2011-03-03 Tony Chang <tony@chromium.org>
[WebKit-https.git] / Tools / ChangeLog
1 2011-03-03  Tony Chang  <tony@chromium.org>
2
3         Reviewed by Dimitri Glazkov.
4
5         [chromium] Run kill-old-processes on the chromium-win bot too
6         https://bugs.webkit.org/show_bug.cgi?id=55630
7
8         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
9         * BuildSlaveSupport/win/kill-old-processes: Add wdiff.exe to the kill list.
10
11 2011-03-03  Carlos Garcia Campos  <cgarcia@igalia.com>
12
13         Reviewed by Martin Robinson.
14
15         [GTK] DRT needs implementation of EventSender.scheduleAsynchronousClick
16         https://bugs.webkit.org/show_bug.cgi?id=53960
17
18         * DumpRenderTree/gtk/EventSender.cpp:
19         (sendClick):
20         (scheduleAsynchronousClickCallback):
21
22 2011-03-03  Adam Roben  <aroben@apple.com>
23
24         Look for Windows-style paths when checking whether a test is an Inspector test
25
26         WTR part of <http://webkit.org/b/55672> <rdar://problem/9080867> All inspector tests time
27         out or crash on Windows 7 Release (WebKit2 Tests)
28
29         Reviewed by Darin Adler.
30
31         * WebKitTestRunner/TestInvocation.cpp:
32         (WTR::shouldOpenWebInspector): Also look for backslashes, since that's what we'll get on
33         Windows.
34
35 2011-03-03  Adam Roben  <aroben@apple.com>
36
37         Make Windows 7 Release (Tests) a core builder
38
39         We can probably make Windows XP Debug (Tests) a core builder, too, but we should probably
40         give it a few more days of greenness first.
41
42         Fixes <http://webkit.org/b/55665> Windows 7 Release (Tests) should be a core builder
43
44         Reviewed by Andreas Kling.
45
46         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
47         (BuildBot): Added a regexp to match Windows 7 Release (Tests).
48
49         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
50         (BuildBotTest): Updated the list of builders to match the current list on build.webkit.org.
51         Updated the core regexps to match the buildbot module. Updated the expected results to match
52         the current set of core builders.
53
54 2011-03-03  Benjamin Poulain  <benjamin.poulain@nokia.com>
55
56         Reviewed by Andreas Kling.
57
58         Adding myself as a reviewer.
59
60         * Scripts/webkitpy/common/config/committers.py:
61
62 2011-03-03  Lukasz Slachciak  <l.slachciak@samsung.com>
63
64         Reviewed by Xan Lopez.
65
66         [GTK] Extended application cache database API and added unit tests file.
67         https://bugs.webkit.org/show_bug.cgi?id=55335
68
69         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: removed external declaration
70         of webkit_application_cache_set_maximum_size due to API exposal.
71
72 2011-03-02  Ojan Vafai  <ojan@chromium.org>
73
74         Reviewed by Adam Barth.
75
76         make webkit-patch upload work for security bugs
77         https://bugs.webkit.org/show_bug.cgi?id=55655
78
79         The problem is that we shell out to prepare-ChangeLog,
80         which does not authenticate with bugzilla. All it needs from
81         bugzilla is the bug title though. We just pass it through from
82         webkit-patch instead.
83
84         * Scripts/prepare-ChangeLog:
85         * Scripts/webkitpy/common/net/bugzilla/bug.py:
86         * Scripts/webkitpy/tool/steps/preparechangelog.py:
87
88 2011-03-01  Ojan Vafai  <ojan@chromium.org>
89
90         Reviewed by Tony Chang.
91
92         [test-dashboard] make gtest JSON match layout-tests JSON
93         https://bugs.webkit.org/show_bug.cgi?id=55556
94
95         http://trac.webkit.org/changeset/80090 stopped converting time_ms to t.
96         Make this code match.
97
98         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
99
100 2011-03-02  Dirk Pranke  <dpranke@chromium.org>
101
102         Reviewed by Tony Chang.
103
104         NRWT - implement Linux Hardy 64-bit port support. This involves
105         adding support for non-'x86' architectures and updating the
106         "deduplicate_tests" script to correctly understand that
107         port.name() is not always equal to basename(port.baseline_path()).
108
109         This change introduces two new port names: 'chromium-linux-x86'
110         and 'chromium-linux-x86_64'. Using the prior 'chromium-linux'
111         flag will cause the code to determine which one to use at
112         runtime, just like we currently do with 'chromium-win'.
113
114         https://bugs.webkit.org/show_bug.cgi?id=55535
115
116         * Scripts/webkitpy/common/system/executive_mock.py:
117           Handle 'return_stderr' keyword arg to run_command()
118         * Scripts/webkitpy/layout_tests/deduplicate_tests.py:
119         * Scripts/webkitpy/layout_tests/deduplicate_tests_unittest.py:
120         * Scripts/webkitpy/layout_tests/port/base.py:
121         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
122         * Scripts/webkitpy/layout_tests/port/chromium_linux_unittest.py:
123         * Scripts/webkitpy/layout_tests/port/factory.py:
124         * Scripts/webkitpy/layout_tests/port/google_chrome.py:
125
126 2011-03-02  Kevin Ollivier  <kevino@theolliviers.com>
127
128         [wx] Build fix, make sure we put the wxWebKit dylib in the right directory and
129         update symlinks accordingly.
130
131         * wx/packaging/build-mac-installer.py:
132
133 2011-03-02  Sheriff Bot  <webkit.review.bot@gmail.com>
134
135         Unreviewed, rolling out r80139.
136         http://trac.webkit.org/changeset/80139
137         https://bugs.webkit.org/show_bug.cgi?id=55620
138
139         failing and crashing tests on Snow Leopard bot (Requested by
140         mihaip1 on #webkit).
141
142         * DumpRenderTree/LayoutTestController.cpp:
143         (LayoutTestController::staticFunctions):
144         * DumpRenderTree/LayoutTestController.h:
145         * DumpRenderTree/chromium/LayoutTestController.cpp:
146         (LayoutTestController::LayoutTestController):
147         * DumpRenderTree/chromium/LayoutTestController.h:
148         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
149         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
150         * DumpRenderTree/mac/UIDelegate.h:
151         * DumpRenderTree/mac/UIDelegate.mm:
152         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
153         * DumpRenderTree/qt/LayoutTestControllerQt.h:
154         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
155         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
156
157 2011-03-02  Adam Roben  <aroben@apple.com>
158
159         Wait for force a paint in WebKitTestRunner until we're actually ready to dump the output
160
161         This matches DumpRenderTree on Windows.
162
163         Fixes <http://webkit.org/b/55469> <rdar://problem/9068539> REGRESSION (r79863): Lots of
164         dom/html/level2/html/HTMLFrameElement*.html tests crashing in FrameView::paintOverhangAreas
165         in WebKit2
166
167         Reviewed by Beth Dakin.
168
169         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
170         (WTR::InjectedBundlePage::dump): Moved code to force a paint here...
171         (WTR::InjectedBundlePage::didFinishLoadForFrame): ...from here.
172
173 2011-03-02  Daniel Cheng  <dcheng@chromium.org>
174
175         Reviewed by David Levin.
176
177         Add feature define for data transfer items
178         https://bugs.webkit.org/show_bug.cgi?id=55510
179
180         * Scripts/build-webkit:
181
182 2011-03-02  Adam Roben  <aroben@apple.com>
183
184         Look in the new-to-SnowLeopard location for crash logs on SnowLeopard and newer
185
186         ReportCrash saves logs to ~/Library/Logs/CrashReporter on Leopard, but to
187         ~/Library/Logs/DiagnosticReports on SnowLeopard. old-run-webkit-tests was only looking in
188         the former location, but was getting lucky on some SnowLeopard machines because symlinks
189         were being created there pointing into DiagnosticReports. For machines without these
190         symlinks, crash logs were not getting captured.
191
192         Fixes <http://webkit.org/b/55607> old-run-webkit-tests doesn't capture crash logs on some
193         SnowLeopard machines
194
195         Reviewed by Eric Seidel.
196
197         * Scripts/old-run-webkit-tests:
198         (captureSavedCrashLog): Look in ~/Library/Logs/DiagnosticReports on SnowLeopard and newer.
199
200 2011-03-02  Adam Roben  <aroben@apple.com>
201
202         Don't try to capture crash logs on platforms that haven't implemented that feature
203
204         Fixes <http://webkit.org/b/55504> Use of uninitialized value in numeric lt (<) at
205         Tools/Scripts/old-run-webkit-tests line 1778 seen on Qt bots after a crash
206
207         Reviewed by Joseph Pecoraro.
208
209         * Scripts/old-run-webkit-tests:
210         (captureSavedCrashLog): Bail if we don't have a crash log glob.
211         (findNewestFileMatchingGlob): Fix a buggy test for when no paths match the glob. This isn't
212         strictly required to fix bug 55504, but seems worthwhile.
213
214 2011-01-26  John Knottenbelt  <jknotten@chromium.org>
215
216         Reviewed by Dmitry Titov.
217
218         Detach Geolocation from Frame when Page destroyed.
219         https://bugs.webkit.org/show_bug.cgi?id=52877
220
221         Extend the layout test controller to expose the number of pending
222         geolocation requests, so that we can test that the requests have
223         been cancelled on page close.
224
225         * DumpRenderTree/LayoutTestController.cpp:
226         (numberOfPendingGeolocationPermissionRequestsCallback):
227         (LayoutTestController::staticFunctions):
228         * DumpRenderTree/LayoutTestController.h:
229         * DumpRenderTree/chromium/LayoutTestController.cpp:
230         (LayoutTestController::LayoutTestController):
231         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
232         * DumpRenderTree/chromium/LayoutTestController.h:
233         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
234         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
235         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
236         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
237         * DumpRenderTree/mac/UIDelegate.h:
238         * DumpRenderTree/mac/UIDelegate.mm:
239         (-[UIDelegate numberOfPendingGeolocationPermissionRequests]):
240         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
241         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
242         * DumpRenderTree/qt/LayoutTestControllerQt.h:
243         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
244         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
245         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
246         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
247
248
249 2011-03-01  Kent Tamura  <tkent@chromium.org>
250
251         Reviewed by Dimitri Glazkov.
252
253         Assertion fails when a form validation bubble appears
254         https://bugs.webkit.org/show_bug.cgi?id=55550
255
256         Test: fast/forms/interactive-validation-attach-assertion.html
257
258         * DumpRenderTree/chromium/WebPreferences.cpp:
259         (WebPreferences::applyTo): Enable the interactive validation feature by
260           default. This matches to Chromium browser.
261
262 2011-03-01  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
263
264         Reviewed by Kent Tamura.
265
266         Remove efl port style exception of Webcore/JavaScriptCore
267         https://bugs.webkit.org/show_bug.cgi?id=55549
268
269         WebCore and JavaScriptCore need to adhere WebKit coding style.
270
271         * Scripts/webkitpy/style/checker.py:
272
273 2011-03-01  Tony Chang  <tony@chromium.org>
274
275         Reviewed by Ojan Vafai.
276
277         [chromium] clean up temp files on chromium-mac builder too
278         https://bugs.webkit.org/show_bug.cgi?id=55525
279
280         * BuildSlaveSupport/chromium/remove-crash-logs:
281
282 2011-02-28  Ojan Vafai  <ojan@chromium.org>
283
284         Reviewed by Tony Chang.
285
286         wrap json in a function call to afford cross-domain loading
287         https://bugs.webkit.org/show_bug.cgi?id=55353
288
289         Also delete the code that compacts the JSON. It turns out this
290         breaks the rebaseline tool and is probably a premature optimization anyways.
