1960f6972b4018aed67f79841cd72ba1c0714fd4
[WebKit-https.git] / Tools / ChangeLog
1 2011-03-03  Carlos Garcia Campos  <cgarcia@igalia.com>
2
3         Reviewed by Martin Robinson.
4
5         [GTK] DRT needs implementation of EventSender.scheduleAsynchronousClick
6         https://bugs.webkit.org/show_bug.cgi?id=53960
7
8         * DumpRenderTree/gtk/EventSender.cpp:
9         (sendClick):
10         (scheduleAsynchronousClickCallback):
11
12 2011-03-03  Adam Roben  <aroben@apple.com>
13
14         Look for Windows-style paths when checking whether a test is an Inspector test
15
16         WTR part of <http://webkit.org/b/55672> <rdar://problem/9080867> All inspector tests time
17         out or crash on Windows 7 Release (WebKit2 Tests)
18
19         Reviewed by Darin Adler.
20
21         * WebKitTestRunner/TestInvocation.cpp:
22         (WTR::shouldOpenWebInspector): Also look for backslashes, since that's what we'll get on
23         Windows.
24
25 2011-03-03  Adam Roben  <aroben@apple.com>
26
27         Make Windows 7 Release (Tests) a core builder
28
29         We can probably make Windows XP Debug (Tests) a core builder, too, but we should probably
30         give it a few more days of greenness first.
31
32         Fixes <http://webkit.org/b/55665> Windows 7 Release (Tests) should be a core builder
33
34         Reviewed by Andreas Kling.
35
36         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
37         (BuildBot): Added a regexp to match Windows 7 Release (Tests).
38
39         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
40         (BuildBotTest): Updated the list of builders to match the current list on build.webkit.org.
41         Updated the core regexps to match the buildbot module. Updated the expected results to match
42         the current set of core builders.
43
44 2011-03-03  Benjamin Poulain  <benjamin.poulain@nokia.com>
45
46         Reviewed by Andreas Kling.
47
48         Adding myself as a reviewer.
49
50         * Scripts/webkitpy/common/config/committers.py:
51
52 2011-03-03  Lukasz Slachciak  <l.slachciak@samsung.com>
53
54         Reviewed by Xan Lopez.
55
56         [GTK] Extended application cache database API and added unit tests file.
57         https://bugs.webkit.org/show_bug.cgi?id=55335
58
59         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp: removed external declaration
60         of webkit_application_cache_set_maximum_size due to API exposal.
61
62 2011-03-02  Ojan Vafai  <ojan@chromium.org>
63
64         Reviewed by Adam Barth.
65
66         make webkit-patch upload work for security bugs
67         https://bugs.webkit.org/show_bug.cgi?id=55655
68
69         The problem is that we shell out to prepare-ChangeLog,
70         which does not authenticate with bugzilla. All it needs from
71         bugzilla is the bug title though. We just pass it through from
72         webkit-patch instead.
73
74         * Scripts/prepare-ChangeLog:
75         * Scripts/webkitpy/common/net/bugzilla/bug.py:
76         * Scripts/webkitpy/tool/steps/preparechangelog.py:
77
78 2011-03-01  Ojan Vafai  <ojan@chromium.org>
79
80         Reviewed by Tony Chang.
81
82         [test-dashboard] make gtest JSON match layout-tests JSON
83         https://bugs.webkit.org/show_bug.cgi?id=55556
84
85         http://trac.webkit.org/changeset/80090 stopped converting time_ms to t.
86         Make this code match.
87
88         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
89
90 2011-03-02  Dirk Pranke  <dpranke@chromium.org>
91
92         Reviewed by Tony Chang.
93
94         NRWT - implement Linux Hardy 64-bit port support. This involves
95         adding support for non-'x86' architectures and updating the
96         "deduplicate_tests" script to correctly understand that
97         port.name() is not always equal to basename(port.baseline_path()).
98
99         This change introduces two new port names: 'chromium-linux-x86'
100         and 'chromium-linux-x86_64'. Using the prior 'chromium-linux'
101         flag will cause the code to determine which one to use at
102         runtime, just like we currently do with 'chromium-win'.
103
104         https://bugs.webkit.org/show_bug.cgi?id=55535
105
106         * Scripts/webkitpy/common/system/executive_mock.py:
107           Handle 'return_stderr' keyword arg to run_command()
108         * Scripts/webkitpy/layout_tests/deduplicate_tests.py:
109         * Scripts/webkitpy/layout_tests/deduplicate_tests_unittest.py:
110         * Scripts/webkitpy/layout_tests/port/base.py:
111         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
112         * Scripts/webkitpy/layout_tests/port/chromium_linux_unittest.py:
113         * Scripts/webkitpy/layout_tests/port/factory.py:
114         * Scripts/webkitpy/layout_tests/port/google_chrome.py:
115
116 2011-03-02  Kevin Ollivier  <kevino@theolliviers.com>
117
118         [wx] Build fix, make sure we put the wxWebKit dylib in the right directory and
119         update symlinks accordingly.
120
121         * wx/packaging/build-mac-installer.py:
122
123 2011-03-02  Sheriff Bot  <webkit.review.bot@gmail.com>
124
125         Unreviewed, rolling out r80139.
126         http://trac.webkit.org/changeset/80139
127         https://bugs.webkit.org/show_bug.cgi?id=55620
128
129         failing and crashing tests on Snow Leopard bot (Requested by
130         mihaip1 on #webkit).
131
132         * DumpRenderTree/LayoutTestController.cpp:
133         (LayoutTestController::staticFunctions):
134         * DumpRenderTree/LayoutTestController.h:
135         * DumpRenderTree/chromium/LayoutTestController.cpp:
136         (LayoutTestController::LayoutTestController):
137         * DumpRenderTree/chromium/LayoutTestController.h:
138         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
139         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
140         * DumpRenderTree/mac/UIDelegate.h:
141         * DumpRenderTree/mac/UIDelegate.mm:
142         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
143         * DumpRenderTree/qt/LayoutTestControllerQt.h:
144         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
145         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
146
147 2011-03-02  Adam Roben  <aroben@apple.com>
148
149         Wait for force a paint in WebKitTestRunner until we're actually ready to dump the output
150
151         This matches DumpRenderTree on Windows.
152
153         Fixes <http://webkit.org/b/55469> <rdar://problem/9068539> REGRESSION (r79863): Lots of
154         dom/html/level2/html/HTMLFrameElement*.html tests crashing in FrameView::paintOverhangAreas
155         in WebKit2
156
157         Reviewed by Beth Dakin.
158
159         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
160         (WTR::InjectedBundlePage::dump): Moved code to force a paint here...
161         (WTR::InjectedBundlePage::didFinishLoadForFrame): ...from here.
162
163 2011-03-02  Daniel Cheng  <dcheng@chromium.org>
164
165         Reviewed by David Levin.
166
167         Add feature define for data transfer items
168         https://bugs.webkit.org/show_bug.cgi?id=55510
169
170         * Scripts/build-webkit:
171
172 2011-03-02  Adam Roben  <aroben@apple.com>
173
174         Look in the new-to-SnowLeopard location for crash logs on SnowLeopard and newer
175
176         ReportCrash saves logs to ~/Library/Logs/CrashReporter on Leopard, but to
177         ~/Library/Logs/DiagnosticReports on SnowLeopard. old-run-webkit-tests was only looking in
178         the former location, but was getting lucky on some SnowLeopard machines because symlinks
179         were being created there pointing into DiagnosticReports. For machines without these
180         symlinks, crash logs were not getting captured.
181
182         Fixes <http://webkit.org/b/55607> old-run-webkit-tests doesn't capture crash logs on some
183         SnowLeopard machines
184
185         Reviewed by Eric Seidel.
186
187         * Scripts/old-run-webkit-tests:
188         (captureSavedCrashLog): Look in ~/Library/Logs/DiagnosticReports on SnowLeopard and newer.
189
190 2011-03-02  Adam Roben  <aroben@apple.com>
191
192         Don't try to capture crash logs on platforms that haven't implemented that feature
193
194         Fixes <http://webkit.org/b/55504> Use of uninitialized value in numeric lt (<) at
195         Tools/Scripts/old-run-webkit-tests line 1778 seen on Qt bots after a crash
196
197         Reviewed by Joseph Pecoraro.
198
199         * Scripts/old-run-webkit-tests:
200         (captureSavedCrashLog): Bail if we don't have a crash log glob.
201         (findNewestFileMatchingGlob): Fix a buggy test for when no paths match the glob. This isn't
202         strictly required to fix bug 55504, but seems worthwhile.
203
204 2011-01-26  John Knottenbelt  <jknotten@chromium.org>
205
206         Reviewed by Dmitry Titov.
207
208         Detach Geolocation from Frame when Page destroyed.
209         https://bugs.webkit.org/show_bug.cgi?id=52877
210
211         Extend the layout test controller to expose the number of pending
212         geolocation requests, so that we can test that the requests have
213         been cancelled on page close.
214
215         * DumpRenderTree/LayoutTestController.cpp:
216         (numberOfPendingGeolocationPermissionRequestsCallback):
217         (LayoutTestController::staticFunctions):
218         * DumpRenderTree/LayoutTestController.h:
219         * DumpRenderTree/chromium/LayoutTestController.cpp:
220         (LayoutTestController::LayoutTestController):
221         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
222         * DumpRenderTree/chromium/LayoutTestController.h:
223         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
224         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
225         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
226         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
227         * DumpRenderTree/mac/UIDelegate.h:
228         * DumpRenderTree/mac/UIDelegate.mm:
229         (-[UIDelegate numberOfPendingGeolocationPermissionRequests]):
230         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
231         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
232         * DumpRenderTree/qt/LayoutTestControllerQt.h:
233         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
234         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
235         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
236         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
237
238
239 2011-03-01  Kent Tamura  <tkent@chromium.org>
240
241         Reviewed by Dimitri Glazkov.
242
243         Assertion fails when a form validation bubble appears
244         https://bugs.webkit.org/show_bug.cgi?id=55550
245
246         Test: fast/forms/interactive-validation-attach-assertion.html
247
248         * DumpRenderTree/chromium/WebPreferences.cpp:
249         (WebPreferences::applyTo): Enable the interactive validation feature by
250           default. This matches to Chromium browser.
251
252 2011-03-01  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
253
254         Reviewed by Kent Tamura.
255
256         Remove efl port style exception of Webcore/JavaScriptCore
257         https://bugs.webkit.org/show_bug.cgi?id=55549
258
259         WebCore and JavaScriptCore need to adhere WebKit coding style.
260
261         * Scripts/webkitpy/style/checker.py:
262
263 2011-03-01  Tony Chang  <tony@chromium.org>
264
265         Reviewed by Ojan Vafai.
266
267         [chromium] clean up temp files on chromium-mac builder too
268         https://bugs.webkit.org/show_bug.cgi?id=55525
269
270         * BuildSlaveSupport/chromium/remove-crash-logs:
271
272 2011-02-28  Ojan Vafai  <ojan@chromium.org>
273
274         Reviewed by Tony Chang.
275
276         wrap json in a function call to afford cross-domain loading
277         https://bugs.webkit.org/show_bug.cgi?id=55353
278
279         Also delete the code that compacts the JSON. It turns out this
280         breaks the rebaseline tool and is probably a premature optimization anyways.
