9932d880cfda70428121e5a6aeca19739439290b
[WebKit-https.git] / Tools / ChangeLog
1 2011-03-02  Sheriff Bot  <webkit.review.bot@gmail.com>
2
3         Unreviewed, rolling out r80139.
4         http://trac.webkit.org/changeset/80139
5         https://bugs.webkit.org/show_bug.cgi?id=55620
6
7         failing and crashing tests on Snow Leopard bot (Requested by
8         mihaip1 on #webkit).
9
10         * DumpRenderTree/LayoutTestController.cpp:
11         (LayoutTestController::staticFunctions):
12         * DumpRenderTree/LayoutTestController.h:
13         * DumpRenderTree/chromium/LayoutTestController.cpp:
14         (LayoutTestController::LayoutTestController):
15         * DumpRenderTree/chromium/LayoutTestController.h:
16         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
17         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
18         * DumpRenderTree/mac/UIDelegate.h:
19         * DumpRenderTree/mac/UIDelegate.mm:
20         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
21         * DumpRenderTree/qt/LayoutTestControllerQt.h:
22         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
23         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
24
25 2011-03-02  Adam Roben  <aroben@apple.com>
26
27         Wait for force a paint in WebKitTestRunner until we're actually ready to dump the output
28
29         This matches DumpRenderTree on Windows.
30
31         Fixes <http://webkit.org/b/55469> <rdar://problem/9068539> REGRESSION (r79863): Lots of
32         dom/html/level2/html/HTMLFrameElement*.html tests crashing in FrameView::paintOverhangAreas
33         in WebKit2
34
35         Reviewed by Beth Dakin.
36
37         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
38         (WTR::InjectedBundlePage::dump): Moved code to force a paint here...
39         (WTR::InjectedBundlePage::didFinishLoadForFrame): ...from here.
40
41 2011-03-02  Daniel Cheng  <dcheng@chromium.org>
42
43         Reviewed by David Levin.
44
45         Add feature define for data transfer items
46         https://bugs.webkit.org/show_bug.cgi?id=55510
47
48         * Scripts/build-webkit:
49
50 2011-03-02  Adam Roben  <aroben@apple.com>
51
52         Look in the new-to-SnowLeopard location for crash logs on SnowLeopard and newer
53
54         ReportCrash saves logs to ~/Library/Logs/CrashReporter on Leopard, but to
55         ~/Library/Logs/DiagnosticReports on SnowLeopard. old-run-webkit-tests was only looking in
56         the former location, but was getting lucky on some SnowLeopard machines because symlinks
57         were being created there pointing into DiagnosticReports. For machines without these
58         symlinks, crash logs were not getting captured.
59
60         Fixes <http://webkit.org/b/55607> old-run-webkit-tests doesn't capture crash logs on some
61         SnowLeopard machines
62
63         Reviewed by Eric Seidel.
64
65         * Scripts/old-run-webkit-tests:
66         (captureSavedCrashLog): Look in ~/Library/Logs/DiagnosticReports on SnowLeopard and newer.
67
68 2011-03-02  Adam Roben  <aroben@apple.com>
69
70         Don't try to capture crash logs on platforms that haven't implemented that feature
71
72         Fixes <http://webkit.org/b/55504> Use of uninitialized value in numeric lt (<) at
73         Tools/Scripts/old-run-webkit-tests line 1778 seen on Qt bots after a crash
74
75         Reviewed by Joseph Pecoraro.
76
77         * Scripts/old-run-webkit-tests:
78         (captureSavedCrashLog): Bail if we don't have a crash log glob.
79         (findNewestFileMatchingGlob): Fix a buggy test for when no paths match the glob. This isn't
80         strictly required to fix bug 55504, but seems worthwhile.
81
82 2011-01-26  John Knottenbelt  <jknotten@chromium.org>
83
84         Reviewed by Dmitry Titov.
85
86         Detach Geolocation from Frame when Page destroyed.
87         https://bugs.webkit.org/show_bug.cgi?id=52877
88
89         Extend the layout test controller to expose the number of pending
90         geolocation requests, so that we can test that the requests have
91         been cancelled on page close.
92
93         * DumpRenderTree/LayoutTestController.cpp:
94         (numberOfPendingGeolocationPermissionRequestsCallback):
95         (LayoutTestController::staticFunctions):
96         * DumpRenderTree/LayoutTestController.h:
97         * DumpRenderTree/chromium/LayoutTestController.cpp:
98         (LayoutTestController::LayoutTestController):
99         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
100         * DumpRenderTree/chromium/LayoutTestController.h:
101         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
102         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
103         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
104         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
105         * DumpRenderTree/mac/UIDelegate.h:
106         * DumpRenderTree/mac/UIDelegate.mm:
107         (-[UIDelegate numberOfPendingGeolocationPermissionRequests]):
108         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
109         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
110         * DumpRenderTree/qt/LayoutTestControllerQt.h:
111         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
112         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
113         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
114         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
115
116
117 2011-03-01  Kent Tamura  <tkent@chromium.org>
118
119         Reviewed by Dimitri Glazkov.
120
121         Assertion fails when a form validation bubble appears
122         https://bugs.webkit.org/show_bug.cgi?id=55550
123
124         Test: fast/forms/interactive-validation-attach-assertion.html
125
126         * DumpRenderTree/chromium/WebPreferences.cpp:
127         (WebPreferences::applyTo): Enable the interactive validation feature by
128           default. This matches to Chromium browser.
129
130 2011-03-01  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
131
132         Reviewed by Kent Tamura.
133
134         Remove efl port style exception of Webcore/JavaScriptCore
135         https://bugs.webkit.org/show_bug.cgi?id=55549
136
137         WebCore and JavaScriptCore need to adhere WebKit coding style.
138
139         * Scripts/webkitpy/style/checker.py:
140
141 2011-03-01  Tony Chang  <tony@chromium.org>
142
143         Reviewed by Ojan Vafai.
144
145         [chromium] clean up temp files on chromium-mac builder too
146         https://bugs.webkit.org/show_bug.cgi?id=55525
147
148         * BuildSlaveSupport/chromium/remove-crash-logs:
149
150 2011-02-28  Ojan Vafai  <ojan@chromium.org>
151
152         Reviewed by Tony Chang.
153
154         wrap json in a function call to afford cross-domain loading
155         https://bugs.webkit.org/show_bug.cgi?id=55353
156
157         Also delete the code that compacts the JSON. It turns out this
158         breaks the rebaseline tool and is probably a premature optimization anyways.
159
160         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
161         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
162         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
163         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
164
165 2011-03-01  Sheriff Bot  <webkit.review.bot@gmail.com>
166
167         Unreviewed, rolling out r80079.
168         http://trac.webkit.org/changeset/80079
169         https://bugs.webkit.org/show_bug.cgi?id=55547
170
171         "Broke the Win debug build?" (Requested by dcheng on #webkit).
172
173         * Scripts/build-webkit:
174
175 2011-03-01  Mark Rowe  <mrowe@apple.com>
176
177         Reviewed by Sam Weinig.
178
179         Update verification scripts to handle WebKit2.
180
181         * Scripts/check-for-global-initializers: Skip object files that use RefCountedLeakChecker.
182         * Scripts/check-for-inappropriate-files-in-framework: Add some inappropriate file types for
183         WebKit2.
184
185 2011-03-01  Daniel Cheng  <dcheng@chromium.org>
186
187         Reviewed by David Levin.
188
189         Add feature define for data transfer items
190         https://bugs.webkit.org/show_bug.cgi?id=55510
191
192         * Scripts/build-webkit:
193
194 2011-02-23  Joseph Pecoraro  <joepeck@webkit.org>
195
196         Reviewed by Timothy Hatcher.
197
198         All Console Messages should be passed to ChromeClients.
199         https://bugs.webkit.org/show_bug.cgi?id=54926
200
201         Update DumpRenderTree's UIDelegate to use the new console message delegate.
202
203         * DumpRenderTree/mac/UIDelegate.mm:
204         (-[UIDelegate webView:addMessageToConsole:withSource:]):
205
206 2011-03-01  Dirk Pranke  <dpranke@chromium.org>
207
208         Reviewed by Adam Barth.
209
210         Tragically, we stopped running the python unit tests
211         on Chromium Win just as I was getting them to be green.
212         This patch turns them back on.
213
214         https://bugs.webkit.org/show_bug.cgi?id=55521
215
216         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
217
218 2011-03-01  Dirk Pranke  <dpranke@chromium.org>
219
220         Reviewed by Adam Roben.
221
222         Disable
223         webkitpy.common.system.executive_unittest.ExecutiveTest.\
224            test_kill_all fails on win until it can be properly triaged.
225
226         https://bugs.webkit.org/show_bug.cgi?id=54790
227
228         * Scripts/webkitpy/common/system/executive_unittest.py:
229
230     2011-03-01  Helder Correia  <helder@sencha.com>
231
232             Unreviewed. Add myself to committers.py.
233
234             * Scripts/webkitpy/common/config/committers.py:
235
236 2011-03-01  Adam Roben  <aroben@apple.com>
237
238         Give up if a crash log for the web process is taking too long to be saved on Windows
239
240         This should keep the bots from getting stuck while waiting for a crash log, if writing a
241         crash log ever hangs or the UI process for some reason doesn't notice it's finished.
242
243         Fixes <http://webkit.org/b/55499> WebKitTestRunner can hang forever waiting for a crash log
244         to be saved for the web process
245
246         Reviewed by Steve Falkenburg.
247
248         * WebKitTestRunner/win/TestControllerWin.cpp:
249         (WTR::runRunLoopUntil): Moved code to run the run loop here here from platformRunUntil.
250         Generalized the code slightly to handle the optional object to wait on.
251         (WTR::TestController::platformRunUntil): Use the new runRunLoopUntil function, and also use
252         it (with a timeout) when waiting for a crash log for the web process to be saved. This will
253         prevent us from waiting forever.
254
255 2011-03-01  Adam Roben  <aroben@apple.com>
256
257         Give WebKitLauncherWin's PDB file a name different from WebKit's
258
259         WebKitLauncherWin (which builds WebKit.exe) and WebKit (which builds WebKit.dll) were
260         fighting over the same PDB file, meaning that one of them was going to lose and not have
261         symbols. WebKitLauncherWin's PDB file is now named WebKitLauncherWin.pdb.
262
263         Fixes <http://webkit.org/b/55495> Crashes in WebKit.dll have bad backtraces for some test
264         runs (due to missing symbols)
265
266         Reviewed by David Kilzer.
267
268         * WebKitLauncherWin/WebKitLauncherWinCommon.vsprops: Use the project name,
269         "WebKitLauncherWin", instead of the target name, "WebKit", when naming the PDB file.
