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 which is caused by r72249.
6
7         https://bugs.webkit.org/show_bug.cgi?id=49740
8
9         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
10
11 2010-11-18  Tony Chang  <tony@chromium.org>
12
13         Reviewed by Adam Barth.
14
15         run platform/chromium/plugins/refcount-leaks.html on all platforms
16         https://bugs.webkit.org/show_bug.cgi?id=49485
17
18         Add PluginObject.testObjectCount which returns the number of allocated
19         TestObjects. Add PluginObject.testCreateTestObject which allocates
20         and returns a TestObject.
21
22         Add TestObject.refCount which returns the number of refs on the
23         TestObject.
24
25         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
26         (pluginGetProperty):
27         (pluginInvoke):
28         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
29         (getTestObjectCount):
30         (testAllocate):
31         (testDeallocate):
32         (testGetProperty):
33         (testScriptObjectInvoke): Release a ref to a plugin object that was
34             previously leaked.
35         * DumpRenderTree/TestNetscapePlugIn/TestObject.h:
36
37 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
38
39         Reviewed by Adam Roben.
40
41         WebKitTools projects (WebKitTestRunner, record-memory-win) should use vsprops for common build settings
42         https://bugs.webkit.org/show_bug.cgi?id=49711
43
44         * WebKitTestRunner/Configurations/WebKitTestRunnerCFLite.vsprops: Added.
45         * WebKitTestRunner/Configurations/WebKitTestRunnerCommon.vsprops: Added.
46         * WebKitTestRunner/Configurations/WebKitTestRunnerCoreFoundation.vsprops: Added.
47         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
48         * record-memory-win/record-memory-win-common.vsprops: Added.
49         * record-memory-win/record-memory-win.vcproj:
50
51 2010-11-18  Andras Becsi  <abecsi@inf.u-szeged.hu>
52
53         Reviewed by Csaba Osztrogonác.
54
55         [Qt][WK2] Only add user agent strings to the list which aren't listed yet.
56
57         * MiniBrowser/qt/BrowserWindow.cpp:
58         (BrowserWindow::updateUserAgentList):
59
60 2010-11-17  Adam Roben  <aroben@apple.com>
61
62         Don't trigger Windows builds when chromium-win test results change
63
64         Reviewed by Mark Rowe.
65
66         * Scripts/webkitpy/common/config/build.py:
67         (_should_file_trigger_build): Changed the regular expression used to
68         search for directory names to only consider full directory names,
69         rather than matching directory names where the string we care about is
70         a suffix of the name.
71
72         * Scripts/webkitpy/common/config/build_unittest.py:
73         (ShouldBuildTest): Added a test.
74
75 2010-11-17  Hayato Ito  <hayato@chromium.org>
76
77         Reviewed by Shinichiro Hamaji.
78
79         Remove a uri member from TestInput class.
80
81         https://bugs.webkit.org/show_bug.cgi?id=49691
82
83         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
84         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
85
86 2010-11-17  Adam Roben  <aroben@apple.com>
87
88         Move FindSafari's settings into a vsprops file
89
90         Fixes <http://webkit.org/b/49699> FindSafari should use a vsprops file
91
92         Reviewed by Steve Falkenburg.
93
94         * FindSafari/FindSafari.vcproj:
95         * FindSafari/FindSafariCommon.vsprops: Added.
96
97 2010-11-17  Adam Roben  <aroben@apple.com>
98
99         Use a minimum font size of 0 in DRT on Windows
100
101         This is the Windows equivalent of r72164.
102
103         Rubber-stamped by Dave Hyatt.
104
105         * DumpRenderTree/win/DumpRenderTree.cpp:
106         (resetDefaultsToConsistentValues):
107
108 2010-11-17  Dirk Pranke  <dpranke@chromium.org>
109
110         Reviewed by Andreas Kling.
111
112         webkit-patch pretty-diff shouldn't need to be run from a checkout root in git
113         https://bugs.webkit.org/show_bug.cgi?id=49639
114
115         * Scripts/webkitpy/common/checkout/scm.py:
116
117 2010-11-17  Mihai Parparita  <mihaip@chromium.org>
118
119         Reviewed by Tony Chang.
120
121         Rebaseline server: add loupe for image diffs
122         https://bugs.webkit.org/show_bug.cgi?id=49692
123         
124         Add a loupe (magnifiying glass) for inspecting image diffs. Shows an
125         enlarged area of the expected, actual and diff images side by side.
126
127         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
128         * Scripts/webkitpy/tool/commands/data/rebaselineserver/loupe.js: Added.
129         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
130         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
131
132 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
133
134         Reviewed by Adam Roben.
135
136         WebKitLauncherWin should use vsprops for shared build settings
137         https://bugs.webkit.org/show_bug.cgi?id=49696
138
139         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
140         * WebKitLauncherWin/WebKitLauncherWinCommon.vsprops: Added.
141
142 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
143
144         Reviewed by Adam Roben.
145
146         WinLauncher should use vsprops for shared build settings
147         https://bugs.webkit.org/show_bug.cgi?id=49695
148
149         * WinLauncher/WinLauncher.cpp:
150         (_tWinMain):
151         (PrintView):
152         (WndProc):
153         (MyEditProc):
154         * WinLauncher/WinLauncher.vcproj:
155         * WinLauncher/WinLauncherCommon.vsprops: Added.
156
157 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
158
159         Delete unused file unintentionally added in last change.
160
161         * DumpRenderTree/win/DumpRenderTreeCURL.vsprops: Removed.
162
163 2010-11-17  Steve Falkenburg  <sfalken@apple.com>
164
165         Reviewed by Adam Roben.
166
167         DumpRenderTree on Windows should use a vsprops file for shared project settings
168         https://bugs.webkit.org/show_bug.cgi?id=49690
169
170         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
171         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePluginCommon.vsprops: Added.
172         * DumpRenderTree/win/DumpRenderTree.vcproj:
173         * DumpRenderTree/win/DumpRenderTreeApple.vsprops: Added.
174         * DumpRenderTree/win/DumpRenderTreeCFLite.vsprops: Added.
175         * DumpRenderTree/win/DumpRenderTreeCURL.vsprops: Added.
176         * DumpRenderTree/win/DumpRenderTreeCairo.vsprops: Added.
177         * DumpRenderTree/win/DumpRenderTreeCommon.vsprops: Added.
178         * DumpRenderTree/win/ImageDiff.vcproj:
179         * DumpRenderTree/win/ImageDiffCommon.vsprops: Added.
180
181 2010-11-17  Hayato Ito  <hayato@chromium.org>
182
183         Refactor TestTypeBase.compare_output().
184
185         Introduce a TestOutput class and update compare_output() of each test
186         types so that they can take both actual and expected TestOutput objects.
187
188         https://bugs.webkit.org/show_bug.cgi?id=49431
189
190         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
191         * Scripts/webkitpy/layout_tests/layout_package/test_output.py: Added.
192         * Scripts/webkitpy/layout_tests/port/base.py:
193         * Scripts/webkitpy/layout_tests/port/chromium.py:
194         * Scripts/webkitpy/layout_tests/port/dryrun.py:
195         * Scripts/webkitpy/layout_tests/port/test.py:
196         * Scripts/webkitpy/layout_tests/port/webkit.py:
197         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
198         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
199         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
200         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
201
202 2010-11-17  Adam Roben  <aroben@apple.com>
203
204         Make each Windows Test builder use the same OS for all its slaves
205
206         Currently each WebKit1 Windows Test builder has one Windows XP slave
207         and one Windows 7 slave. Having different OSes in a single builder
208         meant that the builders would alternate between red and green when
209         there were OS-specific failures.
210
211         Fixes <http://webkit.org/b/49688> Windows Test builders alternate
212         between red and green because of multiple OS versions
213
214         Reviewed by Mark Rowe.
215
216         * BuildSlaveSupport/build.webkit.org-config/config.json: Renamed
217         "Windows Release (Tests)" to "Windows 7 Release (Tests)", "Windows
218         Debug (Tests)" to "Windows XP Debug (Tests)", and "Windows Release
219         (WebKit2 Tests)" to "Windows 7 Release (WebKit2 Tests)". Updated the
220         builders for the two WebKit1 builders so that they would all use the
221         appropriate OS.
222
223 2010-11-17  James Robinson  <jamesr@chromium.org>
224
225         Reviewed by Dan Bernstein.
226
227         [chromium] Set minimum font size to 0 for DRT
228         https://bugs.webkit.org/show_bug.cgi?id=49677
229
230         Followup for http://trac.webkit.org/changeset/72141.
231
232         Test: http/tests/misc/acid3.html
233
234         * DumpRenderTree/chromium/WebPreferences.cpp:
235         (WebPreferences::reset):
236
237 2010-11-17  Andras Becsi  <abecsi@webkit.org>
238
239         Unreviewed build fix.
240
241         [Qt] Add missing file I forgot in r72220.
242
243         * MiniBrowser/MiniBrowser.qrc: Added.
244
245 2010-11-17  Andras Becsi  <abecsi@webkit.org>
246
247         Reviewed by Andreas Kling.
248
249         [Qt][WK2] Make it possible to set custom user agent strings in MiniBrowser.
250         https://bugs.webkit.org/show_bug.cgi?id=49627
251
252         Add a user agent dialog and a resource file for useragentlist.txt to MiniBrowser.
253
254         * MiniBrowser/MiniBrowser.qrc: Added.
255         Add it here to prevent qmake from detecting it
256         since it needs to be copied to the build directory.
257         * MiniBrowser/qt/BrowserWindow.cpp:
258         (BrowserWindow::BrowserWindow):
259         (BrowserWindow::updateUserAgentList):
260         (BrowserWindow::showUserAgentDialog):
261         * MiniBrowser/qt/BrowserWindow.h:
262         * MiniBrowser/qt/MiniBrowser.pro:
263
264 2010-11-17  Adam Roben  <aroben@apple.com>
265
266         Make changes to the Mac WebKit2 Skipped file trigger Windows builds
267
268         We use that file on Windows!
269
270         Fixes <http://webkit.org/b/49643> Changes to
271         LayoutTests/platform/mac-wk2 should trigger a Windows build, but don't
272
273         Reviewed by Steve Falkenburg.
274
275         * Scripts/webkitpy/common/config/build.py:
276         (_should_file_trigger_build): Added the mac-wk2 directory and make it
277         trigger builds on SnowLeopard and Windows.
278
279         * Scripts/webkitpy/common/config/build_unittest.py:
280         (ShouldBuildTest.test_should_build): Added a test.
281
282 2010-11-17  Satish Sampath  <satish@chromium.org>
283
284         Reviewed by Jeremy Orlow.
285
286         Clear the speech input mock explicitly before each test.
287         https://bugs.webkit.org/show_bug.cgi?id=49660
288
289         * DumpRenderTree/chromium/LayoutTestController.cpp:
290         (LayoutTestController::setMockSpeechInputResult):
291         * DumpRenderTree/chromium/LayoutTestController.h:
292         * DumpRenderTree/chromium/WebViewHost.cpp:
293         (WebViewHost::speechInputController):
294         (WebViewHost::reset): Invoke speech input mock's clearResults method.
295         * DumpRenderTree/chromium/WebViewHost.h:
296         (WebViewHost::speechInputControllerMock): Moved mock from LayoutTestController.
297
298 2010-11-17  Alexander Pavlov  <apavlov@chromium.org>
299
300         Unreviewed, specify IRC nickname in committers.py.
301
302         * Scripts/webkitpy/common/config/committers.py:
303
304 2010-11-17  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
305
306         Reviewed by Tony Chang.
307
308         [NRWT] Make http locking similar to perl implementation
309         https://bugs.webkit.org/show_bug.cgi?id=49187
310
311         * Scripts/webkitpy/common/system/file_lock.py: Added.
312         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
313         * Scripts/webkitpy/layout_tests/port/http_lock.py:
314
315 2010-11-17  MORITA Hajime  <morrita@google.com>
316
317         Reviewed by Kent Tamura.
318
319         [Chromium][DRT] EventSender.contextClick() should aware spellchecking
320         https://bugs.webkit.org/show_bug.cgi?id=49366
321
322         - EvenSender: Checked WebContextMenuData.misspelledWord and added extra context menu entries
323           according to the spellchecker suggestion
324         - MockSpellCheck: add fillSuggestionList to provide fake suggestions.
325
326         test_expectations.txt will be changed after this change is ported to test_shell.
327         
328         * DumpRenderTree/chromium/EventSender.cpp:
329         (makeMenuItemStringsFor):
330         (EventSender::contextClick):
331         * DumpRenderTree/chromium/MockSpellCheck.cpp:
332         (MockSpellCheck::fillSuggestionList):
333         (MockSpellCheck::initializeIfNeeded):
334         * DumpRenderTree/chromium/MockSpellCheck.h:
335         * DumpRenderTree/chromium/WebViewHost.cpp:
336         (WebViewHost::mockSpellCheck):
337         * DumpRenderTree/chromium/WebViewHost.h:
338
339 2010-11-16  Dave Hyatt  <hyatt@apple.com>
340
341         Reviewed by Sam Weinig.
342
343         Make sure the pref for minimum font size is just explicitly set to 0 every time,
344         since the bots have 1 stuck in their plist otherwise.
345
346         * DumpRenderTree/mac/DumpRenderTree.mm:
347         (resetDefaultsToConsistentValues):
348
349 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
350
351         Reviewed by Tony Chang.
352
353         Rebaseline server: compute diffs client-side
354         https://bugs.webkit.org/show_bug.cgi?id=49640
355         
356         The image diff output from the DRT is pretty bad on some ports (at
357         least the Mac one), so it's better to compute diffs on the client by
358         using <canvas>.
359
360         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
361         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
362         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
363
364 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
365
366         Reviewed by Tony Chang.
367
368         Rebaseline server: display test results
369         https://bugs.webkit.org/show_bug.cgi?id=49626
370         
371         Adds basic result display to the rebaseline server. On the Python side
372         this involves:
373         - Parsing the unexpected_results.json into a dictionary.
374         - Serving it as JSON under /results.json.
375         (the JSON -> dict -> JSON transform isn't strictly necessary right now,
376         but I'll need to have access to the parsed results on the Python side
377         for follow-up changes).
378         
379         On the web UI side this adds:
380         - Markup for display image and text results (expected, actual, diff),
381           and JS for populating it.
382         - Markup for breaking down test results by failure type and directory,
383           and JS for populating it.
384
385         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
386         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
387         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
388         * Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js: Added.
389         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
390
391 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
392
393         Reviewed by Ojan Vafai.
394
395         See Chromium issue http://codereview.chromium.org/5133001/ - we
396         are modifying the build bots to pass in the builder name with
397         the "GPU" string appended instead of appending it in the code.
398
399         https://bugs.webkit.org/show_bug.cgi?id=49636
400
401         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
402
403 2010-11-16  Steve Falkenburg  <sfalken@apple.com>
404
405         Reviewed by Adam Roben.
406
407         Disable LTCG for Windows Release builds. Add new Release_LTCG configuration.
408         https://bugs.webkit.org/show_bug.cgi?id=49632
409
410         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
411         * DumpRenderTree/win/DumpRenderTree.vcproj:
412         * DumpRenderTree/win/ImageDiff.vcproj:
413         * MiniBrowser/MiniBrowser.vcproj:
414         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
415         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
416         * WebKitTestRunner/win/InjectedBundle.vcproj:
417         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
418         * WinLauncher/WinLauncher.vcproj:
419         * record-memory-win/record-memory-win.vcproj:
420
421 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
422
423         Reviewed by Ojan Vafai.
424
425         new-run-webkit-tests: rename TestInfo to TestInput, move image hash to work thread
426
427         Rename the TestInfo class to TestInput to be clearer about its
428         function, and move the checksum-reading code into dump_render_tree_thread
429         to avoid cross-thread access.
430
431         https://bugs.webkit.org/show_bug.cgi?id=49573
432
433         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
434         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
435         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
436
437 2010-11-16  Dave Hyatt  <hyatt@apple.com>
438
439         Fix bustage. Remove the minimum font size pref setting in DumpRenderTree.
440
441         * DumpRenderTree/mac/DumpRenderTree.mm:
442         (resetDefaultsToConsistentValues):
443
444 2010-11-16  Stephanie Lewis  <slewis@apple.com>
445
446         Reviewed by Geoff Garen.
447
448         <rdar://problem/8624267> Leak creating offscreen webview running fast/dom tests
449
450         Use shared DeviceOrientationProviderMock. The old allocation method created a new WebDeviceOrientationProviderMock 
451         for every WebView, and the WebKit API doesn't support that ownership model.
452
453         I also fixed a leak in setMockDeviceOrientation by adding a missing release.
454
455         * DumpRenderTree/mac/DumpRenderTree.mm:
456         (createWebViewAndOffscreenWindow):
457         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
458         (LayoutTestController::setMockDeviceOrientation):
459
460 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
461
462         Reviewed by Ojan Vafai.
463
464         new-run-webkit-tests: log the process id in --verbose mode
465
466         As part of the switch to multi-process mode, we should log the
467         process id when running w/ --verbose.
468
469         https://bugs.webkit.org/show_bug.cgi?id=49571
470
471         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
472
473 2010-11-16  Adam Roben  <aroben@apple.com>
474
475         Only trigger Windows builds when files that we actually use on Windows
476         are changed
477
478         A new Scheduler subclass, PlatformSpecificScheduler, has been added.
479         It uses the new webkitpy.common.config.build module to determine
480         whether a particular change should trigger a build on a particular
481         platform. The Windows builders have been switched to use a
482         PlatformSpecificScheduler.
483
484         The logic to determine whether or not a particular change should
485         trigger a build on a given platform has only been implemented/tested
486         for Windows. I tried to make it easy to add more platforms in the
487         future, but I don't have enough familiarity with all platforms to be
488         able to implement it for them.