281
282         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
283         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
284         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
285         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
286
287 2011-03-01  Sheriff Bot  <webkit.review.bot@gmail.com>
288
289         Unreviewed, rolling out r80079.
290         http://trac.webkit.org/changeset/80079
291         https://bugs.webkit.org/show_bug.cgi?id=55547
292
293         "Broke the Win debug build?" (Requested by dcheng on #webkit).
294
295         * Scripts/build-webkit:
296
297 2011-03-01  Mark Rowe  <mrowe@apple.com>
298
299         Reviewed by Sam Weinig.
300
301         Update verification scripts to handle WebKit2.
302
303         * Scripts/check-for-global-initializers: Skip object files that use RefCountedLeakChecker.
304         * Scripts/check-for-inappropriate-files-in-framework: Add some inappropriate file types for
305         WebKit2.
306
307 2011-03-01  Daniel Cheng  <dcheng@chromium.org>
308
309         Reviewed by David Levin.
310
311         Add feature define for data transfer items
312         https://bugs.webkit.org/show_bug.cgi?id=55510
313
314         * Scripts/build-webkit:
315
316 2011-02-23  Joseph Pecoraro  <joepeck@webkit.org>
317
318         Reviewed by Timothy Hatcher.
319
320         All Console Messages should be passed to ChromeClients.
321         https://bugs.webkit.org/show_bug.cgi?id=54926
322
323         Update DumpRenderTree's UIDelegate to use the new console message delegate.
324
325         * DumpRenderTree/mac/UIDelegate.mm:
326         (-[UIDelegate webView:addMessageToConsole:withSource:]):
327
328 2011-03-01  Dirk Pranke  <dpranke@chromium.org>
329
330         Reviewed by Adam Barth.
331
332         Tragically, we stopped running the python unit tests
333         on Chromium Win just as I was getting them to be green.
334         This patch turns them back on.
335
336         https://bugs.webkit.org/show_bug.cgi?id=55521
337
338         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
339
340 2011-03-01  Dirk Pranke  <dpranke@chromium.org>
341
342         Reviewed by Adam Roben.
343
344         Disable
345         webkitpy.common.system.executive_unittest.ExecutiveTest.\
346            test_kill_all fails on win until it can be properly triaged.
347
348         https://bugs.webkit.org/show_bug.cgi?id=54790
349
350         * Scripts/webkitpy/common/system/executive_unittest.py:
351
352     2011-03-01  Helder Correia  <helder@sencha.com>
353
354             Unreviewed. Add myself to committers.py.
355
356             * Scripts/webkitpy/common/config/committers.py:
357
358 2011-03-01  Adam Roben  <aroben@apple.com>
359
360         Give up if a crash log for the web process is taking too long to be saved on Windows
361
362         This should keep the bots from getting stuck while waiting for a crash log, if writing a
363         crash log ever hangs or the UI process for some reason doesn't notice it's finished.
364
365         Fixes <http://webkit.org/b/55499> WebKitTestRunner can hang forever waiting for a crash log
366         to be saved for the web process
367
368         Reviewed by Steve Falkenburg.
369
370         * WebKitTestRunner/win/TestControllerWin.cpp:
371         (WTR::runRunLoopUntil): Moved code to run the run loop here here from platformRunUntil.
372         Generalized the code slightly to handle the optional object to wait on.
373         (WTR::TestController::platformRunUntil): Use the new runRunLoopUntil function, and also use
374         it (with a timeout) when waiting for a crash log for the web process to be saved. This will
375         prevent us from waiting forever.
376
377 2011-03-01  Adam Roben  <aroben@apple.com>
378
379         Give WebKitLauncherWin's PDB file a name different from WebKit's
380
381         WebKitLauncherWin (which builds WebKit.exe) and WebKit (which builds WebKit.dll) were
382         fighting over the same PDB file, meaning that one of them was going to lose and not have
383         symbols. WebKitLauncherWin's PDB file is now named WebKitLauncherWin.pdb.
384
385         Fixes <http://webkit.org/b/55495> Crashes in WebKit.dll have bad backtraces for some test
386         runs (due to missing symbols)
387
388         Reviewed by David Kilzer.
389
390         * WebKitLauncherWin/WebKitLauncherWinCommon.vsprops: Use the project name,
391         "WebKitLauncherWin", instead of the target name, "WebKit", when naming the PDB file.
392
393 2011-03-01  Adam Roben  <aroben@apple.com>
394
395         Count web process crashes toward the --exit-after-n-crashes-or-timeouts limit
396
397         Fixes <http://webkit.org/b/55492> Web process crashes aren't counted toward the
398         --exit-after-n-crashes-or-timeouts limit
399
400         Reviewed by David Kilzer.
401
402         * Scripts/old-run-webkit-tests:
403         (stopRunningTestsEarlyIfNeeded): Include web process crashes in the computation, and print
404         the number of tests that caused the web process to crash when exiting early.
405
406 2011-03-01  Adam Roben  <aroben@apple.com>
407
408         Save a crash log when the web process crashes
409
410         On Windows, WebKitTestRunner now detects when the web process is crashing and waits to exit
411         until it has finished crashing, which guarantees that the crash log will have had time to be
412         saved, too. On Mac, we always wait until ReportCrash has exited before capturing the crash
413         log, so all we have to do is choose the right crash log out of the CrashReporter directory.
414
415         Fixes <http://webkit.org/b/44121> <rdar://problem/8320759> When the web process crashes and
416         a crash log is being saved, WebKitTestRunner thinks the web process has become unresponsive
417
418         Reviewed by Sam Weinig.
419
420         * Scripts/old-run-webkit-tests:
421         (testCrashedOrTimedOut): Don't kill WebKitTestRunner when the web process crashes. It will
422         kill itself. On Windows, this will cause us to wait until the crash log has been saved. On
423         Mac, it should have no effect. Capture saved crash logs for web process crashes, too.
424         (captureSavedCrashLog): Added $webProcessCrashed argument. On Mac, look for
425         WebProces_*.crash files when the web process crashes.
426
427         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
428         (WTR::InjectedBundle::initialize): Added an initializationUserData argument. Updated for
429         initializePlatformDefaults -> platformInitialize rename. Pass the initializationUserData
430         along to platformInitialize.
431
432         * WebKitTestRunner/InjectedBundle/InjectedBundle.h: See above.
433
434         * WebKitTestRunner/InjectedBundle/InjectedBundleMain.cpp:
435         (WKBundleInitialize): Pass along the initializationUserData to the InjectedBundle.
436
437         * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm:
438         (WTR::InjectedBundle::platformInitialize):
439         * WebKitTestRunner/InjectedBundle/qt/InjectedBundleQt.cpp:
440         (WTR::InjectedBundle::platformInitialize):
441         Updated function signature.
442
443         * WebKitTestRunner/InjectedBundle/win/InjectedBundleWin.cpp:
444         (WTR::exceptionFilter): Added. Tells the UI process we're crashing by signaling the
445         webProcessCrashingEvent, then lets the crash continue as normal.
446
447         (WTR::InjectedBundle::platformInitialize): Hook up exceptionFilter. Retrieve the name of the
448         event we should use to tell the UI process we're crashing from the initializationUserData,
449         and get a handle to that event.
450
451         * WebKitTestRunner/TestController.cpp:
452         (WTR::TestController::TestController): Initialize new members.
453         (WTR::TestController::processDidCrash): Removed unnecessary WKPageRef argument. Changed to
454         only print the "#CRASHED - WebProcess" message once, since this can be called more than once
455         when a crash log is being saved on Windows. Exit right away if specified. (This is the
456         default.)
457
458         * WebKitTestRunner/TestController.h: Added new members.
459
460         * WebKitTestRunner/win/TestControllerWin.cpp:
461         (WTR::TestController::platformInitialize): Set up the event the web process will use to tell
462         us it's crashing.
463         (WTR::TestController::platformRunUntil): Pass MWMO_INPUTAVAILABLE to
464         ::MsgWaitForMultipleObjectsEx so we'll process messages that have already been seen by
465         ::PeekMessage. (This is unrelated to the bug fix.) Notice when the webProcessCrashingEvent
466         has been signaled. When this happens, print the "#CRASHED - WebProcess" message right away
467         so the test harness will know the web process has crashed and not try to kill us, then wait
468         for the web process to finish crashing so a crash log will have time to be saved.
469         (WTR::toWK): Simple hepler function.
470         (WTR::TestController::platformInitializeContext): Pass along the name of the event the web
471         process should use to tell us it is crashing in the context's initialization user data.
472
473 2011-03-01  Dimitri Glazkov  <dglazkov@chromium.org>
474
475         Reviewed by Tony Gentilcore.
476
477         Move the checks to exclude Python/Perl tests for Chromium Win build to the right place.
478         https://bugs.webkit.org/show_bug.cgi?id=55476
479
480         Chromium Win uses BuildAndTestFactory, not TestFactory.
481
482         * BuildSlaveSupport/build.webkit.org-config/master.cfg: Moved to the right place.
483
484 2011-03-01  Adam Roben  <aroben@apple.com>
485
486         Test that the WebKit2 responsiveness timer doesn't fire too early
487
488         Test for <http://webkit.org/b/55417> <rdar://problem/9065287> RunLoop::Timer fires 1000x too
489         early on Windows
490
491         Reviewed by Anders Carlsson.
492
493         * TestWebKitAPI/PlatformUtilities.h: Added sleep().
494
495         * TestWebKitAPI/Tests/WebKit2/ResponsivenessTimerDoesntFireEarly.cpp: Added.
496         (TestWebKitAPI::didReceiveMessageFromInjectedBundle):
497         (TestWebKitAPI::didFinishLoadForFrame):
498         (TestWebKitAPI::processDidBecomeUnresponsive):
499         (TestWebKitAPI::setInjectedBundleClient):
500         (TestWebKitAPI::setPageLoaderClient):
501         Simple helper functions.
502
503         (TestWebKitAPI::TEST): Load an HTML file to make sure the web process is initialized. Then
504         tell the web process to pause and press the spacebar key. The spacebar keypress should cause
505         the responsiveness timer to start, but the web process should unpause before it has a chance
506         to fire. Run until the web process has unpaused, and assert that the timer didn't fire.
507
508         * TestWebKitAPI/Tests/WebKit2/ResponsivenessTimerDoesntFireEarly_Bundle.cpp: Added.
509         (TestWebKitAPI::ResponsivenessTimerDoesntFireEarlyTest::ResponsivenessTimerDoesntFireEarlyTest):
510         Call up to the base class.
511         (TestWebKitAPI::ResponsivenessTimerDoesntFireEarlyTest::didReceiveMessage): When asked to
512         pause, sleep for 0.5 seconds, then send back a message saying we paused.
513
514         * TestWebKitAPI/mac/PlatformUtilitiesMac.mm:
515         (TestWebKitAPI::Util::sleep): Added. Calls through to usleep.
516         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp:
517         (TestWebKitAPI::Util::sleep): Added. Calls through to ::Sleep.
518
519         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
520         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
521         * TestWebKitAPI/win/TestWebKitAPIInjectedBundle.vcproj:
522         Added new files.
523
524 2011-03-01  Andras Becsi  <abecsi@webkit.org>
525
526         Reviewed by Csaba Osztrogonác.
527
528         [Qt] Clean up the project files and move common options to WebKit.pri.
529
530         * WebKitTestRunner/qt/WebKitTestRunner.pro: Deduplicate.
