44e43fa96a0855d33d8a3d0ee7880532d8dafa8d
[WebKit-https.git] / Tools / ChangeLog
1 2011-07-13  Dimitri Glazkov  <dglazkov@chromium.org>
2
3         Eliminate TestExpectationsFile.
4         https://bugs.webkit.org/show_bug.cgi?id=64458
5
6         Turns out, we can just use a Python list.
7
8         Reviewed by Adam Barth.
9
10         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Folded TestExpectationsFile.append into TestExpectationParser.parse_list,
11             removed TestExpectationsFile.
12         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: Moved tests to better reflect new names, removed iterator test,
13             since there's no more custom iterator machinery.
14
15 2011-07-13  Philippe Normand  <pnormand@igalia.com>
16
17         Unreviewed, added my other email addresses.
18
19         * Scripts/webkitpy/common/config/committers.py:
20
21 2011-07-13  Adam Barth  <abarth@webkit.org>
22
23         Folks should only be listed once.
24
25         * Scripts/webkitpy/common/config/committers.py:
26
27 2011-07-13  Philippe Normand  <pnormand@igalia.com>
28
29         Unreviewed, adding myself as Reviewer.
30
31         * Scripts/webkitpy/common/config/committers.py:
32
33 2011-07-13  Adam Barth  <abarth@webkit.org>
34
35         Fix flickering bug introduced by my previous patch.  I forgot to change
36         the name of the class everywhere.
37
38         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
39
40 2011-07-12  Philippe Normand  <pnormand@igalia.com>
41
42         [GTK] media/media-blocked-by-willsendrequest.html fails
43         https://bugs.webkit.org/show_bug.cgi?id=63699
44
45         Reviewed by Martin Robinson.
46
47         * DumpRenderTree/gtk/DumpRenderTree.cpp:
48         (willSendRequestCallback): Abort the request if explicitely asked
49         by the LayoutTestController's willSendRequestReturnsNull() function.
50
51 2011-07-13  Adam Barth  <abarth@webkit.org>
52
53         Remove "Dismiss" button from garden-o-matic butterbar
54         https://bugs.webkit.org/show_bug.cgi?id=64443
55
56         Reviewed by Dimitri Glazkov.
57
58         We don't have any persistent butter bar messages yet, so the dismiss
59         button is premature (and fairly heavy, visually).
60
61         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
62         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
63
64 2011-07-13  Adam Barth  <abarth@webkit.org>
65
66         Refine garden-o-matic status pane
67         https://bugs.webkit.org/show_bug.cgi?id=64442
68
69         Reviewed by Dimitri Glazkov.
70
71         This patch makes two improvements to the status pane:
72
73         1) We only query the server for failure types that we're expecting.
74            This dramatically reduces the number of HTTP requests, making
75            loading the status pane faster.
76
77         2) The status pane now displays which test and which builder it is
78            showing results for.
79
80         * Scripts/webkitpy/tool/servers/data/gardeningserver/config.js:
81         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
82         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
83         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
84         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
85         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
86         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
87
88 2011-07-13  Adam Barth  <abarth@webkit.org>
89
90         garden-o-matic should show test results
91         https://bugs.webkit.org/show_bug.cgi?id=64440
92
93         Reviewed by Adam Roben.
94
95         This patch introduces the results detail pane, which appears at the
96         bottom of the window and contains test results from the bots.
97         Currently, you can activate the pane by mousing over one of the builder
98         names associated with a failing test.
99
100         This is just a first iteration of the UI.  There's no way to resize or
101         zoom in on elements of the details pane, and images likely aren't sized
102         correctly, but it's a place to start.
103
104         * Scripts/webkitpy/tool/servers/data/gardeningserver/config.js:
105             - Add constants for our data attributes so we don't typo them!
106         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
107             - Add DOM structure for the details pane.
108         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
109         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
110             - Bind events for showing and hiding the details pane.
111         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
112             - Reduce the number of result types that we query for to avoid
113               overfilling the details pane.  We'll probably need another
114               solution here in the long-term.  For example, we could use tabs
115               to pack more results into the pane.
116         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
117         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html:
118             - Now that we're using the config package during testing, we need
119               to include it in the testing HTML.
120         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
121         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
122             - Change the DOM structure of the results detail to make it fit
123               nicely in the details pane (rather than flowing freely in the
124               body, as it did before).
125
126 2011-07-13  Adam Roben  <aroben@apple.com>
127
128         Make TestFailures show every time a possibly-flaky test failed, but hide it by default
129
130         It's useful to be able to see every time a flaky test failed to see whether it failed the
131         same way every time. But doing so takes a lot of space, so the list of failures is now
132         collapsed by default and can be revealed using a disclosure triangle.
133
134         Fixes <http://webkit.org/b/64455> TestFailures page doesn't show as much information for
135         flaky tests as I would like, even though the page is already so long
136
137         Reviewed by Dimitri Glazkov.
138
139         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/FlakyLayoutTestDetector.js:
140         (FlakyLayoutTestDetector.prototype.allFailures): Replaced flakinessExamples with this
141         function. Now returns all failures for the given test.
142
143         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/FlakyLayoutTestDetector_unittests.js:
144         Added. This just contains some simple tests of the FlakyLayoutTestDetector class. We'll add
145         more over time.
146
147         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
148         (LayoutTestHistoryAnalyzer.prototype.start): Updated the documentation comment to reflect
149         that we no longer return passing builds for possibly-flaky tests.
150
151         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
152         (.existing-bugs-list, .suspect-revisions-list, .flakiness-examples-list): Make the list of
153         flakiness examples small, too, since it can get quite long.
154
155         (.disclosure-triangle):
156         (.expanded > .disclosure-triangle):
157         Simple styles for the disclosure triangle.
158
159         (.flakiness-examples-list): Collapse the list by default.
160
161         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
162         (ViewController.prototype._displayBuilder): Pass the total number of builds analyzed to
163         _domForPossiblyFlakyTests.
164         (ViewController.prototype._domForPossiblyFlakyTests): Put a disclosure triangle to the left
165         of each test name, and the number of failures to the right. When the disclosure triangle is
166         clicked for the first time, we build up the list of failures and expand the element. After
167         that we just collapse or expand the element on subsequent clicks.
168
169         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/run-unittests.html:
170         Pulled in new tests.
171
172 2011-07-13  Adam Roben  <aroben@apple.com>
173         
174         Teach TestFailures to understand NRWT's output when it exits early due to too many failures
175
176         Fixes <http://webkit.org/b/64456> TestFailures page reports way too many failures when NRWT
177         exits early
178
179         Reviewed by Dimitri Glazkov.
180
181         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
182         (Builder.prototype.getNumberOfFailingTests): Relaxed the "Exiting early" test not to require
183         it to be at the beginning of the line, since NRWT prints a bunch of junk earlier in the
184         line. Tightened up the regex that's used to parse the number of failing tests to require the
185         leading number to be followed by whitespace so that we won't parse the "2011" in
186         "2011-07-13" as a number of failures.
187
188         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder_unittests.js:
189         Added a test for the above.
190
191 2011-07-12  Dimitri Glazkov  <dglazkov@chromium.org>
192
193         Extract model-like TestExpectationLine and TestExpectationFile from TestExpectations.
194         https://bugs.webkit.org/show_bug.cgi?id=64386
195
196         This is the first step in converting TestExpectations to a real model.
197         * TestExpectationsLine represents a line in the test_expectations.txt file, and
198         * TestExpectationsFile represents the file, which is a collection of lines.
199
200         Reviewed by Adam Barth.
201
202         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
203         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py:
204         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
205
206 2011-07-13  Xan Lopez  <xlopez@igalia.com>
207
208         [GTK] Do not grab focus too early in DRT.
209
210         Reviewed by Gustavo Noronha.
211
212         It causes a layout to happen and a progress signal to be emitted
213         since r90900, but at this point we don't have a
214         LayoutTestController object and we'll eventually crash. Since we
215         already grab focus at the beginning of runTest() this is
216         redundant, so get rid of it to fix the crash.
217
218         * DumpRenderTree/gtk/DumpRenderTree.cpp:
219         (main): remove call to grab_focus
220
221 2011-07-13  Sheriff Bot  <webkit.review.bot@gmail.com>
222
223         Unreviewed, rolling out r90893 and r90894.
224         http://trac.webkit.org/changeset/90893
225         http://trac.webkit.org/changeset/90894
226         https://bugs.webkit.org/show_bug.cgi?id=64441
227
228         NRWT still doesn't work on qt-mac platform (Requested by Ossy
229         on #webkit).
230
231         * Scripts/run-webkit-tests:
232         (useNewRunWebKitTests):
233
234 2011-07-13  Csaba Osztrogonác  <ossy@webkit.org>
235
236         [Qt] NRWT should pick up the right httpd config file
237         https://bugs.webkit.org/show_bug.cgi?id=64086
238
239         * Scripts/run-webkit-tests: Enable NRWT on qt-mac platform after r90810.
240         (useNewRunWebKitTests):
241
242 2011-07-12  Mark Rowe  <mrowe@apple.com>
243
244         Fix the 32-bit build.
245
246         * DumpRenderTree/mac/UIDelegate.mm:
247         (-[UIDelegate webView:exceededApplicationCacheOriginQuotaForSecurityOrigin:totalSpaceNeeded:]):
248         Cast the NSUInteger value to unsigned long to match the format specifier.
249
250 2011-07-12  Adam Barth  <abarth@webkit.org>
251
252         Improve garden-o-matic UI when the bots fail to report revision numbers
253         https://bugs.webkit.org/show_bug.cgi?id=64427
254
255         Reviewed by Dimitri Glazkov.
256
257         At least Win (dbg)(2) seems to fail to report the SVN revision number
258         in full_results.json.  This patch makes garden-o-matic more robust to
259         missing revision numbers.
260
261         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
262         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
263         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
264         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
265         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
266
267 2011-07-12  Adam Barth  <abarth@webkit.org>
268
269         Adjust garden-o-matic layout to use fewer lines
270         https://bugs.webkit.org/show_bug.cgi?id=64422
271
272         Reviewed by Ojan Vafai.
273
274         This patch effectively merges the test name line with the list of
275         builders on which the test fails.  Each test failure now occupies two
276         lines instead of three.
277
278         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
279         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
280         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
281
282 2011-07-12  Adam Barth  <abarth@webkit.org>
283
284         garden-o-matic should label tests failures of new tests as such
285         https://bugs.webkit.org/show_bug.cgi?id=64421
286
287         Reviewed by Ojan Vafai.
288
289         Also, we shouldn't dim failures of new tests, even if we've only seen
290         them once because they're likely to be real problems that need
291         attention.
292
293         * Scripts/webkitpy/tool/servers/data/gardeningserver/base_unittests.js:
294         * Scripts/webkitpy/tool/servers/data/gardeningserver/checkout.js: Added.
295         * Scripts/webkitpy/tool/servers/data/gardeningserver/checkout_unittests.js: Added.
296         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
297         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
298         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
299         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
300         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html:
301         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
302         * Scripts/webkitpy/tool/servers/gardeningserver.py:
303
304 2011-07-12  Adam Barth  <abarth@webkit.org>
305
306         garden-o-matic should display how many times we've seen a failure
307         https://bugs.webkit.org/show_bug.cgi?id=64417
308
309         Reviewed by Ojan Vafai.
310
311         This patch adds some UI to display how many times we've seen a given
312         failure, which can be helpful for determining whether that failure is a
313         real failure or a flaky test.
314
315         When a failure has only been seen once (i.e., only a single run on a
316         single bot), we set the opacity of to 50% to avoid distracting the
317         gardener.
318
319         This patch also refactors the failure walker to have a simpler API
320         internally by moving from an object-oriented paradigm to a functional
321         paradigm.
322
323         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
324         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
325         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
326         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
327         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
328         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
329
330 2011-07-12  Chris Rogers  <crogers@google.com>
331
332         Enable Web Audio for chromium DRT
333         https://bugs.webkit.org/show_bug.cgi?id=64409
334
335         Reviewed by James Robinson.
336
337         * DumpRenderTree/chromium/TestShell.cpp:
338         (TestShell::TestShell):
339
340 2011-07-12  Adam Barth  <abarth@webkit.org>
341
342         Tweak some UI in garden-o-matic now that the tree actually has a
343         failure and I can see what this all looks like.
344
345         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
346         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
347         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
348
349 2011-07-12  Joseph Pecoraro  <joepeck@webkit.org>
350
351         Unreviewed. Skipping a few tests which fail due to differing output
352         in recent ApplicationCache quota tests. Reenabling the tests is
353         tracked by <http://webkit.org/b/64410>.
354
355         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
356         (WebCore::DumpRenderTree::dumpApplicationCacheQuota): This should
357         have been a space, not a newline.
358
359 2011-07-12  Adam Barth  <abarth@webkit.org>
360
361         garden-o-matic should display regression ranges
362         https://bugs.webkit.org/show_bug.cgi?id=64407
363
364         Reviewed by Dimitri Glazkov.
365
366         This patch computes a regression range for a failure by intersecting
367         the regression ranges seen by the various bots.  We make the underlying
368         assumption that a test is only failing due to one revision at any given
369         moment.  If that's not true, this code probably explodes.
370
371         The regression ranges appear asynchronously, which might be a jarring
372         UI.  We'll have to experiment to see.
373
374         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
375         * Scripts/webkitpy/tool/servers/data/gardeningserver/base_unittests.js:
376         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
377         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
378         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
379         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
380         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
381
382 2011-07-12  Adam Barth  <abarth@webkit.org>
383
384         cr-linux-ews should run pixel tests
385         https://bugs.webkit.org/show_bug.cgi?id=64394
386
387         Reviewed by Ojan Vafai.
388
389         This patch enabled pixel test on the cr-linux-ews.  Running a quick
390         test on the bots, it looks like we have five pixel failures, related to
391         some missing fonts.  I've also made the bots skip failing tests to
392         avoid uploading huge zip files to bugs.webkit.org (with all the exected
393         IMAGE failures).
394
395         * Scripts/webkitpy/common/config/ports.py:
396         * Scripts/webkitpy/common/config/ports_unittest.py:
397
398 2011-07-12  Joseph Pecoraro  <joepeck@webkit.org>
399
400         ApplicationCache update should not immediately fail when reaching per-origin quota
401         https://bugs.webkit.org/show_bug.cgi?id=64177
402
403         Reviewed by Alexey Proskuryakov.
404
405         Add layoutTestController.disallowIncreaseForApplicationCacheQuota()
406         to disable the default behavior of raising the default per-origin
407         quota to 5MB when reached. This allows us to test what happens
408         when the quota is reached and not increased.
409
410         * DumpRenderTree/LayoutTestController.cpp:
411         (LayoutTestController::LayoutTestController):
412         (disallowIncreaseForApplicationCacheQuotaCallback):
413         (LayoutTestController::staticFunctions):
414         * DumpRenderTree/LayoutTestController.h:
415         (LayoutTestController::disallowIncreaseForApplicationCacheQuota):
416         (LayoutTestController::setDisallowIncreaseForApplicationCacheQuota):
417         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
418         (LayoutTestController::reset):
419         * DumpRenderTree/qt/LayoutTestControllerQt.h:
420         (LayoutTestController::disallowIncreaseForApplicationCacheQuota):
421         (LayoutTestController::dumpApplicationCacheDelegateCallbacks):
422         Expose a function to prevent automatically increasing the per-origin
423         quota when the quota is reached. This allows us to test what happens
424         when a user would "disallow" a quota increase.
425
426         * DumpRenderTree/mac/UIDelegate.mm:
427         (-[UIDelegate webView:exceededApplicationCacheOriginQuotaForSecurityOrigin:spaceNeeded:]):
428         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
429         (WebCore::DumpRenderTree::dumpApplicationCacheQuota):
430         Respect the new disallowIncrease flag and dump the spaceNeeded values
431         for ports that implement application cache quotas.
432
433 2011-07-12  Ojan Vafai  <ojan@chromium.org>
434
435         fix flakiness dashboard to work with new crash log filenames
436         https://bugs.webkit.org/show_bug.cgi?id=64393
437
438         Reviewed by Adam Barth.
439
440         * TestResultServer/static-dashboards/flakiness_dashboard.html:
441
442 2011-07-12  Adam Barth  <abarth@webkit.org>
443
444         Add a happy animation to garden-o-matic when there are no failures
445         https://bugs.webkit.org/show_bug.cgi?id=64382
446
447         Reviewed by Ojan Vafai.
448
449         As requested by Ojan.
450
451         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css:
452         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
453         * Scripts/webkitpy/tool/servers/data/gardeningserver/partytime.gif: Added.
454         * Scripts/webkitpy/tool/servers/gardeningserver.py:
455
456 2011-07-12  Adam Barth  <abarth@webkit.org>
457
458         sheriffbot can't perform rollouts
459         https://bugs.webkit.org/show_bug.cgi?id=64370
460
461         Reviewed by Eric Seidel.
462
463         Now that we're not updating the working copy before processing every
464         IRC message, the rollout command needs to update it manually.
465
466         * Scripts/webkitpy/tool/bot/irc_command.py:
467         * Scripts/webkitpy/tool/bot/irc_command_unittest.py:
468
469 2011-07-12  Adam Barth  <abarth@webkit.org>
470
471         Fix the grammar in the FIXME comment.
472
473         * Scripts/webkitpy/common/net/resultsjsonparser.py:
474
475 2011-07-12  Eric Seidel  <eric@webkit.org>
476
477         NRWT should open test results page with Safari trunk, not the system provided one on Mac
478         https://bugs.webkit.org/show_bug.cgi?id=64346
479
480         Reviewed by Adam Barth.
481
482         To fix this I implemented Port.show_results_html_file in Mac, Gtk and Qt ports with
483         implementations (mostly) matching those found in old-run-webkit-tests.
484         There are still some minor differences for Qt which Qt hackers may wish to tweak.
485
486         I had to add a WebKitPort._port_flag_for_scripts method (similar to flag() in common.config.ports.py)
487         for the Qt/Gtk ports which always require a flag passed to scripts.
488
489         While trying to test this, I found that FactoryTest.test_chromium_gpu_linux
490         was using a real filesystem (due to assert_platform) which due to
491         global variables in config.py was causing set-webkit-configuration to have
492         an affect on unit test results!  So I fixed this by making FactoryTest.assert_port
493         pass a mock file system whenever calling factory.get.
494
495         Unfortunately TestPort was depending on always being passed a None filesystem
496         and asserting filesystem._tests (only true for unit_test_filesystem()).
497         So I just removed the FactoryTest.test_test and FactoryTest.test_dryrun tests
498         deciding that they were pretty much useless anyway.  If others feel strongly
499         I'm happy to fix this in a different way.
500
501         * Scripts/webkitpy/common/system/executive.py:
502          - default arguments in python are screwy.  They use a single shared
503            instance, so it's better to use argument=None and then argument = argument or Default()
504            if you have any chance of mutating (or returning) the default argument.
505         * Scripts/webkitpy/layout_tests/port/config.py:
506          - This code is wrong.  We don't need to use a global variable here (as far as I can tell).
507            I'm not fixing it in this patch, but I've marked it with a FIXME and we can convert to
508            storing the results of the read on the Config object (which should only be created once during normal operation).
509            Unit tests shouldn't be hitting the disk anyway.  It's possible Config should move off of Port and onto Tool/Host directly.
510         * Scripts/webkitpy/layout_tests/port/factory.py:
511         * Scripts/webkitpy/layout_tests/port/factory_unittest.py:
512         * Scripts/webkitpy/layout_tests/port/gtk.py:
513         * Scripts/webkitpy/layout_tests/port/gtk_unittest.py: Added.
514         * Scripts/webkitpy/layout_tests/port/mac.py:
515         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
516         * Scripts/webkitpy/layout_tests/port/qt.py:
517         * Scripts/webkitpy/layout_tests/port/qt_unittest.py:
518         * Scripts/webkitpy/layout_tests/port/webkit.py:
519
520 2011-07-12  Adam Roben  <aroben@apple.com>
521
522         Teach TestFailures to recognize when run-webkit-tests gets killed by buildbot
523
524         Fixes <http://webkit.org/b/64358> TestFailures page thinks all tests passed in
525         http://build.webkit.org/builders/Windows%207%20Release%20(Tests)/builds/14672
526
527         Reviewed by Daniel Bates.
528
529         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
530         (Builder.prototype.getNumberOfFailingTests): If run-webkit-tests exited with a non-zero
531         exit status but we didn't find any failure counts, assume that there was some error that
532         caused run-webkit-tests to die early (like being killed by buildbot due to a timeout).
533
534         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder_unittests.js:
535         Added a new test that shows that we get a failingTestCount of -1 when run-webkit-tests dies
536         early.
537         (runGetNumberOfFailingTestsTest): Moved most code here from the only pre-existing test in this
538         file.
539
540         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
541         (LayoutTestResultsLoader.prototype.start): Bump the cache number so old cached data that was
542         tainted by the bug fixed in this patch will be evicted.
543
544 2011-07-12  Adam Barth  <abarth@webkit.org>
545
546         cr-linux-ews complains about tests that aren't actually failing
547         https://bugs.webkit.org/show_bug.cgi?id=64367
548
549         The underlying problem here is that full_results.json doesn't have
550         enough information to determine whether a given test result was
551         expected because whether an actual result is expected depends on
552         whether full_results.json was generated duing a test run that included
553         pixel tests.
554
555         The right long-term solution is to make full_results.json a complete
556         description of what happened durning a test run.  In this patch
557         however, to stop the spam, this patch makes the jsonresultsparser
558         ignore pixel failures.
559
560         (I'm landing this patch unreviewed in order to stop the bots from
561         spamming.  I'll happily address any review comments in a follow-up
562         patch.)
563
564         * Scripts/webkitpy/common/net/resultsjsonparser.py:
565         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
566
567 2011-07-12  Adam Barth  <abarth@webkit.org>
568
569         Re-work garden-o-matic UI to begin to look like the mocks
570         https://bugs.webkit.org/show_bug.cgi?id=64334
571
572         Reviewed by Dimitri Glazkov.
573
574         This patch changes how we display failures to be more like the mock the
575         dglazkov made. We now can display the same information much more
576         compactly.