270
271 2011-03-01  Adam Roben  <aroben@apple.com>
272
273         Count web process crashes toward the --exit-after-n-crashes-or-timeouts limit
274
275         Fixes <http://webkit.org/b/55492> Web process crashes aren't counted toward the
276         --exit-after-n-crashes-or-timeouts limit
277
278         Reviewed by David Kilzer.
279
280         * Scripts/old-run-webkit-tests:
281         (stopRunningTestsEarlyIfNeeded): Include web process crashes in the computation, and print
282         the number of tests that caused the web process to crash when exiting early.
283
284 2011-03-01  Adam Roben  <aroben@apple.com>
285
286         Save a crash log when the web process crashes
287
288         On Windows, WebKitTestRunner now detects when the web process is crashing and waits to exit
289         until it has finished crashing, which guarantees that the crash log will have had time to be
290         saved, too. On Mac, we always wait until ReportCrash has exited before capturing the crash
291         log, so all we have to do is choose the right crash log out of the CrashReporter directory.
292
293         Fixes <http://webkit.org/b/44121> <rdar://problem/8320759> When the web process crashes and
294         a crash log is being saved, WebKitTestRunner thinks the web process has become unresponsive
295
296         Reviewed by Sam Weinig.
297
298         * Scripts/old-run-webkit-tests:
299         (testCrashedOrTimedOut): Don't kill WebKitTestRunner when the web process crashes. It will
300         kill itself. On Windows, this will cause us to wait until the crash log has been saved. On
301         Mac, it should have no effect. Capture saved crash logs for web process crashes, too.
302         (captureSavedCrashLog): Added $webProcessCrashed argument. On Mac, look for
303         WebProces_*.crash files when the web process crashes.
304
305         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
306         (WTR::InjectedBundle::initialize): Added an initializationUserData argument. Updated for
307         initializePlatformDefaults -> platformInitialize rename. Pass the initializationUserData
308         along to platformInitialize.
309
310         * WebKitTestRunner/InjectedBundle/InjectedBundle.h: See above.
311
312         * WebKitTestRunner/InjectedBundle/InjectedBundleMain.cpp:
313         (WKBundleInitialize): Pass along the initializationUserData to the InjectedBundle.
314
315         * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm:
316         (WTR::InjectedBundle::platformInitialize):
317         * WebKitTestRunner/InjectedBundle/qt/InjectedBundleQt.cpp:
318         (WTR::InjectedBundle::platformInitialize):
319         Updated function signature.
320
321         * WebKitTestRunner/InjectedBundle/win/InjectedBundleWin.cpp:
322         (WTR::exceptionFilter): Added. Tells the UI process we're crashing by signaling the
323         webProcessCrashingEvent, then lets the crash continue as normal.
324
325         (WTR::InjectedBundle::platformInitialize): Hook up exceptionFilter. Retrieve the name of the
326         event we should use to tell the UI process we're crashing from the initializationUserData,
327         and get a handle to that event.
328
329         * WebKitTestRunner/TestController.cpp:
330         (WTR::TestController::TestController): Initialize new members.
331         (WTR::TestController::processDidCrash): Removed unnecessary WKPageRef argument. Changed to
332         only print the "#CRASHED - WebProcess" message once, since this can be called more than once
333         when a crash log is being saved on Windows. Exit right away if specified. (This is the
334         default.)
335
336         * WebKitTestRunner/TestController.h: Added new members.
337
338         * WebKitTestRunner/win/TestControllerWin.cpp:
339         (WTR::TestController::platformInitialize): Set up the event the web process will use to tell
340         us it's crashing.
341         (WTR::TestController::platformRunUntil): Pass MWMO_INPUTAVAILABLE to
342         ::MsgWaitForMultipleObjectsEx so we'll process messages that have already been seen by
343         ::PeekMessage. (This is unrelated to the bug fix.) Notice when the webProcessCrashingEvent
344         has been signaled. When this happens, print the "#CRASHED - WebProcess" message right away
345         so the test harness will know the web process has crashed and not try to kill us, then wait
346         for the web process to finish crashing so a crash log will have time to be saved.
347         (WTR::toWK): Simple hepler function.
348         (WTR::TestController::platformInitializeContext): Pass along the name of the event the web
349         process should use to tell us it is crashing in the context's initialization user data.
350
351 2011-03-01  Dimitri Glazkov  <dglazkov@chromium.org>
352
353         Reviewed by Tony Gentilcore.
354
355         Move the checks to exclude Python/Perl tests for Chromium Win build to the right place.
356         https://bugs.webkit.org/show_bug.cgi?id=55476
357
358         Chromium Win uses BuildAndTestFactory, not TestFactory.
359
360         * BuildSlaveSupport/build.webkit.org-config/master.cfg: Moved to the right place.
361
362 2011-03-01  Adam Roben  <aroben@apple.com>
363
364         Test that the WebKit2 responsiveness timer doesn't fire too early
365
366         Test for <http://webkit.org/b/55417> <rdar://problem/9065287> RunLoop::Timer fires 1000x too
367         early on Windows
368
369         Reviewed by Anders Carlsson.
370
371         * TestWebKitAPI/PlatformUtilities.h: Added sleep().
372
373         * TestWebKitAPI/Tests/WebKit2/ResponsivenessTimerDoesntFireEarly.cpp: Added.
374         (TestWebKitAPI::didReceiveMessageFromInjectedBundle):
375         (TestWebKitAPI::didFinishLoadForFrame):
376         (TestWebKitAPI::processDidBecomeUnresponsive):
377         (TestWebKitAPI::setInjectedBundleClient):
378         (TestWebKitAPI::setPageLoaderClient):
379         Simple helper functions.
380
381         (TestWebKitAPI::TEST): Load an HTML file to make sure the web process is initialized. Then
382         tell the web process to pause and press the spacebar key. The spacebar keypress should cause
383         the responsiveness timer to start, but the web process should unpause before it has a chance
384         to fire. Run until the web process has unpaused, and assert that the timer didn't fire.
385
386         * TestWebKitAPI/Tests/WebKit2/ResponsivenessTimerDoesntFireEarly_Bundle.cpp: Added.
387         (TestWebKitAPI::ResponsivenessTimerDoesntFireEarlyTest::ResponsivenessTimerDoesntFireEarlyTest):
388         Call up to the base class.
389         (TestWebKitAPI::ResponsivenessTimerDoesntFireEarlyTest::didReceiveMessage): When asked to
390         pause, sleep for 0.5 seconds, then send back a message saying we paused.
391
392         * TestWebKitAPI/mac/PlatformUtilitiesMac.mm:
393         (TestWebKitAPI::Util::sleep): Added. Calls through to usleep.
394         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp:
395         (TestWebKitAPI::Util::sleep): Added. Calls through to ::Sleep.
396
397         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
398         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
399         * TestWebKitAPI/win/TestWebKitAPIInjectedBundle.vcproj:
400         Added new files.
401
402 2011-03-01  Andras Becsi  <abecsi@webkit.org>
403
404         Reviewed by Csaba Osztrogonác.
405
406         [Qt] Clean up the project files and move common options to WebKit.pri.
407
408         * WebKitTestRunner/qt/WebKitTestRunner.pro: Deduplicate.
409
410
411 2011-03-01  Adam Roben  <aroben@apple.com>
412
413         Skip another multiprocessing test on Windows
414
415         It is sometimes failing (and may be leaving child Python processes around).
416         <http://webkit.org/b/55087> tracks the failure.
417
418         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
419         (FuncitonTests.test_get__processes): Skip this test on Windows.
420
421 2011-03-01  Ilya Tikhonovsky  <loislo@chromium.org>
422
423         Reviewed by Pavel Feldman.
424
425         Web Inspector: reduce inspector tests flakiness.
426         https://bugs.webkit.org/show_bug.cgi?id=55401
427
428         * DumpRenderTree/chromium/TestShell.cpp:
429         (TestShell::runFileTest):
430         * DumpRenderTree/chromium/WebPreferences.cpp:
431         (WebPreferences::reset):
432
433 2011-02-28  Mihai Parparita  <mihaip@chromium.org>
434
435         Reviewed by James Robinson.
436
437         NRWT should still upload results when it exits early due to too many failures
438         https://bugs.webkit.org/show_bug.cgi?id=55432
439         
440         Only keyboard interrupts should prevent results from being uploaded.
441
442         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
443         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
444
445 2011-02-28  Alice Liu  <alice.liu@apple.com>
446
447         Reviewed by Adam Roben and Gavin Barraclough.
448
449         Test for https://bugs.webkit.org/show_bug.cgi?id=54898
450         InjectedBundleNodeHandle dies too early in WKBundleHitTestResultGetNodeHandle.
451
452         * TestWebKitAPI/InjectedBundleController.h:
453         (TestWebKitAPI::InjectedBundleController::bundle):
454         * TestWebKitAPI/PlatformWebView.h:
455         * TestWebKitAPI/Tests/WebKit2/HitTestResultNodeHandle.cpp: Added.
456         (TestWebKitAPI::didFinishLoadForFrame):
457         (TestWebKitAPI::didReceiveMessageFromInjectedBundle):
458         (TestWebKitAPI::setPageLoaderClient):
459         (TestWebKitAPI::TEST):
460         * TestWebKitAPI/Tests/WebKit2/HitTestResultNodeHandle_Bundle.cpp: Added.
461         (TestWebKitAPI::HitTestResultNodeHandleTest::HitTestResultNodeHandleTest):
462         (TestWebKitAPI::HitTestResultNodeHandleTest::getContextMenuFromDefaultMenu):
463         (TestWebKitAPI::HitTestResultNodeHandleTest::didCreatePage):
464         * TestWebKitAPI/mac/PlatformWebViewMac.mm:
465         (TestWebKitAPI::PlatformWebView::simulateRightClick):
466         * TestWebKitAPI/win/PlatformWebViewWin.cpp:
467         (TestWebKitAPI::PlatformWebView::simulateRightClick):
468         
469         Added files to:
470         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
471         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
472         * TestWebKitAPI/win/TestWebKitAPIInjectedBundle.vcproj:
473
474 2011-02-28  Adam Roben  <aroben@apple.com>
475
476         Start running InjectedBundle API tests on Windows
477
478         A new project, TestWebKitAPIInjectedBundle, builds the injected bundle DLL.
479
480         Fixes <http://webkit.org/b/55420> run-api-tests doesn't run injected bundle tests on Windows
481
482         Reviewed by Darin Adler.
483
484         * TestWebKitAPI/Configurations/TestWebKitAPIDebug.vsprops: Added.
485         * TestWebKitAPI/Configurations/TestWebKitAPIDebugAll.vsprops: Added.
486         * TestWebKitAPI/Configurations/TestWebKitAPIDebugCairoCFLite.vsprops: Added.
487         * TestWebKitAPI/Configurations/TestWebKitAPIRelease.vsprops: Added.
488         * TestWebKitAPI/Configurations/TestWebKitAPIReleaseCairoCFLite.vsprops: Added.
