2010-11-18 Hayato Ito <hayato@chromium.org>
[WebKit-https.git] / WebKitTools / ChangeLog
1 2010-11-18  Hayato Ito  <hayato@chromium.org>
2
3         Reviewed by Shinichiro Hamaji.
4
5         [NRWT] Fix a break of '--new-baseline' feature in pixel_test
6         which is caused by r72249.
7
8         https://bugs.webkit.org/show_bug.cgi?id=49751
9
10         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
11
12 2010-11-18  Adam Roben  <aroben@apple.com>
13
14         Implement layoutTestController.waitForPolicyDelegate on Windows
15
16         Fixes <http://webkit.org/b/25038> <rdar://problem/6790213>.
17
18         Reviewed by Simon Fraser.
19
20         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
21         (LayoutTestController::waitForPolicyDelegate): Implemented by porting
22         code from LayoutTestControllerMac.mm.
23
24 2010-11-18  Hayato Ito  <hayato@chromium.org>
25
26         Reviewed by Shinichiro Hamaji.
27
28         [NRWT] Fix a break of '--new-baseline' feature which is caused by r72249.
29
30         https://bugs.webkit.org/show_bug.cgi?id=49740
31
32         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
33
34 2010-11-18  Tony Chang  <tony@chromium.org>
35
36         Reviewed by Adam Barth.
37
38         run platform/chromium/plugins/refcount-leaks.html on all platforms
39         https://bugs.webkit.org/show_bug.cgi?id=49485
40
41         Add PluginObject.testObjectCount which returns the number of allocated
42         TestObjects. Add PluginObject.testCreateTestObject which allocates
43         and returns a TestObject.
44
45         Add TestObject.refCount which returns the number of refs on the
46         TestObject.
47
48         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
49         (pluginGetProperty):
50         (pluginInvoke):
51         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
52         (getTestObjectCount):
53         (testAllocate):
54         (testDeallocate):
55         (testGetProperty):
56         (testScriptObjectInvoke): Release a ref to a plugin object that was
57             previously leaked.
58         * DumpRenderTree/TestNetscapePlugIn/TestObject.h:
59
60 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
61
62         Reviewed by Adam Roben.
63
64         WebKitTools projects (WebKitTestRunner, record-memory-win) should use vsprops for common build settings
65         https://bugs.webkit.org/show_bug.cgi?id=49711
66
67         * WebKitTestRunner/Configurations/WebKitTestRunnerCFLite.vsprops: Added.
68         * WebKitTestRunner/Configurations/WebKitTestRunnerCommon.vsprops: Added.
69         * WebKitTestRunner/Configurations/WebKitTestRunnerCoreFoundation.vsprops: Added.
70         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
71         * record-memory-win/record-memory-win-common.vsprops: Added.
72         * record-memory-win/record-memory-win.vcproj:
73
74 2010-11-18  Andras Becsi  <abecsi@inf.u-szeged.hu>
75
76         Reviewed by Csaba Osztrogonác.
77
78         [Qt][WK2] Only add user agent strings to the list which aren't listed yet.
79
80         * MiniBrowser/qt/BrowserWindow.cpp:
81         (BrowserWindow::updateUserAgentList):
82
83 2010-11-17  Adam Roben  <aroben@apple.com>
84
85         Don't trigger Windows builds when chromium-win test results change
86
87         Reviewed by Mark Rowe.
88
89         * Scripts/webkitpy/common/config/build.py:
90         (_should_file_trigger_build): Changed the regular expression used to
91         search for directory names to only consider full directory names,
92         rather than matching directory names where the string we care about is
93         a suffix of the name.
94
95         * Scripts/webkitpy/common/config/build_unittest.py:
96         (ShouldBuildTest): Added a test.
97
98 2010-11-17  Hayato Ito  <hayato@chromium.org>
99
100         Reviewed by Shinichiro Hamaji.
101
102         Remove a uri member from TestInput class.
103
104         https://bugs.webkit.org/show_bug.cgi?id=49691
105
106         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
107         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
108
109 2010-11-17  Adam Roben  <aroben@apple.com>
110
111         Move FindSafari's settings into a vsprops file
112
113         Fixes <http://webkit.org/b/49699> FindSafari should use a vsprops file
114
115         Reviewed by Steve Falkenburg.
116
117         * FindSafari/FindSafari.vcproj:
118         * FindSafari/FindSafariCommon.vsprops: Added.
119
120 2010-11-17  Adam Roben  <aroben@apple.com>
121
122         Use a minimum font size of 0 in DRT on Windows
123
124         This is the Windows equivalent of r72164.
125
126         Rubber-stamped by Dave Hyatt.
127
128         * DumpRenderTree/win/DumpRenderTree.cpp:
129         (resetDefaultsToConsistentValues):
130
131 2010-11-17  Dirk Pranke  <dpranke@chromium.org>
132
133         Reviewed by Andreas Kling.
134
135         webkit-patch pretty-diff shouldn't need to be run from a checkout root in git
136         https://bugs.webkit.org/show_bug.cgi?id=49639
137
138         * Scripts/webkitpy/common/checkout/scm.py:
139
140 2010-11-17  Mihai Parparita  <mihaip@chromium.org>
141
142         Reviewed by Tony Chang.
143
144         Rebaseline server: add loupe for image diffs
145         https://bugs.webkit.org/show_bug.cgi?id=49692
146         
147         Add a loupe (magnifiying glass) for inspecting image diffs. Shows an
148         enlarged area of the expected, actual and diff images side by side.
149
150         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
151         * Scripts/webkitpy/tool/commands/data/rebaselineserver/loupe.js: Added.
152         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
153         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
154
155 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
156
157         Reviewed by Adam Roben.
158
159         WebKitLauncherWin should use vsprops for shared build settings
160         https://bugs.webkit.org/show_bug.cgi?id=49696
161
162         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
163         * WebKitLauncherWin/WebKitLauncherWinCommon.vsprops: Added.
164
165 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
166
167         Reviewed by Adam Roben.
168
169         WinLauncher should use vsprops for shared build settings
170         https://bugs.webkit.org/show_bug.cgi?id=49695
171
172         * WinLauncher/WinLauncher.cpp:
173         (_tWinMain):
174         (PrintView):
175         (WndProc):
176         (MyEditProc):
177         * WinLauncher/WinLauncher.vcproj:
178         * WinLauncher/WinLauncherCommon.vsprops: Added.
179
180 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
181
182         Delete unused file unintentionally added in last change.
183
184         * DumpRenderTree/win/DumpRenderTreeCURL.vsprops: Removed.
185
186 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
187
188         Reviewed by Adam Roben.
189
190         DumpRenderTree on Windows should use a vsprops file for shared project settings
191         https://bugs.webkit.org/show_bug.cgi?id=49690
192
193         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
194         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePluginCommon.vsprops: Added.
195         * DumpRenderTree/win/DumpRenderTree.vcproj:
196         * DumpRenderTree/win/DumpRenderTreeApple.vsprops: Added.
197         * DumpRenderTree/win/DumpRenderTreeCFLite.vsprops: Added.
198         * DumpRenderTree/win/DumpRenderTreeCURL.vsprops: Added.
199         * DumpRenderTree/win/DumpRenderTreeCairo.vsprops: Added.
200         * DumpRenderTree/win/DumpRenderTreeCommon.vsprops: Added.
201         * DumpRenderTree/win/ImageDiff.vcproj:
202         * DumpRenderTree/win/ImageDiffCommon.vsprops: Added.
203
204 2010-11-17  Hayato Ito  <hayato@chromium.org>
205
206         Refactor TestTypeBase.compare_output().
207
208         Introduce a TestOutput class and update compare_output() of each test
209         types so that they can take both actual and expected TestOutput objects.
210
211         https://bugs.webkit.org/show_bug.cgi?id=49431
212
213         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
214         * Scripts/webkitpy/layout_tests/layout_package/test_output.py: Added.
215         * Scripts/webkitpy/layout_tests/port/base.py:
216         * Scripts/webkitpy/layout_tests/port/chromium.py:
217         * Scripts/webkitpy/layout_tests/port/dryrun.py:
218         * Scripts/webkitpy/layout_tests/port/test.py:
219         * Scripts/webkitpy/layout_tests/port/webkit.py:
220         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
221         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
222         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
223         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
224
225 2010-11-17  Adam Roben  <aroben@apple.com>
226
227         Make each Windows Test builder use the same OS for all its slaves
228
229         Currently each WebKit1 Windows Test builder has one Windows XP slave
230         and one Windows 7 slave. Having different OSes in a single builder
231         meant that the builders would alternate between red and green when
232         there were OS-specific failures.
233
234         Fixes <http://webkit.org/b/49688> Windows Test builders alternate
235         between red and green because of multiple OS versions
236
237         Reviewed by Mark Rowe.
238
239         * BuildSlaveSupport/build.webkit.org-config/config.json: Renamed
240         "Windows Release (Tests)" to "Windows 7 Release (Tests)", "Windows
241         Debug (Tests)" to "Windows XP Debug (Tests)", and "Windows Release
242         (WebKit2 Tests)" to "Windows 7 Release (WebKit2 Tests)". Updated the
243         builders for the two WebKit1 builders so that they would all use the
244         appropriate OS.
245
246 2010-11-17  James Robinson  <jamesr@chromium.org>
247
248         Reviewed by Dan Bernstein.
249
250         [chromium] Set minimum font size to 0 for DRT
251         https://bugs.webkit.org/show_bug.cgi?id=49677
252
253         Followup for http://trac.webkit.org/changeset/72141.
254
255         Test: http/tests/misc/acid3.html
256
257         * DumpRenderTree/chromium/WebPreferences.cpp:
258         (WebPreferences::reset):
259
260 2010-11-17  Andras Becsi  <abecsi@webkit.org>
261
262         Unreviewed build fix.
263
264         [Qt] Add missing file I forgot in r72220.
265
266         * MiniBrowser/MiniBrowser.qrc: Added.
267
268 2010-11-17  Andras Becsi  <abecsi@webkit.org>
269
270         Reviewed by Andreas Kling.
271
272         [Qt][WK2] Make it possible to set custom user agent strings in MiniBrowser.
273         https://bugs.webkit.org/show_bug.cgi?id=49627
274
275         Add a user agent dialog and a resource file for useragentlist.txt to MiniBrowser.
276
277         * MiniBrowser/MiniBrowser.qrc: Added.
278         Add it here to prevent qmake from detecting it
279         since it needs to be copied to the build directory.
280         * MiniBrowser/qt/BrowserWindow.cpp:
281         (BrowserWindow::BrowserWindow):
282         (BrowserWindow::updateUserAgentList):
283         (BrowserWindow::showUserAgentDialog):
284         * MiniBrowser/qt/BrowserWindow.h:
285         * MiniBrowser/qt/MiniBrowser.pro:
286
287 2010-11-17  Adam Roben  <aroben@apple.com>
288
289         Make changes to the Mac WebKit2 Skipped file trigger Windows builds
290
291         We use that file on Windows!
292
293         Fixes <http://webkit.org/b/49643> Changes to
294         LayoutTests/platform/mac-wk2 should trigger a Windows build, but don't
295
296         Reviewed by Steve Falkenburg.
297
298         * Scripts/webkitpy/common/config/build.py:
299         (_should_file_trigger_build): Added the mac-wk2 directory and make it
300         trigger builds on SnowLeopard and Windows.
301
302         * Scripts/webkitpy/common/config/build_unittest.py:
303         (ShouldBuildTest.test_should_build): Added a test.
304
305 2010-11-17  Satish Sampath  <satish@chromium.org>
306
307         Reviewed by Jeremy Orlow.
308
309         Clear the speech input mock explicitly before each test.
310         https://bugs.webkit.org/show_bug.cgi?id=49660
311
312         * DumpRenderTree/chromium/LayoutTestController.cpp:
313         (LayoutTestController::setMockSpeechInputResult):
314         * DumpRenderTree/chromium/LayoutTestController.h:
315         * DumpRenderTree/chromium/WebViewHost.cpp:
316         (WebViewHost::speechInputController):
317         (WebViewHost::reset): Invoke speech input mock's clearResults method.
318         * DumpRenderTree/chromium/WebViewHost.h:
319         (WebViewHost::speechInputControllerMock): Moved mock from LayoutTestController.
320
321 2010-11-17  Alexander Pavlov  <apavlov@chromium.org>
322
323         Unreviewed, specify IRC nickname in committers.py.
324
325         * Scripts/webkitpy/common/config/committers.py:
326
327 2010-11-17  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
328
329         Reviewed by Tony Chang.
330
331         [NRWT] Make http locking similar to perl implementation
332         https://bugs.webkit.org/show_bug.cgi?id=49187
333
334         * Scripts/webkitpy/common/system/file_lock.py: Added.
335         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
336         * Scripts/webkitpy/layout_tests/port/http_lock.py:
337
338 2010-11-17  MORITA Hajime  <morrita@google.com>
339
340         Reviewed by Kent Tamura.
341
342         [Chromium][DRT] EventSender.contextClick() should aware spellchecking
343         https://bugs.webkit.org/show_bug.cgi?id=49366
344
345         - EvenSender: Checked WebContextMenuData.misspelledWord and added extra context menu entries
346           according to the spellchecker suggestion
347         - MockSpellCheck: add fillSuggestionList to provide fake suggestions.
348
349         test_expectations.txt will be changed after this change is ported to test_shell.
350         
351         * DumpRenderTree/chromium/EventSender.cpp:
352         (makeMenuItemStringsFor):
353         (EventSender::contextClick):
354         * DumpRenderTree/chromium/MockSpellCheck.cpp:
355         (MockSpellCheck::fillSuggestionList):
356         (MockSpellCheck::initializeIfNeeded):
357         * DumpRenderTree/chromium/MockSpellCheck.h:
358         * DumpRenderTree/chromium/WebViewHost.cpp:
359         (WebViewHost::mockSpellCheck):
360         * DumpRenderTree/chromium/WebViewHost.h:
361
362 2010-11-16  Dave Hyatt  <hyatt@apple.com>
363
364         Reviewed by Sam Weinig.
365
366         Make sure the pref for minimum font size is just explicitly set to 0 every time,
367         since the bots have 1 stuck in their plist otherwise.
368
369         * DumpRenderTree/mac/DumpRenderTree.mm:
370         (resetDefaultsToConsistentValues):
371
372 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
373
374         Reviewed by Tony Chang.
375
376         Rebaseline server: compute diffs client-side
377         https://bugs.webkit.org/show_bug.cgi?id=49640
378         
379         The image diff output from the DRT is pretty bad on some ports (at
380         least the Mac one), so it's better to compute diffs on the client by
381         using <canvas>.
382
383         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
384         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
385         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
386
387 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
388
389         Reviewed by Tony Chang.
390
391         Rebaseline server: display test results
392         https://bugs.webkit.org/show_bug.cgi?id=49626
393         
394         Adds basic result display to the rebaseline server. On the Python side
395         this involves:
396         - Parsing the unexpected_results.json into a dictionary.
397         - Serving it as JSON under /results.json.
398         (the JSON -> dict -> JSON transform isn't strictly necessary right now,
399         but I'll need to have access to the parsed results on the Python side
400         for follow-up changes).
401         
402         On the web UI side this adds:
403         - Markup for display image and text results (expected, actual, diff),
404           and JS for populating it.
405         - Markup for breaking down test results by failure type and directory,
406           and JS for populating it.
407
408         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
409         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
410         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
411         * Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js: Added.
412         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
413
414 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
415
416         Reviewed by Ojan Vafai.
417
418         See Chromium issue http://codereview.chromium.org/5133001/ - we
419         are modifying the build bots to pass in the builder name with
420         the "GPU" string appended instead of appending it in the code.
421
422         https://bugs.webkit.org/show_bug.cgi?id=49636
423
424         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
425
426 2010-11-16  Steve Falkenburg  <sfalken@apple.com>
427
428         Reviewed by Adam Roben.
429
430         Disable LTCG for Windows Release builds. Add new Release_LTCG configuration.
431         https://bugs.webkit.org/show_bug.cgi?id=49632
432
433         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
434         * DumpRenderTree/win/DumpRenderTree.vcproj:
435         * DumpRenderTree/win/ImageDiff.vcproj:
436         * MiniBrowser/MiniBrowser.vcproj:
437         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
438         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
439         * WebKitTestRunner/win/InjectedBundle.vcproj:
440         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
441         * WinLauncher/WinLauncher.vcproj:
442         * record-memory-win/record-memory-win.vcproj:
443
444 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
445
446         Reviewed by Ojan Vafai.
447
448         new-run-webkit-tests: rename TestInfo to TestInput, move image hash to work thread
449
450         Rename the TestInfo class to TestInput to be clearer about its
451         function, and move the checksum-reading code into dump_render_tree_thread
452         to avoid cross-thread access.
453
454         https://bugs.webkit.org/show_bug.cgi?id=49573
455
456         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
457         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
458         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
459
460 2010-11-16  Dave Hyatt  <hyatt@apple.com>
461
462         Fix bustage. Remove the minimum font size pref setting in DumpRenderTree.
463
464         * DumpRenderTree/mac/DumpRenderTree.mm:
465         (resetDefaultsToConsistentValues):
466
467 2010-11-16  Stephanie Lewis  <slewis@apple.com>
468
469         Reviewed by Geoff Garen.
470
471         <rdar://problem/8624267> Leak creating offscreen webview running fast/dom tests
472
473         Use shared DeviceOrientationProviderMock. The old allocation method created a new WebDeviceOrientationProviderMock 
474         for every WebView, and the WebKit API doesn't support that ownership model.
