e01268a0295fb50c9ec5fe89234d9ce453cbf67f
[WebKit-https.git] / Tools / ChangeLog
1 2011-03-01  Daniel Cheng  <dcheng@chromium.org>
2
3         Reviewed by David Levin.
4
5         Add feature define for data transfer items
6         https://bugs.webkit.org/show_bug.cgi?id=55510
7
8         * Scripts/build-webkit:
9
10 2011-02-23  Joseph Pecoraro  <joepeck@webkit.org>
11
12         Reviewed by Timothy Hatcher.
13
14         All Console Messages should be passed to ChromeClients.
15         https://bugs.webkit.org/show_bug.cgi?id=54926
16
17         Update DumpRenderTree's UIDelegate to use the new console message delegate.
18
19         * DumpRenderTree/mac/UIDelegate.mm:
20         (-[UIDelegate webView:addMessageToConsole:withSource:]):
21
22 2011-03-01  Dirk Pranke  <dpranke@chromium.org>
23
24         Reviewed by Adam Barth.
25
26         Tragically, we stopped running the python unit tests
27         on Chromium Win just as I was getting them to be green.
28         This patch turns them back on.
29
30         https://bugs.webkit.org/show_bug.cgi?id=55521
31
32         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
33
34 2011-03-01  Dirk Pranke  <dpranke@chromium.org>
35
36         Reviewed by Adam Roben.
37
38         Disable
39         webkitpy.common.system.executive_unittest.ExecutiveTest.\
40            test_kill_all fails on win until it can be properly triaged.
41
42         https://bugs.webkit.org/show_bug.cgi?id=54790
43
44         * Scripts/webkitpy/common/system/executive_unittest.py:
45
46     2011-03-01  Helder Correia  <helder@sencha.com>
47
48             Unreviewed. Add myself to committers.py.
49
50             * Scripts/webkitpy/common/config/committers.py:
51
52 2011-03-01  Adam Roben  <aroben@apple.com>
53
54         Give up if a crash log for the web process is taking too long to be saved on Windows
55
56         This should keep the bots from getting stuck while waiting for a crash log, if writing a
57         crash log ever hangs or the UI process for some reason doesn't notice it's finished.
58
59         Fixes <http://webkit.org/b/55499> WebKitTestRunner can hang forever waiting for a crash log
60         to be saved for the web process
61
62         Reviewed by Steve Falkenburg.
63
64         * WebKitTestRunner/win/TestControllerWin.cpp:
65         (WTR::runRunLoopUntil): Moved code to run the run loop here here from platformRunUntil.
66         Generalized the code slightly to handle the optional object to wait on.
67         (WTR::TestController::platformRunUntil): Use the new runRunLoopUntil function, and also use
68         it (with a timeout) when waiting for a crash log for the web process to be saved. This will
69         prevent us from waiting forever.
70
71 2011-03-01  Adam Roben  <aroben@apple.com>
72
73         Give WebKitLauncherWin's PDB file a name different from WebKit's
74
75         WebKitLauncherWin (which builds WebKit.exe) and WebKit (which builds WebKit.dll) were
76         fighting over the same PDB file, meaning that one of them was going to lose and not have
77         symbols. WebKitLauncherWin's PDB file is now named WebKitLauncherWin.pdb.
78
79         Fixes <http://webkit.org/b/55495> Crashes in WebKit.dll have bad backtraces for some test
80         runs (due to missing symbols)
81
82         Reviewed by David Kilzer.
83
84         * WebKitLauncherWin/WebKitLauncherWinCommon.vsprops: Use the project name,
85         "WebKitLauncherWin", instead of the target name, "WebKit", when naming the PDB file.
86
87 2011-03-01  Adam Roben  <aroben@apple.com>
88
89         Count web process crashes toward the --exit-after-n-crashes-or-timeouts limit
90
91         Fixes <http://webkit.org/b/55492> Web process crashes aren't counted toward the
92         --exit-after-n-crashes-or-timeouts limit
93
94         Reviewed by David Kilzer.
95
96         * Scripts/old-run-webkit-tests:
97         (stopRunningTestsEarlyIfNeeded): Include web process crashes in the computation, and print
98         the number of tests that caused the web process to crash when exiting early.
99
100 2011-03-01  Adam Roben  <aroben@apple.com>
101
102         Save a crash log when the web process crashes
103
104         On Windows, WebKitTestRunner now detects when the web process is crashing and waits to exit
105         until it has finished crashing, which guarantees that the crash log will have had time to be
106         saved, too. On Mac, we always wait until ReportCrash has exited before capturing the crash
107         log, so all we have to do is choose the right crash log out of the CrashReporter directory.
108
109         Fixes <http://webkit.org/b/44121> <rdar://problem/8320759> When the web process crashes and
110         a crash log is being saved, WebKitTestRunner thinks the web process has become unresponsive
111
112         Reviewed by Sam Weinig.
113
114         * Scripts/old-run-webkit-tests:
115         (testCrashedOrTimedOut): Don't kill WebKitTestRunner when the web process crashes. It will
116         kill itself. On Windows, this will cause us to wait until the crash log has been saved. On
117         Mac, it should have no effect. Capture saved crash logs for web process crashes, too.
118         (captureSavedCrashLog): Added $webProcessCrashed argument. On Mac, look for
119         WebProces_*.crash files when the web process crashes.
120
121         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
122         (WTR::InjectedBundle::initialize): Added an initializationUserData argument. Updated for
123         initializePlatformDefaults -> platformInitialize rename. Pass the initializationUserData
124         along to platformInitialize.
125
126         * WebKitTestRunner/InjectedBundle/InjectedBundle.h: See above.
127
128         * WebKitTestRunner/InjectedBundle/InjectedBundleMain.cpp:
129         (WKBundleInitialize): Pass along the initializationUserData to the InjectedBundle.
130
131         * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm:
132         (WTR::InjectedBundle::platformInitialize):
133         * WebKitTestRunner/InjectedBundle/qt/InjectedBundleQt.cpp:
134         (WTR::InjectedBundle::platformInitialize):
135         Updated function signature.
136
137         * WebKitTestRunner/InjectedBundle/win/InjectedBundleWin.cpp:
138         (WTR::exceptionFilter): Added. Tells the UI process we're crashing by signaling the
139         webProcessCrashingEvent, then lets the crash continue as normal.
140
141         (WTR::InjectedBundle::platformInitialize): Hook up exceptionFilter. Retrieve the name of the
142         event we should use to tell the UI process we're crashing from the initializationUserData,
143         and get a handle to that event.
144
145         * WebKitTestRunner/TestController.cpp:
146         (WTR::TestController::TestController): Initialize new members.
147         (WTR::TestController::processDidCrash): Removed unnecessary WKPageRef argument. Changed to
148         only print the "#CRASHED - WebProcess" message once, since this can be called more than once
149         when a crash log is being saved on Windows. Exit right away if specified. (This is the
150         default.)
151
152         * WebKitTestRunner/TestController.h: Added new members.
153
154         * WebKitTestRunner/win/TestControllerWin.cpp:
155         (WTR::TestController::platformInitialize): Set up the event the web process will use to tell
156         us it's crashing.
157         (WTR::TestController::platformRunUntil): Pass MWMO_INPUTAVAILABLE to
158         ::MsgWaitForMultipleObjectsEx so we'll process messages that have already been seen by
159         ::PeekMessage. (This is unrelated to the bug fix.) Notice when the webProcessCrashingEvent
160         has been signaled. When this happens, print the "#CRASHED - WebProcess" message right away
161         so the test harness will know the web process has crashed and not try to kill us, then wait
162         for the web process to finish crashing so a crash log will have time to be saved.
163         (WTR::toWK): Simple hepler function.
164         (WTR::TestController::platformInitializeContext): Pass along the name of the event the web
165         process should use to tell us it is crashing in the context's initialization user data.
166
167 2011-03-01  Dimitri Glazkov  <dglazkov@chromium.org>
168
169         Reviewed by Tony Gentilcore.
170
171         Move the checks to exclude Python/Perl tests for Chromium Win build to the right place.
172         https://bugs.webkit.org/show_bug.cgi?id=55476
173
174         Chromium Win uses BuildAndTestFactory, not TestFactory.
175
176         * BuildSlaveSupport/build.webkit.org-config/master.cfg: Moved to the right place.
177
178 2011-03-01  Adam Roben  <aroben@apple.com>
179
180         Test that the WebKit2 responsiveness timer doesn't fire too early
181
182         Test for <http://webkit.org/b/55417> <rdar://problem/9065287> RunLoop::Timer fires 1000x too
183         early on Windows
184
185         Reviewed by Anders Carlsson.
186
187         * TestWebKitAPI/PlatformUtilities.h: Added sleep().
188
189         * TestWebKitAPI/Tests/WebKit2/ResponsivenessTimerDoesntFireEarly.cpp: Added.
190         (TestWebKitAPI::didReceiveMessageFromInjectedBundle):
191         (TestWebKitAPI::didFinishLoadForFrame):
192         (TestWebKitAPI::processDidBecomeUnresponsive):
193         (TestWebKitAPI::setInjectedBundleClient):
194         (TestWebKitAPI::setPageLoaderClient):
195         Simple helper functions.
196
197         (TestWebKitAPI::TEST): Load an HTML file to make sure the web process is initialized. Then
198         tell the web process to pause and press the spacebar key. The spacebar keypress should cause
199         the responsiveness timer to start, but the web process should unpause before it has a chance
200         to fire. Run until the web process has unpaused, and assert that the timer didn't fire.
201
202         * TestWebKitAPI/Tests/WebKit2/ResponsivenessTimerDoesntFireEarly_Bundle.cpp: Added.
203         (TestWebKitAPI::ResponsivenessTimerDoesntFireEarlyTest::ResponsivenessTimerDoesntFireEarlyTest):
204         Call up to the base class.
205         (TestWebKitAPI::ResponsivenessTimerDoesntFireEarlyTest::didReceiveMessage): When asked to
206         pause, sleep for 0.5 seconds, then send back a message saying we paused.