291
292         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
293         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
294         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
295         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
296
297 2011-03-01  Sheriff Bot  <webkit.review.bot@gmail.com>
298
299         Unreviewed, rolling out r80079.
300         http://trac.webkit.org/changeset/80079
301         https://bugs.webkit.org/show_bug.cgi?id=55547
302
303         "Broke the Win debug build?" (Requested by dcheng on #webkit).
304
305         * Scripts/build-webkit:
306
307 2011-03-01  Mark Rowe  <mrowe@apple.com>
308
309         Reviewed by Sam Weinig.
310
311         Update verification scripts to handle WebKit2.
312
313         * Scripts/check-for-global-initializers: Skip object files that use RefCountedLeakChecker.
314         * Scripts/check-for-inappropriate-files-in-framework: Add some inappropriate file types for
315         WebKit2.
316
317 2011-03-01  Daniel Cheng  <dcheng@chromium.org>
318
319         Reviewed by David Levin.
320
321         Add feature define for data transfer items
322         https://bugs.webkit.org/show_bug.cgi?id=55510
323
324         * Scripts/build-webkit:
325
326 2011-02-23  Joseph Pecoraro  <joepeck@webkit.org>
327
328         Reviewed by Timothy Hatcher.
329
330         All Console Messages should be passed to ChromeClients.
331         https://bugs.webkit.org/show_bug.cgi?id=54926
332
333         Update DumpRenderTree's UIDelegate to use the new console message delegate.
334
335         * DumpRenderTree/mac/UIDelegate.mm:
336         (-[UIDelegate webView:addMessageToConsole:withSource:]):
337
338 2011-03-01  Dirk Pranke  <dpranke@chromium.org>
339
340         Reviewed by Adam Barth.
341
342         Tragically, we stopped running the python unit tests
343         on Chromium Win just as I was getting them to be green.
344         This patch turns them back on.
345
346         https://bugs.webkit.org/show_bug.cgi?id=55521
347
348         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
349
350 2011-03-01  Dirk Pranke  <dpranke@chromium.org>
351
352         Reviewed by Adam Roben.
353
354         Disable
355         webkitpy.common.system.executive_unittest.ExecutiveTest.\
356            test_kill_all fails on win until it can be properly triaged.
357
358         https://bugs.webkit.org/show_bug.cgi?id=54790
359
360         * Scripts/webkitpy/common/system/executive_unittest.py:
361
362     2011-03-01  Helder Correia  <helder@sencha.com>
363
364             Unreviewed. Add myself to committers.py.
365
366             * Scripts/webkitpy/common/config/committers.py:
367
368 2011-03-01  Adam Roben  <aroben@apple.com>
369
370         Give up if a crash log for the web process is taking too long to be saved on Windows
371
372         This should keep the bots from getting stuck while waiting for a crash log, if writing a
373         crash log ever hangs or the UI process for some reason doesn't notice it's finished.
374
375         Fixes <http://webkit.org/b/55499> WebKitTestRunner can hang forever waiting for a crash log
376         to be saved for the web process
377
378         Reviewed by Steve Falkenburg.
379
380         * WebKitTestRunner/win/TestControllerWin.cpp:
381         (WTR::runRunLoopUntil): Moved code to run the run loop here here from platformRunUntil.
382         Generalized the code slightly to handle the optional object to wait on.
383         (WTR::TestController::platformRunUntil): Use the new runRunLoopUntil function, and also use
384         it (with a timeout) when waiting for a crash log for the web process to be saved. This will
385         prevent us from waiting forever.
386
387 2011-03-01  Adam Roben  <aroben@apple.com>
388
389         Give WebKitLauncherWin's PDB file a name different from WebKit's
390
391         WebKitLauncherWin (which builds WebKit.exe) and WebKit (which builds WebKit.dll) were
392         fighting over the same PDB file, meaning that one of them was going to lose and not have
393         symbols. WebKitLauncherWin's PDB file is now named WebKitLauncherWin.pdb.
394
395         Fixes <http://webkit.org/b/55495> Crashes in WebKit.dll have bad backtraces for some test
396         runs (due to missing symbols)
397
398         Reviewed by David Kilzer.
399
400         * WebKitLauncherWin/WebKitLauncherWinCommon.vsprops: Use the project name,
401         "WebKitLauncherWin", instead of the target name, "WebKit", when naming the PDB file.
402
403 2011-03-01  Adam Roben  <aroben@apple.com>
404
405         Count web process crashes toward the --exit-after-n-crashes-or-timeouts limit
406
407         Fixes <http://webkit.org/b/55492> Web process crashes aren't counted toward the
408         --exit-after-n-crashes-or-timeouts limit
409
410         Reviewed by David Kilzer.
411
412         * Scripts/old-run-webkit-tests:
413         (stopRunningTestsEarlyIfNeeded): Include web process crashes in the computation, and print
414         the number of tests that caused the web process to crash when exiting early.
415
416 2011-03-01  Adam Roben  <aroben@apple.com>
417
418         Save a crash log when the web process crashes
419
420         On Windows, WebKitTestRunner now detects when the web process is crashing and waits to exit
421         until it has finished crashing, which guarantees that the crash log will have had time to be
422         saved, too. On Mac, we always wait until ReportCrash has exited before capturing the crash
423         log, so all we have to do is choose the right crash log out of the CrashReporter directory.
424
425         Fixes <http://webkit.org/b/44121> <rdar://problem/8320759> When the web process crashes and
426         a crash log is being saved, WebKitTestRunner thinks the web process has become unresponsive
427
428         Reviewed by Sam Weinig.
429
430         * Scripts/old-run-webkit-tests:
431         (testCrashedOrTimedOut): Don't kill WebKitTestRunner when the web process crashes. It will
432         kill itself. On Windows, this will cause us to wait until the crash log has been saved. On
433         Mac, it should have no effect. Capture saved crash logs for web process crashes, too.
434         (captureSavedCrashLog): Added $webProcessCrashed argument. On Mac, look for
435         WebProces_*.crash files when the web process crashes.
436
437         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
438         (WTR::InjectedBundle::initialize): Added an initializationUserData argument. Updated for
439         initializePlatformDefaults -> platformInitialize rename. Pass the initializationUserData
440         along to platformInitialize.
441
442         * WebKitTestRunner/InjectedBundle/InjectedBundle.h: See above.
443
444         * WebKitTestRunner/InjectedBundle/InjectedBundleMain.cpp:
445         (WKBundleInitialize): Pass along the initializationUserData to the InjectedBundle.
446
447         * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm:
448         (WTR::InjectedBundle::platformInitialize):
449         * WebKitTestRunner/InjectedBundle/qt/InjectedBundleQt.cpp:
450         (WTR::InjectedBundle::platformInitialize):
451         Updated function signature.
452
453         * WebKitTestRunner/InjectedBundle/win/InjectedBundleWin.cpp:
454         (WTR::exceptionFilter): Added. Tells the UI process we're crashing by signaling the
455         webProcessCrashingEvent, then lets the crash continue as normal.
456
457         (WTR::InjectedBundle::platformInitialize): Hook up exceptionFilter. Retrieve the name of the
458         event we should use to tell the UI process we're crashing from the initializationUserData,
459         and get a handle to that event.
460
461         * WebKitTestRunner/TestController.cpp:
462         (WTR::TestController::TestController): Initialize new members.
463         (WTR::TestController::processDidCrash): Removed unnecessary WKPageRef argument. Changed to
464         only print the "#CRASHED - WebProcess" message once, since this can be called more than once
465         when a crash log is being saved on Windows. Exit right away if specified. (This is the
466         default.)
467
468         * WebKitTestRunner/TestController.h: Added new members.
469
470         * WebKitTestRunner/win/TestControllerWin.cpp:
471         (WTR::TestController::platformInitialize): Set up the event the web process will use to tell
472         us it's crashing.
473         (WTR::TestController::platformRunUntil): Pass MWMO_INPUTAVAILABLE to
474         ::MsgWaitForMultipleObjectsEx so we'll process messages that have already been seen by
475         ::PeekMessage. (This is unrelated to the bug fix.) Notice when the webProcessCrashingEvent
476         has been signaled. When this happens, print the "#CRASHED - WebProcess" message right away
477         so the test harness will know the web process has crashed and not try to kill us, then wait
478         for the web process to finish crashing so a crash log will have time to be saved.
479         (WTR::toWK): Simple hepler function.
480         (WTR::TestController::platformInitializeContext): Pass along the name of the event the web
481         process should use to tell us it is crashing in the context's initialization user data.
482
483 2011-03-01  Dimitri Glazkov  <dglazkov@chromium.org>
484
485         Reviewed by Tony Gentilcore.
486
487         Move the checks to exclude Python/Perl tests for Chromium Win build to the right place.
488         https://bugs.webkit.org/show_bug.cgi?id=55476
489
490         Chromium Win uses BuildAndTestFactory, not TestFactory.
491
492         * BuildSlaveSupport/build.webkit.org-config/master.cfg: Moved to the right place.
493
494 2011-03-01  Adam Roben  <aroben@apple.com>
495
496         Test that the WebKit2 responsiveness timer doesn't fire too early
497
498         Test for <http://webkit.org/b/55417> <rdar://problem/9065287> RunLoop::Timer fires 1000x too
499         early on Windows
500
501         Reviewed by Anders Carlsson.
502
503         * TestWebKitAPI/PlatformUtilities.h: Added sleep().
504
505         * TestWebKitAPI/Tests/WebKit2/ResponsivenessTimerDoesntFireEarly.cpp: Added.
506         (TestWebKitAPI::didReceiveMessageFromInjectedBundle):
507         (TestWebKitAPI::didFinishLoadForFrame):
508         (TestWebKitAPI::processDidBecomeUnresponsive):
509         (TestWebKitAPI::setInjectedBundleClient):
510         (TestWebKitAPI::setPageLoaderClient):
511         Simple helper functions.
512
513         (TestWebKitAPI::TEST): Load an HTML file to make sure the web process is initialized. Then
514         tell the web process to pause and press the spacebar key. The spacebar keypress should cause
515         the responsiveness timer to start, but the web process should unpause before it has a chance
516         to fire. Run until the web process has unpaused, and assert that the timer didn't fire.
517
518         * TestWebKitAPI/Tests/WebKit2/ResponsivenessTimerDoesntFireEarly_Bundle.cpp: Added.
519         (TestWebKitAPI::ResponsivenessTimerDoesntFireEarlyTest::ResponsivenessTimerDoesntFireEarlyTest):
520         Call up to the base class.
521         (TestWebKitAPI::ResponsivenessTimerDoesntFireEarlyTest::didReceiveMessage): When asked to
522         pause, sleep for 0.5 seconds, then send back a message saying we paused.
523
524         * TestWebKitAPI/mac/PlatformUtilitiesMac.mm:
525         (TestWebKitAPI::Util::sleep): Added. Calls through to usleep.
526         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp:
527         (TestWebKitAPI::Util::sleep): Added. Calls through to ::Sleep.
528
529         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
530         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
531         * TestWebKitAPI/win/TestWebKitAPIInjectedBundle.vcproj:
532         Added new files.
533
534 2011-03-01  Andras Becsi  <abecsi@webkit.org>
535
536         Reviewed by Csaba Osztrogonác.
537
538         [Qt] Clean up the project files and move common options to WebKit.pri.