475
476         I also fixed a leak in setMockDeviceOrientation by adding a missing release.
477
478         * DumpRenderTree/mac/DumpRenderTree.mm:
479         (createWebViewAndOffscreenWindow):
480         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
481         (LayoutTestController::setMockDeviceOrientation):
482
483 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
484
485         Reviewed by Ojan Vafai.
486
487         new-run-webkit-tests: log the process id in --verbose mode
488
489         As part of the switch to multi-process mode, we should log the
490         process id when running w/ --verbose.
491
492         https://bugs.webkit.org/show_bug.cgi?id=49571
493
494         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
495
496 2010-11-16  Adam Roben  <aroben@apple.com>
497
498         Only trigger Windows builds when files that we actually use on Windows
499         are changed
500
501         A new Scheduler subclass, PlatformSpecificScheduler, has been added.
502         It uses the new webkitpy.common.config.build module to determine
503         whether a particular change should trigger a build on a particular
504         platform. The Windows builders have been switched to use a
505         PlatformSpecificScheduler.
506
507         The logic to determine whether or not a particular change should
508         trigger a build on a given platform has only been implemented/tested
509         for Windows. I tried to make it easy to add more platforms in the
510         future, but I don't have enough familiarity with all platforms to be
511         able to implement it for them.
512
513         Fixes <http://webkit.org/b/49407> Windows builders kick off builds for
514         lots irrelevant changes (e.g., rebaselining Chromium test results)
515
516         Reviewed by Eric Seidel.
517
518         * BuildSlaveSupport/build.webkit.org-config/config.json: Use a
519         PlatformSpecificScheduler for the Windows builders.
520
521         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
522         (PlatformSpecificScheduler.__init__): Added. Stores our platform, sets
523         up a ChangeFilter that filters to our branch and will call through to
524         our filter method, and calls up to our base class.
525         (PlatformSpecificScheduler.filter): Calls through to
526         build.should_build to find out whether we should trigger a build for
527         this change.
528
529         * Scripts/webkitpy/common/net/build.py: Added.
530         (_should_file_trigger_build): Uses a set of directories and regexp
531         patterns to determine whether the given file should trigger a build on
532         the given platform. As mentioned earlier, this has only been
533         implemented for Windows, though I did try to make some guesses about
534         other platforms.
535         (should_build): Returns true if any of the files should trigger a
536         build on the given platform.
537
538         * Scripts/webkitpy/common/net/build_unittest.py: Added.
539         (ShouldBuildTest.test_should_build): Does some basic testing to make
540         sure we're triggering builds for the right files. It only tests
541         Windows for now, though I tried to make some guesses about other
542         platforms.
543
544 2010-11-16  Simon Hausmann  <simon.hausmann@nokia.com>
545
546         Reviewed by Andreas Kling.
547
548         [Qt] Remove synchronous QWebPage::checkPermissions signal
549         https://bugs.webkit.org/show_bug.cgi?id=46810
550
551         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
552         (WebCore::WebPage::WebPage):
553         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
554         (LayoutTestController::grantDesktopNotificationPermission): When granting
555         permission, grant it directly on the QWebPage/Frame, that will remember it.
556         * QtTestBrowser/webpage.cpp:
557         (WebPage::WebPage):
558         * QtTestBrowser/webpage.h:
559
560
561 2010-11-16  Andras Becsi  <abecsi@inf.u-szeged.hu>
562
563         Reviewed by Csaba Osztrogonác.
564
565         [Qt] Add isPrinting and setPrinting methods to DRT's LayoutTestController.
566
567         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
568         (LayoutTestController::reset):
569         * DumpRenderTree/qt/LayoutTestControllerQt.h:
570         (LayoutTestController::isPrinting):
571         (LayoutTestController::setPrinting):
572
573 2010-11-16  Balazs Kelemen  <kbalazs@webkit.org>
574
575         Reviewed by Andreas Kling.
576
577         [Qt][WK2] Add a way to turn on tiled backing store in MiniBrowser
578         https://bugs.webkit.org/show_bug.cgi?id=49587
579
580         * MiniBrowser/qt/BrowserView.cpp:
581         (BrowserView::BrowserView): Added BackingStoreType
582         argument to ctor. Propagating it to the QGraphicsWKView.
583         * MiniBrowser/qt/BrowserView.h:
584         * MiniBrowser/qt/BrowserWindow.cpp:
585         (BrowserWindow::BrowserWindow):
586         (BrowserWindow::newWindow): Create the new BrowserWindow
587         with the same BackingStoreType that this instance has.
588         * MiniBrowser/qt/BrowserWindow.h: Added BackingStoreType member.
589         * MiniBrowser/qt/main.cpp:
590         (main): Use tiled backing store if got -tiled command line argument.
591         Avoid redundant copy of command line args.
592
593 2010-11-16  John Knottenbelt  <jknotten@chromium.org>
594
595         Reviewed by Jeremy Orlow.
596
597         Move DeviceOrientationClientMock from LayoutTestController to WebViewHost.
598         https://bugs.webkit.org/show_bug.cgi?id=48506
599
600         This ensures that there is one mock per page WebView / WebViewHost. This is a
601         design constraint of the DeviceOrientationClientMock because the m_controller
602         field can only store a pointer to a single instance of the corresponding
603         DeviceOrientationController.
604
605         Test: fast/dom/DeviceOrientation/no-page-cache.html
606
607         * DumpRenderTree/chromium/LayoutTestController.cpp:
608         (LayoutTestController::setMockDeviceOrientation):
609         (LayoutTestController::deviceOrientationClientMock):
610         * DumpRenderTree/chromium/LayoutTestController.h:
611         * DumpRenderTree/chromium/WebViewHost.cpp:
612         (WebViewHost::deviceOrientationClientMock):
613         (WebViewHost::deviceOrientationClient):
614         (WebViewHost::reset):
615         * DumpRenderTree/chromium/WebViewHost.h:
616
617 2010-11-15  Andras Becsi  <abecsi@webkit.org>
618
619         Reviewed by Andreas Kling.
620
621         [Qt][WK2] Avoid polling in WebKitTestRunner.
622         https://bugs.webkit.org/show_bug.cgi?id=49542
623
624         Make the WK2 testing session about 10-15% faster and decrease random flakiness
625         resulting from timing skew by removing the millisecond polling from TestControllerQt.
626
627         * WebKitTestRunner/TestController.cpp:
628         (WTR::TestController::didFinishLoadForFrame):
629         * WebKitTestRunner/TestController.h:
630         * WebKitTestRunner/TestInvocation.cpp:
631         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
632         * WebKitTestRunner/mac/TestControllerMac.mm:
633         (WTR::TestController::notifyDone): Add method stub.
634         * WebKitTestRunner/qt/TestControllerQt.cpp:
635         (WTR::TestControllerRunLoop::instance):
636         (WTR::TestControllerRunLoop::start):
637         (WTR::TestControllerRunLoop::stop):
638         (WTR::TestControllerRunLoop::TestControllerRunLoop):
639         (WTR::TestControllerRunLoop::timerEvent):
640         (WTR::TestController::notifyDone):
641         (WTR::TestController::platformRunUntil):
642         * WebKitTestRunner/win/TestControllerWin.cpp:
643         (WTR::TestController::notifyDone): Add method stub.
644
645 2010-11-15  Darin Adler  <darin@apple.com>
646
647         * Scripts/webkitpy/common/net/bugzilla: Added property svn:ignore.
648
649 2010-11-15  Benjamin Poulain  <benjamin.poulain@nokia.com>
650
651         Reviewed by Andreas Kling.
652
653         [Qt] [WK2] Add Ctrl+L shortcut to Qt Minibrowser
654         https://bugs.webkit.org/show_bug.cgi?id=49544
655
656         Add the shortcut to BrowserWindow.
657
658         * MiniBrowser/qt/BrowserWindow.cpp:
659         (BrowserWindow::BrowserWindow):
660         (BrowserWindow::openLocation):
661         * MiniBrowser/qt/BrowserWindow.h:
662
663 2010-11-14  David Levin  <levin@chromium.org>
664
665         Reviewed by Daniel Bates.
666
667         check-webkit-style should detect PassRefPtr usage in functions.
668         https://bugs.webkit.org/show_bug.cgi?id=49513
669
670         * Scripts/webkitpy/style/checkers/cpp.py:
671         (check_for_function_lengths): Revert a comment change that I
672          accidentally made in r71986.
673         (check_pass_ptr_usage): Added the code to do the check.
674         (process_line): Added the call to check_pass_ptr_usage.
675         (CppChecker): Added the new error category.
676         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
677         (CppStyleTestBase::perform_pass_ptr_check): Runs the new check for
678          testing purposes.
679         (PassPtrTest::*): The class/functions to unit test the new
680          functionality.
681
682 2010-11-14  David Levin  <levin@chromium.org>
683
684         Reviewed by Shinichiro Hamaji.
685
686         check-webkit-style function detection and the line count style checks should be separate.
687         https://bugs.webkit.org/show_bug.cgi?id=49512
688
689         * Scripts/webkitpy/style/checkers/cpp.py: Do the separation.
690         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Adjust the test to
691           call the detection function and fix line counts in two places now that
692           the code really only counts the lines in the body of the function.
693
694 2010-11-14  Andreas Kling  <kling@webkit.org>
695
696         Reviewed by Antonio Gomes.
697
698         http/tests/plugins tests print "Unhandled variable" to stderr under Qt
699         https://bugs.webkit.org/show_bug.cgi?id=33438
700
701         Stifle "Unhandled variable" warning to match what the other
702         TestNetscapePlugin does (NPP_GetValue in TestNetscapePlugIn/main.cpp
703         simply returns NPERR_GENERIC_ERROR for unhandled variables.)
704
705         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
706         (webkit_test_plugin_get_value):
707
708 2010-11-13  David Levin  <levin@chromium.org>
709
710         Reviewed by Eric Seidel.
711
712         check-webkit-style function detection crashes on functions in templates.
713         https://bugs.webkit.org/show_bug.cgi?id=49504
714
715         For example "template <bool shouldClose(const Element*)>".
716
717         * Scripts/webkitpy/style/checkers/cpp.py: If not function is detected, bail out.
718         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Modified the complex
719           function detection test to expose this issue.
720
721 2010-11-12  David Levin  <levin@chromium.org>
722
723         Reviewed by Shinichiro Hamaji.
724
725         check-webkit-style function detection doesn't detect indented functions declaractions.
726         https://bugs.webkit.org/show_bug.cgi?id=49446
727
728         Indented function declarations occur inside class definitions, so
729         they are a pretty common (and worth detecting).
730
731         * Scripts/webkitpy/style/checkers/cpp.py:
732           Changed regex to allow indentation.
733           Changed the function start detection to only happen when not in a
734           function.
735           Changed function end detection to work based on matching braces
736           instead of finding a close brace at the beginning of the line.
737           Fixed close_expression to do what it says when it doesn't find
738           the close.
739         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Indented function test.
740
741 2010-11-12  Daniel Bates  <dbates@rim.com>
742
743         Rubber-stamped by Nikolas Zimmermann.
744
745         Make do-file-rename rename files in the directory WebKit2.
746
747         This also makes the list of searched directories in do-file-rename
748         consistent with the list of searched directories in the script do-webcore-rename.
749
750         * Scripts/do-file-rename:
751
752 2010-11-12  Dirk Pranke  <dpranke@chromium.org>
753
754         Reviewed by James Robinson.
755
756         Attempt yet again to land the fix for bug 49360 (respecting
757         set-webkit-configuration). We need to handle the cases where
758         trying to run webkit-build-directory to find out where the
759         default configuration might be fails (that shows up on some
760         Chromium bots that apparently don't have perl installed).
761
762         https://bugs.webkit.org/show_bug.cgi?id=49360
763
764         * Scripts/webkitpy/layout_tests/port/config.py:
765         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Added.
766         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
767
768 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
769
770         Reviewed by Adam Barth.
771
772         webkit-patch land logs commit message twice
773         https://bugs.webkit.org/show_bug.cgi?id=49482
774         
775         Don't log the comment that's used by bugzilla.close_bug_as_fixed, since
776         that's just the commit message, which we now output following the
777         actual commit.
778
779         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
780
781 2010-11-12  Tony Chang  <tony@chromium.org>
782
783         Reviewed by Adam Barth.
784
785         run platform/chromium/plugins/script-object-invoke.html on all platforms
786         https://bugs.webkit.org/show_bug.cgi?id=49280
787
788         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
789         (testScriptObjectInvoke):
790         (pluginInvoke):
791
792 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
793
794         Reviewed by Adam Barth.
795
796         webkit-patch land-cowboy (and other commands) should display the committed SVN revision after landing a patch
797         https://bugs.webkit.org/show_bug.cgi?id=49471
798
799         Right now there's no indication of a successful commit.
800         
801         * Scripts/webkitpy/tool/commands/download_unittest.py:
802         * Scripts/webkitpy/tool/steps/commit.py:
803
804 2010-11-12  Kinuko Yasuda  <kinuko@chromium.org>
805
806         Reviewed by Ojan Vafai.
807
808         Include detailed test modifiers in results.json and enable incremental uploading for non-layout tests
809         https://bugs.webkit.org/show_bug.cgi?id=49354
810
811         Also moved/integrated the upload method from run_webkit_tests.py to json_results_generator.py.
812
813         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
814         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
815         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
816         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
817
818 2010-11-12  Tony Chang  <tony@chromium.org>
819
820         Reviewed by Kent Tamura.
821
822         [chromium] Load a blank page when closing DRT WebViewHosts
823         https://bugs.webkit.org/show_bug.cgi?id=49417
824
825         In test_shell, when we close a window, we first load about:blank
826         and run GC to fire the destruction logic of the page.  In test_shell,
827         this happens in ~TestShell.  In DRT, we manage virtual windows as
828         WebViewHosts, so we need to replicate this logic in ~WebViewHost.
829
830         Managing focus is a bit trickier, since normally you would get window
831         events to reset the focus to the right window.  We do it manually
832         instead.
833
834         This fixes 2 plugin tests on DRT.
835
836         * DumpRenderTree/chromium/TestShell.cpp:
837         (TestShell::~TestShell): Move window close logic to ~WebViewHost.
838         (TestShell::closeWindow): Move widget close() to WVH destructor and
839             manage focus manually.
840         (TestShell::closeRemainingWindows): Comment typo.
841         * DumpRenderTree/chromium/WebViewHost.cpp:
842         (WebViewHost::closeWidget):
843         (invokeCloseWidget):
844         (WebViewHost::closeWidgetSoon): Close the window asynchronously (matches
845             test_shell).
846         (WebViewHost::~WebViewHost):
847         * DumpRenderTree/chromium/WebViewHost.h:
848
849 2010-11-12  David Levin  <levin@chromium.org>
850
851         Reviewed by Adam Barth.
852
853         cpp.py needs some tidying.
854         https://bugs.webkit.org/show_bug.cgi?id=49443
855
856         * Scripts/webkitpy/style/checkers/cpp.py:
857           Fix parameter ordering for iteratively_replace_matches_with_char
858           to mirror other similar functions (and improved comments, etc.)
859           Also, use a python-ism (for/else) to get rid of a variable that
860           was there to determine if something had been found in the loop.
861
862 2010-11-12  Robert Kroeger  <rjkroege@chromium.org>
863
864         Reviewed by James Robinson.
865
866         [Chromium/DRT] Make EventSendingController send complete touch events.
867         https://bugs.webkit.org/show_bug.cgi?id=49285
868
869         * DumpRenderTree/chromium/EventSender.cpp:
870         (EventSender::addTouchPoint):
871         (EventSender::updateTouchPoint):
872         (EventSender::sendCurrentTouchEvent):
873
874 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
875
876         Unreviewed rollout of r71858.
877
878         Rollout out r71858 since it breaks new-run-webkit-httpd as used by the
879         NaCl tests.
880
881         * Scripts/webkitpy/layout_tests/port/config.py:
882         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Removed.
883         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
884
885 2010-11-12  Benjamin Poulain  <benjamin.poulain@nokia.com>
886
887         Reviewed by Kenneth Rohde Christiansen.
888
889         [Qt] TestController::runUntil doesn't honor the timeout parameter
890         https://bugs.webkit.org/show_bug.cgi?id=48941
891
892         Implement the missing timeout of WebKitTestRunner for Qt.
893
894         A QElapsedTimer has been added to RunUntilConditionLoop in order
895         to measure how long the loop has been running. When the timer
896         is bigger or equal than the timeout, the loop ends even if the
897         condition is not met.
898
899         * WebKitTestRunner/qt/TestControllerQt.cpp:
900         (WTR::RunUntilConditionLoop::start):
901         (WTR::RunUntilConditionLoop::run):
902         (WTR::RunUntilConditionLoop::timerEvent):
903         (WTR::TestController::platformRunUntil):
904
905 2010-11-11  David Levin  <levin@chromium.org>
906
907         Reviewed by Shinichiro Hamaji.
908
909         check-webkit-style function detection doesn't handle templates with spaces.
910         https://bugs.webkit.org/show_bug.cgi?id=49427
911
912         * Scripts/webkitpy/style/checkers/cpp.py: Stripped out templates when
913           finding the function name.
914         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test have a
915           template with a space.
916
917 2010-11-11  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
918
919         Reviewed by Shinichiro Hamaji.
920
921         Enable check-webkit-style on Qt files
922         https://bugs.webkit.org/show_bug.cgi?id=48322
923
924         Adding exception for _q_... private slots and
925         ..._data() methods in tests
926
927         * Scripts/webkitpy/style/checker.py: enable Qt dirs, add exceptions
928         * Scripts/webkitpy/style/checker_unittest.py: improve unittests
929
930 2010-11-11  David Levin  <levin@chromium.org>
931
932         Reviewed by Shinichiro Hamaji.
933
934         check-webkit-style function detection doesn't work for templates and destructors.