531
532
533 2011-03-01  Adam Roben  <aroben@apple.com>
534
535         Skip another multiprocessing test on Windows
536
537         It is sometimes failing (and may be leaving child Python processes around).
538         <http://webkit.org/b/55087> tracks the failure.
539
540         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
541         (FuncitonTests.test_get__processes): Skip this test on Windows.
542
543 2011-03-01  Ilya Tikhonovsky  <loislo@chromium.org>
544
545         Reviewed by Pavel Feldman.
546
547         Web Inspector: reduce inspector tests flakiness.
548         https://bugs.webkit.org/show_bug.cgi?id=55401
549
550         * DumpRenderTree/chromium/TestShell.cpp:
551         (TestShell::runFileTest):
552         * DumpRenderTree/chromium/WebPreferences.cpp:
553         (WebPreferences::reset):
554
555 2011-02-28  Mihai Parparita  <mihaip@chromium.org>
556
557         Reviewed by James Robinson.
558
559         NRWT should still upload results when it exits early due to too many failures
560         https://bugs.webkit.org/show_bug.cgi?id=55432
561         
562         Only keyboard interrupts should prevent results from being uploaded.
563
564         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
565         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
566
567 2011-02-28  Alice Liu  <alice.liu@apple.com>
568
569         Reviewed by Adam Roben and Gavin Barraclough.
570
571         Test for https://bugs.webkit.org/show_bug.cgi?id=54898
572         InjectedBundleNodeHandle dies too early in WKBundleHitTestResultGetNodeHandle.
573
574         * TestWebKitAPI/InjectedBundleController.h:
575         (TestWebKitAPI::InjectedBundleController::bundle):
576         * TestWebKitAPI/PlatformWebView.h:
577         * TestWebKitAPI/Tests/WebKit2/HitTestResultNodeHandle.cpp: Added.
578         (TestWebKitAPI::didFinishLoadForFrame):
579         (TestWebKitAPI::didReceiveMessageFromInjectedBundle):
580         (TestWebKitAPI::setPageLoaderClient):
581         (TestWebKitAPI::TEST):
582         * TestWebKitAPI/Tests/WebKit2/HitTestResultNodeHandle_Bundle.cpp: Added.
583         (TestWebKitAPI::HitTestResultNodeHandleTest::HitTestResultNodeHandleTest):
584         (TestWebKitAPI::HitTestResultNodeHandleTest::getContextMenuFromDefaultMenu):
585         (TestWebKitAPI::HitTestResultNodeHandleTest::didCreatePage):
586         * TestWebKitAPI/mac/PlatformWebViewMac.mm:
587         (TestWebKitAPI::PlatformWebView::simulateRightClick):
588         * TestWebKitAPI/win/PlatformWebViewWin.cpp:
589         (TestWebKitAPI::PlatformWebView::simulateRightClick):
590         
591         Added files to:
592         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
593         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
594         * TestWebKitAPI/win/TestWebKitAPIInjectedBundle.vcproj:
595
596 2011-02-28  Adam Roben  <aroben@apple.com>
597
598         Start running InjectedBundle API tests on Windows
599
600         A new project, TestWebKitAPIInjectedBundle, builds the injected bundle DLL.
601
602         Fixes <http://webkit.org/b/55420> run-api-tests doesn't run injected bundle tests on Windows
603
604         Reviewed by Darin Adler.
605
606         * TestWebKitAPI/Configurations/TestWebKitAPIDebug.vsprops: Added.
607         * TestWebKitAPI/Configurations/TestWebKitAPIDebugAll.vsprops: Added.
608         * TestWebKitAPI/Configurations/TestWebKitAPIDebugCairoCFLite.vsprops: Added.
609         * TestWebKitAPI/Configurations/TestWebKitAPIRelease.vsprops: Added.
610         * TestWebKitAPI/Configurations/TestWebKitAPIReleaseCairoCFLite.vsprops: Added.
611         These are just extracted from TestWebKitAPI.vcproj.
612
613         * TestWebKitAPI/Configurations/TestWebKitAPIInjectedBundleCommon.vsprops: Added. Just sets
614         up the DLL's filename.
615
616         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp: Fixed DLL name.
617         (TestWebKitAPI::Util::createInjectedBundlePath): Changed to use kCFAllocatorNull instead of
618         0 (== kCFAllocatorDefault) so that CF won't try to deallocate our C string.
619
620         * TestWebKitAPI/win/TestWebKitAPI.sln: Added TestWebKitAPIInjectedBundle. Removed the bogus
621         "all" configuration.
622
623         * TestWebKitAPI/win/TestWebKitAPI.vcproj: Changed to use the new .vsprops files. Added
624         injected bundle tests. Let VS resort things.
625
626         * TestWebKitAPI/win/TestWebKitAPIInjectedBundle.vcproj: Copied from Tools/TestWebKitAPI/win/TestWebKitAPI.vcproj.
627
628 2011-02-28  Adam Roben  <aroben@apple.com>
629
630         Add a test for the paint rect passed via WM_PAINT to windowless plugins
631
632         Test for <http://webkit.org/b/55365> <rdar://problem/9031089> REGRESSION (r79040): WebKit2:
633         Large portions of pages with plugins paint black
634
635         Reviewed by Anders Carlsson.
636
637         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
638         (PluginTest::NPP_HandleEvent):
639         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
640         Added NPP_HandleEvent. Just returns 0 at this level.
641
642         * DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowlessPaintRectCoordinates.cpp: Added.
643         (WindowlessPaintRectCoordinates::WindowlessPaintRectCoordinates): Call up to the base class.
644         (WindowlessPaintRectCoordinates::NPP_New): Mark ourselves as windowless.
645         (WindowlessPaintRectCoordinates::NPP_HandleEvent): Check that the paint rect passed via
646         WM_PAINT has the expected coordinates.
647
648         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
649         (NPP_HandleEvent): Give the PluginTest a chance to handle the event before doing anything
650         else.
651
652         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Added new file, let VS
653         resort things.
654
655         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
656         (WTR::InjectedBundlePage::didFinishLoadForFrame): Added a call to WKBundlePageForceRepaint.
657         This is roughly equivalent to the call to -[WebView displayIfNeeded] in -[FrameLoadDelegate
658         webView:didFinishLoadForFrame:] in DumpRenderTree.
659
660         * WebKitTestRunner/win/PlatformWebViewWin.cpp:
661         (WTR::PlatformWebView::PlatformWebView): Tell the WKView it's in a window so that plugins
662         can start running.
663
664 2011-02-28  Adam Roben  <aroben@apple.com>
665
666         Tell MiniBrowser's WKViews that they're in a window
667
668         Fixes <http://webkit.org/b/55364> Plugins don't work in MiniBrowser on Windows
669
670         Reviewed by Anders Carlsson.
671
672         * MiniBrowser/win/BrowserView.cpp:
673         (BrowserView::create): Call WKViewSetIsInWindow after creating the view.
674
675 2011-02-28  Alexis Menard  <alexis.menard@openbossa.org>
676
677         Reviewed by Csaba Osztrogonác.
678
679         The new QML tests require QML_IMPORT_PATH variable to be set in order to
680         load the proper plugin.
681
682         * Scripts/run-qtwebkit-tests:
683
684 2011-02-28  Joone Hur  <joone.hur@collabora.co.uk>
685
686         Unreviewed, fix spelling mistake.
687
688         hardware-accelated => hardware-accelerated.
689
690         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
691
692 2011-02-28  Kristian Amlie  <kristian.amlie@nokia.com>
693
694         Reviewed by Andreas Kling.
695
696         Made sure that the build-webkit --qmake argument is always respected.
697
698         Previously, it would be respected during the build, but not during
699         the early feature detection. The build would also fail if qmake was
700         not in your path, even if you specified --qmake with a valid qmake
701         binary.
702
703         [Qt] WebKit patches required to work with a modularized version of Qt
704         https://bugs.webkit.org/show_bug.cgi?id=53916
705
706         * Scripts/build-webkit:
707         * Scripts/webkitdirs.pm:
708
709 2011-02-27  Ojan Vafai  <ojan@chromium.org>
710
711         Fix failing tests from http://trac.webkit.org/changeset/79837.
712
713         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
714         This was just a mistake that would break if someone tried printing with this configuration.
715         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
716         Method signature changed. Didn't notice that this method was being calling in a unittest.
717
718 2011-02-24  Hayato Ito  <hayato@chromium.org>
719
720         Reviewed by Eric Seidel.
721
722         [NRWT] Remove test_types/* classes.
723
724         This is the last patch in the series of efforts which tried to remove test_types classes.
725         A single_test_runner module will take over responsibilities of test_type classes.
726
727         From this patch, we won't display statistics per test_type even if a '--verbose' option is given.
728
729         https://bugs.webkit.org/show_bug.cgi?id=55123
730
731         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
732         * Scripts/webkitpy/layout_tests/layout_package/test_results.py:
733         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:
734         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
735         * Scripts/webkitpy/layout_tests/layout_package/worker_mixin.py:
736         * Scripts/webkitpy/layout_tests/test_types/__init__.py: Removed.
737         * Scripts/webkitpy/layout_tests/test_types/image_diff.py: Removed.
738         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py: Removed.
739         * Scripts/webkitpy/layout_tests/test_types/test_type_base_unittest.py: Removed.
740         * Scripts/webkitpy/layout_tests/test_types/text_diff.py: Removed.
741
742 2011-02-25  Ojan Vafai  <ojan@chromium.org>
743
744         Reviewed by Tony Chang.
745
746         Change results.json format to the one used by unexpected_results.json
747         https://bugs.webkit.org/show_bug.cgi?id=52267
748
749         Also add runtimes in milliseconds to the JSON and make the output format more compact.
750         Named the file full_results.json to avoid conflicting with the results.json
751         file the test-results server currently serves up.
752
753         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
754         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
755         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
756         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
757         * Scripts/webkitpy/layout_tests/layout_package/result_summary.py:
758         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
759
760 2011-02-27  Adam Roben  <aroben@apple.com>
761
762         Followup fix to r79827
763
764         * DumpRenderTree/win/UIDelegate.cpp:
765         (UIDelegate::webViewAddMessageToConsole): Replaced call to urlSuitableForTestResult with
766         lastPathComponent. This matches what we did here before r79827, and matches Mac.
767
768 2011-02-27  Dan Bernstein  <mitz@apple.com>
769
770         Reviewed by Adam Roben.
771
772         Make bisect-builds print a trac link for the regression range
773         https://bugs.webkit.org/show_bug.cgi?id=55331
774
775         * Scripts/bisect-builds:
776
777 2011-02-27  Adam Roben  <aroben@apple.com>
778
779         Use iswalpha instead of isalpha when dealing with wchar_ts in EditingDelegate
780
781         Fixes <http://webkit.org/b/55062> <rdar://problem/9059907> Crash beneath
782         EditingDelegate::checkSpellingOfString when running fast/forms/input-text-maxlength.html or
783         fast/forms/input-text-paste-maxlength.html on Windows with full page heap enabled
784
785         Reviewed by Anders Carlsson.
786
787         * DumpRenderTree/win/EditingDelegate.cpp:
788         (indexOfFirstWordCharacter):
789         (wordLength):
790         Use iswalpha instead of isalpha, since TCHAR is really just wchar_t.
791
792 2011-02-27  Adam Roben  <aroben@apple.com>
793
794         Make ResourceLoadDelegate print URLs relative to the main resource on Windows
795
796         This matches Mac. (In fact, the code was ported from Mac!)
797
798         Fixes <http://webkit.org/b/55328> URLs printed by ResourceLoadDelegate on Windows don't
799         match Mac
800
801         Reviewed by Anders Carlsson.
802
803         * DumpRenderTree/win/DumpRenderTree.cpp:
804         (substringFromIndex): Added. Emulates -[NSString substringFromIndex:].