207
208         * TestWebKitAPI/mac/PlatformUtilitiesMac.mm:
209         (TestWebKitAPI::Util::sleep): Added. Calls through to usleep.
210         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp:
211         (TestWebKitAPI::Util::sleep): Added. Calls through to ::Sleep.
212
213         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
214         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
215         * TestWebKitAPI/win/TestWebKitAPIInjectedBundle.vcproj:
216         Added new files.
217
218 2011-03-01  Andras Becsi  <abecsi@webkit.org>
219
220         Reviewed by Csaba Osztrogonác.
221
222         [Qt] Clean up the project files and move common options to WebKit.pri.
223
224         * WebKitTestRunner/qt/WebKitTestRunner.pro: Deduplicate.
225
226
227 2011-03-01  Adam Roben  <aroben@apple.com>
228
229         Skip another multiprocessing test on Windows
230
231         It is sometimes failing (and may be leaving child Python processes around).
232         <http://webkit.org/b/55087> tracks the failure.
233
234         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
235         (FuncitonTests.test_get__processes): Skip this test on Windows.
236
237 2011-03-01  Ilya Tikhonovsky  <loislo@chromium.org>
238
239         Reviewed by Pavel Feldman.
240
241         Web Inspector: reduce inspector tests flakiness.
242         https://bugs.webkit.org/show_bug.cgi?id=55401
243
244         * DumpRenderTree/chromium/TestShell.cpp:
245         (TestShell::runFileTest):
246         * DumpRenderTree/chromium/WebPreferences.cpp:
247         (WebPreferences::reset):
248
249 2011-02-28  Mihai Parparita  <mihaip@chromium.org>
250
251         Reviewed by James Robinson.
252
253         NRWT should still upload results when it exits early due to too many failures
254         https://bugs.webkit.org/show_bug.cgi?id=55432
255         
256         Only keyboard interrupts should prevent results from being uploaded.
257
258         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
259         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
260
261 2011-02-28  Alice Liu  <alice.liu@apple.com>
262
263         Reviewed by Adam Roben and Gavin Barraclough.
264
265         Test for https://bugs.webkit.org/show_bug.cgi?id=54898
266         InjectedBundleNodeHandle dies too early in WKBundleHitTestResultGetNodeHandle.
267
268         * TestWebKitAPI/InjectedBundleController.h:
269         (TestWebKitAPI::InjectedBundleController::bundle):
270         * TestWebKitAPI/PlatformWebView.h:
271         * TestWebKitAPI/Tests/WebKit2/HitTestResultNodeHandle.cpp: Added.
272         (TestWebKitAPI::didFinishLoadForFrame):
273         (TestWebKitAPI::didReceiveMessageFromInjectedBundle):
274         (TestWebKitAPI::setPageLoaderClient):
275         (TestWebKitAPI::TEST):
276         * TestWebKitAPI/Tests/WebKit2/HitTestResultNodeHandle_Bundle.cpp: Added.
277         (TestWebKitAPI::HitTestResultNodeHandleTest::HitTestResultNodeHandleTest):
278         (TestWebKitAPI::HitTestResultNodeHandleTest::getContextMenuFromDefaultMenu):
279         (TestWebKitAPI::HitTestResultNodeHandleTest::didCreatePage):
280         * TestWebKitAPI/mac/PlatformWebViewMac.mm:
281         (TestWebKitAPI::PlatformWebView::simulateRightClick):
282         * TestWebKitAPI/win/PlatformWebViewWin.cpp:
283         (TestWebKitAPI::PlatformWebView::simulateRightClick):
284         
285         Added files to:
286         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
287         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
288         * TestWebKitAPI/win/TestWebKitAPIInjectedBundle.vcproj:
289
290 2011-02-28  Adam Roben  <aroben@apple.com>
291
292         Start running InjectedBundle API tests on Windows
293
294         A new project, TestWebKitAPIInjectedBundle, builds the injected bundle DLL.
295
296         Fixes <http://webkit.org/b/55420> run-api-tests doesn't run injected bundle tests on Windows
297
298         Reviewed by Darin Adler.
299
300         * TestWebKitAPI/Configurations/TestWebKitAPIDebug.vsprops: Added.
301         * TestWebKitAPI/Configurations/TestWebKitAPIDebugAll.vsprops: Added.
302         * TestWebKitAPI/Configurations/TestWebKitAPIDebugCairoCFLite.vsprops: Added.
303         * TestWebKitAPI/Configurations/TestWebKitAPIRelease.vsprops: Added.
304         * TestWebKitAPI/Configurations/TestWebKitAPIReleaseCairoCFLite.vsprops: Added.
305         These are just extracted from TestWebKitAPI.vcproj.
306
307         * TestWebKitAPI/Configurations/TestWebKitAPIInjectedBundleCommon.vsprops: Added. Just sets
308         up the DLL's filename.
309
310         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp: Fixed DLL name.
311         (TestWebKitAPI::Util::createInjectedBundlePath): Changed to use kCFAllocatorNull instead of
312         0 (== kCFAllocatorDefault) so that CF won't try to deallocate our C string.
313
314         * TestWebKitAPI/win/TestWebKitAPI.sln: Added TestWebKitAPIInjectedBundle. Removed the bogus
315         "all" configuration.
316
317         * TestWebKitAPI/win/TestWebKitAPI.vcproj: Changed to use the new .vsprops files. Added
318         injected bundle tests. Let VS resort things.
319
320         * TestWebKitAPI/win/TestWebKitAPIInjectedBundle.vcproj: Copied from Tools/TestWebKitAPI/win/TestWebKitAPI.vcproj.
321
322 2011-02-28  Adam Roben  <aroben@apple.com>
323
324         Add a test for the paint rect passed via WM_PAINT to windowless plugins
325
326         Test for <http://webkit.org/b/55365> <rdar://problem/9031089> REGRESSION (r79040): WebKit2:
327         Large portions of pages with plugins paint black
328
329         Reviewed by Anders Carlsson.
330
331         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
332         (PluginTest::NPP_HandleEvent):
333         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
334         Added NPP_HandleEvent. Just returns 0 at this level.
335
336         * DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowlessPaintRectCoordinates.cpp: Added.
337         (WindowlessPaintRectCoordinates::WindowlessPaintRectCoordinates): Call up to the base class.
338         (WindowlessPaintRectCoordinates::NPP_New): Mark ourselves as windowless.
339         (WindowlessPaintRectCoordinates::NPP_HandleEvent): Check that the paint rect passed via
340         WM_PAINT has the expected coordinates.
341
342         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
343         (NPP_HandleEvent): Give the PluginTest a chance to handle the event before doing anything
344         else.
345
346         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Added new file, let VS
347         resort things.
348
349         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
350         (WTR::InjectedBundlePage::didFinishLoadForFrame): Added a call to WKBundlePageForceRepaint.
351         This is roughly equivalent to the call to -[WebView displayIfNeeded] in -[FrameLoadDelegate
352         webView:didFinishLoadForFrame:] in DumpRenderTree.
353
354         * WebKitTestRunner/win/PlatformWebViewWin.cpp:
355         (WTR::PlatformWebView::PlatformWebView): Tell the WKView it's in a window so that plugins
356         can start running.
357
358 2011-02-28  Adam Roben  <aroben@apple.com>
359
360         Tell MiniBrowser's WKViews that they're in a window
361
362         Fixes <http://webkit.org/b/55364> Plugins don't work in MiniBrowser on Windows
363
364         Reviewed by Anders Carlsson.
365
366         * MiniBrowser/win/BrowserView.cpp:
367         (BrowserView::create): Call WKViewSetIsInWindow after creating the view.
368
369 2011-02-28  Alexis Menard  <alexis.menard@openbossa.org>
370
371         Reviewed by Csaba Osztrogonác.
372
373         The new QML tests require QML_IMPORT_PATH variable to be set in order to
374         load the proper plugin.
375
376         * Scripts/run-qtwebkit-tests:
377
378 2011-02-28  Joone Hur  <joone.hur@collabora.co.uk>
379
380         Unreviewed, fix spelling mistake.
381
382         hardware-accelated => hardware-accelerated.
383
384         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
385
386 2011-02-28  Kristian Amlie  <kristian.amlie@nokia.com>
387
388         Reviewed by Andreas Kling.
389
390         Made sure that the build-webkit --qmake argument is always respected.
391
392         Previously, it would be respected during the build, but not during
393         the early feature detection. The build would also fail if qmake was
394         not in your path, even if you specified --qmake with a valid qmake
395         binary.
396
397         [Qt] WebKit patches required to work with a modularized version of Qt
398         https://bugs.webkit.org/show_bug.cgi?id=53916
399
400         * Scripts/build-webkit:
401         * Scripts/webkitdirs.pm:
402
403 2011-02-27  Ojan Vafai  <ojan@chromium.org>
404
405         Fix failing tests from http://trac.webkit.org/changeset/79837.
406
407         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
408         This was just a mistake that would break if someone tried printing with this configuration.
409         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
410         Method signature changed. Didn't notice that this method was being calling in a unittest.
411
412 2011-02-24  Hayato Ito  <hayato@chromium.org>
413
414         Reviewed by Eric Seidel.
415
416         [NRWT] Remove test_types/* classes.
417
418         This is the last patch in the series of efforts which tried to remove test_types classes.
419         A single_test_runner module will take over responsibilities of test_type classes.
420
421         From this patch, we won't display statistics per test_type even if a '--verbose' option is given.
422
423         https://bugs.webkit.org/show_bug.cgi?id=55123
424
425         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
426         * Scripts/webkitpy/layout_tests/layout_package/test_results.py:
427         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:
428         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
429         * Scripts/webkitpy/layout_tests/layout_package/worker_mixin.py:
430         * Scripts/webkitpy/layout_tests/test_types/__init__.py: Removed.
