Adler.
1461
1462         * TestWebKitAPI/Tests/WTF/StringOperators.cpp: Added.
1463         (TestWebKitAPI::TEST): Test that a bunch of different string concatenation expressions don't
1464         create any intermediate WTF::Strings while they're being evaluated.
1465
1466         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
1467         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
1468         Added new file.
1469
1470 2011-07-12  Eric Seidel  <eric@webkit.org>
1471
1472         [Qt] NRWT should pick up the right httpd config file
1473         https://bugs.webkit.org/show_bug.cgi?id=64086
1474
1475         Reviewed by Adam Barth.
1476
1477         This is more fall-out from the Port class inappropriately encapsulating
1478         both platform and port knowledge.  We need to split out some of this
1479         platform knowledge into a separate class which can be better shared
1480         between ports.
1481
1482         The fix was to move all the _path_to_apache_config_file logic down
1483         into the WebKitPort baseclass so that all of the ports can find the
1484         right apache config file, regardless of what platform they may be running on.
1485
1486         I did not move this down into base.Port (even though I considered it).
1487         Chromium duplicates some of this logic, but since they have separate
1488         subclasses for each port-OS combination (e.g. ChromiumMac, ChromiumWin)
1489         they wouldn't notice the change.  Eventually we'll move this logic
1490         out of Port entirely, and then it will be shared by all ports.
1491
1492         I also cleaned up the http-config logic for ORWT while I was there,
1493         although since we're killing that code, I'm happy to revert that part
1494         of the change if changing it is viewed as needlessly risky.
1495
1496         During this cleanup, I noticed that no ports use the "default" httpd.conf
1497         which is left over from Mac OS X Tiger (and old linux distros) which used
1498         Apache 1.3.  I've removed httpd.conf (and associated support in ORWT) since
1499         we no longer support any ports which use this httpd.conf.
1500
1501         * Scripts/webkitperl/httpd.pm:
1502         (getHTTPDConfigPathForTestDirectory):
1503         * Scripts/webkitpy/layout_tests/port/base.py:
1504         * Scripts/webkitpy/layout_tests/port/gtk.py:
1505         * Scripts/webkitpy/layout_tests/port/mac.py:
1506         * Scripts/webkitpy/layout_tests/port/qt.py:
1507         * Scripts/webkitpy/layout_tests/port/webkit.py:
1508         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
1509         * Scripts/webkitpy/layout_tests/port/win.py:
1510         * Scripts/webkitpy/layout_tests/servers/apache_http_server.py:
1511
1512 2011-07-11  Hironori Bono  <hbono@chromium.org>
1513
1514         Reviewed by Adam Roben.
1515
1516         Implement layoutTestController.setTextDirection for WebKit2 and Windows.
1517         https://bugs.webkit.org/show_bug.cgi?id=61931
1518
1519         This change implements layoutTextController.setTextDirection for WebKit2
1520         and Windows so we can run a layout test added by r87770 on them. Each
1521         implementation adds a wrapper function for Editor::setBaseWritingDirection()
1522         so LayoutTestController can call it.
1523
1524         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1525         (LayoutTestController::setTextDirection): Implemented the binding function
1526         for layoutTestController.setTextDirection (Windows).
1527         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
1528         Added a binding function for layoutTestController.setTextDirection (WebKit2).
1529         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: ditto.
1530         (WTR::LayoutTestController::setTextDirection):
1531         * WebKitTestRunner/InjectedBundle/LayoutTestController.h: ditto.
1532
1533 2011-07-11  Raphael Kubo da Costa  <kubo@profusion.mobi>
1534
1535         [EFL] ImageDiff: Make sure gEcoreEvas is destroyed at the right time.
1536         https://bugs.webkit.org/show_bug.cgi?id=64293
1537
1538         Reviewed by Kent Tamura.
1539
1540         gEcoreEvas was being destroyed after shutdownEfl() was called, so the
1541         call to ecore_evas_free() failed.
1542
1543         We now shut it down manually before shutdownEfl() is called.
1544
1545         * DumpRenderTree/efl/ImageDiff.cpp:
1546         (main): Destroy gEcoreEvas before shutting down the EFL.
1547
1548 2011-07-11  Ryosuke Niwa  <rniwa@webkit.org>
1549
1550         Make Chromium Mac and Linux tests bot core builders
1551         https://bugs.webkit.org/show_bug.cgi?id=63196
1552
1553         Reviewed by Adam Barth.
1554
1555         Added Chromium Mac and Linux Release Tests bots core builders.
1556
1557         Also rearranged the order of bots so that test bots show up right next to
1558         respective builders.
1559
1560         * BuildSlaveSupport/build.webkit.org-config/config.json:
1561         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
1562         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1563
1564 2011-07-11  Raphael Kubo da Costa  <kubo@profusion.mobi>
1565
1566         [EFL] DRT: Add code related to font management.
1567         https://bugs.webkit.org/show_bug.cgi?id=63989
1568
1569         Reviewed by Kent Tamura.
1570
1571         The added files are responsible for managing fontconfig and adding the
1572         fonts required by some layout tests to fontconfig's path.
1573
1574         * DumpRenderTree/efl/FontManagement.cpp: Added.
1575         (getFontDirectories):
1576         (getFontFiles):
1577         (addFontDirectories):
1578         (addFontFiles):
1579         (addFontsToEnvironment):
1580         * DumpRenderTree/efl/FontManagement.h: Added.
1581
1582 2011-07-08  Dirk Pranke  <dpranke@chromium.org>
1583
1584         nrwt: linting fixes
1585         https://bugs.webkit.org/show_bug.cgi?id=64225
1586
1587         Reviewed by Eric Siedel.
1588
1589         Miscellaneous linting fixes. The most notable change is that
1590         we add public attributes for user, executive, filesystem, and
1591         options on the Port object, so we don't have to refer to the
1592         "protected" versions all over the place".
1593
1594         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1595         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
1596         * Scripts/webkitpy/layout_tests/controllers/message_broker.py:
1597         * Scripts/webkitpy/layout_tests/controllers/message_broker_unittest.py:
1598         * Scripts/webkitpy/layout_tests/controllers/worker.py:
1599         * Scripts/webkitpy/layout_tests/port/base.py:
1600         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1601
1602 2011-07-11  Ryosuke Niwa  <rniwa@webkit.org>
1603
1604         webkit-patch roll-chromium-deps no longer works
1605         https://bugs.webkit.org/show_bug.cgi?id=64324
1606
1607         Reviewed by James Robinson.
1608
1609         Lower the minimum commit message length to 10.
1610
1611         * Scripts/webkitpy/tool/steps/commit.py:
1612
1613 2011-07-11  Alice Boxhall  <aboxhall@chromium.org>
1614
1615         Convert json_results_generator.py to output version 4 JSON.
1616         https://bugs.webkit.org/show_bug.cgi?id=60869
1617
1618         Reviewed by Ojan Vafai.
1619
1620         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
1621         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1622         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
1623
1624 2011-07-11  Martin Robinson  <mrobinson@igalia.com>
1625
1626         [GTK] [NRWT] Pixel tests do not work
1627         https://bugs.webkit.org/show_bug.cgi?id=64091
1628
1629         Reviewed by Eric Seidel.
1630
1631         Fix pixel test runs for GTK+ new-run-webkit-tests by exposing the
1632         path to the ImageDiff binary.
1633
1634         * Scripts/webkitpy/layout_tests/port/gtk.py: Implement _path_to_image_diff.
1635
1636 2011-07-11  Dirk Pranke  <dpranke@chromium.org>
1637
1638         Fix hang on win32 in a multiprocessing test that shouldn't have been running.
1639
1640         Unreviewed, build fix.
1641
1642         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1643
1644 2011-07-11  Adam Roben  <aroben@apple.com>
1645
1646         Don't count leaks as test failures on TestFailures's front page
1647
1648         As a bonus, this patch adds our first unit test.
1649
1650         Fixes <http://webkit.org/b/64303> REGRESSION (r90489): TestFailures page says far more tests
1651         are failing on the Leaks bot than actually are
1652
1653         Reviewed by Adam Barth.
1654
1655         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
1656         (Builder.prototype.getNumberOfFailingTests): Exclude lines containing "leak" when summing
1657         failure counts.
1658
1659         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder_unittests.js:
1660         Added. Contains a simple test to show that leaks aren't counted as test failures.
1661
1662         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/run-unittests.html:
1663         Added. This is the test harness.
1664
1665 2011-06-14  Mike Stegeman  <mrstegeman@gmail.com>
1666
1667         [GTK] Add API to allow setting local storage database path
1668         https://bugs.webkit.org/show_bug.cgi?id=62091
1669
1670         Reviewed by Martin Robinson and Gustavo Noronha.
1671
1672         Expose the path of the localStorage databases through a setting
1673         to allow HTML5 localStorage to be persistent. New setting is:
1674         html5-local-storage-database-path
1675
1676         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1677         (resetDefaultsToConsistentValues):
1678
1679 2011-07-11  Adam Barth  <abarth@webkit.org>
1680
1681         garden-o-matic should be able to rebaseline tests
1682         https://bugs.webkit.org/show_bug.cgi?id=64186
1683
1684         Reviewed by Eric Seidel.
1685
1686         This patch contains a simple HTTP server binding for the
1687         rebaseline-test command.  The frontend will appear in another patch.
1688
1689         * Scripts/webkitpy/tool/servers/gardeningserver.py:
1690         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py:
1691         * Scripts/webkitpy/tool/servers/reflectionhandler.py:
1692
1693 2011-07-11  Adam Barth  <abarth@webkit.org>
1694
1695         Add some basic UI for showing regression ranges
1696         https://bugs.webkit.org/show_bug.cgi?id=64243
1697
1698         Reviewed by Eric Seidel.
1699
1700         This UI is all temporary.  It exists mostly just to have some buttons
1701         to click to exercise the code.
1702
1703         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
1704         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
1705         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
1706         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
1707         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
1708
1709 2011-07-11  Adam Barth  <abarth@webkit.org>
1710
1711         garden-o-matic should be able to determine which revisions caused a given failure
1712         https://bugs.webkit.org/show_bug.cgi?id=64189
1713
1714         Reviewed by Adam Roben.
1715
1716         Walking the failure history looking for failures turns out to be
1717         slightly tricky because the network requests are asynchronous.
1718         Currently we do all the fetches serially and our cache is unbounded.
1719         We'll probably optimize both those parameters eventually.
1720
1721         This patch also generalizes some functionality in the unit testing
1722         framework to make testing this sort of code easier.
1723
1724         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
1725         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
1726         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
1727
1728 2011-07-11  Adam Barth  <abarth@webkit.org>
1729
1730         Add a webkit-patch command for rebaselining an individual test
1731         https://bugs.webkit.org/show_bug.cgi?id=64246
1732
1733         Reviewed by Eric Seidel.
1734
1735         This patch introduces a command that's able to rebaseline a single
1736         test.  Currently, the command works only with the build.chromium.org
1737         buildbots, but extending it to work with the build.webkit.org bots
1738         shouldn't be that hard.
1739
1740         A complete rebaseling tool should also include an "optimize baselines"
1741         command (which moves/deletes baselines in order to reduce the number of
1742         expected results files), but that will come in a future patch.