577
578         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
579         * Scripts/webkitpy/tool/servers/data/gardeningserver/base_unittests.js:
580         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
581         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.css: Added.
582         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
583         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
584         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
585         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
586         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
587         * Scripts/webkitpy/tool/servers/gardeningserver.py:
588
589 2011-07-11  Adam Roben  <aroben@apple.com>
590
591         Extract TestFailures's bug-filing code into two new classes
592
593         Fixes <http://webkit.org/b/64300> TestFailures page's new-bug-filing code is a mess!
594
595         Reviewed by Darin Adler and Adam Barth.
596
597         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Buildbot.js:
598         (Buildbot.prototype.resultsDirectoryURL): Changed to return a URI-encoded URL. Otherwise the
599         URL isn't valid (and it's harder to mock this function).
600
601         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Buildbot_unittests.js:
602         Added. Just tests the above change (for now).
603
604         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder_unittests.js:
605         Added a license header, enclosed everything in a closure to avoid polluting the global
606         namespace, changed the test name to actually describe the passing condition, and replaced
607         uses of equals() with equal(). (The latter is the actual name of the function, and matches
608         deepEqual, etc.)
609
610         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/NewBugForm.js: Added.
611         (NewBugForm): This class knows how to construct a <form> used for filing new bugs in
612         Bugzilla based on some parameters.
613         (NewBugForm.prototype.domElement): Creates and returns the <form> element. Code came from
614         ViewController.prototype._domForNewAndExistingBugs.
615
616         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/NewBugForm_unittests.js:
617         Added. Tests the above code.
618
619         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailureBugForm.js: Added.
620         (TestFailureBugForm): This class knows how to construct a <form> element used for filing new
621         bugs specifically about test failures. Code came from
622         ViewController.prototype._domForNewAndExistingBugs.
623         (TestFailureBugForm.prototype.domElement): Slightly customizes the <form> element returned
624         by NewBugForm.
625         (TestFailureBugForm.prototype._computeOperatingSystem):
626         (TestFailureBugForm.prototype._computePlatform):
627         (TestFailureBugForm.prototype._createBugTitle):
628         (TestFailureBugForm.prototype._failingResultsHTMLURL):
629         (TestFailureBugForm.prototype._failingRevision):
630         (TestFailureBugForm.prototype._passingRevision):
631         (TestFailureBugForm.prototype._regressionRangeString):
632         Code came from ViewController.prototype._domForNewAndExistingBugs. I broke it out into
633         separate functions to break up the rat's nest a bit.
634
635         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailureBugForm_unittests.js:
636         Added. Tests the above code.
637
638         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
639         (ViewController.prototype._domForNewAndExistingBugs): Moved code from here to
640         TestFailureBugForm/NewBugForm, and changed this code to use a TestFailureBugForm.
641
642         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/index.html: Pull in
643         NewBugForm/TestFailureBugForm.js.
644
645         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/run-unittests.html:
646         Added new tests and required files.
647
648 2011-07-12  Adam Roben  <aroben@apple.com>
649
650         Test that no intermediate WTF::Strings are created when concatenating with string literals
651
652         Test for <http://webkit.org/b/63330> Concatenating string literals and WTF::Strings using
653         operator+ is suboptimal
654
655         Reviewed by Darin Adler.
656
657         * TestWebKitAPI/Tests/WTF/StringOperators.cpp: Added.
658         (TestWebKitAPI::TEST): Test that a bunch of different string concatenation expressions don't
659         create any intermediate WTF::Strings while they're being evaluated.
660
661         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
662         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
663         Added new file.
664
665 2011-07-12  Eric Seidel  <eric@webkit.org>
666
667         [Qt] NRWT should pick up the right httpd config file
668         https://bugs.webkit.org/show_bug.cgi?id=64086
669
670         Reviewed by Adam Barth.
671
672         This is more fall-out from the Port class inappropriately encapsulating
673         both platform and port knowledge.  We need to split out some of this
674         platform knowledge into a separate class which can be better shared
675         between ports.
676
677         The fix was to move all the _path_to_apache_config_file logic down
678         into the WebKitPort baseclass so that all of the ports can find the
679         right apache config file, regardless of what platform they may be running on.
680
681         I did not move this down into base.Port (even though I considered it).
682         Chromium duplicates some of this logic, but since they have separate
683         subclasses for each port-OS combination (e.g. ChromiumMac, ChromiumWin)
684         they wouldn't notice the change.  Eventually we'll move this logic
685         out of Port entirely, and then it will be shared by all ports.
686
687         I also cleaned up the http-config logic for ORWT while I was there,
688         although since we're killing that code, I'm happy to revert that part
689         of the change if changing it is viewed as needlessly risky.
690
691         During this cleanup, I noticed that no ports use the "default" httpd.conf
692         which is left over from Mac OS X Tiger (and old linux distros) which used
693         Apache 1.3.  I've removed httpd.conf (and associated support in ORWT) since
694         we no longer support any ports which use this httpd.conf.
695
696         * Scripts/webkitperl/httpd.pm:
697         (getHTTPDConfigPathForTestDirectory):
698         * Scripts/webkitpy/layout_tests/port/base.py:
699         * Scripts/webkitpy/layout_tests/port/gtk.py:
700         * Scripts/webkitpy/layout_tests/port/mac.py:
701         * Scripts/webkitpy/layout_tests/port/qt.py:
702         * Scripts/webkitpy/layout_tests/port/webkit.py:
703         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
704         * Scripts/webkitpy/layout_tests/port/win.py:
705         * Scripts/webkitpy/layout_tests/servers/apache_http_server.py:
706
707 2011-07-11  Hironori Bono  <hbono@chromium.org>
708
709         Reviewed by Adam Roben.
710
711         Implement layoutTestController.setTextDirection for WebKit2 and Windows.
712         https://bugs.webkit.org/show_bug.cgi?id=61931
713
714         This change implements layoutTextController.setTextDirection for WebKit2
715         and Windows so we can run a layout test added by r87770 on them. Each
716         implementation adds a wrapper function for Editor::setBaseWritingDirection()
717         so LayoutTestController can call it.
718
719         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
720         (LayoutTestController::setTextDirection): Implemented the binding function
721         for layoutTestController.setTextDirection (Windows).
722         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
723         Added a binding function for layoutTestController.setTextDirection (WebKit2).
724         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp: ditto.
725         (WTR::LayoutTestController::setTextDirection):
726         * WebKitTestRunner/InjectedBundle/LayoutTestController.h: ditto.
727
728 2011-07-11  Raphael Kubo da Costa  <kubo@profusion.mobi>
729
730         [EFL] ImageDiff: Make sure gEcoreEvas is destroyed at the right time.
731         https://bugs.webkit.org/show_bug.cgi?id=64293
732
733         Reviewed by Kent Tamura.
734
735         gEcoreEvas was being destroyed after shutdownEfl() was called, so the
736         call to ecore_evas_free() failed.
737
738         We now shut it down manually before shutdownEfl() is called.
739
740         * DumpRenderTree/efl/ImageDiff.cpp:
741         (main): Destroy gEcoreEvas before shutting down the EFL.
742
743 2011-07-11  Ryosuke Niwa  <rniwa@webkit.org>
744
745         Make Chromium Mac and Linux tests bot core builders
746         https://bugs.webkit.org/show_bug.cgi?id=63196
747
748         Reviewed by Adam Barth.
749
750         Added Chromium Mac and Linux Release Tests bots core builders.
751
752         Also rearranged the order of bots so that test bots show up right next to
753         respective builders.
754
755         * BuildSlaveSupport/build.webkit.org-config/config.json:
756         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
757         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
758
759 2011-07-11  Raphael Kubo da Costa  <kubo@profusion.mobi>
760
761         [EFL] DRT: Add code related to font management.
762         https://bugs.webkit.org/show_bug.cgi?id=63989
763
764         Reviewed by Kent Tamura.
765
766         The added files are responsible for managing fontconfig and adding the
767         fonts required by some layout tests to fontconfig's path.
768
769         * DumpRenderTree/efl/FontManagement.cpp: Added.
770         (getFontDirectories):
771         (getFontFiles):
772         (addFontDirectories):
773         (addFontFiles):
774         (addFontsToEnvironment):
775         * DumpRenderTree/efl/FontManagement.h: Added.
776
777 2011-07-08  Dirk Pranke  <dpranke@chromium.org>
778
779         nrwt: linting fixes
780         https://bugs.webkit.org/show_bug.cgi?id=64225
781
782         Reviewed by Eric Siedel.
783
784         Miscellaneous linting fixes. The most notable change is that
785         we add public attributes for user, executive, filesystem, and
786         options on the Port object, so we don't have to refer to the
787         "protected" versions all over the place".
788
789         * Scripts/webkitpy/layout_tests/controllers/manager.py:
790         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
791         * Scripts/webkitpy/layout_tests/controllers/message_broker.py:
792         * Scripts/webkitpy/layout_tests/controllers/message_broker_unittest.py:
793         * Scripts/webkitpy/layout_tests/controllers/worker.py:
794         * Scripts/webkitpy/layout_tests/port/base.py:
795         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
796
797 2011-07-11  Ryosuke Niwa  <rniwa@webkit.org>
798
799         webkit-patch roll-chromium-deps no longer works
800         https://bugs.webkit.org/show_bug.cgi?id=64324
801
802         Reviewed by James Robinson.
803
804         Lower the minimum commit message length to 10.
805
806         * Scripts/webkitpy/tool/steps/commit.py:
807
808 2011-07-11  Alice Boxhall  <aboxhall@chromium.org>
809
810         Convert json_results_generator.py to output version 4 JSON.
811         https://bugs.webkit.org/show_bug.cgi?id=60869
812
813         Reviewed by Ojan Vafai.
814
815         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
816         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
817         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
818
819 2011-07-11  Martin Robinson  <mrobinson@igalia.com>
820
821         [GTK] [NRWT] Pixel tests do not work
822         https://bugs.webkit.org/show_bug.cgi?id=64091
823
824         Reviewed by Eric Seidel.
825
826         Fix pixel test runs for GTK+ new-run-webkit-tests by exposing the
827         path to the ImageDiff binary.
828
829         * Scripts/webkitpy/layout_tests/port/gtk.py: Implement _path_to_image_diff.
830
831 2011-07-11  Dirk Pranke  <dpranke@chromium.org>
832
833         Fix hang on win32 in a multiprocessing test that shouldn't have been running.
834
835         Unreviewed, build fix.
836
837         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
838
839 2011-07-11  Adam Roben  <aroben@apple.com>
840
841         Don't count leaks as test failures on TestFailures's front page
842
843         As a bonus, this patch adds our first unit test.
844
845         Fixes <http://webkit.org/b/64303> REGRESSION (r90489): TestFailures page says far more tests
846         are failing on the Leaks bot than actually are
847
848         Reviewed by Adam Barth.
849
850         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
851         (Builder.prototype.getNumberOfFailingTests): Exclude lines containing "leak" when summing
852         failure counts.
853
854         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder_unittests.js:
855         Added. Contains a simple test to show that leaks aren't counted as test failures.
856
857         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/run-unittests.html:
858         Added. This is the test harness.
859
860 2011-06-14  Mike Stegeman  <mrstegeman@gmail.com>
861
862         [GTK] Add API to allow setting local storage database path
863         https://bugs.webkit.org/show_bug.cgi?id=62091
864
865         Reviewed by Martin Robinson and Gustavo Noronha.
866
867         Expose the path of the localStorage databases through a setting
868         to allow HTML5 localStorage to be persistent. New setting is:
869         html5-local-storage-database-path
870
871         * DumpRenderTree/gtk/DumpRenderTree.cpp:
872         (resetDefaultsToConsistentValues):
873
874 2011-07-11  Adam Barth  <abarth@webkit.org>
875
876         garden-o-matic should be able to rebaseline tests
877         https://bugs.webkit.org/show_bug.cgi?id=64186
878
879         Reviewed by Eric Seidel.
880
881         This patch contains a simple HTTP server binding for the
882         rebaseline-test command.  The frontend will appear in another patch.
883
884         * Scripts/webkitpy/tool/servers/gardeningserver.py:
885         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py:
886         * Scripts/webkitpy/tool/servers/reflectionhandler.py:
887
888 2011-07-11  Adam Barth  <abarth@webkit.org>
889
890         Add some basic UI for showing regression ranges
891         https://bugs.webkit.org/show_bug.cgi?id=64243
892
893         Reviewed by Eric Seidel.
894
895         This UI is all temporary.  It exists mostly just to have some buttons
896         to click to exercise the code.
897
898         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
899         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
900         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
901         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
902         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
903
904 2011-07-11  Adam Barth  <abarth@webkit.org>
905
906         garden-o-matic should be able to determine which revisions caused a given failure
907         https://bugs.webkit.org/show_bug.cgi?id=64189
908
909         Reviewed by Adam Roben.
910
911         Walking the failure history looking for failures turns out to be
912         slightly tricky because the network requests are asynchronous.
913         Currently we do all the fetches serially and our cache is unbounded.
914         We'll probably optimize both those parameters eventually.
915
916         This patch also generalizes some functionality in the unit testing
917         framework to make testing this sort of code easier.
918
919         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
920         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
921         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
922
923 2011-07-11  Adam Barth  <abarth@webkit.org>
924
925         Add a webkit-patch command for rebaselining an individual test
926         https://bugs.webkit.org/show_bug.cgi?id=64246
927
928         Reviewed by Eric Seidel.
929
930         This patch introduces a command that's able to rebaseline a single
931         test.  Currently, the command works only with the build.chromium.org
932         buildbots, but extending it to work with the build.webkit.org bots
933         shouldn't be that hard.
934
935         A complete rebaseling tool should also include an "optimize baselines"
936         command (which moves/deletes baselines in order to reduce the number of
937         expected results files), but that will come in a future patch.
938
939         Really BuilderToPort should be merged into builders.py, but I'm going
940         to save that for a future patch as well.  (We need to stop shaving yaks
941         at some point.)
942
943         * Scripts/webkitpy/tool/commands/rebaseline.py:
944
945 2011-07-11  Adam Barth  <abarth@webkit.org>
946
947         garden-o-matic should be able to roll out patches
948         https://bugs.webkit.org/show_bug.cgi?id=64185
949
950         Reviewed by Eric Seidel.
951
952         This gardening server API is simply a binding to the rollout machinery
953         already present in webkit-patch.
954
955         * Scripts/webkitpy/tool/commands/gardenomatic.py:
956         * Scripts/webkitpy/tool/servers/gardeningserver.py:
957         * Scripts/webkitpy/tool/servers/gardeningserver_unittest.py: Added.
958
959 2011-07-11  Csaba Osztrogonác  <ossy@webkit.org>
960
961         [Qt][Mac] Unreviewed fix after r90746.
962
963         * Scripts/run-webkit-tests:
964         (useNewRunWebKitTests): Really disable NRWT for qt-mac platform.
965
966 2011-07-11  Adam Roben  <aroben@apple.com>
967
968         Teach TestFailures that ORWT's results.html file might be missing due to all tests passing
969
970         There are three reasons why we might fail to fetch ORWT's results.html:
971           - All tests passed, so no results.html was generated
972           - Some error during the test run caused results.html not to be generated (e.g., ORWT
973             timed out)
974           - Some network error occurred when fetching results.html
975
976         We were failing to account for the first possibility in some cases. For test runs before
977         r89610, we first check build.webkit.org/json to determine how many tests failed and whether
978         ORWT exited early due to too many failures; if all tests passed then we don't fetch
979         results.html at all. r89610 changed ORWT to put information in results.html about exiting
980         early due to too many failures, so we no longer needed to check build.webkit.org/json to get
981         that information, and in r89619 I changed TestFailures to do just that. But I forgot that we
982         still needed to check build.webkit.org/json to find out if all tests passed!
983
984         Now, for test runs after r89610, we check results.html first, and then check
985         build.webkit.org/json if we fail to fetch results.html. This lets us distinguish between all
986         tests passing and the error cases.
987
988         Fixes <http://webkit.org/b/64280> TestFailures page can't pinpoint that r90699 caused 13
989         tests to fail on Windows 7 Release (WebKit2 Tests)
990
991         Reviewed by David Kilzer.
992
993         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
994         (LayoutTestResultsLoader.prototype.start): Bumped the cache version so that old, buggy
995         cached data will get evicted. We were marking builds where all tests passed as errors!
996         (LayoutTestResultsLoader.prototype._fetchAndParseORWTResults): Added success/error callback
997         parameters to the fetchAndParseResultsHTML helper function, and added a similar
998         fetchNumberOfFailingTests function that fetches data from build.webkit.org/json (code came
999         from later in the function). For test runs before r89610, we first check
1000         build.webkit.org/json then check results.html. For builds after r89610, we first check
1001         results.html and then check build.webkit.org/json if we couldn't fetch results.html.
1002
1003 2011-07-11  Csaba Osztrogonác  <ossy@webkit.org>
1004
1005         new-run-webkit-tests does not support qt-arm or qt-4.8 results
1006         https://bugs.webkit.org/show_bug.cgi?id=64071
1007
1008         [Qt] NRWT should pick up the right httpd config file
1009         https://bugs.webkit.org/show_bug.cgi?id=64086
1010
1011         Reviewed by Zoltán Herczeg.
1012
1013         * Scripts/run-webkit-tests: Disable NRWT for qt-mac, qt-arm and qt-4.8 platforms until fix.
1014         (useNewRunWebKitTests):
1015
1016 2011-07-11  Kenichi Ishibashi  <bashi@chromium.org>
1017
1018         Add TestNetscapePlugIn/Tests/FormValue.cpp to DRT build files
1019         https://bugs.webkit.org/show_bug.cgi?id=64248
1020
1021         Reviewed by Kent Tamura.
1022
1023         * DumpRenderTree/DumpRenderTree.gypi: Added FormValue.cpp.
1024         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Ditto.
1025         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro: Ditto.
1026         * GNUmakefile.am: Ditto.
1027
1028 2011-07-10  Kenichi Ishibashi  <bashi@chromium.org>
1029
1030         Let plugins participate in form submission
1031         https://bugs.webkit.org/show_bug.cgi?id=13061
1032
1033         Adds NPPVformValue support to the testing plugin.
1034
1035         Reviewed by Anders Carlsson.
1036
1037         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Added FormValue.cpp
1038         * DumpRenderTree/TestNetScapePlugIn/Tests/FormValue.cpp: Added.
1039         (FormValue::FormValue): Ditto.
1040         (FormValue::NPP_GetValue): Ditto.
1041
1042 2011-07-11  Roland Steiner  <rolandsteiner@chromium.org>
1043
1044         Reviewed by Tony Chang.
1045
1046         Complete functions in filesystem.py
1047         https://bugs.webkit.org/show_bug.cgi?id=63528
1048
1049         - missing text file functions added
1050         - functions sorted
1051         - removed 'append' optional parameters (were unused)
1052         - adapted filesystem_mock in the same way
1053
1054         * Scripts/webkitpy/common/system/filesystem.py:
1055         * Scripts/webkitpy/common/system/filesystem_mock.py:
1056         * Scripts/webkitpy/common/system/filesystem_unittest.py:
1057
1058 2011-07-10  Adam Barth  <abarth@webkit.org>
1059
1060         Clean up style in fallback path calculation
1061         https://bugs.webkit.org/show_bug.cgi?id=64239
1062
1063         Reviewed by Daniel Bates.
1064
1065         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1066         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1067         * Scripts/webkitpy/layout_tests/port/mac.py:
1068         * Scripts/webkitpy/layout_tests/port/win.py:
1069
1070 2011-07-10  Balazs Kelemen  <kbalazs@webkit.org>
1071
1072         [Qt][WK2] Implement the rest of WTR::PlatformWebView
1073         https://bugs.webkit.org/show_bug.cgi?id=63630
1074
1075         Reviewed by Kenneth Rohde Christiansen.
1076
1077         * WebKitTestRunner/qt/PlatformWebViewQt.cpp:
1078         (WTR::PlatformWebView::windowFrame):
1079         (WTR::PlatformWebView::setWindowFrame):
1080
1081 2011-07-10  Mark Rowe  <mrowe@apple.com>
1082
1083         Teach build-webkit how to find the latest WebKitSystemInterface binary.
1084
1085         * Scripts/build-webkit:
1086
1087 2011-07-09  Adam Roben  <aroben@apple.com>
1088
1089         Teach TestFailures to abbreviate the examples of test flakiness
1090
1091         These lists can get quite long, and it's not really helpful in most cases to have soooooo
1092         many examples of flakiness.
1093
1094         Fixes <http://webkit.org/b/64203> Lists of flaky revisions on TestFailures page can get so
1095         long they're hard to navigate
1096
1097         Reviewed by Dan Bates.
1098
1099         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/FlakyLayoutTestDetector.js:
1100         (FlakyLayoutTestDetector.prototype.flakinessExamples): If we have more than a certain number
1101         of examples, replace the middle items with a separator.
1102
1103         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
1104         (.flakiness-example-separator): Added styles for the separator.
1105
1106         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1107         (ViewController.prototype._domForPossiblyFlakyTests): Use a vertical ellipsis to represent
1108         the separator.
1109
1110 2011-07-09  Dirk Pranke  <dpranke@chromium.org>
1111
1112         nrwt: stack traces from worker-side exceptions aren't very useful inside test-webkitpy
1113         https://bugs.webkit.org/show_bug.cgi?id=64218
1114
1115         Reviewed by Eric Seidel.
1116
1117         Exceptions aren't picklable and can't be sent across the
1118         manager/worker message queue without losing information. NRWT
1119         handles this by turning the stack trace into a set of strings,
1120         and logging the strings when we receive an exception from the
1121         worker. However, when you are running tests and something
1122         crashes on the worker side, test-webkitpy prints the
1123         manager-side stack trace, which is just confusing and useless.