489
490         Fixes <http://webkit.org/b/49407> Windows builders kick off builds for
491         lots irrelevant changes (e.g., rebaselining Chromium test results)
492
493         Reviewed by Eric Seidel.
494
495         * BuildSlaveSupport/build.webkit.org-config/config.json: Use a
496         PlatformSpecificScheduler for the Windows builders.
497
498         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
499         (PlatformSpecificScheduler.__init__): Added. Stores our platform, sets
500         up a ChangeFilter that filters to our branch and will call through to
501         our filter method, and calls up to our base class.
502         (PlatformSpecificScheduler.filter): Calls through to
503         build.should_build to find out whether we should trigger a build for
504         this change.
505
506         * Scripts/webkitpy/common/net/build.py: Added.
507         (_should_file_trigger_build): Uses a set of directories and regexp
508         patterns to determine whether the given file should trigger a build on
509         the given platform. As mentioned earlier, this has only been
510         implemented for Windows, though I did try to make some guesses about
511         other platforms.
512         (should_build): Returns true if any of the files should trigger a
513         build on the given platform.
514
515         * Scripts/webkitpy/common/net/build_unittest.py: Added.
516         (ShouldBuildTest.test_should_build): Does some basic testing to make
517         sure we're triggering builds for the right files. It only tests
518         Windows for now, though I tried to make some guesses about other
519         platforms.
520
521 2010-11-16  Simon Hausmann  <simon.hausmann@nokia.com>
522
523         Reviewed by Andreas Kling.
524
525         [Qt] Remove synchronous QWebPage::checkPermissions signal
526         https://bugs.webkit.org/show_bug.cgi?id=46810
527
528         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
529         (WebCore::WebPage::WebPage):
530         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
531         (LayoutTestController::grantDesktopNotificationPermission): When granting
532         permission, grant it directly on the QWebPage/Frame, that will remember it.
533         * QtTestBrowser/webpage.cpp:
534         (WebPage::WebPage):
535         * QtTestBrowser/webpage.h:
536
537
538 2010-11-16  Andras Becsi  <abecsi@inf.u-szeged.hu>
539
540         Reviewed by Csaba Osztrogonác.
541
542         [Qt] Add isPrinting and setPrinting methods to DRT's LayoutTestController.
543
544         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
545         (LayoutTestController::reset):
546         * DumpRenderTree/qt/LayoutTestControllerQt.h:
547         (LayoutTestController::isPrinting):
548         (LayoutTestController::setPrinting):
549
550 2010-11-16  Balazs Kelemen  <kbalazs@webkit.org>
551
552         Reviewed by Andreas Kling.
553
554         [Qt][WK2] Add a way to turn on tiled backing store in MiniBrowser
555         https://bugs.webkit.org/show_bug.cgi?id=49587
556
557         * MiniBrowser/qt/BrowserView.cpp:
558         (BrowserView::BrowserView): Added BackingStoreType
559         argument to ctor. Propagating it to the QGraphicsWKView.
560         * MiniBrowser/qt/BrowserView.h:
561         * MiniBrowser/qt/BrowserWindow.cpp:
562         (BrowserWindow::BrowserWindow):
563         (BrowserWindow::newWindow): Create the new BrowserWindow
564         with the same BackingStoreType that this instance has.
565         * MiniBrowser/qt/BrowserWindow.h: Added BackingStoreType member.
566         * MiniBrowser/qt/main.cpp:
567         (main): Use tiled backing store if got -tiled command line argument.
568         Avoid redundant copy of command line args.
569
570 2010-11-16  John Knottenbelt  <jknotten@chromium.org>
571
572         Reviewed by Jeremy Orlow.
573
574         Move DeviceOrientationClientMock from LayoutTestController to WebViewHost.
575         https://bugs.webkit.org/show_bug.cgi?id=48506
576
577         This ensures that there is one mock per page WebView / WebViewHost. This is a
578         design constraint of the DeviceOrientationClientMock because the m_controller
579         field can only store a pointer to a single instance of the corresponding
580         DeviceOrientationController.
581
582         Test: fast/dom/DeviceOrientation/no-page-cache.html
583
584         * DumpRenderTree/chromium/LayoutTestController.cpp:
585         (LayoutTestController::setMockDeviceOrientation):
586         (LayoutTestController::deviceOrientationClientMock):
587         * DumpRenderTree/chromium/LayoutTestController.h:
588         * DumpRenderTree/chromium/WebViewHost.cpp:
589         (WebViewHost::deviceOrientationClientMock):
590         (WebViewHost::deviceOrientationClient):
591         (WebViewHost::reset):
592         * DumpRenderTree/chromium/WebViewHost.h:
593
594 2010-11-15  Andras Becsi  <abecsi@webkit.org>
595
596         Reviewed by Andreas Kling.
597
598         [Qt][WK2] Avoid polling in WebKitTestRunner.
599         https://bugs.webkit.org/show_bug.cgi?id=49542
600
601         Make the WK2 testing session about 10-15% faster and decrease random flakiness
602         resulting from timing skew by removing the millisecond polling from TestControllerQt.
603
604         * WebKitTestRunner/TestController.cpp:
605         (WTR::TestController::didFinishLoadForFrame):
606         * WebKitTestRunner/TestController.h:
607         * WebKitTestRunner/TestInvocation.cpp:
608         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
609         * WebKitTestRunner/mac/TestControllerMac.mm:
610         (WTR::TestController::notifyDone): Add method stub.
611         * WebKitTestRunner/qt/TestControllerQt.cpp:
612         (WTR::TestControllerRunLoop::instance):
613         (WTR::TestControllerRunLoop::start):
614         (WTR::TestControllerRunLoop::stop):
615         (WTR::TestControllerRunLoop::TestControllerRunLoop):
616         (WTR::TestControllerRunLoop::timerEvent):
617         (WTR::TestController::notifyDone):
618         (WTR::TestController::platformRunUntil):
619         * WebKitTestRunner/win/TestControllerWin.cpp:
620         (WTR::TestController::notifyDone): Add method stub.
621
622 2010-11-15  Darin Adler  <darin@apple.com>
623
624         * Scripts/webkitpy/common/net/bugzilla: Added property svn:ignore.
625
626 2010-11-15  Benjamin Poulain  <benjamin.poulain@nokia.com>
627
628         Reviewed by Andreas Kling.
629
630         [Qt] [WK2] Add Ctrl+L shortcut to Qt Minibrowser
631         https://bugs.webkit.org/show_bug.cgi?id=49544
632
633         Add the shortcut to BrowserWindow.
634
635         * MiniBrowser/qt/BrowserWindow.cpp:
636         (BrowserWindow::BrowserWindow):
637         (BrowserWindow::openLocation):
638         * MiniBrowser/qt/BrowserWindow.h:
639
640 2010-11-14  David Levin  <levin@chromium.org>
641
642         Reviewed by Daniel Bates.
643
644         check-webkit-style should detect PassRefPtr usage in functions.
645         https://bugs.webkit.org/show_bug.cgi?id=49513
646
647         * Scripts/webkitpy/style/checkers/cpp.py:
648         (check_for_function_lengths): Revert a comment change that I
649          accidentally made in r71986.
650         (check_pass_ptr_usage): Added the code to do the check.
651         (process_line): Added the call to check_pass_ptr_usage.
652         (CppChecker): Added the new error category.
653         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
654         (CppStyleTestBase::perform_pass_ptr_check): Runs the new check for
655          testing purposes.
656         (PassPtrTest::*): The class/functions to unit test the new
657          functionality.
658
659 2010-11-14  David Levin  <levin@chromium.org>
660
661         Reviewed by Shinichiro Hamaji.
662
663         check-webkit-style function detection and the line count style checks should be separate.
664         https://bugs.webkit.org/show_bug.cgi?id=49512
665
666         * Scripts/webkitpy/style/checkers/cpp.py: Do the separation.
667         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Adjust the test to
668           call the detection function and fix line counts in two places now that
669           the code really only counts the lines in the body of the function.
670
671 2010-11-14  Andreas Kling  <kling@webkit.org>
672
673         Reviewed by Antonio Gomes.
674
675         http/tests/plugins tests print "Unhandled variable" to stderr under Qt
676         https://bugs.webkit.org/show_bug.cgi?id=33438
677
678         Stifle "Unhandled variable" warning to match what the other
679         TestNetscapePlugin does (NPP_GetValue in TestNetscapePlugIn/main.cpp
680         simply returns NPERR_GENERIC_ERROR for unhandled variables.)
681
682         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
683         (webkit_test_plugin_get_value):
684
685 2010-11-13  David Levin  <levin@chromium.org>
686
687         Reviewed by Eric Seidel.
688
689         check-webkit-style function detection crashes on functions in templates.
690         https://bugs.webkit.org/show_bug.cgi?id=49504
691
692         For example "template <bool shouldClose(const Element*)>".
693
694         * Scripts/webkitpy/style/checkers/cpp.py: If not function is detected, bail out.
695         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Modified the complex
696           function detection test to expose this issue.
697
698 2010-11-12  David Levin  <levin@chromium.org>
699
700         Reviewed by Shinichiro Hamaji.
701
702         check-webkit-style function detection doesn't detect indented functions declaractions.
703         https://bugs.webkit.org/show_bug.cgi?id=49446
704
705         Indented function declarations occur inside class definitions, so
706         they are a pretty common (and worth detecting).
707
708         * Scripts/webkitpy/style/checkers/cpp.py:
709           Changed regex to allow indentation.
710           Changed the function start detection to only happen when not in a
711           function.
712           Changed function end detection to work based on matching braces
713           instead of finding a close brace at the beginning of the line.
714           Fixed close_expression to do what it says when it doesn't find
715           the close.
716         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Indented function test.
717
718 2010-11-12  Daniel Bates  <dbates@rim.com>
719
720         Rubber-stamped by Nikolas Zimmermann.
721
722         Make do-file-rename rename files in the directory WebKit2.
723
724         This also makes the list of searched directories in do-file-rename
725         consistent with the list of searched directories in the script do-webcore-rename.
726
727         * Scripts/do-file-rename:
728
729 2010-11-12  Dirk Pranke  <dpranke@chromium.org>
730
731         Reviewed by James Robinson.
732
733         Attempt yet again to land the fix for bug 49360 (respecting
734         set-webkit-configuration). We need to handle the cases where
735         trying to run webkit-build-directory to find out where the
736         default configuration might be fails (that shows up on some
737         Chromium bots that apparently don't have perl installed).
738
739         https://bugs.webkit.org/show_bug.cgi?id=49360
740
741         * Scripts/webkitpy/layout_tests/port/config.py:
742         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Added.
743         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
744
745 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
746
747         Reviewed by Adam Barth.
748
749         webkit-patch land logs commit message twice
750         https://bugs.webkit.org/show_bug.cgi?id=49482
751         
752         Don't log the comment that's used by bugzilla.close_bug_as_fixed, since
753         that's just the commit message, which we now output following the
754         actual commit.
755
756         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
757
758 2010-11-12  Tony Chang  <tony@chromium.org>
759
760         Reviewed by Adam Barth.
761
762         run platform/chromium/plugins/script-object-invoke.html on all platforms
763         https://bugs.webkit.org/show_bug.cgi?id=49280
764
765         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
766         (testScriptObjectInvoke):
767         (pluginInvoke):
768
769 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
770
771         Reviewed by Adam Barth.
772
773         webkit-patch land-cowboy (and other commands) should display the committed SVN revision after landing a patch
774         https://bugs.webkit.org/show_bug.cgi?id=49471
775
776         Right now there's no indication of a successful commit.
777         
778         * Scripts/webkitpy/tool/commands/download_unittest.py:
779         * Scripts/webkitpy/tool/steps/commit.py:
780
781 2010-11-12  Kinuko Yasuda  <kinuko@chromium.org>
782
783         Reviewed by Ojan Vafai.
784
785         Include detailed test modifiers in results.json and enable incremental uploading for non-layout tests
786         https://bugs.webkit.org/show_bug.cgi?id=49354
787
788         Also moved/integrated the upload method from run_webkit_tests.py to json_results_generator.py.
789
790         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
791         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
792         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
793         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
794
795 2010-11-12  Tony Chang  <tony@chromium.org>
796
797         Reviewed by Kent Tamura.
798
799         [chromium] Load a blank page when closing DRT WebViewHosts
800         https://bugs.webkit.org/show_bug.cgi?id=49417
801
802         In test_shell, when we close a window, we first load about:blank
803         and run GC to fire the destruction logic of the page.  In test_shell,
804         this happens in ~TestShell.  In DRT, we manage virtual windows as
805         WebViewHosts, so we need to replicate this logic in ~WebViewHost.
806
807         Managing focus is a bit trickier, since normally you would get window
808         events to reset the focus to the right window.  We do it manually
809         instead.
810
811         This fixes 2 plugin tests on DRT.
812
813         * DumpRenderTree/chromium/TestShell.cpp:
814         (TestShell::~TestShell): Move window close logic to ~WebViewHost.
815         (TestShell::closeWindow): Move widget close() to WVH destructor and
816             manage focus manually.
817         (TestShell::closeRemainingWindows): Comment typo.
818         * DumpRenderTree/chromium/WebViewHost.cpp:
819         (WebViewHost::closeWidget):
820         (invokeCloseWidget):
821         (WebViewHost::closeWidgetSoon): Close the window asynchronously (matches
822             test_shell).
823         (WebViewHost::~WebViewHost):
824         * DumpRenderTree/chromium/WebViewHost.h:
825
826 2010-11-12  David Levin  <levin@chromium.org>
827
828         Reviewed by Adam Barth.
829
830         cpp.py needs some tidying.
831         https://bugs.webkit.org/show_bug.cgi?id=49443
832
833         * Scripts/webkitpy/style/checkers/cpp.py:
834           Fix parameter ordering for iteratively_replace_matches_with_char
835           to mirror other similar functions (and improved comments, etc.)
836           Also, use a python-ism (for/else) to get rid of a variable that
837           was there to determine if something had been found in the loop.
838
839 2010-11-12  Robert Kroeger  <rjkroege@chromium.org>
840
841         Reviewed by James Robinson.
842
843         [Chromium/DRT] Make EventSendingController send complete touch events.
844         https://bugs.webkit.org/show_bug.cgi?id=49285
845
846         * DumpRenderTree/chromium/EventSender.cpp:
847         (EventSender::addTouchPoint):
848         (EventSender::updateTouchPoint):
849         (EventSender::sendCurrentTouchEvent):
850
851 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
852
853         Unreviewed rollout of r71858.
854
855         Rollout out r71858 since it breaks new-run-webkit-httpd as used by the
856         NaCl tests.
857
858         * Scripts/webkitpy/layout_tests/port/config.py:
859         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Removed.
860         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
861
862 2010-11-12  Benjamin Poulain  <benjamin.poulain@nokia.com>
863
864         Reviewed by Kenneth Rohde Christiansen.
865
866         [Qt] TestController::runUntil doesn't honor the timeout parameter
867         https://bugs.webkit.org/show_bug.cgi?id=48941
868
869         Implement the missing timeout of WebKitTestRunner for Qt.
870
871         A QElapsedTimer has been added to RunUntilConditionLoop in order
872         to measure how long the loop has been running. When the timer
873         is bigger or equal than the timeout, the loop ends even if the
874         condition is not met.
875
876         * WebKitTestRunner/qt/TestControllerQt.cpp:
877         (WTR::RunUntilConditionLoop::start):
878         (WTR::RunUntilConditionLoop::run):
879         (WTR::RunUntilConditionLoop::timerEvent):
880         (WTR::TestController::platformRunUntil):
881
882 2010-11-11  David Levin  <levin@chromium.org>
883
884         Reviewed by Shinichiro Hamaji.
885
886         check-webkit-style function detection doesn't handle templates with spaces.
887         https://bugs.webkit.org/show_bug.cgi?id=49427
888
889         * Scripts/webkitpy/style/checkers/cpp.py: Stripped out templates when
890           finding the function name.
891         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test have a
892           template with a space.
893
894 2010-11-11  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
895
896         Reviewed by Shinichiro Hamaji.
897
898         Enable check-webkit-style on Qt files
899         https://bugs.webkit.org/show_bug.cgi?id=48322
900
901         Adding exception for _q_... private slots and
902         ..._data() methods in tests
903
904         * Scripts/webkitpy/style/checker.py: enable Qt dirs, add exceptions
905         * Scripts/webkitpy/style/checker_unittest.py: improve unittests
906
907 2010-11-11  David Levin  <levin@chromium.org>
908
909         Reviewed by Shinichiro Hamaji.
910
911         check-webkit-style function detection doesn't work for templates and destructors.
912         https://bugs.webkit.org/show_bug.cgi?id=49425
913
914         * Scripts/webkitpy/style/checkers/cpp.py: Changed regex for templates and destructors.
915         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test hit this issue.
916
917 2010-11-11  Joone Hur  <joone@kldp.org>
918
919         Reviewed by Antonio Gomes.
920
921         [GTK][DRT] Implement LayoutTestController::nodesFromRect
922         https://bugs.webkit.org/show_bug.cgi?id=46598
923
924         Support nodesFromRect in DRT
925
926         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
927         (LayoutTestController::nodesFromRect):
928
929 2010-11-11  MORITA Hajime  <morrita@google.com>
930
931         Reviewed by Kent Tamura.
932
933         editing/selection/context-menu-on-text.html fails on chromium
934         https://bugs.webkit.org/show_bug.cgi?id=45898
935
936         Enhance Chromium's EventSender::contextClick() to return
937         array of menu strings. The implementation incomplete because
938         Although actual context menu is implemented by the browser,
939         it reflects editability of underlying context thus makes
940         context-menu-on-text.html passed.
