2010-11-17 Hayato Ito <hayato@chromium.org>
[WebKit-https.git] / WebKitTools / ChangeLog
1 2010-11-17  Hayato Ito  <hayato@chromium.org>
2
3         Refactor TestTypeBase.compare_output().
4
5         Introduce a TestOutput class and update compare_output() of each test
6         types so that they can take both actual and expected TestOutput objects.
7
8         https://bugs.webkit.org/show_bug.cgi?id=49431
9
10         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
11         * Scripts/webkitpy/layout_tests/layout_package/test_output.py: Added.
12         * Scripts/webkitpy/layout_tests/port/base.py:
13         * Scripts/webkitpy/layout_tests/port/chromium.py:
14         * Scripts/webkitpy/layout_tests/port/dryrun.py:
15         * Scripts/webkitpy/layout_tests/port/test.py:
16         * Scripts/webkitpy/layout_tests/port/webkit.py:
17         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
18         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
19         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
20         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
21
22 2010-11-17  Adam Roben  <aroben@apple.com>
23
24         Make each Windows Test builder use the same OS for all its slaves
25
26         Currently each WebKit1 Windows Test builder has one Windows XP slave
27         and one Windows 7 slave. Having different OSes in a single builder
28         meant that the builders would alternate between red and green when
29         there were OS-specific failures.
30
31         Fixes <http://webkit.org/b/49688> Windows Test builders alternate
32         between red and green because of multiple OS versions
33
34         Reviewed by Mark Rowe.
35
36         * BuildSlaveSupport/build.webkit.org-config/config.json: Renamed
37         "Windows Release (Tests)" to "Windows 7 Release (Tests)", "Windows
38         Debug (Tests)" to "Windows XP Debug (Tests)", and "Windows Release
39         (WebKit2 Tests)" to "Windows 7 Release (WebKit2 Tests)". Updated the
40         builders for the two WebKit1 builders so that they would all use the
41         appropriate OS.
42
43 2010-11-17  James Robinson  <jamesr@chromium.org>
44
45         Reviewed by Dan Bernstein.
46
47         [chromium] Set minimum font size to 0 for DRT
48         https://bugs.webkit.org/show_bug.cgi?id=49677
49
50         Followup for http://trac.webkit.org/changeset/72141.
51
52         Test: http/tests/misc/acid3.html
53
54         * DumpRenderTree/chromium/WebPreferences.cpp:
55         (WebPreferences::reset):
56
57 2010-11-17  Andras Becsi  <abecsi@webkit.org>
58
59         Unreviewed build fix.
60
61         [Qt] Add missing file I forgot in r72220.
62
63         * MiniBrowser/MiniBrowser.qrc: Added.
64
65 2010-11-17  Andras Becsi  <abecsi@webkit.org>
66
67         Reviewed by Andreas Kling.
68
69         [Qt][WK2] Make it possible to set custom user agent strings in MiniBrowser.
70         https://bugs.webkit.org/show_bug.cgi?id=49627
71
72         Add a user agent dialog and a resource file for useragentlist.txt to MiniBrowser.
73
74         * MiniBrowser/MiniBrowser.qrc: Added.
75         Add it here to prevent qmake from detecting it
76         since it needs to be copied to the build directory.
77         * MiniBrowser/qt/BrowserWindow.cpp:
78         (BrowserWindow::BrowserWindow):
79         (BrowserWindow::updateUserAgentList):
80         (BrowserWindow::showUserAgentDialog):
81         * MiniBrowser/qt/BrowserWindow.h:
82         * MiniBrowser/qt/MiniBrowser.pro:
83
84 2010-11-17  Adam Roben  <aroben@apple.com>
85
86         Make changes to the Mac WebKit2 Skipped file trigger Windows builds
87
88         We use that file on Windows!
89
90         Fixes <http://webkit.org/b/49643> Changes to
91         LayoutTests/platform/mac-wk2 should trigger a Windows build, but don't
92
93         Reviewed by Steve Falkenburg.
94
95         * Scripts/webkitpy/common/config/build.py:
96         (_should_file_trigger_build): Added the mac-wk2 directory and make it
97         trigger builds on SnowLeopard and Windows.
98
99         * Scripts/webkitpy/common/config/build_unittest.py:
100         (ShouldBuildTest.test_should_build): Added a test.
101
102 2010-11-17  Satish Sampath  <satish@chromium.org>
103
104         Reviewed by Jeremy Orlow.
105
106         Clear the speech input mock explicitly before each test.
107         https://bugs.webkit.org/show_bug.cgi?id=49660
108
109         * DumpRenderTree/chromium/LayoutTestController.cpp:
110         (LayoutTestController::setMockSpeechInputResult):
111         * DumpRenderTree/chromium/LayoutTestController.h:
112         * DumpRenderTree/chromium/WebViewHost.cpp:
113         (WebViewHost::speechInputController):
114         (WebViewHost::reset): Invoke speech input mock's clearResults method.
115         * DumpRenderTree/chromium/WebViewHost.h:
116         (WebViewHost::speechInputControllerMock): Moved mock from LayoutTestController.
117
118 2010-11-17  Alexander Pavlov  <apavlov@chromium.org>
119
120         Unreviewed, specify IRC nickname in committers.py.
121
122         * Scripts/webkitpy/common/config/committers.py:
123
124 2010-11-17  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
125
126         Reviewed by Tony Chang.
127
128         [NRWT] Make http locking similar to perl implementation
129         https://bugs.webkit.org/show_bug.cgi?id=49187
130
131         * Scripts/webkitpy/common/system/file_lock.py: Added.
132         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
133         * Scripts/webkitpy/layout_tests/port/http_lock.py:
134
135 2010-11-17  MORITA Hajime  <morrita@google.com>
136
137         Reviewed by Kent Tamura.
138
139         [Chromium][DRT] EventSender.contextClick() should aware spellchecking
140         https://bugs.webkit.org/show_bug.cgi?id=49366
141
142         - EvenSender: Checked WebContextMenuData.misspelledWord and added extra context menu entries
143           according to the spellchecker suggestion
144         - MockSpellCheck: add fillSuggestionList to provide fake suggestions.
145
146         test_expectations.txt will be changed after this change is ported to test_shell.
147         
148         * DumpRenderTree/chromium/EventSender.cpp:
149         (makeMenuItemStringsFor):
150         (EventSender::contextClick):
151         * DumpRenderTree/chromium/MockSpellCheck.cpp:
152         (MockSpellCheck::fillSuggestionList):
153         (MockSpellCheck::initializeIfNeeded):
154         * DumpRenderTree/chromium/MockSpellCheck.h:
155         * DumpRenderTree/chromium/WebViewHost.cpp:
156         (WebViewHost::mockSpellCheck):
157         * DumpRenderTree/chromium/WebViewHost.h:
158
159 2010-11-16  Dave Hyatt  <hyatt@apple.com>
160
161         Reviewed by Sam Weinig.
162
163         Make sure the pref for minimum font size is just explicitly set to 0 every time,
164         since the bots have 1 stuck in their plist otherwise.