1124
1125         This patch changes the logic so that exceptions are passed
1126         through as-is when the worker and manager are in the same
1127         process (the --worker-model=inline option). This increases the
1128         code paths slightly but makes crashes much more useful.
1129
1130         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1131         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
1132         * Scripts/webkitpy/layout_tests/controllers/message_broker.py:
1133         * Scripts/webkitpy/layout_tests/controllers/worker.py:
1134         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1135
1136 2011-07-09  Darin Fisher  <darin@chromium.org>
1137
1138         Eliminate bad dependency on gfx::Point.
1139         https://bugs.webkit.org/show_bug.cgi?id=64228
1140
1141         Reviewed by Kent Tamura.
1142
1143         * DumpRenderTree/chromium/EventSender.cpp:
1144         (initMouseEvent):
1145
1146 2011-07-08  Ojan Vafai  <ojan@chromium.org>
1147
1148         pull static dashboard files into the appengine server from the chromium repository
1149         https://bugs.webkit.org/show_bug.cgi?id=64208
1150
1151         Reviewed by Dirk Pranke.
1152
1153         These files belong in the WebKit repo since they are tied primarily to the webkit tests.
1154         They have some extra bits to support chromium's gtests, but that seems fine.
1155         Mainly, this will allow other WebKit hackers to hack on the dashboards.
1156
1157         As a nice side-effect, we can now push the dashboard files when we do appengine pushes
1158         instead of the weird thing we used to do of pulling them from the Chromium repository
1159         and storing them in the appengine datastore. This allows for cleaning up a lot of code
1160         and will likely make the dashboards load a bit faster.
1161
1162         The new JS files don't fully match WebKit style, but I'd like to clean that up in a
1163         followup patch if possible to maintain my sanity with this patch.
1164
1165         * TestResultServer/app.yaml:
1166         * TestResultServer/handlers/dashboardhandler.py: Removed.
1167         * TestResultServer/handlers/menu.py:
1168         * TestResultServer/main.py:
1169         * TestResultServer/model/dashboardfile.py: Removed.
1170         * TestResultServer/static-dashboards/LICENSE.dygraph.txt: Added.
1171         * TestResultServer/static-dashboards/README.dygraph.txt: Added.
1172         * TestResultServer/static-dashboards/README.webtreemap.txt: Added.
1173         * TestResultServer/static-dashboards/aggregate_results.html: Added.
1174         * TestResultServer/static-dashboards/builders.js: Added.
1175         * TestResultServer/static-dashboards/dashboard_base.js: Added.
1176         * TestResultServer/static-dashboards/dygraph-combined.js: Added.
1177         * TestResultServer/static-dashboards/flakiness_dashboard.html: Added.
1178         * TestResultServer/static-dashboards/flakiness_dashboard_tests.js: Added.
1179         * TestResultServer/static-dashboards/timeline_explorer.html: Added.
1180         * TestResultServer/static-dashboards/treemap.html: Added.
1181         * TestResultServer/static-dashboards/webtreemap.css: Added.
1182         * TestResultServer/static-dashboards/webtreemap.js: Added.
1183         * TestResultServer/stylesheets/dashboardfile.css: Removed.
1184         * TestResultServer/stylesheets/menu.css:
1185         * TestResultServer/templates/dashboardfilelist.html: Removed.
1186         * TestResultServer/templates/menu.html:
1187
1188 2011-07-08  Adam Roben  <aroben@apple.com>
1189
1190         Make TestFailures's list of flaky tests look more like the list of non-flaky tests
1191
1192         Fixes <http://webkit.org/b/64204> TestFailures page's flaky tests list is ugly!
1193
1194         Reviewed by Daniel Bates.
1195
1196         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
1197         (#failure-history, #possibly-flaky-tests): Expanded this rule to apply to the list of flaky
1198         tests.
1199         (#failure-history > li, #possibly-flaky-tests > li): Ditto, but moved the 50px left padding
1200         from here...
1201         (#failure-history > li): ...to here.
1202
1203         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1204         (ViewController.prototype._domForPossiblyFlakyTests): Give the list an id attribute for
1205         styling purposes.
1206
1207 2011-07-08  Jeffrey Pfau  <jpfau@apple.com>
1208
1209         Unreviewed, add myself as committer.
1210
1211         * Scripts/webkitpy/common/config/committers.py:
1212
1213 2011-07-08  Adam Barth  <abarth@webkit.org>
1214
1215         TestResultsServer should keep old test results
1216         https://bugs.webkit.org/show_bug.cgi?id=64199
1217
1218         Reviewed by Ojan Vafai.
1219
1220         Having historical data will help us do failure archeology.
1221
1222         * TestResultServer/handlers/testfilehandler.py:
1223         * TestResultServer/model/testfile.py:
1224         * TestResultServer/templates/showfilelist.jsonp: Added.
1225
1226 2011-07-08  Adam Roben  <aroben@apple.com>
1227
1228         Remove commit-log-editor's dependency on Module::Load::Conditional
1229
1230         This module isn't available in Perl 5.8.8 (the version used on Leopard).
1231
1232         Fixes <http://webkit.org/b/64198> REGRESSION (r90583):
1233         webkitpy.common.checkout.checkout_unittest failing on Leopard
1234
1235         Reviewed by Daniel Bates.
1236
1237         * Scripts/commit-log-editor: Use the new loadTermReadKey() function instead of
1238         Module::Load::Conditional::can_load.
1239         (loadTermReadKey): Added. Tries to load Term::ReadKey and returns true if it's successful.
1240
1241 2011-07-08  Adam Roben  <aroben@apple.com>
1242
1243         Make checkout_unittest more robust against files moving around
1244
1245         Fixes <http://webkit.org/b/64197> checkout_unittest contains ugly, fragile code to find the
1246         Scripts directory
1247
1248         Reviewed by Adam Barth.
1249
1250         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
1251         (CommitMessageForThisCommitTest.test_commit_message_for_this_commit): Instantiate a real SCM
1252         object and use it to get the path to the Scripts directory, rather than hard-coding the
1253         relative path from this file to Scripts.
1254
1255 2011-07-08  Adam Roben  <aroben@apple.com>
1256
1257         Make TestFailures remember that run-webkit-tests timed out, even across reloads
1258
1259         Fixes <http://webkit.org/b/64193> TestFailures page incorrectly thinks all tests passed in
1260         http://build.webkit.org/builders/Windows%207%20Release%20(Tests)/builds/14589 after a reload
1261
1262         Reviewed by Daniel Bates.
1263
1264         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
1265         (LayoutTestResultsLoader.prototype.start): Store an "error" attribute in the cached data.
1266         When true, it indicates that there was an error retrieving the results for this build and
1267         that the errorCallback should be called.
1268
1269 2011-07-08  Adam Barth  <abarth@webkit.org>
1270
1271         Teach garden-o-matic how to display test results
1272         https://bugs.webkit.org/show_bug.cgi?id=64141
1273
1274         Reviewed by Ojan Vafai.
1275
1276         This patch includes basic infrastructure for probing build.chromium.org
1277         for test results.  We only handle text and image tests, not anything
1278         complicated like reftests.  Also, we're using the revision/build
1279         independent results store on the server, so we're avoiding that
1280         complication for now.
1281
1282         It's slightly hacky that we need to probe the server to see what kinds
1283         of results exist.  A better solution would be to add CORS support to
1284         the server or to use the local server to help.
1285
1286         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
1287         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
1288         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
1289         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
1290         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js:
1291         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js:
1292
1293 2011-07-08  Dirk Pranke  <dpranke@chromium.org>
1294
1295         REGRESSION(90419) NRWT's httpd locking is broken for --child-processes=1
1296         https://bugs.webkit.org/show_bug.cgi?id=64092
1297
1298         Reviewed by Tony Chang.
1299
1300         The code was incorrectly creating empty shards in the
1301         shard_in_two case.
1302
1303         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1304         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
1305
1306 2011-07-08  Adam Roben  <aroben@apple.com>
1307
1308         Teach buildbot to figure out how many webkitpy/webkitperl tests failed
1309
1310         Fixes <http://webkit.org/b/64192> It's hard to tell how many test-webkitpy/test-webkitperl
1311         tests failed when looking at build.webkit.org
1312
1313         Reviewed by Eric Seidel.
1314
1315         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1316         (TestWithFailureCount): New class that represents a test build step which has an associated
1317         failure count. Eventually we should move more of our test classes to deriving from this.
1318         (TestWithFailureCount.countFailures): Method for subclasses to override to say how many
1319         failures occurred.
1320
1321         (TestWithFailureCount.commandComplete):
1322         (TestWithFailureCount.evaluateCommand):
1323         (TestWithFailureCount.getText):
1324         (TestWithFailureCount.getText2):
1325         These were all based on RunGtkAPITests.
1326
1327         (RunPythonTests): Changed to inherit from TestWithFailureCount.
1328         (RunPythonTests.countFailures): Parses the test-webkitpy output looking for the count of
1329         failures.
1330         (RunPerlTests): Changed to inherit from TestWithFailureCount.
1331         (RunPerlTests.countFailures): Parses the test-webkitperl output looking for the count of
1332         failures.
1333
1334 2011-07-08  Adam Roben  <aroben@apple.com>
1335
1336         Ensure $CHANGE_LOG_EMAIL_ADDRESS is set when testing webkitpy's commit-log-editor integration
1337
1338         Fixes <http://webkit.org/b/64180> REGRESSION (r90564): test-webkitpy failing on multiple
1339         bots due to commit-log-editor errors
1340
1341         Reviewed by Adam Barth.
1342
1343         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
1344         (CommitMessageForThisCommitTest.test_commit_message_for_this_commit): Set
1345         $CHANGE_LOG_EMAIL_ADDRESS to the patch author's email address. This ensures that
1346         commit-log-editor can find a value for the committer's email, and that the committer and
1347         author email addresses match, which will prevent commit-log-editor from inserting a "Patch
1348         by" line in the commit message.
1349
1350 2011-07-08  Andreas Kling  <kling@webkit.org>
1351
1352         [Qt][WK2] Views should expose QActions for basic navigation.
1353         https://bugs.webkit.org/show_bug.cgi?id=64174
1354
1355         Reviewed by Benjamin Poulain.
1356
1357         Bring the toolbar in Qt's MiniBrowser back to life.
1358
1359         * MiniBrowser/qt/BrowserView.cpp:
1360         (BrowserView::navigationAction):
1361         * MiniBrowser/qt/BrowserView.h:
1362         * MiniBrowser/qt/BrowserWindow.cpp:
1363         (BrowserWindow::BrowserWindow):
1364
1365 2011-07-08  Adam Roben  <aroben@apple.com>
1366
1367         Teach TestFailures how to find test names in commit-log-editor-style commit messages
1368
1369         TestFailures was relying on Trac turning the list of modified files in our commit messages
1370         into an HTML list. But Trac only does this when the list of modified files is indented.
1371         commit-log-editor doesn't indent the file list, so the list wasn't being turned into an HTML
1372         list, which was confusing TestFailures.
1373
1374         TestFailures now does much simpler parsing of the commit message (i.e., just a substring
1375         search) without relying at all on its structure.
1376
1377         Fixes <http://webkit.org/b/64173> TestFailures page fails to blame r90608 for breaking
1378         fast/dom/HTMLProgressElement/progress-element-markup.html on Windows
1379
1380         Reviewed by David Kilzer.
1381
1382         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Trac.js:
1383         (Trac.prototype.getCommitDataForRevisionRange): Instead of trying to parse the commit
1384         message, just return its text.
1385
1386         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1387         (ViewController.prototype._domForRegressionRange): Instead of searching for test names in
1388         each commit's list of modified files, just search for test names anywhere in the commit's
1389         message.
1390
1391 2011-07-08  Adam Barth  <abarth@webkit.org>
1392
1393         sheriffbot is too spammy in IRC
1394         https://bugs.webkit.org/show_bug.cgi?id=64153
1395
1396         Reviewed by Eric Seidel.
1397
1398         Reporting failures in IRC worked well when the tree was greener than it
1399         is today.  Nowadays, this feature mostly just results in spam about
1400         false positives.  If we reach a greener state, we can bring this code
1401         back.
1402
1403         (Another possibility is to restrict this feature to builder bots,
1404         rather than including testers, as we did before this patch.)
1405
1406         * Scripts/webkitpy/tool/commands/sheriffbot.py:
1407         * Scripts/webkitpy/tool/commands/sheriffbot_unittest.py:
1408
1409 2011-07-08  Adam Roben  <aroben@apple.com>
1410
1411         Ensure commit-log-editor adds a "Patch by" line when the author and committer are different
1412
1413         Previously we were only adding a "Patch by" line when the ChangeLog contained a "Reviewed
1414         by" line. But some patches (like rollout patches) don't contain that line. Now we always add
1415         "Patch by" to the commit log regardless of the ChangeLog's contents.
1416
1417         Fixes <http://webkit.org/b/64127> Committer for r90588 is commit-queue@webkit.org, but
1418         should have been abarth@webkit.org
1419
1420         Reviewed by Anders Carlsson.
1421
1422         * Scripts/commit-log-editor:
1423         (createCommitMessage): Try to put the "Patch by" line just above the "Reviewed by" line, as
1424         before. If there is no "Reviewed by" line, try to put it just above the first modified file.
1425         If all else fails, put it at the end of the commit message.
1426         (patchAuthorshipString): Added. Code came from createCommitMessage.
1427
1428 2011-07-08  Herczeg Zoltan  <zherczeg@webkit.org>
1429
1430         Reviewed by Andreas Kling.
1431
1432         Adding myself as a reviewer.
1433
1434         * Scripts/webkitpy/common/config/committers.py:
1435
1436 2011-07-07  Kevin Ollivier  <kevino@theolliviers.com>
1437
1438         [wx] Unreviewed build fix, remove old files from the tree and update paths to new ones.
1439
1440         * wx/build: Removed.
1441         * wx/build/build_utils.py: Removed.
1442         * wx/build/settings.py: Removed.
1443         * wx/build/waf_extensions.py: Removed.
1444         * wx/build/wxpresets.py: Removed.
1445         * wx/packaging/build-mac-installer.py:
1446
1447 2011-07-07  Kevin Ollivier  <kevino@theolliviers.com>
1448
1449         [wx] Unreviewed build fix, more fixes for install name issues.
1450
1451         * wx/packaging/build-mac-installer.py:
1452
1453 2011-07-07  Dirk Pranke  <dpranke@chromium.org>
1454
1455         test-webkitpy fails on chromium win
1456         https://bugs.webkit.org/show_bug.cgi?id=64137
1457
1458         Reviewed by Eric Seidel.
1459
1460         test-webkitpy doesn't play nicely with the multiprocessing
1461         module on win32. An earlier change actually reenabled the tests
1462         on win32 by mistake.
1463
1464         This patch also fixes a few cases where path names will trip
1465         things up on win32 (e.g., by testing unix paths on windows).
1466         We do not lose any real coverage here with those fixes.
1467
1468         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
1469         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
1470         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
1471         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1472
1473 2011-07-07  Leandro Pereira  <leandro@profusion.mobi>
1474
1475         [ImageDiff] Calculate/print difference right after reading baseline image.
1476         https://bugs.webkit.org/show_bug.cgi?id=64117
1477
1478         Reviewed by Kent Tamura.
1479
1480         Otherwise, ImageDiff will block on fgets() until the universe ends or you ^C
1481         it.  Whichever happens first.
1482
1483         * DumpRenderTree/efl/ImageDiff.cpp:
1484         (main):
1485
1486 2011-07-07  Raphael Kubo da Costa  <kubo@profusion.mobi>
1487
1488         [EFL] DRT: Leak cairo_t to make the pixel tests stop crashing
1489         https://bugs.webkit.org/show_bug.cgi?id=64107
1490
1491         Reviewed by Kent Tamura.
1492
1493         The pointer is later managed by BitmapContext, but it was being
1494         de-refed and deleted earlier by our RefPtr, causing crashes in all
1495         pixel tests.
1496
1497         We now call leakRef() to make sure it is not removed when our
1498         createBitmapContextFromWebView goes out of context.
1499
1500         * DumpRenderTree/efl/PixelDumpSupportEfl.cpp:
1501         (createBitmapContextFromWebView):
1502
1503 2011-07-07  Adam Roben  <aroben@apple.com>
1504
1505         Update TestFailures's title and header immediately upon navigation
1506
1507         Fixes <http://webkit.org/b/64125> TestFailures page seems slow to react on link clicks
1508
1509         Reviewed by Daniel Bates.
1510
1511         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1512         (ViewController.prototype.loaded): Create and store an h1 element for displaying the page's
1513         title.
1514         (ViewController.prototype._displayBuilder): Set the page's title and clear out any old
1515         content immediately instead of waiting for the first set of results for the history
1516         analyzer. As we receive new results from the analyzer we'll just clear out the main content
1517         area and replace it with the new info.
1518         (ViewController.prototype._displayTesters): Set the page's title and clear out any old
1519         content immediately instead of waiting for the list of testers from the buildbot. Sprinkled
1520         in a little use of our removeAllChildren helper, too.
1521         (ViewController.prototype._setTitle): Added. Set's the page's title and the header text.
1522
1523 2011-07-07  Adam Roben  <aroben@apple.com>
1524
1525         Make Checkout use SCM's Executive instead of conjuring up its own
1526
1527         This will improve integration with the rest of webkitpy, particularly when invoked via
1528         webkit-patch.
1529
1530         Fixes <http://webkit.org/b/64115> REGRESSION (r90564): webkitpy's Checkout class uses
1531         Executive inappropriately
1532
1533         Reviewed by Adam Barth.
1534
1535         * Scripts/webkitpy/common/checkout/checkout.py:
1536         (Checkout.commit_message_for_this_commit): Use SCM.run instead of creating an Executive for
1537         our own use. SCM might have some super-special Executive that it uses under the covers, and
1538         we want to use it, too!
1539
1540         * Scripts/webkitpy/common/checkout/checkout_unittest.py:
1541         (CommitMessageForThisCommitTest.test_commit_message_for_this_commit): Mock the SCM.run
1542         method to call through to Executive.run_command.
1543
1544 2011-07-07  Adam Roben  <aroben@apple.com>
1545
1546         Make Term/ReadKey.pm an optional dependency of commit-log-editor
1547
1548         When Term/ReadKey.pm isn't present, the --regenerate-log switch will be non-functional.
1549
1550         Fixes <http://webkit.org/b/64113> REGRESSION (r90564):
1551         webkitpy.common.checkout.checkout_unittest failing on GTK bots due to missing
1552         Term/ReadKey.pm module
1553
1554         Reviewed by Xan Lopez.
1555
1556         * Scripts/commit-log-editor: Use Module::Load::Conditional::can_load to only load
1557         Term/ReadKey.pm if it's present. If it isn't present, ignore the --regenerate-log switch so
1558         that existing commit messages will be preserved. (Users can manually remove existing commit
1559         messages using git-reset.)
1560
1561 2011-07-07  Adam Roben  <aroben@apple.com>
1562
1563         Keep testing that commit messages containing Unicode are handled properly by webkitpy
1564
1565         Fixes <http://webkit.org/b/64109> REGRESSION (r90571): test-webkitpy no longer tests that we
1566         correctly handle Unicode in commit messages
1567
1568         Reviewed by Eric Seidel.
1569
1570         * Scripts/webkitpy/common/checkout/checkout_unittest.py: Put some Unicode characters in
1571         places in the ChangeLog that will end up being included in the commit message.
1572
1573 2011-07-07  Andreas Kling  <kling@webkit.org>
1574
1575         [Qt][WK2] Remove Symbian code.
1576         https://bugs.webkit.org/show_bug.cgi?id=64101
1577
1578         Reviewed by Benjamin Poulain.
1579
1580         * MiniBrowser/qt/BrowserWindow.cpp:
1581         (BrowserWindow::BrowserWindow):
1582         (BrowserWindow::screenshot):
1583         * MiniBrowser/qt/MiniBrowser.pro:
1584         * MiniBrowser/qt/MiniBrowserApplication.h:
1585         (WindowOptions::WindowOptions):
1586
1587 2011-07-07  Andreas Kling  <kling@webkit.org>
1588
1589         [Qt][WK2] Don't support ridiculous matrix of QT_NO_FEATURE combinations.
1590         https://bugs.webkit.org/show_bug.cgi?id=64099
1591
1592         Reviewed by Benjamin Poulain.
1593
1594         * MiniBrowser/qt/BrowserWindow.cpp:
1595         (BrowserWindow::openFile):
1596         (BrowserWindow::screenshot):
1597         (BrowserWindow::loadURLListFromFile):
1598
1599 2011-07-07  Adam Roben  <aroben@apple.com>
1600
1601         Completely ignore too-many-failures builds in TestFailures in most circumstances
1602
1603         Because a semi-arbitrary subset of tests are run in a too-many-failures build, we can't
1604         really use them to perform regression analysis. The only time we want to pay attention to
1605         too-many-failures builds is when we're trying to explain when the current bout of
1606         too-many-failures started.
1607
1608         Fixes <http://webkit.org/b/64106> TestFailures page sometimes claims a test started failing
1609         in a build that didn't even run it (because it exited early due to too many failues)
1610
1611         Reviewed by David Kilzer.
1612
1613         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
1614         (LayoutTestHistoryAnalyzer.prototype._incorporateBuildHistory): Removed old, broken
1615         too-many-failures handling that would cause us to blame builds that didn't even run a given
1616         test for breaking it. Instead, skip over all too-many-failures builds unless the most recent
1617         build was itself a too-many-failures build.
1618
1619 2011-07-07  Adam Roben  <aroben@apple.com>
1620
1621         Teach webkitpy's Checkout class to use commit-log-editor to create commit messages
1622
1623         Fixes <http://webkit.org/b/26755> webkit-patch's commit messages are less readable than
1624         commit-log-editor's
1625
1626         Reviewed by David Kilzer.
1627
1628         * Scripts/webkitpy/common/checkout/checkout.py:
1629         (Checkout.commit_message_for_this_commit): Run commit-log-editor, passing it the paths of
1630         the modified ChangeLogs, to generate the commit message, rather than trying to generate one
1631         ourselves.