1743
1744         Really BuilderToPort should be merged into builders.py, but I'm going
1745         to save that for a future patch as well.  (We need to stop shaving yaks
1746         at some point.)
1747
1748         * Scripts/webkitpy/tool/commands/rebaseline.py:
1749
1750 2011-07-11  Adam Barth  <abarth@webkit.org>
1751
1752         garden-o-matic should be able to roll out patches
1753         https://bugs.webkit.org/show_bug.cgi?id=64185
1754
1755         Reviewed by Eric Seidel.
1756
1757         This gardening server API is simply a binding to the rollout machinery
1758         already present in webkit-patch.
1759
1760         * Scripts/webkitpy/tool/commands/gardenomatic.py:
1761         * Scripts/webkitpy/tool/servers/gardeningserver.py:
1762         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py: Added.
1763
1764 2011-07-11  Csaba Osztrogonác  <ossy@webkit.org>
1765
1766         [Qt][Mac] Unreviewed fix after r90746.
1767
1768         * Scripts/run-webkit-tests:
1769         (useNewRunWebKitTests): Really disable NRWT for qt-mac platform.
1770
1771 2011-07-11  Adam Roben  <aroben@apple.com>
1772
1773         Teach TestFailures that ORWT's results.html file might be missing due to all tests passing
1774
1775         There are three reasons why we might fail to fetch ORWT's results.html:
1776           - All tests passed, so no results.html was generated
1777           - Some error during the test run caused results.html not to be generated (e.g., ORWT
1778             timed out)
1779           - Some network error occurred when fetching results.html
1780
1781         We were failing to account for the first possibility in some cases. For test runs before
1782         r89610, we first check build.webkit.org/json to determine how many tests failed and whether
1783         ORWT exited early due to too many failures; if all tests passed then we don't fetch
1784         results.html at all. r89610 changed ORWT to put information in results.html about exiting
1785         early due to too many failures, so we no longer needed to check build.webkit.org/json to get
1786         that information, and in r89619 I changed TestFailures to do just that. But I forgot that we
1787         still needed to check build.webkit.org/json to find out if all tests passed!
1788
1789         Now, for test runs after r89610, we check results.html first, and then check
1790         build.webkit.org/json if we fail to fetch results.html. This lets us distinguish between all
1791         tests passing and the error cases.
1792
1793         Fixes <http://webkit.org/b/64280> TestFailures page can't pinpoint that r90699 caused 13
1794         tests to fail on Windows 7 Release (WebKit2 Tests)
1795
1796         Reviewed by David Kilzer.
1797
1798         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
1799         (LayoutTestResultsLoader.prototype.start): Bumped the cache version so that old, buggy
1800         cached data will get evicted. We were marking builds where all tests passed as errors!
1801         (LayoutTestResultsLoader.prototype._fetchAndParseORWTResults): Added success/error callback
1802         parameters to the fetchAndParseResultsHTML helper function, and added a similar
1803         fetchNumberOfFailingTests function that fetches data from build.webkit.org/json (code came
1804         from later in the function). For test runs before r89610, we first check
1805         build.webkit.org/json then check results.html. For builds after r89610, we first check
1806         results.html and then check build.webkit.org/json if we couldn't fetch results.html.
1807
1808 2011-07-11  Csaba Osztrogonác  <ossy@webkit.org>
1809
1810         new-run-webkit-tests does not support qt-arm or qt-4.8 results
1811         https://bugs.webkit.org/show_bug.cgi?id=64071
1812
1813         [Qt] NRWT should pick up the right httpd config file
1814         https://bugs.webkit.org/show_bug.cgi?id=64086
1815
1816         Reviewed by Zoltán Herczeg.
1817
1818         * Scripts/run-webkit-tests: Disable NRWT for qt-mac, qt-arm and qt-4.8 platforms until fix.
1819         (useNewRunWebKitTests):
1820
1821 2011-07-11  Kenichi Ishibashi  <bashi@chromium.org>
1822
1823         Add TestNetscapePlugIn/Tests/FormValue.cpp to DRT build files
1824         https://bugs.webkit.org/show_bug.cgi?id=64248
1825
1826         Reviewed by Kent Tamura.
1827
1828         * DumpRenderTree/DumpRenderTree.gypi: Added FormValue.cpp.
1829         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Ditto.
1830         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro: Ditto.
1831         * GNUmakefile.am: Ditto.
1832
1833 2011-07-10  Kenichi Ishibashi  <bashi@chromium.org>
1834
1835         Let plugins participate in form submission
1836         https://bugs.webkit.org/show_bug.cgi?id=13061
1837
1838         Adds NPPVformValue support to the testing plugin.
1839
1840         Reviewed by Anders Carlsson.
1841
1842         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Added FormValue.cpp
1843         * DumpRenderTree/TestNetScapePlugIn/Tests/FormValue.cpp: Added.
1844         (FormValue::FormValue): Ditto.
1845         (FormValue::NPP_GetValue): Ditto.
1846
1847 2011-07-11  Roland Steiner  <rolandsteiner@chromium.org>
1848
1849         Reviewed by Tony Chang.
1850
1851         Complete functions in filesystem.py
1852         https://bugs.webkit.org/show_bug.cgi?id=63528
1853
1854         - missing text file functions added
1855         - functions sorted
1856         - removed 'append' optional parameters (were unused)
1857         - adapted filesystem_mock in the same way
1858
1859         * Scripts/webkitpy/common/system/filesystem.py:
1860         * Scripts/webkitpy/common/system/filesystem_mock.py:
1861         * Scripts/webkitpy/common/system/filesystem_unittest.py:
1862
1863 2011-07-10  Adam Barth  <abarth@webkit.org>
1864
1865         Clean up style in fallback path calculation
1866         https://bugs.webkit.org/show_bug.cgi?id=64239
1867
1868         Reviewed by Daniel Bates.
1869
1870         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1871         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1872         * Scripts/webkitpy/layout_tests/port/mac.py:
1873         * Scripts/webkitpy/layout_tests/port/win.py:
1874
1875 2011-07-10  Balazs Kelemen  <kbalazs@webkit.org>
1876
1877         [Qt][WK2] Implement the rest of WTR::PlatformWebView
1878         https://bugs.webkit.org/show_bug.cgi?id=63630
1879
1880         Reviewed by Kenneth Rohde Christiansen.
1881
1882         * WebKitTestRunner/qt/PlatformWebViewQt.cpp:
1883         (WTR::PlatformWebView::windowFrame):
1884         (WTR::PlatformWebView::setWindowFrame):
1885
1886 2011-07-10  Mark Rowe  <mrowe@apple.com>
1887
1888         Teach build-webkit how to find the latest WebKitSystemInterface binary.
1889
1890         * Scripts/build-webkit:
1891
1892 2011-07-09  Adam Roben  <aroben@apple.com>
1893
1894         Teach TestFailures to abbreviate the examples of test flakiness
1895
1896         These lists can get quite long, and it's not really helpful in most cases to have soooooo
1897         many examples of flakiness.
1898
1899         Fixes <http://webkit.org/b/64203> Lists of flaky revisions on TestFailures page can get so
1900         long they're hard to navigate
1901
1902         Reviewed by Dan Bates.
1903
1904         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/FlakyLayoutTestDetector.js:
1905         (FlakyLayoutTestDetector.prototype.flakinessExamples): If we have more than a certain number
1906         of examples, replace the middle items with a separator.
1907
1908         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
1909         (.flakiness-example-separator): Added styles for the separator.
1910
1911         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1912         (ViewController.prototype._domForPossiblyFlakyTests): Use a vertical ellipsis to represent
1913         the separator.
1914
1915 2011-07-09  Dirk Pranke  <dpranke@chromium.org>
1916
1917         nrwt: stack traces from worker-side exceptions aren't very useful inside test-webkitpy
1918         https://bugs.webkit.org/show_bug.cgi?id=64218
1919
1920         Reviewed by Eric Seidel.
1921
1922         Exceptions aren't picklable and can't be sent across the
1923         manager/worker message queue without losing information. NRWT
1924         handles this by turning the stack trace into a set of strings,
1925         and logging the strings when we receive an exception from the
1926         worker. However, when you are running tests and something
1927         crashes on the worker side, test-webkitpy prints the
1928         manager-side stack trace, which is just confusing and useless.
1929
1930         This patch changes the logic so that exceptions are passed
1931         through as-is when the worker and manager are in the same
1932         process (the --worker-model=inline option). This increases the
1933         code paths slightly but makes crashes much more useful.
1934
1935         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1936         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
1937         * Scripts/webkitpy/layout_tests/controllers/message_broker.py:
1938         * Scripts/webkitpy/layout_tests/controllers/worker.py:
1939         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1940
1941 2011-07-09  Darin Fisher  <darin@chromium.org>
1942
1943         Eliminate bad dependency on gfx::Point.
1944         https://bugs.webkit.org/show_bug.cgi?id=64228
1945
1946         Reviewed by Kent Tamura.
1947
1948         * DumpRenderTree/chromium/EventSender.cpp:
1949         (initMouseEvent):
1950
1951 2011-07-08  Ojan Vafai  <ojan@chromium.org>
1952
1953         pull static dashboard files into the appengine server from the chromium repository
1954         https://bugs.webkit.org/show_bug.cgi?id=64208
1955
1956         Reviewed by Dirk Pranke.
1957
1958         These files belong in the WebKit repo since they are tied primarily to the webkit tests.
1959         They have some extra bits to support chromium's gtests, but that seems fine.
1960         Mainly, this will allow other WebKit hackers to hack on the dashboards.
1961
1962         As a nice side-effect, we can now push the dashboard files when we do appengine pushes
1963         instead of the weird thing we used to do of pulling them from the Chromium repository
1964         and storing them in the appengine datastore. This allows for cleaning up a lot of code
1965         and will likely make the dashboards load a bit faster.
1966
1967         The new JS files don't fully match WebKit style, but I'd like to clean that up in a
1968         followup patch if possible to maintain my sanity with this patch.
1969
1970         * TestResultServer/app.yaml:
1971         * TestResultServer/handlers/dashboardhandler.py: Removed.
1972         * TestResultServer/handlers/menu.py:
1973         * TestResultServer/main.py:
1974         * TestResultServer/model/dashboardfile.py: Removed.
1975         * TestResultServer/static-dashboards/LICENSE.dygraph.txt: Added.
1976         * TestResultServer/static-dashboards/README.dygraph.txt: Added.
1977         * TestResultServer/static-dashboards/README.webtreemap.txt: Added.
1978         * TestResultServer/static-dashboards/aggregate_results.html: Added.
1979         * TestResultServer/static-dashboards/builders.js: Added.
1980         * TestResultServer/static-dashboards/dashboard_base.js: Added.
1981         * TestResultServer/static-dashboards/dygraph-combined.js: Added.
1982         * TestResultServer/static-dashboards/flakiness_dashboard.html: Added.
1983         * TestResultServer/static-dashboards/flakiness_dashboard_tests.js: Added.
1984         * TestResultServer/static-dashboards/timeline_explorer.html: Added.