165
166         * DumpRenderTree/mac/DumpRenderTree.mm:
167         (resetDefaultsToConsistentValues):
168
169 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
170
171         Reviewed by Tony Chang.
172
173         Rebaseline server: compute diffs client-side
174         https://bugs.webkit.org/show_bug.cgi?id=49640
175         
176         The image diff output from the DRT is pretty bad on some ports (at
177         least the Mac one), so it's better to compute diffs on the client by
178         using <canvas>.
179
180         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
181         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
182         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
183
184 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
185
186         Reviewed by Tony Chang.
187
188         Rebaseline server: display test results
189         https://bugs.webkit.org/show_bug.cgi?id=49626
190         
191         Adds basic result display to the rebaseline server. On the Python side
192         this involves:
193         - Parsing the unexpected_results.json into a dictionary.
194         - Serving it as JSON under /results.json.
195         (the JSON -> dict -> JSON transform isn't strictly necessary right now,
196         but I'll need to have access to the parsed results on the Python side
197         for follow-up changes).
198         
199         On the web UI side this adds:
200         - Markup for display image and text results (expected, actual, diff),
201           and JS for populating it.
202         - Markup for breaking down test results by failure type and directory,
203           and JS for populating it.
204
205         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
206         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
207         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
208         * Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js: Added.
209         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
210
211 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
212
213         Reviewed by Ojan Vafai.
214
215         See Chromium issue http://codereview.chromium.org/5133001/ - we
216         are modifying the build bots to pass in the builder name with
217         the "GPU" string appended instead of appending it in the code.
218
219         https://bugs.webkit.org/show_bug.cgi?id=49636
220
221         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
222
223 2010-11-16  Steve Falkenburg  <sfalken@apple.com>
224
225         Reviewed by Adam Roben.
226
227         Disable LTCG for Windows Release builds. Add new Release_LTCG configuration.
228         https://bugs.webkit.org/show_bug.cgi?id=49632
229
230         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
231         * DumpRenderTree/win/DumpRenderTree.vcproj:
232         * DumpRenderTree/win/ImageDiff.vcproj:
233         * MiniBrowser/MiniBrowser.vcproj:
234         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
235         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
236         * WebKitTestRunner/win/InjectedBundle.vcproj:
237         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
238         * WinLauncher/WinLauncher.vcproj:
239         * record-memory-win/record-memory-win.vcproj:
240
241 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
242
243         Reviewed by Ojan Vafai.
244
245         new-run-webkit-tests: rename TestInfo to TestInput, move image hash to work thread
246
247         Rename the TestInfo class to TestInput to be clearer about its
248         function, and move the checksum-reading code into dump_render_tree_thread
249         to avoid cross-thread access.
250
251         https://bugs.webkit.org/show_bug.cgi?id=49573
252
253         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
254         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
255         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
256
257 2010-11-16  Dave Hyatt  <hyatt@apple.com>
258
259         Fix bustage. Remove the minimum font size pref setting in DumpRenderTree.
260
261         * DumpRenderTree/mac/DumpRenderTree.mm:
262         (resetDefaultsToConsistentValues):
263
264 2010-11-16  Stephanie Lewis  <slewis@apple.com>
265
266         Reviewed by Geoff Garen.
267
268         <rdar://problem/8624267> Leak creating offscreen webview running fast/dom tests
269
270         Use shared DeviceOrientationProviderMock. The old allocation method created a new WebDeviceOrientationProviderMock 
271         for every WebView, and the WebKit API doesn't support that ownership model.
272
273         I also fixed a leak in setMockDeviceOrientation by adding a missing release.
274
275         * DumpRenderTree/mac/DumpRenderTree.mm:
276         (createWebViewAndOffscreenWindow):
277         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
278         (LayoutTestController::setMockDeviceOrientation):
279
280 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
281
282         Reviewed by Ojan Vafai.
283
284         new-run-webkit-tests: log the process id in --verbose mode
285
286         As part of the switch to multi-process mode, we should log the
287         process id when running w/ --verbose.
288
289         https://bugs.webkit.org/show_bug.cgi?id=49571
290
291         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
292
293 2010-11-16  Adam Roben  <aroben@apple.com>
294
295         Only trigger Windows builds when files that we actually use on Windows
296         are changed
297
298         A new Scheduler subclass, PlatformSpecificScheduler, has been added.
299         It uses the new webkitpy.common.config.build module to determine
300         whether a particular change should trigger a build on a particular
301         platform. The Windows builders have been switched to use a
302         PlatformSpecificScheduler.
303
304         The logic to determine whether or not a particular change should
305         trigger a build on a given platform has only been implemented/tested
306         for Windows. I tried to make it easy to add more platforms in the
307         future, but I don't have enough familiarity with all platforms to be
308         able to implement it for them.
309
310         Fixes <http://webkit.org/b/49407> Windows builders kick off builds for
311         lots irrelevant changes (e.g., rebaselining Chromium test results)
312
313         Reviewed by Eric Seidel.
314
315         * BuildSlaveSupport/build.webkit.org-config/config.json: Use a
316         PlatformSpecificScheduler for the Windows builders.
317
318         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
319         (PlatformSpecificScheduler.__init__): Added. Stores our platform, sets
320         up a ChangeFilter that filters to our branch and will call through to
321         our filter method, and calls up to our base class.
322         (PlatformSpecificScheduler.filter): Calls through to
323         build.should_build to find out whether we should trigger a build for
324         this change.
325
326         * Scripts/webkitpy/common/net/build.py: Added.
327         (_should_file_trigger_build): Uses a set of directories and regexp
328         patterns to determine whether the given file should trigger a build on
329         the given platform. As mentioned earlier, this has only been
330         implemented for Windows, though I did try to make some guesses about
331         other platforms.
332         (should_build): Returns true if any of the files should trigger a
333         build on the given platform.
334
335         * Scripts/webkitpy/common/net/build_unittest.py: Added.
336         (ShouldBuildTest.test_should_build): Does some basic testing to make
337         sure we're triggering builds for the right files. It only tests
338         Windows for now, though I tried to make some guesses about other
339         platforms.
340
341 2010-11-16  Simon Hausmann  <simon.hausmann@nokia.com>
342
343         Reviewed by Andreas Kling.
344
345         [Qt] Remove synchronous QWebPage::checkPermissions signal
346         https://bugs.webkit.org/show_bug.cgi?id=46810
347
348         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
349         (WebCore::WebPage::WebPage):
350         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
351         (LayoutTestController::grantDesktopNotificationPermission): When granting
352         permission, grant it directly on the QWebPage/Frame, that will remember it.