805         (urlSuitableForTestResult): Ported code from -[NSURL(DRTExtras)
806         _drt_descriptionSuitableForTestResult].
807         (cfStringRefToWString): Moved here from LayoutTestControllerWin.cpp.
808
809         * DumpRenderTree/win/DumpRenderTreeWin.h: Added declaration of cfStringRefToWString.
810
811         * DumpRenderTree/win/LayoutTestControllerWin.cpp: Moved cfStringRefToWString to
812         DumpRenderTree.cpp.
813
814 2011-02-27  Adam Roben  <aroben@apple.com>
815
816         Change the order of Chang Shu's email addresses so the first one is their Bugzilla username
817
818         This makes Bugzilla autocompletion work correctly.
819
820         * Scripts/webkitpy/common/config/committers.py:
821
822 2011-02-27  Patrick Gansterer  <paroga@webkit.org>
823
824         Reviewed by Andreas Kling.
825
826         [Qt] fast/loader/user-stylesheet-fast-path.html fails
827         https://bugs.webkit.org/show_bug.cgi?id=50911
828
829         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
830         (LayoutTestController::setUserStyleSheetLocation):
831         Use QUrl::fromEncoded since the passed string is already encoded.
832
833 2011-02-27  Adam Roben  <aroben@apple.com>
834
835         Ensure $testResultsDirectory is an absolute path before setting up Windows crash log saving
836
837         Windows 7 (and maybe Vista?) launches the post-mortem debugger with a working directory that
838         is different from old-run-webkit-test's. Because of this, the path to the crash log file
839         (which is relative to $testResultsDirectory) must be an absolute path. See the bug for more
840         details.
841
842         Fixes <http://webkit.org/b/55318> Crash logs aren't saved on Windows 7 when a relative
843         --results-directory path is used (like on the test slaves)
844
845         Reviewed by Dan Bernstein.
846
847         * Scripts/old-run-webkit-tests: Moved the call to setUpWindowsCrashLogSaving after we've
848         made $testResultsDirectory an absolute path.
849
850 2011-02-26  Dominic Mazzoni  <dmazzoni@google.com>
851
852         Reviewed by Adam Barth.
853
854         m_dumpAccessibilityNotifications should be initialized.
855         https://bugs.webkit.org/show_bug.cgi?id=54874
856
857         * DumpRenderTree/chromium/AccessibilityController.cpp:
858         (AccessibilityController::AccessibilityController):
859
860 2011-02-26  Vsevolod Vlasov  <vsevik@chromium.org>
861
862         Reviewed by Pavel Feldman.
863
864         DumpRenderTree should reset frame opener between tests.
865         https://bugs.webkit.org/show_bug.cgi?id=54874
866
867         Added clearOpener() calls between tests.
868         Some tests are using window.open() to change current window location.
869         This changes window.opener property whose nullity is checked before
870         showing xml viewer.
871
872         * DumpRenderTree/chromium/TestShell.cpp:
873         (TestShell::resetTestController):
874         * DumpRenderTree/gtk/DumpRenderTree.cpp:
875         (resetDefaultsToConsistentValues):
876         * DumpRenderTree/mac/DumpRenderTree.mm:
877         (resetWebViewToConsistentStateBeforeTesting):
878         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
879         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
880         * DumpRenderTree/win/DumpRenderTree.cpp:
881         (resetWebViewToConsistentStateBeforeTesting):
882         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
883         (WTR::InjectedBundlePage::reset):
884
885 2011-02-26  Yael Aharon  <yael.aharon@nokia.com>
886
887         Reviewed by Andreas Kling.
888
889         [Qt] Fix the focus in MiniBrowser.
890         https://bugs.webkit.org/show_bug.cgi?id=55288
891
892         Give focus to the web view when a load starts.
893
894         * MiniBrowser/qt/BrowserWindow.cpp:
895         (BrowserWindow::urlChanged):
896
897 2011-02-26  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
898
899         Reviewed by Adam Barth.
900
901         [EFL] Add coding style exceptions for EFL port
902         https://bugs.webkit.org/show_bug.cgi?id=54733
903
904         The EFL APIs use EFL naming style, which includes both lower-cased and camel-cased,
905         underscore-sparated values. But, style checking rule have found parameter_name errors
906         in efl directories so far.
907
908         * Scripts/webkitpy/style/checker.py: Add prameter_name rule's exception and efl directories.
909
910 2011-02-26  David Levin  <levin@chromium.org>
911
912         Reviewed by Darin Adler.
913
914         check-webkit-style should check for spaces after periods in comments
915         https://bugs.webkit.org/show_bug.cgi?id=55253
916
917         * Scripts/webkitpy/style/checkers/cpp.py: Added the check.
918         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added tests
919           and fixed the name of one related test.
920
921 2011-02-26  Kenneth Rohde Christiansen  <kenneth@webkit.org>
922
923         Reviewed by Andreas Kling.
924
925         Make it possible to test the targetdensity-dpi support
926         https://bugs.webkit.org/show_bug.cgi?id=55142
927
928         Test the viewport meta tag feature targetdensity-dpi by
929         adding extra arguments to dumpConfigurationForViewport
930
931         Add the new arguments to the shared header and all
932         implementation.
933
934         * DumpRenderTree/LayoutTestController.cpp:
935         (dumpConfigurationForViewportCallback):
936         * DumpRenderTree/LayoutTestController.h:
937         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
938         (LayoutTestController::dumpConfigurationForViewport):
939         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
940         (LayoutTestController::dumpConfigurationForViewport):
941         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
942         (LayoutTestController::dumpConfigurationForViewport):
943         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
944         (LayoutTestController::dumpConfigurationForViewport):
945         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
946         (LayoutTestController::dumpConfigurationForViewport):
947         * DumpRenderTree/qt/LayoutTestControllerQt.h:
948
949 2011-02-26  Sheriff Bot  <webkit.review.bot@gmail.com>
950
951         Unreviewed, rolling out r79764.
952         http://trac.webkit.org/changeset/79764
953         https://bugs.webkit.org/show_bug.cgi?id=55295
954
955         "broke Chromium builds" (Requested by rniwa on #webkit).
956
957         * DumpRenderTree/chromium/TestShell.cpp:
958         (TestShell::resetTestController):
959         * DumpRenderTree/gtk/DumpRenderTree.cpp:
960         (resetDefaultsToConsistentValues):
961         * DumpRenderTree/mac/DumpRenderTree.mm:
962         (resetWebViewToConsistentStateBeforeTesting):
963         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
964         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
965         * DumpRenderTree/win/DumpRenderTree.cpp:
966         (resetWebViewToConsistentStateBeforeTesting):
967         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
968         (WTR::InjectedBundlePage::reset):
969
970 2011-02-26  Yongjun Zhang  <yongjun_zhang@apple.com>
971
972         Reviewed by David Kilzer.
973
974         https://bugs.webkit.org/show_bug.cgi?id=48781
975
976         Add a resource load delegate method to query if WebCore should paint the default broken image for failed images.
977
978         Add a new resource load client method (shouldPaintBrokenImage).  WebKit client can decide if WebCore
979         should paint the default broken image when an image fails to load or decode.
980
981         * DumpRenderTree/LayoutTestController.cpp:
982         (LayoutTestController::LayoutTestController):
983         (setShouldPaintBrokenImageCallback):
984         (LayoutTestController::staticFunctions):
985         (LayoutTestController::setShouldPaintBrokenImage):
986         * DumpRenderTree/LayoutTestController.h:
987         (LayoutTestController::shouldPaintBrokenImage):
988         * DumpRenderTree/mac/ResourceLoadDelegate.mm:
989         (-[ResourceLoadDelegate webView:shouldPaintBrokenImageForURL:]):
990
991 2011-02-26  Vsevolod Vlasov  <vsevik@chromium.org>
992
993         Reviewed by Pavel Feldman.
994
995         DumpRenderTree should reset frame opener between tests.
996         https://bugs.webkit.org/show_bug.cgi?id=54874
997
998         Added clearOpener() calls between tests.
999         Some tests are using window.open() to change current window location.
1000         This changes window.opener property whose nullity is checked before
1001         showing xml viewer.
1002
1003         * DumpRenderTree/chromium/TestShell.cpp:
1004         (TestShell::resetTestController):
1005         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1006         (resetDefaultsToConsistentValues):
1007         * DumpRenderTree/mac/DumpRenderTree.mm:
1008         (resetWebViewToConsistentStateBeforeTesting):
1009         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1010         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
1011         * DumpRenderTree/win/DumpRenderTree.cpp:
1012         (resetWebViewToConsistentStateBeforeTesting):
1013         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1014         (WTR::InjectedBundlePage::reset):
1015
1016 2011-02-25  Lukasz Slachciak  <l.slachciak@samsung.com>
1017
1018         Reviewed by Kenneth Rohde Christiansen.
1019
1020         [EFL] Added EWebLauncher command line option to turn on/off frame flattening
1021         https://bugs.webkit.org/show_bug.cgi?id=54809
1022
1023         * EWebLauncher/main.c: Handling -f option added.
1024         (on_key_down):
1025         (browserCreate):
1026         (main):
1027
1028 2011-02-25  Sheriff Bot  <webkit.review.bot@gmail.com>
1029
1030         Unreviewed, rolling out r79324.
1031         http://trac.webkit.org/changeset/79324
1032         https://bugs.webkit.org/show_bug.cgi?id=55260
1033
1034         Leopard Debug is once again running out of address space while
1035         linking (Requested by aroben on #webkit).
1036
1037         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
1038         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1039
1040 2011-02-25  Tony Chang  <tony@chromium.org>
1041
1042         Reviewed by James Robinson.
1043
1044         Clean up temp files left by crashing DRTs on chromium-win
1045         https://bugs.webkit.org/show_bug.cgi?id=55182
1046
1047         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1048         * BuildSlaveSupport/chromium/remove-crash-logs: Added.
1049
1050 2011-02-25  Adam Roben  <aroben@apple.com>
1051
1052         Double-check that modifying the registry worked
1053
1054         On Windows Vista/7 with UAC enabled, regtool will fail to modify the registry, but it will
1055         still return a successful exit code. So we double-check here that the value we tried to
1056         write to the registry was really written.
1057
1058         Fixes <http://webkit.org/b/55227> old-run-webkit-tests falsely claims to be saving crash
1059         logs on Windows Vista/7
1060
1061         Reviewed by Darin Adler.
1062
1063         * Scripts/old-run-webkit-tests:
1064         (writeRegistryString): After trying to write the string value, check that it was really
1065         written.