489         These are just extracted from TestWebKitAPI.vcproj.
490
491         * TestWebKitAPI/Configurations/TestWebKitAPIInjectedBundleCommon.vsprops: Added. Just sets
492         up the DLL's filename.
493
494         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp: Fixed DLL name.
495         (TestWebKitAPI::Util::createInjectedBundlePath): Changed to use kCFAllocatorNull instead of
496         0 (== kCFAllocatorDefault) so that CF won't try to deallocate our C string.
497
498         * TestWebKitAPI/win/TestWebKitAPI.sln: Added TestWebKitAPIInjectedBundle. Removed the bogus
499         "all" configuration.
500
501         * TestWebKitAPI/win/TestWebKitAPI.vcproj: Changed to use the new .vsprops files. Added
502         injected bundle tests. Let VS resort things.
503
504         * TestWebKitAPI/win/TestWebKitAPIInjectedBundle.vcproj: Copied from Tools/TestWebKitAPI/win/TestWebKitAPI.vcproj.
505
506 2011-02-28  Adam Roben  <aroben@apple.com>
507
508         Add a test for the paint rect passed via WM_PAINT to windowless plugins
509
510         Test for <http://webkit.org/b/55365> <rdar://problem/9031089> REGRESSION (r79040): WebKit2:
511         Large portions of pages with plugins paint black
512
513         Reviewed by Anders Carlsson.
514
515         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
516         (PluginTest::NPP_HandleEvent):
517         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
518         Added NPP_HandleEvent. Just returns 0 at this level.
519
520         * DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowlessPaintRectCoordinates.cpp: Added.
521         (WindowlessPaintRectCoordinates::WindowlessPaintRectCoordinates): Call up to the base class.
522         (WindowlessPaintRectCoordinates::NPP_New): Mark ourselves as windowless.
523         (WindowlessPaintRectCoordinates::NPP_HandleEvent): Check that the paint rect passed via
524         WM_PAINT has the expected coordinates.
525
526         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
527         (NPP_HandleEvent): Give the PluginTest a chance to handle the event before doing anything
528         else.
529
530         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Added new file, let VS
531         resort things.
532
533         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
534         (WTR::InjectedBundlePage::didFinishLoadForFrame): Added a call to WKBundlePageForceRepaint.
535         This is roughly equivalent to the call to -[WebView displayIfNeeded] in -[FrameLoadDelegate
536         webView:didFinishLoadForFrame:] in DumpRenderTree.
537
538         * WebKitTestRunner/win/PlatformWebViewWin.cpp:
539         (WTR::PlatformWebView::PlatformWebView): Tell the WKView it's in a window so that plugins
540         can start running.
541
542 2011-02-28  Adam Roben  <aroben@apple.com>
543
544         Tell MiniBrowser's WKViews that they're in a window
545
546         Fixes <http://webkit.org/b/55364> Plugins don't work in MiniBrowser on Windows
547
548         Reviewed by Anders Carlsson.
549
550         * MiniBrowser/win/BrowserView.cpp:
551         (BrowserView::create): Call WKViewSetIsInWindow after creating the view.
552
553 2011-02-28  Alexis Menard  <alexis.menard@openbossa.org>
554
555         Reviewed by Csaba Osztrogonác.
556
557         The new QML tests require QML_IMPORT_PATH variable to be set in order to
558         load the proper plugin.
559
560         * Scripts/run-qtwebkit-tests:
561
562 2011-02-28  Joone Hur  <joone.hur@collabora.co.uk>
563
564         Unreviewed, fix spelling mistake.
565
566         hardware-accelated => hardware-accelerated.
567
568         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
569
570 2011-02-28  Kristian Amlie  <kristian.amlie@nokia.com>
571
572         Reviewed by Andreas Kling.
573
574         Made sure that the build-webkit --qmake argument is always respected.
575
576         Previously, it would be respected during the build, but not during
577         the early feature detection. The build would also fail if qmake was
578         not in your path, even if you specified --qmake with a valid qmake
579         binary.
580
581         [Qt] WebKit patches required to work with a modularized version of Qt
582         https://bugs.webkit.org/show_bug.cgi?id=53916
583
584         * Scripts/build-webkit:
585         * Scripts/webkitdirs.pm:
586
587 2011-02-27  Ojan Vafai  <ojan@chromium.org>
588
589         Fix failing tests from http://trac.webkit.org/changeset/79837.
590
591         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
592         This was just a mistake that would break if someone tried printing with this configuration.
593         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
594         Method signature changed. Didn't notice that this method was being calling in a unittest.
595
596 2011-02-24  Hayato Ito  <hayato@chromium.org>
597
598         Reviewed by Eric Seidel.
599
600         [NRWT] Remove test_types/* classes.
601
602         This is the last patch in the series of efforts which tried to remove test_types classes.
603         A single_test_runner module will take over responsibilities of test_type classes.
604
605         From this patch, we won't display statistics per test_type even if a '--verbose' option is given.
606
607         https://bugs.webkit.org/show_bug.cgi?id=55123
608
609         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
610         * Scripts/webkitpy/layout_tests/layout_package/test_results.py:
611         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:
612         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
613         * Scripts/webkitpy/layout_tests/layout_package/worker_mixin.py:
614         * Scripts/webkitpy/layout_tests/test_types/__init__.py: Removed.
615         * Scripts/webkitpy/layout_tests/test_types/image_diff.py: Removed.
616         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py: Removed.
617         * Scripts/webkitpy/layout_tests/test_types/test_type_base_unittest.py: Removed.
618         * Scripts/webkitpy/layout_tests/test_types/text_diff.py: Removed.
619
620 2011-02-25  Ojan Vafai  <ojan@chromium.org>
621
622         Reviewed by Tony Chang.
623
624         Change results.json format to the one used by unexpected_results.json
625         https://bugs.webkit.org/show_bug.cgi?id=52267
626
627         Also add runtimes in milliseconds to the JSON and make the output format more compact.
628         Named the file full_results.json to avoid conflicting with the results.json
629         file the test-results server currently serves up.
630
631         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
632         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
633         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
634         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
635         * Scripts/webkitpy/layout_tests/layout_package/result_summary.py:
636         * Scripts/webkitpy/layout_tests/layout_package/test_runner.py:
637
638 2011-02-27  Adam Roben  <aroben@apple.com>
639
640         Followup fix to r79827
641
642         * DumpRenderTree/win/UIDelegate.cpp:
643         (UIDelegate::webViewAddMessageToConsole): Replaced call to urlSuitableForTestResult with
644         lastPathComponent. This matches what we did here before r79827, and matches Mac.
645
646 2011-02-27  Dan Bernstein  <mitz@apple.com>
647
648         Reviewed by Adam Roben.
649
650         Make bisect-builds print a trac link for the regression range
651         https://bugs.webkit.org/show_bug.cgi?id=55331
652
653         * Scripts/bisect-builds:
654
655 2011-02-27  Adam Roben  <aroben@apple.com>
656
657         Use iswalpha instead of isalpha when dealing with wchar_ts in EditingDelegate
658
659         Fixes <http://webkit.org/b/55062> <rdar://problem/9059907> Crash beneath
660         EditingDelegate::checkSpellingOfString when running fast/forms/input-text-maxlength.html or
661         fast/forms/input-text-paste-maxlength.html on Windows with full page heap enabled
662
663         Reviewed by Anders Carlsson.
664
665         * DumpRenderTree/win/EditingDelegate.cpp:
666         (indexOfFirstWordCharacter):
667         (wordLength):
668         Use iswalpha instead of isalpha, since TCHAR is really just wchar_t.
669
670 2011-02-27  Adam Roben  <aroben@apple.com>
671
672         Make ResourceLoadDelegate print URLs relative to the main resource on Windows
673
674         This matches Mac. (In fact, the code was ported from Mac!)
675
676         Fixes <http://webkit.org/b/55328> URLs printed by ResourceLoadDelegate on Windows don't
677         match Mac
678
679         Reviewed by Anders Carlsson.
680
681         * DumpRenderTree/win/DumpRenderTree.cpp:
682         (substringFromIndex): Added. Emulates -[NSString substringFromIndex:].
683         (urlSuitableForTestResult): Ported code from -[NSURL(DRTExtras)
684         _drt_descriptionSuitableForTestResult].
685         (cfStringRefToWString): Moved here from LayoutTestControllerWin.cpp.
686
687         * DumpRenderTree/win/DumpRenderTreeWin.h: Added declaration of cfStringRefToWString.
688
689         * DumpRenderTree/win/LayoutTestControllerWin.cpp: Moved cfStringRefToWString to
690         DumpRenderTree.cpp.
691
692 2011-02-27  Adam Roben  <aroben@apple.com>
693
694         Change the order of Chang Shu's email addresses so the first one is their Bugzilla username
695
696         This makes Bugzilla autocompletion work correctly.
697
698         * Scripts/webkitpy/common/config/committers.py:
699
700 2011-02-27  Patrick Gansterer  <paroga@webkit.org>
701
702         Reviewed by Andreas Kling.
703
704         [Qt] fast/loader/user-stylesheet-fast-path.html fails
705         https://bugs.webkit.org/show_bug.cgi?id=50911
706
707         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
708         (LayoutTestController::setUserStyleSheetLocation):
709         Use QUrl::fromEncoded since the passed string is already encoded.
710
711 2011-02-27  Adam Roben  <aroben@apple.com>
712
713         Ensure $testResultsDirectory is an absolute path before setting up Windows crash log saving
714
715         Windows 7 (and maybe Vista?) launches the post-mortem debugger with a working directory that
716         is different from old-run-webkit-test's. Because of this, the path to the crash log file
717         (which is relative to $testResultsDirectory) must be an absolute path. See the bug for more
718         details.
719
720         Fixes <http://webkit.org/b/55318> Crash logs aren't saved on Windows 7 when a relative
721         --results-directory path is used (like on the test slaves)
722
723         Reviewed by Dan Bernstein.
724
725         * Scripts/old-run-webkit-tests: Moved the call to setUpWindowsCrashLogSaving after we've
726         made $testResultsDirectory an absolute path.
727
728 2011-02-26  Dominic Mazzoni  <dmazzoni@google.com>
729
730         Reviewed by Adam Barth.
731
732         m_dumpAccessibilityNotifications should be initialized.
733         https://bugs.webkit.org/show_bug.cgi?id=54874
734
735         * DumpRenderTree/chromium/AccessibilityController.cpp:
736         (AccessibilityController::AccessibilityController):
737
738 2011-02-26  Vsevolod Vlasov  <vsevik@chromium.org>
739
740         Reviewed by Pavel Feldman.
741
742         DumpRenderTree should reset frame opener between tests.
743         https://bugs.webkit.org/show_bug.cgi?id=54874
744
745         Added clearOpener() calls between tests.
746         Some tests are using window.open() to change current window location.
747         This changes window.opener property whose nullity is checked before
748         showing xml viewer.