1985         * TestResultServer/static-dashboards/treemap.html: Added.
1986         * TestResultServer/static-dashboards/webtreemap.css: Added.
1987         * TestResultServer/static-dashboards/webtreemap.js: Added.
1988         * TestResultServer/stylesheets/dashboardfile.css: Removed.
1989         * TestResultServer/stylesheets/menu.css:
1990         * TestResultServer/templates/dashboardfilelist.html: Removed.
1991         * TestResultServer/templates/menu.html:
1992
1993 2011-07-08  Adam Roben  <aroben@apple.com>
1994
1995         Make TestFailures's list of flaky tests look more like the list of non-flaky tests
1996
1997         Fixes <http://webkit.org/b/64204> TestFailures page's flaky tests list is ugly!
1998
1999         Reviewed by Daniel Bates.
2000
2001         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
2002         (#failure-history, #possibly-flaky-tests): Expanded this rule to apply to the list of flaky
2003         tests.
2004         (#failure-history > li, #possibly-flaky-tests > li): Ditto, but moved the 50px left padding
2005         from here...
2006         (#failure-history > li): ...to here.
2007
2008         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
2009         (ViewController.prototype._domForPossiblyFlakyTests): Give the list an id attribute for
2010         styling purposes.
2011
2012 2011-07-08  Jeffrey Pfau  <jpfau@apple.com>
2013
2014         Unreviewed, add myself as committer.
2015
2016         * Scripts/webkitpy/common/config/committers.py:
2017
2018 2011-07-08  Adam Barth  <abarth@webkit.org>
2019
2020         TestResultsServer should keep old test results
2021         https://bugs.webkit.org/show_bug.cgi?id=64199
2022
2023         Reviewed by Ojan Vafai.
2024
2025         Having historical data will help us do failure archeology.
2026
2027         * TestResultServer/handlers/testfilehandler.py:
2028         * TestResultServer/model/testfile.py:
2029         * TestResultServer/templates/showfilelist.jsonp: Added.
2030
2031 2011-07-08  Adam Roben  <aroben@apple.com>
2032
2033         Remove commit-log-editor's dependency on Module::Load::Conditional
2034
2035         This module isn't available in Perl 5.8.8 (the version used on Leopard).
2036
2037         Fixes <http://webkit.org/b/64198> REGRESSION (r90583):
2038         webkitpy.common.checkout.checkout_unittest failing on Leopard
2039
2040         Reviewed by Daniel Bates.
2041
2042         * Scripts/commit-log-editor: Use the new loadTermReadKey() function instead of
2043         Module::Load::Conditional::can_load.
2044         (loadTermReadKey): Added. Tries to load Term::ReadKey and returns true if it's successful.
2045
2046 2011-07-08  Adam Roben  <aroben@apple.com>
2047
2048         Make checkout_unittest more robust against files moving around
2049
2050         Fixes <http://webkit.org/b/64197> checkout_unittest contains ugly, fragile code to find the
2051         Scripts directory
2052
2053         Reviewed by Adam Barth.
2054
2055         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
2056         (CommitMessageForThisCommitTest.test_commit_message_for_this_commit): Instantiate a real SCM
2057         object and use it to get the path to the Scripts directory, rather than hard-coding the
2058         relative path from this file to Scripts.
2059
2060 2011-07-08  Adam Roben  <aroben@apple.com>
2061
2062         Make TestFailures remember that run-webkit-tests timed out, even across reloads
2063
2064         Fixes <http://webkit.org/b/64193> TestFailures page incorrectly thinks all tests passed in
2065         http://build.webkit.org/builders/Windows%207%20Release%20(Tests)/builds/14589 after a reload
2066
2067         Reviewed by Daniel Bates.
2068
2069         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
2070         (LayoutTestResultsLoader.prototype.start): Store an "error" attribute in the cached data.
2071         When true, it indicates that there was an error retrieving the results for this build and
2072         that the errorCallback should be called.
2073
2074 2011-07-08  Adam Barth  <abarth@webkit.org>
2075
2076         Teach garden-o-matic how to display test results
2077         https://bugs.webkit.org/show_bug.cgi?id=64141
2078
2079         Reviewed by Ojan Vafai.
2080
2081         This patch includes basic infrastructure for probing build.chromium.org
2082         for test results.  We only handle text and image tests, not anything
2083         complicated like reftests.  Also, we're using the revision/build
2084         independent results store on the server, so we're avoiding that
2085         complication for now.
2086
2087         It's slightly hacky that we need to probe the server to see what kinds
2088         of results exist.  A better solution would be to add CORS support to
2089         the server or to use the local server to help.
2090
2091         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
2092         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
2093         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
2094         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
2095         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
2096         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
2097
2098 2011-07-08  Dirk Pranke  <dpranke@chromium.org>
2099
2100         REGRESSION(90419) NRWT's httpd locking is broken for --child-processes=1
2101         https://bugs.webkit.org/show_bug.cgi?id=64092
2102
2103         Reviewed by Tony Chang.
2104
2105         The code was incorrectly creating empty shards in the
2106         shard_in_two case.
2107
2108         * Scripts/webkitpy/layout_tests/controllers/manager.py:
2109         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2110
2111 2011-07-08  Adam Roben  <aroben@apple.com>
2112
2113         Teach buildbot to figure out how many webkitpy/webkitperl tests failed
2114
2115         Fixes <http://webkit.org/b/64192> It's hard to tell how many test-webkitpy/test-webkitperl
2116         tests failed when looking at build.webkit.org
2117
2118         Reviewed by Eric Seidel.
2119
2120         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
2121         (TestWithFailureCount): New class that represents a test build step which has an associated
2122         failure count. Eventually we should move more of our test classes to deriving from this.
2123         (TestWithFailureCount.countFailures): Method for subclasses to override to say how many
2124         failures occurred.
2125
2126         (TestWithFailureCount.commandComplete):
2127         (TestWithFailureCount.evaluateCommand):
2128         (TestWithFailureCount.getText):
2129         (TestWithFailureCount.getText2):
2130         These were all based on RunGtkAPITests.
2131
2132         (RunPythonTests): Changed to inherit from TestWithFailureCount.
2133         (RunPythonTests.countFailures): Parses the test-webkitpy output looking for the count of
2134         failures.
2135         (RunPerlTests): Changed to inherit from TestWithFailureCount.
2136         (RunPerlTests.countFailures): Parses the test-webkitperl output looking for the count of
2137         failures.
2138
2139 2011-07-08  Adam Roben  <aroben@apple.com>
2140
2141         Ensure $CHANGE_LOG_EMAIL_ADDRESS is set when testing webkitpy's commit-log-editor integration
2142
2143         Fixes <http://webkit.org/b/64180> REGRESSION (r90564): test-webkitpy failing on multiple
2144         bots due to commit-log-editor errors
2145
2146         Reviewed by Adam Barth.
2147
2148         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
2149         (CommitMessageForThisCommitTest.test_commit_message_for_this_commit): Set
2150         $CHANGE_LOG_EMAIL_ADDRESS to the patch author's email address. This ensures that
2151         commit-log-editor can find a value for the committer's email, and that the committer and
2152         author email addresses match, which will prevent commit-log-editor from inserting a "Patch
2153         by" line in the commit message.
2154
2155 2011-07-08  Andreas Kling  <kling@webkit.org>
2156
2157         [Qt][WK2] Views should expose QActions for basic navigation.
2158         https://bugs.webkit.org/show_bug.cgi?id=64174
2159
2160         Reviewed by Benjamin Poulain.
2161
2162         Bring the toolbar in Qt's MiniBrowser back to life.
2163
2164         * MiniBrowser/qt/BrowserView.cpp:
2165         (BrowserView::navigationAction):
2166         * MiniBrowser/qt/BrowserView.h:
2167         * MiniBrowser/qt/BrowserWindow.cpp:
2168         (BrowserWindow::BrowserWindow):
2169
2170 2011-07-08  Adam Roben  <aroben@apple.com>
2171
2172         Teach TestFailures how to find test names in commit-log-editor-style commit messages
2173
2174         TestFailures was relying on Trac turning the list of modified files in our commit messages
2175         into an HTML list. But Trac only does this when the list of modified files is indented.
2176         commit-log-editor doesn't indent the file list, so the list wasn't being turned into an HTML
2177         list, which was confusing TestFailures.
2178
2179         TestFailures now does much simpler parsing of the commit message (i.e., just a substring
2180         search) without relying at all on its structure.
2181
2182         Fixes <http://webkit.org/b/64173> TestFailures page fails to blame r90608 for breaking
2183         fast/dom/HTMLProgressElement/progress-element-markup.html on Windows
2184
2185         Reviewed by David Kilzer.
2186
2187         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Trac.js:
2188         (Trac.prototype.getCommitDataForRevisionRange): Instead of trying to parse the commit
2189         message, just return its text.
2190
2191         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
2192         (ViewController.prototype._domForRegressionRange): Instead of searching for test names in
2193         each commit's list of modified files, just search for test names anywhere in the commit's
2194         message.
2195
2196 2011-07-08  Adam Barth  <abarth@webkit.org>
2197
2198         sheriffbot is too spammy in IRC
2199         https://bugs.webkit.org/show_bug.cgi?id=64153
2200
2201         Reviewed by Eric Seidel.
2202
2203         Reporting failures in IRC worked well when the tree was greener than it
2204         is today.  Nowadays, this feature mostly just results in spam about
2205         false positives.  If we reach a greener state, we can bring this code
2206         back.
2207
2208         (Another possibility is to restrict this feature to builder bots,
2209         rather than including testers, as we did before this patch.)
2210
2211         * Scripts/webkitpy/tool/commands/sheriffbot.py:
2212         * Scripts/webkitpy/tool/commands/sheriffbot_unittest.py:
2213
2214 2011-07-08  Adam Roben  <aroben@apple.com>
2215
2216         Ensure commit-log-editor adds a "Patch by" line when the author and committer are different
2217
2218         Previously we were only adding a "Patch by" line when the ChangeLog contained a "Reviewed
2219         by" line. But some patches (like rollout patches) don't contain that line. Now we always add
2220         "Patch by" to the commit log regardless of the ChangeLog's contents.
2221
2222         Fixes <http://webkit.org/b/64127> Committer for r90588 is commit-queue@webkit.org, but
2223         should have been abarth@webkit.org
2224
2225         Reviewed by Anders Carlsson.
2226
2227         * Scripts/commit-log-editor:
2228         (createCommitMessage): Try to put the "Patch by" line just above the "Reviewed by" line, as
2229         before. If there is no "Reviewed by" line, try to put it just above the first modified file.
2230         If all else fails, put it at the end of the commit message.
2231         (patchAuthorshipString): Added. Code came from createCommitMessage.
2232
2233 2011-07-08  Herczeg Zoltan  <zherczeg@webkit.org>
2234
2235         Reviewed by Andreas Kling.
2236
2237         Adding myself as a reviewer.
2238
2239         * Scripts/webkitpy/common/config/committers.py:
2240
2241 2011-07-07  Kevin Ollivier  <kevino@theolliviers.com>
2242
2243         [wx] Unreviewed build fix, remove old files from the tree and update paths to new ones.