1066
1067 2011-02-25  Adam Roben  <aroben@apple.com>
1068
1069         Set the 32-bit post-mortem debugger on 64-bit OSes
1070
1071         The post-mortem debugger is controlled by some registry values in
1072         HKLM\SOFTWARE\Microsoft\Windows NT\CurrentVersion\AeDebug. On 64-bit OSes, you can either
1073         set the values in that key, which will be used for 64-bit apps, or in the equivalent key
1074         under Wow6432Node, which will be used for 32-bit apps. Since DumpRenderTree and
1075         WebKitTestRunner are 32-bit, we want to use the latter location.
1076
1077         Fixes <http://webkit.org/b/55225> old-run-webkit-tests sets the post-mortem debugger for
1078         64-bit apps, but should set it for 32-bit apps instead
1079
1080         Reviewed by Darin Adler.
1081
1082         * Scripts/old-run-webkit-tests:
1083         (readRegistryString):
1084         (writeRegistryString):
1085         Pass --wow32 to regtool so it will get/set values beneath Wow6432Node, which is where
1086         Windows looks for the post-mortem debugger values when a 32-bit app crashes on a 64-bit OS.
1087         Passing --wow32 on a 32-bit OS will not have any effect.
1088
1089 2011-02-25  Adam Roben  <aroben@apple.com>
1090
1091         Clean up code to set up crash log saving on Windows
1092
1093         Fixes <http://webkit.org/b/55224> old-run-webkit-test's Windows registry code is ugly
1094
1095         Reviewed by David Kilzer.
1096
1097         * Scripts/old-run-webkit-tests:
1098         (readRegistryString):
1099         (writeRegistryString):
1100         Moved code to read strings from and write strings to the registry here...
1101
1102         (setUpWindowsCrashLogSaving): ...from here.
1103
1104 2011-02-25  Adam Roben  <aroben@apple.com>
1105
1106         Work around Cygwin's crash-suppression behavior
1107
1108         Cygwin calls ::SetErrorMode(SEM_FAILCRITICALERRORS), which any processes it launches will
1109         inherit. This is bad for testing/debugging, as it causes the post-mortem debugger not to be
1110         invoked. (Cygwin does this because it makes crashes more UNIX-y.) We reset the error mode
1111         when our test apps launch to work around Cygwin's behavior.
1112
1113         Fixes <http://webkit.org/b/55222> Test apps crash silently (without invoking post-mortem
1114         debugger) when launched from Cygwin 1.7
1115
1116         Reviewed by Darin Adler.
1117
1118         * DumpRenderTree/win/DumpRenderTree.cpp:
1119         (main):
1120         * TestWebKitAPI/win/main.cpp:
1121         (main):
1122         * WebKitAPITest/main.cpp:
1123         (main):
1124         * WebKitTestRunner/win/TestControllerWin.cpp:
1125         (WTR::TestController::platformInitialize):
1126         Call ::SetErrorMode(0) to undo Cygwin's folly.
1127
1128 2011-02-24  Jocelyn Turcotte  <jocelyn.turcotte@nokia.com>
1129
1130         Reviewed by Andreas Kling.
1131
1132         [Qt] Revert the support for QNAM affined to a different thread.
1133         https://bugs.webkit.org/show_bug.cgi?id=55149
1134
1135         Qt 4.8 will have QNAM use its own thread internally by default,
1136         no need to keep this complexity in WebKit.
1137
1138         This mainly reverts:
1139         http://trac.webkit.org/changeset/73710
1140         http://trac.webkit.org/changeset/73712
1141
1142         * QtTestBrowser/launcherwindow.cpp:
1143         (LauncherWindow::~LauncherWindow):
1144         (LauncherWindow::initializeView):
1145         (LauncherWindow::createChrome):
1146         * QtTestBrowser/launcherwindow.h:
1147         (WindowOptions::WindowOptions):
1148         * QtTestBrowser/webpage.cpp:
1149         (WebPage::WebPage):
1150         * QtTestBrowser/webpage.h:
1151
1152 2011-02-24  Adam Barth  <abarth@webkit.org>
1153
1154         Reviewed by Eric Seidel.
1155
1156         Add a script for creating gypi files
1157         https://bugs.webkit.org/show_bug.cgi?id=55113
1158
1159         This is a trivial script to create gypi files from the source code in
1160         the current directory.
1161
1162         * Scripts/make-gypi: Added.
1163
1164 2011-02-24  Sheriff Bot  <webkit.review.bot@gmail.com>
1165
1166         Unreviewed, rolling out r79628.
1167         http://trac.webkit.org/changeset/79628
1168         https://bugs.webkit.org/show_bug.cgi?id=55195
1169
1170         "broke chromium win tests" (Requested by dpranke on #webkit).
1171
1172         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1173         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1174
1175 2011-02-24  Dirk Pranke  <dpranke@chromium.org>
1176
1177         Reviewed by Mihai Parparita.
1178
1179         rebaseline-chromium-webkit-tests can't handle GPU variants
1180         https://bugs.webkit.org/show_bug.cgi?id=55002
1181
1182         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
1183         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
1184
1185 2011-02-24  Dirk Pranke  <dpranke@chromium.org>
1186
1187         Reviewed by Mihai Parparita.
1188
1189         nrwt: log command used to start web server to --verbose
1190
1191         https://bugs.webkit.org/show_bug.cgi?id=55188
1192
1193         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1194         * Scripts/webkitpy/layout_tests/port/http_server.py:
1195
1196 2011-02-24  Victoria Kirst  <vrk@google.com>
1197
1198         Reviewed by Mihai Parparita.
1199
1200         [chromium] Move media-file.js and video-test.js back to media/ for UILayoutTests
1201         https://bugs.webkit.org/show_bug.cgi?id=55089
1202
1203         This adds an alias into the media/ directory so that http/tests/media
1204         tests can access the media resources when running in an httpd process.
1205
1206         * Scripts/webkitperl/httpd.pm:
1207         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1208         * Scripts/webkitpy/layout_tests/port/http_server.py:
1209
1210 2011-02-24  Dirk Pranke  <dpranke@chromium.org>
1211
1212         Reviewed by Mihai Parparita.
1213
1214         nrwt: enable multiple processes by default on Chromium Win.
1215
1216         https://bugs.webkit.org/show_bug.cgi?id=55163
1217
1218         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1219         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1220
1221 2011-02-24  Sam Weinig  <sam@webkit.org>
1222
1223         Attempt to fix Qt.
1224
1225         * WebKitTestRunner/TestInvocation.cpp:
1226
1227 2011-02-24  James Robinson  <jamesr@chromium.org>
1228
1229         Unreviewed, rolling out r79598.
1230         http://trac.webkit.org/changeset/79598
1231         https://bugs.webkit.org/show_bug.cgi?id=55089
1232
1233         Appears to be causing trouble with httpd on windows
1234
1235         * Scripts/webkitperl/httpd.pm:
1236         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1237         * Scripts/webkitpy/layout_tests/port/http_server.py:
1238
1239 2011-02-24  Sam Weinig  <sam@webkit.org>
1240
1241         Reviewed by Darin Adler.
1242
1243         Add basic pixel testing support to WebKitTestRunner
1244         https://bugs.webkit.org/show_bug.cgi?id=55161
1245
1246         * Scripts/old-run-webkit-tests:
1247         Don't print color space information when using WebKitTestRunner since
1248         we don't need to change the screen.
1249
1250         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
1251         (WTR::InjectedBundle::initialize):
1252         Add call to initializePlatformDefaults.
1253         
1254         (WTR::InjectedBundle::done):
1255         Return a dictionary on completion with both the text output and pixel dump override.
1256
1257         * WebKitTestRunner/InjectedBundle/InjectedBundle.h:
1258         Add declaration for initializePlatformDefaults.
1259
1260         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
1261         (WTR::LayoutTestController::LayoutTestController):
1262         (WTR::LayoutTestController::dumpAsText):
1263         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
1264         (WTR::LayoutTestController::shouldDumpPixels):
1265         Make setting dumpAsText override the dump pixels preference.
1266
1267         * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm: Added.
1268         (WTR::InjectedBundle::initializePlatformDefaults):
1269         Set platform defaults matching DumpRenderTree.
1270
1271         * WebKitTestRunner/InjectedBundle/qt/InjectedBundleQt.cpp: Added.
1272         (WTR::InjectedBundle::initializePlatformDefaults):
1273         * WebKitTestRunner/InjectedBundle/win/InjectedBundleWin.cpp: Added.
1274         (WTR::InjectedBundle::initializePlatformDefaults):
1275         Add stubs for initializePlatformDefaults.
1276
1277         * WebKitTestRunner/TestController.cpp:
1278         (WTR::TestController::runTest):
1279         (WTR::TestController::runTestingServerLoop):
1280         (WTR::TestController::run):
1281         Parse expected pixel results out of stdin and setup the invocation with it.
1282
1283         * WebKitTestRunner/TestInvocation.cpp:
1284         (WTR::TestInvocation::TestInvocation):
1285         Initialize new members.
1286
1287         (WTR::TestInvocation::~TestInvocation):
1288         (WTR::TestInvocation::setIsPixelTest):
1289         (WTR::sizeWebViewForCurrentTest):
1290         (WTR::TestInvocation::invoke):
1291         (WTR::TestInvocation::dump):
1292         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
1293         * WebKitTestRunner/TestInvocation.h:
1294         Dump pixels in addition to text by calling dumpPixelsAndCompareWithExpected.
1295
1296         * WebKitTestRunner/mac/PlatformWebViewMac.mm:
1297         (WTR::PlatformWebView::PlatformWebView):
1298         Use -[NSColorSpace genericRGBColorSpace] instead of the main screen color space.
1299
1300         * WebKitTestRunner/mac/TestInvocationMac.mm: Added.
1301         (WTR::createCGContextFromPlatformView):
1302         (WTR::computeMD5HashStringForContext):
1303         (WTR::dumpBitmap):
1304         (WTR::forceRepaintFunction):
1305         (WTR::TestInvocation::dumpPixelsAndCompareWithExpected):
1306         Add mac specific pixel dumping support. Right now, we always pull pixels
1307         from the window server.
1308
1309         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
1310         * WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro:
1311         * WebKitTestRunner/qt/WebKitTestRunner.pro:
1312         * WebKitTestRunner/win/InjectedBundle.vcproj:
1313         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
1314         Add new files.
1315
1316 2011-02-24  Chang Shu  <cshu@webkit.org>
1317
1318         Reviewed by Kenneth Rohde Christiansen.
1319
1320         [Qt] Launch MiniBrowser if the test run is WebKit2. 
1321         https://bugs.webkit.org/show_bug.cgi?id=55145
1322
1323         * Scripts/old-run-webkit-tests:
1324         * Scripts/run-launcher:
1325         * Scripts/webkitdirs.pm:
1326
1327 2011-02-24  Victoria Kirst  <vrk@google.com>
1328
1329         Reviewed by Mihai Parparita.
1330
1331         [chromium] Move media-file.js and video-test.js back to media/ for UILayoutTests
1332         https://bugs.webkit.org/show_bug.cgi?id=55089
1333
1334         This adds an alias into the media/ directory so that http/tests/media
1335         tests can access the media resources when running in an httpd process.