1632
1633         * Scripts/webkitpy/common/checkout/checkout_unittest.py: Updated the expected commit message
1634         to match commit-log-editor's format.
1635         (CommitMessageForThisCommitTest.setUp): Write the ChangeLogs into Tools and LayoutTests
1636         directories so we can see how the various entries get labeled in the commit message.
1637         (CommitMessageForThisCommitTest.test_commit_message_for_this_commit): Create a mock SCM
1638         instance that knows how to find commit-log-editor and pass it to our Checkout instance.
1639         Don't bother capturing output, since there shouldn't be any.
1640
1641 2011-07-07  Adam Roben  <aroben@apple.com>
1642
1643         Add a --print-log option to commit-log-editor
1644
1645         When specified, commit-log-editor takes a set of ChangeLog paths on the command line,
1646         generates a commit message from those ChangeLogs, prints it to stdout, and exits.
1647
1648         Prep work for fixing <http://webkit.org/b/26755> webkit-patch's commit messages are less
1649         readable than commit-log-editor's
1650
1651         Reviewed by David Kilzer.
1652
1653         * Scripts/commit-log-editor: Changed to use Getopt::Long to parse options. Added --print-log
1654         option, which calls through to createCommitMessage, prints the result, and exits.
1655         (printUsageAndExit): Renamed from "usage" for clarity. Beefed up the usage statement to
1656         explain commit-log-editor's 2.5 (normal, --print-log, and --help) modes of operation.
1657
1658 2011-07-07  Adam Roben  <aroben@apple.com>
1659
1660         Extract commit-log-editor's commit-message-generation code into a separate function
1661
1662         As a bonus, we now skip a bunch of work in the case where we are generating a commit message
1663         for staged git changes.
1664
1665         Prep work for fixing <http://webkit.org/b/26755> webkit-patch's commit messages are less
1666         readable than commit-log-editor's
1667
1668         Reviewed by David Kilzer.
1669
1670         * Scripts/commit-log-editor: Moved the default definition of $endl up toward the top of the
1671         file so that it will be defined even if we aren't operating on an existing log message.
1672         Moved a bunch of top-level code to generate the commit message from here...
1673         (createCommitMessage): ...to here.
1674
1675 2011-07-07  Carlos Garcia Campos  <cgarcia@igalia.com>
1676
1677         [GTK] Open links in a new window when clicking with the middle button in MiniBrowser
1678         https://bugs.webkit.org/show_bug.cgi?id=63988
1679
1680         Reviewed by Martin Robinson.
1681
1682         * MiniBrowser/gtk/BrowserWindow.c:
1683         (browserWindowConstructed):
1684         (decidePolicyForNavigationAction): Ignore the action if a link is
1685         clicked with the middle buttonm and open the link in a new window.
1686         (browserWindowPolicyClientInit):
1687
1688 2011-07-07  Adam Barth  <abarth@webkit.org>
1689
1690         Remove style_references.py
1691         https://bugs.webkit.org/show_bug.cgi?id=64070
1692
1693         Reviewed by Eric Seidel.
1694
1695         We never really adopted the _references design, and this file is a bit
1696         of an orphan at this point.
1697
1698         * Scripts/check-webkit-style:
1699         * Scripts/webkitpy/style/checker.py:
1700         * Scripts/webkitpy/style/checker_unittest.py:
1701         * Scripts/webkitpy/style/checkers/python.py:
1702         * Scripts/webkitpy/style/checkers/test_expectations.py:
1703         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
1704         * Scripts/webkitpy/style/main_unittest.py:
1705         * Scripts/webkitpy/style_references.py: Removed.
1706
1707 2011-07-07  Eric Seidel  <eric@webkit.org>
1708
1709         Fix WebKit2 expected results search paths for Mac and Qt under new-run-webkit-tests
1710         https://bugs.webkit.org/show_bug.cgi?id=64056
1711
1712         Unreviewed.
1713
1714         I had written this unittest prior to landing my change, but
1715         had named it qt_unitest.py, so it wasn't running (nor was
1716         it added to my git repo).
1717
1718         Once I fixe the name of the unittest file, it was very easy
1719         to fix the error in webkit.py (which already had a FIXME).
1720
1721         * Scripts/webkitpy/layout_tests/port/qt_unittest.py: Added.
1722         * Scripts/webkitpy/layout_tests/port/webkit.py:
1723
1724 2011-07-07  Eric Seidel  <eric@webkit.org>
1725
1726         Unreviewed.  Add more debug logging about which test expectations we're using.
1727
1728         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
1729         * Scripts/webkitpy/layout_tests/port/webkit.py:
1730
1731 2011-07-06  Adam Barth  <abarth@webkit.org>
1732
1733         Remove unexpected_results.json
1734         https://bugs.webkit.org/show_bug.cgi?id=64065
1735
1736         Reviewed by Eric Seidel.
1737
1738         No one should be using this file any more.  This patch removes it from
1739         the face of the earth.  This patch also removes two NRWT-specific
1740         command line options that were implemented using
1741         unexpected_results.json.
1742
1743         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
1744         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1745         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
1746         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1747         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1748         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
1749         * Scripts/webkitpy/tool/servers/rebaselineserver_unittest.py:
1750
1751 2011-07-07  Adam Barth  <abarth@webkit.org>
1752
1753         REGRESSION(90520): chromium GPU tests produce output in the LayoutTests directory
1754         https://bugs.webkit.org/show_bug.cgi?id=64067
1755
1756         Reviewed by Eric Seidel.
1757
1758         The author of r90520 forgot to change the chromium_gpu.py implemenation
1759         of this method as well.
1760
1761         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
1762
1763 2011-07-06  Eric Seidel  <eric@webkit.org>
1764
1765         Fix WebKit2 expected results search paths for Mac and Qt under new-run-webkit-tests
1766         https://bugs.webkit.org/show_bug.cgi?id=64056
1767
1768         Reviewed by Adam Barth.
1769
1770         This fixes the MacPort's version of baseline_search_path to include
1771         wk2 fallback like WebKitPort's does.  This also re-works the QtPort
1772         to set self._name (as other ports do) so that WebKitPort's default
1773         implementations will understand Qt operating system flavors correctly
1774         (including when running wk2 tests).
1775
1776         In trying to test this, I realized that nearly all of our port-tests
1777         use the default FileSystem(), User() and Executive() objects, which
1778         is really bad, because they expose details of your system in the unittest results!
1779
1780         Once I fixed many of the Port() calls to pass MockFileSystem(), then
1781         MacPort.test_tests_for_other_platforms started failing, and I had to
1782         teach MockFileSystem.glob how to handle directories to make it work again.
1783
1784         I removed the useless PortTestCase.test_baseline_search_path and replaced
1785         it with more useful test_baseline_search_path tests in MacPortTest and QtPortTest.
1786
1787         * Scripts/webkitpy/common/config/ports.py:
1788         * Scripts/webkitpy/common/system/filesystem_mock.py:
1789         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1790         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
1791         * Scripts/webkitpy/layout_tests/port/mac.py:
1792         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1793         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
1794         * Scripts/webkitpy/layout_tests/port/qt.py:
1795         * Scripts/webkitpy/layout_tests/port/test_files.py:
1796         * Scripts/webkitpy/layout_tests/port/win.py:
1797         * Scripts/webkitpy/to_be_moved/deduplicate_tests.py:
1798         * Scripts/webkitpy/tool/mocktool.py:
1799
1800 2011-07-06  Adam Barth  <abarth@webkit.org>
1801
1802         Wean rebaseline-server off unexpected_results.json
1803         https://bugs.webkit.org/show_bug.cgi?id=64063
1804
1805         Reviewed by Eric Seidel.
1806
1807         In the process of changing this code to use full_results.json, I
1808         noticed that the code was broken (because it wasn't tested).  This
1809         patch also adds test coverage for the broken code.
1810
1811         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
1812         * Scripts/webkitpy/tool/servers/rebaselineserver.py:
1813         * Scripts/webkitpy/tool/servers/rebaselineserver_unittest.py:
1814
1815 2011-07-06  Adam Barth  <abarth@webkit.org>
1816
1817         Wean resultsjsonparser off unexpected_results.json
1818         https://bugs.webkit.org/show_bug.cgi?id=64061
1819
1820         Reviewed by Eric Seidel.
1821
1822         We're about to delete unexpected_results.json, so we need to remove all
1823         the users.
1824
1825         * Scripts/webkitpy/common/config/ports.py:
1826         * Scripts/webkitpy/common/net/resultsjsonparser.py:
1827         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
1828         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
1829
1830 2011-07-06  Adam Barth  <abarth@webkit.org>
1831
1832         Cleanup result_summary.py
1833         https://bugs.webkit.org/show_bug.cgi?id=64057
1834
1835         Reviewed by Eric Seidel.
1836
1837         This class had a bunch of out-of-date docstrings that no longer make
1838         any sense.
1839
1840         * Scripts/webkitpy/layout_tests/models/result_summary.py:
1841
1842 2011-07-06  Adam Barth  <abarth@webkit.org>
1843
1844         Tweak the ChangeLog for DEPS rolls to say "Unreviewed" so that the
1845         patches can be landed by the commit-queue.
1846
1847         * Scripts/webkitpy/tool/steps/preparechangelogfordepsroll.py:
1848
1849 2011-07-06  Adam Barth  <abarth@webkit.org>
1850
1851         Normalize callers of logging.getLogger to use __name__
1852         https://bugs.webkit.org/show_bug.cgi?id=64054
1853
1854         Reviewed by Eric Seidel.
1855
1856         Most callers were already using __name__.  This patch makes everyone
1857         consistent.
1858
1859         * Scripts/webkitpy/common/checkout/diff_parser.py:
1860         * Scripts/webkitpy/common/checkout/scm/svn.py:
1861         * Scripts/webkitpy/common/net/statusserver.py:
1862         * Scripts/webkitpy/common/system/executive.py:
1863         * Scripts/webkitpy/common/system/file_lock.py:
1864         * Scripts/webkitpy/common/system/user.py:
1865         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1866         * Scripts/webkitpy/layout_tests/models/result_summary.py:
1867         * Scripts/webkitpy/layout_tests/models/test_expectations.py:
1868         * Scripts/webkitpy/layout_tests/port/chromium.py:
1869         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1870         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1871         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1872         * Scripts/webkitpy/layout_tests/port/gtk.py:
1873         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1874         * Scripts/webkitpy/layout_tests/port/mac.py:
1875         * Scripts/webkitpy/layout_tests/port/qt.py:
1876         * Scripts/webkitpy/layout_tests/port/server_process.py:
1877         * Scripts/webkitpy/layout_tests/port/webkit.py:
1878         * Scripts/webkitpy/layout_tests/port/win.py:
1879         * Scripts/webkitpy/layout_tests/servers/apache_http_server.py:
1880         * Scripts/webkitpy/layout_tests/servers/http_server.py:
1881         * Scripts/webkitpy/layout_tests/servers/http_server_base.py:
1882         * Scripts/webkitpy/layout_tests/servers/websocket_server.py:
1883         * Scripts/webkitpy/layout_tests/views/printing.py:
1884         * Scripts/webkitpy/python24/versioning.py:
1885         * Scripts/webkitpy/style/checker.py:
1886         * Scripts/webkitpy/style/checkers/test_expectations.py:
1887         * Scripts/webkitpy/style/patchreader.py:
1888         * Scripts/webkitpy/to_be_moved/update_webgl_conformance_tests.py:
1889
1890 2011-07-06  Adam Barth  <abarth@webkit.org>
1891
1892         Move model classes out of layout_package and into models
1893         https://bugs.webkit.org/show_bug.cgi?id=64053
1894
1895         Reviewed by Eric Seidel.
1896
1897         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1898         * Scripts/webkitpy/common/net/layouttestresults.py:
1899         * Scripts/webkitpy/common/net/layouttestresults_unittest.py:
1900         * Scripts/webkitpy/common/net/resultsjsonparser.py:
1901         * Scripts/webkitpy/common/net/resultsjsonparser_unittest.py:
1902         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1903         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
1904         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py:
1905         * Scripts/webkitpy/layout_tests/controllers/worker.py:
1906         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
1907         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1908         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
1909         * Scripts/webkitpy/layout_tests/layout_package/result_summary.py: Removed.
1910         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py: Removed.
1911         * Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py: Removed.
1912         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py: Removed.
1913         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py: Removed.
1914         * Scripts/webkitpy/layout_tests/layout_package/test_input.py: Removed.
1915         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py:
1916         * Scripts/webkitpy/layout_tests/layout_package/test_results.py: Removed.
1917         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py: Removed.
1918         * Scripts/webkitpy/layout_tests/models: Added.
1919         * Scripts/webkitpy/layout_tests/models/__init__.py: Added.
1920         * Scripts/webkitpy/layout_tests/models/result_summary.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/result_summary.py.
1921         * Scripts/webkitpy/layout_tests/models/test_expectations.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_expectations.py.
1922         * Scripts/webkitpy/layout_tests/models/test_expectations_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py.
1923         * Scripts/webkitpy/layout_tests/models/test_failures.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_failures.py.
1924         * Scripts/webkitpy/layout_tests/models/test_failures_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py.
1925         * Scripts/webkitpy/layout_tests/models/test_input.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_input.py.
1926         * Scripts/webkitpy/layout_tests/models/test_results.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_results.py.
1927         * Scripts/webkitpy/layout_tests/models/test_results_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py.
1928         * Scripts/webkitpy/layout_tests/port/chromium.py:
1929         * Scripts/webkitpy/layout_tests/views/printing.py:
1930         * Scripts/webkitpy/layout_tests/views/printing_unittest.py:
1931         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
1932         * Scripts/webkitpy/style_references.py:
1933         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests.py:
1934         * Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
1935         * Scripts/webkitpy/tool/bot/flakytestreporter_unittest.py:
1936         * Scripts/webkitpy/tool/commands/queues_unittest.py:
1937         * Scripts/webkitpy/tool/commands/rebaseline.py:
1938
1939 2011-07-06  Dirk Pranke  <dpranke@chromium.org>
1940
1941         nrwt: fix typo preventing http startup on windows
1942         https://bugs.webkit.org/show_bug.cgi?id=64050
1943
1944         Reviewed by Eric Siedel.
1945
1946         NRWT checks whether it needs to start the servers by looking
1947         for "/http/" in the test names. After r90520, the leading slash
1948         is not present, so the check isn't quite right.
1949
1950         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1951         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
1952
1953 2011-07-06  Dirk Pranke  <dpranke@chromium.org>
1954
1955         nrwt: remove --use-apache from the command line
1956         https://bugs.webkit.org/show_bug.cgi?id=63358
1957
1958         Reviewed by Adam Barth.
1959
1960         Re-land the fix in bug 63358 - the prior fix was flawed in that
1961         we were trying to start the http server during check_sys_deps()
1962         but not actually setting up the environment properly. This
1963         broken prior to this change, but exposed by it, since the patch
1964         changes the chromium port to call the base class's checks.
1965
1966         * Scripts/webkitpy/common/system/executive.py:
1967         * Scripts/webkitpy/layout_tests/port/base.py:
1968         * Scripts/webkitpy/layout_tests/port/chromium.py:
1969         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1970         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1971         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1972         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
1973         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
1974         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1975
1976 2011-07-06  Adam Barth  <abarth@webkit.org>
1977
1978         Move view-related code out of layout_package and into views
1979         https://bugs.webkit.org/show_bug.cgi?id=64051
1980
1981         Reviewed by Eric Seidel.
1982
1983         * Scripts/webkitpy/layout_tests/controllers/manager.py:
1984         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py:
1985         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py:
1986         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py:
1987         * Scripts/webkitpy/layout_tests/layout_package/metered_stream.py: Removed.
1988         * Scripts/webkitpy/layout_tests/layout_package/metered_stream_unittest.py: Removed.
1989         * Scripts/webkitpy/layout_tests/layout_package/printing.py: Removed.
1990         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py: Removed.
1991         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1992         * Scripts/webkitpy/layout_tests/views: Added.
1993         * Scripts/webkitpy/layout_tests/views/__init__.py: Added.
1994         * Scripts/webkitpy/layout_tests/views/metered_stream.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/metered_stream.py.
1995         * Scripts/webkitpy/layout_tests/views/metered_stream_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/metered_stream_unittest.py.
1996         * Scripts/webkitpy/layout_tests/views/printing.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/printing.py.
1997         * Scripts/webkitpy/layout_tests/views/printing_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py.
1998
1999 2011-07-06  Adam Barth  <abarth@webkit.org>
2000
2001         Add goofy comments to explain why we have these files.
2002
2003         * Scripts/webkitpy/layout_tests/controllers/__init__.py:
2004         * Scripts/webkitpy/layout_tests/layout_package/__init__.py:
2005
2006 2011-07-06  Adam Barth  <abarth@webkit.org>
2007
2008         Move a bunch of classes out of layout_tests.layout_package and into layout_tests.controllers
2009         https://bugs.webkit.org/show_bug.cgi?id=64048
2010
2011         Reviewed by Eric Seidel.
2012
2013         * Scripts/webkitpy/layout_tests/controllers: Added.
2014         * Scripts/webkitpy/layout_tests/controllers/__init__.py: Added.
2015         * Scripts/webkitpy/layout_tests/controllers/manager.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/manager.py.
2016         * Scripts/webkitpy/layout_tests/controllers/manager_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py.
2017         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py.
2018         * Scripts/webkitpy/layout_tests/controllers/manager_worker_broker_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py.
2019         * Scripts/webkitpy/layout_tests/controllers/message_broker.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/message_broker.py.
2020         * Scripts/webkitpy/layout_tests/controllers/message_broker_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py.
2021         * Scripts/webkitpy/layout_tests/controllers/single_test_runner.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py.
2022         * Scripts/webkitpy/layout_tests/controllers/worker.py: Copied from Tools/Scripts/webkitpy/layout_tests/layout_package/worker.py.
2023         * Scripts/webkitpy/layout_tests/layout_package/manager.py: Removed.
2024         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py: Removed.
2025         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py: Removed.
2026         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py: Removed.
2027         * Scripts/webkitpy/layout_tests/layout_package/message_broker.py: Removed.
2028         * Scripts/webkitpy/layout_tests/layout_package/message_broker_unittest.py: Removed.
2029         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
2030         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py: Removed.
2031         * Scripts/webkitpy/layout_tests/layout_package/worker.py: Removed.
2032         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2033
2034 2011-07-06  Adam Barth  <abarth@webkit.org>
2035
2036         Attempt to fix unittest on Leopard.  Let's enumerate the keys in sorted
2037         order so the test results are predictable.
2038
2039         * Scripts/webkitpy/tool/bot/irc_command.py:
2040         * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:
2041
2042 2011-07-06  Dirk Pranke  <dpranke@chromium.org>
2043
2044         Reviewed by Eric Siedel.
2045
2046         nrwt: make the code be consistent about using test names instead of filenames or paths
2047
2048         https://bugs.webkit.org/show_bug.cgi?id=63597
2049
2050         It can be hard to figure out where in the code we're assuming
2051         test names are listed as unix-style relative filenames and where
2052         they are either absolute paths or relative paths following the
2053         host filesystem convention.
2054
2055         This patch changes things so that everything outside of the
2056         Port object uses (and must assume) unix-style relative
2057         filenames (with one exception, which is specifying host-local
2058         filenames as a list of test arguments on the command line).
2059
2060         This should make things clearer and more consistent, and also
2061         removes a lot of path conversion calls.
2062
2063         The changes in this patch outside of port/base.py are largely
2064         cosmetic changes of "filename" to "test". We add a few routines
2065         to the base Port implementation to manipulate test names to
2066         split them, figure out if we're referring to a directory of
2067         tests, etc.
2068
2069         * Scripts/webkitpy/common/net/layouttestresults.py:
2070         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
2071         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
2072         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
2073         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
2074         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
2075         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
2076         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
2077         * Scripts/webkitpy/layout_tests/layout_package/result_summary.py:
2078         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
2079         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
2080         * Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
2081         * Scripts/webkitpy/layout_tests/layout_package/test_input.py:
2082         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py:
2083         * Scripts/webkitpy/layout_tests/layout_package/test_results.py:
2084         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:
2085         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
2086         * Scripts/webkitpy/layout_tests/port/base.py:
2087         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2088         * Scripts/webkitpy/layout_tests/port/chromium.py:
2089         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2090         * Scripts/webkitpy/layout_tests/port/dryrun.py:
2091         * Scripts/webkitpy/layout_tests/port/test.py:
2092         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2093         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2094         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests.py:
2095         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests_unittest.py:
2096         * Scripts/webkitpy/tool/bot/flakytestreporter.py:
2097         * Scripts/webkitpy/tool/commands/rebaseline.py:
2098         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
2099
2100 2011-07-06  Chris Guillory  <ctguil@chromium.org>
2101
2102         No review necessary.
2103
2104         Adding myself to committers list in committers.py.
2105         https://bugs.webkit.org/show_bug.cgi?id=63478
2106
2107         * Scripts/webkitpy/common/config/committers.py:
2108
2109 2011-07-06  Johnny Ding  <jnd@chromium.org>
2110
2111         https://bugs.webkit.org/show_bug.cgi?id=63857
2112         Implement createPopupMenu for Chromium DRT.
2113
2114         Reviewed by Tony Chang.
2115
2116         * DumpRenderTree/chromium/WebViewHost.cpp:
2117         (WebViewHost::createPopupMenu):
2118         (WebViewHost::~WebViewHost):
2119         * DumpRenderTree/chromium/WebViewHost.h:
2120
2121 2011-07-06  Mihai Parparita  <mihaip@chromium.org>
2122
2123         Unreviewed, rolling out r90503.
2124         http://trac.webkit.org/changeset/90503
2125         https://bugs.webkit.org/show_bug.cgi?id=63358
2126
2127         Makes Chromium Windows canary bots hang.
2128
2129         * Scripts/webkitpy/layout_tests/port/base.py:
2130         * Scripts/webkitpy/layout_tests/port/chromium.py:
2131         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2132         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2133         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2134         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
2135         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
2136         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2137
2138 2011-07-06  Adam Barth  <abarth@webkit.org>
2139
2140         Move buildbot URL in webkitpy to config.urls
2141         https://bugs.webkit.org/show_bug.cgi?id=64041
2142
2143         Reviewed by Eric Seidel.
2144
2145         This patch prepares for using the buildbot class to talk to the
2146         Chromium buildbots.
2147
2148         * Scripts/webkitpy/common/config/urls.py:
2149         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
2150         * Scripts/webkitpy/tool/commands/download_unittest.py:
2151         * Scripts/webkitpy/tool/commands/queries.py:
2152         * Scripts/webkitpy/tool/mocktool.py:
2153         * Scripts/webkitpy/tool/steps/ensurebuildersaregreen.py:
2154
2155 2011-07-06  Kevin Ollivier  <kevino@theolliviers.com>
2156
2157         [wx] Unreviewed build fix, add the correct path to the waf build settings.