431         * Scripts/webkitpy/layout_tests/test_types/image_diff.py: Removed.
432         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py: Removed.
433         * Scripts/webkitpy/layout_tests/test_types/test_type_base_unittest.py: Removed.
434         * Scripts/webkitpy/layout_tests/test_types/text_diff.py: Removed.
435
436 2011-02-25  Ojan Vafai  <ojan@chromium.org>
437
438         Reviewed by Tony Chang.
439
440         Change results.json format to the one used by unexpected_results.json
441         https://bugs.webkit.org/show_bug.cgi?id=52267
442
443         Also add runtimes in milliseconds to the JSON and make the output format more compact.
444         Named the file full_results.json to avoid conflicting with the results.json
445         file the test-results server currently serves up.
446
447         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
448         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
449         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
450         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
451         * Scripts/webkitpy/layout_tests/layout_package/result_summary.py:
452         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
453
454 2011-02-27  Adam Roben  <aroben@apple.com>
455
456         Followup fix to r79827
457
458         * DumpRenderTree/win/UIDelegate.cpp:
459         (UIDelegate::webViewAddMessageToConsole): Replaced call to urlSuitableForTestResult with
460         lastPathComponent. This matches what we did here before r79827, and matches Mac.
461
462 2011-02-27  Dan Bernstein  <mitz@apple.com>
463
464         Reviewed by Adam Roben.
465
466         Make bisect-builds print a trac link for the regression range
467         https://bugs.webkit.org/show_bug.cgi?id=55331
468
469         * Scripts/bisect-builds:
470
471 2011-02-27  Adam Roben  <aroben@apple.com>
472
473         Use iswalpha instead of isalpha when dealing with wchar_ts in EditingDelegate
474
475         Fixes <http://webkit.org/b/55062> <rdar://problem/9059907> Crash beneath
476         EditingDelegate::checkSpellingOfString when running fast/forms/input-text-maxlength.html or
477         fast/forms/input-text-paste-maxlength.html on Windows with full page heap enabled
478
479         Reviewed by Anders Carlsson.
480
481         * DumpRenderTree/win/EditingDelegate.cpp:
482         (indexOfFirstWordCharacter):
483         (wordLength):
484         Use iswalpha instead of isalpha, since TCHAR is really just wchar_t.
485
486 2011-02-27  Adam Roben  <aroben@apple.com>
487
488         Make ResourceLoadDelegate print URLs relative to the main resource on Windows
489
490         This matches Mac. (In fact, the code was ported from Mac!)
491
492         Fixes <http://webkit.org/b/55328> URLs printed by ResourceLoadDelegate on Windows don't
493         match Mac
494
495         Reviewed by Anders Carlsson.
496
497         * DumpRenderTree/win/DumpRenderTree.cpp:
498         (substringFromIndex): Added. Emulates -[NSString substringFromIndex:].
499         (urlSuitableForTestResult): Ported code from -[NSURL(DRTExtras)
500         _drt_descriptionSuitableForTestResult].
501         (cfStringRefToWString): Moved here from LayoutTestControllerWin.cpp.
502
503         * DumpRenderTree/win/DumpRenderTreeWin.h: Added declaration of cfStringRefToWString.
504
505         * DumpRenderTree/win/LayoutTestControllerWin.cpp: Moved cfStringRefToWString to
506         DumpRenderTree.cpp.
507
508 2011-02-27  Adam Roben  <aroben@apple.com>
509
510         Change the order of Chang Shu's email addresses so the first one is their Bugzilla username
511
512         This makes Bugzilla autocompletion work correctly.
513
514         * Scripts/webkitpy/common/config/committers.py:
515
516 2011-02-27  Patrick Gansterer  <paroga@webkit.org>
517
518         Reviewed by Andreas Kling.
519
520         [Qt] fast/loader/user-stylesheet-fast-path.html fails
521         https://bugs.webkit.org/show_bug.cgi?id=50911
522
523         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
524         (LayoutTestController::setUserStyleSheetLocation):
525         Use QUrl::fromEncoded since the passed string is already encoded.
526
527 2011-02-27  Adam Roben  <aroben@apple.com>
528
529         Ensure $testResultsDirectory is an absolute path before setting up Windows crash log saving
530
531         Windows 7 (and maybe Vista?) launches the post-mortem debugger with a working directory that
532         is different from old-run-webkit-test's. Because of this, the path to the crash log file
533         (which is relative to $testResultsDirectory) must be an absolute path. See the bug for more
534         details.
535
536         Fixes <http://webkit.org/b/55318> Crash logs aren't saved on Windows 7 when a relative
537         --results-directory path is used (like on the test slaves)
538
539         Reviewed by Dan Bernstein.
540
541         * Scripts/old-run-webkit-tests: Moved the call to setUpWindowsCrashLogSaving after we've
542         made $testResultsDirectory an absolute path.
543
544 2011-02-26  Dominic Mazzoni  <dmazzoni@google.com>
545
546         Reviewed by Adam Barth.
547
548         m_dumpAccessibilityNotifications should be initialized.
549         https://bugs.webkit.org/show_bug.cgi?id=54874
550
551         * DumpRenderTree/chromium/AccessibilityController.cpp:
552         (AccessibilityController::AccessibilityController):
553
554 2011-02-26  Vsevolod Vlasov  <vsevik@chromium.org>
555
556         Reviewed by Pavel Feldman.
557
558         DumpRenderTree should reset frame opener between tests.
559         https://bugs.webkit.org/show_bug.cgi?id=54874
560
561         Added clearOpener() calls between tests.
562         Some tests are using window.open() to change current window location.
563         This changes window.opener property whose nullity is checked before
564         showing xml viewer.
565
566         * DumpRenderTree/chromium/TestShell.cpp:
567         (TestShell::resetTestController):
568         * DumpRenderTree/gtk/DumpRenderTree.cpp:
569         (resetDefaultsToConsistentValues):
570         * DumpRenderTree/mac/DumpRenderTree.mm:
571         (resetWebViewToConsistentStateBeforeTesting):
572         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
573         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
574         * DumpRenderTree/win/DumpRenderTree.cpp:
575         (resetWebViewToConsistentStateBeforeTesting):
576         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
577         (WTR::InjectedBundlePage::reset):
578
579 2011-02-26  Yael Aharon  <yael.aharon@nokia.com>
580
581         Reviewed by Andreas Kling.
582
583         [Qt] Fix the focus in MiniBrowser.
584         https://bugs.webkit.org/show_bug.cgi?id=55288
585
586         Give focus to the web view when a load starts.
587
588         * MiniBrowser/qt/BrowserWindow.cpp:
589         (BrowserWindow::urlChanged):
590
591 2011-02-26  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
592
593         Reviewed by Adam Barth.
594
595         [EFL] Add coding style exceptions for EFL port
596         https://bugs.webkit.org/show_bug.cgi?id=54733
597
598         The EFL APIs use EFL naming style, which includes both lower-cased and camel-cased,
599         underscore-sparated values. But, style checking rule have found parameter_name errors
600         in efl directories so far.
601
602         * Scripts/webkitpy/style/checker.py: Add prameter_name rule's exception and efl directories.
603
604 2011-02-26  David Levin  <levin@chromium.org>
605
606         Reviewed by Darin Adler.
607
608         check-webkit-style should check for spaces after periods in comments
609         https://bugs.webkit.org/show_bug.cgi?id=55253
610
611         * Scripts/webkitpy/style/checkers/cpp.py: Added the check.
612         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added tests
613           and fixed the name of one related test.
614
615 2011-02-26  Kenneth Rohde Christiansen  <kenneth@webkit.org>
616
617         Reviewed by Andreas Kling.
618
619         Make it possible to test the targetdensity-dpi support
620         https://bugs.webkit.org/show_bug.cgi?id=55142
621
622         Test the viewport meta tag feature targetdensity-dpi by
623         adding extra arguments to dumpConfigurationForViewport
624
625         Add the new arguments to the shared header and all
626         implementation.
627
628         * DumpRenderTree/LayoutTestController.cpp:
629         (dumpConfigurationForViewportCallback):
630         * DumpRenderTree/LayoutTestController.h:
631         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
632         (LayoutTestController::dumpConfigurationForViewport):
633         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
634         (LayoutTestController::dumpConfigurationForViewport):
635         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
636         (LayoutTestController::dumpConfigurationForViewport):
637         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
638         (LayoutTestController::dumpConfigurationForViewport):
639         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
640         (LayoutTestController::dumpConfigurationForViewport):
641         * DumpRenderTree/qt/LayoutTestControllerQt.h:
642
643 2011-02-26  Sheriff Bot  <webkit.review.bot@gmail.com>
644
645         Unreviewed, rolling out r79764.
646         http://trac.webkit.org/changeset/79764
647         https://bugs.webkit.org/show_bug.cgi?id=55295
648
649         "broke Chromium builds" (Requested by rniwa on #webkit).
650
651         * DumpRenderTree/chromium/TestShell.cpp:
652         (TestShell::resetTestController):
653         * DumpRenderTree/gtk/DumpRenderTree.cpp:
654         (resetDefaultsToConsistentValues):
655         * DumpRenderTree/mac/DumpRenderTree.mm:
656         (resetWebViewToConsistentStateBeforeTesting):
657         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
658         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
659         * DumpRenderTree/win/DumpRenderTree.cpp:
660         (resetWebViewToConsistentStateBeforeTesting):
661         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
662         (WTR::InjectedBundlePage::reset):
663
664 2011-02-26  Yongjun Zhang  <yongjun_zhang@apple.com>
665
666         Reviewed by David Kilzer.
667
668         https://bugs.webkit.org/show_bug.cgi?id=48781
669
670         Add a resource load delegate method to query if WebCore should paint the default broken image for failed images.