539
540         * WebKitTestRunner/qt/WebKitTestRunner.pro: Deduplicate.
541
542
543 2011-03-01  Adam Roben  <aroben@apple.com>
544
545         Skip another multiprocessing test on Windows
546
547         It is sometimes failing (and may be leaving child Python processes around).
548         <http://webkit.org/b/55087> tracks the failure.
549
550         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
551         (FuncitonTests.test_get__processes): Skip this test on Windows.
552
553 2011-03-01  Ilya Tikhonovsky  <loislo@chromium.org>
554
555         Reviewed by Pavel Feldman.
556
557         Web Inspector: reduce inspector tests flakiness.
558         https://bugs.webkit.org/show_bug.cgi?id=55401
559
560         * DumpRenderTree/chromium/TestShell.cpp:
561         (TestShell::runFileTest):
562         * DumpRenderTree/chromium/WebPreferences.cpp:
563         (WebPreferences::reset):
564
565 2011-02-28  Mihai Parparita  <mihaip@chromium.org>
566
567         Reviewed by James Robinson.
568
569         NRWT should still upload results when it exits early due to too many failures
570         https://bugs.webkit.org/show_bug.cgi?id=55432
571         
572         Only keyboard interrupts should prevent results from being uploaded.
573
574         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
575         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
576
577 2011-02-28  Alice Liu  <alice.liu@apple.com>
578
579         Reviewed by Adam Roben and Gavin Barraclough.
580
581         Test for https://bugs.webkit.org/show_bug.cgi?id=54898
582         InjectedBundleNodeHandle dies too early in WKBundleHitTestResultGetNodeHandle.
583
584         * TestWebKitAPI/InjectedBundleController.h:
585         (TestWebKitAPI::InjectedBundleController::bundle):
586         * TestWebKitAPI/PlatformWebView.h:
587         * TestWebKitAPI/Tests/WebKit2/HitTestResultNodeHandle.cpp: Added.
588         (TestWebKitAPI::didFinishLoadForFrame):
589         (TestWebKitAPI::didReceiveMessageFromInjectedBundle):
590         (TestWebKitAPI::setPageLoaderClient):
591         (TestWebKitAPI::TEST):
592         * TestWebKitAPI/Tests/WebKit2/HitTestResultNodeHandle_Bundle.cpp: Added.
593         (TestWebKitAPI::HitTestResultNodeHandleTest::HitTestResultNodeHandleTest):
594         (TestWebKitAPI::HitTestResultNodeHandleTest::getContextMenuFromDefaultMenu):
595         (TestWebKitAPI::HitTestResultNodeHandleTest::didCreatePage):
596         * TestWebKitAPI/mac/PlatformWebViewMac.mm:
597         (TestWebKitAPI::PlatformWebView::simulateRightClick):
598         * TestWebKitAPI/win/PlatformWebViewWin.cpp:
599         (TestWebKitAPI::PlatformWebView::simulateRightClick):
600         
601         Added files to:
602         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
603         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
604         * TestWebKitAPI/win/TestWebKitAPIInjectedBundle.vcproj:
605
606 2011-02-28  Adam Roben  <aroben@apple.com>
607
608         Start running InjectedBundle API tests on Windows
609
610         A new project, TestWebKitAPIInjectedBundle, builds the injected bundle DLL.
611
612         Fixes <http://webkit.org/b/55420> run-api-tests doesn't run injected bundle tests on Windows
613
614         Reviewed by Darin Adler.
615
616         * TestWebKitAPI/Configurations/TestWebKitAPIDebug.vsprops: Added.
617         * TestWebKitAPI/Configurations/TestWebKitAPIDebugAll.vsprops: Added.
618         * TestWebKitAPI/Configurations/TestWebKitAPIDebugCairoCFLite.vsprops: Added.
619         * TestWebKitAPI/Configurations/TestWebKitAPIRelease.vsprops: Added.
620         * TestWebKitAPI/Configurations/TestWebKitAPIReleaseCairoCFLite.vsprops: Added.
621         These are just extracted from TestWebKitAPI.vcproj.
622
623         * TestWebKitAPI/Configurations/TestWebKitAPIInjectedBundleCommon.vsprops: Added. Just sets
624         up the DLL's filename.
625
626         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp: Fixed DLL name.
627         (TestWebKitAPI::Util::createInjectedBundlePath): Changed to use kCFAllocatorNull instead of
628         0 (== kCFAllocatorDefault) so that CF won't try to deallocate our C string.
629
630         * TestWebKitAPI/win/TestWebKitAPI.sln: Added TestWebKitAPIInjectedBundle. Removed the bogus
631         "all" configuration.
632
633         * TestWebKitAPI/win/TestWebKitAPI.vcproj: Changed to use the new .vsprops files. Added
634         injected bundle tests. Let VS resort things.
635
636         * TestWebKitAPI/win/TestWebKitAPIInjectedBundle.vcproj: Copied from Tools/TestWebKitAPI/win/TestWebKitAPI.vcproj.
637
638 2011-02-28  Adam Roben  <aroben@apple.com>
639
640         Add a test for the paint rect passed via WM_PAINT to windowless plugins
641
642         Test for <http://webkit.org/b/55365> <rdar://problem/9031089> REGRESSION (r79040): WebKit2:
643         Large portions of pages with plugins paint black
644
645         Reviewed by Anders Carlsson.
646
647         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
648         (PluginTest::NPP_HandleEvent):
649         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
650         Added NPP_HandleEvent. Just returns 0 at this level.
651
652         * DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowlessPaintRectCoordinates.cpp: Added.
653         (WindowlessPaintRectCoordinates::WindowlessPaintRectCoordinates): Call up to the base class.
654         (WindowlessPaintRectCoordinates::NPP_New): Mark ourselves as windowless.
655         (WindowlessPaintRectCoordinates::NPP_HandleEvent): Check that the paint rect passed via
656         WM_PAINT has the expected coordinates.
657
658         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
659         (NPP_HandleEvent): Give the PluginTest a chance to handle the event before doing anything
660         else.
661
662         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Added new file, let VS
663         resort things.
664
665         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
666         (WTR::InjectedBundlePage::didFinishLoadForFrame): Added a call to WKBundlePageForceRepaint.
667         This is roughly equivalent to the call to -[WebView displayIfNeeded] in -[FrameLoadDelegate
668         webView:didFinishLoadForFrame:] in DumpRenderTree.
669
670         * WebKitTestRunner/win/PlatformWebViewWin.cpp:
671         (WTR::PlatformWebView::PlatformWebView): Tell the WKView it's in a window so that plugins
672         can start running.
673
674 2011-02-28  Adam Roben  <aroben@apple.com>
675
676         Tell MiniBrowser's WKViews that they're in a window
677
678         Fixes <http://webkit.org/b/55364> Plugins don't work in MiniBrowser on Windows
679
680         Reviewed by Anders Carlsson.
681
682         * MiniBrowser/win/BrowserView.cpp:
683         (BrowserView::create): Call WKViewSetIsInWindow after creating the view.
684
685 2011-02-28  Alexis Menard  <alexis.menard@openbossa.org>
686
687         Reviewed by Csaba Osztrogonác.
688
689         The new QML tests require QML_IMPORT_PATH variable to be set in order to
690         load the proper plugin.
691
692         * Scripts/run-qtwebkit-tests:
693
694 2011-02-28  Joone Hur  <joone.hur@collabora.co.uk>
695
696         Unreviewed, fix spelling mistake.
697
698         hardware-accelated => hardware-accelerated.
699
700         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
701
702 2011-02-28  Kristian Amlie  <kristian.amlie@nokia.com>
703
704         Reviewed by Andreas Kling.
705
706         Made sure that the build-webkit --qmake argument is always respected.
707
708         Previously, it would be respected during the build, but not during
709         the early feature detection. The build would also fail if qmake was
710         not in your path, even if you specified --qmake with a valid qmake
711         binary.
712
713         [Qt] WebKit patches required to work with a modularized version of Qt
714         https://bugs.webkit.org/show_bug.cgi?id=53916
715
716         * Scripts/build-webkit:
717         * Scripts/webkitdirs.pm:
718
719 2011-02-27  Ojan Vafai  <ojan@chromium.org>
720
721         Fix failing tests from http://trac.webkit.org/changeset/79837.
722
723         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
724         This was just a mistake that would break if someone tried printing with this configuration.
725         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
726         Method signature changed. Didn't notice that this method was being calling in a unittest.
727
728 2011-02-24  Hayato Ito  <hayato@chromium.org>
729
730         Reviewed by Eric Seidel.
731
732         [NRWT] Remove test_types/* classes.
733
734         This is the last patch in the series of efforts which tried to remove test_types classes.
735         A single_test_runner module will take over responsibilities of test_type classes.
736
737         From this patch, we won't display statistics per test_type even if a '--verbose' option is given.
738
739         https://bugs.webkit.org/show_bug.cgi?id=55123
740
741         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
742         * Scripts/webkitpy/layout_tests/layout_package/test_results.py:
743         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:
744         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
745         * Scripts/webkitpy/layout_tests/layout_package/worker_mixin.py:
746         * Scripts/webkitpy/layout_tests/test_types/__init__.py: Removed.
747         * Scripts/webkitpy/layout_tests/test_types/image_diff.py: Removed.
748         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py: Removed.
749         * Scripts/webkitpy/layout_tests/test_types/test_type_base_unittest.py: Removed.
750         * Scripts/webkitpy/layout_tests/test_types/text_diff.py: Removed.
751
752 2011-02-25  Ojan Vafai  <ojan@chromium.org>
753
754         Reviewed by Tony Chang.
755
756         Change results.json format to the one used by unexpected_results.json
757         https://bugs.webkit.org/show_bug.cgi?id=52267
758
759         Also add runtimes in milliseconds to the JSON and make the output format more compact.
760         Named the file full_results.json to avoid conflicting with the results.json
761         file the test-results server currently serves up.
762
763         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
764         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
765         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
766         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
767         * Scripts/webkitpy/layout_tests/layout_package/result_summary.py:
768         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
769
770 2011-02-27  Adam Roben  <aroben@apple.com>
771
772         Followup fix to r79827
773
774         * DumpRenderTree/win/UIDelegate.cpp:
775         (UIDelegate::webViewAddMessageToConsole): Replaced call to urlSuitableForTestResult with
776         lastPathComponent. This matches what we did here before r79827, and matches Mac.
777
778 2011-02-27  Dan Bernstein  <mitz@apple.com>
779
780         Reviewed by Adam Roben.
781
782         Make bisect-builds print a trac link for the regression range
783         https://bugs.webkit.org/show_bug.cgi?id=55331
784
785         * Scripts/bisect-builds:
786
787 2011-02-27  Adam Roben  <aroben@apple.com>
788
789         Use iswalpha instead of isalpha when dealing with wchar_ts in EditingDelegate
790
791         Fixes <http://webkit.org/b/55062> <rdar://problem/9059907> Crash beneath
792         EditingDelegate::checkSpellingOfString when running fast/forms/input-text-maxlength.html or
793         fast/forms/input-text-paste-maxlength.html on Windows with full page heap enabled
794
795         Reviewed by Anders Carlsson.
796
797         * DumpRenderTree/win/EditingDelegate.cpp:
798         (indexOfFirstWordCharacter):
799         (wordLength):
800         Use iswalpha instead of isalpha, since TCHAR is really just wchar_t.
801
802 2011-02-27  Adam Roben  <aroben@apple.com>
803
804         Make ResourceLoadDelegate print URLs relative to the main resource on Windows
805
806         This matches Mac. (In fact, the code was ported from Mac!)
807
808         Fixes <http://webkit.org/b/55328> URLs printed by ResourceLoadDelegate on Windows don't
809         match Mac
810
811         Reviewed by Anders Carlsson.
812
813         * DumpRenderTree/win/DumpRenderTree.cpp:
814         (substringFromIndex): Added. Emulates -[NSString substringFromIndex:].