941         
942         * DumpRenderTree/chromium/EventSender.cpp:
943         (makeMenuItemStringsFor): Added.
944         (EventSender::contextClick): Now returns array of menu strings.
945         * DumpRenderTree/chromium/WebViewHost.cpp:
946         (WebViewHost::showContextMenu): Now storing given WebContextMenuData object.
947         (WebViewHost::clearContextMenuData): Added.
948         (WebViewHost::lastContextMenuData): Added.
949         * DumpRenderTree/chromium/WebViewHost.h:
950
951 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
952
953         Unreviewed, build fix.
954
955         fix build breakage caused by fix for bug 49380 (r71858).
956
957         config_standalone needs to ensure that the packages it needs are
958         in sys.path.
959
960         https://bugs.webkit.org/show_bug.cgi?id=49419
961
962         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
963
964 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
965
966         Reviewed by Adam Roben.
967
968         Fix NRWT to respect set-webkit-configuration again :(
969
970         This change fixes a typo in config.py that was causing the wrong
971         value to be read initially and us never actually looking into
972         the filesystem to get the default configuration.
973         
974         * Scripts/webkitpy/layout_tests/port/config.py:
975         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
976         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
977
978 2010-11-11  Eric Seidel  <eric@webkit.org>
979
980         Reviewed by Adam Barth.
981
982         Split out Bug, Attachment and CommitterValidator from bugzilla.py
983         https://bugs.webkit.org/show_bug.cgi?id=49403
984
985         CommitterValidator really had nothing to do with bugzilla.py.
986         I've put it next to committers.py for now since its rather tied to that file.
987
988         * Scripts/webkitpy/common/config/committervalidator.py: Added.
989         * Scripts/webkitpy/common/config/committervalidator_unittest.py: Added.
990         * Scripts/webkitpy/common/net/bugzilla/__init__.py:
991         * Scripts/webkitpy/common/net/bugzilla/attachment.py: Added.
992         * Scripts/webkitpy/common/net/bugzilla/bug.py: Added.
993         * Scripts/webkitpy/common/net/bugzilla/bug_unittest.py: Added.
994         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
995         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py:
996         * Scripts/webkitpy/tool/bot/feeders.py:
997         * Scripts/webkitpy/tool/commands/queues.py:
998
999 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
1000
1001         Reviewed by Eric Seidel.
1002
1003         fix missing 'pid' argument for http lockfile checking on win32
1004         https://bugs.webkit.org/show_bug.cgi?id=49363
1005
1006         * Scripts/webkitpy/common/system/executive.py:
1007
1008 2010-11-11  Eric Seidel  <eric@webkit.org>
1009
1010         Rubber-stamped by Adam Barth.
1011
1012         Move bugzilla.py into its own module in preparation for splitting one-file-per-class
1013         https://bugs.webkit.org/show_bug.cgi?id=49402
1014
1015         * Scripts/webkitpy/common/net/bugzilla/__init__.py: Added.
1016         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla.py.
1017         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla_unittest.py.
1018
1019 2010-11-10  Chris Guillory  <chris.guillory@google.com>
1020
1021         Reviewed by Chris Fleizach.
1022
1023         Use IAccessibleComparable to compare accessibility objects.
1024         https://bugs.webkit.org/show_bug.cgi?id=49118
1025
1026         * DumpRenderTree/AccessibilityUIElement.cpp:
1027         (AccessibilityUIElement::isEqual):
1028         * DumpRenderTree/AccessibilityUIElement.h:
1029         * DumpRenderTree/win/AccessibilityUIElementWin.cpp:
1030         (comparableObject):
1031         (AccessibilityUIElement::isEqual):
1032
1033 2010-11-10  Csaba Osztrogonác  <ossy@webkit.org>
1034
1035         Reviewed by David Hyatt.
1036
1037         HTML5 Ruby support should be mandatory feature
1038         https://bugs.webkit.org/show_bug.cgi?id=49272
1039
1040         Remove Ruby as optional feature.
1041
1042         * Scripts/build-webkit:
1043
1044 2010-11-10  Adam Roben  <aroben@apple.com>
1045
1046         Print VC++ Express build logs in the same order that the projects are
1047         built
1048
1049         Fixes <http://webkit.org/b/49326> print-vse-failure-logs prints logs
1050         in an unhelpful order
1051
1052         Reviewed by Eric Seidel.
1053
1054         * Scripts/print-vse-failure-logs:
1055         (PrintVisualStudioExpressLogs._build_order): Added. Uses
1056         print-msvc-project-dependencies to figure out the order in which
1057         projects are built.
1058         (PrintVisualStudioExpressLogs._sort_buildlogs): Added. Sorts the logs
1059         based on their build order and project name.
1060         (PrintVisualStudioExpressLogs._obj_directory): Moved code to find the
1061         scripts directory from here...
1062         (PrintVisualStudioExpressLogs._scripts_directory): ...to here.
1063         (PrintVisualStudioExpressLogs.main): Sort the logs before printing
1064         them.
1065
1066 2010-11-10  Adam Roben  <aroben@apple.com>
1067
1068         Only print the interesting text from VC++ Express build logs
1069
1070         Fixes <http://webkit.org/b/49325> It's hard to find the interesting
1071         output from print-vse-failure-logs
1072
1073         Reviewed by Eric Seidel.
1074
1075         * Scripts/print-vse-failure-logs: Removed unnecessary os.path import,
1076         added newly-required imports. Removed a comment that was essentially a
1077         FIXME, and is now fixed.
1078         (PrintVisualStudioExpressLogs._relevant_text): Added. Uses
1079         BeautifulSoup to extract the relevant text from the build log.
1080         (PrintVisualStudioExpressLogs.main): Only print the relevant text.
1081
1082 2010-11-09  Mihai Parparita  <mihaip@chromium.org>
1083
1084         Reviewed by Tony Chang.
1085
1086         [Chromium] http/tests/security/XFrameOptions fail with Chromium DRT
1087         https://bugs.webkit.org/show_bug.cgi?id=49286
1088         
1089         The tests in http/tests/security/XFrameOptions started to fail after
1090         r71297, since it changed the WebViewHost::assignIdentifierToRequest
1091         logic to always assign identifiers to requests, even if we're not going
1092         to be dumping resource load callbacks at that point. These tests
1093         only call LayoutTestController.dumpResourceLoadCallbacks after the
1094         request is started, thus their expectations have "<unknown>" as the
1095         identifier. Change WebViewHost to the previous behavior, which also
1096         match the other ports'.
1097
1098         * DumpRenderTree/chromium/WebViewHost.cpp:
1099         (WebViewHost::assignIdentifierToRequest):
1100         (WebViewHost::removeIdentifierForRequest):
1101
1102 2010-11-08  Tony Chang  <tony@chromium.org>
1103
1104         Reviewed by Adam Barth.
1105
1106         run platform/chromium/plugins/return-npobject.html on all platforms
1107         https://bugs.webkit.org/show_bug.cgi?id=49103
1108
1109         This tests that we can get an NPObject returned through a method on
1110         an NPAPI Object.
1111
1112         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1113         (testCallbackReturn):
1114         (pluginInvoke):
1115
1116 2010-11-09  Sheriff Bot  <webkit.review.bot@gmail.com>
1117
1118         Unreviewed, rolling out r71672.
1119         http://trac.webkit.org/changeset/71672
1120         https://bugs.webkit.org/show_bug.cgi?id=49276
1121
1122         It doesn't work on Chromium Windows bot (Requested by Ossy on
1123         #webkit).
1124
1125         * Scripts/webkitpy/common/system/file_lock.py: Removed.
1126         * Scripts/webkitpy/common/system/file_lock_unittest.py: Removed.
1127         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1128
1129 2010-11-09  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1130
1131         Reviewed by Tony Chang.
1132
1133         [NRWT] Make http locking similar to perl implementation
1134         https://bugs.webkit.org/show_bug.cgi?id=49187
1135
1136         * Scripts/webkitpy/common/system/file_lock.py: Added.
1137         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
1138         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1139
1140 2010-11-09  James Kozianski  <koz@chromium.org>
1141
1142         Reviewed by Eric Seidel.
1143
1144         Clean up imports for webkit-patch.
1145         https://bugs.webkit.org/show_bug.cgi?id=49083
1146
1147         Move some imports from commands into main.py where they are used.
1148         Move command imports into the __init__ of the commands module.
1149
1150         * Scripts/webkitpy/tool/commands/__init__.py:
1151         * Scripts/webkitpy/tool/commands/download.py:
1152         * Scripts/webkitpy/tool/commands/queues.py:
1153         * Scripts/webkitpy/tool/main.py:
1154
1155 2010-11-09  Yi Shen  <yi.4.shen@nokia.com>
1156
1157         Reviewed by Andreas Kling.
1158
1159         [Qt][QtTestBrowser] Don't erase incorrect url in the Url Bar
1160         https://bugs.webkit.org/show_bug.cgi?id=49047
1161
1162         * QtTestBrowser/mainwindow.cpp:
1163         (MainWindow::setAddressUrl):
1164
1165 2010-11-09  Andras Becsi  <abecsi@webkit.org>
1166
1167         Reviewed by Csaba Osztrogonác.
1168
1169         Improve old-run-webkit-tests --verbose to show DumpTool opening/closing
1170         https://bugs.webkit.org/show_bug.cgi?id=49252
1171
1172         * Scripts/old-run-webkit-tests:
1173
1174 2010-11-09  Csaba Osztrogonác  <ossy@webkit.org>
1175
1176         Reviewed by Tor Arne Vestbø.
1177
1178         Output of old-run-webkit-tests is buggy in verbose mode
1179         https://bugs.webkit.org/show_bug.cgi?id=49249
1180
1181         * Scripts/old-run-webkit-tests: Redundant outputs removed.
1182
1183 2010-11-08  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
1184
1185         Reviewed by Martin Robinson.
1186
1187         [GTK] Link with target name set does not work
1188         https://bugs.webkit.org/show_bug.cgi?id=48865
1189
1190         When a new page is created with a name (target=myFrame), the new
1191         mainFrame could not be found because they where not stored in the
1192         same PageGroup. As PageGroup are not exposed externally so the
1193         simpliest solution is to use a global page group name. This also fixes
1194         issue with visited link coloration across pages. After this change the
1195         private function webkit_web_view_set_group_name() was no longer used
1196         so it was removed completly.
1197
1198         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1199         (createWebView):
1200
1201 2010-11-08  John Knottenbelt  <jknotten@chromium.org>
1202
1203         Reviewed by Kent Tamura.
1204
1205         WebViewHost::reset() uses placement new.
1206         https://bugs.webkit.org/show_bug.cgi?id=49069
1207
1208         WebViewHost is using placement destruction / new to simulate a fresh
1209         WebViewHost object at the same address. This is because the WebView remains
1210         open across tests and maintains a pointer to the WebViewHost.
1211         This change resets member variables explictly instead of the placement new dance.
1212
1213         * DumpRenderTree/chromium/WebViewHost.cpp:
1214         (WebViewHost::WebViewHost):
1215         (WebViewHost::reset):
1216
1217 2010-11-08  Dirk Pranke <dpranke@chromium.org>
1218
1219         Reviewed by Eric Seidel.
1220
1221         Fix the webkit_base_dir logic in webkitpy/layout_tests/port/config.py. 
1222         It turns out that NRWT can't use the code in scm.find_checkout_root()
1223         because the Chromium bots don't do full checkouts of the WebKit
1224         tree; they only check out subdirectories like WebKitTools/Scripts.
1225         Until we can figure out a better approach for this, I've
1226         restored the base_dir-detecting code from NRWT, which works in
1227         any directory tree, scm or no.
1228
1229         This also restores the files modified in r71475 and r71474.
1230
1231         https://bugs.webkit.org/show_bug.cgi?id=49151
1232
1233         * WebKitTools/Scripts/webkitpy/layout_tests/port/config.py
1234         * WebKitTools/Scripts/webkitpy/layout_tests/port/config_unittest.py
1235
1236 2010-11-08  Eric Seidel  <eric@webkit.org>
1237
1238         Reviewed by Adam Barth.
1239
1240         webkit-patch failure-reason explodes if a build is missing
1241         https://bugs.webkit.org/show_bug.cgi?id=49195
1242
1243         This is likely a recent regression.  Adding a unit test for this case.
1244
1245         * Scripts/webkitpy/common/net/buildbot.py:
1246         * Scripts/webkitpy/common/net/buildbot_unittest.py:
1247
1248 2010-11-08  Dan Bernstein  <mitz@apple.com>
1249
1250         Reviewed by Darin Adler.
1251
1252         Ignore the system scrollbar setting.
1253
1254         * DumpRenderTree/mac/DumpRenderTree.mm:
1255         (resetDefaultsToConsistentValues): Set the scrollbar default.
1256
1257 2010-11-05  Adam Roben  <aroben@apple.com>
1258
1259         Make webkitpy.common.system.executive_unittest pass when running under
1260         Win32 Python
1261
1262         Fixes <http://webkit.org/b/49033>.
1263
1264         Reviewed by Dave Levin and Eric Seidel.
1265
1266         * Scripts/webkitpy/common/system/executive.py:
1267         (Executive._run_command_with_teed_output): Pass the arguments through
1268         encode_argument_if_needed rather than using Cygwin-specific code here.
1269         (Executive.run_and_throw_if_fail): Use child_process_encoding to decode
1270         the output.
1271         (Executive.run_command): Use encode_argument_if_needed to encode the
1272         arguments and child_process_encoding to decode the output.
1273         (Executive._child_process_encoding): Returns the encoding that should be
1274         used when communicating with child processes. On Windows we use mbcs,
1275         which maps to the current code page. On all other platforms we use
1276         UTF-8.
1277         (Executive._should_encode_child_process_arguments): Returns True if
1278         arguments to child processes need to be encoded. This is currently
1279         only needed on Cygwin and Win32 Python 2.x.
1280         (Executive._encode_argument_if_needed): Encode the argument using
1281         child_process_encoding if we need to encode arguments to child
1282         processes on this platform.
1283
1284         * Scripts/webkitpy/common/system/executive_unittest.py:
1285         (never_ending_command): Added. Returns arguments to run a command that
1286         will not quit until we kill it. On Windows we use wmic, on other
1287         platforms we use yes.
1288         (ExecutiveTest.test_run_command_with_unicode): Changed to expect the
1289         mbcs encoding to be used and for output from the child processes to
1290         have been roundtripped through encode/decode on Win32 Python. When
1291         UTF-8 is the encoding the roundtripping is undetectable, but with mbcs
1292         it's possible that some characters will not be able to be converted
1293         and will be replaced by question marks; the round-tripping allows us
1294         to expect this result.
1295
1296         (ExecutiveTest.test_kill_process):
1297         (ExecutiveTest.test_kill_all):
1298         Use never_ending_command instead of invoking "yes" directly. Expect an
1299         exit code of 1 when using Win32 Python, as that's what seems to happen.
1300
1301 2010-11-08  Adam Roben  <aroben@apple.com>
1302
1303         Roll out r71532
1304
1305         It broke the build for Cygwin 1.7 installs. Cygwin 1.7's default
1306         .bashrc unsets %TEMP%, which broke copy-tools.cmd.
1307
1308         * Scripts/webkitdirs.pm:
1309
1310 2010-11-08  Tony Chang  <tony@chromium.org>
1311
1312         Reviewed by Adam Barth.
1313
1314         run platform/chromium/plugins/nested-plugin-objects.html on all platforms
1315         https://bugs.webkit.org/show_bug.cgi?id=49094
1316
1317         This tests that objects created by plugins are proplery cleaned up.
1318
1319         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1320         (testPassTestObject):
1321         (pluginInvoke):
1322         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
1323         (testAllocate):
1324         (testDeallocate):
1325         (testGetProperty):
1326         (testConstruct):
1327
1328 2010-11-08  Adam Roben  <aroben@apple.com>
1329
1330         Mark Windows builds triggered from Perl as being non-interactive
1331
1332         This affects whether some of our scripts will show alerts vs. printing
1333         to the build log.
1334
1335         Fixes <http://webkit.org/b/49181> Windows build fail mysteriously when
1336         .vsprops files are updated
1337
1338         Reviewed by Steve Falkenburg.
1339
1340         * Scripts/webkitdirs.pm:
1341         (buildVisualStudioProject): Set WEBKIT_NONINTERACTIVE_BUILD to 1.
1342
1343 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1344
1345         Reviewed by Ojan Vafai.
1346
1347         Make http locking default in NRWT.
1348         https://bugs.webkit.org/show_bug.cgi?id=48053
1349
1350         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1351         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1352
1353 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1354
1355         Reviewed by Csaba Osztrogonác.
1356
1357         [NRWT] If the http lock fails we shouldn't do any locking
1358         https://bugs.webkit.org/show_bug.cgi?id=49164
1359
1360         If something goes wrong with the locking, the test should keep going.
1361
1362         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1363
1364 2010-11-08  Adam Roben  <aroben@apple.com>
1365
1366         Switch back to using kCGImageAlphaPremultipliedFirst when generating
1367         pixel dumps on Windows
1368
1369         I changed this behavior in r71418 thinking that it was required for
1370         getting plugins to show up in pixel dumps. But it doesn't seem to be
1371         necessary, and was making it impossible to compare new Windows pixel
1372         dumps with existing Windows or Mac pixel dumps (because ImageDiff won't
1373         compare an image with alpha to an image without alpha).
1374
1375         Fixes <http://webkit.org/b/49172> REGRESION (r71418): Can't compare
1376         new Windows pixel results to existing Windows or Mac results
1377
1378         Reviewed by Antti Koivisto.
1379
1380         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
1381         (createBitmapContextFromWebView): Replaced kCGImageAlphaNoneSkipFirst
1382         with kCGImageAlphaPremultipliedFirst.