2158         
2159         * wx/packaging/build-win-installer.py:
2160
2161 2011-07-06  Adam Barth  <abarth@webkit.org>
2162
2163         Add missing include.
2164
2165         * Scripts/webkitpy/tool/bot/irc_command.py:
2166
2167 2011-07-06  Adam Barth  <abarth@webkit.org>
2168
2169         Add roll-chromium-deps command to sheriff-bot
2170         https://bugs.webkit.org/show_bug.cgi?id=64037
2171
2172         Reviewed by Eric Seidel.
2173
2174         Lo, and the DEPS were rolled on command.
2175
2176         * Scripts/webkitpy/tool/bot/irc_command.py:
2177         * Scripts/webkitpy/tool/bot/irc_command_unittest.py:
2178         * Scripts/webkitpy/tool/bot/sheriff.py:
2179         * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:
2180         * Scripts/webkitpy/tool/commands/roll.py:
2181         * Scripts/webkitpy/tool/commands/roll_unittest.py:
2182         * Scripts/webkitpy/tool/steps/updatechromiumdeps.py:
2183
2184 2011-07-06  Eric Seidel  <eric@webkit.org>
2185
2186         Split Driver, DriverOutput and DriverInput out into a new driver.py file
2187         https://bugs.webkit.org/show_bug.cgi?id=64031
2188
2189         Reviewed by Adam Barth.
2190
2191         This ended up kinda large and I'm happy to break it up if desired.
2192
2193         As part of cleaning up the Port module, I split Driver, DriverInput and DriverOutput
2194         out into its own driver.py file.
2195
2196         I also cleaned up how we import Port objects, by making port/__init__.py
2197         expose Port, Driver, DriverInput, and DriverOutput.  I removed exposure of
2198         some various test objects as they were causing circular dependencies.
2199         base.py is now an implementation detail of port.  No longer is "base" referenced
2200         anywhere in the code.
2201
2202         I also made Driver.__init__ non-virtual so that sub-classers could all share
2203         the data storage on Driver.  This more closely matches how other subclasses
2204         work throughout our python code.
2205
2206         Standard line-unwrapping and doc-string removal as I went through the code.
2207
2208         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2209         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
2210         * Scripts/webkitpy/layout_tests/port/__init__.py:
2211         * Scripts/webkitpy/layout_tests/port/base.py:
2212         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2213         * Scripts/webkitpy/layout_tests/port/chromium.py:
2214         * Scripts/webkitpy/layout_tests/port/driver.py: Added.
2215         * Scripts/webkitpy/layout_tests/port/dryrun.py:
2216         * Scripts/webkitpy/layout_tests/port/mock_drt.py:
2217         * Scripts/webkitpy/layout_tests/port/test.py:
2218         * Scripts/webkitpy/layout_tests/port/webkit.py:
2219         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2220         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests_unittest.py:
2221         * Scripts/webkitpy/tool/servers/rebaselineserver_unittest.py:
2222
2223 2011-07-06  Adam Roben  <aroben@apple.com>
2224
2225         Clear TestFailures's getNumberOfFailingTests cache
2226
2227         Now that we support NRWT, we need to clear out old zero values for the NRWT bots that might
2228         have been cached by old versions of TestFailures.
2229
2230         Fixes <http://webkit.org/b/64029> TestFailures page still doesn't show NRWT bots if you've
2231         used it recently
2232
2233         Reviewed by Adam Barth.
2234
2235         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
2236         (Builder.prototype.getNumberOfFailingTests): Added a .version property to the cache data
2237         that must match the current version. Otherwise we'll ignore the cached data.
2238
2239 2011-06-24  Dirk Pranke  <dpranke@chromium.org>
2240
2241         Reviewed by Adam Barth.
2242
2243         nrwt: remove --use-apache from the command line
2244         https://bugs.webkit.org/show_bug.cgi?id=63358
2245
2246         This change removes the --use-apache command line argument. It
2247         was initially put in when we were trying to get the cygwin
2248         apache instance to work with Chromium win, but that code has
2249         bitrotted and doesn't work at all now.
2250
2251         Arguably we should remove all of the code to allow for a choice
2252         of web servers, but since we may still want to switch off of
2253         LigHTTPd at some point on Windows, I'll leave the rest in for
2254         now (or at least yank it in a different change.)
2255
2256         * Scripts/webkitpy/layout_tests/port/base.py:
2257         * Scripts/webkitpy/layout_tests/port/chromium.py:
2258         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2259         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2260         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
2261         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2262
2263 2011-07-06  Dimitri Glazkov  <dglazkov@chromium.org>
2264
2265         garden-o-matic should use a favicon to indicate current state.
2266         https://bugs.webkit.org/show_bug.cgi?id=64027
2267
2268         Reviewed by Adam Barth.
2269
2270         * Scripts/webkitpy/tool/servers/data/gardeningserver/favicon-green.png: Added.
2271         * Scripts/webkitpy/tool/servers/data/gardeningserver/favicon-red.png: Added.
2272         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html: Added blurb about icons.
2273         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js: Added setting icon state.
2274         * Scripts/webkitpy/tool/servers/gardeningserver.py: Allowed the use of favicons.
2275
2276 2011-07-06  Daniel Bates  <dbates@rim.com>
2277
2278         Fix webkitpy error after changeset 90483 <http://trac.webkit.org/changeset/90483>; Add missing
2279         parameter to _fetch_file_from_results lamdba function.
2280
2281         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2282
2283 2011-07-06  Kevin Ollivier  <kevino@theolliviers.com>
2284
2285         [wx] Unreviewed build fix. Generate the proper install_name for dylibs on Mac.
2286
2287         * wx/packaging/build-mac-installer.py:
2288
2289 2011-07-06  Xan Lopez  <xlopez@igalia.com>
2290
2291         [GTK] Only set env. variables in start when strictly needed
2292         https://bugs.webkit.org/show_bug.cgi?id=64026
2293
2294         Reviewed by Eric Seidel.
2295
2296         * Scripts/webkitpy/layout_tests/port/gtk.py: add a
2297         setup_environ_for_server method in GtkPort and set as many
2298         env. variables as possible there.
2299
2300 2011-07-06  Adam Roben  <aroben@apple.com>
2301
2302         Teach TestFailures how to load, parse, and interpret NRWT test results
2303
2304         Fixes <http://webkit.org/b/61877> TestFailures page doesn't show testers that use
2305         new-run-webkit-tests
2306
2307         Reviewed by Adam Barth.
2308
2309         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
2310         (Builder.prototype.failureDiagnosisTextAndURL): Added support for the new 'flaky' failure
2311         type. For now we don't account for image-only flakes (but TestFailures doesn't deal with
2312         pixel tests at all currently).
2313         (Builder.prototype.getNumberOfFailingTests): Relaxed the regex used for parsing the number
2314         of failing tests from the buildbot output. Make sure not to count "new passes" (a new
2315         category introduced by NRWT) as failures.
2316
2317         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
2318         (LayoutTestResultsLoader.prototype.start): Moved a bunch of code from here to
2319         _fetchAndParseORWTResults. This function now attempts to load NRWT results, then falls back
2320         to loading ORWT results.
2321         (LayoutTestResultsLoader.prototype._fetchAndParseNRWTResults): Added. Tries to load and
2322         parse the full_results.json file for the given build.
2323         (LayoutTestResultsLoader.prototype._fetchAndParseORWTResults): Added. Code came from start.
2324         Fixed a bug along the way where we were sometimes calling the error callback instead of the
2325         success callback when all tests passed.
2326
2327         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/NRWTResultsParser.js: Added.
2328         (NRWTResultsParser): Do-nothing constructor.
2329         (NRWTResultsParser.prototype.parse): Uses eval() (eww!) to get the test results out of the
2330         JS string, then iterates over all the tests in the results data and builds up a data
2331         structure matching what ORWTResultsParser returns.
2332
2333         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
2334         (Array.prototype.contains):
2335         (String.prototype.contains):
2336         Added these simple helper functions.
2337
2338         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/index.html: Pull in
2339         NRWTResultsParser.js.
2340
2341 2011-07-06  Adam Roben  <aroben@apple.com>
2342
2343         Extract code to parse ORWT's results.html file into its own class
2344
2345         Prep work for fixing <http://webkit.org/b/61877> TestFailures page doesn't show testers that
2346         use new-run-webkit-tests
2347
2348         Reviewed by Adam Barth.
2349
2350         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
2351         (LayoutTestResultsLoader.prototype.start): Moved a bunch of code from here...
2352
2353         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ORWTResultsParser.js: Added.
2354         (ORWTResultsParser):
2355         (ORWTResultsParser.prototype.parse):
2356         ...to here.
2357
2358         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/index.html: Pull in
2359         ORWTResultsParser.js.
2360
2361 2011-07-06  Eric Seidel  <eric@webkit.org>
2362
2363         Unreviewed.  Fix a unittest from hitting the network. :)
2364
2365         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2366
2367 2011-07-06  Eric Seidel  <eric@webkit.org>
2368
2369         Unreviewed.  Fixing an exception in test-webkitpy, caused by bug 64006.
2370
2371         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
2372
2373 2011-07-06  Gustavo Noronha Silva  <gustavo.noronha@collabora.com>
2374
2375         Run an Xvfb instance to run the API tests on, like we do for the
2376         layout tests.
2377
2378         Reviewed by Xan Lopez.
2379
2380         * Scripts/run-gtk-tests:
2381
2382 2011-07-06  Daniel Bates  <dbates@webkit.org>
2383
2384         Make SCM unit tests faster
2385         https://bugs.webkit.org/show_bug.cgi?id=63883
2386
2387         Speeds up the SCM unit tests by a factor of 4. Currently, we create a
2388         mock SVN repo for each test_ method in SVNTest and GitTest and creating
2389         this repo is expensive.
2390
2391         Instead, it is significantly faster to create the mock SVN repo once
2392         and then perform a filesystem copy of it for each test_ method.
2393
2394         Note, Python 2.7's unittest module implements support for per class and
2395         per module setup and tear down methods which could be used to implement
2396         similar functionality. At the time of writing, test-webkitpy is designed
2397         to support Python 2.5. So, we can't take advantage of these Python 2.7
2398         features :(
2399
2400         Reviewed by Eric Seidel.
2401
2402         * Scripts/webkitpy/common/checkout/scm/scm_unittest.py:
2403
2404 2011-07-06  Adam Barth  <abarth@webkit.org>
2405
2406         Remove unused parameter from _add_test in test_expectations.py
2407         https://bugs.webkit.org/show_bug.cgi?id=64008
2408
2409         Reviewed by Eric Seidel.
2410
2411         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
2412
2413 2011-07-06  Adam Barth  <abarth@webkit.org>
2414
2415         Use old-run-webkit-tests when --leaks is present
2416         https://bugs.webkit.org/show_bug.cgi?id=64012
2417
2418         Reviewed by Adam Roben.
2419
2420         We'll remove this branch once support of --leaks is added to
2421         new-run-webkit-tests.
2422
2423         * Scripts/run-webkit-tests:
2424
2425 2011-07-06  Benjamin Poulain  <benjamin@webkit.org>
2426
2427         [Qt] [WK2] Add touch mocking to Qt's MiniBrowser
2428         https://bugs.webkit.org/show_bug.cgi?id=63995
2429
2430         Reviewed by Andreas Kling.
2431
2432         The mouse events are intercepted through QCoreApplication::notify(),
2433         and fake touch events are generated for mouse events when necessary.
2434
2435         If touch events are received from the platform, we assume the current
2436         hardware has a touch screen and we stop generating fake events.
2437
2438         * MiniBrowser/qt/MiniBrowserApplication.cpp:
2439         (isTouchEvent):
2440         (isMouseEvent):
2441         (MiniBrowserApplication::MiniBrowserApplication):
2442         (MiniBrowserApplication::notify):
2443         * MiniBrowser/qt/MiniBrowserApplication.h:
2444
2445 2011-07-06  Xan Lopez  <xlopez@igalia.com>
2446
2447         Reviewed by Gustavo Noronha.
2448
2449         * Scripts/webkitpy/layout_tests/port/gtk.py: use Popen correctly.
2450
2451 2011-07-06  Eric Seidel  <eric@webkit.org>
2452
2453         webkit-patch failure-reason does not understand NRWT results
2454         https://bugs.webkit.org/show_bug.cgi?id=64006
2455
2456         Reviewed by Adam Barth.
2457
2458         Adam Barth tells me failure-reason should be deleted soon,
2459         but aroben's fancy new TestFailures/ page doesn't work for
2460         NRWT yet, and I needed to know when
2461         third-party-cookie-relaxing started failing.
2462
2463         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
2464         * Scripts/webkitpy/tool/commands/queries.py:
2465
2466 2011-07-06  Xan Lopez  <xlopez@igalia.com>
2467
2468         [GTK] Add missing environment variables in NWRT
2469         https://bugs.webkit.org/show_bug.cgi?id=64004
2470
2471         Reviewed by Gustavo Noronha Silva.
2472
2473         Add missing environment variables needed by the GTK+ port.
2474
2475         * Scripts/webkitpy/layout_tests/port/gtk.py: ditto.
2476
2477 2011-07-06  Adam Roben  <aroben@apple.com>
2478
2479         Make run-api-tests output intelligible on the bots
2480
2481         run-api-tests's output on the bots was very hard to read for two reasons:
2482           1) It was mixing its own output with gtest's
2483           2) It was using ANSI escape sequences to print colored text, but the bots don't support
2484              that
2485
2486         Now, the --verbose flag turns off almost all of run-api-tests's own output so that gtest's
2487         output will not be obscured. We still print "Timeout" messages even in verbose mode, since
2488         gtest doesn't have any native support for timeouts. Also, when our output is being
2489         redirected to a file, we don't print ANSI escape sequences.
2490
2491         Fixes <http://webkit.org/b/63996> It's very hard to read run-api-tests output on the bots
2492
2493         Reviewed by Sam Weinig.
2494
2495         * Scripts/run-api-tests: Don't use the :constants interface of Term::ANSIColor. We want to
2496         use the colored function instead.
2497         (runAllTestsInSuite): Don't print out suite names in verbose mode; gtest will give us enough
2498         context that they aren't needed.
2499         (runTest): Don't print out test names or pass/fail messages in verbose mode; gtest will do
2500         that for us. Also, use the new possiblyColored function instead of always coloring output.
2501         (possiblyColored): Added. When printing to a tty, returns the string with the appropriate
2502         ANSI color escape sequences added. Otherwise just returns the string unmodified.
2503
2504 2011-07-06  Xan Lopez  <xlopez@igalia.com>
2505
2506         Reviewed by Gustavo Noronha.
2507
2508         Launch Xvfb (mostly) the same way we were launching it in the
2509         bots, for maximum compatibility.
2510
2511         * Scripts/webkitpy/layout_tests/port/gtk.py:
2512
2513 2011-07-06  Andreas Kling  <kling@webkit.org>
2514
2515         [Qt][WK2] Split Qt API into two different web views (touch and desktop)
2516         https://bugs.webkit.org/show_bug.cgi?id=63950
2517
2518         Co-authored with Benjamin Poulain <benjamin@webkit.org>.
2519
2520         Reviewed by Kenneth Rohde Christiansen.
2521
2522         Port WTR/Qt to use QDesktopWebView and force a 800x600 viewport.
2523
2524         The Qt MiniBrowser now defaults to QDesktopWebView but can be started
2525         in QTouchWebView mode by passing -touch on the command line.
2526
2527         A lot of functionality has been removed and/or disabled from
2528         MiniBrowser until we reintroduce the necessary API's.
2529
2530         * MiniBrowser/qt/BrowserView.cpp:
2531         (BrowserView::BrowserView):
2532         (BrowserView::~BrowserView):
2533         (BrowserView::load):
2534         (BrowserView::view):
2535         (BrowserView::touchWebView):
2536         (BrowserView::desktopWebView):
2537         * MiniBrowser/qt/BrowserView.h:
2538         * MiniBrowser/qt/BrowserWindow.cpp:
2539         (newPageFunction):
2540         (BrowserWindow::BrowserWindow):
2541         (BrowserWindow::webView):
2542         (BrowserWindow::newWindow):
2543         (BrowserWindow::showUserAgentDialog):
2544         (BrowserWindow::updateUserAgentList):
2545         * MiniBrowser/qt/BrowserWindow.h:
2546         * MiniBrowser/qt/MiniBrowserApplication.cpp:
2547         (MiniBrowserApplication::handleUserOptions):
2548         * MiniBrowser/qt/MiniBrowserApplication.h:
2549         (WindowOptions::WindowOptions):
2550         * MiniBrowser/qt/UrlLoader.cpp:
2551         (UrlLoader::UrlLoader):
2552         * MiniBrowser/qt/main.cpp:
2553         (main):
2554         * WebKitTestRunner/qt/PlatformWebViewQt.cpp:
2555         (WTR::WebView::wkView):
2556         (WTR::WebView::pageRef):
2557         (WTR::WebView::WebView):
2558         (WTR::PlatformWebView::page):
2559         (WTR::PlatformWebView::windowFrame):
2560         (WTR::PlatformWebView::setWindowFrame):
2561
2562 2011-07-06  Eric Seidel  <eric@webkit.org>
2563
2564         Unreviewed..
2565
2566         * Scripts/run-webkit-tests:
2567         (useNewRunWebKitTests):
2568          - Fix the pass-thru of --gtk and --qt for old-run-webkit-tests,
2569            it was previously broken by the existence of isGtk/isQt checks which
2570            removed the --gtk/--qt args from @ARGV.
2571
2572 2011-07-06  Eric Seidel  <eric@webkit.org>
2573
2574         new-run-webkit-tests uses a 35s timeout for Mac and 6s for all other ports
2575         https://bugs.webkit.org/show_bug.cgi?id=63983
2576
2577         Reviewed by Xan Lopez.
2578
2579         All WebKit ports want to use a 35s timeout.  Only chromium currently
2580         uses a 6s timeout.  Eventually we'll move all WebKit ports to 6s too
2581         to further speed up new-run-webkit-tests runs, but that's going to be a while.
2582
2583         * Scripts/webkitpy/layout_tests/port/mac.py:
2584         * Scripts/webkitpy/layout_tests/port/webkit.py:
2585         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
2586
2587 2011-07-06  Xan Lopez  <xlopez@igalia.com>
2588
2589         Unreviewed.
2590
2591         NWRT actually seems to work locally, so it should work in the bots
2592         as long as we take care of the TIMEOUT issue (which is being fixed
2593         as I write this). Let's give this another shot and fix whatever
2594         breaks.
2595
2596         * Scripts/run-webkit-tests:
2597         (useNewRunWebKitTests):
2598
2599 2011-07-06  Xan Lopez  <xlopez@igalia.com>
2600
2601         Unreviewed.
2602
2603         Disable NWRT on GTK. At the very least it needs an updated exected
2604         results file since NWRT is more sensitive to timeouts, not sure
2605         what the exact problem is.
2606
2607         * Scripts/run-webkit-tests:
2608         (useNewRunWebKitTests):
2609
2610 2011-07-06  Adam Barth  <abarth@webkit.org>
2611
2612         Fix global variable leak detected by noglobals.
2613
2614         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
2615
2616 2011-07-06  Adam Barth  <abarth@webkit.org>
2617
2618         garden-o-matic shouldn't show flaky tests by default
2619         https://bugs.webkit.org/show_bug.cgi?id=63979
2620
2621         Reviewed by Ojan Vafai.
2622
2623         If any of the actual test results are PASS, the test is flaky and we
2624         can ignore it (at least for now).  If we get awesome at dealing with
2625         failing tests, we might want to tackle flaky tests by refering the user
2626         to the flakiness dashboard, which is way more awesome at dealing with
2627         flaky tests that this tool ever will be.
2628
2629         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
2630         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
2631
2632 2011-07-05  Adam Barth  <abarth@webkit.org>
2633
2634         garden-o-matic should list the failing tests
2635         https://bugs.webkit.org/show_bug.cgi?id=63978
2636
2637         Reviewed by Eric Seidel.
2638
2639         * Scripts/webkitpy/tool/servers/data/gardeningserver/config.js: Added.
2640         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
2641         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js:
2642         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
2643         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js:
2644         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html:
2645         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui.js: Added.
2646         * Scripts/webkitpy/tool/servers/data/gardeningserver/ui_unittests.js: Added.
2647         * Scripts/webkitpy/tool/servers/gardeningserver.py:
2648
2649
2650 2011-07-05  Eric Seidel  <eric@webkit.org>
2651
2652         Re-enable NRWT on Leopard now that bug 63973 is fixed.
2653
2654         Reviewed by Adam Barth.
2655
2656         I also made the should-enable expression a black-list instead
2657         of a white-list since every bot on build.webkit.org is
2658         now using NRWT except windows and WebKit2 bots.