1336
1337         * Scripts/webkitperl/httpd.pm:
1338         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1339         * Scripts/webkitpy/layout_tests/port/http_server.py:
1340
1341 2011-02-24  Andrew Wilson  <atwilson@chromium.org>
1342
1343         Unreviewed, rolling out r79570.
1344         http://trac.webkit.org/changeset/79570
1345         https://bugs.webkit.org/show_bug.cgi?id=54874
1346
1347         Breaks chromium build because glue/mocks/mock_web_frame.h/cc
1348         was not updated
1349
1350         * DumpRenderTree/chromium/TestShell.cpp:
1351         (TestShell::resetTestController):
1352         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1353         (resetDefaultsToConsistentValues):
1354         * DumpRenderTree/mac/DumpRenderTree.mm:
1355         (resetWebViewToConsistentStateBeforeTesting):
1356         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1357         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
1358         * DumpRenderTree/win/DumpRenderTree.cpp:
1359         (resetWebViewToConsistentStateBeforeTesting):
1360         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1361         (WTR::InjectedBundlePage::reset):
1362
1363 2011-02-24  Mahesh Kulkarni  <mahesh.kulkarni@nokia.com>
1364
1365         Reviewed by Laszlo Gombos.
1366
1367         [Qt] add android 2.2 user agent to QtTestBrowser
1368         https://bugs.webkit.org/show_bug.cgi?id=55085
1369
1370         Credit to Forrest Hodgkins to find right user-agent which works for youtube.
1371         * QtTestBrowser/useragentlist.txt:
1372
1373 2011-02-24  Adam Roben  <aroben@apple.com>
1374
1375         Change FrameLoadDelegate to support any number of delegates with delayed work to process
1376
1377         This makes our behavior match Mac more closely, and allows us to remove an incorrect
1378         assertion that was firing during some tests. (The assertion was claiming that there was
1379         never more than one delegate with delayed work to process, but that was not the case.)
1380
1381         Fixes <http://webkit.org/b/55146> Assertion failure in FrameLoadDelegate::locationChangeDone
1382         when running http/tests/navigation/back-twice-without-commit.html
1383
1384         Reviewed by Eric Carlson.
1385
1386         * DumpRenderTree/win/FrameLoadDelegate.cpp:
1387         (delegatesWithDelayedWork): Added. Returns all FrameLoadDelegates that have delayed work to
1388         process. A single delegate may appear in this Vector more than once (just as, on Mac, a
1389         single delegate may have multiple performSelector requests).
1390         (processWorkTimer): Pass the HWND to ::KillTimer, for pedantic brownie points. Added an
1391         assertion that the timer firing is the shared process work timer. Instead of using the
1392         single, global "delegate waiting for timer" delegate, give all delegates that have delayed
1393         work to process a chance to process their work.
1394         (FrameLoadDelegate::locationChangeDone): If we don't already have an active timer for
1395         processing delayed work, create one. Then add ourselves to the delegatesWithDelayedWork
1396         Vector so our processWork function will be called when the timer fires.
1397
1398 2011-02-24  Vsevolod Vlasov  <vsevik@chromium.org>
1399
1400         Reviewed by Alexey Proskuryakov.
1401
1402         DumpRenderTree should reset frame opener between tests.
1403         https://bugs.webkit.org/show_bug.cgi?id=54874
1404
1405         Added clearOpener() calls between tests.
1406         Some tests are using window.open() to change current window location.
1407         This changes window.opener property whose nullity is checked before
1408         showing xml viewer.
1409
1410         * DumpRenderTree/chromium/TestShell.cpp:
1411         (TestShell::resetTestController):
1412         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1413         (resetDefaultsToConsistentValues):
1414         * DumpRenderTree/mac/DumpRenderTree.mm:
1415         (resetWebViewToConsistentStateBeforeTesting):
1416         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1417         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
1418         * DumpRenderTree/win/DumpRenderTree.cpp:
1419         (resetWebViewToConsistentStateBeforeTesting):
1420         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1421         (WTR::InjectedBundlePage::reset):
1422
1423 2011-02-24  Adam Roben  <aroben@apple.com>
1424
1425         Don't trigger Windows builds for changes to WebCore.exp.in or the top-level GNUmakefile.am
1426
1427         Fixes <http://webkit.org/b/55144> Windows builders should not have built r79343 or r79440,
1428         but did
1429
1430         Reviewed by Anders Carlsson.
1431
1432         * Scripts/webkitpy/common/config/build.py:
1433         (_should_file_trigger_build): Loosened the pattern for GNUmakefile.am to include the
1434         top-level GNUmakefile.am, and loosened the pattern for .exp files to include WebCore.exp.in.
1435
1436         * Scripts/webkitpy/common/config/build_unittest.py:
1437         (ShouldBuildTest): Added some more tests.
1438
1439 2011-02-24  Andras Becsi  <abecsi@webkit.org>
1440
1441         Reviewed by Laszlo Gombos.
1442
1443         [Qt] MinGW build fails to link
1444         https://bugs.webkit.org/show_bug.cgi?id=55050
1445
1446         Prepend the libraries of subcomponents instead of appending them
1447         to fix the library order according to the dependency of the libraries
1448
1449         * WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro:
1450         * WebKitTestRunner/qt/WebKitTestRunner.pro:
1451
1452 2011-02-24  Gabor Loki  <loki@webkit.org>
1453
1454         Reviewed by Csaba Osztrogonác.
1455
1456         [Qt] Remove ARMv5 Release buildbot.
1457
1458         Remove ARMv5 Release buildbot, because there is very little difference
1459         between ARMv5 and ARMv7 Release bots, and there is also very little
1460         interest in ARMv5 build slave (compared to ARMv7).
1461
1462         * BuildSlaveSupport/build.webkit.org-config/config.json:
1463         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1464
1465 2011-02-22  Ojan Vafai  <ojan@chromium.org>
1466
1467         Reviewed by Tony Chang.
1468
1469         fix revision number finding when creating git patches
1470         https://bugs.webkit.org/show_bug.cgi?id=55024
1471
1472         HEAD~n does not walk merge commits correctly. git log however does.
1473
1474         * Scripts/webkitpy/common/checkout/scm.py:
1475         * Scripts/webkitpy/common/checkout/scm_unittest.py:
1476
1477 2011-02-23  Kenneth Russell  <kbr@google.com>
1478
1479         Unreviewed, rolling out r79387.
1480         https://bugs.webkit.org/show_bug.cgi?id=54885
1481
1482         Reapply the code changes from r79038 since they weren't the root
1483         cause of the test timeouts.
1484
1485         * DumpRenderTree/chromium/WebViewHost.cpp:
1486         (WebViewHost::reset):
1487
1488 2011-02-23  Siddharth Mathur  <siddharth.mathur@nokia.com>
1489
1490         Reviewed by Laszlo Gombos.
1491
1492         [Qt] Make sure Symbian binary UIDs are unique
1493
1494         * MiniBrowser/qt/MiniBrowser.pro: Resolve the UID collision with
1495         QtTestBrowser.pro by changing the UID.
1496
1497 2011-02-23  Patrick Gansterer  <paroga@webkit.org>
1498
1499         Reviewed by Darin Adler.
1500
1501         Rename PLATFORM(CF) to USE(CF)
1502         https://bugs.webkit.org/show_bug.cgi?id=53540
1503
1504         * DumpRenderTree/config.h:
1505
1506 2011-02-23  Dirk Pranke  <dpranke@chromium.org>
1507
1508         Reviewed by Ojan Vafai.
1509
1510         nrwt: turn on multiprocessing on chromium_linux, clean up unit tests for worker model defaults
1511         https://bugs.webkit.org/show_bug.cgi?id=55016
1512
1513         * Scripts/webkitpy/layout_tests/port/base.py:
1514         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1515         * Scripts/webkitpy/layout_tests/port/chromium_linux_unittest.py: Added.
1516         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1517         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
1518         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1519         * Scripts/webkitpy/layout_tests/port/mac.py:
1520         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1521         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
1522         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
1523
1524 2011-02-23  Tony Chang  <tony@chromium.org>
1525
1526         Reviewed by Kent Tamura.
1527
1528         [chromium] DumpRenderTree --check-layout-test-sys-deps leaks 4 temp dirs
1529         https://bugs.webkit.org/show_bug.cgi?id=55004
1530
1531         * DumpRenderTree/chromium/DumpRenderTree.cpp:
1532         (WebKitSupportTestEnvironment::WebKitSupportTestEnvironment):
1533         (WebKitSupportTestEnvironment::~WebKitSupportTestEnvironment):
1534         (main): Use a C++ object to ensure that TearDownTestEnvironment
1535             gets called when --check-layout-test-sys-deps is called
1536
1537 2011-02-23  Gopal Raghavan  <gopal.1.raghavan@nokia.com>
1538
1539         Reviewed by Anders Carlsson.
1540
1541         [Qt] [Qt] check webkit style fails in qt/tests
1542         https://bugs.webkit.org/show_bug.cgi?id=54998
1543
1544         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.
1545
1546         * Scripts/webkitpy/style/checker.py:
1547
1548 2011-02-22  Vsevolod Vlasov  <vsevik@chromium.org>
1549
1550         Reviewed by Pavel Feldman.
1551
1552         DumpRenderTree should dump as text if test path contains "dumpAsText/"
1553         https://bugs.webkit.org/show_bug.cgi?id=54873
1554
1555         * DumpRenderTree/chromium/LayoutTestController.h:
1556         (LayoutTestController::setShouldDumpAsText):
1557         (LayoutTestController::setShouldGeneratePixelResults):
1558         * DumpRenderTree/chromium/TestShell.cpp:
1559         (TestShell::runFileTest):
1560         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1561         (shouldDumpAsText):
1562         (dump):
1563         (runTest):
1564         * DumpRenderTree/mac/DumpRenderTree.mm:
1565         (shouldDumpAsText):
1566         (runTest):
1567         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1568         (WebCore::isDumpAsTextTest):
1569         (WebCore::DumpRenderTree::open):
1570         * DumpRenderTree/win/DumpRenderTree.cpp:
1571         (shouldDumpAsText):
1572         (runTest):
1573         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1574         (WTR::InjectedBundlePage::dump):
1575
1576 2011-02-22  Kenneth Russell  <kbr@google.com>
1577
1578         Reviewed by James Robinson.
1579
1580         Many tests started timing out on the chrome mac canaries around r79035
1581         https://bugs.webkit.org/show_bug.cgi?id=54885
1582
1583         Revert the code changes, but not the Chromium DEPS roll, from
1584         r79038 to see whether this clears up the timing-out tests. Was not
1585         able to reproduce the timeouts locally with a Release mode build
1586         of DRT (on Snow Leopard).
1587
1588         * DumpRenderTree/chromium/WebViewHost.cpp:
1589         (WebViewHost::reset):
1590
1591 2011-02-22  Dirk Pranke  <dpranke@chromium.org>
1592
1593         Reviewed by James Robinson.
1594
1595         nrwt: --platform chromium-gpu doesn't work with --worker-model=processes
1596         https://bugs.webkit.org/show_bug.cgi?id=55009
1597
1598         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
1599
1600 2011-02-18  Mihai Parparita  <mihaip@chromium.org>
1601
1602         Reviewed by Ojan Vafai.
1603
1604         Default to 500 for --exit-after-n-failures
1605         https://bugs.webkit.org/show_bug.cgi?id=54773
1606
1607         Make 500 the default value for --exit-after-n-failures (ORWT bots
1608         already default to this as of r75726). Similarly, change the default for
1609         --exit-after-n-crashes-or-timeouts to 20.
1610
1611         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1612
1613 2011-02-02  Martin Robinson  <mrobinson@igalia.com>
1614
1615         Reviewed by Xan Lopez.
1616
1617         [GTK] DRT needs an implementation of LayoutTestController.setSelectTrailingWhitespace
1618         https://bugs.webkit.org/show_bug.cgi?id=53603
1619
1620         Add support for LayoutTestController.setSelectTrailingWhitespace by calling through
1621         to DumpRenderTreeSupportGtk for this functionality.