815         (urlSuitableForTestResult): Ported code from -[NSURL(DRTExtras)
816         _drt_descriptionSuitableForTestResult].
817         (cfStringRefToWString): Moved here from LayoutTestControllerWin.cpp.
818
819         * DumpRenderTree/win/DumpRenderTreeWin.h: Added declaration of cfStringRefToWString.
820
821         * DumpRenderTree/win/LayoutTestControllerWin.cpp: Moved cfStringRefToWString to
822         DumpRenderTree.cpp.
823
824 2011-02-27  Adam Roben  <aroben@apple.com>
825
826         Change the order of Chang Shu's email addresses so the first one is their Bugzilla username
827
828         This makes Bugzilla autocompletion work correctly.
829
830         * Scripts/webkitpy/common/config/committers.py:
831
832 2011-02-27  Patrick Gansterer  <paroga@webkit.org>
833
834         Reviewed by Andreas Kling.
835
836         [Qt] fast/loader/user-stylesheet-fast-path.html fails
837         https://bugs.webkit.org/show_bug.cgi?id=50911
838
839         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
840         (LayoutTestController::setUserStyleSheetLocation):
841         Use QUrl::fromEncoded since the passed string is already encoded.
842
843 2011-02-27  Adam Roben  <aroben@apple.com>
844
845         Ensure $testResultsDirectory is an absolute path before setting up Windows crash log saving
846
847         Windows 7 (and maybe Vista?) launches the post-mortem debugger with a working directory that
848         is different from old-run-webkit-test's. Because of this, the path to the crash log file
849         (which is relative to $testResultsDirectory) must be an absolute path. See the bug for more
850         details.
851
852         Fixes <http://webkit.org/b/55318> Crash logs aren't saved on Windows 7 when a relative
853         --results-directory path is used (like on the test slaves)
854
855         Reviewed by Dan Bernstein.
856
857         * Scripts/old-run-webkit-tests: Moved the call to setUpWindowsCrashLogSaving after we've
858         made $testResultsDirectory an absolute path.
859
860 2011-02-26  Dominic Mazzoni  <dmazzoni@google.com>
861
862         Reviewed by Adam Barth.
863
864         m_dumpAccessibilityNotifications should be initialized.
865         https://bugs.webkit.org/show_bug.cgi?id=54874
866
867         * DumpRenderTree/chromium/AccessibilityController.cpp:
868         (AccessibilityController::AccessibilityController):
869
870 2011-02-26  Vsevolod Vlasov  <vsevik@chromium.org>
871
872         Reviewed by Pavel Feldman.
873
874         DumpRenderTree should reset frame opener between tests.
875         https://bugs.webkit.org/show_bug.cgi?id=54874
876
877         Added clearOpener() calls between tests.
878         Some tests are using window.open() to change current window location.
879         This changes window.opener property whose nullity is checked before
880         showing xml viewer.
881
882         * DumpRenderTree/chromium/TestShell.cpp:
883         (TestShell::resetTestController):
884         * DumpRenderTree/gtk/DumpRenderTree.cpp:
885         (resetDefaultsToConsistentValues):
886         * DumpRenderTree/mac/DumpRenderTree.mm:
887         (resetWebViewToConsistentStateBeforeTesting):
888         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
889         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
890         * DumpRenderTree/win/DumpRenderTree.cpp:
891         (resetWebViewToConsistentStateBeforeTesting):
892         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
893         (WTR::InjectedBundlePage::reset):
894
895 2011-02-26  Yael Aharon  <yael.aharon@nokia.com>
896
897         Reviewed by Andreas Kling.
898
899         [Qt] Fix the focus in MiniBrowser.
900         https://bugs.webkit.org/show_bug.cgi?id=55288
901
902         Give focus to the web view when a load starts.
903
904         * MiniBrowser/qt/BrowserWindow.cpp:
905         (BrowserWindow::urlChanged):
906
907 2011-02-26  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
908
909         Reviewed by Adam Barth.
910
911         [EFL] Add coding style exceptions for EFL port
912         https://bugs.webkit.org/show_bug.cgi?id=54733
913
914         The EFL APIs use EFL naming style, which includes both lower-cased and camel-cased,
915         underscore-sparated values. But, style checking rule have found parameter_name errors
916         in efl directories so far.
917
918         * Scripts/webkitpy/style/checker.py: Add prameter_name rule's exception and efl directories.
919
920 2011-02-26  David Levin  <levin@chromium.org>
921
922         Reviewed by Darin Adler.
923
924         check-webkit-style should check for spaces after periods in comments
925         https://bugs.webkit.org/show_bug.cgi?id=55253
926
927         * Scripts/webkitpy/style/checkers/cpp.py: Added the check.
928         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added tests
929           and fixed the name of one related test.
930
931 2011-02-26  Kenneth Rohde Christiansen  <kenneth@webkit.org>
932
933         Reviewed by Andreas Kling.
934
935         Make it possible to test the targetdensity-dpi support
936         https://bugs.webkit.org/show_bug.cgi?id=55142
937
938         Test the viewport meta tag feature targetdensity-dpi by
939         adding extra arguments to dumpConfigurationForViewport
940
941         Add the new arguments to the shared header and all
942         implementation.
943
944         * DumpRenderTree/LayoutTestController.cpp:
945         (dumpConfigurationForViewportCallback):
946         * DumpRenderTree/LayoutTestController.h:
947         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
948         (LayoutTestController::dumpConfigurationForViewport):
949         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
950         (LayoutTestController::dumpConfigurationForViewport):
951         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
952         (LayoutTestController::dumpConfigurationForViewport):
953         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
954         (LayoutTestController::dumpConfigurationForViewport):
955         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
956         (LayoutTestController::dumpConfigurationForViewport):
957         * DumpRenderTree/qt/LayoutTestControllerQt.h:
958
959 2011-02-26  Sheriff Bot  <webkit.review.bot@gmail.com>
960
961         Unreviewed, rolling out r79764.
962         http://trac.webkit.org/changeset/79764
963         https://bugs.webkit.org/show_bug.cgi?id=55295
964
965         "broke Chromium builds" (Requested by rniwa on #webkit).
966
967         * DumpRenderTree/chromium/TestShell.cpp:
968         (TestShell::resetTestController):
969         * DumpRenderTree/gtk/DumpRenderTree.cpp:
970         (resetDefaultsToConsistentValues):
971         * DumpRenderTree/mac/DumpRenderTree.mm:
972         (resetWebViewToConsistentStateBeforeTesting):
973         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
974         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
975         * DumpRenderTree/win/DumpRenderTree.cpp:
976         (resetWebViewToConsistentStateBeforeTesting):
977         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
978         (WTR::InjectedBundlePage::reset):
979
980 2011-02-26  Yongjun Zhang  <yongjun_zhang@apple.com>
981
982         Reviewed by David Kilzer.
983
984         https://bugs.webkit.org/show_bug.cgi?id=48781
985
986         Add a resource load delegate method to query if WebCore should paint the default broken image for failed images.
987
988         Add a new resource load client method (shouldPaintBrokenImage).  WebKit client can decide if WebCore
989         should paint the default broken image when an image fails to load or decode.
990
991         * DumpRenderTree/LayoutTestController.cpp:
992         (LayoutTestController::LayoutTestController):
993         (setShouldPaintBrokenImageCallback):
994         (LayoutTestController::staticFunctions):
995         (LayoutTestController::setShouldPaintBrokenImage):
996         * DumpRenderTree/LayoutTestController.h:
997         (LayoutTestController::shouldPaintBrokenImage):
998         * DumpRenderTree/mac/ResourceLoadDelegate.mm:
999         (-[ResourceLoadDelegate webView:shouldPaintBrokenImageForURL:]):
1000
1001 2011-02-26  Vsevolod Vlasov  <vsevik@chromium.org>
1002
1003         Reviewed by Pavel Feldman.
1004
1005         DumpRenderTree should reset frame opener between tests.
1006         https://bugs.webkit.org/show_bug.cgi?id=54874
1007
1008         Added clearOpener() calls between tests.
1009         Some tests are using window.open() to change current window location.
1010         This changes window.opener property whose nullity is checked before
1011         showing xml viewer.
1012
1013         * DumpRenderTree/chromium/TestShell.cpp:
1014         (TestShell::resetTestController):
1015         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1016         (resetDefaultsToConsistentValues):
1017         * DumpRenderTree/mac/DumpRenderTree.mm:
1018         (resetWebViewToConsistentStateBeforeTesting):
1019         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1020         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
1021         * DumpRenderTree/win/DumpRenderTree.cpp:
1022         (resetWebViewToConsistentStateBeforeTesting):
1023         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1024         (WTR::InjectedBundlePage::reset):
1025
1026 2011-02-25  Lukasz Slachciak  <l.slachciak@samsung.com>
1027
1028         Reviewed by Kenneth Rohde Christiansen.
1029
1030         [EFL] Added EWebLauncher command line option to turn on/off frame flattening
1031         https://bugs.webkit.org/show_bug.cgi?id=54809
1032
1033         * EWebLauncher/main.c: Handling -f option added.
1034         (on_key_down):
1035         (browserCreate):
1036         (main):
1037
1038 2011-02-25  Sheriff Bot  <webkit.review.bot@gmail.com>
1039
1040         Unreviewed, rolling out r79324.
1041         http://trac.webkit.org/changeset/79324
1042         https://bugs.webkit.org/show_bug.cgi?id=55260
1043
1044         Leopard Debug is once again running out of address space while
1045         linking (Requested by aroben on #webkit).
1046
1047         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
1048         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1049
1050 2011-02-25  Tony Chang  <tony@chromium.org>
1051
1052         Reviewed by James Robinson.
1053
1054         Clean up temp files left by crashing DRTs on chromium-win
1055         https://bugs.webkit.org/show_bug.cgi?id=55182
1056
1057         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1058         * BuildSlaveSupport/chromium/remove-crash-logs: Added.
1059
1060 2011-02-25  Adam Roben  <aroben@apple.com>
1061
1062         Double-check that modifying the registry worked
1063
1064         On Windows Vista/7 with UAC enabled, regtool will fail to modify the registry, but it will
1065         still return a successful exit code. So we double-check here that the value we tried to
1066         write to the registry was really written.
1067
1068         Fixes <http://webkit.org/b/55227> old-run-webkit-tests falsely claims to be saving crash
1069         logs on Windows Vista/7
1070
1071         Reviewed by Darin Adler.
1072
1073         * Scripts/old-run-webkit-tests:
1074         (writeRegistryString): After trying to write the string value, check that it was really
1075         written.