2244
2245         * wx/build: Removed.
2246         * wx/build/build_utils.py: Removed.
2247         * wx/build/settings.py: Removed.
2248         * wx/build/waf_extensions.py: Removed.
2249         * wx/build/wxpresets.py: Removed.
2250         * wx/packaging/build-mac-installer.py:
2251
2252 2011-07-07  Kevin Ollivier  <kevino@theolliviers.com>
2253
2254         [wx] Unreviewed build fix, more fixes for install name issues.
2255
2256         * wx/packaging/build-mac-installer.py:
2257
2258 2011-07-07  Dirk Pranke  <dpranke@chromium.org>
2259
2260         test-webkitpy fails on chromium win
2261         https://bugs.webkit.org/show_bug.cgi?id=64137
2262
2263         Reviewed by Eric Seidel.
2264
2265         test-webkitpy doesn't play nicely with the multiprocessing
2266         module on win32. An earlier change actually reenabled the tests
2267         on win32 by mistake.
2268
2269         This patch also fixes a few cases where path names will trip
2270         things up on win32 (e.g., by testing unix paths on windows).
2271         We do not lose any real coverage here with those fixes.
2272
2273         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2274         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2275         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
2276         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2277
2278 2011-07-07  Leandro Pereira  <leandro@profusion.mobi>
2279
2280         [ImageDiff] Calculate/print difference right after reading baseline image.
2281         https://bugs.webkit.org/show_bug.cgi?id=64117
2282
2283         Reviewed by Kent Tamura.
2284
2285         Otherwise, ImageDiff will block on fgets() until the universe ends or you ^C
2286         it.  Whichever happens first.
2287
2288         * DumpRenderTree/efl/ImageDiff.cpp:
2289         (main):
2290
2291 2011-07-07  Raphael Kubo da Costa  <kubo@profusion.mobi>
2292
2293         [EFL] DRT: Leak cairo_t to make the pixel tests stop crashing
2294         https://bugs.webkit.org/show_bug.cgi?id=64107
2295
2296         Reviewed by Kent Tamura.
2297
2298         The pointer is later managed by BitmapContext, but it was being
2299         de-refed and deleted earlier by our RefPtr, causing crashes in all
2300         pixel tests.
2301
2302         We now call leakRef() to make sure it is not removed when our
2303         createBitmapContextFromWebView goes out of context.
2304
2305         * DumpRenderTree/efl/PixelDumpSupportEfl.cpp:
2306         (createBitmapContextFromWebView):
2307
2308 2011-07-07  Adam Roben  <aroben@apple.com>
2309
2310         Update TestFailures's title and header immediately upon navigation
2311
2312         Fixes <http://webkit.org/b/64125> TestFailures page seems slow to react on link clicks
2313
2314         Reviewed by Daniel Bates.
2315
2316         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
2317         (ViewController.prototype.loaded): Create and store an h1 element for displaying the page's
2318         title.
2319         (ViewController.prototype._displayBuilder): Set the page's title and clear out any old
2320         content immediately instead of waiting for the first set of results for the history
2321         analyzer. As we receive new results from the analyzer we'll just clear out the main content
2322         area and replace it with the new info.
2323         (ViewController.prototype._displayTesters): Set the page's title and clear out any old
2324         content immediately instead of waiting for the list of testers from the buildbot. Sprinkled
2325         in a little use of our removeAllChildren helper, too.
2326         (ViewController.prototype._setTitle): Added. Set's the page's title and the header text.
2327
2328 2011-07-07  Adam Roben  <aroben@apple.com>
2329
2330         Make Checkout use SCM's Executive instead of conjuring up its own
2331
2332         This will improve integration with the rest of webkitpy, particularly when invoked via
2333         webkit-patch.
2334
2335         Fixes <http://webkit.org/b/64115> REGRESSION (r90564): webkitpy's Checkout class uses
2336         Executive inappropriately
2337
2338         Reviewed by Adam Barth.
2339
2340         * Scripts/webkitpy/common/checkout/checkout.py:
2341         (Checkout.commit_message_for_this_commit): Use SCM.run instead of creating an Executive for
2342         our own use. SCM might have some super-special Executive that it uses under the covers, and
2343         we want to use it, too!
2344
2345         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
2346         (CommitMessageForThisCommitTest.test_commit_message_for_this_commit): Mock the SCM.run
2347         method to call through to Executive.run_command.
2348
2349 2011-07-07  Adam Roben  <aroben@apple.com>
2350
2351         Make Term/ReadKey.pm an optional dependency of commit-log-editor
2352
2353         When Term/ReadKey.pm isn't present, the --regenerate-log switch will be non-functional.
2354
2355         Fixes <http://webkit.org/b/64113> REGRESSION (r90564):
2356         webkitpy.common.checkout.checkout_unittest failing on GTK bots due to missing
2357         Term/ReadKey.pm module
2358
2359         Reviewed by Xan Lopez.
2360
2361         * Scripts/commit-log-editor: Use Module::Load::Conditional::can_load to only load
2362         Term/ReadKey.pm if it's present. If it isn't present, ignore the --regenerate-log switch so
2363         that existing commit messages will be preserved. (Users can manually remove existing commit
2364         messages using git-reset.)
2365
2366 2011-07-07  Adam Roben  <aroben@apple.com>
2367
2368         Keep testing that commit messages containing Unicode are handled properly by webkitpy
2369
2370         Fixes <http://webkit.org/b/64109> REGRESSION (r90571): test-webkitpy no longer tests that we
2371         correctly handle Unicode in commit messages
2372
2373         Reviewed by Eric Seidel.
2374
2375         * Scripts/webkitpy/common/checkout/checkout_unittest.py: Put some Unicode characters in
2376         places in the ChangeLog that will end up being included in the commit message.
2377
2378 2011-07-07  Andreas Kling  <kling@webkit.org>
2379
2380         [Qt][WK2] Remove Symbian code.
2381         https://bugs.webkit.org/show_bug.cgi?id=64101
2382
2383         Reviewed by Benjamin Poulain.
2384
2385         * MiniBrowser/qt/BrowserWindow.cpp:
2386         (BrowserWindow::BrowserWindow):
2387         (BrowserWindow::screenshot):
2388         * MiniBrowser/qt/MiniBrowser.pro:
2389         * MiniBrowser/qt/MiniBrowserApplication.h:
2390         (WindowOptions::WindowOptions):
2391
2392 2011-07-07  Andreas Kling  <kling@webkit.org>
2393
2394         [Qt][WK2] Don't support ridiculous matrix of QT_NO_FEATURE combinations.
2395         https://bugs.webkit.org/show_bug.cgi?id=64099
2396
2397         Reviewed by Benjamin Poulain.
2398
2399         * MiniBrowser/qt/BrowserWindow.cpp:
2400         (BrowserWindow::openFile):
2401         (BrowserWindow::screenshot):
2402         (BrowserWindow::loadURLListFromFile):
2403
2404 2011-07-07  Adam Roben  <aroben@apple.com>
2405
2406         Completely ignore too-many-failures builds in TestFailures in most circumstances
2407
2408         Because a semi-arbitrary subset of tests are run in a too-many-failures build, we can't
2409         really use them to perform regression analysis. The only time we want to pay attention to
2410         too-many-failures builds is when we're trying to explain when the current bout of
2411         too-many-failures started.
2412
2413         Fixes <http://webkit.org/b/64106> TestFailures page sometimes claims a test started failing
2414         in a build that didn't even run it (because it exited early due to too many failues)
2415
2416         Reviewed by David Kilzer.
2417
2418         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
2419         (LayoutTestHistoryAnalyzer.prototype._incorporateBuildHistory): Removed old, broken
2420         too-many-failures handling that would cause us to blame builds that didn't even run a given
2421         test for breaking it. Instead, skip over all too-many-failures builds unless the most recent
2422         build was itself a too-many-failures build.
2423
2424 2011-07-07  Adam Roben  <aroben@apple.com>
2425
2426         Teach webkitpy's Checkout class to use commit-log-editor to create commit messages
2427
2428         Fixes <http://webkit.org/b/26755> webkit-patch's commit messages are less readable than
2429         commit-log-editor's
2430
2431         Reviewed by David Kilzer.
2432
2433         * Scripts/webkitpy/common/checkout/checkout.py:
2434         (Checkout.commit_message_for_this_commit): Run commit-log-editor, passing it the paths of
2435         the modified ChangeLogs, to generate the commit message, rather than trying to generate one
2436         ourselves.
2437
2438         * Scripts/webkitpy/common/checkout/checkout_unittest.py: Updated the expected commit message
2439         to match commit-log-editor's format.
2440         (CommitMessageForThisCommitTest.setUp): Write the ChangeLogs into Tools and LayoutTests
2441         directories so we can see how the various entries get labeled in the commit message.
2442         (CommitMessageForThisCommitTest.test_commit_message_for_this_commit): Create a mock SCM
2443         instance that knows how to find commit-log-editor and pass it to our Checkout instance.
2444         Don't bother capturing output, since there shouldn't be any.
2445
2446 2011-07-07  Adam Roben  <aroben@apple.com>
2447
2448         Add a --print-log option to commit-log-editor
2449
2450         When specified, commit-log-editor takes a set of ChangeLog paths on the command line,
2451         generates a commit message from those ChangeLogs, prints it to stdout, and exits.
2452
2453         Prep work for fixing <http://webkit.org/b/26755> webkit-patch's commit messages are less
2454         readable than commit-log-editor's
2455
2456         Reviewed by David Kilzer.
2457
2458         * Scripts/commit-log-editor: Changed to use Getopt::Long to parse options. Added --print-log
2459         option, which calls through to createCommitMessage, prints the result, and exits.
2460         (printUsageAndExit): Renamed from "usage" for clarity. Beefed up the usage statement to
2461         explain commit-log-editor's 2.5 (normal, --print-log, and --help) modes of operation.
2462
2463 2011-07-07  Adam Roben  <aroben@apple.com>
2464
2465         Extract commit-log-editor's commit-message-generation code into a separate function
2466
2467         As a bonus, we now skip a bunch of work in the case where we are generating a commit message
2468         for staged git changes.
2469
2470         Prep work for fixing <http://webkit.org/b/26755> webkit-patch's commit messages are less
2471         readable than commit-log-editor's
2472
2473         Reviewed by David Kilzer.
2474
2475         * Scripts/commit-log-editor: Moved the default definition of $endl up toward the top of the
2476         file so that it will be defined even if we aren't operating on an existing log message.
2477         Moved a bunch of top-level code to generate the commit message from here...
2478         (createCommitMessage): ...to here.
2479
2480 2011-07-07  Carlos Garcia Campos  <cgarcia@igalia.com>
2481
2482         [GTK] Open links in a new window when clicking with the middle button in MiniBrowser
2483         https://bugs.webkit.org/show_bug.cgi?id=63988
2484
2485         Reviewed by Martin Robinson.
2486
2487         * MiniBrowser/gtk/BrowserWindow.c:
2488         (browserWindowConstructed):
2489         (decidePolicyForNavigationAction): Ignore the action if a link is
2490         clicked with the middle buttonm and open the link in a new window.