1383
1384 2010-11-08  Csaba Osztrogonac  <ossy@webkit.org>
1385
1386         Unreviewed, rolling out r71466.
1387         http://trac.webkit.org/changeset/71466
1388         https://bugs.webkit.org/show_bug.cgi?id=48865
1389
1390         It broke layout tests on GTK bots.
1391
1392         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1393         (createWebView):
1394
1395 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1396
1397         Reviewed by Csaba Osztrogonác.
1398
1399         Enable running of Qt API tests on BuildBot
1400         https://bugs.webkit.org/show_bug.cgi?id=49004
1401
1402         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1403
1404 2010-11-08  Eric Seidel  <eric@webkit.org>
1405
1406         Reviewed by Adam Barth.
1407
1408         Add clean-review-queue command to remove closed bugs from the webkit.org/pending-review
1409         https://bugs.webkit.org/show_bug.cgi?id=49160
1410
1411         Bugzilla doesn't automatically remove r? when a bug gets closed.
1412         This script takes care of that for webkit.org.
1413
1414         * Scripts/webkitpy/common/net/bugzilla.py:
1415         * Scripts/webkitpy/tool/commands/upload.py:
1416
1417 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
1418
1419         Unreviewed, rolling out r71474.
1420         http://trac.webkit.org/changeset/71474
1421         https://bugs.webkit.org/show_bug.cgi?id=48280
1422
1423         breaks chromium webkit tests
1424         https://bugs.webkit.org/show_bug.cgi?id=49151
1425
1426         * Scripts/webkitpy/layout_tests/port/base.py:
1427         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
1428         * Scripts/webkitpy/layout_tests/port/config.py:
1429         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1430         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
1431         * Scripts/webkitpy/layout_tests/port/webkit.py:
1432         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1433         * Scripts/webkitpy/tool/mocktool.py:
1434
1435 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
1436
1437         Unreviewed, rolling out r71475.
1438         http://trac.webkit.org/changeset/71475
1439  
1440         breaks chromium webkit tests
1441         https://bugs.webkit.org/show_bug.cgi?id=49151
1442
1443         * Scripts/webkitpy/common/newstringio.py: Removed.
1444         * Scripts/webkitpy/common/newstringio_unittest.py: Removed.
1445         * Scripts/webkitpy/common/system/executive_mock.py: Removed.
1446         * Scripts/webkitpy/common/system/filesystem_mock.py: Removed.
1447         * Scripts/webkitpy/layout_tests/port/config_mock.py: Removed.
1448
1449 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1450
1451         Unreviewed, build fix.
1452
1453         Add files inexplicably not committed in r71474 as part of the
1454         fix for bug 48280.
1455
1456         * Scripts/webkitpy/common/newstringio.py: Added.
1457         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
1458         * Scripts/webkitpy/common/system/executive_mock.py: Added.
1459         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
1460         * Scripts/webkitpy/layout_tests/port/config_mock.py: Added.
1461
1462 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1463
1464         Reviewed by Eric Siedel.
1465
1466         new-run-webkit-tests: update port/base and port/webkit to use the
1467         new FileSystem and Config abstractions, pulling more logic out of
1468         the base Port classes into separate, mockable objects.
1469
1470         Also create a MockFileSystem object, a MockConfig object, move
1471         MockExecutive into common/system to be next to executive, and
1472         update the config object to use a FileSystem.
1473
1474         https://bugs.webkit.org/show_bug.cgi?id=48280
1475
1476         * Scripts/webkitpy/common/newstringio.py: Added.
1477         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
1478         * Scripts/webkitpy/common/system/executive_mock.py: Added.
1479         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
1480         * Scripts/webkitpy/layout_tests/port/base.py:
1481         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
1482         * Scripts/webkitpy/layout_tests/port/config.py:
1483         * Scripts/webkitpy/layout_tests/port/config_mock.py:
1484         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1485         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
1486         * Scripts/webkitpy/layout_tests/port/webkit.py:
1487         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1488         * Scripts/webkitpy/tool/mocktool.py:
1489
1490 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1491
1492         Unreviewed, build breakage.
1493
1494         Apparently I uploaded the wrong version of the file to fix 49122
1495         and neither Eric or I noticed - it was missing a dirname() call.
1496         Fixing ...
1497
1498         https://bugs.webkit.org/show_bug.cgi?id=49122
1499
1500         * Scripts/webkitpy/common/checkout/scm.py:
1501
1502 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1503
1504         Reviewed by Eric Seidel.
1505
1506         webkitpy/tool/* unittests change cwd and don't clean up properly
1507
1508         https://bugs.webkit.org/show_bug.cgi?id=49122
1509
1510         * Scripts/webkitpy/common/checkout/scm.py:
1511
1512 2010-11-05  Chris Marrin  <cmarrin@apple.com>
1513
1514         Reviewed by Simon Fraser.
1515
1516         suspendAnimations/resumeAnimations not present in WebKit2
1517         https://bugs.webkit.org/show_bug.cgi?id=49109
1518
1519         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
1520         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
1521         (WTR::LayoutTestController::suspendAnimations):
1522         (WTR::LayoutTestController::resumeAnimations):
1523         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
1524
1525 2010-11-05  Tony Chang  <tony@chromium.org>
1526
1527         Reviewed by David Levin.
1528
1529         cleanup style in TestNetscapePlugIn/PluginObject.cpp
1530         https://bugs.webkit.org/show_bug.cgi?id=49044
1531
1532         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1533         (getPluginClass):
1534         (pluginGetProperty):
1535         (pluginSetProperty):
1536         (variantToIdentifier):
1537         (testCallback):
1538         (getURL):
1539         (getURLNotify):
1540         (testInvokeDefault):
1541         (testGetProperty):
1542         (testDocumentOpen):
1543         (testWindowOpen):
1544         (pluginInvoke):
1545         (handleCallback):
1546         (notifyStream):
1547         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
1548
1549 2010-11-05  Eric Seidel  <eric@webkit.org>
1550
1551         Reviewed by David Levin.
1552
1553         Add QueueStatusServer/__init__.py so others can run the QueueStatusServer tests
1554         https://bugs.webkit.org/show_bug.cgi?id=49032
1555
1556         I wrote this file as part of bug 47847, but I forgot to commit it.
1557         No one else noticed it missing because test-webkitpy knows how
1558         to recover in the case where it can't import QueueStatusServer
1559         (which generally occurs due to not having installed the AppEngine SDK).
1560
1561         * QueueStatusServer/__init__.py: Added.
1562         * QueueStatusServer/model/workitems_unittest.py:
1563          - Remove a test which fails.  This was probably landed (by me)
1564            from my other machine, which since this __init__.py was missing
1565            I never noticed the failure and landed this invalid test.
1566            Sadly we can't really test remove_work_item as it depends
1567            on .key() working.  .key() will throw unless the object
1568            has already been saved it seems.
1569            This may be a mis-design in our remove_work_item implementation,
1570            but for now, just removing the test.
1571
1572 2010-11-04  Adam Roben  <aroben@apple.com>
1573
1574         Teach check-webkit-style about TestNetscapePlugIn
1575
1576         Fixes <http://webkit.org/b/49030> check-webkit-style is ignorant of
1577         TestNetscapePlugIn's coding conventions
1578
1579         Reviewed by John Sullivan.
1580
1581         * Scripts/webkitpy/style/checker.py: Lump TestNetscapePlugIn in with
1582         WebKitAPITest and TestWebKitAPI in having different include order and
1583         naming conventions than WebCore. Added some comments about why the
1584         exceptions exist.
1585
1586 2010-11-04  Adam Roben  <aroben@apple.com>
1587
1588         Add a test that shows that windowed plugins are able to paint
1589
1590         Somehow we never had a test for this before.
1591
1592         Fixes <http://webkit.org/b/49029> Should add a test that shows
1593         windowed plugins can paint
1594
1595         Reviewed by Jon Honeycutt.
1596
1597         * DumpRenderTree/TestNetscapePlugIn/Tests/win/DrawsGradient.cpp: Added.
1598         (DrawsGradient::DrawsGradient):
1599         (DrawsGradient::wndProc): We handle the WM_PAINT and WM_PRINTCLIENT messages.
1600
1601         (DrawsGradient::onPaint):
1602         (DrawsGradient::onPrintClient):
1603         These both just call through to paint.
1604
1605         (DrawsGradient::paint): Fills our client area with some gradients.
1606
1607         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Link
1608         against Msimg32.lib for ::GradientFill and added DrawsGradient.
1609
1610         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
1611         (createBitmapContextFromWebView): Use WM_PRINT instead of
1612         WM_PRINTCLIENT so that ::DefWindowProc will send
1613         WM_PRINT/WM_PRINTCLIENT messages to the WebView's child windows.
1614         Replaced kCGImageAlphaPremultipledFirst with kCGImageAlphaNoneSkipFirst
1615         because GDI doesn't support alpha and kCGImageBitmapOrder32Little with
1616         kCGImageBitmapOrder32Host because, who knows, maybe someday Windows
1617         will run on a big-endian platform.
1618
1619 2010-11-04  Adam Roben  <aroben@apple.com>
1620
1621         Extract much of NPNInvalidateRectInvalidatesWindow's code into a
1622         WindowedPluginTest base class
1623
1624         The base class takes care of subclassing the plugin's window so that a
1625         custom WNDPROC is called. This will make it easier to write tests that
1626         need to handle window messages.
1627
1628         Fixes <http://webkit.org/b/49028> It's hard to write a PluginTest with
1629         a custom WNDPROC
1630
1631         Reviewed by Jon Honeycutt.
1632
1633         * DumpRenderTree/TestNetscapePlugIn/Tests/win/NPNInvalidateRectInvalidatesWindow.cpp:
1634         (NPNInvalidateRectInvalidatesWindow::NPNInvalidateRectInvalidatesWindow):
1635         (NPNInvalidateRectInvalidatesWindow::NPP_SetWindow):
1636         (NPNInvalidateRectInvalidatesWindow::wndProc):
1637         (NPNInvalidateRectInvalidatesWindow::testInvalidateRect):
1638         Moved code from here to WindowedPluginTest. Changed to use window()
1639         instead of m_window.
1640
1641         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1642         Added WindowedPluginTest and added TestNetscapePlugIn/win to the
1643         include path. Also reordered the include path to make a little more
1644         sense and simplified the entry that added TestNetscapePlugIn itself to
1645         the include path.
1646
1647         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.cpp: Added.
1648         (WindowedPluginTest::WindowedPluginTest):
1649         (WindowedPluginTest::NPP_SetWindow):
1650         (WindowedPluginTest::staticWndProc):
1651         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.h: Added.
1652         (WindowedPluginTest::window):
1653         Code came from NPNInvalidateRectInvalidatesWindow.
1654
1655 2010-11-04  Adam Roben  <aroben@apple.com>
1656
1657         Add a plugin test to show that windowed plugins are clipped correctly
1658
1659         Fixes <http://webkit.org/b/49024> <rdar://problem/8487847> Windowed
1660         plugins aren't clipped in WebKit2 on Windows
1661
1662         Reviewed by Jon Honeycutt.
1663
1664         * DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowRegionIsSetToClipRect.cpp: Added.
1665         (WindowRegionIsSetToClipRect::WindowRegionIsSetToClipRect): Initialize members.
1666         (WindowRegionIsSetToClipRect::NPP_SetWindow): Check that our window
1667         region matches the clip rect we know we should have based on
1668         window-region-is-set-to-clip-rect.html, and check that our window class
1669         doesn't have the CS_PARENTDC style.
1670
1671         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1672         Added WindowRegionIsSetToClipRect.
1673
1674 2010-11-05  Alejandro G. Castro  <alex@igalia.com>
1675
1676         Reviewed by Martin Robinson.
1677
1678         [GTK] Avoid font hinting in the DRT
1679         https://bugs.webkit.org/show_bug.cgi?id=48548
1680
1681         Change the settings to avoid font hinting, it was causing
1682         different results depending on the distribution.
1683
1684         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1685         (initializeGtkFontSettings):
1686         * DumpRenderTree/gtk/fonts/fonts.conf:
1687
1688 2010-11-05  Eric Seidel  <eric@webkit.org>
1689
1690         Reviewed by Adam Barth.
1691
1692         build-webkit spams system.log by repeatedly checking for PBXProductDirectory in com.apple.Xcode
1693         https://bugs.webkit.org/show_bug.cgi?id=49051
1694
1695         This is a speculative fix.  The unit tests cover these methods, however
1696         I don't know if this will fully stop the system.log spam.
1697
1698         * Scripts/webkitpy/layout_tests/port/base.py:
1699         * Scripts/webkitpy/layout_tests/port/webkit.py:
1700
1701 2010-11-05  Eric Seidel  <eric@webkit.org>
1702
1703         Reviewed by Adam Barth.
1704
1705         Add basic support for showing bot id on /queue-status/ pages
1706         https://bugs.webkit.org/show_bug.cgi?id=49037
1707
1708         This support is really simple.  Eventually we'll want to
1709         show the bot id in the lock table too, but we don't have
1710         that information stored in the server yet.
1711
1712         * QueueStatusServer/handlers/queuestatus.py:
1713         * QueueStatusServer/handlers/queuestatus_unittest.py: Added.
1714         * QueueStatusServer/templates/includes/singlequeuestatus.html:
1715
1716 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
1717
1718         Reviewed by Adam Barth.
1719
1720         Create a filesystem wrapper that we can use to enforce
1721         particular conventions and use for mocking and dependency
1722         injection down the line.
1723
1724         https://bugs.webkit.org/show_bug.cgi?id=48144
1725
1726         * Scripts/webkitpy/common/system/filesystem.py: Added.
1727         * Scripts/webkitpy/common/system/filesystem_unittest.py: Added.
1728
1729 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1730
1731         Reviewed by Tony Chang.
1732
1733         run_webkit_tests_unittest fails under Python 2.5
1734         https://bugs.webkit.org/show_bug.cgi?id=49043
1735         
1736         Switch from itertools.chain.from_iterable (which was added in 2.6)
1737         to using itertools.chain directly.
1738
1739         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1740
1741 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1742
1743         Unreviewed fix to webkit-patch rebaseline-server so that it runs under
1744         Python 2.5 (it needs an import to use the with statement).
1745
1746         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
1747
1748 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
1749
1750         Reviewed by Adam Barth.
1751
1752         new-run-webkit-tests: split out webkit-specific configuration stuff into a new module
1753
1754         The current NRWT code has webkit-specific configuration code (like
1755         _script_path, default configuration, etc.) mixed in with
1756         layout-test-specific stuff in port/base. The configuration code
1757         should be split out into a separate module for easier mocking,
1758         testing, and isolation.
1759
1760         https://bugs.webkit.org/show_bug.cgi?id=48264
1761
1762         * Scripts/webkitpy/layout_tests/port/config.py: Added.
1763         * Scripts/webkitpy/layout_tests/port/config_unittest.py: Added.
1764
1765 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1766
1767         Reviewed by Tony Chang.
1768
1769         Rebaseline server: initial framework
1770         https://bugs.webkit.org/show_bug.cgi?id=48892
1771         
1772         Adds the basic framework for the rebaseline server (details at 
1773         http://webkit.org/b/47761). Includes the rebaseline-server webkit-patch
1774         command, which starts an HTTP server that can serve static files or
1775         invoke handler methods on a class.
1776
1777         * Scripts/webkitpy/tool/commands/__init__.py:
1778         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html: Added.
1779         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css: Added.
1780         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js: Added.
1781         * Scripts/webkitpy/tool/commands/rebaselineserver.py: Added.
1782
1783 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
1784
1785         Reviewed by James Robinson.
1786
1787         new-run-webkit-tests wasn't using DRT by default for
1788         --platform chromium-gpu
1789
1790         The default value was set to False instead of None, which meant
1791         that the platform specific logic wasn't firing to change the
1792         value to True (b/c we were afraid we'd be overriding the user
1793         preference).
1794
1795         https://bugs.webkit.org/show_bug.cgi?id=49038
1796
1797         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1798
1799 2010-11-04  Tony Chang  <tony@chromium.org>
1800
1801         Reviewed by Adam Barth.
1802
1803         make platform/chromium/plugins/multiple-plugins.html pass on all platforms
1804         https://bugs.webkit.org/show_bug.cgi?id=49026
1805
1806         Have the testObject.bar property return the string "bar".  This lets
1807         us run and pass platform/chromium/plugins/multiple-plugins.html on all
1808         platforms.
1809
1810         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
1811         (testGetProperty):
1812
1813 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1814
1815         Reviewed by Tony Chang.
1816
1817         NRWT doesn't have good test coverage for --run-chunk, --batch-size, --run-part, etc.
1818         https://bugs.webkit.org/show_bug.cgi?id=48878
1819         
1820         Add get_tests_run so that it's easy to see which tests get run (and with
1821         what batching) for a given flag combination. Flesh out the various
1822         test cases that have FIXMEs.
1823         
1824         Also fixes an off-by-one error (batch sizes were one larger than 
1825         expected) and makes --run-part also have wraparound behavior, like
1826         --run-chunk.
1827
1828         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1829         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1830         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1831
1832 2010-11-04  Erik Arvidsson  <arv@chromium.org>
1833
1834         Reviewed by Dimitri Glazkov.
1835
1836         Support box-sizing without the vendor prefix
1837         https://bugs.webkit.org/show_bug.cgi?id=36713
1838
1839         Based on patch by Peter Beverloo <peter@lvp-media.com>
1840
1841         * iExploder/htdocs/cssproperties.in: Change -webkit-box-sizing to box-sizing.