1076
1077 2011-02-25  Adam Roben  <aroben@apple.com>
1078
1079         Set the 32-bit post-mortem debugger on 64-bit OSes
1080
1081         The post-mortem debugger is controlled by some registry values in
1082         HKLM\SOFTWARE\Microsoft\Windows NT\CurrentVersion\AeDebug. On 64-bit OSes, you can either
1083         set the values in that key, which will be used for 64-bit apps, or in the equivalent key
1084         under Wow6432Node, which will be used for 32-bit apps. Since DumpRenderTree and
1085         WebKitTestRunner are 32-bit, we want to use the latter location.
1086
1087         Fixes <http://webkit.org/b/55225> old-run-webkit-tests sets the post-mortem debugger for
1088         64-bit apps, but should set it for 32-bit apps instead
1089
1090         Reviewed by Darin Adler.
1091
1092         * Scripts/old-run-webkit-tests:
1093         (readRegistryString):
1094         (writeRegistryString):
1095         Pass --wow32 to regtool so it will get/set values beneath Wow6432Node, which is where
1096         Windows looks for the post-mortem debugger values when a 32-bit app crashes on a 64-bit OS.
1097         Passing --wow32 on a 32-bit OS will not have any effect.
1098
1099 2011-02-25  Adam Roben  <aroben@apple.com>
1100
1101         Clean up code to set up crash log saving on Windows
1102
1103         Fixes <http://webkit.org/b/55224> old-run-webkit-test's Windows registry code is ugly
1104
1105         Reviewed by David Kilzer.
1106
1107         * Scripts/old-run-webkit-tests:
1108         (readRegistryString):
1109         (writeRegistryString):
1110         Moved code to read strings from and write strings to the registry here...
1111
1112         (setUpWindowsCrashLogSaving): ...from here.
1113
1114 2011-02-25  Adam Roben  <aroben@apple.com>
1115
1116         Work around Cygwin's crash-suppression behavior
1117
1118         Cygwin calls ::SetErrorMode(SEM_FAILCRITICALERRORS), which any processes it launches will
1119         inherit. This is bad for testing/debugging, as it causes the post-mortem debugger not to be
1120         invoked. (Cygwin does this because it makes crashes more UNIX-y.) We reset the error mode
1121         when our test apps launch to work around Cygwin's behavior.
1122
1123         Fixes <http://webkit.org/b/55222> Test apps crash silently (without invoking post-mortem
1124         debugger) when launched from Cygwin 1.7
1125
1126         Reviewed by Darin Adler.
1127
1128         * DumpRenderTree/win/DumpRenderTree.cpp:
1129         (main):
1130         * TestWebKitAPI/win/main.cpp:
1131         (main):
1132         * WebKitAPITest/main.cpp:
1133         (main):
1134         * WebKitTestRunner/win/TestControllerWin.cpp:
1135         (WTR::TestController::platformInitialize):
1136         Call ::SetErrorMode(0) to undo Cygwin's folly.
1137
1138 2011-02-24  Jocelyn Turcotte  <jocelyn.turcotte@nokia.com>
1139
1140         Reviewed by Andreas Kling.
1141
1142         [Qt] Revert the support for QNAM affined to a different thread.
1143         https://bugs.webkit.org/show_bug.cgi?id=55149
1144
1145         Qt 4.8 will have QNAM use its own thread internally by default,
1146         no need to keep this complexity in WebKit.
1147
1148         This mainly reverts:
1149         http://trac.webkit.org/changeset/73710
1150         http://trac.webkit.org/changeset/73712
1151
1152         * QtTestBrowser/launcherwindow.cpp:
1153         (LauncherWindow::~LauncherWindow):
1154         (LauncherWindow::initializeView):
1155         (LauncherWindow::createChrome):
1156         * QtTestBrowser/launcherwindow.h:
1157         (WindowOptions::WindowOptions):
1158         * QtTestBrowser/webpage.cpp:
1159         (WebPage::WebPage):
1160         * QtTestBrowser/webpage.h:
1161
1162 2011-02-24  Adam Barth  <abarth@webkit.org>
1163
1164         Reviewed by Eric Seidel.
1165
1166         Add a script for creating gypi files
1167         https://bugs.webkit.org/show_bug.cgi?id=55113
1168
1169         This is a trivial script to create gypi files from the source code in
1170         the current directory.
1171
1172         * Scripts/make-gypi: Added.
1173
1174 2011-02-24  Sheriff Bot  <webkit.review.bot@gmail.com>
1175
1176         Unreviewed, rolling out r79628.
1177         http://trac.webkit.org/changeset/79628
1178         https://bugs.webkit.org/show_bug.cgi?id=55195
1179
1180         "broke chromium win tests" (Requested by dpranke on #webkit).
1181
1182         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1183         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1184
1185 2011-02-24  Dirk Pranke  <dpranke@chromium.org>
1186
1187         Reviewed by Mihai Parparita.
1188
1189         rebaseline-chromium-webkit-tests can't handle GPU variants
1190         https://bugs.webkit.org/show_bug.cgi?id=55002
1191
1192         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
1193         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
1194
1195 2011-02-24  Dirk Pranke  <dpranke@chromium.org>
1196
1197         Reviewed by Mihai Parparita.
1198
1199         nrwt: log command used to start web server to --verbose
1200
1201         https://bugs.webkit.org/show_bug.cgi?id=55188
1202
1203         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1204         * Scripts/webkitpy/layout_tests/port/http_server.py:
1205
1206 2011-02-24  Victoria Kirst  <vrk@google.com>
1207
1208         Reviewed by Mihai Parparita.
1209
1210         [chromium] Move media-file.js and video-test.js back to media/ for UILayoutTests
1211         https://bugs.webkit.org/show_bug.cgi?id=55089
1212
1213         This adds an alias into the media/ directory so that http/tests/media
1214         tests can access the media resources when running in an httpd process.
1215
1216         * Scripts/webkitperl/httpd.pm:
1217         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1218         * Scripts/webkitpy/layout_tests/port/http_server.py:
1219
1220 2011-02-24  Dirk Pranke  <dpranke@chromium.org>
1221
1222         Reviewed by Mihai Parparita.
1223
1224         nrwt: enable multiple processes by default on Chromium Win.
1225
1226         https://bugs.webkit.org/show_bug.cgi?id=55163
1227
1228         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1229         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1230
1231 2011-02-24  Sam Weinig  <sam@webkit.org>
1232
1233         Attempt to fix Qt.
1234
1235         * WebKitTestRunner/TestInvocation.cpp:
1236
1237 2011-02-24  James Robinson  <jamesr@chromium.org>
1238
1239         Unreviewed, rolling out r79598.
1240         http://trac.webkit.org/changeset/79598
1241         https://bugs.webkit.org/show_bug.cgi?id=55089
1242
1243         Appears to be causing trouble with httpd on windows
1244
1245         * Scripts/webkitperl/httpd.pm:
1246         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1247         * Scripts/webkitpy/layout_tests/port/http_server.py:
1248
1249 2011-02-24  Sam Weinig  <sam@webkit.org>
1250
1251         Reviewed by Darin Adler.
1252
1253         Add basic pixel testing support to WebKitTestRunner
1254         https://bugs.webkit.org/show_bug.cgi?id=55161
1255
1256         * Scripts/old-run-webkit-tests:
1257         Don't print color space information when using WebKitTestRunner since
1258         we don't need to change the screen.
1259
1260         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
1261         (WTR::InjectedBundle::initialize):
1262         Add call to initializePlatformDefaults.
1263         
1264         (WTR::InjectedBundle::done):
1265         Return a dictionary on completion with both the text output and pixel dump override.
1266
1267         * WebKitTestRunner/InjectedBundle/InjectedBundle.h:
1268         Add declaration for initializePlatformDefaults.
1269
1270         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
1271         (WTR::LayoutTestController::LayoutTestController):
1272         (WTR::LayoutTestController::dumpAsText):
1273         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
1274         (WTR::LayoutTestController::shouldDumpPixels):
1275         Make setting dumpAsText override the dump pixels preference.
1276
1277         * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm: Added.
1278         (WTR::InjectedBundle::initializePlatformDefaults):
1279         Set platform defaults matching DumpRenderTree.
1280
1281         * WebKitTestRunner/InjectedBundle/qt/InjectedBundleQt.cpp: Added.
1282         (WTR::InjectedBundle::initializePlatformDefaults):
1283         * WebKitTestRunner/InjectedBundle/win/InjectedBundleWin.cpp: Added.
1284         (WTR::InjectedBundle::initializePlatformDefaults):
1285         Add stubs for initializePlatformDefaults.
1286
1287         * WebKitTestRunner/TestController.cpp:
1288         (WTR::TestController::runTest):
1289         (WTR::TestController::runTestingServerLoop):
1290         (WTR::TestController::run):
1291         Parse expected pixel results out of stdin and setup the invocation with it.
1292
1293         * WebKitTestRunner/TestInvocation.cpp:
1294         (WTR::TestInvocation::TestInvocation):
1295         Initialize new members.
1296
1297         (WTR::TestInvocation::~TestInvocation):
1298         (WTR::TestInvocation::setIsPixelTest):
1299         (WTR::sizeWebViewForCurrentTest):
1300         (WTR::TestInvocation::invoke):
1301         (WTR::TestInvocation::dump):
1302         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
1303         * WebKitTestRunner/TestInvocation.h:
1304         Dump pixels in addition to text by calling dumpPixelsAndCompareWithExpected.
1305
1306         * WebKitTestRunner/mac/PlatformWebViewMac.mm:
1307         (WTR::PlatformWebView::PlatformWebView):
1308         Use -[NSColorSpace genericRGBColorSpace] instead of the main screen color space.
1309
1310         * WebKitTestRunner/mac/TestInvocationMac.mm: Added.
1311         (WTR::createCGContextFromPlatformView):
1312         (WTR::computeMD5HashStringForContext):
1313         (WTR::dumpBitmap):
1314         (WTR::forceRepaintFunction):
1315         (WTR::TestInvocation::dumpPixelsAndCompareWithExpected):
1316         Add mac specific pixel dumping support. Right now, we always pull pixels
1317         from the window server.
1318
1319         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
1320         * WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro:
1321         * WebKitTestRunner/qt/WebKitTestRunner.pro:
1322         * WebKitTestRunner/win/InjectedBundle.vcproj:
1323         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
1324         Add new files.
1325
1326 2011-02-24  Chang Shu  <cshu@webkit.org>
1327
1328         Reviewed by Kenneth Rohde Christiansen.
1329
1330         [Qt] Launch MiniBrowser if the test run is WebKit2. 
1331         https://bugs.webkit.org/show_bug.cgi?id=55145
1332
1333         * Scripts/old-run-webkit-tests:
1334         * Scripts/run-launcher:
1335         * Scripts/webkitdirs.pm:
1336
1337 2011-02-24  Victoria Kirst  <vrk@google.com>
1338
1339         Reviewed by Mihai Parparita.
1340
1341         [chromium] Move media-file.js and video-test.js back to media/ for UILayoutTests
1342         https://bugs.webkit.org/show_bug.cgi?id=55089
1343
1344         This adds an alias into the media/ directory so that http/tests/media
1345         tests can access the media resources when running in an httpd process.