749
750         * DumpRenderTree/chromium/TestShell.cpp:
751         (TestShell::resetTestController):
752         * DumpRenderTree/gtk/DumpRenderTree.cpp:
753         (resetDefaultsToConsistentValues):
754         * DumpRenderTree/mac/DumpRenderTree.mm:
755         (resetWebViewToConsistentStateBeforeTesting):
756         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
757         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
758         * DumpRenderTree/win/DumpRenderTree.cpp:
759         (resetWebViewToConsistentStateBeforeTesting):
760         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
761         (WTR::InjectedBundlePage::reset):
762
763 2011-02-26  Yael Aharon  <yael.aharon@nokia.com>
764
765         Reviewed by Andreas Kling.
766
767         [Qt] Fix the focus in MiniBrowser.
768         https://bugs.webkit.org/show_bug.cgi?id=55288
769
770         Give focus to the web view when a load starts.
771
772         * MiniBrowser/qt/BrowserWindow.cpp:
773         (BrowserWindow::urlChanged):
774
775 2011-02-26  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
776
777         Reviewed by Adam Barth.
778
779         [EFL] Add coding style exceptions for EFL port
780         https://bugs.webkit.org/show_bug.cgi?id=54733
781
782         The EFL APIs use EFL naming style, which includes both lower-cased and camel-cased,
783         underscore-sparated values. But, style checking rule have found parameter_name errors
784         in efl directories so far.
785
786         * Scripts/webkitpy/style/checker.py: Add prameter_name rule's exception and efl directories.
787
788 2011-02-26  David Levin  <levin@chromium.org>
789
790         Reviewed by Darin Adler.
791
792         check-webkit-style should check for spaces after periods in comments
793         https://bugs.webkit.org/show_bug.cgi?id=55253
794
795         * Scripts/webkitpy/style/checkers/cpp.py: Added the check.
796         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added tests
797           and fixed the name of one related test.
798
799 2011-02-26  Kenneth Rohde Christiansen  <kenneth@webkit.org>
800
801         Reviewed by Andreas Kling.
802
803         Make it possible to test the targetdensity-dpi support
804         https://bugs.webkit.org/show_bug.cgi?id=55142
805
806         Test the viewport meta tag feature targetdensity-dpi by
807         adding extra arguments to dumpConfigurationForViewport
808
809         Add the new arguments to the shared header and all
810         implementation.
811
812         * DumpRenderTree/LayoutTestController.cpp:
813         (dumpConfigurationForViewportCallback):
814         * DumpRenderTree/LayoutTestController.h:
815         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
816         (LayoutTestController::dumpConfigurationForViewport):
817         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
818         (LayoutTestController::dumpConfigurationForViewport):
819         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
820         (LayoutTestController::dumpConfigurationForViewport):
821         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
822         (LayoutTestController::dumpConfigurationForViewport):
823         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
824         (LayoutTestController::dumpConfigurationForViewport):
825         * DumpRenderTree/qt/LayoutTestControllerQt.h:
826
827 2011-02-26  Sheriff Bot  <webkit.review.bot@gmail.com>
828
829         Unreviewed, rolling out r79764.
830         http://trac.webkit.org/changeset/79764
831         https://bugs.webkit.org/show_bug.cgi?id=55295
832
833         "broke Chromium builds" (Requested by rniwa on #webkit).
834
835         * DumpRenderTree/chromium/TestShell.cpp:
836         (TestShell::resetTestController):
837         * DumpRenderTree/gtk/DumpRenderTree.cpp:
838         (resetDefaultsToConsistentValues):
839         * DumpRenderTree/mac/DumpRenderTree.mm:
840         (resetWebViewToConsistentStateBeforeTesting):
841         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
842         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
843         * DumpRenderTree/win/DumpRenderTree.cpp:
844         (resetWebViewToConsistentStateBeforeTesting):
845         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
846         (WTR::InjectedBundlePage::reset):
847
848 2011-02-26  Yongjun Zhang  <yongjun_zhang@apple.com>
849
850         Reviewed by David Kilzer.
851
852         https://bugs.webkit.org/show_bug.cgi?id=48781
853
854         Add a resource load delegate method to query if WebCore should paint the default broken image for failed images.
855
856         Add a new resource load client method (shouldPaintBrokenImage).  WebKit client can decide if WebCore
857         should paint the default broken image when an image fails to load or decode.
858
859         * DumpRenderTree/LayoutTestController.cpp:
860         (LayoutTestController::LayoutTestController):
861         (setShouldPaintBrokenImageCallback):
862         (LayoutTestController::staticFunctions):
863         (LayoutTestController::setShouldPaintBrokenImage):
864         * DumpRenderTree/LayoutTestController.h:
865         (LayoutTestController::shouldPaintBrokenImage):
866         * DumpRenderTree/mac/ResourceLoadDelegate.mm:
867         (-[ResourceLoadDelegate webView:shouldPaintBrokenImageForURL:]):
868
869 2011-02-26  Vsevolod Vlasov  <vsevik@chromium.org>
870
871         Reviewed by Pavel Feldman.
872
873         DumpRenderTree should reset frame opener between tests.
874         https://bugs.webkit.org/show_bug.cgi?id=54874
875
876         Added clearOpener() calls between tests.
877         Some tests are using window.open() to change current window location.
878         This changes window.opener property whose nullity is checked before
879         showing xml viewer.
880
881         * DumpRenderTree/chromium/TestShell.cpp:
882         (TestShell::resetTestController):
883         * DumpRenderTree/gtk/DumpRenderTree.cpp:
884         (resetDefaultsToConsistentValues):
885         * DumpRenderTree/mac/DumpRenderTree.mm:
886         (resetWebViewToConsistentStateBeforeTesting):
887         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
888         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
889         * DumpRenderTree/win/DumpRenderTree.cpp:
890         (resetWebViewToConsistentStateBeforeTesting):
891         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
892         (WTR::InjectedBundlePage::reset):
893
894 2011-02-25  Lukasz Slachciak  <l.slachciak@samsung.com>
895
896         Reviewed by Kenneth Rohde Christiansen.
897
898         [EFL] Added EWebLauncher command line option to turn on/off frame flattening
899         https://bugs.webkit.org/show_bug.cgi?id=54809
900
901         * EWebLauncher/main.c: Handling -f option added.
902         (on_key_down):
903         (browserCreate):
904         (main):
905
906 2011-02-25  Sheriff Bot  <webkit.review.bot@gmail.com>
907
908         Unreviewed, rolling out r79324.
909         http://trac.webkit.org/changeset/79324
910         https://bugs.webkit.org/show_bug.cgi?id=55260
911
912         Leopard Debug is once again running out of address space while
913         linking (Requested by aroben on #webkit).
914
915         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
916         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
917
918 2011-02-25  Tony Chang  <tony@chromium.org>
919
920         Reviewed by James Robinson.
921
922         Clean up temp files left by crashing DRTs on chromium-win
923         https://bugs.webkit.org/show_bug.cgi?id=55182
924
925         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
926         * BuildSlaveSupport/chromium/remove-crash-logs: Added.
927
928 2011-02-25  Adam Roben  <aroben@apple.com>
929
930         Double-check that modifying the registry worked
931
932         On Windows Vista/7 with UAC enabled, regtool will fail to modify the registry, but it will
933         still return a successful exit code. So we double-check here that the value we tried to
934         write to the registry was really written.
935
936         Fixes <http://webkit.org/b/55227> old-run-webkit-tests falsely claims to be saving crash
937         logs on Windows Vista/7
938
939         Reviewed by Darin Adler.
940
941         * Scripts/old-run-webkit-tests:
942         (writeRegistryString): After trying to write the string value, check that it was really
943         written.
944
945 2011-02-25  Adam Roben  <aroben@apple.com>
946
947         Set the 32-bit post-mortem debugger on 64-bit OSes
948
949         The post-mortem debugger is controlled by some registry values in
950         HKLM\SOFTWARE\Microsoft\Windows NT\CurrentVersion\AeDebug. On 64-bit OSes, you can either
951         set the values in that key, which will be used for 64-bit apps, or in the equivalent key
952         under Wow6432Node, which will be used for 32-bit apps. Since DumpRenderTree and
953         WebKitTestRunner are 32-bit, we want to use the latter location.
954
955         Fixes <http://webkit.org/b/55225> old-run-webkit-tests sets the post-mortem debugger for
956         64-bit apps, but should set it for 32-bit apps instead
957
958         Reviewed by Darin Adler.
959
960         * Scripts/old-run-webkit-tests:
961         (readRegistryString):
962         (writeRegistryString):
963         Pass --wow32 to regtool so it will get/set values beneath Wow6432Node, which is where
964         Windows looks for the post-mortem debugger values when a 32-bit app crashes on a 64-bit OS.
965         Passing --wow32 on a 32-bit OS will not have any effect.
966
967 2011-02-25  Adam Roben  <aroben@apple.com>
968
969         Clean up code to set up crash log saving on Windows
970
971         Fixes <http://webkit.org/b/55224> old-run-webkit-test's Windows registry code is ugly
972
973         Reviewed by David Kilzer.
974
975         * Scripts/old-run-webkit-tests:
976         (readRegistryString):
977         (writeRegistryString):
978         Moved code to read strings from and write strings to the registry here...
979
980         (setUpWindowsCrashLogSaving): ...from here.
981
982 2011-02-25  Adam Roben  <aroben@apple.com>
983
984         Work around Cygwin's crash-suppression behavior
985
986         Cygwin calls ::SetErrorMode(SEM_FAILCRITICALERRORS), which any processes it launches will
987         inherit. This is bad for testing/debugging, as it causes the post-mortem debugger not to be
988         invoked. (Cygwin does this because it makes crashes more UNIX-y.) We reset the error mode
989         when our test apps launch to work around Cygwin's behavior.
990
991         Fixes <http://webkit.org/b/55222> Test apps crash silently (without invoking post-mortem
992         debugger) when launched from Cygwin 1.7
993
994         Reviewed by Darin Adler.
995
996         * DumpRenderTree/win/DumpRenderTree.cpp:
997         (main):
998         * TestWebKitAPI/win/main.cpp:
999         (main):
1000         * WebKitAPITest/main.cpp:
1001         (main):
1002         * WebKitTestRunner/win/TestControllerWin.cpp:
1003         (WTR::TestController::platformInitialize):
1004         Call ::SetErrorMode(0) to undo Cygwin's folly.
1005
1006 2011-02-24  Jocelyn Turcotte  <jocelyn.turcotte@nokia.com>
1007
1008         Reviewed by Andreas Kling.
1009
1010         [Qt] Revert the support for QNAM affined to a different thread.