353         * QtTestBrowser/webpage.cpp:
354         (WebPage::WebPage):
355         * QtTestBrowser/webpage.h:
356
357
358 2010-11-16  Andras Becsi  <abecsi@inf.u-szeged.hu>
359
360         Reviewed by Csaba Osztrogonác.
361
362         [Qt] Add isPrinting and setPrinting methods to DRT's LayoutTestController.
363
364         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
365         (LayoutTestController::reset):
366         * DumpRenderTree/qt/LayoutTestControllerQt.h:
367         (LayoutTestController::isPrinting):
368         (LayoutTestController::setPrinting):
369
370 2010-11-16  Balazs Kelemen  <kbalazs@webkit.org>
371
372         Reviewed by Andreas Kling.
373
374         [Qt][WK2] Add a way to turn on tiled backing store in MiniBrowser
375         https://bugs.webkit.org/show_bug.cgi?id=49587
376
377         * MiniBrowser/qt/BrowserView.cpp:
378         (BrowserView::BrowserView): Added BackingStoreType
379         argument to ctor. Propagating it to the QGraphicsWKView.
380         * MiniBrowser/qt/BrowserView.h:
381         * MiniBrowser/qt/BrowserWindow.cpp:
382         (BrowserWindow::BrowserWindow):
383         (BrowserWindow::newWindow): Create the new BrowserWindow
384         with the same BackingStoreType that this instance has.
385         * MiniBrowser/qt/BrowserWindow.h: Added BackingStoreType member.
386         * MiniBrowser/qt/main.cpp:
387         (main): Use tiled backing store if got -tiled command line argument.
388         Avoid redundant copy of command line args.
389
390 2010-11-16  John Knottenbelt  <jknotten@chromium.org>
391
392         Reviewed by Jeremy Orlow.
393
394         Move DeviceOrientationClientMock from LayoutTestController to WebViewHost.
395         https://bugs.webkit.org/show_bug.cgi?id=48506
396
397         This ensures that there is one mock per page WebView / WebViewHost. This is a
398         design constraint of the DeviceOrientationClientMock because the m_controller
399         field can only store a pointer to a single instance of the corresponding
400         DeviceOrientationController.
401
402         Test: fast/dom/DeviceOrientation/no-page-cache.html
403
404         * DumpRenderTree/chromium/LayoutTestController.cpp:
405         (LayoutTestController::setMockDeviceOrientation):
406         (LayoutTestController::deviceOrientationClientMock):
407         * DumpRenderTree/chromium/LayoutTestController.h:
408         * DumpRenderTree/chromium/WebViewHost.cpp:
409         (WebViewHost::deviceOrientationClientMock):
410         (WebViewHost::deviceOrientationClient):
411         (WebViewHost::reset):
412         * DumpRenderTree/chromium/WebViewHost.h:
413
414 2010-11-15  Andras Becsi  <abecsi@webkit.org>
415
416         Reviewed by Andreas Kling.
417
418         [Qt][WK2] Avoid polling in WebKitTestRunner.
419         https://bugs.webkit.org/show_bug.cgi?id=49542
420
421         Make the WK2 testing session about 10-15% faster and decrease random flakiness
422         resulting from timing skew by removing the millisecond polling from TestControllerQt.
423
424         * WebKitTestRunner/TestController.cpp:
425         (WTR::TestController::didFinishLoadForFrame):
426         * WebKitTestRunner/TestController.h:
427         * WebKitTestRunner/TestInvocation.cpp:
428         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
429         * WebKitTestRunner/mac/TestControllerMac.mm:
430         (WTR::TestController::notifyDone): Add method stub.
431         * WebKitTestRunner/qt/TestControllerQt.cpp:
432         (WTR::TestControllerRunLoop::instance):
433         (WTR::TestControllerRunLoop::start):
434         (WTR::TestControllerRunLoop::stop):
435         (WTR::TestControllerRunLoop::TestControllerRunLoop):
436         (WTR::TestControllerRunLoop::timerEvent):
437         (WTR::TestController::notifyDone):
438         (WTR::TestController::platformRunUntil):
439         * WebKitTestRunner/win/TestControllerWin.cpp:
440         (WTR::TestController::notifyDone): Add method stub.
441
442 2010-11-15  Darin Adler  <darin@apple.com>
443
444         * Scripts/webkitpy/common/net/bugzilla: Added property svn:ignore.
445
446 2010-11-15  Benjamin Poulain  <benjamin.poulain@nokia.com>
447
448         Reviewed by Andreas Kling.
449
450         [Qt] [WK2] Add Ctrl+L shortcut to Qt Minibrowser
451         https://bugs.webkit.org/show_bug.cgi?id=49544
452
453         Add the shortcut to BrowserWindow.
454
455         * MiniBrowser/qt/BrowserWindow.cpp:
456         (BrowserWindow::BrowserWindow):
457         (BrowserWindow::openLocation):
458         * MiniBrowser/qt/BrowserWindow.h:
459
460 2010-11-14  David Levin  <levin@chromium.org>
461
462         Reviewed by Daniel Bates.
463
464         check-webkit-style should detect PassRefPtr usage in functions.
465         https://bugs.webkit.org/show_bug.cgi?id=49513
466
467         * Scripts/webkitpy/style/checkers/cpp.py:
468         (check_for_function_lengths): Revert a comment change that I
469          accidentally made in r71986.
470         (check_pass_ptr_usage): Added the code to do the check.
471         (process_line): Added the call to check_pass_ptr_usage.
472         (CppChecker): Added the new error category.
473         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
474         (CppStyleTestBase::perform_pass_ptr_check): Runs the new check for
475          testing purposes.
476         (PassPtrTest::*): The class/functions to unit test the new
477          functionality.
478
479 2010-11-14  David Levin  <levin@chromium.org>
480
481         Reviewed by Shinichiro Hamaji.
482
483         check-webkit-style function detection and the line count style checks should be separate.
484         https://bugs.webkit.org/show_bug.cgi?id=49512
485
486         * Scripts/webkitpy/style/checkers/cpp.py: Do the separation.
487         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Adjust the test to
488           call the detection function and fix line counts in two places now that
489           the code really only counts the lines in the body of the function.
490
491 2010-11-14  Andreas Kling  <kling@webkit.org>
492
493         Reviewed by Antonio Gomes.
494
495         http/tests/plugins tests print "Unhandled variable" to stderr under Qt
496         https://bugs.webkit.org/show_bug.cgi?id=33438
497
498         Stifle "Unhandled variable" warning to match what the other
499         TestNetscapePlugin does (NPP_GetValue in TestNetscapePlugIn/main.cpp
500         simply returns NPERR_GENERIC_ERROR for unhandled variables.)
501
502         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
503         (webkit_test_plugin_get_value):
504
505 2010-11-13  David Levin  <levin@chromium.org>
506
507         Reviewed by Eric Seidel.
508
509         check-webkit-style function detection crashes on functions in templates.