671
672         Add a new resource load client method (shouldPaintBrokenImage).  WebKit client can decide if WebCore
673         should paint the default broken image when an image fails to load or decode.
674
675         * DumpRenderTree/LayoutTestController.cpp:
676         (LayoutTestController::LayoutTestController):
677         (setShouldPaintBrokenImageCallback):
678         (LayoutTestController::staticFunctions):
679         (LayoutTestController::setShouldPaintBrokenImage):
680         * DumpRenderTree/LayoutTestController.h:
681         (LayoutTestController::shouldPaintBrokenImage):
682         * DumpRenderTree/mac/ResourceLoadDelegate.mm:
683         (-[ResourceLoadDelegate webView:shouldPaintBrokenImageForURL:]):
684
685 2011-02-26  Vsevolod Vlasov  <vsevik@chromium.org>
686
687         Reviewed by Pavel Feldman.
688
689         DumpRenderTree should reset frame opener between tests.
690         https://bugs.webkit.org/show_bug.cgi?id=54874
691
692         Added clearOpener() calls between tests.
693         Some tests are using window.open() to change current window location.
694         This changes window.opener property whose nullity is checked before
695         showing xml viewer.
696
697         * DumpRenderTree/chromium/TestShell.cpp:
698         (TestShell::resetTestController):
699         * DumpRenderTree/gtk/DumpRenderTree.cpp:
700         (resetDefaultsToConsistentValues):
701         * DumpRenderTree/mac/DumpRenderTree.mm:
702         (resetWebViewToConsistentStateBeforeTesting):
703         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
704         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
705         * DumpRenderTree/win/DumpRenderTree.cpp:
706         (resetWebViewToConsistentStateBeforeTesting):
707         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
708         (WTR::InjectedBundlePage::reset):
709
710 2011-02-25  Lukasz Slachciak  <l.slachciak@samsung.com>
711
712         Reviewed by Kenneth Rohde Christiansen.
713
714         [EFL] Added EWebLauncher command line option to turn on/off frame flattening
715         https://bugs.webkit.org/show_bug.cgi?id=54809
716
717         * EWebLauncher/main.c: Handling -f option added.
718         (on_key_down):
719         (browserCreate):
720         (main):
721
722 2011-02-25  Sheriff Bot  <webkit.review.bot@gmail.com>
723
724         Unreviewed, rolling out r79324.
725         http://trac.webkit.org/changeset/79324
726         https://bugs.webkit.org/show_bug.cgi?id=55260
727
728         Leopard Debug is once again running out of address space while
729         linking (Requested by aroben on #webkit).
730
731         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
732         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
733
734 2011-02-25  Tony Chang  <tony@chromium.org>
735
736         Reviewed by James Robinson.
737
738         Clean up temp files left by crashing DRTs on chromium-win
739         https://bugs.webkit.org/show_bug.cgi?id=55182
740
741         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
742         * BuildSlaveSupport/chromium/remove-crash-logs: Added.
743
744 2011-02-25  Adam Roben  <aroben@apple.com>
745
746         Double-check that modifying the registry worked
747
748         On Windows Vista/7 with UAC enabled, regtool will fail to modify the registry, but it will
749         still return a successful exit code. So we double-check here that the value we tried to
750         write to the registry was really written.
751
752         Fixes <http://webkit.org/b/55227> old-run-webkit-tests falsely claims to be saving crash
753         logs on Windows Vista/7
754
755         Reviewed by Darin Adler.
756
757         * Scripts/old-run-webkit-tests:
758         (writeRegistryString): After trying to write the string value, check that it was really
759         written.
760
761 2011-02-25  Adam Roben  <aroben@apple.com>
762
763         Set the 32-bit post-mortem debugger on 64-bit OSes
764
765         The post-mortem debugger is controlled by some registry values in
766         HKLM\SOFTWARE\Microsoft\Windows NT\CurrentVersion\AeDebug. On 64-bit OSes, you can either
767         set the values in that key, which will be used for 64-bit apps, or in the equivalent key
768         under Wow6432Node, which will be used for 32-bit apps. Since DumpRenderTree and
769         WebKitTestRunner are 32-bit, we want to use the latter location.
770
771         Fixes <http://webkit.org/b/55225> old-run-webkit-tests sets the post-mortem debugger for
772         64-bit apps, but should set it for 32-bit apps instead
773
774         Reviewed by Darin Adler.
775
776         * Scripts/old-run-webkit-tests:
777         (readRegistryString):
778         (writeRegistryString):
779         Pass --wow32 to regtool so it will get/set values beneath Wow6432Node, which is where
780         Windows looks for the post-mortem debugger values when a 32-bit app crashes on a 64-bit OS.
781         Passing --wow32 on a 32-bit OS will not have any effect.
782
783 2011-02-25  Adam Roben  <aroben@apple.com>
784
785         Clean up code to set up crash log saving on Windows
786
787         Fixes <http://webkit.org/b/55224> old-run-webkit-test's Windows registry code is ugly
788
789         Reviewed by David Kilzer.
790
791         * Scripts/old-run-webkit-tests:
792         (readRegistryString):
793         (writeRegistryString):
794         Moved code to read strings from and write strings to the registry here...
795
796         (setUpWindowsCrashLogSaving): ...from here.
797
798 2011-02-25  Adam Roben  <aroben@apple.com>
799
800         Work around Cygwin's crash-suppression behavior
801
802         Cygwin calls ::SetErrorMode(SEM_FAILCRITICALERRORS), which any processes it launches will
803         inherit. This is bad for testing/debugging, as it causes the post-mortem debugger not to be
804         invoked. (Cygwin does this because it makes crashes more UNIX-y.) We reset the error mode
805         when our test apps launch to work around Cygwin's behavior.
806
807         Fixes <http://webkit.org/b/55222> Test apps crash silently (without invoking post-mortem
808         debugger) when launched from Cygwin 1.7
809
810         Reviewed by Darin Adler.
811
812         * DumpRenderTree/win/DumpRenderTree.cpp:
813         (main):
814         * TestWebKitAPI/win/main.cpp:
815         (main):
816         * WebKitAPITest/main.cpp:
817         (main):
818         * WebKitTestRunner/win/TestControllerWin.cpp:
819         (WTR::TestController::platformInitialize):
820         Call ::SetErrorMode(0) to undo Cygwin's folly.
821
822 2011-02-24  Jocelyn Turcotte  <jocelyn.turcotte@nokia.com>
823
824         Reviewed by Andreas Kling.
825
826         [Qt] Revert the support for QNAM affined to a different thread.
827         https://bugs.webkit.org/show_bug.cgi?id=55149
828
829         Qt 4.8 will have QNAM use its own thread internally by default,
830         no need to keep this complexity in WebKit.
831
832         This mainly reverts:
833         http://trac.webkit.org/changeset/73710
834         http://trac.webkit.org/changeset/73712
835
836         * QtTestBrowser/launcherwindow.cpp:
837         (LauncherWindow::~LauncherWindow):
838         (LauncherWindow::initializeView):
839         (LauncherWindow::createChrome):
840         * QtTestBrowser/launcherwindow.h:
841         (WindowOptions::WindowOptions):
842         * QtTestBrowser/webpage.cpp:
843         (WebPage::WebPage):
844         * QtTestBrowser/webpage.h:
845
846 2011-02-24  Adam Barth  <abarth@webkit.org>
847
848         Reviewed by Eric Seidel.
849
850         Add a script for creating gypi files
851         https://bugs.webkit.org/show_bug.cgi?id=55113
852
853         This is a trivial script to create gypi files from the source code in
854         the current directory.
855
856         * Scripts/make-gypi: Added.
857
858 2011-02-24  Sheriff Bot  <webkit.review.bot@gmail.com>
859
860         Unreviewed, rolling out r79628.
861         http://trac.webkit.org/changeset/79628
862         https://bugs.webkit.org/show_bug.cgi?id=55195
863
864         "broke chromium win tests" (Requested by dpranke on #webkit).
865
866         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
867         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
868
869 2011-02-24  Dirk Pranke  <dpranke@chromium.org>
870
871         Reviewed by Mihai Parparita.
872
873         rebaseline-chromium-webkit-tests can't handle GPU variants
874         https://bugs.webkit.org/show_bug.cgi?id=55002
875
876         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
877         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
878
879 2011-02-24  Dirk Pranke  <dpranke@chromium.org>
880
881         Reviewed by Mihai Parparita.
882
883         nrwt: log command used to start web server to --verbose
884
885         https://bugs.webkit.org/show_bug.cgi?id=55188
886
887         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
888         * Scripts/webkitpy/layout_tests/port/http_server.py:
889
890 2011-02-24  Victoria Kirst  <vrk@google.com>
891
892         Reviewed by Mihai Parparita.
893
894         [chromium] Move media-file.js and video-test.js back to media/ for UILayoutTests
895         https://bugs.webkit.org/show_bug.cgi?id=55089
896
897         This adds an alias into the media/ directory so that http/tests/media
898         tests can access the media resources when running in an httpd process.
899
900         * Scripts/webkitperl/httpd.pm:
901         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
902         * Scripts/webkitpy/layout_tests/port/http_server.py:
903
904 2011-02-24  Dirk Pranke  <dpranke@chromium.org>
905
906         Reviewed by Mihai Parparita.
907
908         nrwt: enable multiple processes by default on Chromium Win.