2491         (browserWindowPolicyClientInit):
2492
2493 2011-07-07  Adam Barth  <abarth@webkit.org>
2494
2495         Remove style_references.py
2496         https://bugs.webkit.org/show_bug.cgi?id=64070
2497
2498         Reviewed by Eric Seidel.
2499
2500         We never really adopted the _references design, and this file is a bit
2501         of an orphan at this point.
2502
2503         * Scripts/check-webkit-style:
2504         * Scripts/webkitpy/style/checker.py:
2505         * Scripts/webkitpy/style/checker_unittest.py:
2506         * Scripts/webkitpy/style/checkers/python.py:
2507         * Scripts/webkitpy/style/checkers/test_expectations.py:
2508         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2509         * Scripts/webkitpy/style/main_unittest.py:
2510         * Scripts/webkitpy/style_references.py: Removed.
2511
2512 2011-07-07  Eric Seidel  <eric@webkit.org>
2513
2514         Fix WebKit2 expected results search paths for Mac and Qt under new-run-webkit-tests
2515         https://bugs.webkit.org/show_bug.cgi?id=64056
2516
2517         Unreviewed.
2518
2519         I had written this unittest prior to landing my change, but
2520         had named it qt_unitest.py, so it wasn't running (nor was
2521         it added to my git repo).
2522
2523         Once I fixe the name of the unittest file, it was very easy
2524         to fix the error in webkit.py (which already had a FIXME).
2525
2526         * Scripts/webkitpy/layout_tests/port/qt_unittest.py: Added.
2527         * Scripts/webkitpy/layout_tests/port/webkit.py:
2528
2529 2011-07-07  Eric Seidel  <eric@webkit.org>
2530
2531         Unreviewed.  Add more debug logging about which test expectations we're using.
2532
2533         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2534         * Scripts/webkitpy/layout_tests/port/webkit.py:
2535
2536 2011-07-06  Adam Barth  <abarth@webkit.org>
2537
2538         Remove unexpected_results.json
2539         https://bugs.webkit.org/show_bug.cgi?id=64065
2540
2541         Reviewed by Eric Seidel.
2542
2543         No one should be using this file any more.  This patch removes it from
2544         the face of the earth.  This patch also removes two NRWT-specific
2545         command line options that were implemented using
2546         unexpected_results.json.
2547
2548         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
2549         * Scripts/webkitpy/layout_tests/controllers/manager.py:
2550         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2551         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2552         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2553         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
2554         * Scripts/webkitpy/tool/servers/rebaselineserver_unittest.py:
2555
2556 2011-07-07  Adam Barth  <abarth@webkit.org>
2557
2558         REGRESSION(90520): chromium GPU tests produce output in the LayoutTests directory
2559         https://bugs.webkit.org/show_bug.cgi?id=64067
2560
2561         Reviewed by Eric Seidel.
2562
2563         The author of r90520 forgot to change the chromium_gpu.py implemenation
2564         of this method as well.
2565
2566         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2567
2568 2011-07-06  Eric Seidel  <eric@webkit.org>
2569
2570         Fix WebKit2 expected results search paths for Mac and Qt under new-run-webkit-tests
2571         https://bugs.webkit.org/show_bug.cgi?id=64056
2572
2573         Reviewed by Adam Barth.
2574
2575         This fixes the MacPort's version of baseline_search_path to include
2576         wk2 fallback like WebKitPort's does.  This also re-works the QtPort
2577         to set self._name (as other ports do) so that WebKitPort's default
2578         implementations will understand Qt operating system flavors correctly
2579         (including when running wk2 tests).
2580
2581         In trying to test this, I realized that nearly all of our port-tests
2582         use the default FileSystem(), User() and Executive() objects, which
2583         is really bad, because they expose details of your system in the unittest results!
2584
2585         Once I fixed many of the Port() calls to pass MockFileSystem(), then
2586         MacPort.test_tests_for_other_platforms started failing, and I had to
2587         teach MockFileSystem.glob how to handle directories to make it work again.
2588
2589         I removed the useless PortTestCase.test_baseline_search_path and replaced
2590         it with more useful test_baseline_search_path tests in MacPortTest and QtPortTest.
2591
2592         * Scripts/webkitpy/common/config/ports.py:
2593         * Scripts/webkitpy/common/system/filesystem_mock.py:
2594         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2595         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2596         * Scripts/webkitpy/layout_tests/port/mac.py:
2597         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
2598         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
2599         * Scripts/webkitpy/layout_tests/port/qt.py:
2600         * Scripts/webkitpy/layout_tests/port/test_files.py:
2601         * Scripts/webkitpy/layout_tests/port/win.py:
2602         * Scripts/webkitpy/to_be_moved/deduplicate_tests.py:
2603         * Scripts/webkitpy/tool/mocktool.py:
2604
2605 2011-07-06  Adam Barth  <abarth@webkit.org>
2606
2607         Wean rebaseline-server off unexpected_results.json
2608         https://bugs.webkit.org/show_bug.cgi?id=64063
2609
2610         Reviewed by Eric Seidel.
2611
2612         In the process of changing this code to use full_results.json, I
2613         noticed that the code was broken (because it wasn't tested).  This
2614         patch also adds test coverage for the broken code.
2615
2616         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
2617         * Scripts/webkitpy/tool/servers/rebaselineserver.py:
2618         * Scripts/webkitpy/tool/servers/rebaselineserver_unittest.py:
2619
2620 2011-07-06  Adam Barth  <abarth@webkit.org>
2621
2622         Wean resultsjsonparser off unexpected_results.json
2623         https://bugs.webkit.org/show_bug.cgi?id=64061
2624
2625         Reviewed by Eric Seidel.
2626
2627         We're about to delete unexpected_results.json, so we need to remove all
2628         the users.
2629
2630         * Scripts/webkitpy/common/config/ports.py:
2631         * Scripts/webkitpy/common/net/resultsjsonparser.py:
2632         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
2633         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2634
2635 2011-07-06  Adam Barth  <abarth@webkit.org>
2636
2637         Cleanup result_summary.py
2638         https://bugs.webkit.org/show_bug.cgi?id=64057
2639
2640         Reviewed by Eric Seidel.
2641
2642         This class had a bunch of out-of-date docstrings that no longer make
2643         any sense.
2644
2645         * Scripts/webkitpy/layout_tests/models/result_summary.py:
2646
2647 2011-07-06  Adam Barth  <abarth@webkit.org>
2648
2649         Tweak the ChangeLog for DEPS rolls to say "Unreviewed" so that the
2650         patches can be landed by the commit-queue.
2651
2652         * Scripts/webkitpy/tool/steps/preparechangelogfordepsroll.py:
2653
2654 2011-07-06  Adam Barth  <abarth@webkit.org>
2655
2656         Normalize callers of logging.getLogger to use __name__
2657         https://bugs.webkit.org/show_bug.cgi?id=64054
2658
2659         Reviewed by Eric Seidel.
2660
2661         Most callers were already using __name__.  This patch makes everyone
2662         consistent.
2663
2664         * Scripts/webkitpy/common/checkout/diff_parser.py:
2665         * Scripts/webkitpy/common/checkout/scm/svn.py:
2666         * Scripts/webkitpy/common/net/statusserver.py:
2667         * Scripts/webkitpy/common/system/executive.py:
2668         * Scripts/webkitpy/common/system/file_lock.py:
2669         * Scripts/webkitpy/common/system/user.py:
2670         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
2671         * Scripts/webkitpy/layout_tests/models/result_summary.py:
2672         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
2673         * Scripts/webkitpy/layout_tests/port/chromium.py:
2674         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2675         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
2676         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2677         * Scripts/webkitpy/layout_tests/port/gtk.py:
2678         * Scripts/webkitpy/layout_tests/port/http_lock.py:
2679         * Scripts/webkitpy/layout_tests/port/mac.py:
2680         * Scripts/webkitpy/layout_tests/port/qt.py:
2681         * Scripts/webkitpy/layout_tests/port/server_process.py:
2682         * Scripts/webkitpy/layout_tests/port/webkit.py:
2683         * Scripts/webkitpy/layout_tests/port/win.py:
2684         * Scripts/webkitpy/layout_tests/servers/apache_http_server.py:
2685         * Scripts/webkitpy/layout_tests/servers/http_server.py:
2686         * Scripts/webkitpy/layout_tests/servers/http_server_base.py:
2687         * Scripts/webkitpy/layout_tests/servers/websocket_server.py:
2688         * Scripts/webkitpy/layout_tests/views/printing.py:
2689         * Scripts/webkitpy/python24/versioning.py:
2690         * Scripts/webkitpy/style/checker.py:
2691         * Scripts/webkitpy/style/checkers/test_expectations.py:
2692         * Scripts/webkitpy/style/patchreader.py:
2693         * Scripts/webkitpy/to_be_moved/update_webgl_conformance_tests.py:
2694
2695 2011-07-06  Adam Barth  <abarth@webkit.org>
2696
2697         Move model classes out of layout_package and into models
2698         https://bugs.webkit.org/show_bug.cgi?id=64053
2699
2700         Reviewed by Eric Seidel.
2701
2702         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2703         * Scripts/webkitpy/common/net/layouttestresults.py:
2704         * Scripts/webkitpy/common/net/layouttestresults_unittest.py:
2705         * Scripts/webkitpy/common/net/resultsjsonparser.py:
2706         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
2707         * Scripts/webkitpy/layout_tests/controllers/manager.py:
2708         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2709         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py:
2710         * Scripts/webkitpy/layout_tests/controllers/worker.py:
2711         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
2712         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
2713         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
2714         * Scripts/webkitpy/layout_tests/layout_package/result_summary.py: Removed.
2715         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py: Removed.
2716         * Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py: Removed.
2717         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py: Removed.
2718         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py: Removed.
2719         * Scripts/webkitpy/layout_tests/layout_package/test_input.py: Removed.
2720         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py:
2721         * Scripts/webkitpy/layout_tests/layout_package/test_results.py: Removed.
2722         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py: Removed.
2723         * Scripts/webkitpy/layout_tests/models: Added.
2724         * Scripts/webkitpy/layout_tests/models/__init__.py: Added.
2725         * Scripts/webkitpy/layout_tests/models/result_summary.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/result_summary.py.
2726         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_expectations.py.
2727         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py.
2728         * Scripts/webkitpy/layout_tests/models/test_failures.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_failures.py.
2729         * Scripts/webkitpy/layout_tests/models/test_failures_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py.
2730         * Scripts/webkitpy/layout_tests/models/test_input.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_input.py.
2731         * Scripts/webkitpy/layout_tests/models/test_results.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_results.py.
2732         * Scripts/webkitpy/layout_tests/models/test_results_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py.
2733         * Scripts/webkitpy/layout_tests/port/chromium.py:
2734         * Scripts/webkitpy/layout_tests/views/printing.py:
2735         * Scripts/webkitpy/layout_tests/views/printing_unittest.py:
2736         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2737         * Scripts/webkitpy/style_references.py:
2738         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests.py:
2739         * Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
2740         * Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py:
2741         * Scripts/webkitpy/tool/commands/queues_unittest.py:
2742         * Scripts/webkitpy/tool/commands/rebaseline.py:
2743
2744 2011-07-06  Dirk Pranke  <dpranke@chromium.org>
2745
2746         nrwt: fix typo preventing http startup on windows
2747         https://bugs.webkit.org/show_bug.cgi?id=64050
2748
2749         Reviewed by Eric Siedel.
2750
2751         NRWT checks whether it needs to start the servers by looking
2752         for "/http/" in the test names. After r90520, the leading slash
2753         is not present, so the check isn't quite right.