1622
1623         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1624         (resetDefaultsToConsistentValues): Always call LayoutTestController.setSelectTrailingWhitespace to
1625         false when reseting values between tests.
1626         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1627         (LayoutTestController::setSelectTrailingWhitespaceEnabled): Call LayoutTestController.setSelectTrailingWhitespace
1628         with the appropriate value.
1629
1630 2011-02-22  Adam Roben  <aroben@apple.com>
1631
1632         Don't trigger a Windows build when the top-level ChangeLog changes
1633
1634         r79320 is an example of a change that shouldn't have built on Windows, but did, because of
1635         this bug.
1636
1637         Fixes <http://webkit.org/b/54957> Windows bots build when top-level ChangeLog changes, but
1638         shouldn't
1639
1640         Reviewed by Anders Carlsson.
1641
1642         * Scripts/webkitpy/common/config/build.py:
1643         (_should_file_trigger_build): Modified the ChangeLog pattern to match the top-level
1644         ChangeLog, and re-sorted it in the list.
1645
1646         * Scripts/webkitpy/common/config/build_unittest.py:
1647         (ShouldBuildTest): Added a test that ChangeLogs don't trigger builds.
1648
1649 2011-02-22  Sheriff Bot  <webkit.review.bot@gmail.com>
1650
1651         Unreviewed, rolling out r78691.
1652         http://trac.webkit.org/changeset/78691
1653         https://bugs.webkit.org/show_bug.cgi?id=54956
1654
1655         Leopard Debug is successfully building again (Requested by
1656         aroben on #webkit).
1657
1658         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
1659         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1660
1661 2011-02-22  Ilya Tikhonovsky  <loislo@chromium.org>
1662
1663         Reviewed by Pavel Feldman.
1664
1665         Web Inspector: flakyness of inspector tests.
1666         https://bugs.webkit.org/show_bug.cgi?id=54729
1667
1668         As far as we have the protocol with sequence numbers we can simplify test harness support
1669         and drop out chomium specific methods.
1670
1671         * DumpRenderTree/DumpRenderTree.gypi:
1672         * DumpRenderTree/chromium/DRTDevToolsAgent.cpp:
1673         (DRTDevToolsAgent::sendMessageToInspectorFrontend):
1674         (DRTDevToolsAgent::runtimePropertyChanged):
1675         (DRTDevToolsAgent::asyncCall):
1676         (DRTDevToolsAgent::call):
1677         * DumpRenderTree/chromium/DRTDevToolsAgent.h:
1678         * DumpRenderTree/chromium/DRTDevToolsCallArgs.cpp: Removed.
1679         * DumpRenderTree/chromium/DRTDevToolsCallArgs.h: Removed.
1680         * DumpRenderTree/chromium/DRTDevToolsClient.cpp:
1681         (DRTDevToolsClient::sendMessageToBackend):
1682         (DRTDevToolsClient::asyncCall):
1683         (DRTDevToolsClient::call):
1684         * DumpRenderTree/chromium/DRTDevToolsClient.h:
1685
1686 2011-02-22  Andras Becsi  <abecsi@webkit.org>
1687
1688         Reviewed by Laszlo Gombos.
1689
1690         [Qt] Redesign the build system
1691         https://bugs.webkit.org/show_bug.cgi?id=51339
1692
1693         Part 2.
1694
1695         Build WebCore as a static library, compile the WebKit API and WebKit2 API
1696         in a final step and link to WebKit2, WebCore and JSC libraries to fix
1697         linking issues resulting from stripped away symbols.
1698
1699         * WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro: Remove superfluous includepaths.
1700
1701 2011-02-21  Lukasz Slachciak  <l.slachciak@samsung.com>
1702
1703         Reviewed by Antonio Gomes.
1704
1705         [EFL] Restored spatial navigation option in EWebLauncher since #18662 bug is closed
1706         https://bugs.webkit.org/show_bug.cgi?id=54806
1707
1708         * EWebLauncher/main.c:
1709         (on_key_down): calling ewk_view_setting_spatial_navigation_set when "F12" pressed added.
1710
1711 2011-02-21  Dirk Pranke  <dpranke@chromium.org>
1712
1713         Reviewed by Kent Tamura.
1714
1715         nrwt: revert behavior on 10.5 to "old-threads". It's possible
1716         that r79062 has introduced a bunch of flakiness, although I
1717         can't explain why right now. This change should restore the
1718         old behavior, and then we can look at the build logs to see
1719         if the tests stop timing out or being flaky.
1720
1721         https://bugs.webkit.org/show_bug.cgi?id=54925
1722
1723         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1724         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
1725         * Scripts/webkitpy/layout_tests/port/mac.py:
1726         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1727
1728 2011-02-21  Dirk Pranke  <dpranke@chromium.org>
1729
1730         Reviewed by Dimitri Glazkov.
1731
1732         nrwt: fix typo causing us to try and repeatedly shut down the servers
1733         https://bugs.webkit.org/show_bug.cgi?id=54904
1734
1735         * Scripts/webkitpy/layout_tests/layout_package/worker_mixin.py:
1736
1737 2011-02-21  Alejandro G. Castro  <alex@igalia.com>
1738
1739         Reviewed by Csaba Osztrogonác.
1740
1741         REGRESSION(r79157):
1742         plugins/get-url-with-javascript-destroying-plugin.html make the
1743         next test crash
1744         https://bugs.webkit.org/show_bug.cgi?id=54863
1745
1746         * GNUmakefile.am:
1747
1748 2011-02-20  Dirk Pranke  <dpranke@chromium.org>
1749
1750         Reviewed by Ojan Vafai.
1751
1752         test-webkitpy: stop skipping a bunch of tests on win32
1753         https://bugs.webkit.org/show_bug.cgi?id=54788
1754
1755         * Scripts/webkitpy/common/net/testoutputset_unittest.py:
1756         * Scripts/webkitpy/test/main.py:
1757
1758 2011-02-20  Dirk Pranke  <dpranke@chromium.org>
1759
1760         Reviewed by Ojan Vafai.
1761
1762         Fix various bugs related to handling of dir separators and real
1763         vs. fake filesystems that were causing test-webkitpy to fail
1764         tests on win32. These bugs have a variety of causes but are
1765         lumped together because they should be easy to review.
1766
1767         https://bugs.webkit.org/show_bug.cgi?id=54700
1768
1769         * Scripts/webkitpy/common/net/testoutput.py:
1770           Here we were using os.path.sep but not escaping it prior to
1771           using in a regexp, which  caused bad things on win32.
1772         * Scripts/webkitpy/common/net/testoutput_unittest.py:
1773           Same here.
1774         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
1775           Here we were using the native filesystem instead of the mock
1776           filesystem, making some tests unnecessarily platform dependent.
1777           Also we weren't setting the default configuration for the
1778           TestChromiumWinPort.
1779         * Scripts/webkitpy/layout_tests/update_webgl_conformance_tests.py:
1780           Here we *shouldn't* use os.path.join(), just so the code becomes
1781           less platform dependent.
1782         * Scripts/webkitpy/style/checker.py:
1783           Here we should be using os.path.sep instead of "/"; the tests
1784           require the native filesystem.
1785         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
1786           Here the underlying code is using os.path.sep, so we probably
1787           should as well.
1788         * Scripts/webkitpy/style/test_expectations_unittest.py:
1789           Here we should be using the filesystem hanging of the port
1790           object (which is the mocked filesystem) rather than the native
1791           one.
1792
1793 2011-02-20  Anders Carlsson  <andersca@apple.com>
1794
1795         Reviewed by Maciej Stachowiak.
1796
1797         Crash when a plug-in requests a javascript: url that destroys the plug-in
1798         https://bugs.webkit.org/show_bug.cgi?id=54837
1799         <rdar://problem/9005475>
1800
1801         Add new plug-in test.
1802
1803         * DumpRenderTree/DumpRenderTree.gypi:
1804         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
1805         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
1806         (PluginTest::NPN_GetURL):
1807         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
1808         * DumpRenderTree/TestNetscapePlugIn/Tests/GetURLWithJavaScriptURLDestroyingPlugin.cpp: Added.
1809         (GetURLWithJavaScriptURLDestroyingPlugin::GetURLWithJavaScriptURLDestroyingPlugin):
1810         (GetURLWithJavaScriptURLDestroyingPlugin::NPP_New):
1811         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1812         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
1813
1814 2011-02-19  Patrick Gansterer  <paroga@webkit.org>
1815
1816         Reviewed by Andreas Kling.
1817
1818         Make WinCE a core builder
1819         https://bugs.webkit.org/show_bug.cgi?id=54801
1820
1821         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
1822         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1823
1824 2011-02-19  Lukasz Slachciak  <l.slachciak@samsung.com>
1825
1826         Reviewed by Andreas Kling.
1827
1828         Fixed problem with launching EWebLauncher with "run-launcher --efl". Incorrect path referring to non-existing .libs directory ws set.
1829         https://webkit.org/b/54778
1830
1831         * Scripts/webkitdirs.pm:
1832
1833 2011-02-18  David Levin  <levin@chromium.org>
1834
1835         Reviewed by Eric Seidel.
1836
1837         check-webkit-style: Misses brace style error when the line with the { has a } in it.
1838         https://bugs.webkit.org/show_bug.cgi?id=54769
1839
1840         * Scripts/webkitpy/style/checkers/cpp.py: Fixed the check for the close brace
1841           to only look after the last open brace, so that the open brace in this line
1842           "} else {" will still be able to trigger the error.
1843         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added related unit tests.
1844
1845 2011-02-18  Zan Dobersek  <zandobersek@gmail.com>
1846
1847         Reviewed by Martin Robinson.
1848
1849         [Gtk] Re-enable meter tag support
1850         https://bugs.webkit.org/show_bug.cgi?id=54762
1851
1852         Meter tag is supported on the Gtk port, enable it by default.
1853
1854         * Scripts/build-webkit:
1855
1856 2011-02-18  David Levin  <levin@chromium.org>
1857
1858         Reviewed by Eric Seidel.
1859
1860         check-webkit-style falsely complains about WebKitGTK+ public headers
1861         https://bugs.webkit.org/show_bug.cgi?id=54650
1862
1863         * Scripts/webkitpy/style/checker.py: Add some exceptions for the gtk files.
1864         * Scripts/webkitpy/style/checkers/cpp.py: Restrict the checks for WEBKIT_API
1865           to the chromium directory (and improved the checks slightly).
1866         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added corresponding tests.
1867
1868 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1869
1870         Reviewed by James Robinson.
1871
1872         new-run-webkit-tests: enable multiple processes by default on
1873         mac. This change removes the artificial restrictions we placed
1874         on the # of child processes we used with the old threading
1875         model, and switches to the new message-based model and multiple
1876         processes, where available. If multiple processes are not
1877         available (Leopard / Python 2.5), then we use the 'inline' model
1878         instead of the 'old-threads' model on the 'mac' port or one
1879         process on the 'chromium-mac' port. We need additional testing
1880         to see if the new 'threads' model works reliably and is worth
1881         supporting, or if we should just live with things being slightly
1882         slow.