510         https://bugs.webkit.org/show_bug.cgi?id=49504
511
512         For example "template <bool shouldClose(const Element*)>".
513
514         * Scripts/webkitpy/style/checkers/cpp.py: If not function is detected, bail out.
515         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Modified the complex
516           function detection test to expose this issue.
517
518 2010-11-12  David Levin  <levin@chromium.org>
519
520         Reviewed by Shinichiro Hamaji.
521
522         check-webkit-style function detection doesn't detect indented functions declaractions.
523         https://bugs.webkit.org/show_bug.cgi?id=49446
524
525         Indented function declarations occur inside class definitions, so
526         they are a pretty common (and worth detecting).
527
528         * Scripts/webkitpy/style/checkers/cpp.py:
529           Changed regex to allow indentation.
530           Changed the function start detection to only happen when not in a
531           function.
532           Changed function end detection to work based on matching braces
533           instead of finding a close brace at the beginning of the line.
534           Fixed close_expression to do what it says when it doesn't find
535           the close.
536         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Indented function test.
537
538 2010-11-12  Daniel Bates  <dbates@rim.com>
539
540         Rubber-stamped by Nikolas Zimmermann.
541
542         Make do-file-rename rename files in the directory WebKit2.
543
544         This also makes the list of searched directories in do-file-rename
545         consistent with the list of searched directories in the script do-webcore-rename.
546
547         * Scripts/do-file-rename:
548
549 2010-11-12  Dirk Pranke  <dpranke@chromium.org>
550
551         Reviewed by James Robinson.
552
553         Attempt yet again to land the fix for bug 49360 (respecting
554         set-webkit-configuration). We need to handle the cases where
555         trying to run webkit-build-directory to find out where the
556         default configuration might be fails (that shows up on some
557         Chromium bots that apparently don't have perl installed).
558
559         https://bugs.webkit.org/show_bug.cgi?id=49360
560
561         * Scripts/webkitpy/layout_tests/port/config.py:
562         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Added.
563         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
564
565 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
566
567         Reviewed by Adam Barth.
568
569         webkit-patch land logs commit message twice
570         https://bugs.webkit.org/show_bug.cgi?id=49482
571         
572         Don't log the comment that's used by bugzilla.close_bug_as_fixed, since
573         that's just the commit message, which we now output following the
574         actual commit.
575
576         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
577
578 2010-11-12  Tony Chang  <tony@chromium.org>
579
580         Reviewed by Adam Barth.
581
582         run platform/chromium/plugins/script-object-invoke.html on all platforms
583         https://bugs.webkit.org/show_bug.cgi?id=49280
584
585         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
586         (testScriptObjectInvoke):
587         (pluginInvoke):
588
589 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
590
591         Reviewed by Adam Barth.
592
593         webkit-patch land-cowboy (and other commands) should display the committed SVN revision after landing a patch
594         https://bugs.webkit.org/show_bug.cgi?id=49471
595
596         Right now there's no indication of a successful commit.
597         
598         * Scripts/webkitpy/tool/commands/download_unittest.py:
599         * Scripts/webkitpy/tool/steps/commit.py:
600
601 2010-11-12  Kinuko Yasuda  <kinuko@chromium.org>
602
603         Reviewed by Ojan Vafai.
604
605         Include detailed test modifiers in results.json and enable incremental uploading for non-layout tests
606         https://bugs.webkit.org/show_bug.cgi?id=49354
607
608         Also moved/integrated the upload method from run_webkit_tests.py to json_results_generator.py.
609
610         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
611         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
612         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
613         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
614
615 2010-11-12  Tony Chang  <tony@chromium.org>
616
617         Reviewed by Kent Tamura.
618
619         [chromium] Load a blank page when closing DRT WebViewHosts
620         https://bugs.webkit.org/show_bug.cgi?id=49417
621
622         In test_shell, when we close a window, we first load about:blank
623         and run GC to fire the destruction logic of the page.  In test_shell,
624         this happens in ~TestShell.  In DRT, we manage virtual windows as
625         WebViewHosts, so we need to replicate this logic in ~WebViewHost.
626
627         Managing focus is a bit trickier, since normally you would get window
628         events to reset the focus to the right window.  We do it manually
629         instead.
630
631         This fixes 2 plugin tests on DRT.
632
633         * DumpRenderTree/chromium/TestShell.cpp:
634         (TestShell::~TestShell): Move window close logic to ~WebViewHost.
635         (TestShell::closeWindow): Move widget close() to WVH destructor and
636             manage focus manually.
637         (TestShell::closeRemainingWindows): Comment typo.
638         * DumpRenderTree/chromium/WebViewHost.cpp:
639         (WebViewHost::closeWidget):
640         (invokeCloseWidget):
641         (WebViewHost::closeWidgetSoon): Close the window asynchronously (matches
642             test_shell).
643         (WebViewHost::~WebViewHost):
644         * DumpRenderTree/chromium/WebViewHost.h:
645
646 2010-11-12  David Levin  <levin@chromium.org>
647
648         Reviewed by Adam Barth.
649
650         cpp.py needs some tidying.
651         https://bugs.webkit.org/show_bug.cgi?id=49443
652
653         * Scripts/webkitpy/style/checkers/cpp.py:
654           Fix parameter ordering for iteratively_replace_matches_with_char
655           to mirror other similar functions (and improved comments, etc.)
656           Also, use a python-ism (for/else) to get rid of a variable that
657           was there to determine if something had been found in the loop.
658
659 2010-11-12  Robert Kroeger  <rjkroege@chromium.org>
660
661         Reviewed by James Robinson.
662
663         [Chromium/DRT] Make EventSendingController send complete touch events.
664         https://bugs.webkit.org/show_bug.cgi?id=49285
665
666         * DumpRenderTree/chromium/EventSender.cpp:
667         (EventSender::addTouchPoint):
668         (EventSender::updateTouchPoint):
669         (EventSender::sendCurrentTouchEvent):
670
671 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
672
673         Unreviewed rollout of r71858.
674
675         Rollout out r71858 since it breaks new-run-webkit-httpd as used by the
676         NaCl tests.
677
678         * Scripts/webkitpy/layout_tests/port/config.py:
679         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Removed.
680         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
681
682 2010-11-12  Benjamin Poulain  <benjamin.poulain@nokia.com>
683
684         Reviewed by Kenneth Rohde Christiansen.
685
686         [Qt] TestController::runUntil doesn't honor the timeout parameter
687         https://bugs.webkit.org/show_bug.cgi?id=48941
688
689         Implement the missing timeout of WebKitTestRunner for Qt.
690
691         A QElapsedTimer has been added to RunUntilConditionLoop in order
692         to measure how long the loop has been running. When the timer
693         is bigger or equal than the timeout, the loop ends even if the
694         condition is not met.