2754
2755         * Scripts/webkitpy/layout_tests/controllers/manager.py:
2756         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2757
2758 2011-07-06  Dirk Pranke  <dpranke@chromium.org>
2759
2760         nrwt: remove --use-apache from the command line
2761         https://bugs.webkit.org/show_bug.cgi?id=63358
2762
2763         Reviewed by Adam Barth.
2764
2765         Re-land the fix in bug 63358 - the prior fix was flawed in that
2766         we were trying to start the http server during check_sys_deps()
2767         but not actually setting up the environment properly. This
2768         broken prior to this change, but exposed by it, since the patch
2769         changes the chromium port to call the base class's checks.
2770
2771         * Scripts/webkitpy/common/system/executive.py:
2772         * Scripts/webkitpy/layout_tests/port/base.py:
2773         * Scripts/webkitpy/layout_tests/port/chromium.py:
2774         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2775         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2776         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2777         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
2778         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
2779         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2780
2781 2011-07-06  Adam Barth  <abarth@webkit.org>
2782
2783         Move view-related code out of layout_package and into views
2784         https://bugs.webkit.org/show_bug.cgi?id=64051
2785
2786         Reviewed by Eric Seidel.
2787
2788         * Scripts/webkitpy/layout_tests/controllers/manager.py:
2789         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
2790         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
2791         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py:
2792         * Scripts/webkitpy/layout_tests/layout_package/metered_stream.py: Removed.
2793         * Scripts/webkitpy/layout_tests/layout_package/metered_stream_unittest.py: Removed.
2794         * Scripts/webkitpy/layout_tests/layout_package/printing.py: Removed.
2795         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py: Removed.
2796         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2797         * Scripts/webkitpy/layout_tests/views: Added.
2798         * Scripts/webkitpy/layout_tests/views/__init__.py: Added.
2799         * Scripts/webkitpy/layout_tests/views/metered_stream.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/metered_stream.py.
2800         * Scripts/webkitpy/layout_tests/views/metered_stream_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/metered_stream_unittest.py.
2801         * Scripts/webkitpy/layout_tests/views/printing.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/printing.py.
2802         * Scripts/webkitpy/layout_tests/views/printing_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py.
2803
2804 2011-07-06  Adam Barth  <abarth@webkit.org>
2805
2806         Add goofy comments to explain why we have these files.
2807
2808         * Scripts/webkitpy/layout_tests/controllers/__init__.py:
2809         * Scripts/webkitpy/layout_tests/layout_package/__init__.py:
2810
2811 2011-07-06  Adam Barth  <abarth@webkit.org>
2812
2813         Move a bunch of classes out of layout_tests.layout_package and into layout_tests.controllers
2814         https://bugs.webkit.org/show_bug.cgi?id=64048
2815
2816         Reviewed by Eric Seidel.
2817
2818         * Scripts/webkitpy/layout_tests/controllers: Added.
2819         * Scripts/webkitpy/layout_tests/controllers/__init__.py: Added.
2820         * Scripts/webkitpy/layout_tests/controllers/manager.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/manager.py.
2821         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py.
2822         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py.
2823         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py.
2824         * Scripts/webkitpy/layout_tests/controllers/message_broker.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/message_broker.py.
2825         * Scripts/webkitpy/layout_tests/controllers/message_broker_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py.
2826         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py.
2827         * Scripts/webkitpy/layout_tests/controllers/worker.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/worker.py.
2828         * Scripts/webkitpy/layout_tests/layout_package/manager.py: Removed.
2829         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py: Removed.
2830         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py: Removed.
2831         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py: Removed.
2832         * Scripts/webkitpy/layout_tests/layout_package/message_broker.py: Removed.
2833         * Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py: Removed.
2834         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
2835         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py: Removed.
2836         * Scripts/webkitpy/layout_tests/layout_package/worker.py: Removed.
2837         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2838
2839 2011-07-06  Adam Barth  <abarth@webkit.org>
2840
2841         Attempt to fix unittest on Leopard.  Let's enumerate the keys in sorted
2842         order so the test results are predictable.
2843
2844         * Scripts/webkitpy/tool/bot/irc_command.py:
2845         * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:
2846
2847 2011-07-06  Dirk Pranke  <dpranke@chromium.org>
2848
2849         Reviewed by Eric Siedel.
2850
2851         nrwt: make the code be consistent about using test names instead of filenames or paths
2852
2853         https://bugs.webkit.org/show_bug.cgi?id=63597
2854
2855         It can be hard to figure out where in the code we're assuming
2856         test names are listed as unix-style relative filenames and where
2857         they are either absolute paths or relative paths following the
2858         host filesystem convention.
2859
2860         This patch changes things so that everything outside of the
2861         Port object uses (and must assume) unix-style relative
2862         filenames (with one exception, which is specifying host-local
2863         filenames as a list of test arguments on the command line).
2864
2865         This should make things clearer and more consistent, and also
2866         removes a lot of path conversion calls.
2867
2868         The changes in this patch outside of port/base.py are largely
2869         cosmetic changes of "filename" to "test". We add a few routines
2870         to the base Port implementation to manipulate test names to
2871         split them, figure out if we're referring to a directory of
2872         tests, etc.
2873
2874         * Scripts/webkitpy/common/net/layouttestresults.py:
2875         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
2876         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
2877         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
2878         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
2879         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
2880         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
2881         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
2882         * Scripts/webkitpy/layout_tests/layout_package/result_summary.py:
2883         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
2884         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
2885         * Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
2886         * Scripts/webkitpy/layout_tests/layout_package/test_input.py:
2887         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py:
2888         * Scripts/webkitpy/layout_tests/layout_package/test_results.py:
2889         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:
2890         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
2891         * Scripts/webkitpy/layout_tests/port/base.py:
2892         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2893         * Scripts/webkitpy/layout_tests/port/chromium.py:
2894         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2895         * Scripts/webkitpy/layout_tests/port/dryrun.py:
2896         * Scripts/webkitpy/layout_tests/port/test.py:
2897         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2898         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2899         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests.py:
2900         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests_unittest.py:
2901         * Scripts/webkitpy/tool/bot/flakytestreporter.py:
2902         * Scripts/webkitpy/tool/commands/rebaseline.py:
2903         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
2904
2905 2011-07-06  Chris Guillory  <ctguil@chromium.org>
2906
2907         No review necessary.
2908
2909         Adding myself to committers list in committers.py.
2910         https://bugs.webkit.org/show_bug.cgi?id=63478
2911
2912         * Scripts/webkitpy/common/config/committers.py:
2913
2914 2011-07-06  Johnny Ding  <jnd@chromium.org>
2915
2916         https://bugs.webkit.org/show_bug.cgi?id=63857
2917         Implement createPopupMenu for Chromium DRT.
2918
2919         Reviewed by Tony Chang.
2920
2921         * DumpRenderTree/chromium/WebViewHost.cpp:
2922         (WebViewHost::createPopupMenu):
2923         (WebViewHost::~WebViewHost):
2924         * DumpRenderTree/chromium/WebViewHost.h:
2925
2926 2011-07-06  Mihai Parparita  <mihaip@chromium.org>
2927
2928         Unreviewed, rolling out r90503.
2929         http://trac.webkit.org/changeset/90503
2930         https://bugs.webkit.org/show_bug.cgi?id=63358
2931
2932         Makes Chromium Windows canary bots hang.
2933
2934         * Scripts/webkitpy/layout_tests/port/base.py:
2935         * Scripts/webkitpy/layout_tests/port/chromium.py:
2936         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2937         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2938         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2939         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
2940         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
2941         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2942
2943 2011-07-06  Adam Barth  <abarth@webkit.org>
2944
2945         Move buildbot URL in webkitpy to config.urls
2946         https://bugs.webkit.org/show_bug.cgi?id=64041
2947
2948         Reviewed by Eric Seidel.
2949
2950         This patch prepares for using the buildbot class to talk to the
2951         Chromium buildbots.
2952
2953         * Scripts/webkitpy/common/config/urls.py:
2954         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
2955         * Scripts/webkitpy/tool/commands/download_unittest.py:
2956         * Scripts/webkitpy/tool/commands/queries.py:
2957         * Scripts/webkitpy/tool/mocktool.py:
2958         * Scripts/webkitpy/tool/steps/ensurebuildersaregreen.py:
2959
2960 2011-07-06  Kevin Ollivier  <kevino@theolliviers.com>
2961
2962         [wx] Unreviewed build fix, add the correct path to the waf build settings.
2963         
2964         * wx/packaging/build-win-installer.py:
2965
2966 2011-07-06  Adam Barth  <abarth@webkit.org>
2967
2968         Add missing include.
2969
2970         * Scripts/webkitpy/tool/bot/irc_command.py:
2971
2972 2011-07-06  Adam Barth  <abarth@webkit.org>
2973
2974         Add roll-chromium-deps command to sheriff-bot
2975         https://bugs.webkit.org/show_bug.cgi?id=64037
2976
2977         Reviewed by Eric Seidel.
2978
2979         Lo, and the DEPS were rolled on command.
2980
2981         * Scripts/webkitpy/tool/bot/irc_command.py:
2982         * Scripts/webkitpy/tool/bot/irc_command_unittest.py:
2983         * Scripts/webkitpy/tool/bot/sheriff.py:
2984         * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:
2985         * Scripts/webkitpy/tool/commands/roll.py:
2986         * Scripts/webkitpy/tool/commands/roll_unittest.py:
2987         * Scripts/webkitpy/tool/steps/updatechromiumdeps.py:
2988
2989 2011-07-06  Eric Seidel  <eric@webkit.org>
2990
2991         Split Driver, DriverOutput and DriverInput out into a new driver.py file
2992         https://bugs.webkit.org/show_bug.cgi?id=64031
2993
2994         Reviewed by Adam Barth.
2995
2996         This ended up kinda large and I'm happy to break it up if desired.
2997
2998         As part of cleaning up the Port module, I split Driver, DriverInput and DriverOutput
2999         out into its own driver.py file.
3000
3001         I also cleaned up how we import Port objects, by making port/__init__.py
3002         expose Port, Driver, DriverInput, and DriverOutput.  I removed exposure of
3003         some various test objects as they were causing circular dependencies.
3004         base.py is now an implementation detail of port.  No longer is "base" referenced
3005         anywhere in the code.
3006
3007         I also made Driver.__init__ non-virtual so that sub-classers could all share
3008         the data storage on Driver.  This more closely matches how other subclasses
3009         work throughout our python code.
3010
3011         Standard line-unwrapping and doc-string removal as I went through the code.
3012
3013         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
3014         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
3015         * Scripts/webkitpy/layout_tests/port/__init__.py:
3016         * Scripts/webkitpy/layout_tests/port/base.py:
3017         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
3018         * Scripts/webkitpy/layout_tests/port/chromium.py:
3019         * Scripts/webkitpy/layout_tests/port/driver.py: Added.