1842
1843 2010-11-04  Csaba Osztrogonác  <ossy@webkit.org>
1844
1845         Unreviewed rollout r71340, because it broke Chromium Windows bot.
1846
1847         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1848         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1849
1850 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1851
1852         Reviewed by Ojan Vafai.
1853
1854         Make http locking default in NRWT.
1855         https://bugs.webkit.org/show_bug.cgi?id=48053
1856
1857         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1858         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1859
1860 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1861
1862         Reviewed by Eric Seidel.
1863
1864         [NRWT] Clear invalid http locks on Windows platform as well
1865         https://bugs.webkit.org/show_bug.cgi?id=48515
1866
1867         * Scripts/webkitpy/common/system/executive.py:
1868         * Scripts/webkitpy/common/system/executive_unittest.py:
1869         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1870
1871 2010-11-02  Adam Roben  <aroben@apple.com>
1872
1873         Reduce our dependence on coreutils when running Python tests
1874
1875         This patch introduces versions of the standard echo and cat utilities
1876         implemented in Python. They are probably missing features of their
1877         coreutils equivalents, but they can do what's necessary for our Python
1878         tests. This is useful on Windows, where these utilities typically
1879         aren't available.
1880
1881         Fixes <http://webkit.org/b/48883> executive_unittest relies on echo
1882         and cat utilities from coreutils, which are not present on Windows
1883
1884         Reviewed by Eric Seidel.
1885
1886         * Scripts/webkitpy/common/system/executive_unittest.py: Changed to use
1887         our Python-based echo and cat.
1888
1889         * Scripts/webkitpy/common/system/fileutils.py: Added.
1890         (make_stdout_binary): On Windows, puts sys.stdout into binary mode so
1891         that \n won't be translated into \r\n. I couldn't think of a good way
1892         to test this directly without touching the filesystem, but it is tested
1893         indirectly by echo_unittest.
1894
1895         * Scripts/webkitpy/test/cat.py: Added.
1896         (command_arguments): Returns a list for invoking cat with the given arguments.
1897         (main): Acts like a simplified version of the coreutils cat utility.
1898
1899         * Scripts/webkitpy/test/cat_unittest.py: Added.
1900         (CatTest.assert_cat): Runs cat with the given input and ensures the
1901         output matches the input.
1902         (CatTest.test_basic): Performs a simple test of cat.
1903         (CatTest.test_no_newline): Tests what happens when the input string
1904         doesn't have a trailing newline.
1905         (CatTest.test_unicode): Tests passing a unicode string to cat.
1906         (CatTest.test_as_command): Tests running cat as a separate command.
1907
1908         * Scripts/webkitpy/test/echo.py: Added.
1909         (command_arguments): Returns a list for invoking echo with the given arguments.
1910         (main): Acts like a simplified version of the coreutils echo utility.
1911
1912         * Scripts/webkitpy/test/echo_unittest.py: Added.
1913         (EchoTest.test_basic): Performs a simple test of echo.
1914         (EchoTest.test_no_newline): Tests passing -n to echo to suppress the
1915         trailing newline.
1916         (EchoTest.test_unicode): Tests passing unicode and non-unicode strings
1917         to echo.
1918         (EchoTest.test_argument_order): Tests what happens when -n is not the
1919         first argument.
1920         (EchoTest.test_empty_arguments): Tests what happens when you pass [] to
1921         echo.main.
1922         (EchoTest.test_no_arguments): Tests what happens when you call
1923         echo.main with no arguments.
1924         (EchoTest.test_as_command): Tests running echo as a separate command.
1925
1926 2010-11-04  Renata Hodovan  <reni@inf.u-szeged.hu>
1927
1928         Unreviewed: Add myself to the list of Committers.
1929
1930         * Scripts/webkitpy/common/config/committers.py:
1931
1932 2010-11-04  Andreas Kling  <kling@webkit.org>
1933
1934         Reviewed by Kenneth Rohde Christiansen.
1935
1936         [Qt] Hook into QWebPage::scrollRequested for resizes-to-contents mode
1937
1938         In RTC mode the QGraphicsWebView item is the size of the contents,
1939         scrolling works a bit differently  (we need to react to scrollRequested.)
1940
1941         Normally QGraphicsView will replay the last mouse event when scrolling,
1942         so to prevent WebKit from getting confused by this we temporarily make
1943         the QGraphicsView non-interactive.
1944
1945         * QtTestBrowser/webview.cpp:
1946         (WebViewGraphicsBased::setPage):
1947         (WebViewGraphicsBased::scrollRequested):
1948         * QtTestBrowser/webview.h:
1949
1950 2010-11-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
1951
1952         Reviewed by Andreas Kling.
1953
1954         Use OS(MAC_OS_X) rather than PLATFORM(MAC) in TestControllerQt
1955
1956         PLATFORM(MAC) is not defined for the Qt port, as it refers to
1957         the Mac-port, not the Mac OS X operating system.
1958
1959         * WebKitTestRunner/qt/TestControllerQt.cpp:
1960         (WTR::isExistingLibrary):
1961
1962 2010-11-04  Leonid Ebril  <leonid.ebril@nokia.com>
1963
1964         Reviewed by Andreas Kling.
1965
1966         [Qt] Adding iphone user agent string the user agent list for QtTestBrowser
1967         https://bugs.webkit.org/show_bug.cgi?id=48869
1968
1969         * QtTestBrowser/useragentlist.txt:
1970
1971 2010-11-03  Adam Roben  <aroben@apple.com>
1972
1973         Always use uppercase drive names in strings returned by abspath_to_uri
1974
1975         Some versions of cygpath use lowercase drive letters while others use
1976         uppercase, which makes it hard to test the output of code that uses
1977         cygpath.
1978
1979         Fixes <http://webkit.org/b/48914> webkitpy.common.system.path_unittest
1980         fails with Cygwin 1.5
1981
1982         Reviewed by Eric Seidel.
1983
1984         * Scripts/webkitpy/common/system/path.py:
1985         (cygpath): Updated the docstring to indicate that only absolute paths
1986         should be passed for now (though relative paths will work fine).
1987         (_Cygpath.convert): Upper-case the first letter of the converted Windows path.
1988
1989 2010-11-03  George Guo  <George.Guo@Nokia.com>
1990
1991         Reviewed by Andreas Kling.
1992
1993         [Qt] QtTestBrowser : set mmp rule pageddata in Symbian
1994         https://bugs.webkit.org/show_bug.cgi?id=48767
1995     
1996         Paging is needd on Symbian devices to support benchmarks tests like
1997         dromaeo.com and Celtic Kane that need a lot of memory to run  
1998
1999         * QtTestBrowser/QtTestBrowser.pro:
2000
2001 2010-11-03  Jenn Braithwaite  <jennb@chromium.org>
2002
2003         Reviewed by Dmitry Titov.
2004
2005         Chromium: Update resource tracking when moving a frame between documents
2006         https://bugs.webkit.org/show_bug.cgi?id=48363
2007
2008         * DumpRenderTree/chromium/WebViewHost.cpp:
2009         (WebViewHost::assignIdentifierToRequest):
2010         Always put resource id in map so we can make assumptions about its
2011         presence.
2012         (WebViewHost::removeIdentifierForRequest):
2013         Added.
2014         (WebViewHost::didFinishResourceLoad):
2015         (WebViewHost::didFailResourceLoad):
2016         Check existence of resource id before removing from map.
2017         * DumpRenderTree/chromium/WebViewHost.h:
2018
2019 2010-11-03  Victor Wang  <victorw@chromium.org>
2020
2021         Reviewed by Adam Barth.
2022
2023         [Chromium] update buildbot names in chromium rebaseline tool.
2024
2025         https://bugs.webkit.org/show_bug.cgi?id=48881
2026
2027         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2028
2029 2010-11-03  Darin Adler  <darin@apple.com>
2030
2031         Updated Xcode projects by opening them with Xcode 3.2.4.
2032         Updated svn:ignore for Xcode projects.
2033
2034         * MiniBrowser/MiniBrowser.xcodeproj: Added property svn:ignore.
2035         * TestWebKitAPI/TestWebKitAPI.xcodeproj: Added property svn:ignore.
2036         * WebKitLauncher/WebKitLauncher.xcodeproj: Modified property svn:ignore.
2037         * WebKitLauncher/WebKitLauncher.xcodeproj/project.pbxproj: Updated with Xcode 3.2.4.
2038
2039 2010-11-03  Adam Roben  <aroben@apple.com>
2040
2041         Stop waiting for messages from the web process after a timeout elapses
2042
2043         TestController::runUntil now takes a timeout parameter that specifies
2044         how long to wait for the condition to become true. WebKitTestRunner
2045         takes a --timeout flag that specifies how long the timeout should be.
2046         run-webkit-tests passes this flag to WebKitTestRunner so its timeout
2047         will be similar to run-webkit-tests.
2048
2049         Fixes <http://webkit.org/b/43047> <rdar://problem/8365833>
2050         WebKitTestRunner waits forever if the web process crashes
2051
2052         Reviewed by Darin Adler and Anders Carlsson.
2053
2054         * Scripts/old-run-webkit-tests:
2055         (top level): Moved the GuardMalloc timeout adjustment here from
2056         readFromDumpToolWithTimer.
2057         (openDumpTool): Make WTR use a timeout similar to but slightly shorter
2058         than the one that was specified on the command line.
2059
2060         * WebKitTestRunner/TestController.cpp:
2061         (WTR::TestController::TestController): Initialize our timeout values.
2062         (WTR::TestController::initialize): Parse the --timeout flag and use it
2063         to modify our timeout values.
2064         (WTR::TestController::resetStateToConsistentValues): Changed to use a
2065         short timeout while waiting for the web process to reset and to return
2066         a boolean indicating whether we were able to reset the web process.
2067         Uses a 5-second timeout while waiting for the process to be reset.
2068         (WTR::TestController::runTest): Changed to return a boolean indicating
2069         whether we were able to reset the web process (and thus run the test).
2070         (WTR::TestController::runUntil): Call through to platformRunUntil.
2071
2072         (WTR::TestController::runTestingServerLoop):
2073         (WTR::TestController::run):
2074         Changed to bail if any test can't be run. This will cause the process
2075         to exit. (Unfortunately this will make run-webkit-tests think we
2076         crashed; see <http://webkit.org/b/48943>.)
2077
2078         * WebKitTestRunner/TestController.h: Added platformRunUntil,
2079         m_longTimeout, and m_shortTimeout.
2080
2081         * WebKitTestRunner/TestInvocation.cpp:
2082         (WTR::TestInvocation::invoke): Use a short timeout when waiting for the
2083         initial response and a long timeout when waiting for the test to
2084         complete. Check whether runUntil timed out and print an error message
2085         if so.
2086
2087         * WebKitTestRunner/mac/TestControllerMac.mm:
2088         (WTR::TestController::platformRunUntil): Renamed from runUntil. Pass
2089         [NSDate distantPast] to -[NSRunLoop runMode:beforeDate:] so that we
2090         won't block waiting for the run loop. Only loop until the timeout
2091         elapses.
2092
2093         * WebKitTestRunner/qt/TestControllerQt.cpp:
2094         (WTR::TestController::platformRunUntil): Renamed from runUntil. Added a
2095         FIXME about honoring the timeout.
2096
2097         * WebKitTestRunner/win/TestControllerWin.cpp:
2098         (WTR::TestController::platformRunUntil): Renamed from runUntil. Use
2099         ::MsgWaitForMultipleObjectsEx to implement the timeout. Changed to use
2100         ::PeekMessageW so that we don't block waiting for messages to become
2101         available.
2102
2103 2010-11-03  Adam Roben  <aroben@apple.com>
2104
2105         Add a plugin test that evaluates JS after removing the plugin element
2106         from the document
2107
2108         This test replaces platform/win/plugins/plugin-delayed-destroy.html.
2109         That test was made to prevent a crash very similar to this one, but
2110         unfortunately tested only the mechanism that prevented the crash and
2111         not whether the crash itself was prevented. Since WebKit2 uses a
2112         different mechanism to prevent the crash, the test was failing even
2113         though WebKit2 was not vulnerable to the crash. This new test crashes
2114         if there is no mechanism in place to prevent it and passes in both
2115         WebKit1 and WebKit2.
2116
2117         Fixes <http://webkit.org/b/46711> <rdar://problem/8485903>
2118         platform/win/plugins/plugin-delayed-destroy.html fails in WebKit2
2119
2120         Reviewed by Anders Carlsson.
2121
2122         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
2123         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2124         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
2125         * GNUmakefile.am:
2126         Added new file.
2127
2128         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
2129         (pluginDeallocate): Make sure we delete the PluginTest object. This
2130         prevents a leak and also allows us to test the crash.
2131
2132         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
2133         (PluginTest::executeScript): Made this into a non-static member
2134         function.
2135
2136         (PluginTest::waitUntilDone):
2137         (PluginTest::notifyDone):
2138         Updated for changes to executeScript.
2139
2140         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Added executeScript.
2141
2142         * DumpRenderTree/TestNetscapePlugIn/Tests/EvaluateJSAfterRemovingPluginElement.cpp: Added.
2143         (EvaluateJSAfterRemovingPluginElement::EvaluateJSAfterRemovingPluginElement):
2144         Initialize ourselves and tell the test harness to wait.
2145         (EvaluateJSAfterRemovingPluginElement::NPP_DestroyStream): Remove our
2146         plugin element from the document, then execute some JavaScript. If
2147         WebKit does not have appropriate mechanisms in place, we'll be
2148         destroyed inside the first call to executeScript and crash on the
2149         second call.
2150
2151 2010-11-02  Stephen White  <senorblanco@chromium.org>
2152
2153         Reviewed by Tony Chang.
2154
2155         [chromium] Fix LayoutTestController UMRs.
2156         https://bugs.webkit.org/show_bug.cgi?id=48872
2157
2158         * DumpRenderTree/chromium/LayoutTestController.cpp:
2159         (LayoutTestController::LayoutTestController):
2160
2161 2010-11-03  Kent Tamura  <tkent@chromium.org>
2162
2163         Reviewed by Dimitri Glazkov.
2164
2165         REGRESSION: rebaseline-chromium-webkit-tests uses non-zero tolerance for
2166         image dup detection
2167         https://bugs.webkit.org/show_bug.cgi?id=48744
2168
2169         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2170          - Make a function for option parsing for ease of test
2171          - Set 0 to options.tolerance
2172         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
2173          - Add a test for this change
2174
2175 2010-11-02  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
2176
2177         Unreviewed: Add myself to the list of Committers.
2178
2179         * Scripts/webkitpy/common/config/committers.py:
2180
2181 2010-11-02  Anders Carlsson  <andersca@apple.com>
2182
2183         Fix build.
2184
2185         * TestWebKitAPI/PlatformUtilities.cpp:
2186         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
2187
2188 2010-11-02  Sam Weinig  <sam@webkit.org>
2189
2190         Reviewed by Anders Carlsson.
2191
2192         Add a way to send startup messages on the context which can be posted when a process launches
2193         <rdar://problem/8617928>
2194         https://bugs.webkit.org/show_bug.cgi?id=48838
2195
2196         * MiniBrowser/mac/WebBundle/WebBundleMain.m:
2197         (WKBundleInitialize):
2198         * TestWebKitAPI/InjectedBundleController.cpp:
2199         (TestWebKitAPI::InjectedBundleController::initialize):
2200         (TestWebKitAPI::InjectedBundleController::didReceiveMessage):
2201         * TestWebKitAPI/InjectedBundleController.h:
2202         * TestWebKitAPI/InjectedBundleMain.cpp:
2203         (WKBundleInitialize):
2204         * TestWebKitAPI/PlatformUtilities.cpp:
2205         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
2206         * WebKitTestRunner/InjectedBundle/InjectedBundleMain.cpp:
2207         (WKBundleInitialize):
2208         Update implementations of WKBundleInitialize to take an initial userData
2209         argument.  Change TestWebKitAPI to use the new initial userData to initialize
2210         each test's bundle.
2211
2212 2010-11-02  Benjamin Kalman  <kalman@google.com>
2213
2214         Reviewed by Ojan Vafai.
2215
2216         new-run-webkit-tests doesn't strip "LayoutTests/" from prefix, unlike old-run-webkit-tests
2217         https://bugs.webkit.org/show_bug.cgi?id=48794
2218
2219         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: Strip the "LayoutTests/" prefix from test argument paths.
2220
2221 2010-11-02  Adam Roben  <aroben@apple.com>
2222
2223         Skip webkitpy.layout_tests.run_webkit_tests_unittest.MainTest on Cygwin
2224         Python 2.5.x
2225
2226         It is known to hang on that version of Python. See
2227         <http://webkit.org/b/48614>.
2228
2229         Reviewed by Adam Barth.
2230
2231         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py: Use
2232         skip_if to skip MainTest on Cygwin Python 2.5.x.
2233
2234         * Scripts/webkitpy/test/skip.py: Added.
2235         (skip_if): If the passed-in condition is false, find all the test_*
2236         methods of the given class and replace them with a function that just
2237         logs that we're skipping these tests. This is loosely based on the
2238         unittest.skip_if decorator added in Python 3.1.
2239         (_skipped_method): Returns a function that just logs that the tests are
2240         being skipped. This is used in place of the actual test_* functions.
2241
2242         * Scripts/webkitpy/test/skip_unittest.py: Added.
2243         (SkipTest.setUp): Create a logger for ourselves and give it a handler
2244         that logs to a private stream.
2245         (SkipTest.tearDown): Reset the logger.
2246         (SkipTest.create_fixture_class): Returns a class that we use to test
2247         skip_if. It has a single test_* method, test_foo, that just calls
2248         through to a callback.
2249         (SkipTest.foo_callback): Record that test_foo was called.