1346
1347         * Scripts/webkitperl/httpd.pm:
1348         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1349         * Scripts/webkitpy/layout_tests/port/http_server.py:
1350
1351 2011-02-24  Andrew Wilson  <atwilson@chromium.org>
1352
1353         Unreviewed, rolling out r79570.
1354         http://trac.webkit.org/changeset/79570
1355         https://bugs.webkit.org/show_bug.cgi?id=54874
1356
1357         Breaks chromium build because glue/mocks/mock_web_frame.h/cc
1358         was not updated
1359
1360         * DumpRenderTree/chromium/TestShell.cpp:
1361         (TestShell::resetTestController):
1362         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1363         (resetDefaultsToConsistentValues):
1364         * DumpRenderTree/mac/DumpRenderTree.mm:
1365         (resetWebViewToConsistentStateBeforeTesting):
1366         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1367         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
1368         * DumpRenderTree/win/DumpRenderTree.cpp:
1369         (resetWebViewToConsistentStateBeforeTesting):
1370         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1371         (WTR::InjectedBundlePage::reset):
1372
1373 2011-02-24  Mahesh Kulkarni  <mahesh.kulkarni@nokia.com>
1374
1375         Reviewed by Laszlo Gombos.
1376
1377         [Qt] add android 2.2 user agent to QtTestBrowser
1378         https://bugs.webkit.org/show_bug.cgi?id=55085
1379
1380         Credit to Forrest Hodgkins to find right user-agent which works for youtube.
1381         * QtTestBrowser/useragentlist.txt:
1382
1383 2011-02-24  Adam Roben  <aroben@apple.com>
1384
1385         Change FrameLoadDelegate to support any number of delegates with delayed work to process
1386
1387         This makes our behavior match Mac more closely, and allows us to remove an incorrect
1388         assertion that was firing during some tests. (The assertion was claiming that there was
1389         never more than one delegate with delayed work to process, but that was not the case.)
1390
1391         Fixes <http://webkit.org/b/55146> Assertion failure in FrameLoadDelegate::locationChangeDone
1392         when running http/tests/navigation/back-twice-without-commit.html
1393
1394         Reviewed by Eric Carlson.
1395
1396         * DumpRenderTree/win/FrameLoadDelegate.cpp:
1397         (delegatesWithDelayedWork): Added. Returns all FrameLoadDelegates that have delayed work to
1398         process. A single delegate may appear in this Vector more than once (just as, on Mac, a
1399         single delegate may have multiple performSelector requests).
1400         (processWorkTimer): Pass the HWND to ::KillTimer, for pedantic brownie points. Added an
1401         assertion that the timer firing is the shared process work timer. Instead of using the
1402         single, global "delegate waiting for timer" delegate, give all delegates that have delayed
1403         work to process a chance to process their work.
1404         (FrameLoadDelegate::locationChangeDone): If we don't already have an active timer for
1405         processing delayed work, create one. Then add ourselves to the delegatesWithDelayedWork
1406         Vector so our processWork function will be called when the timer fires.
1407
1408 2011-02-24  Vsevolod Vlasov  <vsevik@chromium.org>
1409
1410         Reviewed by Alexey Proskuryakov.
1411
1412         DumpRenderTree should reset frame opener between tests.
1413         https://bugs.webkit.org/show_bug.cgi?id=54874
1414
1415         Added clearOpener() calls between tests.
1416         Some tests are using window.open() to change current window location.
1417         This changes window.opener property whose nullity is checked before
1418         showing xml viewer.
1419
1420         * DumpRenderTree/chromium/TestShell.cpp:
1421         (TestShell::resetTestController):
1422         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1423         (resetDefaultsToConsistentValues):
1424         * DumpRenderTree/mac/DumpRenderTree.mm:
1425         (resetWebViewToConsistentStateBeforeTesting):
1426         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1427         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
1428         * DumpRenderTree/win/DumpRenderTree.cpp:
1429         (resetWebViewToConsistentStateBeforeTesting):
1430         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1431         (WTR::InjectedBundlePage::reset):
1432
1433 2011-02-24  Adam Roben  <aroben@apple.com>
1434
1435         Don't trigger Windows builds for changes to WebCore.exp.in or the top-level GNUmakefile.am
1436
1437         Fixes <http://webkit.org/b/55144> Windows builders should not have built r79343 or r79440,
1438         but did
1439
1440         Reviewed by Anders Carlsson.
1441
1442         * Scripts/webkitpy/common/config/build.py:
1443         (_should_file_trigger_build): Loosened the pattern for GNUmakefile.am to include the
1444         top-level GNUmakefile.am, and loosened the pattern for .exp files to include WebCore.exp.in.
1445
1446         * Scripts/webkitpy/common/config/build_unittest.py:
1447         (ShouldBuildTest): Added some more tests.
1448
1449 2011-02-24  Andras Becsi  <abecsi@webkit.org>
1450
1451         Reviewed by Laszlo Gombos.
1452
1453         [Qt] MinGW build fails to link
1454         https://bugs.webkit.org/show_bug.cgi?id=55050
1455
1456         Prepend the libraries of subcomponents instead of appending them
1457         to fix the library order according to the dependency of the libraries
1458
1459         * WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro:
1460         * WebKitTestRunner/qt/WebKitTestRunner.pro:
1461
1462 2011-02-24  Gabor Loki  <loki@webkit.org>
1463
1464         Reviewed by Csaba Osztrogonác.
1465
1466         [Qt] Remove ARMv5 Release buildbot.
1467
1468         Remove ARMv5 Release buildbot, because there is very little difference
1469         between ARMv5 and ARMv7 Release bots, and there is also very little
1470         interest in ARMv5 build slave (compared to ARMv7).
1471
1472         * BuildSlaveSupport/build.webkit.org-config/config.json:
1473         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1474
1475 2011-02-22  Ojan Vafai  <ojan@chromium.org>
1476
1477         Reviewed by Tony Chang.
1478
1479         fix revision number finding when creating git patches
1480         https://bugs.webkit.org/show_bug.cgi?id=55024
1481
1482         HEAD~n does not walk merge commits correctly. git log however does.
1483
1484         * Scripts/webkitpy/common/checkout/scm.py:
1485         * Scripts/webkitpy/common/checkout/scm_unittest.py:
1486
1487 2011-02-23  Kenneth Russell  <kbr@google.com>
1488
1489         Unreviewed, rolling out r79387.
1490         https://bugs.webkit.org/show_bug.cgi?id=54885
1491
1492         Reapply the code changes from r79038 since they weren't the root
1493         cause of the test timeouts.
1494
1495         * DumpRenderTree/chromium/WebViewHost.cpp:
1496         (WebViewHost::reset):
1497
1498 2011-02-23  Siddharth Mathur  <siddharth.mathur@nokia.com>
1499
1500         Reviewed by Laszlo Gombos.
1501
1502         [Qt] Make sure Symbian binary UIDs are unique
1503
1504         * MiniBrowser/qt/MiniBrowser.pro: Resolve the UID collision with
1505         QtTestBrowser.pro by changing the UID.
1506
1507 2011-02-23  Patrick Gansterer  <paroga@webkit.org>
1508
1509         Reviewed by Darin Adler.
1510
1511         Rename PLATFORM(CF) to USE(CF)
1512         https://bugs.webkit.org/show_bug.cgi?id=53540
1513
1514         * DumpRenderTree/config.h:
1515
1516 2011-02-23  Dirk Pranke  <dpranke@chromium.org>
1517
1518         Reviewed by Ojan Vafai.
1519
1520         nrwt: turn on multiprocessing on chromium_linux, clean up unit tests for worker model defaults
1521         https://bugs.webkit.org/show_bug.cgi?id=55016
1522
1523         * Scripts/webkitpy/layout_tests/port/base.py:
1524         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1525         * Scripts/webkitpy/layout_tests/port/chromium_linux_unittest.py: Added.
1526         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1527         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
1528         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1529         * Scripts/webkitpy/layout_tests/port/mac.py:
1530         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1531         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
1532         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
1533
1534 2011-02-23  Tony Chang  <tony@chromium.org>
1535
1536         Reviewed by Kent Tamura.
1537
1538         [chromium] DumpRenderTree --check-layout-test-sys-deps leaks 4 temp dirs
1539         https://bugs.webkit.org/show_bug.cgi?id=55004
1540
1541         * DumpRenderTree/chromium/DumpRenderTree.cpp:
1542         (WebKitSupportTestEnvironment::WebKitSupportTestEnvironment):
1543         (WebKitSupportTestEnvironment::~WebKitSupportTestEnvironment):
1544         (main): Use a C++ object to ensure that TearDownTestEnvironment
1545             gets called when --check-layout-test-sys-deps is called
1546
1547 2011-02-23  Gopal Raghavan  <gopal.1.raghavan@nokia.com>
1548
1549         Reviewed by Anders Carlsson.
1550
1551         [Qt] [Qt] check webkit style fails in qt/tests
1552         https://bugs.webkit.org/show_bug.cgi?id=54998
1553
1554         check-webkit-style was complaining about missing config.h in WebKit/qt/test/qdeclarativewebview. This patch was originally part of 50222. Reviewer suggested to split and create a separate patch for checker.
1555
1556         * Scripts/webkitpy/style/checker.py:
1557
1558 2011-02-22  Vsevolod Vlasov  <vsevik@chromium.org>
1559
1560         Reviewed by Pavel Feldman.
1561
1562         DumpRenderTree should dump as text if test path contains "dumpAsText/"
1563         https://bugs.webkit.org/show_bug.cgi?id=54873
1564
1565         * DumpRenderTree/chromium/LayoutTestController.h:
1566         (LayoutTestController::setShouldDumpAsText):
1567         (LayoutTestController::setShouldGeneratePixelResults):
1568         * DumpRenderTree/chromium/TestShell.cpp:
1569         (TestShell::runFileTest):
1570         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1571         (shouldDumpAsText):
1572         (dump):
1573         (runTest):
1574         * DumpRenderTree/mac/DumpRenderTree.mm:
1575         (shouldDumpAsText):
1576         (runTest):
1577         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1578         (WebCore::isDumpAsTextTest):
1579         (WebCore::DumpRenderTree::open):
1580         * DumpRenderTree/win/DumpRenderTree.cpp:
1581         (shouldDumpAsText):
1582         (runTest):
1583         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1584         (WTR::InjectedBundlePage::dump):
1585
1586 2011-02-22  Kenneth Russell  <kbr@google.com>
1587
1588         Reviewed by James Robinson.
1589
1590         Many tests started timing out on the chrome mac canaries around r79035
1591         https://bugs.webkit.org/show_bug.cgi?id=54885
1592
1593         Revert the code changes, but not the Chromium DEPS roll, from
1594         r79038 to see whether this clears up the timing-out tests. Was not
1595         able to reproduce the timeouts locally with a Release mode build
1596         of DRT (on Snow Leopard).
1597
1598         * DumpRenderTree/chromium/WebViewHost.cpp:
1599         (WebViewHost::reset):
1600
1601 2011-02-22  Dirk Pranke  <dpranke@chromium.org>
1602
1603         Reviewed by James Robinson.
1604
1605         nrwt: --platform chromium-gpu doesn't work with --worker-model=processes
1606         https://bugs.webkit.org/show_bug.cgi?id=55009
1607
1608         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
1609
1610 2011-02-18  Mihai Parparita  <mihaip@chromium.org>
1611
1612         Reviewed by Ojan Vafai.
1613
1614         Default to 500 for --exit-after-n-failures
1615         https://bugs.webkit.org/show_bug.cgi?id=54773
1616
1617         Make 500 the default value for --exit-after-n-failures (ORWT bots
1618         already default to this as of r75726). Similarly, change the default for
1619         --exit-after-n-crashes-or-timeouts to 20.
1620
1621         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1622
1623 2011-02-02  Martin Robinson  <mrobinson@igalia.com>
1624
1625         Reviewed by Xan Lopez.
1626
1627         [GTK] DRT needs an implementation of LayoutTestController.setSelectTrailingWhitespace
1628         https://bugs.webkit.org/show_bug.cgi?id=53603
1629
1630         Add support for LayoutTestController.setSelectTrailingWhitespace by calling through
1631         to DumpRenderTreeSupportGtk for this functionality.