935         https://bugs.webkit.org/show_bug.cgi?id=49425
936
937         * Scripts/webkitpy/style/checkers/cpp.py: Changed regex for templates and destructors.
938         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test hit this issue.
939
940 2010-11-11  Joone Hur  <joone@kldp.org>
941
942         Reviewed by Antonio Gomes.
943
944         [GTK][DRT] Implement LayoutTestController::nodesFromRect
945         https://bugs.webkit.org/show_bug.cgi?id=46598
946
947         Support nodesFromRect in DRT
948
949         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
950         (LayoutTestController::nodesFromRect):
951
952 2010-11-11  MORITA Hajime  <morrita@google.com>
953
954         Reviewed by Kent Tamura.
955
956         editing/selection/context-menu-on-text.html fails on chromium
957         https://bugs.webkit.org/show_bug.cgi?id=45898
958
959         Enhance Chromium's EventSender::contextClick() to return
960         array of menu strings. The implementation incomplete because
961         Although actual context menu is implemented by the browser,
962         it reflects editability of underlying context thus makes
963         context-menu-on-text.html passed.
964         
965         * DumpRenderTree/chromium/EventSender.cpp:
966         (makeMenuItemStringsFor): Added.
967         (EventSender::contextClick): Now returns array of menu strings.
968         * DumpRenderTree/chromium/WebViewHost.cpp:
969         (WebViewHost::showContextMenu): Now storing given WebContextMenuData object.
970         (WebViewHost::clearContextMenuData): Added.
971         (WebViewHost::lastContextMenuData): Added.
972         * DumpRenderTree/chromium/WebViewHost.h:
973
974 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
975
976         Unreviewed, build fix.
977
978         fix build breakage caused by fix for bug 49380 (r71858).
979
980         config_standalone needs to ensure that the packages it needs are
981         in sys.path.
982
983         https://bugs.webkit.org/show_bug.cgi?id=49419
984
985         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
986
987 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
988
989         Reviewed by Adam Roben.
990
991         Fix NRWT to respect set-webkit-configuration again :(
992
993         This change fixes a typo in config.py that was causing the wrong
994         value to be read initially and us never actually looking into
995         the filesystem to get the default configuration.
996         
997         * Scripts/webkitpy/layout_tests/port/config.py:
998         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
999         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1000
1001 2010-11-11  Eric Seidel  <eric@webkit.org>
1002
1003         Reviewed by Adam Barth.
1004
1005         Split out Bug, Attachment and CommitterValidator from bugzilla.py
1006         https://bugs.webkit.org/show_bug.cgi?id=49403
1007
1008         CommitterValidator really had nothing to do with bugzilla.py.
1009         I've put it next to committers.py for now since its rather tied to that file.
1010
1011         * Scripts/webkitpy/common/config/committervalidator.py: Added.
1012         * Scripts/webkitpy/common/config/committervalidator_unittest.py: Added.
1013         * Scripts/webkitpy/common/net/bugzilla/__init__.py:
1014         * Scripts/webkitpy/common/net/bugzilla/attachment.py: Added.
1015         * Scripts/webkitpy/common/net/bugzilla/bug.py: Added.
1016         * Scripts/webkitpy/common/net/bugzilla/bug_unittest.py: Added.
1017         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
1018         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py:
1019         * Scripts/webkitpy/tool/bot/feeders.py:
1020         * Scripts/webkitpy/tool/commands/queues.py:
1021
1022 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1023
1024         Reviewed by Eric Seidel.
1025
1026         fix missing 'pid' argument for http lockfile checking on win32
1027         https://bugs.webkit.org/show_bug.cgi?id=49363
1028
1029         * Scripts/webkitpy/common/system/executive.py:
1030
1031 2010-11-11  Eric Seidel  <eric@webkit.org>
1032
1033         Rubber-stamped by Adam Barth.
1034
1035         Move bugzilla.py into its own module in preparation for splitting one-file-per-class
1036         https://bugs.webkit.org/show_bug.cgi?id=49402
1037
1038         * Scripts/webkitpy/common/net/bugzilla/__init__.py: Added.
1039         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla.py.
1040         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla_unittest.py.
1041
1042 2010-11-10  Chris Guillory  <chris.guillory@google.com>
1043
1044         Reviewed by Chris Fleizach.
1045
1046         Use IAccessibleComparable to compare accessibility objects.
1047         https://bugs.webkit.org/show_bug.cgi?id=49118
1048
1049         * DumpRenderTree/AccessibilityUIElement.cpp:
1050         (AccessibilityUIElement::isEqual):
1051         * DumpRenderTree/AccessibilityUIElement.h:
1052         * DumpRenderTree/win/AccessibilityUIElementWin.cpp:
1053         (comparableObject):
1054         (AccessibilityUIElement::isEqual):
1055
1056 2010-11-10  Csaba Osztrogonác  <ossy@webkit.org>
1057
1058         Reviewed by David Hyatt.
1059
1060         HTML5 Ruby support should be mandatory feature
1061         https://bugs.webkit.org/show_bug.cgi?id=49272
1062
1063         Remove Ruby as optional feature.
1064
1065         * Scripts/build-webkit:
1066
1067 2010-11-10  Adam Roben  <aroben@apple.com>
1068
1069         Print VC++ Express build logs in the same order that the projects are
1070         built
1071
1072         Fixes <http://webkit.org/b/49326> print-vse-failure-logs prints logs
1073         in an unhelpful order
1074
1075         Reviewed by Eric Seidel.
1076
1077         * Scripts/print-vse-failure-logs:
1078         (PrintVisualStudioExpressLogs._build_order): Added. Uses
1079         print-msvc-project-dependencies to figure out the order in which
1080         projects are built.
1081         (PrintVisualStudioExpressLogs._sort_buildlogs): Added. Sorts the logs
1082         based on their build order and project name.
1083         (PrintVisualStudioExpressLogs._obj_directory): Moved code to find the
1084         scripts directory from here...
1085         (PrintVisualStudioExpressLogs._scripts_directory): ...to here.
1086         (PrintVisualStudioExpressLogs.main): Sort the logs before printing
1087         them.
1088
1089 2010-11-10  Adam Roben  <aroben@apple.com>
1090
1091         Only print the interesting text from VC++ Express build logs
1092
1093         Fixes <http://webkit.org/b/49325> It's hard to find the interesting
1094         output from print-vse-failure-logs
1095
1096         Reviewed by Eric Seidel.
1097
1098         * Scripts/print-vse-failure-logs: Removed unnecessary os.path import,
1099         added newly-required imports. Removed a comment that was essentially a
1100         FIXME, and is now fixed.
1101         (PrintVisualStudioExpressLogs._relevant_text): Added. Uses
1102         BeautifulSoup to extract the relevant text from the build log.
1103         (PrintVisualStudioExpressLogs.main): Only print the relevant text.
1104
1105 2010-11-09  Mihai Parparita  <mihaip@chromium.org>
1106
1107         Reviewed by Tony Chang.
1108
1109         [Chromium] http/tests/security/XFrameOptions fail with Chromium DRT
1110         https://bugs.webkit.org/show_bug.cgi?id=49286
1111         
1112         The tests in http/tests/security/XFrameOptions started to fail after
1113         r71297, since it changed the WebViewHost::assignIdentifierToRequest
1114         logic to always assign identifiers to requests, even if we're not going
1115         to be dumping resource load callbacks at that point. These tests
1116         only call LayoutTestController.dumpResourceLoadCallbacks after the
1117         request is started, thus their expectations have "<unknown>" as the
1118         identifier. Change WebViewHost to the previous behavior, which also
1119         match the other ports'.
1120
1121         * DumpRenderTree/chromium/WebViewHost.cpp:
1122         (WebViewHost::assignIdentifierToRequest):
1123         (WebViewHost::removeIdentifierForRequest):
1124
1125 2010-11-08  Tony Chang  <tony@chromium.org>
1126
1127         Reviewed by Adam Barth.
1128
1129         run platform/chromium/plugins/return-npobject.html on all platforms
1130         https://bugs.webkit.org/show_bug.cgi?id=49103
1131
1132         This tests that we can get an NPObject returned through a method on
1133         an NPAPI Object.
1134
1135         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1136         (testCallbackReturn):
1137         (pluginInvoke):
1138
1139 2010-11-09  Sheriff Bot  <webkit.review.bot@gmail.com>
1140
1141         Unreviewed, rolling out r71672.
1142         http://trac.webkit.org/changeset/71672
1143         https://bugs.webkit.org/show_bug.cgi?id=49276
1144
1145         It doesn't work on Chromium Windows bot (Requested by Ossy on
1146         #webkit).
1147
1148         * Scripts/webkitpy/common/system/file_lock.py: Removed.
1149         * Scripts/webkitpy/common/system/file_lock_unittest.py: Removed.
1150         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1151
1152 2010-11-09  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1153
1154         Reviewed by Tony Chang.
1155
1156         [NRWT] Make http locking similar to perl implementation
1157         https://bugs.webkit.org/show_bug.cgi?id=49187
1158
1159         * Scripts/webkitpy/common/system/file_lock.py: Added.
1160         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
1161         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1162
1163 2010-11-09  James Kozianski  <koz@chromium.org>
1164
1165         Reviewed by Eric Seidel.
1166
1167         Clean up imports for webkit-patch.
1168         https://bugs.webkit.org/show_bug.cgi?id=49083
1169
1170         Move some imports from commands into main.py where they are used.
1171         Move command imports into the __init__ of the commands module.
1172
1173         * Scripts/webkitpy/tool/commands/__init__.py:
1174         * Scripts/webkitpy/tool/commands/download.py:
1175         * Scripts/webkitpy/tool/commands/queues.py:
1176         * Scripts/webkitpy/tool/main.py:
1177
1178 2010-11-09  Yi Shen  <yi.4.shen@nokia.com>
1179
1180         Reviewed by Andreas Kling.
1181
1182         [Qt][QtTestBrowser] Don't erase incorrect url in the Url Bar
1183         https://bugs.webkit.org/show_bug.cgi?id=49047
1184
1185         * QtTestBrowser/mainwindow.cpp:
1186         (MainWindow::setAddressUrl):
1187
1188 2010-11-09  Andras Becsi  <abecsi@webkit.org>
1189
1190         Reviewed by Csaba Osztrogonác.
1191
1192         Improve old-run-webkit-tests --verbose to show DumpTool opening/closing
1193         https://bugs.webkit.org/show_bug.cgi?id=49252
1194
1195         * Scripts/old-run-webkit-tests:
1196
1197 2010-11-09  Csaba Osztrogonác  <ossy@webkit.org>
1198
1199         Reviewed by Tor Arne Vestbø.
1200
1201         Output of old-run-webkit-tests is buggy in verbose mode
1202         https://bugs.webkit.org/show_bug.cgi?id=49249
1203
1204         * Scripts/old-run-webkit-tests: Redundant outputs removed.
1205
1206 2010-11-08  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
1207
1208         Reviewed by Martin Robinson.
1209
1210         [GTK] Link with target name set does not work
1211         https://bugs.webkit.org/show_bug.cgi?id=48865
1212
1213         When a new page is created with a name (target=myFrame), the new
1214         mainFrame could not be found because they where not stored in the
1215         same PageGroup. As PageGroup are not exposed externally so the
1216         simpliest solution is to use a global page group name. This also fixes
1217         issue with visited link coloration across pages. After this change the
1218         private function webkit_web_view_set_group_name() was no longer used
1219         so it was removed completly.
1220
1221         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1222         (createWebView):
1223
1224 2010-11-08  John Knottenbelt  <jknotten@chromium.org>
1225
1226         Reviewed by Kent Tamura.
1227
1228         WebViewHost::reset() uses placement new.
1229         https://bugs.webkit.org/show_bug.cgi?id=49069
1230
1231         WebViewHost is using placement destruction / new to simulate a fresh
1232         WebViewHost object at the same address. This is because the WebView remains
1233         open across tests and maintains a pointer to the WebViewHost.
1234         This change resets member variables explictly instead of the placement new dance.
1235
1236         * DumpRenderTree/chromium/WebViewHost.cpp:
1237         (WebViewHost::WebViewHost):
1238         (WebViewHost::reset):
1239
1240 2010-11-08  Dirk Pranke <dpranke@chromium.org>
1241
1242         Reviewed by Eric Seidel.
1243
1244         Fix the webkit_base_dir logic in webkitpy/layout_tests/port/config.py. 
1245         It turns out that NRWT can't use the code in scm.find_checkout_root()
1246         because the Chromium bots don't do full checkouts of the WebKit
1247         tree; they only check out subdirectories like WebKitTools/Scripts.
1248         Until we can figure out a better approach for this, I've
1249         restored the base_dir-detecting code from NRWT, which works in
1250         any directory tree, scm or no.
1251
1252         This also restores the files modified in r71475 and r71474.
1253
1254         https://bugs.webkit.org/show_bug.cgi?id=49151
1255
1256         * WebKitTools/Scripts/webkitpy/layout_tests/port/config.py
1257         * WebKitTools/Scripts/webkitpy/layout_tests/port/config_unittest.py
1258
1259 2010-11-08  Eric Seidel  <eric@webkit.org>
1260
1261         Reviewed by Adam Barth.
1262
1263         webkit-patch failure-reason explodes if a build is missing
1264         https://bugs.webkit.org/show_bug.cgi?id=49195
1265
1266         This is likely a recent regression.  Adding a unit test for this case.
1267
1268         * Scripts/webkitpy/common/net/buildbot.py:
1269         * Scripts/webkitpy/common/net/buildbot_unittest.py:
1270
1271 2010-11-08  Dan Bernstein  <mitz@apple.com>
1272
1273         Reviewed by Darin Adler.
1274
1275         Ignore the system scrollbar setting.
1276
1277         * DumpRenderTree/mac/DumpRenderTree.mm:
1278         (resetDefaultsToConsistentValues): Set the scrollbar default.
1279
1280 2010-11-05  Adam Roben  <aroben@apple.com>
1281
1282         Make webkitpy.common.system.executive_unittest pass when running under
1283         Win32 Python
1284
1285         Fixes <http://webkit.org/b/49033>.
1286
1287         Reviewed by Dave Levin and Eric Seidel.
1288
1289         * Scripts/webkitpy/common/system/executive.py:
1290         (Executive._run_command_with_teed_output): Pass the arguments through
1291         encode_argument_if_needed rather than using Cygwin-specific code here.
1292         (Executive.run_and_throw_if_fail): Use child_process_encoding to decode
1293         the output.
1294         (Executive.run_command): Use encode_argument_if_needed to encode the
1295         arguments and child_process_encoding to decode the output.
1296         (Executive._child_process_encoding): Returns the encoding that should be
1297         used when communicating with child processes. On Windows we use mbcs,
1298         which maps to the current code page. On all other platforms we use
1299         UTF-8.
1300         (Executive._should_encode_child_process_arguments): Returns True if
1301         arguments to child processes need to be encoded. This is currently
1302         only needed on Cygwin and Win32 Python 2.x.
1303         (Executive._encode_argument_if_needed): Encode the argument using
1304         child_process_encoding if we need to encode arguments to child
1305         processes on this platform.
1306
1307         * Scripts/webkitpy/common/system/executive_unittest.py:
1308         (never_ending_command): Added. Returns arguments to run a command that
1309         will not quit until we kill it. On Windows we use wmic, on other
1310         platforms we use yes.
1311         (ExecutiveTest.test_run_command_with_unicode): Changed to expect the
1312         mbcs encoding to be used and for output from the child processes to
1313         have been roundtripped through encode/decode on Win32 Python. When
1314         UTF-8 is the encoding the roundtripping is undetectable, but with mbcs
1315         it's possible that some characters will not be able to be converted
1316         and will be replaced by question marks; the round-tripping allows us
1317         to expect this result.
1318
1319         (ExecutiveTest.test_kill_process):
1320         (ExecutiveTest.test_kill_all):
1321         Use never_ending_command instead of invoking "yes" directly. Expect an
1322         exit code of 1 when using Win32 Python, as that's what seems to happen.
1323
1324 2010-11-08  Adam Roben  <aroben@apple.com>
1325
1326         Roll out r71532
1327
1328         It broke the build for Cygwin 1.7 installs. Cygwin 1.7's default
1329         .bashrc unsets %TEMP%, which broke copy-tools.cmd.
1330
1331         * Scripts/webkitdirs.pm:
1332
1333 2010-11-08  Tony Chang  <tony@chromium.org>
1334
1335         Reviewed by Adam Barth.
1336
1337         run platform/chromium/plugins/nested-plugin-objects.html on all platforms
1338         https://bugs.webkit.org/show_bug.cgi?id=49094
1339
1340         This tests that objects created by plugins are proplery cleaned up.
1341
1342         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1343         (testPassTestObject):
1344         (pluginInvoke):
1345         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
1346         (testAllocate):
1347         (testDeallocate):
1348         (testGetProperty):
1349         (testConstruct):
1350
1351 2010-11-08  Adam Roben  <aroben@apple.com>
1352
1353         Mark Windows builds triggered from Perl as being non-interactive
1354
1355         This affects whether some of our scripts will show alerts vs. printing
1356         to the build log.
1357
1358         Fixes <http://webkit.org/b/49181> Windows build fail mysteriously when
1359         .vsprops files are updated
1360
1361         Reviewed by Steve Falkenburg.
1362
1363         * Scripts/webkitdirs.pm:
1364         (buildVisualStudioProject): Set WEBKIT_NONINTERACTIVE_BUILD to 1.