695
696         * WebKitTestRunner/qt/TestControllerQt.cpp:
697         (WTR::RunUntilConditionLoop::start):
698         (WTR::RunUntilConditionLoop::run):
699         (WTR::RunUntilConditionLoop::timerEvent):
700         (WTR::TestController::platformRunUntil):
701
702 2010-11-11  David Levin  <levin@chromium.org>
703
704         Reviewed by Shinichiro Hamaji.
705
706         check-webkit-style function detection doesn't handle templates with spaces.
707         https://bugs.webkit.org/show_bug.cgi?id=49427
708
709         * Scripts/webkitpy/style/checkers/cpp.py: Stripped out templates when
710           finding the function name.
711         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test have a
712           template with a space.
713
714 2010-11-11  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
715
716         Reviewed by Shinichiro Hamaji.
717
718         Enable check-webkit-style on Qt files
719         https://bugs.webkit.org/show_bug.cgi?id=48322
720
721         Adding exception for _q_... private slots and
722         ..._data() methods in tests
723
724         * Scripts/webkitpy/style/checker.py: enable Qt dirs, add exceptions
725         * Scripts/webkitpy/style/checker_unittest.py: improve unittests
726
727 2010-11-11  David Levin  <levin@chromium.org>
728
729         Reviewed by Shinichiro Hamaji.
730
731         check-webkit-style function detection doesn't work for templates and destructors.
732         https://bugs.webkit.org/show_bug.cgi?id=49425
733
734         * Scripts/webkitpy/style/checkers/cpp.py: Changed regex for templates and destructors.
735         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test hit this issue.
736
737 2010-11-11  Joone Hur  <joone@kldp.org>
738
739         Reviewed by Antonio Gomes.
740
741         [GTK][DRT] Implement LayoutTestController::nodesFromRect
742         https://bugs.webkit.org/show_bug.cgi?id=46598
743
744         Support nodesFromRect in DRT
745
746         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
747         (LayoutTestController::nodesFromRect):
748
749 2010-11-11  MORITA Hajime  <morrita@google.com>
750
751         Reviewed by Kent Tamura.
752
753         editing/selection/context-menu-on-text.html fails on chromium
754         https://bugs.webkit.org/show_bug.cgi?id=45898
755
756         Enhance Chromium's EventSender::contextClick() to return
757         array of menu strings. The implementation incomplete because
758         Although actual context menu is implemented by the browser,
759         it reflects editability of underlying context thus makes
760         context-menu-on-text.html passed.
761         
762         * DumpRenderTree/chromium/EventSender.cpp:
763         (makeMenuItemStringsFor): Added.
764         (EventSender::contextClick): Now returns array of menu strings.
765         * DumpRenderTree/chromium/WebViewHost.cpp:
766         (WebViewHost::showContextMenu): Now storing given WebContextMenuData object.
767         (WebViewHost::clearContextMenuData): Added.
768         (WebViewHost::lastContextMenuData): Added.
769         * DumpRenderTree/chromium/WebViewHost.h:
770
771 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
772
773         Unreviewed, build fix.
774
775         fix build breakage caused by fix for bug 49380 (r71858).
776
777         config_standalone needs to ensure that the packages it needs are
778         in sys.path.
779
780         https://bugs.webkit.org/show_bug.cgi?id=49419
781
782         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
783
784 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
785
786         Reviewed by Adam Roben.
787
788         Fix NRWT to respect set-webkit-configuration again :(
789
790         This change fixes a typo in config.py that was causing the wrong
791         value to be read initially and us never actually looking into
792         the filesystem to get the default configuration.
793         
794         * Scripts/webkitpy/layout_tests/port/config.py:
795         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
796         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
797
798 2010-11-11  Eric Seidel  <eric@webkit.org>
799
800         Reviewed by Adam Barth.
801
802         Split out Bug, Attachment and CommitterValidator from bugzilla.py
803         https://bugs.webkit.org/show_bug.cgi?id=49403
804
805         CommitterValidator really had nothing to do with bugzilla.py.
806         I've put it next to committers.py for now since its rather tied to that file.
807
808         * Scripts/webkitpy/common/config/committervalidator.py: Added.
809         * Scripts/webkitpy/common/config/committervalidator_unittest.py: Added.
810         * Scripts/webkitpy/common/net/bugzilla/__init__.py:
811         * Scripts/webkitpy/common/net/bugzilla/attachment.py: Added.
812         * Scripts/webkitpy/common/net/bugzilla/bug.py: Added.
813         * Scripts/webkitpy/common/net/bugzilla/bug_unittest.py: Added.
814         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
815         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py:
816         * Scripts/webkitpy/tool/bot/feeders.py:
817         * Scripts/webkitpy/tool/commands/queues.py:
818
819 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
820
821         Reviewed by Eric Seidel.
822
823         fix missing 'pid' argument for http lockfile checking on win32
824         https://bugs.webkit.org/show_bug.cgi?id=49363
825
826         * Scripts/webkitpy/common/system/executive.py:
827
828 2010-11-11  Eric Seidel  <eric@webkit.org>
829
830         Rubber-stamped by Adam Barth.
831
832         Move bugzilla.py into its own module in preparation for splitting one-file-per-class
833         https://bugs.webkit.org/show_bug.cgi?id=49402
834
835         * Scripts/webkitpy/common/net/bugzilla/__init__.py: Added.
836         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla.py.
837         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla_unittest.py.
838
839 2010-11-10  Chris Guillory  <chris.guillory@google.com>
840
841         Reviewed by Chris Fleizach.
842
843         Use IAccessibleComparable to compare accessibility objects.
844         https://bugs.webkit.org/show_bug.cgi?id=49118
845
846         * DumpRenderTree/AccessibilityUIElement.cpp:
847         (AccessibilityUIElement::isEqual):
848         * DumpRenderTree/AccessibilityUIElement.h:
849         * DumpRenderTree/win/AccessibilityUIElementWin.cpp:
850         (comparableObject):
851         (AccessibilityUIElement::isEqual):
852
853 2010-11-10  Csaba Osztrogonác  <ossy@webkit.org>
854
855         Reviewed by David Hyatt.
856
857         HTML5 Ruby support should be mandatory feature
858         https://bugs.webkit.org/show_bug.cgi?id=49272
859
860         Remove Ruby as optional feature.
861
862         * Scripts/build-webkit:
863
864 2010-11-10  Adam Roben  <aroben@apple.com>
865
866         Print VC++ Express build logs in the same order that the projects are
867         built
868
869         Fixes <http://webkit.org/b/49326> print-vse-failure-logs prints logs
870         in an unhelpful order
871
872         Reviewed by Eric Seidel.
873
874         * Scripts/print-vse-failure-logs:
875         (PrintVisualStudioExpressLogs._build_order): Added. Uses
876         print-msvc-project-dependencies to figure out the order in which
877         projects are built.
878         (PrintVisualStudioExpressLogs._sort_buildlogs): Added. Sorts the logs
879         based on their build order and project name.