1632
1633         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1634         (resetDefaultsToConsistentValues): Always call LayoutTestController.setSelectTrailingWhitespace to
1635         false when reseting values between tests.
1636         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1637         (LayoutTestController::setSelectTrailingWhitespaceEnabled): Call LayoutTestController.setSelectTrailingWhitespace
1638         with the appropriate value.
1639
1640 2011-02-22  Adam Roben  <aroben@apple.com>
1641
1642         Don't trigger a Windows build when the top-level ChangeLog changes
1643
1644         r79320 is an example of a change that shouldn't have built on Windows, but did, because of
1645         this bug.
1646
1647         Fixes <http://webkit.org/b/54957> Windows bots build when top-level ChangeLog changes, but
1648         shouldn't
1649
1650         Reviewed by Anders Carlsson.
1651
1652         * Scripts/webkitpy/common/config/build.py:
1653         (_should_file_trigger_build): Modified the ChangeLog pattern to match the top-level
1654         ChangeLog, and re-sorted it in the list.
1655
1656         * Scripts/webkitpy/common/config/build_unittest.py:
1657         (ShouldBuildTest): Added a test that ChangeLogs don't trigger builds.
1658
1659 2011-02-22  Sheriff Bot  <webkit.review.bot@gmail.com>
1660
1661         Unreviewed, rolling out r78691.
1662         http://trac.webkit.org/changeset/78691
1663         https://bugs.webkit.org/show_bug.cgi?id=54956
1664
1665         Leopard Debug is successfully building again (Requested by
1666         aroben on #webkit).
1667
1668         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
1669         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1670
1671 2011-02-22  Ilya Tikhonovsky  <loislo@chromium.org>
1672
1673         Reviewed by Pavel Feldman.
1674
1675         Web Inspector: flakyness of inspector tests.
1676         https://bugs.webkit.org/show_bug.cgi?id=54729
1677
1678         As far as we have the protocol with sequence numbers we can simplify test harness support
1679         and drop out chomium specific methods.
1680
1681         * DumpRenderTree/DumpRenderTree.gypi:
1682         * DumpRenderTree/chromium/DRTDevToolsAgent.cpp:
1683         (DRTDevToolsAgent::sendMessageToInspectorFrontend):
1684         (DRTDevToolsAgent::runtimePropertyChanged):
1685         (DRTDevToolsAgent::asyncCall):
1686         (DRTDevToolsAgent::call):
1687         * DumpRenderTree/chromium/DRTDevToolsAgent.h:
1688         * DumpRenderTree/chromium/DRTDevToolsCallArgs.cpp: Removed.
1689         * DumpRenderTree/chromium/DRTDevToolsCallArgs.h: Removed.
1690         * DumpRenderTree/chromium/DRTDevToolsClient.cpp:
1691         (DRTDevToolsClient::sendMessageToBackend):
1692         (DRTDevToolsClient::asyncCall):
1693         (DRTDevToolsClient::call):
1694         * DumpRenderTree/chromium/DRTDevToolsClient.h:
1695
1696 2011-02-22  Andras Becsi  <abecsi@webkit.org>
1697
1698         Reviewed by Laszlo Gombos.
1699
1700         [Qt] Redesign the build system
1701         https://bugs.webkit.org/show_bug.cgi?id=51339
1702
1703         Part 2.
1704
1705         Build WebCore as a static library, compile the WebKit API and WebKit2 API
1706         in a final step and link to WebKit2, WebCore and JSC libraries to fix
1707         linking issues resulting from stripped away symbols.
1708
1709         * WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro: Remove superfluous includepaths.
1710
1711 2011-02-21  Lukasz Slachciak  <l.slachciak@samsung.com>
1712
1713         Reviewed by Antonio Gomes.
1714
1715         [EFL] Restored spatial navigation option in EWebLauncher since #18662 bug is closed
1716         https://bugs.webkit.org/show_bug.cgi?id=54806
1717
1718         * EWebLauncher/main.c:
1719         (on_key_down): calling ewk_view_setting_spatial_navigation_set when "F12" pressed added.
1720
1721 2011-02-21  Dirk Pranke  <dpranke@chromium.org>
1722
1723         Reviewed by Kent Tamura.
1724
1725         nrwt: revert behavior on 10.5 to "old-threads". It's possible
1726         that r79062 has introduced a bunch of flakiness, although I
1727         can't explain why right now. This change should restore the
1728         old behavior, and then we can look at the build logs to see
1729         if the tests stop timing out or being flaky.
1730
1731         https://bugs.webkit.org/show_bug.cgi?id=54925
1732
1733         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1734         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
1735         * Scripts/webkitpy/layout_tests/port/mac.py:
1736         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1737
1738 2011-02-21  Dirk Pranke  <dpranke@chromium.org>
1739
1740         Reviewed by Dimitri Glazkov.
1741
1742         nrwt: fix typo causing us to try and repeatedly shut down the servers
1743         https://bugs.webkit.org/show_bug.cgi?id=54904
1744
1745         * Scripts/webkitpy/layout_tests/layout_package/worker_mixin.py:
1746
1747 2011-02-21  Alejandro G. Castro  <alex@igalia.com>
1748
1749         Reviewed by Csaba Osztrogonác.
1750
1751         REGRESSION(r79157):
1752         plugins/get-url-with-javascript-destroying-plugin.html make the
1753         next test crash
1754         https://bugs.webkit.org/show_bug.cgi?id=54863
1755
1756         * GNUmakefile.am:
1757
1758 2011-02-20  Dirk Pranke  <dpranke@chromium.org>
1759
1760         Reviewed by Ojan Vafai.
1761
1762         test-webkitpy: stop skipping a bunch of tests on win32
1763         https://bugs.webkit.org/show_bug.cgi?id=54788
1764
1765         * Scripts/webkitpy/common/net/testoutputset_unittest.py:
1766         * Scripts/webkitpy/test/main.py:
1767
1768 2011-02-20  Dirk Pranke  <dpranke@chromium.org>
1769
1770         Reviewed by Ojan Vafai.
1771
1772         Fix various bugs related to handling of dir separators and real
1773         vs. fake filesystems that were causing test-webkitpy to fail
1774         tests on win32. These bugs have a variety of causes but are
1775         lumped together because they should be easy to review.
1776
1777         https://bugs.webkit.org/show_bug.cgi?id=54700
1778
1779         * Scripts/webkitpy/common/net/testoutput.py:
1780           Here we were using os.path.sep but not escaping it prior to
1781           using in a regexp, which  caused bad things on win32.
1782         * Scripts/webkitpy/common/net/testoutput_unittest.py:
1783           Same here.
1784         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
1785           Here we were using the native filesystem instead of the mock
1786           filesystem, making some tests unnecessarily platform dependent.
1787           Also we weren't setting the default configuration for the
1788           TestChromiumWinPort.
1789         * Scripts/webkitpy/layout_tests/update_webgl_conformance_tests.py:
1790           Here we *shouldn't* use os.path.join(), just so the code becomes
1791           less platform dependent.
1792         * Scripts/webkitpy/style/checker.py:
1793           Here we should be using os.path.sep instead of "/"; the tests
1794           require the native filesystem.
1795         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
1796           Here the underlying code is using os.path.sep, so we probably
1797           should as well.
1798         * Scripts/webkitpy/style/test_expectations_unittest.py:
1799           Here we should be using the filesystem hanging of the port
1800           object (which is the mocked filesystem) rather than the native
1801           one.
1802
1803 2011-02-20  Anders Carlsson  <andersca@apple.com>
1804
1805         Reviewed by Maciej Stachowiak.
1806
1807         Crash when a plug-in requests a javascript: url that destroys the plug-in
1808         https://bugs.webkit.org/show_bug.cgi?id=54837
1809         <rdar://problem/9005475>
1810
1811         Add new plug-in test.
1812
1813         * DumpRenderTree/DumpRenderTree.gypi:
1814         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
1815         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
1816         (PluginTest::NPN_GetURL):
1817         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
1818         * DumpRenderTree/TestNetscapePlugIn/Tests/GetURLWithJavaScriptURLDestroyingPlugin.cpp: Added.
1819         (GetURLWithJavaScriptURLDestroyingPlugin::GetURLWithJavaScriptURLDestroyingPlugin):
1820         (GetURLWithJavaScriptURLDestroyingPlugin::NPP_New):
1821         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1822         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
1823
1824 2011-02-19  Patrick Gansterer  <paroga@webkit.org>
1825
1826         Reviewed by Andreas Kling.
1827
1828         Make WinCE a core builder
1829         https://bugs.webkit.org/show_bug.cgi?id=54801
1830
1831         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
1832         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1833
1834 2011-02-19  Lukasz Slachciak  <l.slachciak@samsung.com>
1835
1836         Reviewed by Andreas Kling.
1837
1838         Fixed problem with launching EWebLauncher with "run-launcher --efl". Incorrect path referring to non-existing .libs directory ws set.
1839         https://webkit.org/b/54778
1840
1841         * Scripts/webkitdirs.pm:
1842
1843 2011-02-18  David Levin  <levin@chromium.org>
1844
1845         Reviewed by Eric Seidel.
1846
1847         check-webkit-style: Misses brace style error when the line with the { has a } in it.
1848         https://bugs.webkit.org/show_bug.cgi?id=54769
1849
1850         * Scripts/webkitpy/style/checkers/cpp.py: Fixed the check for the close brace
1851           to only look after the last open brace, so that the open brace in this line
1852           "} else {" will still be able to trigger the error.
1853         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added related unit tests.
1854
1855 2011-02-18  Zan Dobersek  <zandobersek@gmail.com>
1856
1857         Reviewed by Martin Robinson.
1858
1859         [Gtk] Re-enable meter tag support
1860         https://bugs.webkit.org/show_bug.cgi?id=54762
1861
1862         Meter tag is supported on the Gtk port, enable it by default.
1863
1864         * Scripts/build-webkit:
1865
1866 2011-02-18  David Levin  <levin@chromium.org>
1867
1868         Reviewed by Eric Seidel.
1869
1870         check-webkit-style falsely complains about WebKitGTK+ public headers
1871         https://bugs.webkit.org/show_bug.cgi?id=54650
1872
1873         * Scripts/webkitpy/style/checker.py: Add some exceptions for the gtk files.
1874         * Scripts/webkitpy/style/checkers/cpp.py: Restrict the checks for WEBKIT_API
1875           to the chromium directory (and improved the checks slightly).
1876         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added corresponding tests.
1877
1878 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1879
1880         Reviewed by James Robinson.
1881
1882         new-run-webkit-tests: enable multiple processes by default on
1883         mac. This change removes the artificial restrictions we placed
1884         on the # of child processes we used with the old threading
1885         model, and switches to the new message-based model and multiple
1886         processes, where available. If multiple processes are not
1887         available (Leopard / Python 2.5), then we use the 'inline' model
1888         instead of the 'old-threads' model on the 'mac' port or one
1889         process on the 'chromium-mac' port. We need additional testing
1890         to see if the new 'threads' model works reliably and is worth
1891         supporting, or if we should just live with things being slightly
1892         slow.