1365
1366 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1367
1368         Reviewed by Ojan Vafai.
1369
1370         Make http locking default in NRWT.
1371         https://bugs.webkit.org/show_bug.cgi?id=48053
1372
1373         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1374         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1375
1376 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1377
1378         Reviewed by Csaba Osztrogonác.
1379
1380         [NRWT] If the http lock fails we shouldn't do any locking
1381         https://bugs.webkit.org/show_bug.cgi?id=49164
1382
1383         If something goes wrong with the locking, the test should keep going.
1384
1385         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1386
1387 2010-11-08  Adam Roben  <aroben@apple.com>
1388
1389         Switch back to using kCGImageAlphaPremultipliedFirst when generating
1390         pixel dumps on Windows
1391
1392         I changed this behavior in r71418 thinking that it was required for
1393         getting plugins to show up in pixel dumps. But it doesn't seem to be
1394         necessary, and was making it impossible to compare new Windows pixel
1395         dumps with existing Windows or Mac pixel dumps (because ImageDiff won't
1396         compare an image with alpha to an image without alpha).
1397
1398         Fixes <http://webkit.org/b/49172> REGRESION (r71418): Can't compare
1399         new Windows pixel results to existing Windows or Mac results
1400
1401         Reviewed by Antti Koivisto.
1402
1403         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
1404         (createBitmapContextFromWebView): Replaced kCGImageAlphaNoneSkipFirst
1405         with kCGImageAlphaPremultipliedFirst.
1406
1407 2010-11-08  Csaba Osztrogonac  <ossy@webkit.org>
1408
1409         Unreviewed, rolling out r71466.
1410         http://trac.webkit.org/changeset/71466
1411         https://bugs.webkit.org/show_bug.cgi?id=48865
1412
1413         It broke layout tests on GTK bots.
1414
1415         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1416         (createWebView):
1417
1418 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1419
1420         Reviewed by Csaba Osztrogonác.
1421
1422         Enable running of Qt API tests on BuildBot
1423         https://bugs.webkit.org/show_bug.cgi?id=49004
1424
1425         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1426
1427 2010-11-08  Eric Seidel  <eric@webkit.org>
1428
1429         Reviewed by Adam Barth.
1430
1431         Add clean-review-queue command to remove closed bugs from the webkit.org/pending-review
1432         https://bugs.webkit.org/show_bug.cgi?id=49160
1433
1434         Bugzilla doesn't automatically remove r? when a bug gets closed.
1435         This script takes care of that for webkit.org.
1436
1437         * Scripts/webkitpy/common/net/bugzilla.py:
1438         * Scripts/webkitpy/tool/commands/upload.py:
1439
1440 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
1441
1442         Unreviewed, rolling out r71474.
1443         http://trac.webkit.org/changeset/71474
1444         https://bugs.webkit.org/show_bug.cgi?id=48280
1445
1446         breaks chromium webkit tests
1447         https://bugs.webkit.org/show_bug.cgi?id=49151
1448
1449         * Scripts/webkitpy/layout_tests/port/base.py:
1450         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
1451         * Scripts/webkitpy/layout_tests/port/config.py:
1452         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1453         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
1454         * Scripts/webkitpy/layout_tests/port/webkit.py:
1455         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1456         * Scripts/webkitpy/tool/mocktool.py:
1457
1458 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
1459
1460         Unreviewed, rolling out r71475.
1461         http://trac.webkit.org/changeset/71475
1462  
1463         breaks chromium webkit tests
1464         https://bugs.webkit.org/show_bug.cgi?id=49151
1465
1466         * Scripts/webkitpy/common/newstringio.py: Removed.
1467         * Scripts/webkitpy/common/newstringio_unittest.py: Removed.
1468         * Scripts/webkitpy/common/system/executive_mock.py: Removed.
1469         * Scripts/webkitpy/common/system/filesystem_mock.py: Removed.
1470         * Scripts/webkitpy/layout_tests/port/config_mock.py: Removed.
1471
1472 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1473
1474         Unreviewed, build fix.
1475
1476         Add files inexplicably not committed in r71474 as part of the
1477         fix for bug 48280.
1478
1479         * Scripts/webkitpy/common/newstringio.py: Added.
1480         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
1481         * Scripts/webkitpy/common/system/executive_mock.py: Added.
1482         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
1483         * Scripts/webkitpy/layout_tests/port/config_mock.py: Added.
1484
1485 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1486
1487         Reviewed by Eric Siedel.
1488
1489         new-run-webkit-tests: update port/base and port/webkit to use the
1490         new FileSystem and Config abstractions, pulling more logic out of
1491         the base Port classes into separate, mockable objects.
1492
1493         Also create a MockFileSystem object, a MockConfig object, move
1494         MockExecutive into common/system to be next to executive, and
1495         update the config object to use a FileSystem.
1496
1497         https://bugs.webkit.org/show_bug.cgi?id=48280
1498
1499         * Scripts/webkitpy/common/newstringio.py: Added.
1500         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
1501         * Scripts/webkitpy/common/system/executive_mock.py: Added.
1502         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
1503         * Scripts/webkitpy/layout_tests/port/base.py:
1504         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
1505         * Scripts/webkitpy/layout_tests/port/config.py:
1506         * Scripts/webkitpy/layout_tests/port/config_mock.py:
1507         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1508         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
1509         * Scripts/webkitpy/layout_tests/port/webkit.py:
1510         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1511         * Scripts/webkitpy/tool/mocktool.py:
1512
1513 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1514
1515         Unreviewed, build breakage.
1516
1517         Apparently I uploaded the wrong version of the file to fix 49122
1518         and neither Eric or I noticed - it was missing a dirname() call.
1519         Fixing ...
1520
1521         https://bugs.webkit.org/show_bug.cgi?id=49122
1522
1523         * Scripts/webkitpy/common/checkout/scm.py:
1524
1525 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1526
1527         Reviewed by Eric Seidel.
1528
1529         webkitpy/tool/* unittests change cwd and don't clean up properly
1530
1531         https://bugs.webkit.org/show_bug.cgi?id=49122
1532
1533         * Scripts/webkitpy/common/checkout/scm.py:
1534
1535 2010-11-05  Chris Marrin  <cmarrin@apple.com>
1536
1537         Reviewed by Simon Fraser.
1538
1539         suspendAnimations/resumeAnimations not present in WebKit2
1540         https://bugs.webkit.org/show_bug.cgi?id=49109
1541
1542         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
1543         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
1544         (WTR::LayoutTestController::suspendAnimations):
1545         (WTR::LayoutTestController::resumeAnimations):
1546         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
1547
1548 2010-11-05  Tony Chang  <tony@chromium.org>
1549
1550         Reviewed by David Levin.
1551
1552         cleanup style in TestNetscapePlugIn/PluginObject.cpp
1553         https://bugs.webkit.org/show_bug.cgi?id=49044
1554
1555         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1556         (getPluginClass):
1557         (pluginGetProperty):
1558         (pluginSetProperty):
1559         (variantToIdentifier):
1560         (testCallback):
1561         (getURL):
1562         (getURLNotify):
1563         (testInvokeDefault):
1564         (testGetProperty):
1565         (testDocumentOpen):
1566         (testWindowOpen):
1567         (pluginInvoke):
1568         (handleCallback):
1569         (notifyStream):
1570         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
1571
1572 2010-11-05  Eric Seidel  <eric@webkit.org>
1573
1574         Reviewed by David Levin.
1575
1576         Add QueueStatusServer/__init__.py so others can run the QueueStatusServer tests
1577         https://bugs.webkit.org/show_bug.cgi?id=49032
1578
1579         I wrote this file as part of bug 47847, but I forgot to commit it.
1580         No one else noticed it missing because test-webkitpy knows how
1581         to recover in the case where it can't import QueueStatusServer
1582         (which generally occurs due to not having installed the AppEngine SDK).
1583
1584         * QueueStatusServer/__init__.py: Added.
1585         * QueueStatusServer/model/workitems_unittest.py:
1586          - Remove a test which fails.  This was probably landed (by me)
1587            from my other machine, which since this __init__.py was missing
1588            I never noticed the failure and landed this invalid test.
1589            Sadly we can't really test remove_work_item as it depends
1590            on .key() working.  .key() will throw unless the object
1591            has already been saved it seems.
1592            This may be a mis-design in our remove_work_item implementation,
1593            but for now, just removing the test.
1594
1595 2010-11-04  Adam Roben  <aroben@apple.com>
1596
1597         Teach check-webkit-style about TestNetscapePlugIn
1598
1599         Fixes <http://webkit.org/b/49030> check-webkit-style is ignorant of
1600         TestNetscapePlugIn's coding conventions
1601
1602         Reviewed by John Sullivan.
1603
1604         * Scripts/webkitpy/style/checker.py: Lump TestNetscapePlugIn in with
1605         WebKitAPITest and TestWebKitAPI in having different include order and
1606         naming conventions than WebCore. Added some comments about why the
1607         exceptions exist.
1608
1609 2010-11-04  Adam Roben  <aroben@apple.com>
1610
1611         Add a test that shows that windowed plugins are able to paint
1612
1613         Somehow we never had a test for this before.
1614
1615         Fixes <http://webkit.org/b/49029> Should add a test that shows
1616         windowed plugins can paint
1617
1618         Reviewed by Jon Honeycutt.
1619
1620         * DumpRenderTree/TestNetscapePlugIn/Tests/win/DrawsGradient.cpp: Added.
1621         (DrawsGradient::DrawsGradient):
1622         (DrawsGradient::wndProc): We handle the WM_PAINT and WM_PRINTCLIENT messages.
1623
1624         (DrawsGradient::onPaint):
1625         (DrawsGradient::onPrintClient):
1626         These both just call through to paint.
1627
1628         (DrawsGradient::paint): Fills our client area with some gradients.
1629
1630         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Link
1631         against Msimg32.lib for ::GradientFill and added DrawsGradient.
1632
1633         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
1634         (createBitmapContextFromWebView): Use WM_PRINT instead of
1635         WM_PRINTCLIENT so that ::DefWindowProc will send
1636         WM_PRINT/WM_PRINTCLIENT messages to the WebView's child windows.
1637         Replaced kCGImageAlphaPremultipledFirst with kCGImageAlphaNoneSkipFirst
1638         because GDI doesn't support alpha and kCGImageBitmapOrder32Little with
1639         kCGImageBitmapOrder32Host because, who knows, maybe someday Windows
1640         will run on a big-endian platform.
1641
1642 2010-11-04  Adam Roben  <aroben@apple.com>
1643
1644         Extract much of NPNInvalidateRectInvalidatesWindow's code into a
1645         WindowedPluginTest base class
1646
1647         The base class takes care of subclassing the plugin's window so that a
1648         custom WNDPROC is called. This will make it easier to write tests that
1649         need to handle window messages.
1650
1651         Fixes <http://webkit.org/b/49028> It's hard to write a PluginTest with
1652         a custom WNDPROC
1653
1654         Reviewed by Jon Honeycutt.
1655
1656         * DumpRenderTree/TestNetscapePlugIn/Tests/win/NPNInvalidateRectInvalidatesWindow.cpp:
1657         (NPNInvalidateRectInvalidatesWindow::NPNInvalidateRectInvalidatesWindow):
1658         (NPNInvalidateRectInvalidatesWindow::NPP_SetWindow):
1659         (NPNInvalidateRectInvalidatesWindow::wndProc):
1660         (NPNInvalidateRectInvalidatesWindow::testInvalidateRect):
1661         Moved code from here to WindowedPluginTest. Changed to use window()
1662         instead of m_window.
1663
1664         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1665         Added WindowedPluginTest and added TestNetscapePlugIn/win to the
1666         include path. Also reordered the include path to make a little more
1667         sense and simplified the entry that added TestNetscapePlugIn itself to
1668         the include path.
1669
1670         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.cpp: Added.
1671         (WindowedPluginTest::WindowedPluginTest):
1672         (WindowedPluginTest::NPP_SetWindow):
1673         (WindowedPluginTest::staticWndProc):
1674         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.h: Added.
1675         (WindowedPluginTest::window):
1676         Code came from NPNInvalidateRectInvalidatesWindow.
1677
1678 2010-11-04  Adam Roben  <aroben@apple.com>
1679
1680         Add a plugin test to show that windowed plugins are clipped correctly
1681
1682         Fixes <http://webkit.org/b/49024> <rdar://problem/8487847> Windowed
1683         plugins aren't clipped in WebKit2 on Windows
1684
1685         Reviewed by Jon Honeycutt.
1686
1687         * DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowRegionIsSetToClipRect.cpp: Added.
1688         (WindowRegionIsSetToClipRect::WindowRegionIsSetToClipRect): Initialize members.
1689         (WindowRegionIsSetToClipRect::NPP_SetWindow): Check that our window
1690         region matches the clip rect we know we should have based on
1691         window-region-is-set-to-clip-rect.html, and check that our window class
1692         doesn't have the CS_PARENTDC style.
1693
1694         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1695         Added WindowRegionIsSetToClipRect.
1696
1697 2010-11-05  Alejandro G. Castro  <alex@igalia.com>
1698
1699         Reviewed by Martin Robinson.
1700
1701         [GTK] Avoid font hinting in the DRT
1702         https://bugs.webkit.org/show_bug.cgi?id=48548
1703
1704         Change the settings to avoid font hinting, it was causing
1705         different results depending on the distribution.
1706
1707         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1708         (initializeGtkFontSettings):
1709         * DumpRenderTree/gtk/fonts/fonts.conf:
1710
1711 2010-11-05  Eric Seidel  <eric@webkit.org>
1712
1713         Reviewed by Adam Barth.
1714
1715         build-webkit spams system.log by repeatedly checking for PBXProductDirectory in com.apple.Xcode
1716         https://bugs.webkit.org/show_bug.cgi?id=49051
1717
1718         This is a speculative fix.  The unit tests cover these methods, however
1719         I don't know if this will fully stop the system.log spam.
1720
1721         * Scripts/webkitpy/layout_tests/port/base.py:
1722         * Scripts/webkitpy/layout_tests/port/webkit.py:
1723
1724 2010-11-05  Eric Seidel  <eric@webkit.org>
1725
1726         Reviewed by Adam Barth.
1727
1728         Add basic support for showing bot id on /queue-status/ pages
1729         https://bugs.webkit.org/show_bug.cgi?id=49037
1730
1731         This support is really simple.  Eventually we'll want to
1732         show the bot id in the lock table too, but we don't have
1733         that information stored in the server yet.
1734
1735         * QueueStatusServer/handlers/queuestatus.py:
1736         * QueueStatusServer/handlers/queuestatus_unittest.py: Added.
1737         * QueueStatusServer/templates/includes/singlequeuestatus.html:
1738
1739 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
1740
1741         Reviewed by Adam Barth.
1742
1743         Create a filesystem wrapper that we can use to enforce
1744         particular conventions and use for mocking and dependency
1745         injection down the line.
1746
1747         https://bugs.webkit.org/show_bug.cgi?id=48144
1748
1749         * Scripts/webkitpy/common/system/filesystem.py: Added.
1750         * Scripts/webkitpy/common/system/filesystem_unittest.py: Added.
1751
1752 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1753
1754         Reviewed by Tony Chang.
1755
1756         run_webkit_tests_unittest fails under Python 2.5
1757         https://bugs.webkit.org/show_bug.cgi?id=49043
1758         
1759         Switch from itertools.chain.from_iterable (which was added in 2.6)
1760         to using itertools.chain directly.
1761
1762         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1763
1764 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1765
1766         Unreviewed fix to webkit-patch rebaseline-server so that it runs under
1767         Python 2.5 (it needs an import to use the with statement).
1768
1769         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
1770
1771 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
1772
1773         Reviewed by Adam Barth.
1774
1775         new-run-webkit-tests: split out webkit-specific configuration stuff into a new module
1776
1777         The current NRWT code has webkit-specific configuration code (like
1778         _script_path, default configuration, etc.) mixed in with
1779         layout-test-specific stuff in port/base. The configuration code
1780         should be split out into a separate module for easier mocking,
1781         testing, and isolation.
1782
1783         https://bugs.webkit.org/show_bug.cgi?id=48264
1784
1785         * Scripts/webkitpy/layout_tests/port/config.py: Added.
1786         * Scripts/webkitpy/layout_tests/port/config_unittest.py: Added.
1787
1788 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1789
1790         Reviewed by Tony Chang.
1791
1792         Rebaseline server: initial framework
1793         https://bugs.webkit.org/show_bug.cgi?id=48892
1794         
1795         Adds the basic framework for the rebaseline server (details at 
1796         http://webkit.org/b/47761). Includes the rebaseline-server webkit-patch
1797         command, which starts an HTTP server that can serve static files or
1798         invoke handler methods on a class.
1799
1800         * Scripts/webkitpy/tool/commands/__init__.py:
1801         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html: Added.
1802         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css: Added.
1803         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js: Added.
1804         * Scripts/webkitpy/tool/commands/rebaselineserver.py: Added.
1805
1806 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
1807
1808         Reviewed by James Robinson.
1809
1810         new-run-webkit-tests wasn't using DRT by default for
1811         --platform chromium-gpu
1812
1813         The default value was set to False instead of None, which meant
1814         that the platform specific logic wasn't firing to change the
1815         value to True (b/c we were afraid we'd be overriding the user
1816         preference).
1817
1818         https://bugs.webkit.org/show_bug.cgi?id=49038
1819
1820         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1821
1822 2010-11-04  Tony Chang  <tony@chromium.org>
1823
1824         Reviewed by Adam Barth.
1825
1826         make platform/chromium/plugins/multiple-plugins.html pass on all platforms
1827         https://bugs.webkit.org/show_bug.cgi?id=49026
1828
1829         Have the testObject.bar property return the string "bar".  This lets
1830         us run and pass platform/chromium/plugins/multiple-plugins.html on all
1831         platforms.
1832
1833         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
1834         (testGetProperty):
1835
1836 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1837
1838         Reviewed by Tony Chang.
1839
1840         NRWT doesn't have good test coverage for --run-chunk, --batch-size, --run-part, etc.
1841         https://bugs.webkit.org/show_bug.cgi?id=48878
1842         
1843         Add get_tests_run so that it's easy to see which tests get run (and with
1844         what batching) for a given flag combination. Flesh out the various
1845         test cases that have FIXMEs.