3020         * Scripts/webkitpy/layout_tests/port/dryrun.py:
3021         * Scripts/webkitpy/layout_tests/port/mock_drt.py:
3022         * Scripts/webkitpy/layout_tests/port/test.py:
3023         * Scripts/webkitpy/layout_tests/port/webkit.py:
3024         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
3025         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests_unittest.py:
3026         * Scripts/webkitpy/tool/servers/rebaselineserver_unittest.py:
3027
3028 2011-07-06  Adam Roben  <aroben@apple.com>
3029
3030         Clear TestFailures's getNumberOfFailingTests cache
3031
3032         Now that we support NRWT, we need to clear out old zero values for the NRWT bots that might
3033         have been cached by old versions of TestFailures.
3034
3035         Fixes <http://webkit.org/b/64029> TestFailures page still doesn't show NRWT bots if you've
3036         used it recently
3037
3038         Reviewed by Adam Barth.
3039
3040         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
3041         (Builder.prototype.getNumberOfFailingTests): Added a .version property to the cache data
3042         that must match the current version. Otherwise we'll ignore the cached data.
3043
3044 2011-06-24  Dirk Pranke  <dpranke@chromium.org>
3045
3046         Reviewed by Adam Barth.
3047
3048         nrwt: remove --use-apache from the command line
3049         https://bugs.webkit.org/show_bug.cgi?id=63358
3050
3051         This change removes the --use-apache command line argument. It
3052         was initially put in when we were trying to get the cygwin
3053         apache instance to work with Chromium win, but that code has
3054         bitrotted and doesn't work at all now.
3055
3056         Arguably we should remove all of the code to allow for a choice
3057         of web servers, but since we may still want to switch off of
3058         LigHTTPd at some point on Windows, I'll leave the rest in for
3059         now (or at least yank it in a different change.)
3060
3061         * Scripts/webkitpy/layout_tests/port/base.py:
3062         * Scripts/webkitpy/layout_tests/port/chromium.py:
3063         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
3064         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
3065         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
3066         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3067
3068 2011-07-06  Dimitri Glazkov  <dglazkov@chromium.org>
3069
3070         garden-o-matic should use a favicon to indicate current state.
3071         https://bugs.webkit.org/show_bug.cgi?id=64027
3072
3073         Reviewed by Adam Barth.
3074
3075         * Scripts/webkitpy/tool/servers/data/gardeningserver/favicon-green.png: Added.
3076         * Scripts/webkitpy/tool/servers/data/gardeningserver/favicon-red.png: Added.
3077         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html: Added blurb about icons.
3078         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js: Added setting icon state.
3079         * Scripts/webkitpy/tool/servers/gardeningserver.py: Allowed the use of favicons.
3080
3081 2011-07-06  Daniel Bates  <dbates@rim.com>
3082
3083         Fix webkitpy error after changeset 90483 <http://trac.webkit.org/changeset/90483>; Add missing
3084         parameter to _fetch_file_from_results lamdba function.
3085
3086         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
3087
3088 2011-07-06  Kevin Ollivier  <kevino@theolliviers.com>
3089
3090         [wx] Unreviewed build fix. Generate the proper install_name for dylibs on Mac.
3091
3092         * wx/packaging/build-mac-installer.py:
3093
3094 2011-07-06  Xan Lopez  <xlopez@igalia.com>
3095
3096         [GTK] Only set env. variables in start when strictly needed
3097         https://bugs.webkit.org/show_bug.cgi?id=64026
3098
3099         Reviewed by Eric Seidel.
3100
3101         * Scripts/webkitpy/layout_tests/port/gtk.py: add a
3102         setup_environ_for_server method in GtkPort and set as many
3103         env. variables as possible there.
3104
3105 2011-07-06  Adam Roben  <aroben@apple.com>
3106
3107         Teach TestFailures how to load, parse, and interpret NRWT test results
3108
3109         Fixes <http://webkit.org/b/61877> TestFailures page doesn't show testers that use
3110         new-run-webkit-tests
3111
3112         Reviewed by Adam Barth.
3113
3114         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
3115         (Builder.prototype.failureDiagnosisTextAndURL): Added support for the new 'flaky' failure
3116         type. For now we don't account for image-only flakes (but TestFailures doesn't deal with
3117         pixel tests at all currently).
3118         (Builder.prototype.getNumberOfFailingTests): Relaxed the regex used for parsing the number
3119         of failing tests from the buildbot output. Make sure not to count "new passes" (a new
3120         category introduced by NRWT) as failures.
3121
3122         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
3123         (LayoutTestResultsLoader.prototype.start): Moved a bunch of code from here to
3124         _fetchAndParseORWTResults. This function now attempts to load NRWT results, then falls back
3125         to loading ORWT results.
3126         (LayoutTestResultsLoader.prototype._fetchAndParseNRWTResults): Added. Tries to load and
3127         parse the full_results.json file for the given build.
3128         (LayoutTestResultsLoader.prototype._fetchAndParseORWTResults): Added. Code came from start.
3129         Fixed a bug along the way where we were sometimes calling the error callback instead of the
3130         success callback when all tests passed.
3131
3132         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/NRWTResultsParser.js: Added.
3133         (NRWTResultsParser): Do-nothing constructor.
3134         (NRWTResultsParser.prototype.parse): Uses eval() (eww!) to get the test results out of the
3135         JS string, then iterates over all the tests in the results data and builds up a data
3136         structure matching what ORWTResultsParser returns.
3137
3138         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
3139         (Array.prototype.contains):
3140         (String.prototype.contains):
3141         Added these simple helper functions.
3142
3143         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/index.html: Pull in
3144         NRWTResultsParser.js.
3145
3146 2011-07-06  Adam Roben  <aroben@apple.com>
3147
3148         Extract code to parse ORWT's results.html file into its own class
3149
3150         Prep work for fixing <http://webkit.org/b/61877> TestFailures page doesn't show testers that
3151         use new-run-webkit-tests
3152
3153         Reviewed by Adam Barth.
3154
3155         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
3156         (LayoutTestResultsLoader.prototype.start): Moved a bunch of code from here...
3157
3158         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ORWTResultsParser.js: Added.
3159         (ORWTResultsParser):
3160         (ORWTResultsParser.prototype.parse):
3161         ...to here.
3162
3163         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/index.html: Pull in
3164         ORWTResultsParser.js.
3165
3166 2011-07-06  Eric Seidel  <eric@webkit.org>
3167
3168         Unreviewed.  Fix a unittest from hitting the network. :)
3169
3170         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
3171
3172 2011-07-06  Eric Seidel  <eric@webkit.org>
3173
3174         Unreviewed.  Fixing an exception in test-webkitpy, caused by bug 64006.
3175
3176         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
3177
3178 2011-07-06  Gustavo Noronha Silva  <gustavo.noronha@collabora.com>
3179
3180         Run an Xvfb instance to run the API tests on, like we do for the
3181         layout tests.
3182
3183         Reviewed by Xan Lopez.
3184
3185         * Scripts/run-gtk-tests:
3186
3187 2011-07-06  Daniel Bates  <dbates@webkit.org>
3188
3189         Make SCM unit tests faster
3190         https://bugs.webkit.org/show_bug.cgi?id=63883
3191
3192         Speeds up the SCM unit tests by a factor of 4. Currently, we create a
3193         mock SVN repo for each test_ method in SVNTest and GitTest and creating
3194         this repo is expensive.
3195
3196         Instead, it is significantly faster to create the mock SVN repo once
3197         and then perform a filesystem copy of it for each test_ method.
3198
3199         Note, Python 2.7's unittest module implements support for per class and
3200         per module setup and tear down methods which could be used to implement
3201         similar functionality. At the time of writing, test-webkitpy is designed
3202         to support Python 2.5. So, we can't take advantage of these Python 2.7
3203         features :(
3204
3205         Reviewed by Eric Seidel.
3206
3207         * Scripts/webkitpy/common/checkout/scm/scm_unittest.py:
3208
3209 2011-07-06  Adam Barth  <abarth@webkit.org>
3210
3211         Remove unused parameter from _add_test in test_expectations.py
3212         https://bugs.webkit.org/show_bug.cgi?id=64008
3213
3214         Reviewed by Eric Seidel.
3215
3216         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
3217
3218 2011-07-06  Adam Barth  <abarth@webkit.org>
3219
3220         Use old-run-webkit-tests when --leaks is present
3221         https://bugs.webkit.org/show_bug.cgi?id=64012
3222
3223         Reviewed by Adam Roben.
3224
3225         We'll remove this branch once support of --leaks is added to
3226         new-run-webkit-tests.
3227
3228         * Scripts/run-webkit-tests:
3229
3230 2011-07-06  Benjamin Poulain  <benjamin@webkit.org>
3231
3232         [Qt] [WK2] Add touch mocking to Qt's MiniBrowser
3233         https://bugs.webkit.org/show_bug.cgi?id=63995
3234
3235         Reviewed by Andreas Kling.
3236
3237         The mouse events are intercepted through QCoreApplication::notify(),
3238         and fake touch events are generated for mouse events when necessary.
3239
3240         If touch events are received from the platform, we assume the current
3241         hardware has a touch screen and we stop generating fake events.
3242
3243         * MiniBrowser/qt/MiniBrowserApplication.cpp:
3244         (isTouchEvent):
3245         (isMouseEvent):
3246         (MiniBrowserApplication::MiniBrowserApplication):
3247         (MiniBrowserApplication::notify):
3248         * MiniBrowser/qt/MiniBrowserApplication.h:
3249
3250 2011-07-06  Xan Lopez  <xlopez@igalia.com>
3251
3252         Reviewed by Gustavo Noronha.
3253
3254         * Scripts/webkitpy/layout_tests/port/gtk.py: use Popen correctly.
3255
3256 2011-07-06  Eric Seidel  <eric@webkit.org>
3257
3258         webkit-patch failure-reason does not understand NRWT results
3259         https://bugs.webkit.org/show_bug.cgi?id=64006
3260
3261         Reviewed by Adam Barth.
3262
3263         Adam Barth tells me failure-reason should be deleted soon,
3264         but aroben's fancy new TestFailures/ page doesn't work for
3265         NRWT yet, and I needed to know when
3266         third-party-cookie-relaxing started failing.
3267
3268         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
3269         * Scripts/webkitpy/tool/commands/queries.py:
3270
3271 2011-07-06  Xan Lopez  <xlopez@igalia.com>
3272
3273         [GTK] Add missing environment variables in NWRT
3274         https://bugs.webkit.org/show_bug.cgi?id=64004
3275
3276         Reviewed by Gustavo Noronha Silva.
3277
3278         Add missing environment variables needed by the GTK+ port.
3279
3280         * Scripts/webkitpy/layout_tests/port/gtk.py: ditto.