880         (PrintVisualStudioExpressLogs._obj_directory): Moved code to find the
881         scripts directory from here...
882         (PrintVisualStudioExpressLogs._scripts_directory): ...to here.
883         (PrintVisualStudioExpressLogs.main): Sort the logs before printing
884         them.
885
886 2010-11-10  Adam Roben  <aroben@apple.com>
887
888         Only print the interesting text from VC++ Express build logs
889
890         Fixes <http://webkit.org/b/49325> It's hard to find the interesting
891         output from print-vse-failure-logs
892
893         Reviewed by Eric Seidel.
894
895         * Scripts/print-vse-failure-logs: Removed unnecessary os.path import,
896         added newly-required imports. Removed a comment that was essentially a
897         FIXME, and is now fixed.
898         (PrintVisualStudioExpressLogs._relevant_text): Added. Uses
899         BeautifulSoup to extract the relevant text from the build log.
900         (PrintVisualStudioExpressLogs.main): Only print the relevant text.
901
902 2010-11-09  Mihai Parparita  <mihaip@chromium.org>
903
904         Reviewed by Tony Chang.
905
906         [Chromium] http/tests/security/XFrameOptions fail with Chromium DRT
907         https://bugs.webkit.org/show_bug.cgi?id=49286
908         
909         The tests in http/tests/security/XFrameOptions started to fail after
910         r71297, since it changed the WebViewHost::assignIdentifierToRequest
911         logic to always assign identifiers to requests, even if we're not going
912         to be dumping resource load callbacks at that point. These tests
913         only call LayoutTestController.dumpResourceLoadCallbacks after the
914         request is started, thus their expectations have "<unknown>" as the
915         identifier. Change WebViewHost to the previous behavior, which also
916         match the other ports'.
917
918         * DumpRenderTree/chromium/WebViewHost.cpp:
919         (WebViewHost::assignIdentifierToRequest):
920         (WebViewHost::removeIdentifierForRequest):
921
922 2010-11-08  Tony Chang  <tony@chromium.org>
923
924         Reviewed by Adam Barth.
925
926         run platform/chromium/plugins/return-npobject.html on all platforms
927         https://bugs.webkit.org/show_bug.cgi?id=49103
928
929         This tests that we can get an NPObject returned through a method on
930         an NPAPI Object.
931
932         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
933         (testCallbackReturn):
934         (pluginInvoke):
935
936 2010-11-09  Sheriff Bot  <webkit.review.bot@gmail.com>
937
938         Unreviewed, rolling out r71672.
939         http://trac.webkit.org/changeset/71672
940         https://bugs.webkit.org/show_bug.cgi?id=49276
941
942         It doesn't work on Chromium Windows bot (Requested by Ossy on
943         #webkit).
944
945         * Scripts/webkitpy/common/system/file_lock.py: Removed.
946         * Scripts/webkitpy/common/system/file_lock_unittest.py: Removed.
947         * Scripts/webkitpy/layout_tests/port/http_lock.py:
948
949 2010-11-09  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
950
951         Reviewed by Tony Chang.
952
953         [NRWT] Make http locking similar to perl implementation
954         https://bugs.webkit.org/show_bug.cgi?id=49187
955
956         * Scripts/webkitpy/common/system/file_lock.py: Added.
957         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
958         * Scripts/webkitpy/layout_tests/port/http_lock.py:
959
960 2010-11-09  James Kozianski  <koz@chromium.org>
961
962         Reviewed by Eric Seidel.
963
964         Clean up imports for webkit-patch.
965         https://bugs.webkit.org/show_bug.cgi?id=49083
966
967         Move some imports from commands into main.py where they are used.
968         Move command imports into the __init__ of the commands module.
969
970         * Scripts/webkitpy/tool/commands/__init__.py:
971         * Scripts/webkitpy/tool/commands/download.py:
972         * Scripts/webkitpy/tool/commands/queues.py:
973         * Scripts/webkitpy/tool/main.py:
974
975 2010-11-09  Yi Shen  <yi.4.shen@nokia.com>
976
977         Reviewed by Andreas Kling.
978
979         [Qt][QtTestBrowser] Don't erase incorrect url in the Url Bar
980         https://bugs.webkit.org/show_bug.cgi?id=49047
981
982         * QtTestBrowser/mainwindow.cpp:
983         (MainWindow::setAddressUrl):
984
985 2010-11-09  Andras Becsi  <abecsi@webkit.org>
986
987         Reviewed by Csaba Osztrogonác.
988
989         Improve old-run-webkit-tests --verbose to show DumpTool opening/closing
990         https://bugs.webkit.org/show_bug.cgi?id=49252
991
992         * Scripts/old-run-webkit-tests:
993
994 2010-11-09  Csaba Osztrogonác  <ossy@webkit.org>
995
996         Reviewed by Tor Arne Vestbø.
997
998         Output of old-run-webkit-tests is buggy in verbose mode
999         https://bugs.webkit.org/show_bug.cgi?id=49249
1000
1001         * Scripts/old-run-webkit-tests: Redundant outputs removed.
1002
1003 2010-11-08  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
1004
1005         Reviewed by Martin Robinson.
1006
1007         [GTK] Link with target name set does not work
1008         https://bugs.webkit.org/show_bug.cgi?id=48865
1009
1010         When a new page is created with a name (target=myFrame), the new
1011         mainFrame could not be found because they where not stored in the
1012         same PageGroup. As PageGroup are not exposed externally so the
1013         simpliest solution is to use a global page group name. This also fixes
1014         issue with visited link coloration across pages. After this change the
1015         private function webkit_web_view_set_group_name() was no longer used
1016         so it was removed completly.
1017
1018         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1019         (createWebView):
1020
1021 2010-11-08  John Knottenbelt  <jknotten@chromium.org>
1022
1023         Reviewed by Kent Tamura.
1024
1025         WebViewHost::reset() uses placement new.
1026         https://bugs.webkit.org/show_bug.cgi?id=49069
1027
1028         WebViewHost is using placement destruction / new to simulate a fresh
1029         WebViewHost object at the same address. This is because the WebView remains
1030         open across tests and maintains a pointer to the WebViewHost.
1031         This change resets member variables explictly instead of the placement new dance.
1032
1033         * DumpRenderTree/chromium/WebViewHost.cpp:
1034         (WebViewHost::WebViewHost):
1035         (WebViewHost::reset):
1036
1037 2010-11-08  Dirk Pranke <dpranke@chromium.org>
1038
1039         Reviewed by Eric Seidel.
1040
1041         Fix the webkit_base_dir logic in webkitpy/layout_tests/port/config.py. 
1042         It turns out that NRWT can't use the code in scm.find_checkout_root()
1043         because the Chromium bots don't do full checkouts of the WebKit
1044         tree; they only check out subdirectories like WebKitTools/Scripts.