1893
1894         https://bugs.webkit.org/show_bug.cgi?id=54596
1895
1896         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1897         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
1898         * Scripts/webkitpy/layout_tests/port/mac.py:
1899         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1900         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1901
1902 2011-02-18  Andrew Wilson  <atwilson@chromium.org>
1903
1904         Unreviewed, rolling out r79047.
1905         http://trac.webkit.org/changeset/79047
1906         https://bugs.webkit.org/show_bug.cgi?id=54596
1907
1908         Broke canary bots - please watch the canaries the next time
1909         you land this
1910
1911         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1912         * Scripts/webkitpy/layout_tests/port/mac.py:
1913
1914 2011-02-18  Emil A Eklund  <eae@chromium.org>
1915
1916         Reviewed by Darin Adler.
1917
1918         Crash in EventHandler::sendContextMenuEventForKey
1919         https://bugs.webkit.org/show_bug.cgi?id=54495
1920
1921         Add support for the context menu key (VK_APPS) to EventSender for
1922         platforms that support the key.
1923
1924         * DumpRenderTree/chromium/EventSender.cpp:
1925         (EventSender::keyDown):
1926         * DumpRenderTree/gtk/EventSender.cpp:
1927         (keyDownCallback):
1928         * DumpRenderTree/qt/EventSenderQt.cpp:
1929         (EventSender::keyDown):
1930         * DumpRenderTree/win/EventSender.cpp:
1931         (keyDownCallback):
1932
1933 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1934
1935         Reviewed by Ojan Vafai.
1936
1937         new-run-webkit-tests: enable multiple processes by default on mac
1938         https://bugs.webkit.org/show_bug.cgi?id=54596
1939
1940         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1941         * Scripts/webkitpy/layout_tests/port/mac.py:
1942
1943 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1944
1945         Reviewed by Adam Roben.
1946
1947         Skip pretty patch unit tests if ruby isn't installed.
1948
1949         https://bugs.webkit.org/show_bug.cgi?id=54699
1950
1951         * Scripts/webkitpy/common/prettypatch_unittest.py:
1952
1953 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1954
1955         Reviewed by Adam Roben.
1956
1957         Skip mac.test_skipped_file_paths() when running on win32; the
1958         test has hardcoded '/' directory separators and since it's a
1959         port-specific test and we have coverage on other ports it's not
1960         worth it to make the test more generic.
1961
1962         https://bugs.webkit.org/show_bug.cgi?id=54698
1963
1964         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1965
1966 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1967
1968         Reviewed by Adam Roben.
1969
1970         Fix webkitpy.layout_tests.port.config_unittest to work on Win32.
1971
1972         https://bugs.webkit.org/show_bug.cgi?id=54703
1973
1974         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1975
1976 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1977
1978         Reviewed by Adam Roben.
1979
1980         Change the default port we're mocking out on win32 when we
1981         run mock_drt_unittest. Normally we'd default to the
1982         port-specific default, but that would be "win", which doesn't
1983         currently work.
1984
1985         https://bugs.webkit.org/show_bug.cgi?id=54708
1986
1987         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
1988
1989 2011-02-18  Kenneth Russell  <kbr@google.com>
1990
1991         Reviewed by James Robinson.
1992
1993         [chromium] Clean up initialization of minimum timer interval
1994         https://bugs.webkit.org/show_bug.cgi?id=54772
1995
1996         Use new GetForegroundTabTimerInterval function in webkit_support
1997         to reset the minimum timer interval before each test.
1998
1999         * DumpRenderTree/chromium/WebViewHost.cpp:
2000         (WebViewHost::reset):
2001
2002 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
2003
2004         Reviewed by Adam Roben.
2005
2006         r78494 fixed the chromium_win unittest failing under win32,
2007         but not cygwin. This fixes cygwin as well.
2008
2009         https://bugs.webkit.org/show_bug.cgi?id=54652
2010
2011         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2012
2013 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
2014
2015         Reviewed by Ojan Vafai, Mihai Parparita.
2016
2017         new-run-webkit-tests: r78522 made it impossible to create a
2018         'chromium-win' or 'chromium-mac' port; they would always get the
2019         version tacked on to the end. It turns out that we actually need
2020         to be able to create these ports in order for
2021         rebaseline-chromium-webkit-tests to work correctly given the
2022         broken coverage of our layout bots. As soon as we get SL and
2023         Win 7 bots running reliably on the canaries, we can get rid of
2024         this change. Note that we only need to change the chromium_mac
2025         and win ports because linux doesn't yet use versions and that
2026         script doesn't work with non-chromium ports.
2027
2028         Also, this change fixes a bug where we were using
2029         "platform/chromium-win-win7" and
2030         "platform/chromium-mac-snowleopard" as the baseline dirs for
2031         Win 7 and SL; we should've been using the generic directories
2032         instead since they are the newest versions of the os's.
2033
2034         https://bugs.webkit.org/show_bug.cgi?id=54691
2035
2036         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
2037         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
2038         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2039         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2040         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2041
2042 2011-02-16  MORITA Hajime  <morrita@google.com>
2043
2044         Reviewed by Kent Tamura.
2045
2046         [Chromium][DRT] some PostDelayedTask() accesses possibly dangling this pointer.
2047         http://webkit.org/b/53899
2048
2049         Replaced static invocation wrappers with MethodTask objects,
2050         which are removed on WebViewHost destructor.
2051         
2052         * DumpRenderTree/chromium/WebViewHost.cpp:
2053         (WebViewHost::requestCheckingOfText):
2054         (WebViewHost::scheduleAnimation):
2055         (WebViewHost::closeWidgetSoon):
2056         * DumpRenderTree/chromium/WebViewHost.h:
2057         (WebViewHost::HostMethodTask::HostMethodTask):
2058         (WebViewHost::HostMethodTask::runIfValid):
2059         (WebViewHost::taskList):
2060
2061 2011-02-07  Hayato Ito  <hayato@chromium.org>
2062
2063         Reviewed by Ojan Vafai.
2064
2065         Get rid of code which writes test results from test_type's
2066         compare_output() method.
2067         Writing test results now happens in single_test_runner, using
2068         newly introduced test_result_writer module.
2069
2070         Subsequent patches will eliminate test_type/* classes.
2071
2072         https://bugs.webkit.org/show_bug.cgi?id=53908
2073
2074         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
2075         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py: Added.
2076         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
2077         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
2078         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2079
2080 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
2081
2082         Reviewed by Adam Roben.
2083
2084         Add blacklist of remaining webkitpy/* modules that fail on
2085         Win32, so that test-webkitpy will at least run cleanly.
2086
2087         https://bugs.webkit.org/show_bug.cgi?id=54709
2088
2089         * Scripts/webkitpy/layout_tests/test/main.py:
2090
2091 2011-02-17  Simon Fraser  <simon.fraser@apple.com>
2092
2093         Reviewed by Sam Weinig.
2094
2095         Set a preference to ensure that scrollbars are in a predictable
2096         state for DRT.
2097
2098         * DumpRenderTree/mac/DumpRenderTree.mm:
2099         (resetDefaultsToConsistentValues):
2100
2101 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
2102
2103         Reviewed by Adam Roben.
2104
2105         test-webkitpy doesn't work with the multiprocessing module on win32
2106         or cygwin, so we skip it for now.
2107
2108         https://bugs.webkit.org/show_bug.cgi?id=54520
2109
2110         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
2111         * Scripts/webkitpy/layout_tests/layout_package/run_webkit_tests_unittest.py:
2112
2113 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
2114
2115         Reviewed by Adam Roben.
2116
2117         Fix version, baseline_search_path for chromium-win-win7. Also
2118         fix chromium_win_unittests, and fix port_testcase.make_port()
2119         to correctly use the options passed into it and only set
2120         results_directory() if necessary.
2121
2122         https://bugs.webkit.org/show_bug.cgi?id=54652
2123
2124         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2125         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2126         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
2127
2128 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
2129
2130         Reviewed by Ojan Vafai.
2131
2132         Fix a hang in new-run-webkit-tests when run with
2133         --dry-run --worker-model=processes
2134
2135         https://bugs.webkit.org/show_bug.cgi?id=54595
2136
2137         * Scripts/webkitpy/layout_tests/layout_package/test_runner2.py:
2138         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2139
2140 2011-02-17  Kevin Ollivier  <kevino@theolliviers.com>
2141
2142         [wx] Build fixes after recent changes.
2143
2144         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2145         (LayoutTestController::setMinimumTimerInterval):
2146         * wx/build/settings.py:
2147
2148 2011-02-17  Csaba Osztrogonác  <ossy@webkit.org>
2149
2150         Reviewed by Andreas Kling.
2151
2152         [Qt] ImageDiff sometimes hangs
2153         https://bugs.webkit.org/show_bug.cgi?id=54641
2154
2155         * DumpRenderTree/qt/ImageDiff.cpp: Clear actualImage and baselineImage after all comparison.
2156         (main):
2157
2158 2011-02-17  Peter Varga  <pvarga@webkit.org>
2159
2160         Reviewed by Csaba Osztrogonác.
2161
2162         The run-sunspider --v8-suite doesn't work
2163         https://bugs.webkit.org/show_bug.cgi?id=54660
2164
2165         * Scripts/run-sunspider:
2166
2167 2011-02-17  Gabor Rapcsanyi  <rgabor@webkit.org>
2168
2169         Reviewed by Csaba Osztrogonác.
2170
2171         [NRWT] Fix environment setup on Qt port.
2172         https://bugs.webkit.org/show_bug.cgi?id=54549
2173
2174         * Scripts/webkitpy/layout_tests/port/qt.py:
2175
2176 2011-02-16  Yael Aharon  <yael.aharon@nokia.com>
2177
2178         Reviewed by Andreas Kling.
2179
2180         [Qt] Enable tiled backing store by default in MiniBrowser.
2181         https://bugs.webkit.org/show_bug.cgi?id=54581
2182
2183         Instead of a flag to turn it on, a new flag is defined to turn
2184         tiled backing store off.
2185
2186         * MiniBrowser/qt/MiniBrowserApplication.cpp:
2187         (MiniBrowserApplication::handleUserOptions):
2188         * MiniBrowser/qt/MiniBrowserApplication.h:
2189         (WindowOptions::WindowOptions):
2190
2191 2011-02-16  Eric Seidel  <eric@webkit.org>
2192
2193         Reviewed by Mihai Parparita.
2194
2195         flaky test archive uploads are always 22 bytes long
2196         https://bugs.webkit.org/show_bug.cgi?id=54593
2197
2198         Speculative fix for the 22-bytes bug.
2199
2200         * Scripts/webkitpy/tool/bot/flakytestreporter.py:
2201
2202 2011-02-16  Anna Cavender  <annacc@chromium.org>
2203
2204         Reviewed by Mihai Parparita.
2205
2206         http/tests/media should be served over HTTP (not from local file)
2207         https://bugs.webkit.org/show_bug.cgi?id=54028
2208
2209         Removed special file:// treatment of http/tests/media
2210
2211         * Scripts/old-run-webkit-tests:
2212         * Scripts/webkitpy/layout_tests/port/base.py:
2213
2214 2011-02-16  Mihai Parparita  <mihaip@chromium.org>
2215
2216         Reviewed by Ojan Vafai.
2217
2218         ChangeLog.parse_latest_entry_from_file does not handle rolled over ChangeLogs
2219         https://bugs.webkit.org/show_bug.cgi?id=54609
2220         
2221         r78737 rolled over ChangeLogs, and webkit-patch complains when landing
2222         patches such as this one that are the first to add an entry to the
2223         ChangeLog (since they don't see a second date line to know where the
2224         entry ends).
2225
2226         * Scripts/webkitpy/common/checkout/changelog.py:
2227         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
2228
2229 == Rolled over to ChangeLog-2011-02-16 ==