2250         (SkipTest.test_skip_if_false): Pass skip_if a False condition and test
2251         that test_foo does get called.
2252         (SkipTest.test_skip_if_true): Pass skip_if a True condition and test
2253         that test_foo does not get called and the appropriate message gets
2254         logged.
2255
2256 2010-11-02  Adam Barth  <abarth@webkit.org>
2257
2258         Reviewed by Eric Seidel.
2259
2260         webkit-patch should tell check-webkit-style which files were changed so
2261         check-webkit-style doesn't have to stat the whole working copy again
2262         https://bugs.webkit.org/show_bug.cgi?id=48792
2263
2264         * Scripts/webkitpy/tool/mocktool.py:
2265         * Scripts/webkitpy/tool/steps/checkstyle.py:
2266
2267 2010-11-02  Robert Kroeger  <rjkroege@chromium.org>
2268
2269         Reviewed by James Robinson.
2270
2271         [Chromium/DRT] Make EventSendingController honour leapForward for touch events.
2272         https://bugs.webkit.org/show_bug.cgi?id=48777
2273
2274         * DumpRenderTree/chromium/EventSender.cpp:
2275         (EventSender::sendCurrentTouchEvent):
2276
2277 2010-11-02  Adam Roben  <aroben@apple.com>
2278
2279         Only track resource identifiers in DRT when dumpResourceLoadCallbacks
2280         is on
2281
2282         This reverts Windows to our pre-r71097 behavior. That patch made us
2283         track all resource identifiers, including the main resource, so the
2284         main resource's URL started appearing in test output instead of
2285         "<unknown>". Arguably having the main resource's URL is better, but all
2286         other platforms print "<unknown>" and we want to match.
2287
2288         Fixes <http://webkit.org/b/48837> <rdar://problem/8620351> REGRESSION
2289         (r71097): Many http tests failing on Windows
2290
2291         Reviewed by Anders Carlsson.
2292
2293         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
2294         (ResourceLoadDelegate::identifierForInitialRequest): Don't add the
2295         identifier to the URL map if we're not supposed to dump resource load
2296         callbacks.
2297         (ResourceLoadDelegate::removeIdentifierForRequest): Always remove the
2298         identifier from the URL map even if we're already "done". There's no
2299         point in keeping out-of-date identifiers around.
2300
2301 2010-11-01  Jenn Braithwaite  <jennb@chromium.org>
2302
2303         Reviewed by Adam Roben.
2304
2305         Windows: Update resource tracking when moving a frame between documents
2306         https://bugs.webkit.org/show_bug.cgi?id=48364
2307
2308         * DumpRenderTree/win/DumpRenderTree.cpp:
2309         (createWebViewAndOffscreenWindow):
2310         (main):
2311         Give each WebView its own ResourceLoadDelegate instance in order to
2312         make assertions about resource ids on a particular WebView.
2313         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
2314         (ResourceLoadDelegate::identifierForInitialRequest):
2315         Always add id to the map.
2316         (ResourceLoadDelegate::removeIdentifierForRequest):
2317         Added.
2318         (ResourceLoadDelegate::willSendRequest):
2319         (ResourceLoadDelegate::didReceiveAuthenticationChallenge):
2320         (ResourceLoadDelegate::didReceiveResponse):
2321         (ResourceLoadDelegate::didFinishLoadingFromDataSource):
2322         (ResourceLoadDelegate::didFailLoadingWithError):
2323         (ResourceLoadDelegate::descriptionSuitableForTestResult):
2324         Replace static descriptionSuitableForTestResult with static member function to access identifier map.
2325         * DumpRenderTree/win/ResourceLoadDelegate.h:
2326         (ResourceLoadDelegate::urlMap):
2327         Moved within class so that each WebView has its own id map.
2328
2329 2010-11-01  Ojan Vafai  <ojan@chromium.org>
2330
2331         Reviewed by Dimitri Glazkov.
2332
2333         remove debug code from run_webkit_tests.py
2334         https://bugs.webkit.org/show_bug.cgi?id=48800
2335
2336         Remove temporary debug code and make --master-name required
2337         if --test-results-server is set now that all clients set
2338         --master-name.
2339
2340         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2341
2342 2010-11-01  Tony Chang  <tony@chromium.org>
2343
2344         Reviewed by Kent Tamura.
2345
2346         add plugin event logging for linux
2347         https://bugs.webkit.org/show_bug.cgi?id=48779
2348
2349         This is taken from Chromium's fork of the layout test plugin:
2350         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
2351
2352         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
2353         (handleEventX11):
2354         (NPP_HandleEvent):
2355
2356 2010-11-01  Dirk Pranke  <dpranke@chromium.org>
2357
2358         Reviewed by James Robinson.
2359
2360         new-run-webkit-tests: use DRT, child-processes=1 for GPU tests by default
2361         https://bugs.webkit.org/show_bug.cgi?id=48790
2362
2363         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2364         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2365
2366 2010-11-01  Adam Barth  <abarth@webkit.org>
2367
2368         Reviewed by Eric Seidel.
2369
2370         Teach check-webkit-style how to accept a list of files to diff on the
2371         command line
2372         https://bugs.webkit.org/show_bug.cgi?id=48784
2373
2374         In a future patch, webkit-patch will use this option to improve
2375         performance.  I'm landing this in two pieces to avoid causing a version
2376         skew problem for the style-bot.
2377
2378         * Scripts/check-webkit-style:
2379         * Scripts/webkitpy/style/optparser.py:
2380         * Scripts/webkitpy/style/optparser_unittest.py:
2381         * Scripts/webkitpy/style_references.py:
2382
2383 2010-11-01  Anders Carlsson  <andersca@apple.com>
2384
2385         Reviewed by John Sullivan.
2386
2387         Tear down the related WebProcessProxy when a WebContext is deallocated
2388         https://bugs.webkit.org/show_bug.cgi?id=48769
2389
2390         * TestWebKitAPI/Tests/WebKit2/FailedLoad.cpp:
2391         (TestWebKitAPI::didFailProvisionalLoadWithErrorForFrame):
2392         We don't support empty URLs anymore, update test to expect a null URL instead.
2393
2394 2010-11-01  Søren Gjesse  <sgjesse@chromium.org>
2395
2396         Reviewed by Andreas Kling.
2397
2398         Fix warning when compiling the chromium port of DumpRenderShell
2399         with clang.
2400         https://bugs.webkit.org/show_bug.cgi?id=48414
2401
2402         * DumpRenderTree/chromium/TestShell.h:
2403         (TestShell::javaScriptFlagsForLoad):
2404
2405 2010-11-01  Adam Roben  <aroben@apple.com>
2406
2407         Fix typo from r71022
2408
2409         * BuildSlaveSupport/build.webkit.org-config/config.json:
2410
2411 2010-11-01  Adam Roben  <aroben@apple.com>
2412
2413         Trigger the Windows Release WebKit2 tests when a Release build
2414         finishes, not when a Debug build finishes
2415
2416         Fixes <http://webkit.org/b/48754> Windows Release WebKit2 tests are
2417         triggered at the wrong time
2418
2419         Reviewed by Sam Weinig.
2420
2421         * BuildSlaveSupport/build.webkit.org-config/config.json: Fixed
2422         triggerable name and trigger.
2423
2424 2010-11-01  Mario Sanchez Prada  <msanchez@igalia.com>
2425
2426         Unreviewed. Adding my IRC nickname to the list of committers.
2427
2428         * Scripts/webkitpy/common/config/committers.py:
2429
2430 2010-10-31  Robert Hogan  <robert@webkit.org>
2431
2432         Reviewed by Antonio Gomes.
2433
2434         [Qt] [Gtk] Plug-ins having upper case in mime type are failing to load
2435
2436         Qt and Gtk are case-sensitive when storing the declared mime-type
2437         of plugins. Since plugin mime-types are lowercased prior to searching
2438         for them in the plugin database, ensure they are loaded with the
2439         mime-type in lower case too.
2440
2441         Change the test netscape plugin to declare its mimetype in sentence
2442         case so that the correct behaviour is enforced.
2443
2444         https://bugs.webkit.org/show_bug.cgi?id=36815
2445
2446         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
2447         (NP_GetMIMEDescription):
2448
2449 2010-10-31  Robert Hogan  <robert@webkit.org>
2450
2451         Reviewed by Antonio Gomes.
2452
2453         [Qt] Support nodesFromRect in DRT
2454
2455         https://bugs.webkit.org/show_bug.cgi?id=48716
2456
2457         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2458         (LayoutTestController::nodesFromRect):
2459         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2460
2461 2010-10-30  Andreas Kling  <kling@webkit.org>
2462
2463         Reviewed by Antonio Gomes.
2464
2465         [Qt] QtTestBrowser: Switching view type moves the embedded inspector
2466         https://bugs.webkit.org/show_bug.cgi?id=48705
2467
2468         Reinsert the embedded inspector into the splitter after changing
2469         between QWebView/QGraphicsWebView.
2470
2471         * QtTestBrowser/launcherwindow.cpp:
2472         (LauncherWindow::init):
2473         (LauncherWindow::initializeView):
2474         * QtTestBrowser/webinspector.h:
2475         (WebInspector::WebInspector):
2476
2477 2010-10-28  Antonio Gomes  <agomes@rim.com>
2478
2479         Reviewed by Ojan Vafai.
2480
2481         Needs a "LinuxEditingBehavior", perhaps with a better name
2482         https://bugs.webkit.org/show_bug.cgi?id=36627
2483
2484         Adding support to Mac's, GTK+'s, Windows' and Chromium's LayoutTestController class to test the newly introduced Unix editing behavior.
2485
2486         * DumpRenderTree/chromium/LayoutTestController.cpp:
2487         (LayoutTestController::setEditingBehavior):
2488         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2489         (LayoutTestController::setEditingBehavior):
2490         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2491         (LayoutTestController::setEditingBehavior):
2492         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2493         (LayoutTestController::setEditingBehavior):
2494
2495 2010-10-29  Andreas Kling  <kling@webkit.org>
2496
2497         Reviewed by Kenneth Rohde Christiansen.
2498
2499         [Qt] QtTestBrowser: Add keyboard shortcut to toggle full screen (F11)
2500         https://bugs.webkit.org/show_bug.cgi?id=48695
2501
2502         * QtTestBrowser/launcherwindow.cpp:
2503         (LauncherWindow::createChrome):
2504
2505 2010-10-29  Andreas Kling  <kling@webkit.org>
2506
2507         Reviewed by Kenneth Rohde Christiansen.
2508
2509         [Qt] QtTestBrowser: Fix uninitialized read in FpsTimer
2510         https://bugs.webkit.org/show_bug.cgi?id=48675
2511
2512         FpsTimer::m_timer was never initialized and passed to QObject::killTimer()
2513         on startup with in -graphicsbased mode.
2514
2515         * QtTestBrowser/fpstimer.cpp:
2516         (FpsTimer::FpsTimer):
2517         (FpsTimer::numFrames):
2518         (FpsTimer::stop):
2519         (FpsTimer::timerEvent):
2520         * QtTestBrowser/fpstimer.h:
2521
2522 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
2523
2524         Reviewed by Tony Chang.
2525
2526         new-run-webkit-tests: change TestResults to be serializable
2527
2528         In preparation for changing new-run-webkit-tests from
2529         multithreaded to multiprocess, we need to make sure the data
2530         going between the threads is easily serialized over a socket.
2531
2532         This change adds serialization/pickling for the TestResults and
2533         TestFailure objects (using cPickle).
2534
2535         The TestFailure objects included a "has_wdiff" flag for Text
2536         results, but the flag wasn't being used, so I've removed it,
2537         simplifying the state to basically a set of enum objects with
2538         associated methods.
2539
2540         https://bugs.webkit.org/show_bug.cgi?id=48616
2541
2542         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2543         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
2544         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
2545         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
2546         * Scripts/webkitpy/layout_tests/layout_package/test_results.py: Added.
2547         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:Added. 
2548         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2549         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2550
2551 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
2552
2553         Reviewed by Ojan Vafai.
2554
2555         new-run-webkit-tests: need to provide separate dashboard results for GPU tests
2556         https://bugs.webkit.org/show_bug.cgi?id=48687
2557
2558         Modify the steps to upload the JSON files to the dashboards so
2559         that the GPU tests don't conflict with the regular tests on a
2560         port. We do this by modifying the --builder-name parameter to
2561         append " - GPU", which should cause the app to treat the results
2562         as a completely new builder. This is a little non-obvious, but
2563         keeps us from having to restructure the app.
2564
2565         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2566         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2567         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2568
2569 2010-10-29  Ojan Vafai  <ojan@chromium.org>
2570
2571         Reviewed by Tony Chang.
2572
2573         [chromium] add debug logging to help diagnose flakiness dashboard issues
2574         https://bugs.webkit.org/show_bug.cgi?id=48657
2575
2576         The appengine app thinks it's getting empty files uploaded. Add some logging
2577         to see if new-run-webkit-tests agrees.
2578         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2579
2580 2010-10-29  Dimitri Glazkov  <dglazkov@chromium.org>
2581
2582         Reviewed by Ojan Vafai.
2583
2584         [Chromium/DRT] Add master-name flag to new-run-webkit-tests.
2585         https://bugs.webkit.org/show_bug.cgi?id=48649
2586
2587         * BuildSlaveSupport/build.webkit.org-config/master.cfg: Added flag.
2588
2589 2010-10-29  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
2590
2591         Reviewed by Martin Robinson.
2592
2593         Enable popup window in GtkLauncher
2594         https://bugs.webkit.org/show_bug.cgi?id=48335
2595
2596         The GtkLauncher application does not support opening new window when
2597         a link with "target=_blank" is clicked or similar call to
2598         window.open(). Instead, GtkLauncher does nothing which breaks
2599         navigation of some websites.
2600
2601         * GtkLauncher/main.c:
2602         (activate_uri_entry_cb):
2603         (update_title):
2604         (link_hover_cb):
2605         (notify_title_cb):
2606         (notify_load_status_cb):
2607         (notify_progress_cb):
2608         (destroy_cb):
2609         (go_back_cb):
2610         (go_forward_cb):
2611         (create_web_view_cb):
2612         (web_view_ready_cb):
2613         (close_web_view_cb):
2614         (create_browser):
2615         (create_statusbar):
2616         (create_toolbar):
2617         (create_window):
2618         (main):
2619
2620 2010-10-29  Adam Roben  <aroben@apple.com>
2621
2622         Teach check-webkit-style about WebKit2's idiosyncracies
2623
2624         Fixes <http://webkit.org/b/48638> Style bot complains about a number
2625         of WebKit2 conventions
2626
2627         Reviewed by Anders Carlsson.
2628
2629         * Scripts/webkitpy/style/checker.py: Excluded some rules for various
2630         WebKit2-related files. Also updated the excluded rules for
2631         WebKitAPITest to match the current code.
2632
2633 2010-10-29  Csaba Osztrogonác  <ossy@webkit.org>
2634
2635         Reviewed by Adam Roben and David Kilzer.
2636
2637         Fix and cleanup of build systems
2638         https://bugs.webkit.org/show_bug.cgi?id=48342
2639
2640         * Scripts/build-webkit:
2641          - Remove unnecessary ENABLE_SANDBOX option.
2642          - Add ENABLE_FULLSCREEN_API option.
2643
2644 2010-10-28  Adam Roben  <aroben@apple.com>
2645
2646         Switch the Windows WebKit2 bot to the Release configuration
2647
2648         We only have one machine testing WebKit2 on Windows right now, and
2649         Debug is just too slow for it to keep up.
2650
2651         Fixes (hopefully!) <http://webkit.org/b/48615> Windows WebKit2 bot is
2652         always way behind
2653
2654         Reviewed by Jon Honeycutt.
2655
2656         * BuildSlaveSupport/build.webkit.org-config/config.json:
2657
2658 2010-10-29  Leandro Gracia Gil  <leandrogracia@google.com>
2659
2660         Reviewed by Jeremy Orlow.
2661
2662         Added a second parameter to setMockSpeechInputResult for
2663         the language used in speech input.
2664         https://bugs.webkit.org/show_bug.cgi?id=47089
2665
2666         * DumpRenderTree/LayoutTestController.cpp:
2667         (setMockSpeechInputResultCallback):
2668         * DumpRenderTree/LayoutTestController.h:
2669         * DumpRenderTree/chromium/LayoutTestController.cpp:
2670         (LayoutTestController::setMockSpeechInputResult):
2671         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2672         (LayoutTestController::setMockSpeechInputResult):
2673         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2674         (LayoutTestController::setMockSpeechInputResult):
2675         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2676         (LayoutTestController::setMockSpeechInputResult):
2677         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2678         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2679         (LayoutTestController::setMockSpeechInputResult):
2680         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2681         (LayoutTestController::setMockSpeechInputResult):
2682         (LayoutTestController::pageSizeAndMarginsInPixels):
2683
2684 2010-10-28  Ojan Vafai  <ojan@chromium.org>
2685
2686         Reviewed by Tony Chang.
2687
2688         [chromium] add a result-small.json file for the test dashboard
2689         https://bugs.webkit.org/show_bug.cgi?id=48547
2690
2691         Output both a results.json file and a results-small.json file.
2692         The dashboard will load results-small.json by default so it loads faster.
2693
2694         * TestResultServer/model/jsonresults.py:
2695         * TestResultServer/model/jsonresults_unittest.py:
2696         Added a bunch of sys.path hackery. Unforunately, this uses hardcoded
2697         paths. That obviously needs to be fixed, but at least this way it
2698         clearly documents what paths are necessary.