1883
1884         https://bugs.webkit.org/show_bug.cgi?id=54596
1885
1886         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1887         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
1888         * Scripts/webkitpy/layout_tests/port/mac.py:
1889         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1890         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1891
1892 2011-02-18  Andrew Wilson  <atwilson@chromium.org>
1893
1894         Unreviewed, rolling out r79047.
1895         http://trac.webkit.org/changeset/79047
1896         https://bugs.webkit.org/show_bug.cgi?id=54596
1897
1898         Broke canary bots - please watch the canaries the next time
1899         you land this
1900
1901         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1902         * Scripts/webkitpy/layout_tests/port/mac.py:
1903
1904 2011-02-18  Emil A Eklund  <eae@chromium.org>
1905
1906         Reviewed by Darin Adler.
1907
1908         Crash in EventHandler::sendContextMenuEventForKey
1909         https://bugs.webkit.org/show_bug.cgi?id=54495
1910
1911         Add support for the context menu key (VK_APPS) to EventSender for
1912         platforms that support the key.
1913
1914         * DumpRenderTree/chromium/EventSender.cpp:
1915         (EventSender::keyDown):
1916         * DumpRenderTree/gtk/EventSender.cpp:
1917         (keyDownCallback):
1918         * DumpRenderTree/qt/EventSenderQt.cpp:
1919         (EventSender::keyDown):
1920         * DumpRenderTree/win/EventSender.cpp:
1921         (keyDownCallback):
1922
1923 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1924
1925         Reviewed by Ojan Vafai.
1926
1927         new-run-webkit-tests: enable multiple processes by default on mac
1928         https://bugs.webkit.org/show_bug.cgi?id=54596
1929
1930         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1931         * Scripts/webkitpy/layout_tests/port/mac.py:
1932
1933 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1934
1935         Reviewed by Adam Roben.
1936
1937         Skip pretty patch unit tests if ruby isn't installed.
1938
1939         https://bugs.webkit.org/show_bug.cgi?id=54699
1940
1941         * Scripts/webkitpy/common/prettypatch_unittest.py:
1942
1943 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1944
1945         Reviewed by Adam Roben.
1946
1947         Skip mac.test_skipped_file_paths() when running on win32; the
1948         test has hardcoded '/' directory separators and since it's a
1949         port-specific test and we have coverage on other ports it's not
1950         worth it to make the test more generic.
1951
1952         https://bugs.webkit.org/show_bug.cgi?id=54698
1953
1954         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1955
1956 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1957
1958         Reviewed by Adam Roben.
1959
1960         Fix webkitpy.layout_tests.port.config_unittest to work on Win32.
1961
1962         https://bugs.webkit.org/show_bug.cgi?id=54703
1963
1964         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1965
1966 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1967
1968         Reviewed by Adam Roben.
1969
1970         Change the default port we're mocking out on win32 when we
1971         run mock_drt_unittest. Normally we'd default to the
1972         port-specific default, but that would be "win", which doesn't
1973         currently work.
1974
1975         https://bugs.webkit.org/show_bug.cgi?id=54708
1976
1977         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
1978
1979 2011-02-18  Kenneth Russell  <kbr@google.com>
1980
1981         Reviewed by James Robinson.
1982
1983         [chromium] Clean up initialization of minimum timer interval
1984         https://bugs.webkit.org/show_bug.cgi?id=54772
1985
1986         Use new GetForegroundTabTimerInterval function in webkit_support
1987         to reset the minimum timer interval before each test.
1988
1989         * DumpRenderTree/chromium/WebViewHost.cpp:
1990         (WebViewHost::reset):
1991
1992 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1993
1994         Reviewed by Adam Roben.
1995
1996         r78494 fixed the chromium_win unittest failing under win32,
1997         but not cygwin. This fixes cygwin as well.
1998
1999         https://bugs.webkit.org/show_bug.cgi?id=54652
2000
2001         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2002
2003 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
2004
2005         Reviewed by Ojan Vafai, Mihai Parparita.
2006
2007         new-run-webkit-tests: r78522 made it impossible to create a
2008         'chromium-win' or 'chromium-mac' port; they would always get the
2009         version tacked on to the end. It turns out that we actually need
2010         to be able to create these ports in order for
2011         rebaseline-chromium-webkit-tests to work correctly given the
2012         broken coverage of our layout bots. As soon as we get SL and
2013         Win 7 bots running reliably on the canaries, we can get rid of
2014         this change. Note that we only need to change the chromium_mac
2015         and win ports because linux doesn't yet use versions and that
2016         script doesn't work with non-chromium ports.
2017
2018         Also, this change fixes a bug where we were using
2019         "platform/chromium-win-win7" and
2020         "platform/chromium-mac-snowleopard" as the baseline dirs for
2021         Win 7 and SL; we should've been using the generic directories
2022         instead since they are the newest versions of the os's.
2023
2024         https://bugs.webkit.org/show_bug.cgi?id=54691
2025
2026         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
2027         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
2028         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2029         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2030         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2031
2032 2011-02-16  MORITA Hajime  <morrita@google.com>
2033
2034         Reviewed by Kent Tamura.
2035
2036         [Chromium][DRT] some PostDelayedTask() accesses possibly dangling this pointer.
2037         http://webkit.org/b/53899
2038
2039         Replaced static invocation wrappers with MethodTask objects,
2040         which are removed on WebViewHost destructor.
2041         
2042         * DumpRenderTree/chromium/WebViewHost.cpp:
2043         (WebViewHost::requestCheckingOfText):
2044         (WebViewHost::scheduleAnimation):
2045         (WebViewHost::closeWidgetSoon):
2046         * DumpRenderTree/chromium/WebViewHost.h:
2047         (WebViewHost::HostMethodTask::HostMethodTask):
2048         (WebViewHost::HostMethodTask::runIfValid):
2049         (WebViewHost::taskList):
2050
2051 2011-02-07  Hayato Ito  <hayato@chromium.org>
2052
2053         Reviewed by Ojan Vafai.
2054
2055         Get rid of code which writes test results from test_type's
2056         compare_output() method.
2057         Writing test results now happens in single_test_runner, using
2058         newly introduced test_result_writer module.
2059
2060         Subsequent patches will eliminate test_type/* classes.
2061
2062         https://bugs.webkit.org/show_bug.cgi?id=53908
2063
2064         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
2065         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py: Added.
2066         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
2067         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
2068         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2069
2070 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
2071
2072         Reviewed by Adam Roben.
2073
2074         Add blacklist of remaining webkitpy/* modules that fail on
2075         Win32, so that test-webkitpy will at least run cleanly.
2076
2077         https://bugs.webkit.org/show_bug.cgi?id=54709
2078
2079         * Scripts/webkitpy/layout_tests/test/main.py:
2080
2081 2011-02-17  Simon Fraser  <simon.fraser@apple.com>
2082
2083         Reviewed by Sam Weinig.
2084
2085         Set a preference to ensure that scrollbars are in a predictable
2086         state for DRT.
2087
2088         * DumpRenderTree/mac/DumpRenderTree.mm:
2089         (resetDefaultsToConsistentValues):
2090
2091 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
2092
2093         Reviewed by Adam Roben.
2094
2095         test-webkitpy doesn't work with the multiprocessing module on win32
2096         or cygwin, so we skip it for now.
2097
2098         https://bugs.webkit.org/show_bug.cgi?id=54520
2099
2100         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
2101         * Scripts/webkitpy/layout_tests/layout_package/run_webkit_tests_unittest.py:
2102
2103 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
2104
2105         Reviewed by Adam Roben.
2106
2107         Fix version, baseline_search_path for chromium-win-win7. Also
2108         fix chromium_win_unittests, and fix port_testcase.make_port()
2109         to correctly use the options passed into it and only set
2110         results_directory() if necessary.
2111
2112         https://bugs.webkit.org/show_bug.cgi?id=54652
2113
2114         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2115         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2116         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
2117
2118 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
2119
2120         Reviewed by Ojan Vafai.
2121
2122         Fix a hang in new-run-webkit-tests when run with
2123         --dry-run --worker-model=processes
2124
2125         https://bugs.webkit.org/show_bug.cgi?id=54595
2126
2127         * Scripts/webkitpy/layout_tests/layout_package/test_runner2.py:
2128         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2129
2130 2011-02-17  Kevin Ollivier  <kevino@theolliviers.com>
2131
2132         [wx] Build fixes after recent changes.
2133
2134         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2135         (LayoutTestController::setMinimumTimerInterval):
2136         * wx/build/settings.py:
2137
2138 2011-02-17  Csaba Osztrogonác  <ossy@webkit.org>
2139
2140         Reviewed by Andreas Kling.
2141
2142         [Qt] ImageDiff sometimes hangs
2143         https://bugs.webkit.org/show_bug.cgi?id=54641
2144
2145         * DumpRenderTree/qt/ImageDiff.cpp: Clear actualImage and baselineImage after all comparison.
2146         (main):
2147
2148 2011-02-17  Peter Varga  <pvarga@webkit.org>
2149
2150         Reviewed by Csaba Osztrogonác.
2151
2152         The run-sunspider --v8-suite doesn't work
2153         https://bugs.webkit.org/show_bug.cgi?id=54660
2154
2155         * Scripts/run-sunspider:
2156
2157 2011-02-17  Gabor Rapcsanyi  <rgabor@webkit.org>
2158
2159         Reviewed by Csaba Osztrogonác.
2160
2161         [NRWT] Fix environment setup on Qt port.
2162         https://bugs.webkit.org/show_bug.cgi?id=54549
2163
2164         * Scripts/webkitpy/layout_tests/port/qt.py:
2165
2166 2011-02-16  Yael Aharon  <yael.aharon@nokia.com>
2167
2168         Reviewed by Andreas Kling.
2169
2170         [Qt] Enable tiled backing store by default in MiniBrowser.
2171         https://bugs.webkit.org/show_bug.cgi?id=54581
2172
2173         Instead of a flag to turn it on, a new flag is defined to turn
2174         tiled backing store off.
2175
2176         * MiniBrowser/qt/MiniBrowserApplication.cpp:
2177         (MiniBrowserApplication::handleUserOptions):
2178         * MiniBrowser/qt/MiniBrowserApplication.h:
2179         (WindowOptions::WindowOptions):
2180
2181 2011-02-16  Eric Seidel  <eric@webkit.org>
2182
2183         Reviewed by Mihai Parparita.
2184
2185         flaky test archive uploads are always 22 bytes long
2186         https://bugs.webkit.org/show_bug.cgi?id=54593
2187
2188         Speculative fix for the 22-bytes bug.
2189
2190         * Scripts/webkitpy/tool/bot/flakytestreporter.py:
2191
2192 2011-02-16  Anna Cavender  <annacc@chromium.org>
2193
2194         Reviewed by Mihai Parparita.
2195
2196         http/tests/media should be served over HTTP (not from local file)
2197         https://bugs.webkit.org/show_bug.cgi?id=54028
2198
2199         Removed special file:// treatment of http/tests/media
2200
2201         * Scripts/old-run-webkit-tests:
2202         * Scripts/webkitpy/layout_tests/port/base.py:
2203
2204 2011-02-16  Mihai Parparita  <mihaip@chromium.org>
2205
2206         Reviewed by Ojan Vafai.
2207
2208         ChangeLog.parse_latest_entry_from_file does not handle rolled over ChangeLogs
2209         https://bugs.webkit.org/show_bug.cgi?id=54609
2210         
2211         r78737 rolled over ChangeLogs, and webkit-patch complains when landing
2212         patches such as this one that are the first to add an entry to the
2213         ChangeLog (since they don't see a second date line to know where the
2214         entry ends).
2215
2216         * Scripts/webkitpy/common/checkout/changelog.py:
2217         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
2218
2219 == Rolled over to ChangeLog-2011-02-16 ==