1846         
1847         Also fixes an off-by-one error (batch sizes were one larger than 
1848         expected) and makes --run-part also have wraparound behavior, like
1849         --run-chunk.
1850
1851         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1852         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1853         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1854
1855 2010-11-04  Erik Arvidsson  <arv@chromium.org>
1856
1857         Reviewed by Dimitri Glazkov.
1858
1859         Support box-sizing without the vendor prefix
1860         https://bugs.webkit.org/show_bug.cgi?id=36713
1861
1862         Based on patch by Peter Beverloo <peter@lvp-media.com>
1863
1864         * iExploder/htdocs/cssproperties.in: Change -webkit-box-sizing to box-sizing.
1865
1866 2010-11-04  Csaba Osztrogonác  <ossy@webkit.org>
1867
1868         Unreviewed rollout r71340, because it broke Chromium Windows bot.
1869
1870         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1871         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1872
1873 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1874
1875         Reviewed by Ojan Vafai.
1876
1877         Make http locking default in NRWT.
1878         https://bugs.webkit.org/show_bug.cgi?id=48053
1879
1880         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1881         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1882
1883 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1884
1885         Reviewed by Eric Seidel.
1886
1887         [NRWT] Clear invalid http locks on Windows platform as well
1888         https://bugs.webkit.org/show_bug.cgi?id=48515
1889
1890         * Scripts/webkitpy/common/system/executive.py:
1891         * Scripts/webkitpy/common/system/executive_unittest.py:
1892         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1893
1894 2010-11-02  Adam Roben  <aroben@apple.com>
1895
1896         Reduce our dependence on coreutils when running Python tests
1897
1898         This patch introduces versions of the standard echo and cat utilities
1899         implemented in Python. They are probably missing features of their
1900         coreutils equivalents, but they can do what's necessary for our Python
1901         tests. This is useful on Windows, where these utilities typically
1902         aren't available.
1903
1904         Fixes <http://webkit.org/b/48883> executive_unittest relies on echo
1905         and cat utilities from coreutils, which are not present on Windows
1906
1907         Reviewed by Eric Seidel.
1908
1909         * Scripts/webkitpy/common/system/executive_unittest.py: Changed to use
1910         our Python-based echo and cat.
1911
1912         * Scripts/webkitpy/common/system/fileutils.py: Added.
1913         (make_stdout_binary): On Windows, puts sys.stdout into binary mode so
1914         that \n won't be translated into \r\n. I couldn't think of a good way
1915         to test this directly without touching the filesystem, but it is tested
1916         indirectly by echo_unittest.
1917
1918         * Scripts/webkitpy/test/cat.py: Added.
1919         (command_arguments): Returns a list for invoking cat with the given arguments.
1920         (main): Acts like a simplified version of the coreutils cat utility.
1921
1922         * Scripts/webkitpy/test/cat_unittest.py: Added.
1923         (CatTest.assert_cat): Runs cat with the given input and ensures the
1924         output matches the input.
1925         (CatTest.test_basic): Performs a simple test of cat.
1926         (CatTest.test_no_newline): Tests what happens when the input string
1927         doesn't have a trailing newline.
1928         (CatTest.test_unicode): Tests passing a unicode string to cat.
1929         (CatTest.test_as_command): Tests running cat as a separate command.
1930
1931         * Scripts/webkitpy/test/echo.py: Added.
1932         (command_arguments): Returns a list for invoking echo with the given arguments.
1933         (main): Acts like a simplified version of the coreutils echo utility.
1934
1935         * Scripts/webkitpy/test/echo_unittest.py: Added.
1936         (EchoTest.test_basic): Performs a simple test of echo.
1937         (EchoTest.test_no_newline): Tests passing -n to echo to suppress the
1938         trailing newline.
1939         (EchoTest.test_unicode): Tests passing unicode and non-unicode strings
1940         to echo.
1941         (EchoTest.test_argument_order): Tests what happens when -n is not the
1942         first argument.
1943         (EchoTest.test_empty_arguments): Tests what happens when you pass [] to
1944         echo.main.
1945         (EchoTest.test_no_arguments): Tests what happens when you call
1946         echo.main with no arguments.
1947         (EchoTest.test_as_command): Tests running echo as a separate command.
1948
1949 2010-11-04  Renata Hodovan  <reni@inf.u-szeged.hu>
1950
1951         Unreviewed: Add myself to the list of Committers.
1952
1953         * Scripts/webkitpy/common/config/committers.py:
1954
1955 2010-11-04  Andreas Kling  <kling@webkit.org>
1956
1957         Reviewed by Kenneth Rohde Christiansen.
1958
1959         [Qt] Hook into QWebPage::scrollRequested for resizes-to-contents mode
1960
1961         In RTC mode the QGraphicsWebView item is the size of the contents,
1962         scrolling works a bit differently  (we need to react to scrollRequested.)
1963
1964         Normally QGraphicsView will replay the last mouse event when scrolling,
1965         so to prevent WebKit from getting confused by this we temporarily make
1966         the QGraphicsView non-interactive.
1967
1968         * QtTestBrowser/webview.cpp:
1969         (WebViewGraphicsBased::setPage):
1970         (WebViewGraphicsBased::scrollRequested):
1971         * QtTestBrowser/webview.h:
1972
1973 2010-11-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
1974
1975         Reviewed by Andreas Kling.
1976
1977         Use OS(MAC_OS_X) rather than PLATFORM(MAC) in TestControllerQt
1978
1979         PLATFORM(MAC) is not defined for the Qt port, as it refers to
1980         the Mac-port, not the Mac OS X operating system.
1981
1982         * WebKitTestRunner/qt/TestControllerQt.cpp:
1983         (WTR::isExistingLibrary):
1984
1985 2010-11-04  Leonid Ebril  <leonid.ebril@nokia.com>
1986
1987         Reviewed by Andreas Kling.
1988
1989         [Qt] Adding iphone user agent string the user agent list for QtTestBrowser
1990         https://bugs.webkit.org/show_bug.cgi?id=48869
1991
1992         * QtTestBrowser/useragentlist.txt:
1993
1994 2010-11-03  Adam Roben  <aroben@apple.com>
1995
1996         Always use uppercase drive names in strings returned by abspath_to_uri
1997
1998         Some versions of cygpath use lowercase drive letters while others use
1999         uppercase, which makes it hard to test the output of code that uses
2000         cygpath.
2001
2002         Fixes <http://webkit.org/b/48914> webkitpy.common.system.path_unittest
2003         fails with Cygwin 1.5
2004
2005         Reviewed by Eric Seidel.
2006
2007         * Scripts/webkitpy/common/system/path.py:
2008         (cygpath): Updated the docstring to indicate that only absolute paths
2009         should be passed for now (though relative paths will work fine).
2010         (_Cygpath.convert): Upper-case the first letter of the converted Windows path.
2011
2012 2010-11-03  George Guo  <George.Guo@Nokia.com>
2013
2014         Reviewed by Andreas Kling.
2015
2016         [Qt] QtTestBrowser : set mmp rule pageddata in Symbian
2017         https://bugs.webkit.org/show_bug.cgi?id=48767
2018     
2019         Paging is needd on Symbian devices to support benchmarks tests like
2020         dromaeo.com and Celtic Kane that need a lot of memory to run  
2021
2022         * QtTestBrowser/QtTestBrowser.pro:
2023
2024 2010-11-03  Jenn Braithwaite  <jennb@chromium.org>
2025
2026         Reviewed by Dmitry Titov.
2027
2028         Chromium: Update resource tracking when moving a frame between documents
2029         https://bugs.webkit.org/show_bug.cgi?id=48363
2030
2031         * DumpRenderTree/chromium/WebViewHost.cpp:
2032         (WebViewHost::assignIdentifierToRequest):
2033         Always put resource id in map so we can make assumptions about its
2034         presence.
2035         (WebViewHost::removeIdentifierForRequest):
2036         Added.
2037         (WebViewHost::didFinishResourceLoad):
2038         (WebViewHost::didFailResourceLoad):
2039         Check existence of resource id before removing from map.
2040         * DumpRenderTree/chromium/WebViewHost.h:
2041
2042 2010-11-03  Victor Wang  <victorw@chromium.org>
2043
2044         Reviewed by Adam Barth.
2045
2046         [Chromium] update buildbot names in chromium rebaseline tool.
2047
2048         https://bugs.webkit.org/show_bug.cgi?id=48881
2049
2050         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2051
2052 2010-11-03  Darin Adler  <darin@apple.com>
2053
2054         Updated Xcode projects by opening them with Xcode 3.2.4.
2055         Updated svn:ignore for Xcode projects.
2056
2057         * MiniBrowser/MiniBrowser.xcodeproj: Added property svn:ignore.
2058         * TestWebKitAPI/TestWebKitAPI.xcodeproj: Added property svn:ignore.
2059         * WebKitLauncher/WebKitLauncher.xcodeproj: Modified property svn:ignore.
2060         * WebKitLauncher/WebKitLauncher.xcodeproj/project.pbxproj: Updated with Xcode 3.2.4.
2061
2062 2010-11-03  Adam Roben  <aroben@apple.com>
2063
2064         Stop waiting for messages from the web process after a timeout elapses
2065
2066         TestController::runUntil now takes a timeout parameter that specifies
2067         how long to wait for the condition to become true. WebKitTestRunner
2068         takes a --timeout flag that specifies how long the timeout should be.
2069         run-webkit-tests passes this flag to WebKitTestRunner so its timeout
2070         will be similar to run-webkit-tests.
2071
2072         Fixes <http://webkit.org/b/43047> <rdar://problem/8365833>
2073         WebKitTestRunner waits forever if the web process crashes
2074
2075         Reviewed by Darin Adler and Anders Carlsson.
2076
2077         * Scripts/old-run-webkit-tests:
2078         (top level): Moved the GuardMalloc timeout adjustment here from
2079         readFromDumpToolWithTimer.
2080         (openDumpTool): Make WTR use a timeout similar to but slightly shorter
2081         than the one that was specified on the command line.
2082
2083         * WebKitTestRunner/TestController.cpp:
2084         (WTR::TestController::TestController): Initialize our timeout values.
2085         (WTR::TestController::initialize): Parse the --timeout flag and use it
2086         to modify our timeout values.
2087         (WTR::TestController::resetStateToConsistentValues): Changed to use a
2088         short timeout while waiting for the web process to reset and to return
2089         a boolean indicating whether we were able to reset the web process.
2090         Uses a 5-second timeout while waiting for the process to be reset.
2091         (WTR::TestController::runTest): Changed to return a boolean indicating
2092         whether we were able to reset the web process (and thus run the test).
2093         (WTR::TestController::runUntil): Call through to platformRunUntil.
2094
2095         (WTR::TestController::runTestingServerLoop):
2096         (WTR::TestController::run):
2097         Changed to bail if any test can't be run. This will cause the process
2098         to exit. (Unfortunately this will make run-webkit-tests think we
2099         crashed; see <http://webkit.org/b/48943>.)
2100
2101         * WebKitTestRunner/TestController.h: Added platformRunUntil,
2102         m_longTimeout, and m_shortTimeout.
2103
2104         * WebKitTestRunner/TestInvocation.cpp:
2105         (WTR::TestInvocation::invoke): Use a short timeout when waiting for the
2106         initial response and a long timeout when waiting for the test to
2107         complete. Check whether runUntil timed out and print an error message
2108         if so.
2109
2110         * WebKitTestRunner/mac/TestControllerMac.mm:
2111         (WTR::TestController::platformRunUntil): Renamed from runUntil. Pass
2112         [NSDate distantPast] to -[NSRunLoop runMode:beforeDate:] so that we
2113         won't block waiting for the run loop. Only loop until the timeout
2114         elapses.
2115
2116         * WebKitTestRunner/qt/TestControllerQt.cpp:
2117         (WTR::TestController::platformRunUntil): Renamed from runUntil. Added a
2118         FIXME about honoring the timeout.
2119
2120         * WebKitTestRunner/win/TestControllerWin.cpp:
2121         (WTR::TestController::platformRunUntil): Renamed from runUntil. Use
2122         ::MsgWaitForMultipleObjectsEx to implement the timeout. Changed to use
2123         ::PeekMessageW so that we don't block waiting for messages to become
2124         available.
2125
2126 2010-11-03  Adam Roben  <aroben@apple.com>
2127
2128         Add a plugin test that evaluates JS after removing the plugin element
2129         from the document
2130
2131         This test replaces platform/win/plugins/plugin-delayed-destroy.html.
2132         That test was made to prevent a crash very similar to this one, but
2133         unfortunately tested only the mechanism that prevented the crash and
2134         not whether the crash itself was prevented. Since WebKit2 uses a
2135         different mechanism to prevent the crash, the test was failing even
2136         though WebKit2 was not vulnerable to the crash. This new test crashes
2137         if there is no mechanism in place to prevent it and passes in both
2138         WebKit1 and WebKit2.
2139
2140         Fixes <http://webkit.org/b/46711> <rdar://problem/8485903>
2141         platform/win/plugins/plugin-delayed-destroy.html fails in WebKit2
2142
2143         Reviewed by Anders Carlsson.
2144
2145         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
2146         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2147         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
2148         * GNUmakefile.am:
2149         Added new file.
2150
2151         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
2152         (pluginDeallocate): Make sure we delete the PluginTest object. This
2153         prevents a leak and also allows us to test the crash.
2154
2155         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
2156         (PluginTest::executeScript): Made this into a non-static member
2157         function.
2158
2159         (PluginTest::waitUntilDone):
2160         (PluginTest::notifyDone):
2161         Updated for changes to executeScript.
2162
2163         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Added executeScript.
2164
2165         * DumpRenderTree/TestNetscapePlugIn/Tests/EvaluateJSAfterRemovingPluginElement.cpp: Added.
2166         (EvaluateJSAfterRemovingPluginElement::EvaluateJSAfterRemovingPluginElement):
2167         Initialize ourselves and tell the test harness to wait.
2168         (EvaluateJSAfterRemovingPluginElement::NPP_DestroyStream): Remove our
2169         plugin element from the document, then execute some JavaScript. If
2170         WebKit does not have appropriate mechanisms in place, we'll be
2171         destroyed inside the first call to executeScript and crash on the
2172         second call.
2173
2174 2010-11-02  Stephen White  <senorblanco@chromium.org>
2175
2176         Reviewed by Tony Chang.
2177
2178         [chromium] Fix LayoutTestController UMRs.
2179         https://bugs.webkit.org/show_bug.cgi?id=48872
2180
2181         * DumpRenderTree/chromium/LayoutTestController.cpp:
2182         (LayoutTestController::LayoutTestController):
2183
2184 2010-11-03  Kent Tamura  <tkent@chromium.org>
2185
2186         Reviewed by Dimitri Glazkov.
2187
2188         REGRESSION: rebaseline-chromium-webkit-tests uses non-zero tolerance for
2189         image dup detection
2190         https://bugs.webkit.org/show_bug.cgi?id=48744
2191
2192         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2193          - Make a function for option parsing for ease of test
2194          - Set 0 to options.tolerance
2195         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
2196          - Add a test for this change
2197
2198 2010-11-02  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
2199
2200         Unreviewed: Add myself to the list of Committers.
2201
2202         * Scripts/webkitpy/common/config/committers.py:
2203
2204 2010-11-02  Anders Carlsson  <andersca@apple.com>
2205
2206         Fix build.
2207
2208         * TestWebKitAPI/PlatformUtilities.cpp:
2209         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
2210
2211 2010-11-02  Sam Weinig  <sam@webkit.org>
2212
2213         Reviewed by Anders Carlsson.
2214
2215         Add a way to send startup messages on the context which can be posted when a process launches
2216         <rdar://problem/8617928>
2217         https://bugs.webkit.org/show_bug.cgi?id=48838
2218
2219         * MiniBrowser/mac/WebBundle/WebBundleMain.m:
2220         (WKBundleInitialize):
2221         * TestWebKitAPI/InjectedBundleController.cpp:
2222         (TestWebKitAPI::InjectedBundleController::initialize):
2223         (TestWebKitAPI::InjectedBundleController::didReceiveMessage):
2224         * TestWebKitAPI/InjectedBundleController.h:
2225         * TestWebKitAPI/InjectedBundleMain.cpp:
2226         (WKBundleInitialize):
2227         * TestWebKitAPI/PlatformUtilities.cpp:
2228         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
2229         * WebKitTestRunner/InjectedBundle/InjectedBundleMain.cpp:
2230         (WKBundleInitialize):
2231         Update implementations of WKBundleInitialize to take an initial userData
2232         argument.  Change TestWebKitAPI to use the new initial userData to initialize
2233         each test's bundle.
2234
2235 2010-11-02  Benjamin Kalman  <kalman@google.com>
2236
2237         Reviewed by Ojan Vafai.
2238
2239         new-run-webkit-tests doesn't strip "LayoutTests/" from prefix, unlike old-run-webkit-tests
2240         https://bugs.webkit.org/show_bug.cgi?id=48794
2241
2242         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: Strip the "LayoutTests/" prefix from test argument paths.
2243
2244 2010-11-02  Adam Roben  <aroben@apple.com>
2245
2246         Skip webkitpy.layout_tests.run_webkit_tests_unittest.MainTest on Cygwin
2247         Python 2.5.x
2248
2249         It is known to hang on that version of Python. See
2250         <http://webkit.org/b/48614>.
2251
2252         Reviewed by Adam Barth.
2253
2254         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py: Use
2255         skip_if to skip MainTest on Cygwin Python 2.5.x.
2256
2257         * Scripts/webkitpy/test/skip.py: Added.