909
910         https://bugs.webkit.org/show_bug.cgi?id=55163
911
912         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
913         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
914
915 2011-02-24  Sam Weinig  <sam@webkit.org>
916
917         Attempt to fix Qt.
918
919         * WebKitTestRunner/TestInvocation.cpp:
920
921 2011-02-24  James Robinson  <jamesr@chromium.org>
922
923         Unreviewed, rolling out r79598.
924         http://trac.webkit.org/changeset/79598
925         https://bugs.webkit.org/show_bug.cgi?id=55089
926
927         Appears to be causing trouble with httpd on windows
928
929         * Scripts/webkitperl/httpd.pm:
930         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
931         * Scripts/webkitpy/layout_tests/port/http_server.py:
932
933 2011-02-24  Sam Weinig  <sam@webkit.org>
934
935         Reviewed by Darin Adler.
936
937         Add basic pixel testing support to WebKitTestRunner
938         https://bugs.webkit.org/show_bug.cgi?id=55161
939
940         * Scripts/old-run-webkit-tests:
941         Don't print color space information when using WebKitTestRunner since
942         we don't need to change the screen.
943
944         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
945         (WTR::InjectedBundle::initialize):
946         Add call to initializePlatformDefaults.
947         
948         (WTR::InjectedBundle::done):
949         Return a dictionary on completion with both the text output and pixel dump override.
950
951         * WebKitTestRunner/InjectedBundle/InjectedBundle.h:
952         Add declaration for initializePlatformDefaults.
953
954         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
955         (WTR::LayoutTestController::LayoutTestController):
956         (WTR::LayoutTestController::dumpAsText):
957         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
958         (WTR::LayoutTestController::shouldDumpPixels):
959         Make setting dumpAsText override the dump pixels preference.
960
961         * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm: Added.
962         (WTR::InjectedBundle::initializePlatformDefaults):
963         Set platform defaults matching DumpRenderTree.
964
965         * WebKitTestRunner/InjectedBundle/qt/InjectedBundleQt.cpp: Added.
966         (WTR::InjectedBundle::initializePlatformDefaults):
967         * WebKitTestRunner/InjectedBundle/win/InjectedBundleWin.cpp: Added.
968         (WTR::InjectedBundle::initializePlatformDefaults):
969         Add stubs for initializePlatformDefaults.
970
971         * WebKitTestRunner/TestController.cpp:
972         (WTR::TestController::runTest):
973         (WTR::TestController::runTestingServerLoop):
974         (WTR::TestController::run):
975         Parse expected pixel results out of stdin and setup the invocation with it.
976
977         * WebKitTestRunner/TestInvocation.cpp:
978         (WTR::TestInvocation::TestInvocation):
979         Initialize new members.
980
981         (WTR::TestInvocation::~TestInvocation):
982         (WTR::TestInvocation::setIsPixelTest):
983         (WTR::sizeWebViewForCurrentTest):
984         (WTR::TestInvocation::invoke):
985         (WTR::TestInvocation::dump):
986         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
987         * WebKitTestRunner/TestInvocation.h:
988         Dump pixels in addition to text by calling dumpPixelsAndCompareWithExpected.
989
990         * WebKitTestRunner/mac/PlatformWebViewMac.mm:
991         (WTR::PlatformWebView::PlatformWebView):
992         Use -[NSColorSpace genericRGBColorSpace] instead of the main screen color space.
993
994         * WebKitTestRunner/mac/TestInvocationMac.mm: Added.
995         (WTR::createCGContextFromPlatformView):
996         (WTR::computeMD5HashStringForContext):
997         (WTR::dumpBitmap):
998         (WTR::forceRepaintFunction):
999         (WTR::TestInvocation::dumpPixelsAndCompareWithExpected):
1000         Add mac specific pixel dumping support. Right now, we always pull pixels
1001         from the window server.
1002
1003         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
1004         * WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro:
1005         * WebKitTestRunner/qt/WebKitTestRunner.pro:
1006         * WebKitTestRunner/win/InjectedBundle.vcproj:
1007         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
1008         Add new files.
1009
1010 2011-02-24  Chang Shu  <cshu@webkit.org>
1011
1012         Reviewed by Kenneth Rohde Christiansen.
1013
1014         [Qt] Launch MiniBrowser if the test run is WebKit2. 
1015         https://bugs.webkit.org/show_bug.cgi?id=55145
1016
1017         * Scripts/old-run-webkit-tests:
1018         * Scripts/run-launcher:
1019         * Scripts/webkitdirs.pm:
1020
1021 2011-02-24  Victoria Kirst  <vrk@google.com>
1022
1023         Reviewed by Mihai Parparita.
1024
1025         [chromium] Move media-file.js and video-test.js back to media/ for UILayoutTests
1026         https://bugs.webkit.org/show_bug.cgi?id=55089
1027
1028         This adds an alias into the media/ directory so that http/tests/media
1029         tests can access the media resources when running in an httpd process.
1030
1031         * Scripts/webkitperl/httpd.pm:
1032         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1033         * Scripts/webkitpy/layout_tests/port/http_server.py:
1034
1035 2011-02-24  Andrew Wilson  <atwilson@chromium.org>
1036
1037         Unreviewed, rolling out r79570.
1038         http://trac.webkit.org/changeset/79570
1039         https://bugs.webkit.org/show_bug.cgi?id=54874
1040
1041         Breaks chromium build because glue/mocks/mock_web_frame.h/cc
1042         was not updated
1043
1044         * DumpRenderTree/chromium/TestShell.cpp:
1045         (TestShell::resetTestController):
1046         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1047         (resetDefaultsToConsistentValues):
1048         * DumpRenderTree/mac/DumpRenderTree.mm:
1049         (resetWebViewToConsistentStateBeforeTesting):
1050         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1051         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
1052         * DumpRenderTree/win/DumpRenderTree.cpp:
1053         (resetWebViewToConsistentStateBeforeTesting):
1054         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1055         (WTR::InjectedBundlePage::reset):
1056
1057 2011-02-24  Mahesh Kulkarni  <mahesh.kulkarni@nokia.com>
1058
1059         Reviewed by Laszlo Gombos.
1060
1061         [Qt] add android 2.2 user agent to QtTestBrowser
1062         https://bugs.webkit.org/show_bug.cgi?id=55085
1063
1064         Credit to Forrest Hodgkins to find right user-agent which works for youtube.
1065         * QtTestBrowser/useragentlist.txt:
1066
1067 2011-02-24  Adam Roben  <aroben@apple.com>
1068
1069         Change FrameLoadDelegate to support any number of delegates with delayed work to process
1070
1071         This makes our behavior match Mac more closely, and allows us to remove an incorrect
1072         assertion that was firing during some tests. (The assertion was claiming that there was
1073         never more than one delegate with delayed work to process, but that was not the case.)
1074
1075         Fixes <http://webkit.org/b/55146> Assertion failure in FrameLoadDelegate::locationChangeDone
1076         when running http/tests/navigation/back-twice-without-commit.html
1077
1078         Reviewed by Eric Carlson.
1079
1080         * DumpRenderTree/win/FrameLoadDelegate.cpp:
1081         (delegatesWithDelayedWork): Added. Returns all FrameLoadDelegates that have delayed work to
1082         process. A single delegate may appear in this Vector more than once (just as, on Mac, a
1083         single delegate may have multiple performSelector requests).
1084         (processWorkTimer): Pass the HWND to ::KillTimer, for pedantic brownie points. Added an
1085         assertion that the timer firing is the shared process work timer. Instead of using the
1086         single, global "delegate waiting for timer" delegate, give all delegates that have delayed
1087         work to process a chance to process their work.
1088         (FrameLoadDelegate::locationChangeDone): If we don't already have an active timer for
1089         processing delayed work, create one. Then add ourselves to the delegatesWithDelayedWork
1090         Vector so our processWork function will be called when the timer fires.
1091
1092 2011-02-24  Vsevolod Vlasov  <vsevik@chromium.org>
1093
1094         Reviewed by Alexey Proskuryakov.
1095
1096         DumpRenderTree should reset frame opener between tests.
1097         https://bugs.webkit.org/show_bug.cgi?id=54874
1098
1099         Added clearOpener() calls between tests.
1100         Some tests are using window.open() to change current window location.
1101         This changes window.opener property whose nullity is checked before
1102         showing xml viewer.
1103
1104         * DumpRenderTree/chromium/TestShell.cpp:
1105         (TestShell::resetTestController):
1106         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1107         (resetDefaultsToConsistentValues):
1108         * DumpRenderTree/mac/DumpRenderTree.mm:
1109         (resetWebViewToConsistentStateBeforeTesting):
1110         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1111         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
1112         * DumpRenderTree/win/DumpRenderTree.cpp:
1113         (resetWebViewToConsistentStateBeforeTesting):
1114         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1115         (WTR::InjectedBundlePage::reset):
1116
1117 2011-02-24  Adam Roben  <aroben@apple.com>
1118
1119         Don't trigger Windows builds for changes to WebCore.exp.in or the top-level GNUmakefile.am
1120
1121         Fixes <http://webkit.org/b/55144> Windows builders should not have built r79343 or r79440,
1122         but did
1123
1124         Reviewed by Anders Carlsson.
1125
1126         * Scripts/webkitpy/common/config/build.py:
1127         (_should_file_trigger_build): Loosened the pattern for GNUmakefile.am to include the
1128         top-level GNUmakefile.am, and loosened the pattern for .exp files to include WebCore.exp.in.
1129
1130         * Scripts/webkitpy/common/config/build_unittest.py:
1131         (ShouldBuildTest): Added some more tests.
1132
1133 2011-02-24  Andras Becsi  <abecsi@webkit.org>
1134
1135         Reviewed by Laszlo Gombos.
1136
1137         [Qt] MinGW build fails to link
1138         https://bugs.webkit.org/show_bug.cgi?id=55050
1139
1140         Prepend the libraries of subcomponents instead of appending them
1141         to fix the library order according to the dependency of the libraries
1142
1143         * WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro:
1144         * WebKitTestRunner/qt/WebKitTestRunner.pro:
1145
1146 2011-02-24  Gabor Loki  <loki@webkit.org>
1147
1148         Reviewed by Csaba Osztrogonác.
1149
1150         [Qt] Remove ARMv5 Release buildbot.