1011         https://bugs.webkit.org/show_bug.cgi?id=55149
1012
1013         Qt 4.8 will have QNAM use its own thread internally by default,
1014         no need to keep this complexity in WebKit.
1015
1016         This mainly reverts:
1017         http://trac.webkit.org/changeset/73710
1018         http://trac.webkit.org/changeset/73712
1019
1020         * QtTestBrowser/launcherwindow.cpp:
1021         (LauncherWindow::~LauncherWindow):
1022         (LauncherWindow::initializeView):
1023         (LauncherWindow::createChrome):
1024         * QtTestBrowser/launcherwindow.h:
1025         (WindowOptions::WindowOptions):
1026         * QtTestBrowser/webpage.cpp:
1027         (WebPage::WebPage):
1028         * QtTestBrowser/webpage.h:
1029
1030 2011-02-24  Adam Barth  <abarth@webkit.org>
1031
1032         Reviewed by Eric Seidel.
1033
1034         Add a script for creating gypi files
1035         https://bugs.webkit.org/show_bug.cgi?id=55113
1036
1037         This is a trivial script to create gypi files from the source code in
1038         the current directory.
1039
1040         * Scripts/make-gypi: Added.
1041
1042 2011-02-24  Sheriff Bot  <webkit.review.bot@gmail.com>
1043
1044         Unreviewed, rolling out r79628.
1045         http://trac.webkit.org/changeset/79628
1046         https://bugs.webkit.org/show_bug.cgi?id=55195
1047
1048         "broke chromium win tests" (Requested by dpranke on #webkit).
1049
1050         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1051         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1052
1053 2011-02-24  Dirk Pranke  <dpranke@chromium.org>
1054
1055         Reviewed by Mihai Parparita.
1056
1057         rebaseline-chromium-webkit-tests can't handle GPU variants
1058         https://bugs.webkit.org/show_bug.cgi?id=55002
1059
1060         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
1061         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
1062
1063 2011-02-24  Dirk Pranke  <dpranke@chromium.org>
1064
1065         Reviewed by Mihai Parparita.
1066
1067         nrwt: log command used to start web server to --verbose
1068
1069         https://bugs.webkit.org/show_bug.cgi?id=55188
1070
1071         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1072         * Scripts/webkitpy/layout_tests/port/http_server.py:
1073
1074 2011-02-24  Victoria Kirst  <vrk@google.com>
1075
1076         Reviewed by Mihai Parparita.
1077
1078         [chromium] Move media-file.js and video-test.js back to media/ for UILayoutTests
1079         https://bugs.webkit.org/show_bug.cgi?id=55089
1080
1081         This adds an alias into the media/ directory so that http/tests/media
1082         tests can access the media resources when running in an httpd process.
1083
1084         * Scripts/webkitperl/httpd.pm:
1085         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1086         * Scripts/webkitpy/layout_tests/port/http_server.py:
1087
1088 2011-02-24  Dirk Pranke  <dpranke@chromium.org>
1089
1090         Reviewed by Mihai Parparita.
1091
1092         nrwt: enable multiple processes by default on Chromium Win.
1093
1094         https://bugs.webkit.org/show_bug.cgi?id=55163
1095
1096         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1097         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1098
1099 2011-02-24  Sam Weinig  <sam@webkit.org>
1100
1101         Attempt to fix Qt.
1102
1103         * WebKitTestRunner/TestInvocation.cpp:
1104
1105 2011-02-24  James Robinson  <jamesr@chromium.org>
1106
1107         Unreviewed, rolling out r79598.
1108         http://trac.webkit.org/changeset/79598
1109         https://bugs.webkit.org/show_bug.cgi?id=55089
1110
1111         Appears to be causing trouble with httpd on windows
1112
1113         * Scripts/webkitperl/httpd.pm:
1114         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1115         * Scripts/webkitpy/layout_tests/port/http_server.py:
1116
1117 2011-02-24  Sam Weinig  <sam@webkit.org>
1118
1119         Reviewed by Darin Adler.
1120
1121         Add basic pixel testing support to WebKitTestRunner
1122         https://bugs.webkit.org/show_bug.cgi?id=55161
1123
1124         * Scripts/old-run-webkit-tests:
1125         Don't print color space information when using WebKitTestRunner since
1126         we don't need to change the screen.
1127
1128         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
1129         (WTR::InjectedBundle::initialize):
1130         Add call to initializePlatformDefaults.
1131         
1132         (WTR::InjectedBundle::done):
1133         Return a dictionary on completion with both the text output and pixel dump override.
1134
1135         * WebKitTestRunner/InjectedBundle/InjectedBundle.h:
1136         Add declaration for initializePlatformDefaults.
1137
1138         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
1139         (WTR::LayoutTestController::LayoutTestController):
1140         (WTR::LayoutTestController::dumpAsText):
1141         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
1142         (WTR::LayoutTestController::shouldDumpPixels):
1143         Make setting dumpAsText override the dump pixels preference.
1144
1145         * WebKitTestRunner/InjectedBundle/mac/InjectedBundleMac.mm: Added.
1146         (WTR::InjectedBundle::initializePlatformDefaults):
1147         Set platform defaults matching DumpRenderTree.
1148
1149         * WebKitTestRunner/InjectedBundle/qt/InjectedBundleQt.cpp: Added.
1150         (WTR::InjectedBundle::initializePlatformDefaults):
1151         * WebKitTestRunner/InjectedBundle/win/InjectedBundleWin.cpp: Added.
1152         (WTR::InjectedBundle::initializePlatformDefaults):
1153         Add stubs for initializePlatformDefaults.
1154
1155         * WebKitTestRunner/TestController.cpp:
1156         (WTR::TestController::runTest):
1157         (WTR::TestController::runTestingServerLoop):
1158         (WTR::TestController::run):
1159         Parse expected pixel results out of stdin and setup the invocation with it.
1160
1161         * WebKitTestRunner/TestInvocation.cpp:
1162         (WTR::TestInvocation::TestInvocation):
1163         Initialize new members.
1164
1165         (WTR::TestInvocation::~TestInvocation):
1166         (WTR::TestInvocation::setIsPixelTest):
1167         (WTR::sizeWebViewForCurrentTest):
1168         (WTR::TestInvocation::invoke):
1169         (WTR::TestInvocation::dump):
1170         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
1171         * WebKitTestRunner/TestInvocation.h:
1172         Dump pixels in addition to text by calling dumpPixelsAndCompareWithExpected.
1173
1174         * WebKitTestRunner/mac/PlatformWebViewMac.mm:
1175         (WTR::PlatformWebView::PlatformWebView):
1176         Use -[NSColorSpace genericRGBColorSpace] instead of the main screen color space.
1177
1178         * WebKitTestRunner/mac/TestInvocationMac.mm: Added.
1179         (WTR::createCGContextFromPlatformView):
1180         (WTR::computeMD5HashStringForContext):
1181         (WTR::dumpBitmap):
1182         (WTR::forceRepaintFunction):
1183         (WTR::TestInvocation::dumpPixelsAndCompareWithExpected):
1184         Add mac specific pixel dumping support. Right now, we always pull pixels
1185         from the window server.
1186
1187         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
1188         * WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro:
1189         * WebKitTestRunner/qt/WebKitTestRunner.pro:
1190         * WebKitTestRunner/win/InjectedBundle.vcproj:
1191         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
1192         Add new files.
1193
1194 2011-02-24  Chang Shu  <cshu@webkit.org>
1195
1196         Reviewed by Kenneth Rohde Christiansen.
1197
1198         [Qt] Launch MiniBrowser if the test run is WebKit2. 
1199         https://bugs.webkit.org/show_bug.cgi?id=55145
1200
1201         * Scripts/old-run-webkit-tests:
1202         * Scripts/run-launcher:
1203         * Scripts/webkitdirs.pm:
1204
1205 2011-02-24  Victoria Kirst  <vrk@google.com>
1206
1207         Reviewed by Mihai Parparita.
1208
1209         [chromium] Move media-file.js and video-test.js back to media/ for UILayoutTests
1210         https://bugs.webkit.org/show_bug.cgi?id=55089
1211
1212         This adds an alias into the media/ directory so that http/tests/media
1213         tests can access the media resources when running in an httpd process.
1214
1215         * Scripts/webkitperl/httpd.pm:
1216         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1217         * Scripts/webkitpy/layout_tests/port/http_server.py:
1218
1219 2011-02-24  Andrew Wilson  <atwilson@chromium.org>
1220
1221         Unreviewed, rolling out r79570.
1222         http://trac.webkit.org/changeset/79570
1223         https://bugs.webkit.org/show_bug.cgi?id=54874
1224
1225         Breaks chromium build because glue/mocks/mock_web_frame.h/cc
1226         was not updated
1227
1228         * DumpRenderTree/chromium/TestShell.cpp:
1229         (TestShell::resetTestController):
1230         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1231         (resetDefaultsToConsistentValues):
1232         * DumpRenderTree/mac/DumpRenderTree.mm:
1233         (resetWebViewToConsistentStateBeforeTesting):
1234         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1235         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
1236         * DumpRenderTree/win/DumpRenderTree.cpp:
1237         (resetWebViewToConsistentStateBeforeTesting):
1238         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1239         (WTR::InjectedBundlePage::reset):
1240
1241 2011-02-24  Mahesh Kulkarni  <mahesh.kulkarni@nokia.com>
1242
1243         Reviewed by Laszlo Gombos.
1244
1245         [Qt] add android 2.2 user agent to QtTestBrowser
1246         https://bugs.webkit.org/show_bug.cgi?id=55085
1247
1248         Credit to Forrest Hodgkins to find right user-agent which works for youtube.
1249         * QtTestBrowser/useragentlist.txt:
1250
1251 2011-02-24  Adam Roben  <aroben@apple.com>
1252
1253         Change FrameLoadDelegate to support any number of delegates with delayed work to process
1254
1255         This makes our behavior match Mac more closely, and allows us to remove an incorrect
1256         assertion that was firing during some tests. (The assertion was claiming that there was
1257         never more than one delegate with delayed work to process, but that was not the case.)
1258
1259         Fixes <http://webkit.org/b/55146> Assertion failure in FrameLoadDelegate::locationChangeDone
1260         when running http/tests/navigation/back-twice-without-commit.html
1261
1262         Reviewed by Eric Carlson.
1263
1264         * DumpRenderTree/win/FrameLoadDelegate.cpp:
1265         (delegatesWithDelayedWork): Added. Returns all FrameLoadDelegates that have delayed work to
1266         process. A single delegate may appear in this Vector more than once (just as, on Mac, a
1267         single delegate may have multiple performSelector requests).
1268         (processWorkTimer): Pass the HWND to ::KillTimer, for pedantic brownie points. Added an
1269         assertion that the timer firing is the shared process work timer. Instead of using the
1270         single, global "delegate waiting for timer" delegate, give all delegates that have delayed
1271         work to process a chance to process their work.