2258         (skip_if): If the passed-in condition is false, find all the test_*
2259         methods of the given class and replace them with a function that just
2260         logs that we're skipping these tests. This is loosely based on the
2261         unittest.skip_if decorator added in Python 3.1.
2262         (_skipped_method): Returns a function that just logs that the tests are
2263         being skipped. This is used in place of the actual test_* functions.
2264
2265         * Scripts/webkitpy/test/skip_unittest.py: Added.
2266         (SkipTest.setUp): Create a logger for ourselves and give it a handler
2267         that logs to a private stream.
2268         (SkipTest.tearDown): Reset the logger.
2269         (SkipTest.create_fixture_class): Returns a class that we use to test
2270         skip_if. It has a single test_* method, test_foo, that just calls
2271         through to a callback.
2272         (SkipTest.foo_callback): Record that test_foo was called.
2273         (SkipTest.test_skip_if_false): Pass skip_if a False condition and test
2274         that test_foo does get called.
2275         (SkipTest.test_skip_if_true): Pass skip_if a True condition and test
2276         that test_foo does not get called and the appropriate message gets
2277         logged.
2278
2279 2010-11-02  Adam Barth  <abarth@webkit.org>
2280
2281         Reviewed by Eric Seidel.
2282
2283         webkit-patch should tell check-webkit-style which files were changed so
2284         check-webkit-style doesn't have to stat the whole working copy again
2285         https://bugs.webkit.org/show_bug.cgi?id=48792
2286
2287         * Scripts/webkitpy/tool/mocktool.py:
2288         * Scripts/webkitpy/tool/steps/checkstyle.py:
2289
2290 2010-11-02  Robert Kroeger  <rjkroege@chromium.org>
2291
2292         Reviewed by James Robinson.
2293
2294         [Chromium/DRT] Make EventSendingController honour leapForward for touch events.
2295         https://bugs.webkit.org/show_bug.cgi?id=48777
2296
2297         * DumpRenderTree/chromium/EventSender.cpp:
2298         (EventSender::sendCurrentTouchEvent):
2299
2300 2010-11-02  Adam Roben  <aroben@apple.com>
2301
2302         Only track resource identifiers in DRT when dumpResourceLoadCallbacks
2303         is on
2304
2305         This reverts Windows to our pre-r71097 behavior. That patch made us
2306         track all resource identifiers, including the main resource, so the
2307         main resource's URL started appearing in test output instead of
2308         "<unknown>". Arguably having the main resource's URL is better, but all
2309         other platforms print "<unknown>" and we want to match.
2310
2311         Fixes <http://webkit.org/b/48837> <rdar://problem/8620351> REGRESSION
2312         (r71097): Many http tests failing on Windows
2313
2314         Reviewed by Anders Carlsson.
2315
2316         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
2317         (ResourceLoadDelegate::identifierForInitialRequest): Don't add the
2318         identifier to the URL map if we're not supposed to dump resource load
2319         callbacks.
2320         (ResourceLoadDelegate::removeIdentifierForRequest): Always remove the
2321         identifier from the URL map even if we're already "done". There's no
2322         point in keeping out-of-date identifiers around.
2323
2324 2010-11-01  Jenn Braithwaite  <jennb@chromium.org>
2325
2326         Reviewed by Adam Roben.
2327
2328         Windows: Update resource tracking when moving a frame between documents
2329         https://bugs.webkit.org/show_bug.cgi?id=48364
2330
2331         * DumpRenderTree/win/DumpRenderTree.cpp:
2332         (createWebViewAndOffscreenWindow):
2333         (main):
2334         Give each WebView its own ResourceLoadDelegate instance in order to
2335         make assertions about resource ids on a particular WebView.
2336         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
2337         (ResourceLoadDelegate::identifierForInitialRequest):
2338         Always add id to the map.
2339         (ResourceLoadDelegate::removeIdentifierForRequest):
2340         Added.
2341         (ResourceLoadDelegate::willSendRequest):
2342         (ResourceLoadDelegate::didReceiveAuthenticationChallenge):
2343         (ResourceLoadDelegate::didReceiveResponse):
2344         (ResourceLoadDelegate::didFinishLoadingFromDataSource):
2345         (ResourceLoadDelegate::didFailLoadingWithError):
2346         (ResourceLoadDelegate::descriptionSuitableForTestResult):
2347         Replace static descriptionSuitableForTestResult with static member function to access identifier map.
2348         * DumpRenderTree/win/ResourceLoadDelegate.h:
2349         (ResourceLoadDelegate::urlMap):
2350         Moved within class so that each WebView has its own id map.
2351
2352 2010-11-01  Ojan Vafai  <ojan@chromium.org>
2353
2354         Reviewed by Dimitri Glazkov.
2355
2356         remove debug code from run_webkit_tests.py
2357         https://bugs.webkit.org/show_bug.cgi?id=48800
2358
2359         Remove temporary debug code and make --master-name required
2360         if --test-results-server is set now that all clients set
2361         --master-name.
2362
2363         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2364
2365 2010-11-01  Tony Chang  <tony@chromium.org>
2366
2367         Reviewed by Kent Tamura.
2368
2369         add plugin event logging for linux
2370         https://bugs.webkit.org/show_bug.cgi?id=48779
2371
2372         This is taken from Chromium's fork of the layout test plugin:
2373         http://git.chromium.org/gitweb/?p=chromium.git/.git;a=blob;f=webkit/tools/npapi_layout_test_plugin/main.cpp;h=3ebdada2f049b3624756438cff852364f86a2ede;hb=HEAD#l348
2374
2375         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
2376         (handleEventX11):
2377         (NPP_HandleEvent):
2378
2379 2010-11-01  Dirk Pranke  <dpranke@chromium.org>
2380
2381         Reviewed by James Robinson.
2382
2383         new-run-webkit-tests: use DRT, child-processes=1 for GPU tests by default
2384         https://bugs.webkit.org/show_bug.cgi?id=48790
2385
2386         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2387         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2388
2389 2010-11-01  Adam Barth  <abarth@webkit.org>
2390
2391         Reviewed by Eric Seidel.
2392
2393         Teach check-webkit-style how to accept a list of files to diff on the
2394         command line
2395         https://bugs.webkit.org/show_bug.cgi?id=48784
2396
2397         In a future patch, webkit-patch will use this option to improve
2398         performance.  I'm landing this in two pieces to avoid causing a version
2399         skew problem for the style-bot.
2400
2401         * Scripts/check-webkit-style:
2402         * Scripts/webkitpy/style/optparser.py:
2403         * Scripts/webkitpy/style/optparser_unittest.py:
2404         * Scripts/webkitpy/style_references.py:
2405
2406 2010-11-01  Anders Carlsson  <andersca@apple.com>
2407
2408         Reviewed by John Sullivan.
2409
2410         Tear down the related WebProcessProxy when a WebContext is deallocated
2411         https://bugs.webkit.org/show_bug.cgi?id=48769
2412
2413         * TestWebKitAPI/Tests/WebKit2/FailedLoad.cpp:
2414         (TestWebKitAPI::didFailProvisionalLoadWithErrorForFrame):
2415         We don't support empty URLs anymore, update test to expect a null URL instead.
2416
2417 2010-11-01  Søren Gjesse  <sgjesse@chromium.org>
2418
2419         Reviewed by Andreas Kling.
2420
2421         Fix warning when compiling the chromium port of DumpRenderShell
2422         with clang.
2423         https://bugs.webkit.org/show_bug.cgi?id=48414
2424
2425         * DumpRenderTree/chromium/TestShell.h:
2426         (TestShell::javaScriptFlagsForLoad):
2427
2428 2010-11-01  Adam Roben  <aroben@apple.com>
2429
2430         Fix typo from r71022
2431
2432         * BuildSlaveSupport/build.webkit.org-config/config.json:
2433
2434 2010-11-01  Adam Roben  <aroben@apple.com>
2435
2436         Trigger the Windows Release WebKit2 tests when a Release build
2437         finishes, not when a Debug build finishes
2438
2439         Fixes <http://webkit.org/b/48754> Windows Release WebKit2 tests are
2440         triggered at the wrong time
2441
2442         Reviewed by Sam Weinig.
2443
2444         * BuildSlaveSupport/build.webkit.org-config/config.json: Fixed
2445         triggerable name and trigger.
2446
2447 2010-11-01  Mario Sanchez Prada  <msanchez@igalia.com>
2448
2449         Unreviewed. Adding my IRC nickname to the list of committers.
2450
2451         * Scripts/webkitpy/common/config/committers.py:
2452
2453 2010-10-31  Robert Hogan  <robert@webkit.org>
2454
2455         Reviewed by Antonio Gomes.
2456
2457         [Qt] [Gtk] Plug-ins having upper case in mime type are failing to load
2458
2459         Qt and Gtk are case-sensitive when storing the declared mime-type
2460         of plugins. Since plugin mime-types are lowercased prior to searching
2461         for them in the plugin database, ensure they are loaded with the
2462         mime-type in lower case too.
2463
2464         Change the test netscape plugin to declare its mimetype in sentence
2465         case so that the correct behaviour is enforced.
2466
2467         https://bugs.webkit.org/show_bug.cgi?id=36815
2468
2469         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
2470         (NP_GetMIMEDescription):
2471
2472 2010-10-31  Robert Hogan  <robert@webkit.org>
2473
2474         Reviewed by Antonio Gomes.
2475
2476         [Qt] Support nodesFromRect in DRT
2477
2478         https://bugs.webkit.org/show_bug.cgi?id=48716
2479
2480         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2481         (LayoutTestController::nodesFromRect):
2482         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2483
2484 2010-10-30  Andreas Kling  <kling@webkit.org>
2485
2486         Reviewed by Antonio Gomes.
2487
2488         [Qt] QtTestBrowser: Switching view type moves the embedded inspector
2489         https://bugs.webkit.org/show_bug.cgi?id=48705
2490
2491         Reinsert the embedded inspector into the splitter after changing
2492         between QWebView/QGraphicsWebView.
2493
2494         * QtTestBrowser/launcherwindow.cpp:
2495         (LauncherWindow::init):
2496         (LauncherWindow::initializeView):
2497         * QtTestBrowser/webinspector.h:
2498         (WebInspector::WebInspector):
2499
2500 2010-10-28  Antonio Gomes  <agomes@rim.com>
2501
2502         Reviewed by Ojan Vafai.
2503
2504         Needs a "LinuxEditingBehavior", perhaps with a better name
2505         https://bugs.webkit.org/show_bug.cgi?id=36627
2506
2507         Adding support to Mac's, GTK+'s, Windows' and Chromium's LayoutTestController class to test the newly introduced Unix editing behavior.
2508
2509         * DumpRenderTree/chromium/LayoutTestController.cpp:
2510         (LayoutTestController::setEditingBehavior):
2511         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2512         (LayoutTestController::setEditingBehavior):
2513         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2514         (LayoutTestController::setEditingBehavior):
2515         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2516         (LayoutTestController::setEditingBehavior):
2517
2518 2010-10-29  Andreas Kling  <kling@webkit.org>
2519
2520         Reviewed by Kenneth Rohde Christiansen.
2521
2522         [Qt] QtTestBrowser: Add keyboard shortcut to toggle full screen (F11)
2523         https://bugs.webkit.org/show_bug.cgi?id=48695
2524
2525         * QtTestBrowser/launcherwindow.cpp:
2526         (LauncherWindow::createChrome):
2527
2528 2010-10-29  Andreas Kling  <kling@webkit.org>
2529
2530         Reviewed by Kenneth Rohde Christiansen.
2531
2532         [Qt] QtTestBrowser: Fix uninitialized read in FpsTimer
2533         https://bugs.webkit.org/show_bug.cgi?id=48675
2534
2535         FpsTimer::m_timer was never initialized and passed to QObject::killTimer()
2536         on startup with in -graphicsbased mode.
2537
2538         * QtTestBrowser/fpstimer.cpp:
2539         (FpsTimer::FpsTimer):
2540         (FpsTimer::numFrames):
2541         (FpsTimer::stop):
2542         (FpsTimer::timerEvent):
2543         * QtTestBrowser/fpstimer.h:
2544
2545 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
2546
2547         Reviewed by Tony Chang.
2548
2549         new-run-webkit-tests: change TestResults to be serializable
2550
2551         In preparation for changing new-run-webkit-tests from
2552         multithreaded to multiprocess, we need to make sure the data
2553         going between the threads is easily serialized over a socket.
2554
2555         This change adds serialization/pickling for the TestResults and
2556         TestFailure objects (using cPickle).
2557
2558         The TestFailure objects included a "has_wdiff" flag for Text
2559         results, but the flag wasn't being used, so I've removed it,
2560         simplifying the state to basically a set of enum objects with
2561         associated methods.
2562
2563         https://bugs.webkit.org/show_bug.cgi?id=48616
2564
2565         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2566         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
2567         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
2568         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
2569         * Scripts/webkitpy/layout_tests/layout_package/test_results.py: Added.
2570         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:Added. 
2571         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2572         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2573
2574 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
2575
2576         Reviewed by Ojan Vafai.
2577
2578         new-run-webkit-tests: need to provide separate dashboard results for GPU tests
2579         https://bugs.webkit.org/show_bug.cgi?id=48687
2580
2581         Modify the steps to upload the JSON files to the dashboards so
2582         that the GPU tests don't conflict with the regular tests on a
2583         port. We do this by modifying the --builder-name parameter to
2584         append " - GPU", which should cause the app to treat the results
2585         as a completely new builder. This is a little non-obvious, but
2586         keeps us from having to restructure the app.
2587
2588         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2589         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2590         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2591
2592 2010-10-29  Ojan Vafai  <ojan@chromium.org>
2593
2594         Reviewed by Tony Chang.
2595
2596         [chromium] add debug logging to help diagnose flakiness dashboard issues
2597         https://bugs.webkit.org/show_bug.cgi?id=48657
2598
2599         The appengine app thinks it's getting empty files uploaded. Add some logging
2600         to see if new-run-webkit-tests agrees.
2601         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2602
2603 2010-10-29  Dimitri Glazkov  <dglazkov@chromium.org>
2604
2605         Reviewed by Ojan Vafai.
2606
2607         [Chromium/DRT] Add master-name flag to new-run-webkit-tests.
2608         https://bugs.webkit.org/show_bug.cgi?id=48649
2609
2610         * BuildSlaveSupport/build.webkit.org-config/master.cfg: Added flag.
2611
2612 2010-10-29  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
2613
2614         Reviewed by Martin Robinson.
2615
2616         Enable popup window in GtkLauncher
2617         https://bugs.webkit.org/show_bug.cgi?id=48335
2618
2619         The GtkLauncher application does not support opening new window when
2620         a link with "target=_blank" is clicked or similar call to
2621         window.open(). Instead, GtkLauncher does nothing which breaks
2622         navigation of some websites.
2623
2624         * GtkLauncher/main.c:
2625         (activate_uri_entry_cb):
2626         (update_title):
2627         (link_hover_cb):
2628         (notify_title_cb):
2629         (notify_load_status_cb):
2630         (notify_progress_cb):
2631         (destroy_cb):
2632         (go_back_cb):
2633         (go_forward_cb):
2634         (create_web_view_cb):
2635         (web_view_ready_cb):
2636         (close_web_view_cb):
2637         (create_browser):
2638         (create_statusbar):
2639         (create_toolbar):
2640         (create_window):
2641         (main):
2642
2643 2010-10-29  Adam Roben  <aroben@apple.com>
2644
2645         Teach check-webkit-style about WebKit2's idiosyncracies
2646
2647         Fixes <http://webkit.org/b/48638> Style bot complains about a number
2648         of WebKit2 conventions
2649
2650         Reviewed by Anders Carlsson.
2651
2652         * Scripts/webkitpy/style/checker.py: Excluded some rules for various
2653         WebKit2-related files. Also updated the excluded rules for
2654         WebKitAPITest to match the current code.
2655
2656 2010-10-29  Csaba Osztrogonác  <ossy@webkit.org>
2657
2658         Reviewed by Adam Roben and David Kilzer.
2659
2660         Fix and cleanup of build systems
2661         https://bugs.webkit.org/show_bug.cgi?id=48342
2662
2663         * Scripts/build-webkit:
2664          - Remove unnecessary ENABLE_SANDBOX option.
2665          - Add ENABLE_FULLSCREEN_API option.
2666
2667 2010-10-28  Adam Roben  <aroben@apple.com>
2668
2669         Switch the Windows WebKit2 bot to the Release configuration
2670
2671         We only have one machine testing WebKit2 on Windows right now, and
2672         Debug is just too slow for it to keep up.
2673
2674         Fixes (hopefully!) <http://webkit.org/b/48615> Windows WebKit2 bot is
2675         always way behind
2676
2677         Reviewed by Jon Honeycutt.
2678
2679         * BuildSlaveSupport/build.webkit.org-config/config.json:
2680
2681 2010-10-29  Leandro Gracia Gil  <leandrogracia@google.com>
2682
2683         Reviewed by Jeremy Orlow.
2684
2685         Added a second parameter to setMockSpeechInputResult for
2686         the language used in speech input.