2699
2700 2010-10-28  Eric Seidel  <eric@webkit.org>
2701
2702         Reviewed by Adam Barth.
2703
2704         webkit-patch upload calls changed_files more often than it should
2705         https://bugs.webkit.org/show_bug.cgi?id=48567
2706
2707         Passing changed_files around everywhere isn't a very elegant solution
2708         but it's the one we have for the moment.  I think keeping an explicit
2709         cache on Checkout (or making StepState() a real class) is a better
2710         long-term option.
2711
2712         Previously bug_id_for_this_commit was calling changed_files and the
2713         result was never getting cached on the state.  Now we're explicitly
2714         caching the result on the state and passing that to the bug_id_for_this_commit call.
2715
2716         I looked into building unit tests for this.  Doing so would require
2717         using a real Checkout object with a MockSCM and overriding the appropriate
2718         calls on SCM to count how often we're stating the file system.
2719         That's a useful set of tests to build for a separate change.
2720
2721         * Scripts/webkitpy/common/checkout/api.py:
2722         * Scripts/webkitpy/tool/commands/download.py:
2723         * Scripts/webkitpy/tool/commands/upload.py:
2724         * Scripts/webkitpy/tool/mocktool.py:
2725
2726 2010-10-28  Eric Seidel  <eric@webkit.org>
2727
2728         Reviewed by Adam Barth.
2729
2730         Make suggest-reviewers slightly faster
2731         https://bugs.webkit.org/show_bug.cgi?id=48562
2732
2733         Add @memoized to one more common call.
2734
2735         * Scripts/webkitpy/common/checkout/api.py:
2736
2737 2010-10-28  Eric Seidel  <eric@webkit.org>
2738
2739         Reviewed by Adam Barth.
2740
2741         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is still very slow
2742         https://bugs.webkit.org/show_bug.cgi?id=48536
2743
2744         This patch makes it about 40 seconds faster, but it still
2745         takes 1:40.  This will require more refinement.
2746
2747         The suggested reviewers list appears to be the same.
2748
2749         I think the next step may be to have it stop the search after
2750         5 reviewers are found.  We never want to suggest 30 people.
2751
2752         * Scripts/webkitpy/common/checkout/scm.py:
2753          - Using --remove-empty to theoretically stop lookups past
2754            when a file is removed.  I'm not entirely clear that the option
2755            does what it says it does.  Example:
2756            git log --pretty=format:%H -5 --remove-empty -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
2757            returns only one commit
2758            vs.
2759            git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
2760            which returns 5.  I was not aware that wx files were ever removed from the repository?
2761
2762 2010-10-28  Kinuko Yasuda  <kinuko@chromium.org>
2763
2764         Reviewed by David Levin.
2765
2766         [Chromium] Support FileSystem in chromium DRT
2767         https://bugs.webkit.org/show_bug.cgi?id=47643
2768
2769         * DumpRenderTree/chromium/WebViewHost.cpp:
2770         (WebViewHost::openFileSystem): Added.
2771         * DumpRenderTree/chromium/WebViewHost.h:
2772         (WebViewHost::openFileSystem): Added.
2773
2774 2010-10-28  Sam Weinig  <sam@webkit.org>
2775
2776         Reviewed by Anders Carlsson.
2777
2778         WKURLRefs should be allowed to be null
2779         <rdar://problem/8575621>
2780         https://bugs.webkit.org/show_bug.cgi?id=48535
2781
2782         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
2783         * TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp:
2784         (TestWebKitAPI::didStartProvisionalLoadForFrame):
2785         (TestWebKitAPI::didCommitLoadForFrame):
2786         (TestWebKitAPI::didFinishLoadForFrame):
2787         Test that URLs are null pointers when unset.
2788
2789 2010-10-28  Chris Fleizach  <cfleizach@apple.com>
2790
2791         Reviewed by Adele Peterson.
2792
2793         AX: multi select group option does not handle setting of AXSelectedChildren correctly
2794         https://bugs.webkit.org/show_bug.cgi?id=48464
2795
2796         Add support for querying information about selected children to DRT, including:
2797            selectedChildAtIndex
2798            selectedChildrenCount
2799            setSelectedChild
2800
2801         * DumpRenderTree/AccessibilityUIElement.cpp:
2802         (selectedChildAtIndexCallback):
2803         (setSelectedChildCallback):
2804         (selectedChildrenCountCallback):
2805         (AccessibilityUIElement::setSelectedChild):
2806         (AccessibilityUIElement::selectedChildrenCount):
2807         (AccessibilityUIElement::selectedChildAtIndex):
2808         (AccessibilityUIElement::getJSClass):
2809         * DumpRenderTree/AccessibilityUIElement.h:
2810         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
2811         (AccessibilityUIElement::selectedChildAtIndex):
2812         (AccessibilityUIElement::selectedChildrenCount):
2813         (AccessibilityUIElement::setSelectedChild):
2814
2815 2010-10-27  Sam Weinig  <sam@webkit.org>
2816
2817         Reviewed by Anders Carlsson.
2818
2819         Add WebKit2 API for window feature getter/setters
2820         <rdar://problem/8590373>
2821         https://bugs.webkit.org/show_bug.cgi?id=48496
2822
2823         * MiniBrowser/mac/BrowserWindowController.m:
2824         (-[BrowserWindowController awakeFromNib]):
2825         * MiniBrowser/win/BrowserView.cpp:
2826         (BrowserView::create):
2827         * WebKitTestRunner/TestController.cpp:
2828         (WTR::createOtherPage):
2829         (WTR::TestController::initialize):
2830
2831 2010-10-28  Søren Gjesse  <sgjesse@chromium.org>
2832
2833         Reviewed by Tony Chang.
2834
2835         Added support for the DumpRenderTree flags --multiple-loads and --js-flags to the Python test runner.
2836         https://bugs.webkit.org/show_bug.cgi?id=48236
2837
2838         * Scripts/webkitpy/layout_tests/port/chromium.py:
2839         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2840
2841 2010-10-28  Sergio Villar Senin  <svillar@igalia.com>
2842
2843         Reviewed by Martin Robinson.
2844
2845         [GTK] WebKitWebFrame's load-status is not properly notified to the tests
2846         https://bugs.webkit.org/show_bug.cgi?id=48048
2847
2848         DRT now listens to WebKitWebFrame load-status signals for each
2849         created frame instead of the load-status signal of the
2850         WebKitWebView that only notifies about changes in the main frame.
2851
2852         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2853         (webViewLoadFinished):
2854         (webFrameLoadStatusNotified):
2855         (frameCreatedCallback):
2856         (createWebView):
2857         (main):
2858
2859 2010-10-28  Tony Chang  <tony@chromium.org>
2860
2861         Reviewed by Kent Tamura.
2862
2863         [chromium] fix textInputController.{selectedRange,markedRange}
2864         https://bugs.webkit.org/show_bug.cgi?id=48487
2865
2866         * DumpRenderTree/chromium/TextInputController.cpp:
2867         (TextInputController::markedRange): Return arrays of ints, rather than a string
2868         (TextInputController::selectedRange): Ditto.
2869
2870 2010-10-27  Ojan Vafai  <ojan@chromium.org>
2871
2872         Reviewed by Tony Chang.
2873
2874         [chromium] add a master-name flag to new-run-webkit-tests
2875         https://bugs.webkit.org/show_bug.cgi?id=48488
2876
2877         The test results server now allows adding a master name to
2878         the uploaded files. This lets us distinguish bots that have
2879         the same name, but are on different masters.
2880
2881         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2882
2883 2010-10-28  Adam Roben  <aroben@apple.com>
2884
2885         Don't append a newline to the test output if the frame has no document
2886         element in WebKitTestRunner
2887
2888         Fixes <http://webkit.org/b/48526> Extra trailing newline when running
2889         plugins/document-open.html in WebKitTestRunner
2890
2891         Reviewed by Anders Carlsson.
2892
2893         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2894         (WTR::toJS): Added. Turns a UTF-8 C string into a JSStringRef.
2895         (WTR::hasDocumentElement): Added. Uses the JSC API to figure out
2896         whether the frame has a document element.
2897         (WTR::dumpFrameText): Match DRT by bailing (rather than appending an
2898         empty string and a newline) if the frame has no document element.
2899
2900 2010-10-28  Adam Roben  <aroben@apple.com>
2901
2902         Skip npn-invalidate-rect-invalidates-window.html on headless XP
2903         machines
2904
2905         TestNetscapePlugIn never receives a WM_PAINT message on headless XP
2906         machines, so this test times out. Fixing the test is covered by
2907         <http://webkit.org/b/48333>.
2908
2909         * Scripts/old-run-webkit-tests: Skip
2910         npn-invalidate-rect-invalidates-window.html on Windows if accelerated
2911         compositing support is disabled, which likely means we're on a headless
2912         XP machine.
2913
2914 2010-10-28  Kimmo Kinnunen  <kimmok@iki.fi>
2915
2916         Adding myself as a committer.
2917
2918         * Scripts/webkitpy/common/config/committers.py:
2919
2920 2010-10-27  Eric Seidel  <eric@webkit.org>
2921
2922         Reviewed by Adam Barth.
2923
2924         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is very slow
2925         https://bugs.webkit.org/show_bug.cgi?id=48500
2926
2927         This doesn't fix the problem, but it makes things slightly better.
2928         Each git svn find-rev call takes about .25 seconds on my desktop.
2929         This patch uses a new memoized class to avoid those calls when possible.
2930
2931         The real slowness is still git log on some files, like:
2932         git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
2933         I'm not yet sure how to make the pathological git logs better.
2934
2935         * Scripts/webkitpy/common/checkout/scm.py:
2936         * Scripts/webkitpy/common/memoized.py: Added.
2937         * Scripts/webkitpy/common/memoized_unittest.py: Added.
2938
2939 2010-10-27  Eric Seidel  <eric@webkit.org>
2940
2941         Reviewed by Ojan Vafai.
2942
2943         EWS bots should not use --quiet when running build-webkit
2944         https://bugs.webkit.org/show_bug.cgi?id=48482
2945
2946         --quiet is only correct when the sub-process does the error reporting.
2947         In the case of _can_build() the parent process is reporting the error.
2948         We'd like the full build log at queues.webkit.org so someone can look
2949         at the log and understand why the EWS is failing to build trunk.
2950
2951         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2952         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
2953         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
2954         * Scripts/webkitpy/tool/mocktool.py:
2955
2956 2010-10-27  Brian Weinstein  <bweinstein@apple.com>
2957
2958         More Windows build fixage. Rename a variable that was named string.
2959
2960         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2961         (WTR::hasPrefix):
2962
2963 2010-10-27  Ojan Vafai  <ojan@chromium.org>
2964
2965         Reviewed by Dimitri Glazkov.
2966
2967         [chromium] Make the test results server store which master the bot is on
2968         https://bugs.webkit.org/show_bug.cgi?id=48478
2969
2970         The chromium bots recently changed so that there are multiple slaves with
2971         the same name on different masters. Up till now, the test results server
2972         assumed slave names were unique. Adds a master field to the file in order
2973         to distinguish.
2974
2975         Also, for files that currently lack a master or testtype, set them appropriately.
2976
2977         * TestResultServer/handlers/testfilehandler.py:
2978         * TestResultServer/index.yaml:
2979         * TestResultServer/model/jsonresults.py:
2980         * TestResultServer/model/testfile.py:
2981         * TestResultServer/templates/showfilelist.html:
2982         * TestResultServer/templates/uploadform.html:
2983
2984 2010-10-26  Darin Adler  <darin@apple.com>
2985
2986         Reviewed by Sam Weinig.
2987
2988         WebKitTestRunner needs to support layoutTestController.dumpBackForwardList
2989         https://bugs.webkit.org/show_bug.cgi?id=42322
2990         rdar://problem/8193631
2991
2992         WebKitTestRunner needs to support layoutTestController.clearBackForwardList
2993         https://bugs.webkit.org/show_bug.cgi?id=42333
2994         rdar://problem/8193643
2995
2996         * WebKitTestRunner/Configurations/InjectedBundle.xcconfig:
2997         Renamed the product to WebKitTestRunnerInjectedBundle to avoid
2998         name conflicts in the build directory.
2999
3000         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
3001         Added dumpBackForwardList and clearBackForwardList.
3002
3003         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
3004         (WTR::InjectedBundle::InjectedBundle): Removed initialization of m_mainPage.
3005         (WTR::InjectedBundle::didCreatePage): Removed unneeded underscore.
3006         (WTR::InjectedBundle::willDestroyPage): Ditto.
3007         (WTR::InjectedBundle::didReceiveMessage): Ditto.
3008         (WTR::InjectedBundle::initialize): Ditto.
3009         (WTR::InjectedBundle::didCreatePage): Changed code to use m_pages
3010         instead of m_mainPage and m_otherPages.
3011         (WTR::InjectedBundle::willDestroyPage): Ditto.
3012         (WTR::InjectedBundle::page): Ditto.
3013         (WTR::InjectedBundle::beginTesting): Ditto.
3014         (WTR::InjectedBundle::done): Ditto.
3015         (WTR::InjectedBundle::closeOtherPages): Ditto.
3016         (WTR::InjectedBundle::dumpBackForwardListsForAllPages): Added.
3017
3018         * WebKitTestRunner/InjectedBundle/InjectedBundle.h: Changed
3019         page function to longer be inline, and pageCount function
3020         to use m_pages. Added dumpBackForwardListsForAllPages, and
3021         removed some underscores. Replaced m_mainPage and m_otherPageas
3022         with m_pages.
3023
3024         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3025         (WTR::adoptWK): Added.
3026         (WTR::hasPrefix): Added.
3027         (WTR::InjectedBundlePage::reset): Added code to set up
3028         m_previousTestBackForwardListItem.
3029         (WTR::InjectedBundlePage::dump): Added code to call
3030         dumpBackForwardListsForAllPages.
3031         (WTR::compareByTargetName): Added.
3032         (WTR::dumpBackForwardListItem): Added.
3033         (WTR::InjectedBundlePage::dumpBackForwardList): Added.
3034
3035         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h: Added
3036         dumpBackForwardList and m_previousTestBackForwardListItem.
3037
3038         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
3039         (WTR::LayoutTestController::LayoutTestController): Initialize
3040         m_shouldDumpBackForwardListsForAllWindows to false.
3041         (WTR::LayoutTestController::clearBackForwardList): Added.
3042
3043         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
3044         Added dumpBackForwardList, clearBackForwardList,
3045         shouldDumpBackForwardListsForAllWindows, and
3046         m_shouldDumpBackForwardListsForAllWindows.
3047
3048         * WebKitTestRunner/TestController.cpp:
3049         (WTR::TestController::initialize): Set up
3050         didReceiveSynchronousMessageFromInjectedBundle.
3051         (WTR::TestController::didReceiveSynchronousMessageFromInjectedBundle):
3052         Added.
3053         * WebKitTestRunner/TestController.h: Ditto.
3054
3055         * WebKitTestRunner/TestInvocation.cpp:
3056         (WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):
3057         Added.
3058         * WebKitTestRunner/TestInvocation.h: Ditto.
3059
3060         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
3061         Renamed bundle to WebKitTestRunnerInjectedBundle (see above).
3062         * WebKitTestRunner/mac/TestControllerMac.mm:
3063         (WTR::TestController::initializeInjectedBundlePath): Ditto.
3064
3065 2010-10-27  Chris Rogers  <crogers@google.com>
3066
3067         Reviewed by Chris Marrin.
3068
3069         Add ENABLE_WEB_AUDIO feature enable flag (initially disabled) to build-webkit
3070         https://bugs.webkit.org/show_bug.cgi?id=48279
3071
3072         * Scripts/build-webkit:
3073
3074 2010-10-27  Eric Seidel  <eric@webkit.org>
3075
3076         Unreviewed.
3077
3078         build-webkit should collect Visual Studio Express logs and display them
3079         https://bugs.webkit.org/show_bug.cgi?id=39199
3080
3081         It turns out my previous patch wasn't actually working on the win-ews
3082         machine.  So I've fixed my mistakes from before.
3083
3084         * Scripts/build-webkit:
3085          - Windows VSE builds change the CWD while building.  Why?  Who knows.
3086         * Scripts/print-vse-failure-logs:
3087          - windows VSE builds don't use Debug/Release as I expected, so I've
3088            moved off of --configuration to --top-level and added the /obj
3089            optimization while I was there.
3090
3091 2010-10-27  Dimitri Glazkov  <dglazkov@chromium.org>
3092
3093         Unreviewed, rolling out r70674.
3094         http://trac.webkit.org/changeset/70674
3095         https://bugs.webkit.org/show_bug.cgi?id=48053
3096
3097         Broke Chromium Windows build.
3098
3099         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3100         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3101
3102 2010-10-25  Tony Chang  <tony@chromium.org>
3103
3104         Reviewed by Anders Carlsson.
3105
3106         compile TestNetscapePlugIn on chromium linux
3107         https://bugs.webkit.org/show_bug.cgi?id=48274
3108
3109         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
3110         (NP_Initialize): On Linux, plugin funcs are set in initialize.
3111         (NPP_New): Mark the plugin as windowless.
3112         (NPP_GetValue): Handle mime type values.
3113         (NP_GetMIMEDescription):
3114         (NP_GetValue):
3115         * DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npapi.h: define TRUE and FALSE, which are in webkit's npapi.h.
3116         * DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npfunctions.h: Pull in npapi.h to get TRUE/FALSE (matches webkit's npfunctions.h)
3117
3118 2010-10-27  Kenneth Rohde Christiansen  <kenneth@webkit.org>
3119
3120         Reviewed by Anders Carlsson.
3121
3122         Remove contentSizeChanged callbacks as it is no longer
3123         part of the public UIClient.