1045         Until we can figure out a better approach for this, I've
1046         restored the base_dir-detecting code from NRWT, which works in
1047         any directory tree, scm or no.
1048
1049         This also restores the files modified in r71475 and r71474.
1050
1051         https://bugs.webkit.org/show_bug.cgi?id=49151
1052
1053         * WebKitTools/Scripts/webkitpy/layout_tests/port/config.py
1054         * WebKitTools/Scripts/webkitpy/layout_tests/port/config_unittest.py
1055
1056 2010-11-08  Eric Seidel  <eric@webkit.org>
1057
1058         Reviewed by Adam Barth.
1059
1060         webkit-patch failure-reason explodes if a build is missing
1061         https://bugs.webkit.org/show_bug.cgi?id=49195
1062
1063         This is likely a recent regression.  Adding a unit test for this case.
1064
1065         * Scripts/webkitpy/common/net/buildbot.py:
1066         * Scripts/webkitpy/common/net/buildbot_unittest.py:
1067
1068 2010-11-08  Dan Bernstein  <mitz@apple.com>
1069
1070         Reviewed by Darin Adler.
1071
1072         Ignore the system scrollbar setting.
1073
1074         * DumpRenderTree/mac/DumpRenderTree.mm:
1075         (resetDefaultsToConsistentValues): Set the scrollbar default.
1076
1077 2010-11-05  Adam Roben  <aroben@apple.com>
1078
1079         Make webkitpy.common.system.executive_unittest pass when running under
1080         Win32 Python
1081
1082         Fixes <http://webkit.org/b/49033>.
1083
1084         Reviewed by Dave Levin and Eric Seidel.
1085
1086         * Scripts/webkitpy/common/system/executive.py:
1087         (Executive._run_command_with_teed_output): Pass the arguments through
1088         encode_argument_if_needed rather than using Cygwin-specific code here.
1089         (Executive.run_and_throw_if_fail): Use child_process_encoding to decode
1090         the output.
1091         (Executive.run_command): Use encode_argument_if_needed to encode the
1092         arguments and child_process_encoding to decode the output.
1093         (Executive._child_process_encoding): Returns the encoding that should be
1094         used when communicating with child processes. On Windows we use mbcs,
1095         which maps to the current code page. On all other platforms we use
1096         UTF-8.
1097         (Executive._should_encode_child_process_arguments): Returns True if
1098         arguments to child processes need to be encoded. This is currently
1099         only needed on Cygwin and Win32 Python 2.x.
1100         (Executive._encode_argument_if_needed): Encode the argument using
1101         child_process_encoding if we need to encode arguments to child
1102         processes on this platform.
1103
1104         * Scripts/webkitpy/common/system/executive_unittest.py:
1105         (never_ending_command): Added. Returns arguments to run a command that
1106         will not quit until we kill it. On Windows we use wmic, on other
1107         platforms we use yes.
1108         (ExecutiveTest.test_run_command_with_unicode): Changed to expect the
1109         mbcs encoding to be used and for output from the child processes to
1110         have been roundtripped through encode/decode on Win32 Python. When
1111         UTF-8 is the encoding the roundtripping is undetectable, but with mbcs
1112         it's possible that some characters will not be able to be converted
1113         and will be replaced by question marks; the round-tripping allows us
1114         to expect this result.
1115
1116         (ExecutiveTest.test_kill_process):
1117         (ExecutiveTest.test_kill_all):
1118         Use never_ending_command instead of invoking "yes" directly. Expect an
1119         exit code of 1 when using Win32 Python, as that's what seems to happen.
1120
1121 2010-11-08  Adam Roben  <aroben@apple.com>
1122
1123         Roll out r71532
1124
1125         It broke the build for Cygwin 1.7 installs. Cygwin 1.7's default
1126         .bashrc unsets %TEMP%, which broke copy-tools.cmd.
1127
1128         * Scripts/webkitdirs.pm:
1129
1130 2010-11-08  Tony Chang  <tony@chromium.org>
1131
1132         Reviewed by Adam Barth.
1133
1134         run platform/chromium/plugins/nested-plugin-objects.html on all platforms
1135         https://bugs.webkit.org/show_bug.cgi?id=49094
1136
1137         This tests that objects created by plugins are proplery cleaned up.
1138
1139         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1140         (testPassTestObject):
1141         (pluginInvoke):
1142         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
1143         (testAllocate):
1144         (testDeallocate):
1145         (testGetProperty):
1146         (testConstruct):
1147
1148 2010-11-08  Adam Roben  <aroben@apple.com>
1149
1150         Mark Windows builds triggered from Perl as being non-interactive
1151
1152         This affects whether some of our scripts will show alerts vs. printing
1153         to the build log.
1154
1155         Fixes <http://webkit.org/b/49181> Windows build fail mysteriously when
1156         .vsprops files are updated
1157
1158         Reviewed by Steve Falkenburg.
1159
1160         * Scripts/webkitdirs.pm:
1161         (buildVisualStudioProject): Set WEBKIT_NONINTERACTIVE_BUILD to 1.
1162
1163 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1164
1165         Reviewed by Ojan Vafai.
1166
1167         Make http locking default in NRWT.
1168         https://bugs.webkit.org/show_bug.cgi?id=48053
1169
1170         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1171         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1172
1173 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1174
1175         Reviewed by Csaba Osztrogonác.
1176
1177         [NRWT] If the http lock fails we shouldn't do any locking
1178         https://bugs.webkit.org/show_bug.cgi?id=49164
1179
1180         If something goes wrong with the locking, the test should keep going.
1181
1182         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1183
1184 2010-11-08  Adam Roben  <aroben@apple.com>
1185
1186         Switch back to using kCGImageAlphaPremultipliedFirst when generating
1187         pixel dumps on Windows
1188
1189         I changed this behavior in r71418 thinking that it was required for
1190         getting plugins to show up in pixel dumps. But it doesn't seem to be
1191         necessary, and was making it impossible to compare new Windows pixel
1192         dumps with existing Windows or Mac pixel dumps (because ImageDiff won't
1193         compare an image with alpha to an image without alpha).
1194
1195         Fixes <http://webkit.org/b/49172> REGRESION (r71418): Can't compare
1196         new Windows pixel results to existing Windows or Mac results
1197
1198         Reviewed by Antti Koivisto.
1199
1200         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
1201         (createBitmapContextFromWebView): Replaced kCGImageAlphaNoneSkipFirst
1202         with kCGImageAlphaPremultipliedFirst.