2687         https://bugs.webkit.org/show_bug.cgi?id=47089
2688
2689         * DumpRenderTree/LayoutTestController.cpp:
2690         (setMockSpeechInputResultCallback):
2691         * DumpRenderTree/LayoutTestController.h:
2692         * DumpRenderTree/chromium/LayoutTestController.cpp:
2693         (LayoutTestController::setMockSpeechInputResult):
2694         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2695         (LayoutTestController::setMockSpeechInputResult):
2696         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2697         (LayoutTestController::setMockSpeechInputResult):
2698         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2699         (LayoutTestController::setMockSpeechInputResult):
2700         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2701         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2702         (LayoutTestController::setMockSpeechInputResult):
2703         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2704         (LayoutTestController::setMockSpeechInputResult):
2705         (LayoutTestController::pageSizeAndMarginsInPixels):
2706
2707 2010-10-28  Ojan Vafai  <ojan@chromium.org>
2708
2709         Reviewed by Tony Chang.
2710
2711         [chromium] add a result-small.json file for the test dashboard
2712         https://bugs.webkit.org/show_bug.cgi?id=48547
2713
2714         Output both a results.json file and a results-small.json file.
2715         The dashboard will load results-small.json by default so it loads faster.
2716
2717         * TestResultServer/model/jsonresults.py:
2718         * TestResultServer/model/jsonresults_unittest.py:
2719         Added a bunch of sys.path hackery. Unforunately, this uses hardcoded
2720         paths. That obviously needs to be fixed, but at least this way it
2721         clearly documents what paths are necessary.
2722
2723 2010-10-28  Eric Seidel  <eric@webkit.org>
2724
2725         Reviewed by Adam Barth.
2726
2727         webkit-patch upload calls changed_files more often than it should
2728         https://bugs.webkit.org/show_bug.cgi?id=48567
2729
2730         Passing changed_files around everywhere isn't a very elegant solution
2731         but it's the one we have for the moment.  I think keeping an explicit
2732         cache on Checkout (or making StepState() a real class) is a better
2733         long-term option.
2734
2735         Previously bug_id_for_this_commit was calling changed_files and the
2736         result was never getting cached on the state.  Now we're explicitly
2737         caching the result on the state and passing that to the bug_id_for_this_commit call.
2738
2739         I looked into building unit tests for this.  Doing so would require
2740         using a real Checkout object with a MockSCM and overriding the appropriate
2741         calls on SCM to count how often we're stating the file system.
2742         That's a useful set of tests to build for a separate change.
2743
2744         * Scripts/webkitpy/common/checkout/api.py:
2745         * Scripts/webkitpy/tool/commands/download.py:
2746         * Scripts/webkitpy/tool/commands/upload.py:
2747         * Scripts/webkitpy/tool/mocktool.py:
2748
2749 2010-10-28  Eric Seidel  <eric@webkit.org>
2750
2751         Reviewed by Adam Barth.
2752
2753         Make suggest-reviewers slightly faster
2754         https://bugs.webkit.org/show_bug.cgi?id=48562
2755
2756         Add @memoized to one more common call.
2757
2758         * Scripts/webkitpy/common/checkout/api.py:
2759
2760 2010-10-28  Eric Seidel  <eric@webkit.org>
2761
2762         Reviewed by Adam Barth.
2763
2764         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is still very slow
2765         https://bugs.webkit.org/show_bug.cgi?id=48536
2766
2767         This patch makes it about 40 seconds faster, but it still
2768         takes 1:40.  This will require more refinement.
2769
2770         The suggested reviewers list appears to be the same.
2771
2772         I think the next step may be to have it stop the search after
2773         5 reviewers are found.  We never want to suggest 30 people.
2774
2775         * Scripts/webkitpy/common/checkout/scm.py:
2776          - Using --remove-empty to theoretically stop lookups past
2777            when a file is removed.  I'm not entirely clear that the option
2778            does what it says it does.  Example:
2779            git log --pretty=format:%H -5 --remove-empty -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
2780            returns only one commit
2781            vs.
2782            git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
2783            which returns 5.  I was not aware that wx files were ever removed from the repository?
2784
2785 2010-10-28  Kinuko Yasuda  <kinuko@chromium.org>
2786
2787         Reviewed by David Levin.
2788
2789         [Chromium] Support FileSystem in chromium DRT
2790         https://bugs.webkit.org/show_bug.cgi?id=47643
2791
2792         * DumpRenderTree/chromium/WebViewHost.cpp:
2793         (WebViewHost::openFileSystem): Added.
2794         * DumpRenderTree/chromium/WebViewHost.h:
2795         (WebViewHost::openFileSystem): Added.
2796
2797 2010-10-28  Sam Weinig  <sam@webkit.org>
2798
2799         Reviewed by Anders Carlsson.
2800
2801         WKURLRefs should be allowed to be null
2802         <rdar://problem/8575621>
2803         https://bugs.webkit.org/show_bug.cgi?id=48535
2804
2805         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
2806         * TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp:
2807         (TestWebKitAPI::didStartProvisionalLoadForFrame):
2808         (TestWebKitAPI::didCommitLoadForFrame):
2809         (TestWebKitAPI::didFinishLoadForFrame):
2810         Test that URLs are null pointers when unset.
2811
2812 2010-10-28  Chris Fleizach  <cfleizach@apple.com>
2813
2814         Reviewed by Adele Peterson.
2815
2816         AX: multi select group option does not handle setting of AXSelectedChildren correctly
2817         https://bugs.webkit.org/show_bug.cgi?id=48464
2818
2819         Add support for querying information about selected children to DRT, including:
2820            selectedChildAtIndex
2821            selectedChildrenCount
2822            setSelectedChild
2823
2824         * DumpRenderTree/AccessibilityUIElement.cpp:
2825         (selectedChildAtIndexCallback):
2826         (setSelectedChildCallback):
2827         (selectedChildrenCountCallback):
2828         (AccessibilityUIElement::setSelectedChild):
2829         (AccessibilityUIElement::selectedChildrenCount):
2830         (AccessibilityUIElement::selectedChildAtIndex):
2831         (AccessibilityUIElement::getJSClass):
2832         * DumpRenderTree/AccessibilityUIElement.h:
2833         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
2834         (AccessibilityUIElement::selectedChildAtIndex):
2835         (AccessibilityUIElement::selectedChildrenCount):
2836         (AccessibilityUIElement::setSelectedChild):
2837
2838 2010-10-27  Sam Weinig  <sam@webkit.org>
2839
2840         Reviewed by Anders Carlsson.
2841
2842         Add WebKit2 API for window feature getter/setters
2843         <rdar://problem/8590373>
2844         https://bugs.webkit.org/show_bug.cgi?id=48496
2845
2846         * MiniBrowser/mac/BrowserWindowController.m:
2847         (-[BrowserWindowController awakeFromNib]):
2848         * MiniBrowser/win/BrowserView.cpp:
2849         (BrowserView::create):
2850         * WebKitTestRunner/TestController.cpp:
2851         (WTR::createOtherPage):
2852         (WTR::TestController::initialize):
2853
2854 2010-10-28  Søren Gjesse  <sgjesse@chromium.org>
2855
2856         Reviewed by Tony Chang.
2857
2858         Added support for the DumpRenderTree flags --multiple-loads and --js-flags to the Python test runner.
2859         https://bugs.webkit.org/show_bug.cgi?id=48236
2860
2861         * Scripts/webkitpy/layout_tests/port/chromium.py:
2862         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2863
2864 2010-10-28  Sergio Villar Senin  <svillar@igalia.com>
2865
2866         Reviewed by Martin Robinson.
2867
2868         [GTK] WebKitWebFrame's load-status is not properly notified to the tests
2869         https://bugs.webkit.org/show_bug.cgi?id=48048
2870
2871         DRT now listens to WebKitWebFrame load-status signals for each
2872         created frame instead of the load-status signal of the
2873         WebKitWebView that only notifies about changes in the main frame.
2874
2875         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2876         (webViewLoadFinished):
2877         (webFrameLoadStatusNotified):
2878         (frameCreatedCallback):
2879         (createWebView):
2880         (main):
2881
2882 2010-10-28  Tony Chang  <tony@chromium.org>
2883
2884         Reviewed by Kent Tamura.
2885
2886         [chromium] fix textInputController.{selectedRange,markedRange}
2887         https://bugs.webkit.org/show_bug.cgi?id=48487
2888
2889         * DumpRenderTree/chromium/TextInputController.cpp:
2890         (TextInputController::markedRange): Return arrays of ints, rather than a string
2891         (TextInputController::selectedRange): Ditto.
2892
2893 2010-10-27  Ojan Vafai  <ojan@chromium.org>
2894
2895         Reviewed by Tony Chang.
2896
2897         [chromium] add a master-name flag to new-run-webkit-tests
2898         https://bugs.webkit.org/show_bug.cgi?id=48488
2899
2900         The test results server now allows adding a master name to
2901         the uploaded files. This lets us distinguish bots that have
2902         the same name, but are on different masters.
2903
2904         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2905
2906 2010-10-28  Adam Roben  <aroben@apple.com>
2907
2908         Don't append a newline to the test output if the frame has no document
2909         element in WebKitTestRunner
2910
2911         Fixes <http://webkit.org/b/48526> Extra trailing newline when running
2912         plugins/document-open.html in WebKitTestRunner
2913
2914         Reviewed by Anders Carlsson.
2915
2916         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2917         (WTR::toJS): Added. Turns a UTF-8 C string into a JSStringRef.
2918         (WTR::hasDocumentElement): Added. Uses the JSC API to figure out
2919         whether the frame has a document element.
2920         (WTR::dumpFrameText): Match DRT by bailing (rather than appending an
2921         empty string and a newline) if the frame has no document element.
2922
2923 2010-10-28  Adam Roben  <aroben@apple.com>
2924
2925         Skip npn-invalidate-rect-invalidates-window.html on headless XP
2926         machines
2927
2928         TestNetscapePlugIn never receives a WM_PAINT message on headless XP
2929         machines, so this test times out. Fixing the test is covered by
2930         <http://webkit.org/b/48333>.
2931
2932         * Scripts/old-run-webkit-tests: Skip
2933         npn-invalidate-rect-invalidates-window.html on Windows if accelerated
2934         compositing support is disabled, which likely means we're on a headless
2935         XP machine.
2936
2937 2010-10-28  Kimmo Kinnunen  <kimmok@iki.fi>
2938
2939         Adding myself as a committer.
2940
2941         * Scripts/webkitpy/common/config/committers.py:
2942
2943 2010-10-27  Eric Seidel  <eric@webkit.org>
2944
2945         Reviewed by Adam Barth.
2946
2947         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is very slow
2948         https://bugs.webkit.org/show_bug.cgi?id=48500
2949
2950         This doesn't fix the problem, but it makes things slightly better.
2951         Each git svn find-rev call takes about .25 seconds on my desktop.
2952         This patch uses a new memoized class to avoid those calls when possible.
2953
2954         The real slowness is still git log on some files, like:
2955         git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
2956         I'm not yet sure how to make the pathological git logs better.
2957
2958         * Scripts/webkitpy/common/checkout/scm.py:
2959         * Scripts/webkitpy/common/memoized.py: Added.
2960         * Scripts/webkitpy/common/memoized_unittest.py: Added.
2961
2962 2010-10-27  Eric Seidel  <eric@webkit.org>
2963
2964         Reviewed by Ojan Vafai.
2965
2966         EWS bots should not use --quiet when running build-webkit
2967         https://bugs.webkit.org/show_bug.cgi?id=48482
2968
2969         --quiet is only correct when the sub-process does the error reporting.
2970         In the case of _can_build() the parent process is reporting the error.
2971         We'd like the full build log at queues.webkit.org so someone can look
2972         at the log and understand why the EWS is failing to build trunk.
2973
2974         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2975         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
2976         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
2977         * Scripts/webkitpy/tool/mocktool.py:
2978
2979 2010-10-27  Brian Weinstein  <bweinstein@apple.com>
2980
2981         More Windows build fixage. Rename a variable that was named string.
2982
2983         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2984         (WTR::hasPrefix):
2985
2986 2010-10-27  Ojan Vafai  <ojan@chromium.org>
2987
2988         Reviewed by Dimitri Glazkov.
2989
2990         [chromium] Make the test results server store which master the bot is on
2991         https://bugs.webkit.org/show_bug.cgi?id=48478
2992
2993         The chromium bots recently changed so that there are multiple slaves with
2994         the same name on different masters. Up till now, the test results server
2995         assumed slave names were unique. Adds a master field to the file in order
2996         to distinguish.
2997
2998         Also, for files that currently lack a master or testtype, set them appropriately.
2999
3000         * TestResultServer/handlers/testfilehandler.py:
3001         * TestResultServer/index.yaml:
3002         * TestResultServer/model/jsonresults.py:
3003         * TestResultServer/model/testfile.py:
3004         * TestResultServer/templates/showfilelist.html:
3005         * TestResultServer/templates/uploadform.html:
3006
3007 2010-10-26  Darin Adler  <darin@apple.com>
3008
3009         Reviewed by Sam Weinig.
3010
3011         WebKitTestRunner needs to support layoutTestController.dumpBackForwardList
3012         https://bugs.webkit.org/show_bug.cgi?id=42322
3013         rdar://problem/8193631
3014
3015         WebKitTestRunner needs to support layoutTestController.clearBackForwardList
3016         https://bugs.webkit.org/show_bug.cgi?id=42333
3017         rdar://problem/8193643
3018
3019         * WebKitTestRunner/Configurations/InjectedBundle.xcconfig:
3020         Renamed the product to WebKitTestRunnerInjectedBundle to avoid
3021         name conflicts in the build directory.
3022
3023         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
3024         Added dumpBackForwardList and clearBackForwardList.
3025
3026         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
3027         (WTR::InjectedBundle::InjectedBundle): Removed initialization of m_mainPage.
3028         (WTR::InjectedBundle::didCreatePage): Removed unneeded underscore.
3029         (WTR::InjectedBundle::willDestroyPage): Ditto.
3030         (WTR::InjectedBundle::didReceiveMessage): Ditto.
3031         (WTR::InjectedBundle::initialize): Ditto.
3032         (WTR::InjectedBundle::didCreatePage): Changed code to use m_pages
3033         instead of m_mainPage and m_otherPages.
3034         (WTR::InjectedBundle::willDestroyPage): Ditto.
3035         (WTR::InjectedBundle::page): Ditto.
3036         (WTR::InjectedBundle::beginTesting): Ditto.
3037         (WTR::InjectedBundle::done): Ditto.
3038         (WTR::InjectedBundle::closeOtherPages): Ditto.
3039         (WTR::InjectedBundle::dumpBackForwardListsForAllPages): Added.
3040
3041         * WebKitTestRunner/InjectedBundle/InjectedBundle.h: Changed
3042         page function to longer be inline, and pageCount function
3043         to use m_pages. Added dumpBackForwardListsForAllPages, and
3044         removed some underscores. Replaced m_mainPage and m_otherPageas
3045         with m_pages.
3046
3047         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3048         (WTR::adoptWK): Added.
3049         (WTR::hasPrefix): Added.
3050         (WTR::InjectedBundlePage::reset): Added code to set up
3051         m_previousTestBackForwardListItem.
3052         (WTR::InjectedBundlePage::dump): Added code to call
3053         dumpBackForwardListsForAllPages.
3054         (WTR::compareByTargetName): Added.
3055         (WTR::dumpBackForwardListItem): Added.
3056         (WTR::InjectedBundlePage::dumpBackForwardList): Added.
3057
3058         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h: Added
3059         dumpBackForwardList and m_previousTestBackForwardListItem.
3060
3061         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
3062         (WTR::LayoutTestController::LayoutTestController): Initialize
3063         m_shouldDumpBackForwardListsForAllWindows to false.
3064         (WTR::LayoutTestController::clearBackForwardList): Added.
3065
3066         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
3067         Added dumpBackForwardList, clearBackForwardList,
3068         shouldDumpBackForwardListsForAllWindows, and
3069         m_shouldDumpBackForwardListsForAllWindows.
3070
3071         * WebKitTestRunner/TestController.cpp:
3072         (WTR::TestController::initialize): Set up
3073         didReceiveSynchronousMessageFromInjectedBundle.
3074         (WTR::TestController::didReceiveSynchronousMessageFromInjectedBundle):
3075         Added.
3076         * WebKitTestRunner/TestController.h: Ditto.
3077
3078         * WebKitTestRunner/TestInvocation.cpp:
3079         (WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):
3080         Added.
3081         * WebKitTestRunner/TestInvocation.h: Ditto.
3082
3083         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
3084         Renamed bundle to WebKitTestRunnerInjectedBundle (see above).
3085         * WebKitTestRunner/mac/TestControllerMac.mm:
3086         (WTR::TestController::initializeInjectedBundlePath): Ditto.
3087
3088 2010-10-27  Chris Rogers  <crogers@google.com>
3089
3090         Reviewed by Chris Marrin.
3091
3092         Add ENABLE_WEB_AUDIO feature enable flag (initially disabled) to build-webkit
3093         https://bugs.webkit.org/show_bug.cgi?id=48279
3094
3095         * Scripts/build-webkit:
3096
3097 2010-10-27  Eric Seidel  <eric@webkit.org>
3098
3099         Unreviewed.
3100
3101         build-webkit should collect Visual Studio Express logs and display them
3102         https://bugs.webkit.org/show_bug.cgi?id=39199
3103
3104         It turns out my previous patch wasn't actually working on the win-ews
3105         machine.  So I've fixed my mistakes from before.
3106
3107         * Scripts/build-webkit:
3108          - Windows VSE builds change the CWD while building.  Why?  Who knows.
3109         * Scripts/print-vse-failure-logs:
3110          - windows VSE builds don't use Debug/Release as I expected, so I've
3111            moved off of --configuration to --top-level and added the /obj
3112            optimization while I was there.
3113
3114 2010-10-27  Dimitri Glazkov  <dglazkov@chromium.org>
3115
3116         Unreviewed, rolling out r70674.
3117         http://trac.webkit.org/changeset/70674
3118         https://bugs.webkit.org/show_bug.cgi?id=48053
3119
3120         Broke Chromium Windows build.
3121
3122         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3123         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3124
3125 2010-10-25  Tony Chang  <tony@chromium.org>
3126
3127         Reviewed by Anders Carlsson.
3128
3129         compile TestNetscapePlugIn on chromium linux
3130         https://bugs.webkit.org/show_bug.cgi?id=48274
3131
3132         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
3133         (NP_Initialize): On Linux, plugin funcs are set in initialize.