2659
2660         * Scripts/run-webkit-tests:
2661         (useNewRunWebKitTests):
2662
2663 2011-07-05  Eric Seidel  <eric@webkit.org>
2664
2665         Perl string concat failure on Leopard (causes NRWT to fail)
2666         https://bugs.webkit.org/show_bug.cgi?id=63973
2667
2668         Reviewed by Daniel Bates.
2669
2670         Leopard's "xcodebuild -version" has a different format from Snow Leopard/Lion
2671         thus the version parsing logic in webkitdirs.pm seems to fail.
2672
2673         Default to "3.0" instead of undef when we can't parse the version number.
2674         This should send us down the correct path for Leopard.
2675
2676         * Scripts/webkitdirs.pm:
2677         (determineBaseProductDir):
2678
2679 2011-07-05  Kevin Ollivier  <kevino@theolliviers.com>
2680
2681         [wx] Unreviewed build fix after addition of Source/WebCore/xml/parser dir.
2682
2683         * waf/build/settings.py:
2684
2685 2011-07-05  Eric Seidel  <eric@webkit.org>
2686
2687         Disable NRWT for Leopard until https://bugs.webkit.org/show_bug.cgi?id=63973 can be fixed.
2688
2689         Reviewed by Adam Barth.
2690
2691         * Scripts/run-webkit-tests:
2692         (useNewRunWebKitTests):
2693
2694 2011-07-05  Eric Seidel  <eric@webkit.org>
2695
2696         Make run-webkit-tests pass --verbose to new-run-webkit-tests on qt/gtk bots.
2697
2698         Reviewed by Adam Barth.
2699
2700         * Scripts/run-webkit-tests:
2701         (runningOnBuildBot):
2702
2703 2011-07-05  Adam Barth  <abarth@webkit.org>
2704
2705         Add results_unittests.js to garden-o-matic
2706         https://bugs.webkit.org/show_bug.cgi?id=63971
2707
2708         Reviewed by Eric Seidel.
2709
2710         Adding these unit tests forced me to commit to an API for the results
2711         module.  I'm not sure if this is the API we'll end up with, but it's
2712         something not entirely unreasonable.
2713
2714         Also, I've changed results.js over to WebKit style, as discussed previously.
2715
2716         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
2717         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
2718         * Scripts/webkitpy/tool/servers/data/gardeningserver/results_unittests.js: Added.
2719         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html:
2720
2721 2011-07-05  Eric Seidel  <eric@webkit.org>
2722
2723         Teach run-webkit-tests how to translate --qt and --gtk for new-run-webkit-tests.
2724
2725         Reviewed by Adam Barth.
2726
2727         * Scripts/run-webkit-tests:
2728
2729 2011-07-05  Adam Barth  <abarth@webkit.org>
2730
2731         Add first unit tests for garden-o-matic
2732         https://bugs.webkit.org/show_bug.cgi?id=63969
2733
2734         Reviewed by Eric Seidel.
2735
2736         I've also reformated base.js to match WebKit style, as discussed in the
2737         previous bug.
2738
2739         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js:
2740         * Scripts/webkitpy/tool/servers/data/gardeningserver/base_unittests.js: Added.
2741         * Scripts/webkitpy/tool/servers/data/gardeningserver/run-unittests.html: Added.
2742
2743 2011-07-05  Eric Seidel  <eric@webkit.org>
2744
2745         Move Qt, Gtk and Leopard to NRWT.
2746
2747         We'll roll out any of these if we see problems.
2748
2749         Reviewed by Adam Barth.
2750
2751         * Scripts/run-webkit-tests:
2752         (useNewRunWebKitTests):
2753
2754 2011-07-05  Adam Barth  <abarth@webkit.org>
2755
2756         garden-o-matic should understand which tests have unexpected failures
2757         https://bugs.webkit.org/show_bug.cgi?id=63965
2758
2759         Reviewed by Eric Seidel.
2760
2761         Currently, this code just logs the list of unexpected failures to the
2762         console.  A future patch will do something useful with this
2763         information.  A testing harness will also come in the next patch.
2764
2765         * Scripts/webkitpy/tool/servers/data/gardeningserver/base.js: Added.
2766         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
2767         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js:
2768         * Scripts/webkitpy/tool/servers/gardeningserver.py:
2769
2770 2011-07-05  Adam Barth  <abarth@webkit.org>
2771
2772         garden-o-matic should know how to fetch test results from the (NRWT) bots
2773         https://bugs.webkit.org/show_bug.cgi?id=63959
2774
2775         Reviewed by Eric Seidel.
2776
2777         Currently, this code just fetches the results and expectations JSON
2778         blob for Linux from the server and logs it to the console.
2779
2780         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
2781         * Scripts/webkitpy/tool/servers/data/gardeningserver/results.js: Added.
2782         * Scripts/webkitpy/tool/servers/gardeningserver.py:
2783
2784 2011-07-05  Dirk Pranke  <dpranke@chromium.org>
2785
2786         Reviewed by Eric Seidel.
2787
2788         Python tests are failing on leopard
2789         https://bugs.webkit.org/show_bug.cgi?id=63842
2790
2791         Fix the missing flush() call that appears to be needed by the
2792         logging package in python 2.5. Also fix the regressions
2793         introduced in the run_webkit_test tests when we switched from
2794         thread to processes (since processes aren't available on 2.5).
2795
2796         Also fix a missing "from __future__ import with_statement' in
2797         testfilehandler (unreviewed).
2798
2799         * Scripts/webkitpy/layout_tests/layout_package/metered_stream.py:
2800         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
2801         * TestResultServer/handlers/testfilehandler.py:
2802
2803 2011-07-05  Eric Seidel  <eric@webkit.org>
2804
2805         Reviewed by Adam Barth.
2806
2807         re-enabling new-run-webkit-tests on the Snow Leopard WebKit1 bots
2808         now that we believe the stray http-server issue to be solved.
2809
2810         * Scripts/run-webkit-tests:
2811         (useNewRunWebKitTests):
2812
2813 2011-07-05  Adam Barth  <abarth@webkit.org>
2814
2815         Turns out we need to support numerals too because that's what jQuery
2816         uses by default.
2817
2818         * TestResultServer/handlers/testfilehandler.py:
2819
2820 2011-07-05  Dirk Pranke  <dpranke@chromium.org>
2821
2822         nrwt: allow for multiple http shards
2823         https://bugs.webkit.org/show_bug.cgi?id=63116
2824
2825         Reviewed by Tony Chang.
2826
2827         This modifies the sharding logic to support multiple http
2828         shards, but for now we clamp to one shard until we can test
2829         perf impact and flakiness impact.
2830
2831         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
2832         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
2833
2834 2011-07-05  Dirk Pranke  <dpranke@chromium.org>
2835
2836         Re-land nrwt: make sharding tests needing locks less hard-coded
2837         https://bugs.webkit.org/show_bug.cgi?id=63112
2838
2839         Reviewed by Ojan Vafai.
2840
2841         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
2842         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
2843
2844 2011-07-05  Adam Barth  <abarth@webkit.org>
2845
2846         TestResultServer should support callback parameter for JSON
2847         https://bugs.webkit.org/show_bug.cgi?id=63961
2848
2849         Reviewed by Ojan Vafai.
2850
2851         This makes using jQuery.ajax so much more pleasant.
2852
2853         * TestResultServer/handlers/testfilehandler.py:
2854
2855 2011-07-05  Eric Seidel  <eric@webkit.org>
2856
2857         buildbot needs to understand whether NRWT exited early after having too many failures
2858         https://bugs.webkit.org/show_bug.cgi?id=63839
2859
2860         Reviewed by Adam Barth.
2861
2862         Fix-up the exited early messages printed by NRWT so that
2863         the buildbot can parse them as expected.
2864         It looks for lines using "if line.find('Exiting early') >= 0:"
2865
2866         I also plumbed the "early exit" status through to results.json
2867         in the form of an "interrupted" bool.  It was unclear to me
2868         if results.json already had enough information to compute this bool
2869         itself.  It's possible Ojan could come up with a better fix.
2870
2871         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
2872         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
2873         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
2874         * Scripts/webkitpy/layout_tests/layout_package/result_summary.py:
2875
2876 2011-07-05  Eric Seidel  <eric@webkit.org>
2877
2878         new-run-webkit-tests fails to start http server if one is already running
2879         https://bugs.webkit.org/show_bug.cgi?id=63956
2880
2881         Reviewed by Adam Barth.
2882
2883         * BuildSlaveSupport/kill-old-processes:
2884          - Removed ^M line endings and sorted the process names.
2885          - There are no functional changes to this file.
2886            (except that now that its using unix line endings it can be executed directly!)
2887         * Scripts/webkitpy/layout_tests/servers/apache_http_server.py:
2888          - Change the name to 'httpd' to match old-run-webkit-test paths.
2889
2890 2011-07-05  Adam Barth  <abarth@webkit.org>
2891
2892         Add basic ajax support to garden-o-matic
2893         https://bugs.webkit.org/show_bug.cgi?id=63874
2894
2895         Reviewed by Eric Seidel.
2896
2897         This patch adds some basic infrastructure to garden-o-matic.
2898         Currently, the infrastructure is just used to make the "quit" command
2899         use Ajax, but in the future, this infrastructure will be used to do
2900         more sophistocated remote proceedure calls.
2901
2902         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
2903         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js: Added.
2904         * Scripts/webkitpy/tool/servers/gardeningserver.py:
2905         * Scripts/webkitpy/tool/servers/reflectionhandler.py:
2906
2907 2011-07-05  Adam Barth  <abarth@webkit.org>
2908
2909         Add trivial garden-o-matic command (with server)
2910         https://bugs.webkit.org/show_bug.cgi?id=63872
2911
2912         Reviewed by Eric Seidel.
2913
2914         This patch adds a trivial "Hello, world" garden-o-matic command,
2915         complete with an HTTP server.  This command re-uses a bunch of code
2916         from the existing rebaseline-server command.  Over time, this command
2917         will grow to be a tool useful for gardening the WebKit tree.
2918
2919         * Scripts/webkitpy/tool/commands/__init__.py:
2920         * Scripts/webkitpy/tool/commands/abstractservercommand.py: Added.
2921         * Scripts/webkitpy/tool/commands/gardenomatic.py: Added.
2922         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
2923         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html: Added.
2924         * Scripts/webkitpy/tool/servers/gardeningserver.py: Added.
2925         * Scripts/webkitpy/tool/servers/rebaselineserver.py:
2926
2927 2011-07-05  Adam Roben  <aroben@apple.com>
2928
2929         Make prepare-ChangeLog include modified Perl functions in its ChangeLog template
2930
2931         This is a very simple first cut. Functions must start with a line that starts with "sub "
2932         and end with a line that starts with a closing brace. No leading whitespace is allowed.
2933         Package names aren't parsed at all.
2934
2935         Fixes <http://webkit.org/b/21591> prepare-ChangeLog should know how to find functions in
2936         Perl files
2937
2938         Reviewed by David Kilzer.
2939
2940         * Scripts/prepare-ChangeLog: Removed redundant code that ignored certain files when
2941         searching for function line ranges. This is already done inside the get_function_line_ranges
2942         function.
2943         (get_function_line_ranges): Cleaned up coding style a little bit. Call
2944         get_function_line_ranges_for_perl for files with .pl and .pm extensions. For files with an
2945         unknown extension or no extension, read the shebang line to try to determine the script
2946         interpreter. Call get_function_line_ranges_for_perl if the interpreter seems to be Perl.
2947         (get_function_line_ranges_for_perl): Added. Does extremely basic parsing of the file to find
2948         lines starting with "sub " or "}".
2949
2950 2011-07-05  Adam Roben  <aroben@apple.com>
2951
2952         Clean up run-api-tests output on Windows
2953
2954         We were mixing run-api-tests output with gtest output, and the result was a mess.
2955
2956         Fixes <http://webkit.org/b/63954> run-api-tests output is very confusing on Windows
2957
2958         Reviewed by Dan Bates.
2959
2960         * Scripts/run-api-tests:
2961         (runTest):
2962         (populateTests):
2963         Made Windows use the formerly-Mac-only codepaths which correctly suppress gtest output
2964         except in verbose mode. Most changes are due to indentation. Use of the arch utility is now
2965         the only Mac-specific thing in these functions. Some other platform-specific code was moved
2966         from here...
2967
2968         (prepareEnvironmentForRunningTestTool):
2969         (testToolPath):
2970         ...to here.
2971
2972 2011-07-04  Gavin Barraclough  <barraclough@apple.com>
2973
2974         Reviewed by Sam Weinig.
2975
2976         https://bugs.webkit.org/show_bug.cgi?id=16652
2977         Firefox and JavaScriptCore differ in Number.toString(integer)
2978
2979         Added forwarding header.
2980
2981         * DumpRenderTree/ForwardingHeaders/wtf/StdLibExtras.h: Added.
2982
2983 2011-07-04  Anders Carlsson  <andersca@apple.com>
2984
2985         Try to fix Qt build.
2986
2987         * DumpRenderTree/TestNetscapePlugIn/Tests/NPRuntimeRemoveProperty.cpp:
2988
2989 2011-07-04  Anders Carlsson  <andersca@apple.com>
2990
2991         NP_RemoveProperty is not called back by Safari when delete npObject.prop is encountered in JavaScript
2992         https://bugs.webkit.org/show_bug.cgi?id=63915
2993         <rdar://problem/7124300>
2994
2995         Reviewed by Sam Weinig.
2996
2997         Extend the NPRuntimeRemoveProperty to handle delete object.property from JavaScript.
2998
2999         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
3000         (PluginTest::NPN_IdentifierIsString):
3001         (PluginTest::NPN_UTF8FromIdentifier):
3002         (PluginTest::NPN_IntFromIdentifier):
3003         (PluginTest::NPN_RetainObject):
3004         (PluginTest::NPN_ReleaseObject):
3005         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
3006         (PluginTest::Object::removeProperty):
3007         (PluginTest::Object::identifierIs):
3008         (PluginTest::Object::NP_RemoveProperty):
3009         (PluginTest::Object::npClass):
3010         * DumpRenderTree/TestNetscapePlugIn/Tests/NPRuntimeRemoveProperty.cpp:
3011         (NPRuntimeRemoveProperty::TestObject::TestObject):
3012         (NPRuntimeRemoveProperty::TestObject::hasProperty):
3013         (NPRuntimeRemoveProperty::TestObject::getProperty):
3014         (NPRuntimeRemoveProperty::TestObject::removeProperty):
3015         (NPRuntimeRemoveProperty::PluginObject::PluginObject):
3016         (NPRuntimeRemoveProperty::PluginObject::~PluginObject):
3017         (NPRuntimeRemoveProperty::PluginObject::hasMethod):
3018         (NPRuntimeRemoveProperty::PluginObject::invoke):
3019         (NPRuntimeRemoveProperty::PluginObject::hasProperty):
3020         (NPRuntimeRemoveProperty::PluginObject::getProperty):
3021         (NPRuntimeRemoveProperty::NPP_GetValue):
3022
3023 2011-07-04  Leandro Pereira  <leandro@profusion.mobi>
3024
3025         Reviewed by Kent Tamura.
3026
3027         [EFL] Add LayoutTestControllerEfl.
3028         https://bugs.webkit.org/show_bug.cgi?id=61974
3029
3030         Only the core features are implemented. Most functions are actually
3031         stubs and will be implemented as soon as its features are needed.
3032         Build system changes will be made as soon as DRT is completely
3033         upstreamed.
3034
3035         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: Added.
3036         (LayoutTestController::~LayoutTestController):
3037         (LayoutTestController::addDisallowedURL):
3038         (LayoutTestController::clearBackForwardList):
3039         (LayoutTestController::copyDecodedHostName):
3040         (LayoutTestController::copyEncodedHostName):
3041         (LayoutTestController::dispatchPendingLoadRequests):
3042         (LayoutTestController::display):
3043         (LayoutTestController::counterValueForElementById):
3044         (LayoutTestController::keepWebHistory):
3045         (LayoutTestController::computedStyleIncludingVisitedInfo):
3046         (LayoutTestController::nodesFromRect):
3047         (LayoutTestController::layerTreeAsText):
3048         (LayoutTestController::pageNumberForElementById):
3049         (LayoutTestController::numberOfPages):
3050         (LayoutTestController::pageProperty):
3051         (LayoutTestController::isPageBoxVisible):
3052         (LayoutTestController::pageSizeAndMarginsInPixels):
3053         (LayoutTestController::webHistoryItemCount):
3054         (LayoutTestController::workerThreadCount):
3055         (LayoutTestController::notifyDone):
3056         (LayoutTestController::pathToLocalResource):
3057         (LayoutTestController::queueLoad):
3058         (LayoutTestController::setAcceptsEditing):
3059         (LayoutTestController::setAlwaysAcceptCookies):
3060         (LayoutTestController::setCustomPolicyDelegate):
3061         (LayoutTestController::waitForPolicyDelegate):
3062         (LayoutTestController::setScrollbarPolicy):
3063         (LayoutTestController::addOriginAccessWhitelistEntry):
3064         (LayoutTestController::removeOriginAccessWhitelistEntry):
3065         (LayoutTestController::setMainFrameIsFirstResponder):
3066         (LayoutTestController::setTabKeyCyclesThroughElements):
3067         (LayoutTestController::setUseDashboardCompatibilityMode):
3068         (LayoutTestController::setUserStyleSheetEnabled):
3069         (LayoutTestController::setUserStyleSheetLocation):
3070         (LayoutTestController::setValueForUser):
3071         (LayoutTestController::setViewModeMediaFeature):
3072         (LayoutTestController::setWindowIsKey):
3073         (LayoutTestController::setSmartInsertDeleteEnabled):
3074         (waitToDumpWatchdogFired):
3075         (LayoutTestController::setWaitToDump):
3076         (LayoutTestController::windowCount):
3077         (LayoutTestController::setPrivateBrowsingEnabled):
3078         (LayoutTestController::setJavaScriptCanAccessClipboard):
3079         (LayoutTestController::setXSSAuditorEnabled):
3080         (LayoutTestController::setFrameFlatteningEnabled):
3081         (LayoutTestController::setSpatialNavigationEnabled):
3082         (LayoutTestController::setAllowUniversalAccessFromFileURLs):
3083         (LayoutTestController::setAllowFileAccessFromFileURLs):
3084         (LayoutTestController::setAuthorAndUserStylesEnabled):
3085         (LayoutTestController::setAutofilled):
3086         (LayoutTestController::disableImageLoading):
3087         (LayoutTestController::setMockDeviceOrientation):
3088         (LayoutTestController::setMockGeolocationPosition):
3089         (LayoutTestController::setMockGeolocationError):
3090         (LayoutTestController::setGeolocationPermission):
3091         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
3092         (LayoutTestController::addMockSpeechInputResult):
3093         (LayoutTestController::setIconDatabaseEnabled):
3094         (LayoutTestController::setJavaScriptProfilingEnabled):
3095         (LayoutTestController::setSelectTrailingWhitespaceEnabled):
3096         (LayoutTestController::setPopupBlockingEnabled):
3097         (LayoutTestController::setPluginsEnabled):
3098         (LayoutTestController::elementDoesAutoCompleteForElementWithId):
3099         (LayoutTestController::execCommand):
3100         (LayoutTestController::findString):
3101         (LayoutTestController::isCommandEnabled):
3102         (LayoutTestController::setCacheModel):
3103         (LayoutTestController::setPersistentUserStyleSheetLocation):
3104         (LayoutTestController::clearPersistentUserStyleSheet):
3105         (LayoutTestController::clearAllApplicationCaches):
3106         (LayoutTestController::setApplicationCacheOriginQuota):
3107         (LayoutTestController::clearApplicationCacheForOrigin):
3108         (LayoutTestController::localStorageDiskUsageForOrigin):
3109         (LayoutTestController::originsWithApplicationCache):
3110         (LayoutTestController::applicationCacheDiskUsageForOrigin):
3111         (LayoutTestController::clearAllDatabases):
3112         (LayoutTestController::setDatabaseQuota):
3113         (LayoutTestController::originsWithLocalStorage):
3114         (LayoutTestController::deleteAllLocalStorage):
3115         (LayoutTestController::deleteLocalStorageForOrigin):
3116         (LayoutTestController::observeStorageTrackerNotifications):
3117         (LayoutTestController::syncLocalStorage):
3118         (LayoutTestController::setDomainRelaxationForbiddenForURLScheme):
3119         (LayoutTestController::goBack):
3120         (LayoutTestController::setDefersLoading):
3121         (LayoutTestController::setAppCacheMaximumSize):
3122         (LayoutTestController::pauseAnimationAtTimeOnElementWithId):
3123         (LayoutTestController::pauseTransitionAtTimeOnElementWithId):
3124         (LayoutTestController::sampleSVGAnimationForElementAtTime):
3125         (LayoutTestController::numberOfActiveAnimations):
3126         (LayoutTestController::suspendAnimations):
3127         (LayoutTestController::resumeAnimations):
3128         (LayoutTestController::overridePreference):
3129         (LayoutTestController::addUserScript):
3130         (LayoutTestController::addUserStyleSheet):
3131         (LayoutTestController::setDeveloperExtrasEnabled):
3132         (LayoutTestController::setAsynchronousSpellCheckingEnabled):
3133         (LayoutTestController::showWebInspector):
3134         (LayoutTestController::closeWebInspector):
3135         (LayoutTestController::evaluateInWebInspector):
3136         (LayoutTestController::evaluateScriptInIsolatedWorld):
3137         (LayoutTestController::removeAllVisitedLinks):
3138         (LayoutTestController::callShouldCloseOnWebView):
3139         (LayoutTestController::apiTestNewWindowDataLoadBaseURL):
3140         (LayoutTestController::apiTestGoToCurrentBackForwardItem):
3141         (LayoutTestController::setWebViewEditable):
3142         (LayoutTestController::markerTextForListItem):
3143         (LayoutTestController::authenticateSession):
3144         (LayoutTestController::setEditingBehavior):
3145         (LayoutTestController::abortModal):
3146         (LayoutTestController::hasSpellingMarker):
3147         (LayoutTestController::hasGrammarMarker):
3148         (LayoutTestController::dumpConfigurationForViewport):
3149         (LayoutTestController::setSerializeHTTPLoads):
3150         (LayoutTestController::setMinimumTimerInterval):
3151         (LayoutTestController::setTextDirection):
3152         (LayoutTestController::allowRoundingHacks):
3153
3154 2011-06-24  Vsevolod Vlasov  <vsevik@chromium.org>
3155
3156         Unreviewed, add myself as committer.
3157
3158         * Scripts/webkitpy/common/config/committers.py:
3159
3160 2011-07-03  Sheriff Bot  <webkit.review.bot@gmail.com>
3161
3162         Unreviewed, rolling out r90347.
3163         http://trac.webkit.org/changeset/90347
3164         https://bugs.webkit.org/show_bug.cgi?id=63886
3165
3166         Build breaks on Leopard, Chromium-win, WinCairo, and WinCE.
3167         (Requested by tkent on #webkit).
