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