3124
3125         Make WKPageContentsSizeChangedCallback be a private API
3126         https://bugs.webkit.org/show_bug.cgi?id=48409
3127
3128         * MiniBrowser/mac/BrowserWindowController.m:
3129         (-[BrowserWindowController awakeFromNib]):
3130         * MiniBrowser/win/BrowserView.cpp:
3131         (BrowserView::create):
3132         * WebKitTestRunner/TestController.cpp:
3133         (WTR::createOtherPage):
3134         (WTR::TestController::initialize):
3135
3136 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3137
3138         Reviewed by Ojan Vafai.
3139
3140         Make http locking default in NRWT.
3141         https://bugs.webkit.org/show_bug.cgi?id=48053
3142
3143         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3144         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3145
3146 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3147
3148         Reviewed by Tony Chang.
3149
3150         [NRWT] Fix http lock on Windows platform
3151         https://bugs.webkit.org/show_bug.cgi?id=48321
3152
3153         * Scripts/webkitpy/layout_tests/port/http_lock.py:
3154
3155 2010-10-27  Satish Sampath  <satish@chromium.org>
3156
3157         Unreviewed, rolling out r70665.
3158         http://trac.webkit.org/changeset/70665
3159         https://bugs.webkit.org/show_bug.cgi?id=47089
3160
3161         Need to address Alexey's review comments.
3162
3163         * DumpRenderTree/LayoutTestController.cpp:
3164         (setMockSpeechInputResultCallback):
3165         * DumpRenderTree/LayoutTestController.h:
3166         * DumpRenderTree/chromium/LayoutTestController.cpp:
3167         (LayoutTestController::setMockSpeechInputResult):
3168         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3169         (LayoutTestController::setMockSpeechInputResult):
3170         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3171         (LayoutTestController::setMockSpeechInputResult):
3172         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3173         (LayoutTestController::setMockSpeechInputResult):
3174         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3175         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3176         (LayoutTestController::setMockSpeechInputResult):
3177         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3178         (LayoutTestController::setMockSpeechInputResult):
3179
3180 2010-10-27  Leandro Gracia Gil  <leandrogracia@google.com>
3181
3182         Reviewed by Jeremy Orlow.
3183
3184         Added a second parameter to setMockSpeechInputResult for
3185         the language used in speech input.
3186         https://bugs.webkit.org/show_bug.cgi?id=47089
3187
3188         * DumpRenderTree/LayoutTestController.cpp:
3189         (setMockSpeechInputResultCallback):
3190         * DumpRenderTree/LayoutTestController.h:
3191         * DumpRenderTree/chromium/LayoutTestController.cpp:
3192         (LayoutTestController::setMockSpeechInputResult):
3193         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3194         (LayoutTestController::setMockSpeechInputResult):
3195         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3196         (LayoutTestController::setMockSpeechInputResult):
3197         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3198         (LayoutTestController::setMockSpeechInputResult):
3199         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3200         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3201         (LayoutTestController::setMockSpeechInputResult):
3202         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3203         (LayoutTestController::setMockSpeechInputResult):
3204         (LayoutTestController::pageSizeAndMarginsInPixels):
3205
3206 2010-10-27  Adam Roben  <aroben@apple.com>
3207
3208         Reset TestNetscapePlugIn's NPP_GetValue pointer when
3209         NullNPPGetValuePointer finishes running
3210
3211         Fixes <http://webkit.org/b/48435> REGRESSION (r70655): Many plugins
3212         tests are failing on Qt
3213
3214         Reviewed by Anders Carlsson.
3215
3216         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
3217         (PluginTest::NPP_Destroy):
3218         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
3219         Added a do-nothing NPP_Destroy implementation.
3220
3221         * DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp:
3222         (NullNPPGetValuePointer::NullNPPGetValuePointer): Save the original
3223         NPP_GetValue pointer in m_originalNPPGetValuePointer so we can restore
3224         it later.
3225         (NullNPPGetValuePointer::NPP_Destroy): Added. Restores the original
3226         NPP_GetValue pointer so it can be used in other tests.
3227
3228         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
3229         (NPP_Destroy):
3230         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
3231         (webkit_test_plugin_destroy_instance):
3232         Call through to the PluginTest when NPP_Destroy is called.
3233
3234 2010-10-26  Brian Weinstein  <bweinstein@apple.com>
3235
3236         Reviewed by Adam Roben.
3237
3238         WebKit2 shouldn't try to send an empty user agent
3239         https://bugs.webkit.org/show_bug.cgi?id=48397
3240         
3241         Add a test that when we set our custom user agent to an empty string, we don't send
3242         an empty user agent.
3243
3244         * TestWebKitAPI/Tests/WebKit2/PreventEmptyUserAgent.cpp: Added.
3245         (TestWebKitAPI::didRunJavaScript): Make sure that the result of navigator.userAgent isn't empty.
3246         (TestWebKitAPI::TEST): Set our custom user agent to the empty string, and run navigator.userAgent.
3247         
3248         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Add the new file.
3249         * TestWebKitAPI/win/TestWebKitAPI.vcproj: Ditto.
3250
3251 2010-10-27  Adam Roben  <aroben@apple.com>
3252
3253         Check in file I forgot in r70653
3254
3255         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
3256         (NP_Initialize): Save the NPPluginFuncs struct the browser passed to us
3257         so we can be naughty and modify it later.
3258
3259 2010-10-27  Adam Roben  <aroben@apple.com>
3260
3261         Test that WebKit doesn't crash if the plugin passes 0 for its
3262         NPP_GetValue pointer
3263
3264         Test for <http://webkit.org/b/48433> Crash in
3265         NetscapePlugin::shouldLoadSrcURL when using Shockwave Director 10.3 in
3266         WebKit2 on Windows
3267
3268         Reviewed by Eric Carlson.
3269
3270         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3271         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
3272         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
3273         * GNUmakefile.am:
3274         Added NullNPPGetValuePointer.cpp.
3275
3276         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
3277         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
3278         Set up a place to store the NPPluginFuncs struct the browser passed to us.
3279
3280         * DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp: Added.
3281         (NullNPPGetValuePointer::NullNPPGetValuePointer): Null out the
3282         NPP_GetValue pointer we passed to the browser to simulate a plugin that
3283         doesn't implement NPP_GetValue.
3284         (NullNPPGetValuePointer::NPP_GetValue): Print an error message. If this
3285         function is called, it means that WebKit has changed in a way that
3286         makes this test invalid.
3287
3288         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
3289         (NP_GetEntryPoints): Save the NPPluginFuncs struct the browser passed
3290         to us so we can be naughty and modify it later.
3291
3292 2010-10-27  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
3293
3294         Reviewed by Andreas Kling.
3295
3296         Remove references to ancient QGVLauncher and QtLauncher
3297         https://bugs.webkit.org/show_bug.cgi?id=48430
3298
3299         QtTestBrowser substitutes both and has checks enabled.
3300
3301         * Scripts/webkitpy/style/checker.py: remove references
3302         * Scripts/webkitpy/style/checker_unittest.py: ditto
3303
3304 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3305
3306         Reviewed by Ojan Vafai.
3307
3308         [NRWT] Don't use image hash when it's no need in single test mode.
3309         https://bugs.webkit.org/show_bug.cgi?id=48326
3310
3311         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3312
3313 2010-10-27  David Kilzer  <ddkilzer@apple.com>
3314
3315         Fix leak of CFMutableDictionaryRef in createXMLStringFromWebArchiveData()
3316
3317         Reviewed by Adam Roben.
3318
3319         Follow-up fix for: <https://bugs.webkit.org/show_bug.cgi?id=48278>
3320
3321         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
3322         (createXMLStringFromWebArchiveData): Use RetainPtr<> to fix a
3323         leak introduced in r70613.
3324
3325 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3326
3327         Reviewed by Csaba Osztrogonác.
3328
3329         [NRWT] Add platform specific baseline search paths for Qt port
3330         https://bugs.webkit.org/show_bug.cgi?id=48428
3331
3332         * Scripts/webkitpy/layout_tests/port/qt.py:
3333
3334 2010-10-27  Adam Roben  <aroben@apple.com>
3335
3336         Catch exceptions when checking if we're inside a git working directory
3337
3338         Fixes <http://webkit.org/b/48420> REGRESSION (r70562): test-webkitpy
3339         fails on systems without git installed
3340
3341         Reviewed by Anders Carlsson.
3342
3343         * Scripts/webkitpy/common/net/credentials.py:
3344         (Credentials._credentials_from_git): Put the call to
3345         Git.in_working_directory inside the try/except since it, too, attempts
3346         to execute git and thus will throw on systems that don't have git
3347         installed.
3348
3349 2010-10-27  Nikolas Zimmermann  <nzimmermann@rim.com>
3350
3351         Reviewed by David Kilzer.
3352
3353         Convert DumpRenderTree webarchive code to CoreFoundation
3354         https://bugs.webkit.org/show_bug.cgi?id=48278
3355
3356         CFPropertyListCreateWithData and CFPropertyListCreateData are only available in 10.6+.
3357         Replace CFPropertyListCreateWithData by a combination of CFReadStreamCreateWithBytesNoCopy and CFPropertyListCreateFromStream.
3358         Replace CFPropertyListCreateData by CFPropertyListCreateXMLData.
3359
3360         These changes are wrapped in BUILDING_ON_LEOPARD, as the methods are deprecated on 10.6+.
3361
3362         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
3363         (createXMLStringFromWebArchiveData):
3364
3365 2010-10-26  Antonio Gomes  <agomes@rim.com>
3366
3367         Reviewed by Martin Robinson.
3368
3369         Remove the absolute path used to include DumpRenderTreeSupportGtk.h from LayoutTestController.h
3370
3371         It turns out this is rather unneeded since WebKitTools/GNUMakefile.am has WebKit/gtk/ in its include
3372         path, and then we can just do #include "WebCoreSupport/DumpRenderTreeSupportGtk.h"
3373
3374         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3375
3376 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3377
3378         Rename WebArchiveDumpSupport.mm to WebArchiveDumpSupport.cpp
3379
3380         Reviewed by Adam Roben.
3381
3382         Part 4 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3383
3384         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3385         Updated for file moves and renames.
3386         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.mm.
3387         * DumpRenderTree/cf/WebArchiveDumpSupport.h: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.h.
3388
3389 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3390
3391         Convert WebArchiveDumpSupport.mm from NS objects to CF types
3392
3393         Reviewed by Adam Roben.
3394
3395         Part 3 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3396
3397         In order to share WebArchive code between the Mac and Windows
3398         ports, the code in WebArchiveDumpSupport.mm was converted from
3399         Cocoa to C++ using CoreFoundation (CF) types.
3400
3401         * DumpRenderTree/mac/DumpRenderTree.mm:
3402         (dump): Renamed serializeWebArchiveToXML() to
3403         createXMLStringFromWebArchiveData() and added HardAutorelease()
3404         to prevent leaks.
3405         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
3406         (createXMLStringFromWebArchiveData): Renamed from
3407         serializeWebArchiveToXML().  Changed to use CF types.
3408         (createCFURLResponseFromResponseData): Renamed from
3409         unarchiveNSURLResponseFromResponseData().  Changed to use CF
3410         types for its parameter and return type.
3411         * DumpRenderTree/mac/WebArchiveDumpSupport.mm: Replaced use of
3412         NS objects with CF types.  It will be renamed to *.cpp in a
3413         future commit.
3414         (convertMIMEType): Changed to use case-insensitive string
3415         comparisons.
3416         (convertWebResourceDataToString):
3417         (normalizeHTTPResponseHeaderFields):
3418         (normalizeWebResourceURL):
3419         (convertWebResourceResponseToDictionary):
3420         (compareResourceURLs):
3421         (createXMLStringFromWebArchiveData):
3422         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
3423         (createCFURLResponseFromResponseData): Renamed from
3424         unarchiveNSURLResponseFromResponseData().  Changed to use CF
3425         types for its parameter and return type.
3426
3427 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3428
3429         Extract use of NSKeyedUnarchiver from WebArchiveDumpSupport.mm
3430
3431         Reviewed by Adam Roben.
3432
3433         Part 2 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3434
3435         There is no equivalent to NSKeyedUnarchiver in CoreFoundation,
3436         so extract it into a platform-specific source file.
3437
3438         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
3439         (unarchiveNSURLResponseFromResponseData): Added declaration.
3440         * DumpRenderTree/mac/WebArchiveDumpSupport.mm:
3441         (convertWebResourceResponseToDictionary): Extracted code to
3442         unarchiveNSURLResponseFromResponseData() in
3443         WebArchiveDumpSupportMac.mm.  Updated logic to return early if
3444         nil is returned from unarchiveNSURLResponseFromResponseData().
3445         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
3446         (unarchiveNSURLResponseFromResponseData): Added.  Extracted
3447         code from convertWebResourceResponseToDictionary() in
3448         WebArchiveDumpSupport.mm.
3449
3450 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3451
3452         Extract call to -[WebHTMLRepresentation supportedNonImageMIMETypes] from WebArchiveDumpSupport.mm
3453
3454         Reviewed by Adam Roben.
3455
3456         Part 1 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3457
3458         The call to -[WebHTMLRepresentation supportedNonImageMIMETypes]
3459         is not cross-platform between Mac and Windows, so extract it
3460         into a platform-specific source file.
3461
3462         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Added
3463         WebArchiveDumpSupportMac.mm to the project.
3464         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
3465         (supportedNonImageMIMETypes): Added declaration.
3466         * DumpRenderTree/mac/WebArchiveDumpSupport.mm:
3467         (convertWebResourceDataToString): Replaced call to
3468         -[WebHTMLRepresentation supportedNonImageMIMETypes] with
3469         supportedNonImageMIMETypes().
3470         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm: Added.
3471         (supportedNonImageMIMETypes): Added.  Extracted from
3472         WebArchiveDumpSupport.mm.
3473
3474 2010-10-26  Antonio Gomes  <agomes@rim.com>
3475
3476         Reviewed by Martin Robinson.
3477
3478         [GTK] Implement DumpRenderTreeSupportGtk (similarly to DumpRenderTreeSupportQt idea)
3479         https://bugs.webkit.org/show_bug.cgi?id=48199
3480
3481         Implements support to WebKitTabToLinksPreferenceKey through LayoutTestController::overridePreference.
3482         The corresponding DumpRenderTreeSupportGtk method is called in the DRT context only.
3483
3484         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3485         (resetDefaultsToConsistentValues):
3486         (createWebView):
3487         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3488         (LayoutTestController::overridePreference):
3489
3490 2010-10-26  Ariya Hidayat  <ariya@sencha.com>
3491
3492         Unreviewed, change the order of my emails for bugzilla autocompletion.
3493
3494         * Scripts/webkitpy/common/config/committers.py:
3495
3496 2010-10-26  Eric Seidel  <eric@webkit.org>
3497
3498         Reviewed by David Kilzer.
3499
3500         build-webkit should collect Visual Studio Express logs and display them
3501         https://bugs.webkit.org/show_bug.cgi?id=39199
3502
3503         * Scripts/build-webkit:
3504         * Scripts/print-vse-failure-logs: Added.
3505         * Scripts/webkitdirs.pm:
3506
3507 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3508
3509         <http://webkit.org/b/48224> build-webkit: add support for --meter-tag switch
3510
3511         Reviewed by Csaba Osztrogonác.
3512
3513         * Scripts/build-webkit: Added support for --meter-tag switch.
3514         It should have been added with r60820.
3515
3516 2010-10-26  Ademar de Souza Reis Jr.  <ademar.reis@openbossa.org>
3517
3518         Reviewed by Dumitru Daniliuc.
3519
3520         check-webkit-style fails on operator+=, operator-=, ... methods
3521         https://bugs.webkit.org/show_bug.cgi?id=48258
3522
3523         * Scripts/webkitpy/style/checkers/cpp.py: Added exceptions
3524         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added unit-tests
3525
3526 2010-10-26  Ojan Vafai  <ojan@chromium.org>
3527
3528         Reviewed by Tony Chang.
3529
3530         remove DEFER support from new-run-webkit-tests
3531         https://bugs.webkit.org/show_bug.cgi?id=48387
3532
3533         DEFER was needed when we were trying to ship Chrome beta.
3534         Now it's just extra complication.
3535
3536         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
3537         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
3538         * Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
3539         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
3540         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3541         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
3542
3543 2010-10-26  Eric Seidel  <eric@webkit.org>
3544
3545         Unreviewed.  Adding import with_statement to fix python 2.5.
3546
3547         Teach webkit-patch how to read credentials from the environment
3548         https://bugs.webkit.org/show_bug.cgi?id=48275
3549
3550         * Scripts/webkitpy/common/net/credentials_unittest.py:
3551          - import with_statement to fix python 2.5.
3552
3553 2010-10-26  Alexey Proskuryakov  <ap@apple.com>
3554
3555         Reviewed by Sam Weinig.
3556
3557         https://bugs.webkit.org/show_bug.cgi?id=48375
3558         <rdar://problem/8392724> Need delegate calls in PageLoaderClient to indicate if we have
3559         loaded insecure content
3560
3561         Updated for WebKit2 changes.
3562
3563         * MiniBrowser/mac/BrowserWindowController.m:
3564         (didDisplayInsecureContentForFrame):
3565         (didRunInsecureContentForFrame):
3566         (-[BrowserWindowController awakeFromNib]):
3567         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3568         (WTR::InjectedBundlePage::InjectedBundlePage):
3569         (WTR::InjectedBundlePage::didDisplayInsecureContentForFrame):
3570         (WTR::InjectedBundlePage::didRunInsecureContentForFrame):
3571         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
3572         * WebKitTestRunner/TestController.cpp:
3573         (WTR::TestController::initialize):
3574
3575 2010-10-26  Adam Roben  <aroben@apple.com>
3576
3577         Clear up confusion between "3D rendering" and "accelerated compositing"
3578
3579         WebKi