3168
3169         * DumpRenderTree/ForwardingHeaders/wtf/StdLibExtras.h: Removed.
3170
3171 2011-07-03  Daniel Bates  <dbates@webkit.org>
3172
3173         Import Checkout in scm_unittests.py and logging in svn.py to fix
3174         Python missing global variable errors
3175
3176         Rubber-stamped by Adam Barth.
3177
3178         * Scripts/webkitpy/common/checkout/scm/scm_unittest.py: Import Checkout.
3179         * Scripts/webkitpy/common/checkout/scm/svn.py: Import logging.
3180
3181 2011-07-02  Gavin Barraclough  <barraclough@apple.com>
3182
3183         Reviewed by Sam Weinig.
3184
3185         https://bugs.webkit.org/show_bug.cgi?id=16652
3186         Firefox and JavaScriptCore differ in Number.toString(integer)
3187
3188         Added forwarding header.
3189
3190         * DumpRenderTree/ForwardingHeaders/wtf/StdLibExtras.h: Added.
3191
3192 2011-04-02  Robert Hogan  <robert@webkit.org>
3193
3194         Reviewed by Benjamin Poulain.
3195
3196         [Qt] Support third-party cookie policy for Qt clients
3197         https://bugs.webkit.org/show_bug.cgi?id=45455
3198
3199         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3200         (LayoutTestController::reset):
3201         (LayoutTestController::setAlwaysAcceptCookies):
3202         (LayoutTestController::setAlwaysBlockCookies):
3203         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3204
3205 2011-07-03  Adam Barth  <abarth@webkit.org>
3206
3207         Factor ReflectionHandler out of the rebaseline server.  Just code motion.
3208
3209         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
3210         * Scripts/webkitpy/tool/servers/rebaselineserver.py:
3211         * Scripts/webkitpy/tool/servers/reflectionhandler.py: Added.
3212
3213 2011-07-03  Adam Barth  <abarth@webkit.org>
3214
3215         Move the rebaseline server out of the commands package and into the
3216         (new) servers package.  This patch prepares for refactoring this code
3217         to share server infrastructure with the (forthcoming) garden-o-matic
3218         command.
3219
3220         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
3221         * Scripts/webkitpy/tool/servers/__init__.py: Added.
3222         * Scripts/webkitpy/tool/servers/data/rebaselineserver/loupe.js:
3223         * Scripts/webkitpy/tool/servers/data/rebaselineserver/main.css:
3224         * Scripts/webkitpy/tool/servers/data/rebaselineserver/queue.js:
3225         * Scripts/webkitpy/tool/servers/rebaselineserver.py: Added.
3226
3227 2011-07-01  Adam Barth  <abarth@webkit.org>
3228
3229         Fix typo in master.cfg, hoping to trigger a master restart.
3230
3231         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3232
3233 2011-07-01  Adam Barth  <abarth@webkit.org>
3234
3235         Turn off new-run-webkit-tests and remove some debug logging.
3236
3237         * Scripts/run-webkit-tests:
3238         * Scripts/webkitpy/layout_tests/port/http_lock.py:
3239
3240 2011-07-01  Adam Barth  <abarth@webkit.org>
3241
3242         Add some more debug logging to try to understand why we're getting
3243         stuck acquring the HTTP lock.
3244
3245         * Scripts/webkitpy/layout_tests/port/http_lock.py:
3246
3247 2011-07-01  Adam Barth  <abarth@webkit.org>
3248
3249         Refactor NRWT parsing logic in master.cfg to understand that run-webkit-tests might run NRWT
3250         https://bugs.webkit.org/show_bug.cgi?id=63854
3251
3252         Reviewed by Dirk Pranke.
3253
3254         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3255
3256 2011-07-01  Eric Seidel  <eric@webkit.org>
3257
3258         new-run-webkit-tests results does not understand that mac uses test_expectations files
3259         https://bugs.webkit.org/show_bug.cgi?id=63838
3260
3261         Reviewed by Adam Barth.
3262
3263         We've decided to "enable" test_expectations for the Mac port for now.
3264         It makes the results.html results much more confusing to read, but at least they're
3265         no longer lying to us.
3266
3267         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3268         * Scripts/webkitpy/layout_tests/port/base.py:
3269         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
3270
3271 2011-07-01  Adam Barth  <abarth@webkit.org>
3272
3273         Temporarily disable waiting for ReportCrash to see how this code
3274         affects performance on the bots.
3275
3276         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3277         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
3278
3279 2011-07-01  Dmitry Lomov  <dslomov@google.com>
3280
3281         Reviewed by Adam Roben.
3282
3283         https://bugs.webkit.org/show_bug.cgi?id=63818
3284         Test bots shouldn't attempt to build TestWebKitAPI on Windows.
3285         TestWebKitAPI builds as a part of general webkit build and is a
3286         part of build products archive.
3287
3288         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3289         * Scripts/run-api-tests:
3290
3291 2011-07-01  Adam Barth  <abarth@webkit.org>
3292
3293         We shouldn't wait for ReportCrash while the timeout is running
3294         https://bugs.webkit.org/show_bug.cgi?id=63837
3295
3296         Reviewed by Eric Seidel.
3297
3298         Instead, we should move this check to worker.py.
3299
3300         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
3301         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
3302
3303 2011-07-01  Adam Roben  <aroben@apple.com>
3304
3305         Stop setting unsafe headers on POSTs via XHR from TestFailures page
3306
3307         Fixes <http://webkit.org/b/63833> TestFailures page causes a bunch of spew in the
3308         Inspector's console about unsafe headers
3309
3310         Reviewed by Dan Bates.
3311
3312         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
3313         (fetchResource): Removed unnecessary code to set Content-Length and Connection headers.
3314
3315 2011-07-01  Eric Seidel  <eric@webkit.org>
3316
3317         Reviewed by Adam Barth.
3318
3319         Switch Snow Leopard over to new-run-webkit-tests (except for wk2)
3320         https://bugs.webkit.org/show_bug.cgi?id=63786
3321
3322         We're down to one bug blocking the "move to NRWT" bug:
3323         https://bugs.webkit.org/show_bug.cgi?id=34984
3324         so we're trying transitioning Snow Leopard to NRWT as an experiment.
3325
3326         If we run into any blocking issues we'll of course roll this back out
3327         (or anyone reading this should feel encouraged to).
3328
3329         * Scripts/run-webkit-tests:
3330
3331 2011-07-01  Adam Barth  <abarth@webkit.org>
3332
3333         Add a compat shim for --leaks.  Support for --leaks will be added in
3334         https://bugs.webkit.org/show_bug.cgi?id=63832.
3335
3336         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3337
3338 2011-07-01  Adam Barth  <abarth@webkit.org>
3339
3340         Reviewed by Darin Adler.
3341
3342         ReportCrash destabilizes new-run-webkit-tests
3343         https://bugs.webkit.org/show_bug.cgi?id=63795
3344
3345         ReportCrash chews up too many system resources that it destablizes
3346         tests run concurrently.  This patch causes us to spin down all the
3347         worker processes when ReportCrash is running.
3348
3349         Also, this patch causes the master process to tell the user we're
3350         waiting for the crash reporter rather than having that be the worker's
3351         job.  This stops the user from getting spammed with too many "waiting
3352         for crash reporter" messages and also cleans up some minor UI issues
3353         with the meter.
3354
3355         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3356         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
3357         * Scripts/webkitpy/layout_tests/port/base.py:
3358         * Scripts/webkitpy/layout_tests/port/server_process.py:
3359
3360 2011-07-01  Adam Roben  <aroben@apple.com>
3361
3362         Teach webkitpy about the new format of our ChangeLog template
3363
3364         r90229 moved the "Need a short description and bug URL" line to the top of the ChangeLog
3365         template. But webkitpy didn't know this, so the various rollout-related commands were
3366         leaving that line in the ChangeLog.
3367
3368         Fixes <http://webkit.org/b/63815> REGRESSION (r90229): webkit-patch rollout and sheriffbot
3369         rollout are broken
3370
3371         Reviewed by Dan Bates.
3372
3373         * Scripts/webkitpy/common/checkout/changelog.py:
3374         (ChangeLog.update_with_unreviewed_message): Changed to look for the "Need a short
3375         description and bug URL" line instead of the "Reviewed by" line as the first line of the
3376         ChangeLog template.
3377
3378         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
3379         (ChangeLogTest): Updated _new_entry_boilerplate to match what prepare-ChangeLog now
3380         generates.
3381
3382 2011-07-01  Adam Roben  <aroben@apple.com>
3383
3384         Make TestFailures stop loading old builds once we've explained all current failures
3385
3386         We used to do this, but some refactoring in r90165 broke this logic.
3387
3388         Fixes <http://webkit.org/b/63808> REGRESSION (r90165): TestFailures page always loads all
3389         builds back to the beginning of time
3390
3391         Reviewed by Dan Bates.
3392
3393         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
3394         (LayoutTestHistoryAnalyzer.prototype.start): Stop fetching old builds once we're told that
3395         all current failures have been explained.
3396
3397 2011-07-01  Adam Roben  <aroben@apple.com>
3398
3399         commit-log-editor reorders ChangeLog entries in unexpected ways
3400         https://bugs.webkit.org/show_bug.cgi?id=63804
3401
3402         Reviewed by Darin Adler.
3403
3404         commit-log-editor was changed in r46899 to move the bug title and URL to the top of the
3405         commit message in order to make git-based tools, which expect the first line of the commit
3406         message to be a summary of the change, to work better. But this reordering of the ChangeLog
3407         entry was unpredictable and unexpected.
3408
3409         I changed prepare-ChangeLog to put the bug title and URL at the top of the ChangeLog entry
3410         so that commit-log-editor doesn't have to move them. Then I changed commit-log-editor not to
3411         try to move them.
3412
3413         * Scripts/commit-log-editor: Partially reverted r46899. We no longer save the "Reviewed by"
3414         line and try to move it down below the bug title and URL. The order of the text in the
3415         ChangeLog is now preserved, except that we insert a "Patch by" line just above the "Reviewed
3416         by" line in cases where the committer and author are not the same person.
3417
3418         * Scripts/prepare-ChangeLog: Moved the "Reviewed by" line down below the bug title and URL.
3419
3420 2011-07-01  Mihnea Ovidenie  <mihnea@adobe.com>
3421
3422         Reviewed by Adam Roben.
3423
3424         [CSSRegions] Add build slave
3425         https://bugs.webkit.org/show_bug.cgi?id=62372
3426
3427         * BuildSlaveSupport/build.webkit.org-config/config.json:
3428         Add an entry for a release build of Mac.
3429
3430 2011-07-01  Sheriff Bot  <webkit.review.bot@gmail.com>
3431
3432         Unreviewed, rolling out r90213.
3433         http://trac.webkit.org/changeset/90213
3434         https://bugs.webkit.org/show_bug.cgi?id=63792
3435
3436         The SL bot has an errant HTTP server holding port 8080
3437         (Requested by abarth on #webkit).
3438
3439         * Scripts/run-webkit-tests:
3440
3441 2011-07-01  Carlos Garcia Campos  <cgarcia@igalia.com>
3442
3443         Unreviewed. Fix WebKit2 GTK build after r90163.
3444
3445         * MiniBrowser/gtk/WebBundle/WebBundleMain.c:
3446         (didCreatePage):
3447
3448 2011-07-01  Adam Barth  <abarth@webkit.org>
3449
3450         Exit faster when we get a keyboard interrupt.  Previously, we'd open
3451         Safari if you ^C NRWT after it found a failure.
3452
3453         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3454
3455 2011-07-01  Eric Seidel  <eric@webkit.org>
3456
3457         Reviewed by Adam Barth.
3458
3459         Switch Snow Leopard over to new-run-webkit-tests (except for wk2)
3460         https://bugs.webkit.org/show_bug.cgi?id=63786
3461
3462         We're down to one bug blocking the "move to NRWT" bug:
3463         https://bugs.webkit.org/show_bug.cgi?id=34984
3464         so we're trying transitioning Snow Leopard to NRWT as an experiment.
3465
3466         If we run into any blocking issues we'll of course roll this back out
3467         (or anyone reading this should feel encouraged to).
3468
3469         * Scripts/run-webkit-tests:
3470
3471 2011-07-01  Sheriff Bot  <webkit.review.bot@gmail.com>
3472
3473         Unreviewed, rolling out r90192.
3474         http://trac.webkit.org/changeset/90192
3475         https://bugs.webkit.org/show_bug.cgi?id=63788
3476
3477         Appears to have caused NRWT on Chromium WebKit Vista to hang
3478         (Requested by abarth on #webkit).
3479
3480         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3481         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
3482
3483 2011-06-30  Adam Barth  <abarth@webkit.org>
3484
3485         Reviewed by Eric Seidel.
3486
3487         Remove the concept of "being wedged" from new-run-webkit-tests
3488         https://bugs.webkit.org/show_bug.cgi?id=63767
3489
3490         Worker processes shouldn't ever become wedged.  My understanding is
3491         that this code was originally motivated by the old threading-based
3492         design but no longer servers any purpose.
3493
3494         Note: If we actually have a problem with the test harness getting
3495         stuck, buildbot will kill us.
3496
3497         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3498         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py:
3499         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
3500
3501 2011-06-30  Adam Barth  <abarth@webkit.org>
3502
3503         The win32 implementation of fork is sad about forking test-webkitpy.
3504         Hopefully this will heal the bot.
3505
3506         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
3507
3508 2011-06-30  Eric Seidel  <eric@webkit.org>
3509
3510         Reviewed by Adam Barth.
3511
3512         Update new-run-webkit-tests now that the wk2 shared skipped list has moved
3513         https://bugs.webkit.org/show_bug.cgi?id=63780
3514
3515         The shared list was moved from mac-wk2 to wk2 as part of
3516         http://trac.webkit.org/changeset/90125.
3517
3518         * Scripts/webkitpy/layout_tests/port/webkit.py:
3519         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
3520
3521 2011-06-30  Kenichi Ishibashi  <bashi@chromium.org>
3522
3523         Reviewed by Darin Adler.
3524
3525         Style Checker should flag removal of "developmentRegion = English;" from project.pbxproj.
3526         https://bugs.webkit.org/show_bug.cgi?id=62022
3527
3528         Adds XcodeProjectFileChecker to check removal of "developmentRegion".
3529
3530         * Scripts/webkitpy/style/checker.py: Added XcodeProjectFileChecker as a checker for *.pbxproj.
3531         * Scripts/webkitpy/style/checkers/xcodeproj.py: Added.
3532         * Scripts/webkitpy/style/checkers/xcodeproj_unittest.py: Added.
3533
3534 2011-06-30  Dirk Pranke  <dpranke@chromium.org>
3535
3536         Reviewed by Ojan Vafai.
3537
3538         nrwt: make sharding tests needing locks less hard-coded
3539         https://bugs.webkit.org/show_bug.cgi?id=63112
3540
3541         This change also changes the manager logic so that it will
3542         drop the server lock as soon as all of the shards requiring
3543         the lock have completed.
3544
3545         This change includes some minor namespace/import changes in the
3546         unit tests, and also makes the Manager a new-style object, which
3547         it should've been all along.
3548
3549         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3550         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
3551
3552 2011-06-30  Adam Barth  <abarth@webkit.org>
3553
3554         Reviewed by Eric Seidel.
3555
3556         Remove threaded mode from new-run-webkit-tests
3557         https://bugs.webkit.org/show_bug.cgi?id=63771
3558
3559         This mode is not used and is buggy.  Rather than carry around a bunch
3560         of unused buggy code, we should rip it out and focus on the
3561         multiprocess implementation.
3562
3563         * Scripts/webkitpy/common/system/executive.py:
3564         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py:
3565         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
3566         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
3567         * Scripts/webkitpy/layout_tests/port/base.py:
3568         * Scripts/webkitpy/layout_tests/port/chromium.py:
3569         * Scripts/webkitpy/layout_tests/port/mac.py:
3570         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
3571         * Scripts/webkitpy/layout_tests/port/server_process.py:
3572         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3573         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
3574
3575 2011-06-30  Adam Barth  <abarth@webkit.org>
3576
3577         Reviewed by Darin Adler.
3578
3579         Clean up output from new-run-webkit-tests
3580         https://bugs.webkit.org/show_bug.cgi?id=63759
3581
3582         Printing messages from the child process looks super ugly because of
3583         the way the pretty-printer works.  Printing a blank line first is a
3584         hack, but it makes things at least partially sane.
3585
3586         Also, handle the case where calling sample throws an exception.
3587
3588         * Scripts/webkitpy/layout_tests/port/server_process.py:
3589         * Scripts/webkitpy/layout_tests/port/server_process_unittest.py:
3590
3591 2011-06-27  Diego Gonzalez  <diegohcg@webkit.org>
3592
3593         Reviewed by Antonio Gomes.
3594
3595         [Qt] DRT: Provide LayoutTestController::setDefersLoading and ::goBack support
3596         https://bugs.webkit.org/show_bug.cgi?id=63489
3597
3598         Unskipped:
3599             loader/navigation-while-deferring-loads.html
3600
3601         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3602         (LayoutTestController::goBack):
3603         (LayoutTestController::setDefersLoading):
3604         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3605
3606 2011-06-30  Adam Barth  <abarth@webkit.org>
3607
3608         Reviewed by Dirk Pranke.
3609
3610         new-run-webkit-tests should stop when I ^C it
3611         https://bugs.webkit.org/show_bug.cgi?id=63763
3612
3613         It still doesn't stop fast enough, but at least with this patch it
3614         doesn't run more tests.
3615
3616         Also, I tweaked the UI that happens when there's a keyboard interrupt
3617         to not be amazingly ugly.  There's still a garbage "Te" left on the
3618         console, but that's a problem for another day.
3619
3620         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3621
3622 2011-06-30  Adam Roben  <aroben@apple.com>
3623
3624         Make TestFailures never treat a too-many-failures build as a last-passing build for a test
3625
3626         Fixes <http://webkit.org/b/63758> TestFailures sometimes incorrectly says that a build was
3627         the last passing build for a test, when in fact the test wasn't run due to too many failures
3628         causing old-run-webkit-tests to exit early
3629
3630         Reviewed by Anders Carlsson.
3631
3632         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
3633         (LayoutTestHistoryAnalyzer.prototype._incorporateBuildHistory): If there were too many
3634         failures in this build, assume that all the tests failed in the last build would have failed
3635         in this build, too, had they been run. Introduced some new local variables to reduce typing,
3636         and used them throughout.
3637
3638         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
3639         (LayoutTestResultsLoader.prototype.start): Fix a regression from r89841 where we wouldn't
3640         correctly remember whether old-run-webkit-tests exited early due to too many failures for
3641         builds older than r89610. Also removed an unused parameter while I was at it.
3642
3643 2011-06-30  Adam Roben  <aroben@apple.com>
3644
3645         Make TestFaiulres only load old-results directories as needed
3646
3647         Fixes <http://webkit.org/b/63752> Tester pages on TestFailures page load very slowly
3648
3649         Reviewed by Anders Carlsson.
3650
3651         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
3652         (Builder.prototype.getBuildNames): Moved up into the API section of the class. Now just
3653         calls through to _getBuildNamesFromResultsDirectory.
3654         (Builder.prototype.getOldBuildNames): Added. Just calls through to
3655         _getBuildNamesFromResultsDirectory.
3656         (Builder.prototype._getBuildNamesFromResultsDirectory): Renamed from getBuildNames. Now
3657         takes the directory URL as an argument and only fetches that single URL.
3658
3659         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
3660         (LayoutTestHistoryAnalyzer.prototype.start): Moved most logic to _analyzeBuilds. First
3661         analyzes builds from Builder.getBuildNames, then from Builder.getOldBuildNames if needed.
3662         (LayoutTestHistoryAnalyzer.prototype._analyzeBuilds): Moved logic here from start. (Most
3663         changes are just indentation.) Now takes a callback to call when we've finished analyzing
3664         all builds in buildNames so that we can try to fetch more build names if needed.