1151
1152         Remove ARMv5 Release buildbot, because there is very little difference
1153         between ARMv5 and ARMv7 Release bots, and there is also very little
1154         interest in ARMv5 build slave (compared to ARMv7).
1155
1156         * BuildSlaveSupport/build.webkit.org-config/config.json:
1157         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1158
1159 2011-02-22  Ojan Vafai  <ojan@chromium.org>
1160
1161         Reviewed by Tony Chang.
1162
1163         fix revision number finding when creating git patches
1164         https://bugs.webkit.org/show_bug.cgi?id=55024
1165
1166         HEAD~n does not walk merge commits correctly. git log however does.
1167
1168         * Scripts/webkitpy/common/checkout/scm.py:
1169         * Scripts/webkitpy/common/checkout/scm_unittest.py:
1170
1171 2011-02-23  Kenneth Russell  <kbr@google.com>
1172
1173         Unreviewed, rolling out r79387.
1174         https://bugs.webkit.org/show_bug.cgi?id=54885
1175
1176         Reapply the code changes from r79038 since they weren't the root
1177         cause of the test timeouts.
1178
1179         * DumpRenderTree/chromium/WebViewHost.cpp:
1180         (WebViewHost::reset):
1181
1182 2011-02-23  Siddharth Mathur  <siddharth.mathur@nokia.com>
1183
1184         Reviewed by Laszlo Gombos.
1185
1186         [Qt] Make sure Symbian binary UIDs are unique
1187
1188         * MiniBrowser/qt/MiniBrowser.pro: Resolve the UID collision with
1189         QtTestBrowser.pro by changing the UID.
1190
1191 2011-02-23  Patrick Gansterer  <paroga@webkit.org>
1192
1193         Reviewed by Darin Adler.
1194
1195         Rename PLATFORM(CF) to USE(CF)
1196         https://bugs.webkit.org/show_bug.cgi?id=53540
1197
1198         * DumpRenderTree/config.h:
1199
1200 2011-02-23  Dirk Pranke  <dpranke@chromium.org>
1201
1202         Reviewed by Ojan Vafai.
1203
1204         nrwt: turn on multiprocessing on chromium_linux, clean up unit tests for worker model defaults
1205         https://bugs.webkit.org/show_bug.cgi?id=55016
1206
1207         * Scripts/webkitpy/layout_tests/port/base.py:
1208         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1209         * Scripts/webkitpy/layout_tests/port/chromium_linux_unittest.py: Added.
1210         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1211         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
1212         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1213         * Scripts/webkitpy/layout_tests/port/mac.py:
1214         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1215         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
1216         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
1217
1218 2011-02-23  Tony Chang  <tony@chromium.org>
1219
1220         Reviewed by Kent Tamura.
1221
1222         [chromium] DumpRenderTree --check-layout-test-sys-deps leaks 4 temp dirs
1223         https://bugs.webkit.org/show_bug.cgi?id=55004
1224
1225         * DumpRenderTree/chromium/DumpRenderTree.cpp:
1226         (WebKitSupportTestEnvironment::WebKitSupportTestEnvironment):
1227         (WebKitSupportTestEnvironment::~WebKitSupportTestEnvironment):
1228         (main): Use a C++ object to ensure that TearDownTestEnvironment
1229             gets called when --check-layout-test-sys-deps is called
1230
1231 2011-02-23  Gopal Raghavan  <gopal.1.raghavan@nokia.com>
1232
1233         Reviewed by Anders Carlsson.
1234
1235         [Qt] [Qt] check webkit style fails in qt/tests
1236         https://bugs.webkit.org/show_bug.cgi?id=54998
1237
1238         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.
1239
1240         * Scripts/webkitpy/style/checker.py:
1241
1242 2011-02-22  Vsevolod Vlasov  <vsevik@chromium.org>
1243
1244         Reviewed by Pavel Feldman.
1245
1246         DumpRenderTree should dump as text if test path contains "dumpAsText/"
1247         https://bugs.webkit.org/show_bug.cgi?id=54873
1248
1249         * DumpRenderTree/chromium/LayoutTestController.h:
1250         (LayoutTestController::setShouldDumpAsText):
1251         (LayoutTestController::setShouldGeneratePixelResults):
1252         * DumpRenderTree/chromium/TestShell.cpp:
1253         (TestShell::runFileTest):
1254         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1255         (shouldDumpAsText):
1256         (dump):
1257         (runTest):
1258         * DumpRenderTree/mac/DumpRenderTree.mm:
1259         (shouldDumpAsText):
1260         (runTest):
1261         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1262         (WebCore::isDumpAsTextTest):
1263         (WebCore::DumpRenderTree::open):
1264         * DumpRenderTree/win/DumpRenderTree.cpp:
1265         (shouldDumpAsText):
1266         (runTest):
1267         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1268         (WTR::InjectedBundlePage::dump):
1269
1270 2011-02-22  Kenneth Russell  <kbr@google.com>
1271
1272         Reviewed by James Robinson.
1273
1274         Many tests started timing out on the chrome mac canaries around r79035
1275         https://bugs.webkit.org/show_bug.cgi?id=54885
1276
1277         Revert the code changes, but not the Chromium DEPS roll, from
1278         r79038 to see whether this clears up the timing-out tests. Was not
1279         able to reproduce the timeouts locally with a Release mode build
1280         of DRT (on Snow Leopard).
1281
1282         * DumpRenderTree/chromium/WebViewHost.cpp:
1283         (WebViewHost::reset):
1284
1285 2011-02-22  Dirk Pranke  <dpranke@chromium.org>
1286
1287         Reviewed by James Robinson.
1288
1289         nrwt: --platform chromium-gpu doesn't work with --worker-model=processes
1290         https://bugs.webkit.org/show_bug.cgi?id=55009
1291
1292         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
1293
1294 2011-02-18  Mihai Parparita  <mihaip@chromium.org>
1295
1296         Reviewed by Ojan Vafai.
1297
1298         Default to 500 for --exit-after-n-failures
1299         https://bugs.webkit.org/show_bug.cgi?id=54773
1300
1301         Make 500 the default value for --exit-after-n-failures (ORWT bots
1302         already default to this as of r75726). Similarly, change the default for
1303         --exit-after-n-crashes-or-timeouts to 20.
1304
1305         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1306
1307 2011-02-02  Martin Robinson  <mrobinson@igalia.com>
1308
1309         Reviewed by Xan Lopez.
1310
1311         [GTK] DRT needs an implementation of LayoutTestController.setSelectTrailingWhitespace
1312         https://bugs.webkit.org/show_bug.cgi?id=53603
1313
1314         Add support for LayoutTestController.setSelectTrailingWhitespace by calling through
1315         to DumpRenderTreeSupportGtk for this functionality.
1316
1317         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1318         (resetDefaultsToConsistentValues): Always call LayoutTestController.setSelectTrailingWhitespace to
1319         false when reseting values between tests.
1320         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1321         (LayoutTestController::setSelectTrailingWhitespaceEnabled): Call LayoutTestController.setSelectTrailingWhitespace
1322         with the appropriate value.
1323
1324 2011-02-22  Adam Roben  <aroben@apple.com>
1325
1326         Don't trigger a Windows build when the top-level ChangeLog changes
1327
1328         r79320 is an example of a change that shouldn't have built on Windows, but did, because of
1329         this bug.
1330
1331         Fixes <http://webkit.org/b/54957> Windows bots build when top-level ChangeLog changes, but
1332         shouldn't
1333
1334         Reviewed by Anders Carlsson.
1335
1336         * Scripts/webkitpy/common/config/build.py:
1337         (_should_file_trigger_build): Modified the ChangeLog pattern to match the top-level
1338         ChangeLog, and re-sorted it in the list.
1339
1340         * Scripts/webkitpy/common/config/build_unittest.py:
1341         (ShouldBuildTest): Added a test that ChangeLogs don't trigger builds.
1342
1343 2011-02-22  Sheriff Bot  <webkit.review.bot@gmail.com>
1344
1345         Unreviewed, rolling out r78691.
1346         http://trac.webkit.org/changeset/78691
1347         https://bugs.webkit.org/show_bug.cgi?id=54956
1348
1349         Leopard Debug is successfully building again (Requested by
1350         aroben on #webkit).
1351
1352         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
1353         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1354
1355 2011-02-22  Ilya Tikhonovsky  <loislo@chromium.org>
1356
1357         Reviewed by Pavel Feldman.
1358
1359         Web Inspector: flakyness of inspector tests.
1360         https://bugs.webkit.org/show_bug.cgi?id=54729
1361
1362         As far as we have the protocol with sequence numbers we can simplify test harness support
1363         and drop out chomium specific methods.
1364
1365         * DumpRenderTree/DumpRenderTree.gypi:
1366         * DumpRenderTree/chromium/DRTDevToolsAgent.cpp:
1367         (DRTDevToolsAgent::sendMessageToInspectorFrontend):
1368         (DRTDevToolsAgent::runtimePropertyChanged):
1369         (DRTDevToolsAgent::asyncCall):
1370         (DRTDevToolsAgent::call):
1371         * DumpRenderTree/chromium/DRTDevToolsAgent.h:
1372         * DumpRenderTree/chromium/DRTDevToolsCallArgs.cpp: Removed.
1373         * DumpRenderTree/chromium/DRTDevToolsCallArgs.h: Removed.
1374         * DumpRenderTree/chromium/DRTDevToolsClient.cpp:
1375         (DRTDevToolsClient::sendMessageToBackend):
1376         (DRTDevToolsClient::asyncCall):
1377         (DRTDevToolsClient::call):
1378         * DumpRenderTree/chromium/DRTDevToolsClient.h:
1379
1380 2011-02-22  Andras Becsi  <abecsi@webkit.org>
1381
1382         Reviewed by Laszlo Gombos.
1383
1384         [Qt] Redesign the build system
1385         https://bugs.webkit.org/show_bug.cgi?id=51339
1386
1387         Part 2.
1388
1389         Build WebCore as a static library, compile the WebKit API and WebKit2 API
1390         in a final step and link to WebKit2, WebCore and JSC libraries to fix
1391         linking issues resulting from stripped away symbols.