3134         (NPP_New): Mark the plugin as windowless.
3135         (NPP_GetValue): Handle mime type values.
3136         (NP_GetMIMEDescription):
3137         (NP_GetValue):
3138         * DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npapi.h: define TRUE and FALSE, which are in webkit's npapi.h.
3139         * DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npfunctions.h: Pull in npapi.h to get TRUE/FALSE (matches webkit's npfunctions.h)
3140
3141 2010-10-27  Kenneth Rohde Christiansen  <kenneth@webkit.org>
3142
3143         Reviewed by Anders Carlsson.
3144
3145         Remove contentSizeChanged callbacks as it is no longer
3146         part of the public UIClient.
3147
3148         Make WKPageContentsSizeChangedCallback be a private API
3149         https://bugs.webkit.org/show_bug.cgi?id=48409
3150
3151         * MiniBrowser/mac/BrowserWindowController.m:
3152         (-[BrowserWindowController awakeFromNib]):
3153         * MiniBrowser/win/BrowserView.cpp:
3154         (BrowserView::create):
3155         * WebKitTestRunner/TestController.cpp:
3156         (WTR::createOtherPage):
3157         (WTR::TestController::initialize):
3158
3159 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3160
3161         Reviewed by Ojan Vafai.
3162
3163         Make http locking default in NRWT.
3164         https://bugs.webkit.org/show_bug.cgi?id=48053
3165
3166         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3167         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3168
3169 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3170
3171         Reviewed by Tony Chang.
3172
3173         [NRWT] Fix http lock on Windows platform
3174         https://bugs.webkit.org/show_bug.cgi?id=48321
3175
3176         * Scripts/webkitpy/layout_tests/port/http_lock.py:
3177
3178 2010-10-27  Satish Sampath  <satish@chromium.org>
3179
3180         Unreviewed, rolling out r70665.
3181         http://trac.webkit.org/changeset/70665
3182         https://bugs.webkit.org/show_bug.cgi?id=47089
3183
3184         Need to address Alexey's review comments.
3185
3186         * DumpRenderTree/LayoutTestController.cpp:
3187         (setMockSpeechInputResultCallback):
3188         * DumpRenderTree/LayoutTestController.h:
3189         * DumpRenderTree/chromium/LayoutTestController.cpp:
3190         (LayoutTestController::setMockSpeechInputResult):
3191         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3192         (LayoutTestController::setMockSpeechInputResult):
3193         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3194         (LayoutTestController::setMockSpeechInputResult):
3195         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3196         (LayoutTestController::setMockSpeechInputResult):
3197         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3198         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3199         (LayoutTestController::setMockSpeechInputResult):
3200         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3201         (LayoutTestController::setMockSpeechInputResult):
3202
3203 2010-10-27  Leandro Gracia Gil  <leandrogracia@google.com>
3204
3205         Reviewed by Jeremy Orlow.
3206
3207         Added a second parameter to setMockSpeechInputResult for
3208         the language used in speech input.
3209         https://bugs.webkit.org/show_bug.cgi?id=47089
3210
3211         * DumpRenderTree/LayoutTestController.cpp:
3212         (setMockSpeechInputResultCallback):
3213         * DumpRenderTree/LayoutTestController.h:
3214         * DumpRenderTree/chromium/LayoutTestController.cpp:
3215         (LayoutTestController::setMockSpeechInputResult):
3216         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3217         (LayoutTestController::setMockSpeechInputResult):
3218         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3219         (LayoutTestController::setMockSpeechInputResult):
3220         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3221         (LayoutTestController::setMockSpeechInputResult):
3222         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3223         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3224         (LayoutTestController::setMockSpeechInputResult):
3225         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3226         (LayoutTestController::setMockSpeechInputResult):
3227         (LayoutTestController::pageSizeAndMarginsInPixels):
3228
3229 2010-10-27  Adam Roben  <aroben@apple.com>
3230
3231         Reset TestNetscapePlugIn's NPP_GetValue pointer when
3232         NullNPPGetValuePointer finishes running
3233
3234         Fixes <http://webkit.org/b/48435> REGRESSION (r70655): Many plugins
3235         tests are failing on Qt
3236
3237         Reviewed by Anders Carlsson.
3238
3239         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
3240         (PluginTest::NPP_Destroy):
3241         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
3242         Added a do-nothing NPP_Destroy implementation.
3243
3244         * DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp:
3245         (NullNPPGetValuePointer::NullNPPGetValuePointer): Save the original
3246         NPP_GetValue pointer in m_originalNPPGetValuePointer so we can restore
3247         it later.
3248         (NullNPPGetValuePointer::NPP_Destroy): Added. Restores the original
3249         NPP_GetValue pointer so it can be used in other tests.
3250
3251         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
3252         (NPP_Destroy):
3253         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
3254         (webkit_test_plugin_destroy_instance):
3255         Call through to the PluginTest when NPP_Destroy is called.
3256
3257 2010-10-26  Brian Weinstein  <bweinstein@apple.com>
3258
3259         Reviewed by Adam Roben.
3260
3261         WebKit2 shouldn't try to send an empty user agent
3262         https://bugs.webkit.org/show_bug.cgi?id=48397
3263         
3264         Add a test that when we set our custom user agent to an empty string, we don't send
3265         an empty user agent.
3266
3267         * TestWebKitAPI/Tests/WebKit2/PreventEmptyUserAgent.cpp: Added.
3268         (TestWebKitAPI::didRunJavaScript): Make sure that the result of navigator.userAgent isn't empty.
3269         (TestWebKitAPI::TEST): Set our custom user agent to the empty string, and run navigator.userAgent.
3270         
3271         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Add the new file.
3272         * TestWebKitAPI/win/TestWebKitAPI.vcproj: Ditto.
3273
3274 2010-10-27  Adam Roben  <aroben@apple.com>
3275
3276         Check in file I forgot in r70653
3277
3278         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
3279         (NP_Initialize): Save the NPPluginFuncs struct the browser passed to us
3280         so we can be naughty and modify it later.
3281
3282 2010-10-27  Adam Roben  <aroben@apple.com>
3283
3284         Test that WebKit doesn't crash if the plugin passes 0 for its
3285         NPP_GetValue pointer
3286
3287         Test for <http://webkit.org/b/48433> Crash in
3288         NetscapePlugin::shouldLoadSrcURL when using Shockwave Director 10.3 in
3289         WebKit2 on Windows
3290
3291         Reviewed by Eric Carlson.
3292
3293         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3294         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
3295         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
3296         * GNUmakefile.am:
3297         Added NullNPPGetValuePointer.cpp.
3298
3299         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
3300         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
3301         Set up a place to store the NPPluginFuncs struct the browser passed to us.
3302
3303         * DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp: Added.
3304         (NullNPPGetValuePointer::NullNPPGetValuePointer): Null out the
3305         NPP_GetValue pointer we passed to the browser to simulate a plugin that
3306         doesn't implement NPP_GetValue.
3307         (NullNPPGetValuePointer::NPP_GetValue): Print an error message. If this
3308         function is called, it means that WebKit has changed in a way that
3309         makes this test invalid.
3310
3311         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
3312         (NP_GetEntryPoints): Save the NPPluginFuncs struct the browser passed
3313         to us so we can be naughty and modify it later.
3314
3315 2010-10-27  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
3316
3317         Reviewed by Andreas Kling.
3318
3319         Remove references to ancient QGVLauncher and QtLauncher
3320         https://bugs.webkit.org/show_bug.cgi?id=48430
3321
3322         QtTestBrowser substitutes both and has checks enabled.
3323
3324         * Scripts/webkitpy/style/checker.py: remove references
3325         * Scripts/webkitpy/style/checker_unittest.py: ditto
3326
3327 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3328
3329         Reviewed by Ojan Vafai.
3330
3331         [NRWT] Don't use image hash when it's no need in single test mode.
3332         https://bugs.webkit.org/show_bug.cgi?id=48326
3333
3334         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3335
3336 2010-10-27  David Kilzer  <ddkilzer@apple.com>
3337
3338         Fix leak of CFMutableDictionaryRef in createXMLStringFromWebArchiveData()
3339
3340         Reviewed by Adam Roben.
3341
3342         Follow-up fix for: <https://bugs.webkit.org/show_bug.cgi?id=48278>
3343
3344         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
3345         (createXMLStringFromWebArchiveData): Use RetainPtr<> to fix a
3346         leak introduced in r70613.
3347
3348 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3349
3350         Reviewed by Csaba Osztrogonác.
3351
3352         [NRWT] Add platform specific baseline search paths for Qt port
3353         https://bugs.webkit.org/show_bug.cgi?id=48428
3354
3355         * Scripts/webkitpy/layout_tests/port/qt.py:
3356
3357 2010-10-27  Adam Roben  <aroben@apple.com>
3358
3359         Catch exceptions when checking if we're inside a git working directory
3360
3361         Fixes <http://webkit.org/b/48420> REGRESSION (r70562): test-webkitpy
3362         fails on systems without git installed
3363
3364         Reviewed by Anders Carlsson.
3365
3366         * Scripts/webkitpy/common/net/credentials.py:
3367         (Credentials._credentials_from_git): Put the call to
3368         Git.in_working_directory inside the try/except since it, too, attempts
3369         to execute git and thus will throw on systems that don't have git
3370         installed.
3371
3372 2010-10-27  Nikolas Zimmermann  <nzimmermann@rim.com>
3373
3374         Reviewed by David Kilzer.
3375
3376         Convert DumpRenderTree webarchive code to CoreFoundation
3377         https://bugs.webkit.org/show_bug.cgi?id=48278
3378
3379         CFPropertyListCreateWithData and CFPropertyListCreateData are only available in 10.6+.
3380         Replace CFPropertyListCreateWithData by a combination of CFReadStreamCreateWithBytesNoCopy and CFPropertyListCreateFromStream.
3381         Replace CFPropertyListCreateData by CFPropertyListCreateXMLData.
3382
3383         These changes are wrapped in BUILDING_ON_LEOPARD, as the methods are deprecated on 10.6+.
3384
3385         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
3386         (createXMLStringFromWebArchiveData):
3387
3388 2010-10-26  Antonio Gomes  <agomes@rim.com>
3389
3390         Reviewed by Martin Robinson.
3391
3392         Remove the absolute path used to include DumpRenderTreeSupportGtk.h from LayoutTestController.h
3393
3394         It turns out this is rather unneeded since WebKitTools/GNUMakefile.am has WebKit/gtk/ in its include
3395         path, and then we can just do #include "WebCoreSupport/DumpRenderTreeSupportGtk.h"
3396
3397         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3398
3399 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3400
3401         Rename WebArchiveDumpSupport.mm to WebArchiveDumpSupport.cpp
3402
3403         Reviewed by Adam Roben.
3404
3405         Part 4 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3406
3407         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3408         Updated for file moves and renames.
3409         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.mm.
3410         * DumpRenderTree/cf/WebArchiveDumpSupport.h: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.h.
3411
3412 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3413
3414         Convert WebArchiveDumpSupport.mm from NS objects to CF types
3415
3416         Reviewed by Adam Roben.
3417
3418         Part 3 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3419
3420         In order to share WebArchive code between the Mac and Windows
3421         ports, the code in WebArchiveDumpSupport.mm was converted from
3422         Cocoa to C++ using CoreFoundation (CF) types.
3423
3424         * DumpRenderTree/mac/DumpRenderTree.mm:
3425         (dump): Renamed serializeWebArchiveToXML() to
3426         createXMLStringFromWebArchiveData() and added HardAutorelease()
3427         to prevent leaks.
3428         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
3429         (createXMLStringFromWebArchiveData): Renamed from
3430         serializeWebArchiveToXML().  Changed to use CF types.
3431         (createCFURLResponseFromResponseData): Renamed from
3432         unarchiveNSURLResponseFromResponseData().  Changed to use CF
3433         types for its parameter and return type.
3434         * DumpRenderTree/mac/WebArchiveDumpSupport.mm: Replaced use of
3435         NS objects with CF types.  It will be renamed to *.cpp in a
3436         future commit.
3437         (convertMIMEType): Changed to use case-insensitive string
3438         comparisons.
3439         (convertWebResourceDataToString):
3440         (normalizeHTTPResponseHeaderFields):
3441         (normalizeWebResourceURL):
3442         (convertWebResourceResponseToDictionary):
3443         (compareResourceURLs):
3444         (createXMLStringFromWebArchiveData):
3445         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
3446         (createCFURLResponseFromResponseData): Renamed from
3447         unarchiveNSURLResponseFromResponseData().  Changed to use CF
3448         types for its parameter and return type.
3449
3450 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3451
3452         Extract use of NSKeyedUnarchiver from WebArchiveDumpSupport.mm
3453
3454         Reviewed by Adam Roben.
3455
3456         Part 2 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3457
3458         There is no equivalent to NSKeyedUnarchiver in CoreFoundation,
3459         so extract it into a platform-specific source file.
3460
3461         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
3462         (unarchiveNSURLResponseFromResponseData): Added declaration.
3463         * DumpRenderTree/mac/WebArchiveDumpSupport.mm:
3464         (convertWebResourceResponseToDictionary): Extracted code to
3465         unarchiveNSURLResponseFromResponseData() in
3466         WebArchiveDumpSupportMac.mm.  Updated logic to return early if
3467         nil is returned from unarchiveNSURLResponseFromResponseData().
3468         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
3469         (unarchiveNSURLResponseFromResponseData): Added.  Extracted
3470         code from convertWebResourceResponseToDictionary() in
3471         WebArchiveDumpSupport.mm.
3472
3473 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3474
3475         Extract call to -[WebHTMLRepresentation supportedNonImageMIMETypes] from WebArchiveDumpSupport.mm
3476
3477         Reviewed by Adam Roben.
3478
3479         Part 1 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3480
3481         The call to -[WebHTMLRepresentation supportedNonImageMIMETypes]
3482         is not cross-platform between Mac and Windows, so extract it
3483         into a platform-specific source file.
3484
3485         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Added
3486         WebArchiveDumpSupportMac.mm to the project.
3487         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
3488         (supportedNonImageMIMETypes): Added declaration.
3489         * DumpRenderTree/mac/WebArchiveDumpSupport.mm:
3490         (convertWebResourceDataToString): Replaced call to
3491         -[WebHTMLRepresentation supportedNonImageMIMETypes] with
3492         supportedNonImageMIMETypes().
3493         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm: Added.
3494         (supportedNonImageMIMETypes): Added.  Extracted from
3495         WebArchiveDumpSupport.mm.
3496
3497 2010-10-26  Antonio Gomes  <agomes@rim.com>
3498
3499         Reviewed by Martin Robinson.
3500
3501         [GTK] Implement DumpRenderTreeSupportGtk (similarly to DumpRenderTreeSupportQt idea)
3502         https://bugs.webkit.org/show_bug.cgi?id=48199
3503
3504         Implements support to WebKitTabToLinksPreferenceKey through LayoutTestController::overridePreference.
3505         The corresponding DumpRenderTreeSupportGtk method is called in the DRT context only.
3506
3507         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3508         (resetDefaultsToConsistentValues):
3509         (createWebView):
3510         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3511         (LayoutTestController::overridePreference):
3512
3513 2010-10-26  Ariya Hidayat  <ariya@sencha.com>
3514
3515         Unreviewed, change the order of my emails for bugzilla autocompletion.
3516
3517         * Scripts/webkitpy/common/config/committers.py:
3518
3519 2010-10-26  Eric Seidel  <eric@webkit.org>
3520
3521         Reviewed by David Kilzer.
3522
3523         build-webkit should collect Visual Studio Express logs and display them
3524         https://bugs.webkit.org/show_bug.cgi?id=39199
3525
3526         * Scripts/build-webkit:
3527         * Scripts/print-vse-failure-logs: Added.
3528         * Scripts/webkitdirs.pm:
3529
3530 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3531
3532         <http://webkit.org/b/48224> build-webkit: add support for --meter-tag switch
3533
3534         Reviewed by Csaba Osztrogonác.
3535
3536         * Scripts/build-webkit: Added support for --meter-tag switch.
3537         It should have been added with r60820.
3538
3539 2010-10-26  Ademar de Souza Reis Jr.  <ademar.reis@openbossa.org>
3540
3541         Reviewed by Dumitru Daniliuc.
3542
3543         check-webkit-style fails on operator+=, operator-=, ... methods
3544         https://bugs.webkit.org/show_bug.cgi?id=48258
3545
3546         * Scripts/webkitpy/style/checkers/cpp.py: Added exceptions
3547         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added unit-tests
3548
3549 2010-10-26  Ojan Vafai  <ojan@chromium.org>
3550
3551         Reviewed by Tony Chang.
3552
3553         remove DEFER support from new-run-webkit-tests
3554         https://bugs.webkit.org/show_bug.cgi?id=48387
3555
3556         DEFER was needed when we were trying to ship Chrome beta.
3557         Now it's just extra complication.
3558
3559         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
3560         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
3561         * Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
3562         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
3563         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3564         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
3565
3566 2010-10-26  Eric Seidel  <eric@webkit.org>
3567
3568         Unreviewed.  Adding import with_statement to fix python 2.5.
3569
3570         Teach webkit-patch how to read credentials from the environment
3571         https://bugs.webkit.org/show_bug.cgi?id=48275
3572
3573         * Scripts/webkitpy/common/net/credentials_unittest.py:
3574          - import with_statement to fix python 2.5.
3575
3576 2010-10-26  Alexey Proskuryakov  <ap@apple.com>
3577
3578         Reviewed by Sam Weinig.
3579
3580         https://bugs.webkit.org/show_bug.cgi?id=48375