3665
3666 2011-06-30  Mark Rowe  <mrowe@apple.com>
3667
3668         Reviewed by Anders Carlsson.
3669
3670         Update MiniBrowser and WebKitTestRunner to use the client version constants.
3671
3672         * MiniBrowser/gtk/BrowserWindow.c:
3673         (browserWindowLoaderClientInit):
3674         (browserWindowUIClientInit):
3675         * MiniBrowser/gtk/WebBundle/WebBundleMain.c:
3676         (didCreatePage):
3677         (WKBundleInitialize):
3678         * MiniBrowser/gtk/main.c:
3679         (createWKContextWithInjectedBundle):
3680         * MiniBrowser/mac/AppDelegate.m:
3681         (-[BrowserAppDelegate init]):
3682         * MiniBrowser/mac/BrowserWindowController.m:
3683         (-[BrowserWindowController awakeFromNib]):
3684         * MiniBrowser/mac/WebBundle/WebBundleMain.m:
3685         (WKBundleInitialize):
3686         * MiniBrowser/win/BrowserView.cpp:
3687         (BrowserView::create):
3688         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
3689         (WTR::InjectedBundle::initialize):
3690         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3691         (WTR::InjectedBundlePage::InjectedBundlePage):
3692         * WebKitTestRunner/TestController.cpp:
3693         (WTR::TestController::createOtherPage):
3694         (WTR::TestController::initialize):
3695
3696 2011-06-30  Adam Roben  <aroben@apple.com>
3697
3698         Identify revisions that modified failing tests on TestFailures page
3699
3700         Fixes <http://webkit.org/b/63716> TestFailures page should try to infer which commit caused
3701         a failure by looking at commit logs
3702
3703         Reviewed by David Kilzer.
3704
3705         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
3706         (.existing-bugs-list, .suspect-revisions-list): Made this rule apply to the suspect
3707         revisions list, too.
3708
3709         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Trac.js:
3710         (Trac): Initialize our memory cache.
3711         (Trac.prototype.getCommitDataForRevisionRange): Added. Fetches the commit log in RSS form,
3712         then parses out the revision number, Trac's idea of the commit title, and the files modified
3713         by the commit. (To get the modified files, we rely on the commit message including a
3714         prepare-ChangeLog-style file list.)
3715         (Trac.prototype.logURL): Added a new formatAsRSS parameter.
3716
3717         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
3718         (removePathExtension): Added. Returns a new string with the last dot and everything after it
3719         removed.
3720         (sorted): Added sortFunction parameter.
3721         (Node.prototype.removeAllChildren): Added. Does what it says.
3722
3723         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
3724         (ViewController.prototype._displayBuilder): Updated for change to _domForRegressionRange.
3725         (ViewController.prototype._domForRegressionRange): Changed argument order to match
3726         _domForNewAndExistingBugs. Now takes the list of failing tests and uses it to figure out
3727         which revisions modified the failing tests and displays the suspect revisions in a list.
3728
3729 2011-06-30  Eric Seidel  <eric@webkit.org>
3730
3731         Reviewed by Adam Barth.
3732
3733         Rename should_kill_dump_render_tree to driver_needs_restart in preparation for supporting WebKitTestRunner
3734         https://bugs.webkit.org/show_bug.cgi?id=63743
3735
3736         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
3737         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
3738
3739 2011-06-30  Martin Robinson  <mrobinson@igalia.com>
3740
3741         Reviewed by Anders Carlsson.
3742
3743         [GTK] Crash observed with nspluginwrapper and flash
3744         https://bugs.webkit.org/show_bug.cgi?id=62249
3745
3746         Added a TestNetscapePlugin test which verifies that WebKit properly
3747         handles situations where InvalidateRect is called with a null instance.
3748
3749         * DumpRenderTree/TestNetscapePlugIn/Tests/x11/CallInvalidateRectWithNullNPPArgument.cpp: Added.
3750         (CallInvalidateRectWithNullNPPArgument::CallInvalidateRectWithNullNPPArgument):
3751         (CallInvalidateRectWithNullNPPArgument::NPP_New):
3752         * GNUmakefile.am: Add the new file to sources list.
3753
3754 2011-06-30  Eric Seidel  <eric@webkit.org>
3755
3756         Reviewed by Adam Barth.
3757
3758         Remove redundant tests in TestFailuresTest
3759         https://bugs.webkit.org/show_bug.cgi?id=63742
3760
3761         These tests are all already covered by TestFailuresTest.test_loads.
3762         I also added the missing "T" from TestFailuresest. :)
3763
3764         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
3765
3766 2011-06-30  Adam Barth  <abarth@webkit.org>
3767
3768         Reviewed by Eric Seidel.
3769
3770         REGRESSION(r84294): new-run-webkit-tests results.html generate links to diffs.html or diff.png that don't exist
3771         https://bugs.webkit.org/show_bug.cgi?id=63494
3772
3773         This patch does two things:
3774
3775         1) Refactor diff_image to return the image diff instead of writing the
3776         diff to a file.  This fixes a bunch of hacks and disentangles a bunch
3777         of code.
3778
3779         2) When there's a checksum mismatch but not image diff, we no longer
3780         report an IMAGE failure to results.html.  That fixes the bug in
3781         question because results.html won't try to link to a non-existent image
3782         diff.
3783
3784         * Scripts/webkitpy/common/system/executive_mock.py:
3785         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
3786         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py:
3787         * Scripts/webkitpy/layout_tests/port/base.py:
3788         * Scripts/webkitpy/layout_tests/port/chromium.py:
3789         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
3790         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
3791         * Scripts/webkitpy/layout_tests/port/test.py:
3792         * Scripts/webkitpy/layout_tests/port/webkit.py:
3793         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
3794         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests.py:
3795
3796 2011-06-28  Hans Wennborg  <hans@chromium.org>
3797
3798         Reviewed by Tony Chang.
3799
3800         IndexedDB: Prepare for running layout tests with LevelDB
3801         https://bugs.webkit.org/show_bug.cgi?id=63593
3802
3803         The TestShell must always provide a temporary folder for LevelDB,
3804         not just when the backing store type is overridden.
3805
3806         * DumpRenderTree/chromium/LayoutTestController.cpp:
3807         (LayoutTestController::setOverrideIndexedDBBackingStore):
3808         (LayoutTestController::clearAllDatabases):
3809         * DumpRenderTree/chromium/LayoutTestController.h:
3810         * DumpRenderTree/chromium/TestShell.cpp:
3811         (TestShell::TestShell):
3812         * DumpRenderTree/chromium/TestShell.h:
3813
3814 2011-06-30  Zsolt Fehér  <h490267@stud.u-szeged.hu>
3815
3816         Reviewed by Csaba Osztrogonác.
3817
3818         [WK2] Create a general wk2 specific LayoutTests/platform/wk2 platform
3819         qt-wk2, win-wk2, mac-wk2 use this common Skipped list too
3820         https://bugs.webkit.org/show_bug.cgi?id=61785 
3821
3822         * Scripts/old-run-webkit-tests:
3823
3824 2011-06-30  Adam Roben  <aroben@apple.com>
3825
3826         Use a colon to separate test names from failure info on TestFailures
3827
3828         This matches how the "Possibly Flaky Tests" list works, and is much prettier when the
3829         failure info itself includes parentheses.
3830
3831         Fixes <http://webkit.org/b/63722> Crashing symbols on TestFailures page look ugly due to
3832         double parentheses
3833
3834         Reviewed by Darin Adler.
3835
3836         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
3837         (ViewController.prototype._domForFailedTest): Use a colon instead of surrounding the failure
3838         info in parentheses.
3839
3840 2011-06-30  Adam Roben  <aroben@apple.com>
3841
3842         Robustify PersistentCache against cached data that contains the string ': '
3843
3844         Fixes <http://webkit.org/b/63721> Exceptions thrown while loading TestFailures page due to
3845         passing a bad string to JSON.parse
3846
3847         Reviewed by Darin Adler.
3848
3849         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/PersistentCache.js:
3850         (PersistentCache._parseDateAndJSONFromString): Manually split the string on the separator
3851         using indexOf and substring. (String.prototype.split always splits on all instances of the
3852         separator.)
3853
3854 2011-06-30  Adam Roben  <aroben@apple.com>
3855
3856         Include the crashing symbol in crash logs links on TestFailures
3857
3858         Fixes <http://webkit.org/b/63465> Links to crash logs on TestFailures page should include
3859         the crashing symbol (like links in results.html do)
3860
3861         Reviewed by David Kilzer.
3862
3863         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
3864         (LayoutTestResultsLoader.prototype.start): Bumped the cache version because we now store
3865         crashing symbols for crashing tests. Renamed testsForResultTable to parseResultTable because
3866         it now returns more than just the test names. Specifically, it now looks for crash log links
3867         and extracts the crashing symbol name from them. Updated callers of parseResultTable to
3868         match its new behavior. Changed to store the crashing symbol along with the failure type in
3869         the data we pass to the callback for tests which crashed.
3870
3871         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
3872         (code): Make <code> elements a little smaller because their contents can be quite long.
3873
3874         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
3875         (ViewController.prototype._domForFailureDiagnosis): Include the crashing symbol inside a
3876         <code> element in the link, if there is a crashing symbol.
3877
3878 2011-06-30  Adam Roben  <aroben@apple.com>
3879
3880         Use objects instead of strings to represent a test result in TestFailures code
3881
3882         This will eventually allow us to store more than just the type of failure for each test.
3883         (E.g., we can store the name of the crashing symbol for tests which crashed.)
3884
3885         Prep work for <http://webkit.org/b/63465> Links to crash logs on TestFailures page should
3886         include the crashing symbol (like links in results.html do)
3887
3888         Reviewed by David Kilzer.
3889
3890         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
3891         (Builder.prototype.failureDiagnosisTextAndURL): Changed to expect a testResult object
3892         instead of just a failureType string.
3893
3894         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/FlakyLayoutTestDetector.js:
3895         (FlakyLayoutTestDetector.prototype.incorporateTestResults): Changed to store a
3896         testResult-like object for passing tests.
3897         (FlakyLayoutTestDetector.prototype.flakinessExamples): Changed to expect testResult-like
3898         objects.
3899
3900         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
3901         (LayoutTestResultsLoader.prototype.start): Store a version number along with the cached data
3902         so we can throw away cached data that's in an old format. Store a testResult object for each
3903         test instead of just its failure type.
3904
3905         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
3906         (ViewController.prototype._domForFailedTest):
3907         (ViewController.prototype._domForFailureDiagnosis):
3908         Changed to expect testResult objects instead of failureType strings.
3909
3910 2011-06-30  Adam Roben  <aroben@apple.com>
3911
3912         Show full commit logs when visiting Trac from TestFailures page
3913
3914         Fixes <http://webkit.org/b/63715> Links to Trac on TestFailures page
3915         should show full commit logs
3916
3917         Reviewed by David Kilzer.
3918
3919         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Trac.js:
3920         (Trac.prototype.logURL): Added new 'showFullCommitLogs' parameter. When specified, we add
3921         verbose=on to the log URL to tell Trac to show full commit logs.
3922
3923         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
3924         (ViewController.prototype._domForRegressionRange): Specify true for showFullCommitLogs.
3925
3926 2011-06-30  Adam Roben  <aroben@apple.com>
3927
3928         Add a link to file a bug about TestFailures to the TestFailures page
3929
3930         Fixes <http://webkit.org/b/63466> TestFailures page should include a link to file a bug
3931         about TestFailures itself
3932
3933         Reviewed by Anders Carlsson.
3934
3935         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
3936         (aside): Added some styles for the new link. It is positioned in the upper-right and
3937         hopefully not to glaringly obvious.
3938
3939         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
3940         (ViewController.prototype.loaded): Set up our main content element and auxiliary UI.
3941
3942         (ViewController.prototype._displayBuilder):
3943         (ViewController.prototype._displayTesters):
3944         Modify the main content element instead of the entire body, as there are other things in the
3945         body that we want to preserve.
3946
3947         (ViewController.prototype._domForAuxiliaryUIElements): Create and return an <aside> element
3948         that contains a link to file a bug about the TestFailures page.
3949
3950 2011-06-30  Adam Roben  <aroben@apple.com>
3951
3952         Ignore callbacks that correspond to old navigations of the TestFailures page
3953
3954         Fixes <http://webkit.org/b/63711> TestFailures page sometimes ignores back/forward
3955         navigations
3956
3957         Reviewed by Anders Carlsson.
3958
3959         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
3960         (LayoutTestHistoryAnalyzer.prototype.start): Check the return value of the callback. If the
3961         callback returns false, stop fetching more data.
3962
3963         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
3964         (ViewController): Initialize new _navigationID member. This member is used to assign each
3965         navigation by the user a unique ID.
3966         (ViewController.prototype.parseHash): Increment the navigation ID.
3967
3968         (ViewController.prototype._displayBuilder):
3969         (ViewController.prototype._displayTesters):
3970         At the start of these functions, record the current navigation ID. Bail out from callbacks
3971         if the navigation ID has changed, indicating that the user has navigated elsewhere.
3972
3973 2011-06-30  Carlos Garcia Campos  <cgarcia@igalia.com>
3974
3975         Reviewed by Martin Robinson.
3976
3977         [GTK] Add back/forward menu to MiniBrowser toolbar
3978         https://bugs.webkit.org/show_bug.cgi?id=63445
3979
3980         * MiniBrowser/gtk/BrowserWindow.c:
3981         (browser_window_init): Use a GtkMenuToolButton for navigation
3982         widgets.
3983         (browserWindowConstructed): Get the page BackForwardList.
3984         (browserWindowHistoryItemActivated): Go to activated history item.
3985         (browserWindowHistoryItemSelected): Show the url of currently
3986         selected history item in the statusbar.
3987         (createGtkActionFromBackForwardItem): Create a GtkAction from a
3988         WKBackForwardListItemRef.
3989         (browserWindowCreateMenuItemFromBackForwardItem): Create a
3990         GtkMenuItem from a WKBackForwardListItemRef.
3991         (browserWindowCreateBackForwardMenu): Create a GtkMenu with the
3992         given history items.
3993         (browserWindowUpdateNavigationActions): Update navigation widgets
3994         sensitivity and history menus.
3995         (didChangeBackForwardList): Call
3996         browserWindowUpdateNavigationActions().
3997         (browserWindowLoaderClientInit): Add didChangeBackForwardList
3998         callback.
3999
4000 2011-06-30  Dmitry Lomov  <dslomov@google.com>
4001
4002         Reviewed by Adam Roben.
4003
4004         https://bugs.webkit.org/show_bug.cgi?id=63660
4005         This enables unit-tests step on Windows and MacOS SnowLeopard.
4006
4007         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
4008
4009 2011-06-30  Balazs Kelemen  <kbalazs@webkit.org>
4010
4011         Reviewed by Kenneth Rohde Christiansen.
4012
4013         [Qt][WK2] Implement the rest of WTR::PlatformWebView
4014         https://bugs.webkit.org/show_bug.cgi?id=63630
4015
4016         * WebKitTestRunner/qt/PlatformWebViewQt.cpp:
4017         (WTR::PlatformWebView::windowFrame):
4018         (WTR::PlatformWebView::setWindowFrame):
4019
4020 2011-06-29  Dirk Pranke  <dpranke@chromium.org>
4021
4022         Reviewed by Adam Barth.
4023
4024         test-webkitpy is spamming lots of messages about updating test expectations
4025         https://bugs.webkit.org/show_bug.cgi?id=63680
4026
4027         This patch changes the passing_run() function to capture the
4028         output to stdout and stderr and assert that it is empty, and
4029         also specifies --no-new-test-results by default to avoid getting
4030         "generating baseline" messages for the expected-missing results.
4031
4032         Also check that we are logging messages when we generate new
4033         baselines.
4034
4035         Also fix a minor bug in the way test_tolerance() was
4036         checking for whether the tests were passing or not.
4037
4038         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
4039
4040 2011-06-29  Yuta Kitamura  <yutak@chromium.org>
4041
4042         Reviewed by Kent Tamura.
4043
4044         DRT: Make Hixie76WebSocketProtocolEnabled preference flag configurable from LayoutTestController
4045         https://bugs.webkit.org/show_bug.cgi?id=63532
4046
4047         LayoutTestController::overridePreference() will be used to change the value of
4048         the Hixie76WebSocketProtocolEnabled preference flag from layout tests. In this way,
4049         we can avoid writing a lot of boilerplate code in each port's DumpRenderTree, but
4050         at least we need to reset preference values in DumpRenderTree every time we start
4051         a new test because preference values changed in some test should not be leaked
4052         to the next test.
4053
4054         * DumpRenderTree/chromium/LayoutTestController.cpp:
4055         (LayoutTestController::overridePreference):
4056         The name of preference key is defined in WebKit/{mac/WebView,win}/WebPreferenceKeysPrivate.h.
4057         * DumpRenderTree/chromium/WebPreferences.cpp:
4058         (WebPreferences::reset):
4059         Set the default value. This will be called from TestShell::resetTestController().
4060         (WebPreferences::applyTo):
4061         * DumpRenderTree/chromium/WebPreferences.h:
4062         * DumpRenderTree/mac/DumpRenderTree.mm:
4063         (resetDefaultsToConsistentValues): Set the default value.
4064         * DumpRenderTree/win/DumpRenderTree.cpp:
4065         (resetDefaultsToConsistentValues): Ditto.
4066
4067 2011-06-29  Kenichi Ishibashi  <bashi@chromium.org>
4068
4069         Reviewed by Adam Barth.
4070
4071         check-webkit-style fails to detect https://webkit.org/b/XXXXX as a bug number
4072         https://webkit.org/b/57360
4073
4074         Modify the regular expression to match https scheme.
4075
4076         * Scripts/webkitpy/common/checkout/changelog_unittest.py: Added a test case for https://webkit.org/b/XXXXX.
4077         * Scripts/webkitpy/common/config/urls.py: Modified bug_url_short to match https scheme.
4078
4079 2011-06-29  Adam Barth  <abarth@webkit.org>
4080
4081         Reviewed by Eric Seidel.
4082
4083         new-run-webkit-tests: images page switch actual and expected images before they're loaded
4084         https://bugs.webkit.org/show_bug.cgi?id=63199
4085
4086         We now preload the images and keep the DOM nodes in memory instead of
4087         going back to the MemoryCache all the time.  (Also, remove some nutty
4088         </img> HTML.)
4089
4090         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py:
4091
4092 2011-06-29  Adam Barth  <abarth@webkit.org>
4093
4094         Reviewed by Eric Seidel.
4095
4096         new-run-webkit-tests should tell you when it creates new expectations
4097         https://bugs.webkit.org/show_bug.cgi?id=63675
4098
4099         Unforunately, this patch creates a bunch of test-webkitpy spam because
4100         we don't understand how Python logging works well enough to stop the
4101         spam.  :(
4102
4103         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
4104
4105 2011-06-29  Eric Seidel  <eric@webkit.org>
4106
4107         Reviewed by Adam Barth.
4108
4109         Refactor WebKitDriver.run_test into smaller pieces
4110         https://bugs.webkit.org/show_bug.cgi?id=63673
4111
4112         There is no functional change here.  Just moving code around.
4113         This is in preparation for adding support for #CRASHED and #CRASHED - WebProcess
4114
4115         * Scripts/webkitpy/layout_tests/port/webkit.py:
4116
4117 2011-06-29  Eric Seidel  <eric@webkit.org>
4118
4119         Reviewed by Adam Barth.
4120
4121         Simplify skipped file finding in preparation for adding wk2 skipped list fallback
4122         https://bugs.webkit.org/show_bug.cgi?id=63501
4123
4124         Re-landing after fixing a bug in WebKitPort.baseline_search_path which broke
4125         the rebaselineserver unit tests.  Also fixed nits brought up by Dirk after previous landing.
4126
4127         The goal was to add support for wk2 skipped lists.
4128         However, I found that our skipped list computation was a
4129         manual hack (which only worked for the mac port).
4130
4131         So I fixed a FIXME to move the skipped list finding
4132         into WebKitPort instead of MacPort.
4133         Doing so required the concept of a "port_name", since previously
4134         the only name accessible from a port object was name()
4135         which includes many things beyond "mac" or "chromium", etc.
4136
4137         Eventually I believe we'll want to re-think the way that we pass
4138         in a port_name argument to Port subclasses and expect them to parse
4139         it.  But for now I just added a cls.port_name variable which contains
4140         the static information needed to compute wk2 names as well as
4141         compute Skipped list fallback which works for Mac/Win/Qt and Gtk.
4142
4143         In order to test my new _skipped_file_search_paths method, I
4144         fixed another FIXME by making it return relative paths.
4145
4146         I also fixed the test_expectations_path code in WebKitPort to use port_name.
4147         It was using name() which would return PORT-VERSION so MacPort was overriding
4148         it to use just PORT.  After fixing test_expectations_path to use port_name
4149         (and making it aware of webkit2) I was able to remove the MacPort implementation.
4150
4151         * Scripts/webkitpy/layout_tests/port/base.py:
4152          - Add port_name() to access "mac", since name() returns "mac-leopard" etc.
4153          - Document that real_name() seems to have no purpose.
4154         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
4155          - Add _parse_port_name(), eventually we might call this from WebKitPort directly.
4156         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
4157          - Add _parse_port_name.
4158         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
4159          - Add _parse_port_name.
4160         * Scripts/webkitpy/layout_tests/port/gtk.py:
4161         * Scripts/webkitpy/layout_tests/port/mac.py:
4162          - Move Skipped-file finding code down to WebKitPort
4163         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
4164         * Scripts/webkitpy/layout_tests/port/qt.py:
4165         * Scripts/webkitpy/layout_tests/port/webkit.py:
4166         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
4167
4168 2011-06-29  Sheriff Bot  <webkit.review.bot@gmail.com>
4169
4170         Unreviewed, rolling out r90070.
4171         http://trac.webkit.org/changeset/90070
4172         https://bugs.webkit.org/show_bug.cgi?id=63676
4173
4174         Broke test-webkitpy rebaseline tests (Requested by eseidel2 on
4175         #webkit).
4176
4177         * Scripts/webkitpy/layout_tests/port/base.py:
4178         * Scripts/webkitpy/layout_tests/port/chromium.py:
4179         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
4180         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
4181         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
4182         * Scripts/webkitpy/layout_tests/port/gtk.py:
4183         * Scripts/webkitpy/layout_tests/port/mac.py:
4184         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
4185         * Scripts/webkitpy/layout_tests/port/qt.py:
4186         * Scripts/webkitpy/layout_tests/port/webkit.py:
4187         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
4188         * Scripts/webkitpy/layout_tests/port/win.py:
4189         * Scripts/webkitpy/tool/commands/rebaselineserver_unittest.py:
4190<