1392
1393         * WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro: Remove superfluous includepaths.
1394
1395 2011-02-21  Lukasz Slachciak  <l.slachciak@samsung.com>
1396
1397         Reviewed by Antonio Gomes.
1398
1399         [EFL] Restored spatial navigation option in EWebLauncher since #18662 bug is closed
1400         https://bugs.webkit.org/show_bug.cgi?id=54806
1401
1402         * EWebLauncher/main.c:
1403         (on_key_down): calling ewk_view_setting_spatial_navigation_set when "F12" pressed added.
1404
1405 2011-02-21  Dirk Pranke  <dpranke@chromium.org>
1406
1407         Reviewed by Kent Tamura.
1408
1409         nrwt: revert behavior on 10.5 to "old-threads". It's possible
1410         that r79062 has introduced a bunch of flakiness, although I
1411         can't explain why right now. This change should restore the
1412         old behavior, and then we can look at the build logs to see
1413         if the tests stop timing out or being flaky.
1414
1415         https://bugs.webkit.org/show_bug.cgi?id=54925
1416
1417         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1418         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
1419         * Scripts/webkitpy/layout_tests/port/mac.py:
1420         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1421
1422 2011-02-21  Dirk Pranke  <dpranke@chromium.org>
1423
1424         Reviewed by Dimitri Glazkov.
1425
1426         nrwt: fix typo causing us to try and repeatedly shut down the servers
1427         https://bugs.webkit.org/show_bug.cgi?id=54904
1428
1429         * Scripts/webkitpy/layout_tests/layout_package/worker_mixin.py:
1430
1431 2011-02-21  Alejandro G. Castro  <alex@igalia.com>
1432
1433         Reviewed by Csaba Osztrogonác.
1434
1435         REGRESSION(r79157):
1436         plugins/get-url-with-javascript-destroying-plugin.html make the
1437         next test crash
1438         https://bugs.webkit.org/show_bug.cgi?id=54863
1439
1440         * GNUmakefile.am:
1441
1442 2011-02-20  Dirk Pranke  <dpranke@chromium.org>
1443
1444         Reviewed by Ojan Vafai.
1445
1446         test-webkitpy: stop skipping a bunch of tests on win32
1447         https://bugs.webkit.org/show_bug.cgi?id=54788
1448
1449         * Scripts/webkitpy/common/net/testoutputset_unittest.py:
1450         * Scripts/webkitpy/test/main.py:
1451
1452 2011-02-20  Dirk Pranke  <dpranke@chromium.org>
1453
1454         Reviewed by Ojan Vafai.
1455
1456         Fix various bugs related to handling of dir separators and real
1457         vs. fake filesystems that were causing test-webkitpy to fail
1458         tests on win32. These bugs have a variety of causes but are
1459         lumped together because they should be easy to review.
1460
1461         https://bugs.webkit.org/show_bug.cgi?id=54700
1462
1463         * Scripts/webkitpy/common/net/testoutput.py:
1464           Here we were using os.path.sep but not escaping it prior to
1465           using in a regexp, which  caused bad things on win32.
1466         * Scripts/webkitpy/common/net/testoutput_unittest.py:
1467           Same here.
1468         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
1469           Here we were using the native filesystem instead of the mock
1470           filesystem, making some tests unnecessarily platform dependent.
1471           Also we weren't setting the default configuration for the
1472           TestChromiumWinPort.
1473         * Scripts/webkitpy/layout_tests/update_webgl_conformance_tests.py:
1474           Here we *shouldn't* use os.path.join(), just so the code becomes
1475           less platform dependent.
1476         * Scripts/webkitpy/style/checker.py:
1477           Here we should be using os.path.sep instead of "/"; the tests
1478           require the native filesystem.
1479         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
1480           Here the underlying code is using os.path.sep, so we probably
1481           should as well.
1482         * Scripts/webkitpy/style/test_expectations_unittest.py:
1483           Here we should be using the filesystem hanging of the port
1484           object (which is the mocked filesystem) rather than the native
1485           one.
1486
1487 2011-02-20  Anders Carlsson  <andersca@apple.com>
1488
1489         Reviewed by Maciej Stachowiak.
1490
1491         Crash when a plug-in requests a javascript: url that destroys the plug-in
1492         https://bugs.webkit.org/show_bug.cgi?id=54837
1493         <rdar://problem/9005475>
1494
1495         Add new plug-in test.
1496
1497         * DumpRenderTree/DumpRenderTree.gypi:
1498         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
1499         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
1500         (PluginTest::NPN_GetURL):
1501         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
1502         * DumpRenderTree/TestNetscapePlugIn/Tests/GetURLWithJavaScriptURLDestroyingPlugin.cpp: Added.
1503         (GetURLWithJavaScriptURLDestroyingPlugin::GetURLWithJavaScriptURLDestroyingPlugin):
1504         (GetURLWithJavaScriptURLDestroyingPlugin::NPP_New):
1505         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1506         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
1507
1508 2011-02-19  Patrick Gansterer  <paroga@webkit.org>
1509
1510         Reviewed by Andreas Kling.
1511
1512         Make WinCE a core builder
1513         https://bugs.webkit.org/show_bug.cgi?id=54801
1514
1515         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
1516         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1517
1518 2011-02-19  Lukasz Slachciak  <l.slachciak@samsung.com>
1519
1520         Reviewed by Andreas Kling.
1521
1522         Fixed problem with launching EWebLauncher with "run-launcher --efl". Incorrect path referring to non-existing .libs directory ws set.
1523         https://webkit.org/b/54778
1524
1525         * Scripts/webkitdirs.pm:
1526
1527 2011-02-18  David Levin  <levin@chromium.org>
1528
1529         Reviewed by Eric Seidel.
1530
1531         check-webkit-style: Misses brace style error when the line with the { has a } in it.
1532         https://bugs.webkit.org/show_bug.cgi?id=54769
1533
1534         * Scripts/webkitpy/style/checkers/cpp.py: Fixed the check for the close brace
1535           to only look after the last open brace, so that the open brace in this line
1536           "} else {" will still be able to trigger the error.
1537         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added related unit tests.
1538
1539 2011-02-18  Zan Dobersek  <zandobersek@gmail.com>
1540
1541         Reviewed by Martin Robinson.
1542
1543         [Gtk] Re-enable meter tag support
1544         https://bugs.webkit.org/show_bug.cgi?id=54762
1545
1546         Meter tag is supported on the Gtk port, enable it by default.
1547
1548         * Scripts/build-webkit:
1549
1550 2011-02-18  David Levin  <levin@chromium.org>
1551
1552         Reviewed by Eric Seidel.
1553
1554         check-webkit-style falsely complains about WebKitGTK+ public headers
1555         https://bugs.webkit.org/show_bug.cgi?id=54650
1556
1557         * Scripts/webkitpy/style/checker.py: Add some exceptions for the gtk files.
1558         * Scripts/webkitpy/style/checkers/cpp.py: Restrict the checks for WEBKIT_API
1559           to the chromium directory (and improved the checks slightly).
1560         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added corresponding tests.
1561
1562 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1563
1564         Reviewed by James Robinson.
1565
1566         new-run-webkit-tests: enable multiple processes by default on
1567         mac. This change removes the artificial restrictions we placed
1568         on the # of child processes we used with the old threading
1569         model, and switches to the new message-based model and multiple
1570         processes, where available. If multiple processes are not
1571         available (Leopard / Python 2.5), then we use the 'inline' model
1572         instead of the 'old-threads' model on the 'mac' port or one
1573         process on the 'chromium-mac' port. We need additional testing
1574         to see if the new 'threads' model works reliably and is worth
1575         supporting, or if we should just live with things being slightly
1576         slow.
1577
1578         https://bugs.webkit.org/show_bug.cgi?id=54596
1579
1580         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1581         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
1582         * Scripts/webkitpy/layout_tests/port/mac.py:
1583         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1584         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1585
1586 2011-02-18  Andrew Wilson  <atwilson@chromium.org>
1587
1588         Unreviewed, rolling out r79047.
1589         http://trac.webkit.org/changeset/79047
1590         https://bugs.webkit.org/show_bug.cgi?id=54596
1591
1592         Broke canary bots - please watch the canaries the next time
1593         you land this
1594
1595         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1596         * Scripts/webkitpy/layout_tests/port/mac.py:
1597
1598 2011-02-18  Emil A Eklund  <eae@chromium.org>
1599
1600         Reviewed by Darin Adler.
1601
1602         Crash in EventHandler::sendContextMenuEventForKey
1603         https://bugs.webkit.org/show_bug.cgi?id=54495
1604
1605         Add support for the context menu key (VK_APPS) to EventSender for
1606         platforms that support the key.
1607
1608         * DumpRenderTree/chromium/EventSender.cpp:
1609         (EventSender::keyDown):
1610         * DumpRenderTree/gtk/EventSender.cpp:
1611         (keyDownCallback):
1612         * DumpRenderTree/qt/EventSenderQt.cpp:
1613         (EventSender::keyDown):
1614         * DumpRenderTree/win/EventSender.cpp:
1615         (keyDownCallback):
1616
1617 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1618
1619         Reviewed by Ojan Vafai.
1620
1621         new-run-webkit-tests: enable multiple processes by default on mac
1622         https://bugs.webkit.org/show_bug.cgi?id=54596
1623
1624         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1625         * Scripts/webkitpy/layout_tests/port/mac.py:
1626
1627 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1628
1629         Reviewed by Adam Roben.
1630
1631         Skip pretty patch unit tests if ruby isn't installed.
1632
1633         https://bugs.webkit.org/show_bug.cgi?id=54699
1634
1635         * Scripts/webkitpy/common/prettypatch_unittest.py:
1636
1637 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1638
1639         Reviewed by Adam Roben.
1640
1641         Skip mac.test_skipped_file_paths() when running on win32; the
1642         test has hardcoded '/' directory separators and since it's a
1643         port-specific test and we have coverage on other ports it's not
1644         worth it to make the test more generic.