3281
3282 2011-07-06  Adam Roben  <aroben@apple.com>
3283
3284         Make run-api-tests output intelligible on the bots
3285
3286         run-api-tests's output on the bots was very hard to read for two reasons:
3287           1) It was mixing its own output with gtest's
3288           2) It was using ANSI escape sequences to print colored text, but the bots don't support
3289              that
3290
3291         Now, the --verbose flag turns off almost all of run-api-tests's own output so that gtest's
3292         output will not be obscured. We still print "Timeout" messages even in verbose mode, since
3293         gtest doesn't have any native support for timeouts. Also, when our output is being
3294         redirected to a file, we don't print ANSI escape sequences.
3295
3296         Fixes <http://webkit.org/b/63996> It's very hard to read run-api-tests output on the bots
3297
3298         Reviewed by Sam Weinig.
3299
3300         * Scripts/run-api-tests: Don't use the :constants interface of Term::ANSIColor. We want to
3301         use the colored function instead.
3302         (runAllTestsInSuite): Don't print out suite names in verbose mode; gtest will give us enough
3303         context that they aren't needed.
3304         (runTest): Don't print out test names or pass/fail messages in verbose mode; gtest will do
3305         that for us. Also, use the new possiblyColored function instead of always coloring output.
3306         (possiblyColored): Added. When printing to a tty, returns the string with the appropriate
3307         ANSI color escape sequences added. Otherwise just returns the string unmodified.
3308
3309 2011-07-06  Xan Lopez  <xlopez@igalia.com>
3310
3311         Reviewed by Gustavo Noronha.
3312
3313         Launch Xvfb (mostly) the same way we were launching it in the
3314         bots, for maximum compatibility.
3315
3316         * Scripts/webkitpy/layout_tests/port/gtk.py:
3317
3318 2011-07-06  Andreas Kling  <kling@webkit.org>
3319
3320         [Qt][WK2] Split Qt API into two different web views (touch and desktop)
3321         https://bugs.webkit.org/show_bug.cgi?id=63950
3322
3323         Co-authored with Benjamin Poulain <benjamin@webkit.org>.
3324
3325         Reviewed by Kenneth Rohde Christiansen.
3326
3327         Port WTR/Qt to use QDesktopWebView and force a 800x600 viewport.
3328
3329         The Qt MiniBrowser now defaults to QDesktopWebView but can be started
3330         in QTouchWebView mode by passing -touch on the command line.
3331
3332         A lot of functionality has been removed and/or disabled from
3333         MiniBrowser until we reintroduce the necessary API's.
3334
3335         * MiniBrowser/qt/BrowserView.cpp:
3336         (BrowserView::BrowserView):
3337         (BrowserView::~BrowserView):
3338         (BrowserView::load):
3339         (BrowserView::view):
3340         (BrowserView::touchWebView):
3341         (BrowserView::desktopWebView):
3342         * MiniBrowser/qt/BrowserView.h:
3343         * MiniBrowser/qt/BrowserWindow.cpp:
3344         (newPageFunction):
3345         (BrowserWindow::BrowserWindow):
3346         (BrowserWindow::webView):
3347         (BrowserWindow::newWindow):
3348         (BrowserWindow::showUserAgentDialog):
3349         (BrowserWindow::updateUserAgentList):
3350         * MiniBrowser/qt/BrowserWindow.h:
3351         * MiniBrowser/qt/MiniBrowserApplication.cpp:
3352         (MiniBrowserApplication::handleUserOptions):
3353         * MiniBrowser/qt/MiniBrowserApplication.h:
3354         (WindowOptions::WindowOptions):
3355         * MiniBrowser/qt/UrlLoader.cpp:
3356         (UrlLoader::UrlLoader):
3357         * MiniBrowser/qt/main.cpp:
3358         (main):
3359         * WebKitTestRunner/qt/PlatformWebViewQt.cpp:
3360         (WTR::WebView::wkView):
3361         (WTR::WebView::pageRef):
3362         (WTR::WebView::WebView):
3363         (WTR::PlatformWebView::page):
3364         (WTR::PlatformWebView::windowFrame):
3365         (WTR::PlatformWebView::setWindowFrame):
3366
3367 2011-07-06  Eric Seidel  <eric@webkit.org>
3368
3369         Unreviewed..
3370
3371         * Scripts/run-webkit-tests:
3372         (useNewRunWebKitTests):
3373          - Fix the pass-thru of --gtk and --qt for old-run-webkit-tests,
3374            it was previously broken by the existence of isGtk/isQt checks which
3375            removed the --gtk/--qt args from @ARGV.
3376
3377 2011-07-06  Eric Seidel  <eric@webkit.org>
3378
3379         new-run-webkit-tests uses a 35s timeout for Mac and 6s for all other ports
3380         https://bugs.webkit.org/show_bug.cgi?id=63983
3381
3382         Reviewed by Xan Lopez.
3383
3384         All WebKit ports want to use a 35s timeout.  Only chromium currently
3385         uses a 6s timeout.  Eventually we'll move all WebKit ports to 6s too
3386         to further speed up new-run-webkit-tests runs, but that's going to be a while.
3387
3388         * Scripts/webkitpy/layout_tests/port/mac.py:
3389         * Scripts/webkitpy/layout_tests/port/webkit.py:
3390         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
3391
3392 2011-07-06  Xan Lopez  <xlopez@igalia.com>
3393
3394         Unreviewed.
3395
3396         NWRT actually seems to work locally, so it should work in the bots
3397         as long as we take care of the TIMEOUT issue (which is being fixed
3398         as I write this). Let's give this another shot and fix whatever
3399         breaks.
3400
3401         * Scripts/run-webkit-tests:
3402         (useNewRunWebKitTests):
3403
3404 2011-07-06  Xan Lopez  <xlopez@igalia.com>
3405
3406         Unreviewed.
3407
3408         Disable NWRT on GTK. At the very least it needs an updated exected
3409         results file since NWRT is more sensitive to timeouts, not sure
3410         what the exact problem is.
3411
3412         * Scripts/run-webkit-tests:
3413         (useNewRunWebKitTests):
3414
3415 2011-07-06  Adam Barth  <abarth@webkit.org>
3416
3417         Fix global variable leak detected by noglobals.
3418
3419         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
3420
3421 2011-07-06  Adam Barth  <abarth@webkit.org>
3422
3423         garden-o-matic shouldn't show flaky tests by default
3424         https://bugs.webkit.org/show_bug.cgi?id=63979
3425
3426         Reviewed by Ojan Vafai.
3427
3428         If any of the actual test results are PASS, the test is flaky and we
3429         can ignore it (at least for now).  If we get awesome at dealing with
3430         failing tests, we might want to tackle flaky tests by refering the user
3431         to the flakiness dashboard, which is way more awesome at dealing with
3432         flaky tests that this tool ever will be.
3433
3434         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
3435         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
3436
3437 2011-07-05  Adam Barth  <abarth@webkit.org>
3438
3439         garden-o-matic should list the failing tests
3440         https://bugs.webkit.org/show_bug.cgi?id=63978
3441
3442         Reviewed by Eric Seidel.
3443
3444         * Scripts/webkitpy/tool/servers/data/gardeningserver/config.js: Added.
3445         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
3446         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
3447         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
3448         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
3449         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html:
3450         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js: Added.
3451         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js: Added.
3452         * Scripts/webkitpy/tool/servers/gardeningserver.py:
3453
3454
3455 2011-07-05  Eric Seidel  <eric@webkit.org>
3456
3457         Re-enable NRWT on Leopard now that bug 63973 is fixed.
3458
3459         Reviewed by Adam Barth.
3460
3461         I also made the should-enable expression a black-list instead
3462         of a white-list since every bot on build.webkit.org is
3463         now using NRWT except windows and WebKit2 bots.
3464
3465         * Scripts/run-webkit-tests:
3466         (useNewRunWebKitTests):
3467
3468 2011-07-05  Eric Seidel  <eric@webkit.org>
3469
3470         Perl string concat failure on Leopard (causes NRWT to fail)
3471         https://bugs.webkit.org/show_bug.cgi?id=63973
3472
3473         Reviewed by Daniel Bates.
3474
3475         Leopard's "xcodebuild -version" has a different format from Snow Leopard/Lion
3476         thus the version parsing logic in webkitdirs.pm seems to fail.
3477
3478         Default to "3.0" instead of undef when we can't parse the version number.
3479         This should send us down the correct path for Leopard.
3480
3481         * Scripts/webkitdirs.pm:
3482         (determineBaseProductDir):
3483
3484 2011-07-05  Kevin Ollivier  <kevino@theolliviers.com>
3485
3486         [wx] Unreviewed build fix after addition of Source/WebCore/xml/parser dir.
3487
3488         * waf/build/settings.py:
3489
3490 2011-07-05  Eric Seidel  <eric@webkit.org>
3491
3492         Disable NRWT for Leopard until https://bugs.webkit.org/show_bug.cgi?id=63973 can be fixed.
3493
3494         Reviewed by Adam Barth.
3495
3496         * Scripts/run-webkit-tests:
3497         (useNewRunWebKitTests):
3498
3499 2011-07-05  Eric Seidel  <eric@webkit.org>
3500
3501         Make run-webkit-tests pass --verbose to new-run-webkit-tests on qt/gtk bots.
3502
3503         Reviewed by Adam Barth.
3504
3505         * Scripts/run-webkit-tests:
3506         (runningOnBuildBot):
3507
3508 2011-07-05  Adam Barth  <abarth@webkit.org>
3509
3510         Add results_unittests.js to garden-o-matic
3511         https://bugs.webkit.org/show_bug.cgi?id=63971
3512
3513         Reviewed by Eric Seidel.
3514
3515         Adding these unit tests forced me to commit to an API for the results
3516         module.  I'm not sure if this is the API we'll end up with, but it's
3517         something not entirely unreasonable.
3518
3519         Also, I've changed results.js over to WebKit style, as discussed previously.
3520
3521         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
3522         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
3523         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js: Added.
3524         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html:
3525
3526 2011-07-05  Eric Seidel  <eric@webkit.org>
3527
3528         Teach run-webkit-tests how to translate --qt and --gtk for new-run-webkit-tests.
3529
3530         Reviewed by Adam Barth.
3531
3532         * Scripts/run-webkit-tests:
3533
3534 2011-07-05  Adam Barth  <abarth@webkit.org>
3535
3536         Add first unit tests for garden-o-matic
3537         https://bugs.webkit.org/show_bug.cgi?id=63969
3538
3539         Reviewed by Eric Seidel.
3540
3541         I've also reformated base.js to match WebKit style, as discussed in the
3542         previous bug.
3543
3544         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
3545         * Scripts/webkitpy/tool/servers/data/gardeningserver/base_unittests.js: Added.
3546         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html: Added.
3547
3548 2011-07-05  Eric Seidel  <eric@webkit.org>
3549
3550         Move Qt, Gtk and Leopard to NRWT.
3551
3552         We'll roll out any of these if we see problems.
3553
3554         Reviewed by Adam Barth.
3555
3556         * Scripts/run-webkit-tests:
3557         (useNewRunWebKitTests):
3558
3559 2011-07-05  Adam Barth  <abarth@webkit.org>
3560
3561         garden-o-matic should understand which tests have unexpected failures
3562         https://bugs.webkit.org/show_bug.cgi?id=63965
3563
3564         Reviewed by Eri