1272         (FrameLoadDelegate::locationChangeDone): If we don't already have an active timer for
1273         processing delayed work, create one. Then add ourselves to the delegatesWithDelayedWork
1274         Vector so our processWork function will be called when the timer fires.
1275
1276 2011-02-24  Vsevolod Vlasov  <vsevik@chromium.org>
1277
1278         Reviewed by Alexey Proskuryakov.
1279
1280         DumpRenderTree should reset frame opener between tests.
1281         https://bugs.webkit.org/show_bug.cgi?id=54874
1282
1283         Added clearOpener() calls between tests.
1284         Some tests are using window.open() to change current window location.
1285         This changes window.opener property whose nullity is checked before
1286         showing xml viewer.
1287
1288         * DumpRenderTree/chromium/TestShell.cpp:
1289         (TestShell::resetTestController):
1290         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1291         (resetDefaultsToConsistentValues):
1292         * DumpRenderTree/mac/DumpRenderTree.mm:
1293         (resetWebViewToConsistentStateBeforeTesting):
1294         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1295         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
1296         * DumpRenderTree/win/DumpRenderTree.cpp:
1297         (resetWebViewToConsistentStateBeforeTesting):
1298         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1299         (WTR::InjectedBundlePage::reset):
1300
1301 2011-02-24  Adam Roben  <aroben@apple.com>
1302
1303         Don't trigger Windows builds for changes to WebCore.exp.in or the top-level GNUmakefile.am
1304
1305         Fixes <http://webkit.org/b/55144> Windows builders should not have built r79343 or r79440,
1306         but did
1307
1308         Reviewed by Anders Carlsson.
1309
1310         * Scripts/webkitpy/common/config/build.py:
1311         (_should_file_trigger_build): Loosened the pattern for GNUmakefile.am to include the
1312         top-level GNUmakefile.am, and loosened the pattern for .exp files to include WebCore.exp.in.
1313
1314         * Scripts/webkitpy/common/config/build_unittest.py:
1315         (ShouldBuildTest): Added some more tests.
1316
1317 2011-02-24  Andras Becsi  <abecsi@webkit.org>
1318
1319         Reviewed by Laszlo Gombos.
1320
1321         [Qt] MinGW build fails to link
1322         https://bugs.webkit.org/show_bug.cgi?id=55050
1323
1324         Prepend the libraries of subcomponents instead of appending them
1325         to fix the library order according to the dependency of the libraries
1326
1327         * WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro:
1328         * WebKitTestRunner/qt/WebKitTestRunner.pro:
1329
1330 2011-02-24  Gabor Loki  <loki@webkit.org>
1331
1332         Reviewed by Csaba Osztrogonác.
1333
1334         [Qt] Remove ARMv5 Release buildbot.
1335
1336         Remove ARMv5 Release buildbot, because there is very little difference
1337         between ARMv5 and ARMv7 Release bots, and there is also very little
1338         interest in ARMv5 build slave (compared to ARMv7).
1339
1340         * BuildSlaveSupport/build.webkit.org-config/config.json:
1341         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1342
1343 2011-02-22  Ojan Vafai  <ojan@chromium.org>
1344
1345         Reviewed by Tony Chang.
1346
1347         fix revision number finding when creating git patches
1348         https://bugs.webkit.org/show_bug.cgi?id=55024
1349
1350         HEAD~n does not walk merge commits correctly. git log however does.
1351
1352         * Scripts/webkitpy/common/checkout/scm.py:
1353         * Scripts/webkitpy/common/checkout/scm_unittest.py:
1354
1355 2011-02-23  Kenneth Russell  <kbr@google.com>
1356
1357         Unreviewed, rolling out r79387.
1358         https://bugs.webkit.org/show_bug.cgi?id=54885
1359
1360         Reapply the code changes from r79038 since they weren't the root
1361         cause of the test timeouts.
1362
1363         * DumpRenderTree/chromium/WebViewHost.cpp:
1364         (WebViewHost::reset):
1365
1366 2011-02-23  Siddharth Mathur  <siddharth.mathur@nokia.com>
1367
1368         Reviewed by Laszlo Gombos.
1369
1370         [Qt] Make sure Symbian binary UIDs are unique
1371
1372         * MiniBrowser/qt/MiniBrowser.pro: Resolve the UID collision with
1373         QtTestBrowser.pro by changing the UID.
1374
1375 2011-02-23  Patrick Gansterer  <paroga@webkit.org>
1376
1377         Reviewed by Darin Adler.
1378
1379         Rename PLATFORM(CF) to USE(CF)
1380         https://bugs.webkit.org/show_bug.cgi?id=53540
1381
1382         * DumpRenderTree/config.h:
1383
1384 2011-02-23  Dirk Pranke  <dpranke@chromium.org>
1385
1386         Reviewed by Ojan Vafai.
1387
1388         nrwt: turn on multiprocessing on chromium_linux, clean up unit tests for worker model defaults
1389         https://bugs.webkit.org/show_bug.cgi?id=55016
1390
1391         * Scripts/webkitpy/layout_tests/port/base.py:
1392         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1393         * Scripts/webkitpy/layout_tests/port/chromium_linux_unittest.py: Added.
1394         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1395         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
1396         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1397         * Scripts/webkitpy/layout_tests/port/mac.py:
1398         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1399         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
1400         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
1401
1402 2011-02-23  Tony Chang  <tony@chromium.org>
1403
1404         Reviewed by Kent Tamura.
1405
1406         [chromium] DumpRenderTree --check-layout-test-sys-deps leaks 4 temp dirs
1407         https://bugs.webkit.org/show_bug.cgi?id=55004
1408
1409         * DumpRenderTree/chromium/DumpRenderTree.cpp:
1410         (WebKitSupportTestEnvironment::WebKitSupportTestEnvironment):
1411         (WebKitSupportTestEnvironment::~WebKitSupportTestEnvironment):
1412         (main): Use a C++ object to ensure that TearDownTestEnvironment
1413             gets called when --check-layout-test-sys-deps is called
1414
1415 2011-02-23  Gopal Raghavan  <gopal.1.raghavan@nokia.com>
1416
1417         Reviewed by Anders Carlsson.
1418
1419         [Qt] [Qt] check webkit style fails in qt/tests
1420         https://bugs.webkit.org/show_bug.cgi?id=54998
1421
1422         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.
1423
1424         * Scripts/webkitpy/style/checker.py:
1425
1426 2011-02-22  Vsevolod Vlasov  <vsevik@chromium.org>
1427
1428         Reviewed by Pavel Feldman.
1429
1430         DumpRenderTree should dump as text if test path contains "dumpAsText/"
1431         https://bugs.webkit.org/show_bug.cgi?id=54873
1432
1433         * DumpRenderTree/chromium/LayoutTestController.h:
1434         (LayoutTestController::setShouldDumpAsText):
1435         (LayoutTestController::setShouldGeneratePixelResults):
1436         * DumpRenderTree/chromium/TestShell.cpp:
1437         (TestShell::runFileTest):
1438         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1439         (shouldDumpAsText):
1440         (dump):
1441         (runTest):
1442         * DumpRenderTree/mac/DumpRenderTree.mm:
1443         (shouldDumpAsText):
1444         (runTest):
1445         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1446         (WebCore::isDumpAsTextTest):
1447         (WebCore::DumpRenderTree::open):
1448         * DumpRenderTree/win/DumpRenderTree.cpp:
1449         (shouldDumpAsText):
1450         (runTest):
1451         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1452         (WTR::InjectedBundlePage::dump):
1453
1454 2011-02-22  Kenneth Russell  <kbr@google.com>
1455
1456         Reviewed by James Robinson.
1457
1458         Many tests started timing out on the chrome mac canaries around r79035
1459         https://bugs.webkit.org/show_bug.cgi?id=54885
1460
1461         Revert the code changes, but not the Chromium DEPS roll, from
1462         r79038 to see whether this clears up the timing-out tests. Was not
1463         able to reproduce the timeouts locally with a Release mode build
1464         of DRT (on Snow Leopard).
1465
1466         * DumpRenderTree/chromium/WebViewHost.cpp:
1467         (WebViewHost::reset):
1468
1469 2011-02-22  Dirk Pranke  <dpranke@chromium.org>
1470
1471         Reviewed by James Robinson.
1472
1473         nrwt: --platform chromium-gpu doesn't work with --worker-model=processes
1474         https://bugs.webkit.org/show_bug.cgi?id=55009
1475
1476         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
1477
1478 2011-02-18  Mihai Parparita  <mihaip@chromium.org>
1479
1480         Reviewed by Ojan Vafai.
1481
1482         Default to 500 for --exit-after-n-failures
1483         https://bugs.webkit.org/show_bug.cgi?id=54773
1484
1485         Make 500 the default value for --exit-after-n-failures (ORWT bots
1486         already default to this as of r75726). Similarly, change the default for
1487         --exit-after-n-crashes-or-timeouts to 20.
1488
1489         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1490
1491 2011-02-02  Martin Robinson  <mrobinson@igalia.com>
1492
1493         Reviewed by Xan Lopez.
1494
1495         [GTK] DRT needs an implementation of LayoutTestController.setSelectTrailingWhitespace
1496         https://bugs.webkit.org/show_bug.cgi?id=53603
1497
1498         Add support for LayoutTestController.setSelectTrailingWhitespace by calling through
1499         to DumpRenderTreeSupportGtk for this functionality.
1500
1501         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1502         (resetDefaultsToConsistentValues): Always call LayoutTestController.setSelectTrailingWhitespace to
1503         false when reseting values between tests.
1504         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1505         (LayoutTestController::setSelectTrailingWhitespaceEnabled): Call LayoutTestController.setSelectTrailingWhitespace
1506         with the appropriate value.
1507
1508 2011-02-22  Adam Roben  <aroben@apple.com>
1509
1510         Don't trigger a Windows build when the top-level ChangeLog changes
1511
1512         r79320 is an example of a change that shouldn't have built on Windows, but did, because of
1513         this bug.
1514
1515         Fixes <http://webkit.org/b/54957> Windows bots build when top-level ChangeLog changes, but
1516         shouldn't
1517
1518         Reviewed by Anders Carlsson.
1519
1520         * Scripts/webkitpy/common/config/build.py:
1521         (_should_file_trigger_build): Modified the ChangeLog pattern to match the top-level
1522         ChangeLog, and re-sorted it in the list.
1523
1524         * Scripts/webkitpy/common/config/build_unittest.py:
1525         (ShouldBuildTest): Added a test that ChangeLogs don't trigger builds.
1526
1527 2011-02-22  Sheriff Bot  <webkit.review.bot@gmail.com>
1528
1529         Unreviewed, rolling out r78691.
1530         http://trac.webkit.org/changeset/78691
1531         https://bugs.webkit.org/show_bug.cgi?id=54956
1532
1533         Leopard Debug is successfully building again (Requested by
1534         aroben on #webkit).
