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