1645
1646         https://bugs.webkit.org/show_bug.cgi?id=54698
1647
1648         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1649
1650 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1651
1652         Reviewed by Adam Roben.
1653
1654         Fix webkitpy.layout_tests.port.config_unittest to work on Win32.
1655
1656         https://bugs.webkit.org/show_bug.cgi?id=54703
1657
1658         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1659
1660 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1661
1662         Reviewed by Adam Roben.
1663
1664         Change the default port we're mocking out on win32 when we
1665         run mock_drt_unittest. Normally we'd default to the
1666         port-specific default, but that would be "win", which doesn't
1667         currently work.
1668
1669         https://bugs.webkit.org/show_bug.cgi?id=54708
1670
1671         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
1672
1673 2011-02-18  Kenneth Russell  <kbr@google.com>
1674
1675         Reviewed by James Robinson.
1676
1677         [chromium] Clean up initialization of minimum timer interval
1678         https://bugs.webkit.org/show_bug.cgi?id=54772
1679
1680         Use new GetForegroundTabTimerInterval function in webkit_support
1681         to reset the minimum timer interval before each test.
1682
1683         * DumpRenderTree/chromium/WebViewHost.cpp:
1684         (WebViewHost::reset):
1685
1686 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1687
1688         Reviewed by Adam Roben.
1689
1690         r78494 fixed the chromium_win unittest failing under win32,
1691         but not cygwin. This fixes cygwin as well.
1692
1693         https://bugs.webkit.org/show_bug.cgi?id=54652
1694
1695         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1696
1697 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1698
1699         Reviewed by Ojan Vafai, Mihai Parparita.
1700
1701         new-run-webkit-tests: r78522 made it impossible to create a
1702         'chromium-win' or 'chromium-mac' port; they would always get the
1703         version tacked on to the end. It turns out that we actually need
1704         to be able to create these ports in order for
1705         rebaseline-chromium-webkit-tests to work correctly given the
1706         broken coverage of our layout bots. As soon as we get SL and
1707         Win 7 bots running reliably on the canaries, we can get rid of
1708         this change. Note that we only need to change the chromium_mac
1709         and win ports because linux doesn't yet use versions and that
1710         script doesn't work with non-chromium ports.
1711
1712         Also, this change fixes a bug where we were using
1713         "platform/chromium-win-win7" and
1714         "platform/chromium-mac-snowleopard" as the baseline dirs for
1715         Win 7 and SL; we should've been using the generic directories
1716         instead since they are the newest versions of the os's.
1717
1718         https://bugs.webkit.org/show_bug.cgi?id=54691
1719
1720         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1721         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
1722         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1723         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1724         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
1725
1726 2011-02-16  MORITA Hajime  <morrita@google.com>
1727
1728         Reviewed by Kent Tamura.
1729
1730         [Chromium][DRT] some PostDelayedTask() accesses possibly dangling this pointer.
1731         http://webkit.org/b/53899
1732
1733         Replaced static invocation wrappers with MethodTask objects,
1734         which are removed on WebViewHost destructor.
1735         
1736         * DumpRenderTree/chromium/WebViewHost.cpp:
1737         (WebViewHost::requestCheckingOfText):
1738         (WebViewHost::scheduleAnimation):
1739         (WebViewHost::closeWidgetSoon):
1740         * DumpRenderTree/chromium/WebViewHost.h:
1741         (WebViewHost::HostMethodTask::HostMethodTask):
1742         (WebViewHost::HostMethodTask::runIfValid):
1743         (WebViewHost::taskList):
1744
1745 2011-02-07  Hayato Ito  <hayato@chromium.org>
1746
1747         Reviewed by Ojan Vafai.
1748
1749         Get rid of code which writes test results from test_type's
1750         compare_output() method.
1751         Writing test results now happens in single_test_runner, using
1752         newly introduced test_result_writer module.
1753
1754         Subsequent patches will eliminate test_type/* classes.
1755
1756         https://bugs.webkit.org/show_bug.cgi?id=53908
1757
1758         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
1759         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py: Added.
1760         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
1761         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
1762         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
1763
1764 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
1765
1766         Reviewed by Adam Roben.
1767
1768         Add blacklist of remaining webkitpy/* modules that fail on
1769         Win32, so that test-webkitpy will at least run cleanly.
1770
1771         https://bugs.webkit.org/show_bug.cgi?id=54709
1772
1773         * Scripts/webkitpy/layout_tests/test/main.py:
1774
1775 2011-02-17  Simon Fraser  <simon.fraser@apple.com>
1776
1777         Reviewed by Sam Weinig.
1778
1779         Set a preference to ensure that scrollbars are in a predictable
1780         state for DRT.
1781
1782         * DumpRenderTree/mac/DumpRenderTree.mm:
1783         (resetDefaultsToConsistentValues):
1784
1785 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
1786
1787         Reviewed by Adam Roben.
1788
1789         test-webkitpy doesn't work with the multiprocessing module on win32
1790         or cygwin, so we skip it for now.
1791
1792         https://bugs.webkit.org/show_bug.cgi?id=54520
1793
1794         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
1795         * Scripts/webkitpy/layout_tests/layout_package/run_webkit_tests_unittest.py:
1796
1797 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
1798
1799         Reviewed by Adam Roben.
1800
1801         Fix version, baseline_search_path for chromium-win-win7. Also
1802         fix chromium_win_unittests, and fix port_testcase.make_port()
1803         to correctly use the options passed into it and only set
1804         results_directory() if necessary.
1805
1806         https://bugs.webkit.org/show_bug.cgi?id=54652
1807
1808         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1809         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1810         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
1811
1812 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
1813
1814         Reviewed by Ojan Vafai.
1815
1816         Fix a hang in new-run-webkit-tests when run with
1817         --dry-run --worker-model=processes
1818
1819         https://bugs.webkit.org/show_bug.cgi?id=54595
1820
1821         * Scripts/webkitpy/layout_tests/layout_package/test_runner2.py:
1822         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1823
1824 2011-02-17  Kevin Ollivier  <kevino@theolliviers.com>
1825
1826         [wx] Build fixes after recent changes.
1827
1828         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
1829         (LayoutTestController::setMinimumTimerInterval):
1830         * wx/build/settings.py:
1831
1832 2011-02-17  Csaba Osztrogonác  <ossy@webkit.org>
1833
1834         Reviewed by Andreas Kling.
1835
1836         [Qt] ImageDiff sometimes hangs
1837         https://bugs.webkit.org/show_bug.cgi?id=54641
1838
1839         * DumpRenderTree/qt/ImageDiff.cpp: Clear actualImage and baselineImage after all comparison.
1840         (main):
1841
1842 2011-02-17  Peter Varga  <pvarga@webkit.org>
1843
1844         Reviewed by Csaba Osztrogonác.
1845
1846         The run-sunspider --v8-suite doesn't work
1847         https://bugs.webkit.org/show_bug.cgi?id=54660
1848
1849         * Scripts/run-sunspider:
1850
1851 2011-02-17  Gabor Rapcsanyi  <rgabor@webkit.org>
1852
1853         Reviewed by Csaba Osztrogonác.
1854
1855         [NRWT] Fix environment setup on Qt port.
1856         https://bugs.webkit.org/show_bug.cgi?id=54549
1857
1858         * Scripts/webkitpy/layout_tests/port/qt.py:
1859
1860 2011-02-16  Yael Aharon  <yael.aharon@nokia.com>
1861
1862         Reviewed by Andreas Kling.
1863
1864         [Qt] Enable tiled backing store by default in MiniBrowser.
1865         https://bugs.webkit.org/show_bug.cgi?id=54581
1866
1867         Instead of a flag to turn it on, a new flag is defined to turn
1868         tiled backing store off.
1869
1870         * MiniBrowser/qt/MiniBrowserApplication.cpp:
1871         (MiniBrowserApplication::handleUserOptions):
1872         * MiniBrowser/qt/MiniBrowserApplication.h:
1873         (WindowOptions::WindowOptions):
1874
1875 2011-02-16  Eric Seidel  <eric@webkit.org>
1876
1877         Reviewed by Mihai Parparita.
1878
1879         flaky test archive uploads are always 22 bytes long
1880         https://bugs.webkit.org/show_bug.cgi?id=54593
1881
1882         Speculative fix for the 22-bytes bug.
1883
1884         * Scripts/webkitpy/tool/bot/flakytestreporter.py:
1885
1886 2011-02-16  Anna Cavender  <annacc@chromium.org>
1887
1888         Reviewed by Mihai Parparita.
1889
1890         http/tests/media should be served over HTTP (not from local file)
1891         https://bugs.webkit.org/show_bug.cgi?id=54028
1892
1893         Removed special file:// treatment of http/tests/media
1894
1895         * Scripts/old-run-webkit-tests:
1896         * Scripts/webkitpy/layout_tests/port/base.py:
1897
1898 2011-02-16  Mihai Parparita  <mihaip@chromium.org>
1899
1900         Reviewed by Ojan Vafai.
1901
1902         ChangeLog.parse_latest_entry_from_file does not handle rolled over ChangeLogs
1903         https://bugs.webkit.org/show_bug.cgi?id=54609
1904         
1905         r78737 rolled over ChangeLogs, and webkit-patch complains when landing
1906         patches such as this one that are the first to add an entry to the
1907         ChangeLog (since they don't see a second date line to know where the
1908         entry ends).
1909
1910         * Scripts/webkitpy/common/checkout/changelog.py:
1911         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
1912
1913 == Rolled over to ChangeLog-2011-02-16 ==