1535
1536         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
1537         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1538
1539 2011-02-22  Ilya Tikhonovsky  <loislo@chromium.org>
1540
1541         Reviewed by Pavel Feldman.
1542
1543         Web Inspector: flakyness of inspector tests.
1544         https://bugs.webkit.org/show_bug.cgi?id=54729
1545
1546         As far as we have the protocol with sequence numbers we can simplify test harness support
1547         and drop out chomium specific methods.
1548
1549         * DumpRenderTree/DumpRenderTree.gypi:
1550         * DumpRenderTree/chromium/DRTDevToolsAgent.cpp:
1551         (DRTDevToolsAgent::sendMessageToInspectorFrontend):
1552         (DRTDevToolsAgent::runtimePropertyChanged):
1553         (DRTDevToolsAgent::asyncCall):
1554         (DRTDevToolsAgent::call):
1555         * DumpRenderTree/chromium/DRTDevToolsAgent.h:
1556         * DumpRenderTree/chromium/DRTDevToolsCallArgs.cpp: Removed.
1557         * DumpRenderTree/chromium/DRTDevToolsCallArgs.h: Removed.
1558         * DumpRenderTree/chromium/DRTDevToolsClient.cpp:
1559         (DRTDevToolsClient::sendMessageToBackend):
1560         (DRTDevToolsClient::asyncCall):
1561         (DRTDevToolsClient::call):
1562         * DumpRenderTree/chromium/DRTDevToolsClient.h:
1563
1564 2011-02-22  Andras Becsi  <abecsi@webkit.org>
1565
1566         Reviewed by Laszlo Gombos.
1567
1568         [Qt] Redesign the build system
1569         https://bugs.webkit.org/show_bug.cgi?id=51339
1570
1571         Part 2.
1572
1573         Build WebCore as a static library, compile the WebKit API and WebKit2 API
1574         in a final step and link to WebKit2, WebCore and JSC libraries to fix
1575         linking issues resulting from stripped away symbols.
1576
1577         * WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro: Remove superfluous includepaths.
1578
1579 2011-02-21  Lukasz Slachciak  <l.slachciak@samsung.com>
1580
1581         Reviewed by Antonio Gomes.
1582
1583         [EFL] Restored spatial navigation option in EWebLauncher since #18662 bug is closed
1584         https://bugs.webkit.org/show_bug.cgi?id=54806
1585
1586         * EWebLauncher/main.c:
1587         (on_key_down): calling ewk_view_setting_spatial_navigation_set when "F12" pressed added.
1588
1589 2011-02-21  Dirk Pranke  <dpranke@chromium.org>
1590
1591         Reviewed by Kent Tamura.
1592
1593         nrwt: revert behavior on 10.5 to "old-threads". It's possible
1594         that r79062 has introduced a bunch of flakiness, although I
1595         can't explain why right now. This change should restore the
1596         old behavior, and then we can look at the build logs to see
1597         if the tests stop timing out or being flaky.
1598
1599         https://bugs.webkit.org/show_bug.cgi?id=54925
1600
1601         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1602         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
1603         * Scripts/webkitpy/layout_tests/port/mac.py:
1604         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1605
1606 2011-02-21  Dirk Pranke  <dpranke@chromium.org>
1607
1608         Reviewed by Dimitri Glazkov.
1609
1610         nrwt: fix typo causing us to try and repeatedly shut down the servers
1611         https://bugs.webkit.org/show_bug.cgi?id=54904
1612
1613         * Scripts/webkitpy/layout_tests/layout_package/worker_mixin.py:
1614
1615 2011-02-21  Alejandro G. Castro  <alex@igalia.com>
1616
1617         Reviewed by Csaba Osztrogonác.
1618
1619         REGRESSION(r79157):
1620         plugins/get-url-with-javascript-destroying-plugin.html make the
1621         next test crash
1622         https://bugs.webkit.org/show_bug.cgi?id=54863
1623
1624         * GNUmakefile.am:
1625
1626 2011-02-20  Dirk Pranke  <dpranke@chromium.org>
1627
1628         Reviewed by Ojan Vafai.
1629
1630         test-webkitpy: stop skipping a bunch of tests on win32
1631         https://bugs.webkit.org/show_bug.cgi?id=54788
1632
1633         * Scripts/webkitpy/common/net/testoutputset_unittest.py:
1634         * Scripts/webkitpy/test/main.py:
1635
1636 2011-02-20  Dirk Pranke  <dpranke@chromium.org>
1637
1638         Reviewed by Ojan Vafai.
1639
1640         Fix various bugs related to handling of dir separators and real
1641         vs. fake filesystems that were causing test-webkitpy to fail
1642         tests on win32. These bugs have a variety of causes but are
1643         lumped together because they should be easy to review.
1644
1645         https://bugs.webkit.org/show_bug.cgi?id=54700
1646
1647         * Scripts/webkitpy/common/net/testoutput.py:
1648           Here we were using os.path.sep but not escaping it prior to
1649           using in a regexp, which  caused bad things on win32.
1650         * Scripts/webkitpy/common/net/testoutput_unittest.py:
1651           Same here.
1652         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
1653           Here we were using the native filesystem instead of the mock
1654           filesystem, making some tests unnecessarily platform dependent.
1655           Also we weren't setting the default configuration for the
1656           TestChromiumWinPort.
1657         * Scripts/webkitpy/layout_tests/update_webgl_conformance_tests.py:
1658           Here we *shouldn't* use os.path.join(), just so the code becomes
1659           less platform dependent.
1660         * Scripts/webkitpy/style/checker.py:
1661           Here we should be using os.path.sep instead of "/"; the tests
1662           require the native filesystem.
1663         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
1664           Here the underlying code is using os.path.sep, so we probably
1665           should as well.
1666         * Scripts/webkitpy/style/test_expectations_unittest.py:
1667           Here we should be using the filesystem hanging of the port
1668           object (which is the mocked filesystem) rather than the native
1669           one.
1670
1671 2011-02-20  Anders Carlsson  <andersca@apple.com>
1672
1673         Reviewed by Maciej Stachowiak.
1674
1675         Crash when a plug-in requests a javascript: url that destroys the plug-in
1676         https://bugs.webkit.org/show_bug.cgi?id=54837
1677         <rdar://problem/9005475>
1678
1679         Add new plug-in test.
1680
1681         * DumpRenderTree/DumpRenderTree.gypi:
1682         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
1683         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
1684         (PluginTest::NPN_GetURL):
1685         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
1686         * DumpRenderTree/TestNetscapePlugIn/Tests/GetURLWithJavaScriptURLDestroyingPlugin.cpp: Added.
1687         (GetURLWithJavaScriptURLDestroyingPlugin::GetURLWithJavaScriptURLDestroyingPlugin):
1688         (GetURLWithJavaScriptURLDestroyingPlugin::NPP_New):
1689         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1690         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
1691
1692 2011-02-19  Patrick Gansterer  <paroga@webkit.org>
1693
1694         Reviewed by Andreas Kling.
1695
1696         Make WinCE a core builder
1697         https://bugs.webkit.org/show_bug.cgi?id=54801
1698
1699         * Scripts/webkitpy/common/net/buildbot/buildbot.py:
1700         * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:
1701
1702 2011-02-19  Lukasz Slachciak  <l.slachciak@samsung.com>
1703
1704         Reviewed by Andreas Kling.
1705
1706         Fixed problem with launching EWebLauncher with "run-launcher --efl". Incorrect path referring to non-existing .libs directory ws set.
1707         https://webkit.org/b/54778
1708
1709         * Scripts/webkitdirs.pm:
1710
1711 2011-02-18  David Levin  <levin@chromium.org>
1712
1713         Reviewed by Eric Seidel.
1714
1715         check-webkit-style: Misses brace style error when the line with the { has a } in it.
1716         https://bugs.webkit.org/show_bug.cgi?id=54769
1717
1718         * Scripts/webkitpy/style/checkers/cpp.py: Fixed the check for the close brace
1719           to only look after the last open brace, so that the open brace in this line
1720           "} else {" will still be able to trigger the error.
1721         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added related unit tests.
1722
1723 2011-02-18  Zan Dobersek  <zandobersek@gmail.com>
1724
1725         Reviewed by Martin Robinson.
1726
1727         [Gtk] Re-enable meter tag support
1728         https://bugs.webkit.org/show_bug.cgi?id=54762
1729
1730         Meter tag is supported on the Gtk port, enable it by default.
1731
1732         * Scripts/build-webkit:
1733
1734 2011-02-18  David Levin  <levin@chromium.org>
1735
1736         Reviewed by Eric Seidel.
1737
1738         check-webkit-style falsely complains about WebKitGTK+ public headers
1739         https://bugs.webkit.org/show_bug.cgi?id=54650
1740
1741         * Scripts/webkitpy/style/checker.py: Add some exceptions for the gtk files.
1742         * Scripts/webkitpy/style/checkers/cpp.py: Restrict the checks for WEBKIT_API
1743           to the chromium directory (and improved the checks slightly).
1744         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added corresponding tests.
1745
1746 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1747
1748         Reviewed by James Robinson.
1749
1750         new-run-webkit-tests: enable multiple processes by default on
1751         mac. This change removes the artificial restrictions we placed
1752         on the # of child processes we used with the old threading
1753         model, and switches to the new message-based model and multiple
1754         processes, where available. If multiple processes are not
1755         available (Leopard / Python 2.5), then we use the 'inline' model
1756         instead of the 'old-threads' model on the 'mac' port or one
1757         process on the 'chromium-mac' port. We need additional testing
1758         to see if the new 'threads' model works reliably and is worth
1759         supporting, or if we should just live with things being slightly
1760         slow.
1761
1762         https://bugs.webkit.org/show_bug.cgi?id=54596
1763
1764         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1765         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
1766         * Scripts/webkitpy/layout_tests/port/mac.py:
1767         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1768         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1769
1770 2011-02-18  Andrew Wilson  <atwilson@chromium.org>
1771
1772         Unreviewed, rolling out r79047.
1773         http://trac.webkit.org/changeset/79047
1774         https://bugs.webkit.org/show_bug.cgi?id=54596
1775
1776         Broke canary bots - please watch the canaries the next time
1777         you land this
1778
1779         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1780         * Scripts/webkitpy/layout_tests/port/mac.py:
1781
1782 2011-02-18  Emil A Eklund  <eae@chromium.org>
1783
1784         Reviewed by Darin Adler.
1785
1786         Crash in EventHandler::sendContextMenuEventForKey
1787         https://bugs.webkit.org/show_bug.cgi?id=54495
1788
1789         Add support for the context menu key (VK_APPS) to EventSender for
1790         platforms that support the key.
1791
1792         * DumpRenderTree/chromium/EventSender.cpp:
1793         (EventSender::keyDown):
1794         * DumpRenderTree/gtk/EventSender.cpp:
1795         (keyDownCallback):
1796         * DumpRenderTree/qt/EventSenderQt.cpp:
1797         (EventSender::keyDown):
1798         * DumpRenderTree/win/EventSender.cpp:
1799         (keyDownCallback):
1800
1801 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1802
1803         Reviewed by Ojan Vafai.
1804
1805         new-run-webkit-tests: enable multiple processes by default on mac
1806         https://bugs.webkit.org/show_bug.cgi?id=54596
1807
1808         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1809         * Scripts/webkitpy/layout_tests/port/mac.py:
1810
1811 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1812
1813         Reviewed by Adam Roben.
1814
1815         Skip pretty patch unit tests if ruby isn't installed.
1816
1817         https://bugs.webkit.org/show_bug.cgi?id=54699
1818
1819         * Scripts/webkitpy/common/prettypatch_unittest.py:
1820
1821 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1822
1823         Reviewed by Adam Roben.
1824
1825         Skip mac.test_skipped_file_paths() when running on win32; the
1826         test has hardcoded '/' directory separators and since it's a
1827         port-specific test and we have coverage on other ports it's not
1828         worth it to make the test more generic.
1829
1830         https://bugs.webkit.org/show_bug.cgi?id=54698
1831
1832         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1833
1834 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1835
1836         Reviewed by Adam Roben.
1837
1838         Fix webkitpy.layout_tests.port.config_unittest to work on Win32.
1839
1840         https://bugs.webkit.org/show_bug.cgi?id=54703
1841
1842         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1843
1844 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1845
1846         Reviewed by Adam Roben.
1847
1848         Change the default port we're mocking out on win32 when we
1849         run mock_drt_unittest. Normally we'd default to the
1850         port-specific default, but that would be "win", which doesn't
1851         currently work.
1852
1853         https://bugs.webkit.org/show_bug.cgi?id=54708
1854
1855         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
1856
1857 2011-02-18  Kenneth Russell  <kbr@google.com>
1858
1859         Reviewed by James Robinson.
1860
1861         [chromium] Clean up initialization of minimum timer interval
1862         https://bugs.webkit.org/show_bug.cgi?id=54772
1863
1864         Use new GetForegroundTabTimerInterval function in webkit_support
1865         to reset the minimum timer interval before each test.
1866
1867         * DumpRenderTree/chromium/WebViewHost.cpp:
1868         (WebViewHost::reset):
1869
1870 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1871
1872         Reviewed by Adam Roben.
1873
1874         r78494 fixed the chromium_win unittest failing under win32,
1875         but not cygwin. This fixes cygwin as well.
1876
1877         https://bugs.webkit.org/show_bug.cgi?id=54652
1878
1879         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1880
1881 2011-02-18  Dirk Pranke  <dpranke@chromium.org>
1882
1883         Reviewed by Ojan Vafai, Mihai Parparita.
1884
1885         new-run-webkit-tests: r78522 made it impossible to create a
1886         'chromium-win' or 'chromium-mac' port; they would always get the
1887         version tacked on to the end. It turns out that we actually need
1888         to be able to create these ports in order for
1889         rebaseline-chromium-webkit-tests to work correctly given the
1890         broken coverage of our layout bots. As soon as we get SL and
1891         Win 7 bots running reliably on the canaries, we can get rid of
1892         this change. Note that we only need to change the chromium_mac
1893         and win ports because linux doesn't yet use versions and that
1894         script doesn't work with non-chromium ports.
1895
1896         Also, this change fixes a bug where we were using
1897         "platform/chromium-win-win7" and
1898         "platform/chromium-mac-snowleopard" as the baseline dirs for
1899         Win 7 and SL; we should've been using the generic directories
1900         instead since they are the newest versions of the os's.
1901
1902         https://bugs.webkit.org/show_bug.cgi?id=54691
1903
1904         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1905         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
1906         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1907         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1908         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
1909
1910 2011-02-16  MORITA Hajime  <morrita@google.com>
1911
1912         Reviewed by Kent Tamura.
1913
1914         [Chromium][DRT] some PostDelayedTask() accesses possibly dangling this pointer.
1915         http://webkit.org/b/53899
1916
1917         Replaced static invocation wrappers with MethodTask objects,
1918         which are removed on WebViewHost destructor.
1919         
1920         * DumpRenderTree/chromium/WebViewHost.cpp:
1921         (WebViewHost::requestCheckingOfText):
1922         (WebViewHost::scheduleAnimation):
1923         (WebViewHost::closeWidgetSoon):
1924         * DumpRenderTree/chromium/WebViewHost.h:
1925         (WebViewHost::HostMethodTask::HostMethodTask):
1926         (WebViewHost::HostMethodTask::runIfValid):
1927         (WebViewHost::taskList):
1928
1929 2011-02-07  Hayato Ito  <hayato@chromium.org>
1930
1931         Reviewed by Ojan Vafai.
1932
1933         Get rid of code which writes test results from test_type's
1934         compare_output() method.
1935         Writing test results now happens in single_test_runner, using
1936         newly introduced test_result_writer module.
1937
1938         Subsequent patches will eliminate test_type/* classes.
1939
1940         https://bugs.webkit.org/show_bug.cgi?id=53908
1941
1942         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
1943         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py: Added.
1944         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
1945         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
1946         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
1947
1948 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
1949
1950         Reviewed by Adam Roben.
1951
1952         Add blacklist of remaining webkitpy/* modules that fail on
1953         Win32, so that test-webkitpy will at least run cleanly.
1954
1955         https://bugs.webkit.org/show_bug.cgi?id=54709
1956
1957         * Scripts/webkitpy/layout_tests/test/main.py:
1958
1959 2011-02-17  Simon Fraser  <simon.fraser@apple.com>
1960
1961         Reviewed by Sam Weinig.
1962
1963         Set a preference to ensure that scrollbars are in a predictable
1964         state for DRT.
1965
1966         * DumpRenderTree/mac/DumpRenderTree.mm:
1967         (resetDefaultsToConsistentValues):
1968
1969 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
1970
1971         Reviewed by Adam Roben.
1972
1973         test-webkitpy doesn't work with the multiprocessing module on win32
1974         or cygwin, so we skip it for now.
1975
1976         https://bugs.webkit.org/show_bug.cgi?id=54520
1977
1978         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
1979         * Scripts/webkitpy/layout_tests/layout_package/run_webkit_tests_unittest.py:
1980
1981 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
1982
1983         Reviewed by Adam Roben.
1984
1985         Fix version, baseline_search_path for chromium-win-win7. Also
1986         fix chromium_win_unittests, and fix port_testcase.make_port()
1987         to correctly use the options passed into it and only set
1988         results_directory() if necessary.
1989
1990         https://bugs.webkit.org/show_bug.cgi?id=54652
1991
1992         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1993         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1994         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
1995
1996 2011-02-17  Dirk Pranke  <dpranke@chromium.org>
1997
1998         Reviewed by Ojan Vafai.
1999
2000         Fix a hang in new-run-webkit-tests when run with
2001         --dry-run --worker-model=processes
2002
2003         https://bugs.webkit.org/show_bug.cgi?id=54595
2004
2005         * Scripts/webkitpy/layout_tests/layout_package/test_runner2.py:
2006         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2007
2008 2011-02-17  Kevin Ollivier  <kevino@theolliviers.com>
2009
2010         [wx] Build fixes after recent changes.
2011
2012         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2013         (LayoutTestController::setMinimumTimerInterval):
2014         * wx/build/settings.py:
2015
2016 2011-02-17  Csaba Osztrogonác  <ossy@webkit.org>
2017
2018         Reviewed by Andreas Kling.
2019
2020         [Qt] ImageDiff sometimes hangs
2021         https://bugs.webkit.org/show_bug.cgi?id=54641
2022
2023         * DumpRenderTree/qt/ImageDiff.cpp: Clear actualImage and baselineImage after all comparison.
2024         (main):
2025
2026 2011-02-17  Peter Varga  <pvarga@webkit.org>
2027
2028         Reviewed by Csaba Osztrogonác.
2029
2030         The run-sunspider --v8-suite doesn't work
2031         https://bugs.webkit.org/show_bug.cgi?id=54660
2032
2033         * Scripts/run-sunspider:
2034
2035 2011-02-17  Gabor Rapcsanyi  <rgabor@webkit.org>
2036
2037         Reviewed by Csaba Osztrogonác.
2038
2039         [NRWT] Fix environment setup on Qt port.
2040         https://bugs.webkit.org/show_bug.cgi?id=54549
2041
2042         * Scripts/webkitpy/layout_tests/port/qt.py:
2043
2044 2011-02-16  Yael Aharon  <yael.aharon@nokia.com>
2045
2046         Reviewed by Andreas Kling.
2047
2048         [Qt] Enable tiled backing store by default in MiniBrowser.
2049         https://bugs.webkit.org/show_bug.cgi?id=54581
2050
2051         Instead of a flag to turn it on, a new flag is defined to turn
2052         tiled backing store off.
2053
2054         * MiniBrowser/qt/MiniBrowserApplication.cpp:
2055         (MiniBrowserApplication::handleUserOptions):
2056         * MiniBrowser/qt/MiniBrowserApplication.h:
2057         (WindowOptions::WindowOptions):
2058
2059 2011-02-16  Eric Seidel  <eric@webkit.org>
2060
2061         Reviewed by Mihai Parparita.
2062
2063         flaky test archive uploads are always 22 bytes long
2064         https://bugs.webkit.org/show_bug.cgi?id=54593
2065
2066         Speculative fix for the 22-bytes bug.
2067
2068         * Scripts/webkitpy/tool/bot/flakytestreporter.py:
2069
2070 2011-02-16  Anna Cavender  <annacc@chromium.org>
2071
2072         Reviewed by Mihai Parparita.
2073
2074         http/tests/media should be served over HTTP (not from local file)
2075         https://bugs.webkit.org/show_bug.cgi?id=54028
2076
2077         Removed special file:// treatment of http/tests/media
2078
2079         * Scripts/old-run-webkit-tests:
2080         * Scripts/webkitpy/layout_tests/port/base.py:
2081
2082 2011-02-16  Mihai Parparita  <mihaip@chromium.org>
2083
2084         Reviewed by Ojan Vafai.
2085
2086         ChangeLog.parse_latest_entry_from_file does not handle rolled over ChangeLogs
2087         https://bugs.webkit.org/show_bug.cgi?id=54609
2088         
2089         r78737 rolled over ChangeLogs, and webkit-patch complains when landing
2090         patches such as this one that are the first to add an entry to the
2091         ChangeLog (since they don't see a second date line to know where the
2092         entry ends).
2093
2094         * Scripts/webkitpy/common/checkout/changelog.py:
2095         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
2096
2097 == Rolled over to ChangeLog-2011-02-16 ==