1203
1204 2010-11-08  Csaba Osztrogonac  <ossy@webkit.org>
1205
1206         Unreviewed, rolling out r71466.
1207         http://trac.webkit.org/changeset/71466
1208         https://bugs.webkit.org/show_bug.cgi?id=48865
1209
1210         It broke layout tests on GTK bots.
1211
1212         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1213         (createWebView):
1214
1215 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1216
1217         Reviewed by Csaba Osztrogonác.
1218
1219         Enable running of Qt API tests on BuildBot
1220         https://bugs.webkit.org/show_bug.cgi?id=49004
1221
1222         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1223
1224 2010-11-08  Eric Seidel  <eric@webkit.org>
1225
1226         Reviewed by Adam Barth.
1227
1228         Add clean-review-queue command to remove closed bugs from the webkit.org/pending-review
1229         https://bugs.webkit.org/show_bug.cgi?id=49160
1230
1231         Bugzilla doesn't automatically remove r? when a bug gets closed.
1232         This script takes care of that for webkit.org.
1233
1234         * Scripts/webkitpy/common/net/bugzilla.py:
1235         * Scripts/webkitpy/tool/commands/upload.py:
1236
1237 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
1238
1239         Unreviewed, rolling out r71474.
1240         http://trac.webkit.org/changeset/71474
1241         https://bugs.webkit.org/show_bug.cgi?id=48280
1242
1243         breaks chromium webkit tests
1244         https://bugs.webkit.org/show_bug.cgi?id=49151
1245
1246         * Scripts/webkitpy/layout_tests/port/base.py:
1247         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
1248         * Scripts/webkitpy/layout_tests/port/config.py:
1249         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1250         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
1251         * Scripts/webkitpy/layout_tests/port/webkit.py:
1252         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1253         * Scripts/webkitpy/tool/mocktool.py:
1254
1255 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
1256
1257         Unreviewed, rolling out r71475.
1258         http://trac.webkit.org/changeset/71475
1259  
1260         breaks chromium webkit tests
1261         https://bugs.webkit.org/show_bug.cgi?id=49151
1262
1263         * Scripts/webkitpy/common/newstringio.py: Removed.
1264         * Scripts/webkitpy/common/newstringio_unittest.py: Removed.
1265         * Scripts/webkitpy/common/system/executive_mock.py: Removed.
1266         * Scripts/webkitpy/common/system/filesystem_mock.py: Removed.
1267         * Scripts/webkitpy/layout_tests/port/config_mock.py: Removed.
1268
1269 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1270
1271         Unreviewed, build fix.
1272
1273         Add files inexplicably not committed in r71474 as part of the
1274         fix for bug 48280.
1275
1276         * Scripts/webkitpy/common/newstringio.py: Added.
1277         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
1278         * Scripts/webkitpy/common/system/executive_mock.py: Added.
1279         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
1280         * Scripts/webkitpy/layout_tests/port/config_mock.py: Added.
1281
1282 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1283
1284         Reviewed by Eric Siedel.
1285
1286         new-run-webkit-tests: update port/base and port/webkit to use the
1287         new FileSystem and Config abstractions, pulling more logic out of
1288         the base Port classes into separate, mockable objects.
1289
1290         Also create a MockFileSystem object, a MockConfig object, move
1291         MockExecutive into common/system to be next to executive, and
1292         update the config object to use a FileSystem.
1293
1294         https://bugs.webkit.org/show_bug.cgi?id=48280
1295
1296         * Scripts/webkitpy/common/newstringio.py: Added.
1297         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
1298         * Scripts/webkitpy/common/system/executive_mock.py: Added.
1299         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
1300         * Scripts/webkitpy/layout_tests/port/base.py:
1301         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
1302         * Scripts/webkitpy/layout_tests/port/config.py:
1303         * Scripts/webkitpy/layout_tests/port/config_mock.py:
1304         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1305         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
1306         * Scripts/webkitpy/layout_tests/port/webkit.py:
1307         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1308         * Scripts/webkitpy/tool/mocktool.py:
1309
1310 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1311
1312         Unreviewed, build breakage.
1313
1314         Apparently I uploaded the wrong version of the file to fix 49122
1315         and neither Eric or I noticed - it was missing a dirname() call.
1316         Fixing ...
1317
1318         https://bugs.webkit.org/show_bug.cgi?id=49122
1319
1320         * Scripts/webkitpy/common/checkout/scm.py:
1321
1322 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1323
1324         Reviewed by Eric Seidel.
1325
1326         webkitpy/tool/* unittests change cwd and don't clean up properly
1327
1328         https://bugs.webkit.org/show_bug.cgi?id=49122
1329
1330         * Scripts/webkitpy/common/checkout/scm.py:
1331
1332 2010-11-05  Chris Marrin  <cmarrin@apple.com>
1333
1334         Reviewed by Simon Fraser.
1335
1336         suspendAnimations/resumeAnimations not present in WebKit2
1337         https://bugs.webkit.org/show_bug.cgi?id=49109
1338
1339         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
1340         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
1341         (WTR::LayoutTestController::suspendAnimations):
1342         (WTR::LayoutTestController::resumeAnimations):
1343         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
1344
1345 2010-11-05  Tony Chang  <tony@chromium.org>
1346
1347         Reviewed by David Levin.
1348
1349         cleanup style in TestNetscapePlugIn/PluginObject.cpp
1350         https://bugs.webkit.org/show_bug.cgi?id=49044
1351
1352         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1353         (getPluginClass):
1354         (pluginGetProperty):
1355         (pluginSetProperty):
1356         (variantToIdentifier):
1357         (testCallback):
1358         (getURL):
1359         (getURLNotify):
1360         (testInvokeDefault):
1361         (testGetProperty):
1362         (testDocumentOpen):
1363         (testWindowOpen):
1364         (pluginInvoke):
1365         (handleCallback):
1366         (notifyStream):
1367         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
1368
1369 2010-11-05  Eric Seidel  <eric@webkit.org>
1370
1371         Reviewed by David Levin.
1372
1373         Add QueueStatusServer/__init__.py so others can run the QueueStatusServer tests
1374         https://bugs.webkit.org/show_bug.cgi?id=49032
1375
1376         I wrote this file as part of bug 47847, but I forgot to commit it.
1377         No one else noticed it missing because test-webkitpy knows how
1378         to recover in the case where it can't import QueueStatusServer
1379         (which generally occurs due to not having installed the AppEngine SDK).
1380
1381         * QueueStatusServer/__init__.py: Added.
1382         * QueueStatusServer/model/workitems_unittest.py:
1383          - Remove a test which fails.  This was probably landed (by me)
1384            from my other machine, which since this __init__.py was missing
1385            I never noticed the failure and landed this invalid test.
1386            Sadly we can't really test remove_work_item as it depends
1387            on .key() working.  .key() will throw unless the object
1388            has already been saved it seems.
1389            This may be a mis-design in our remove_work_item implementation,
1390            but for now, just removing the test.
1391
1392 2010-11-04  Adam Roben  <aroben@apple.com>
1393
1394         Teach check-webkit-style about TestNetscapePlugIn
1395
1396         Fixes <http://webkit.org/b/49030> check-webkit-style is ignorant of
1397         TestNetscapePlugIn's coding conventions
1398
1399         Reviewed by John Sullivan.
1400
1401         * Scripts/webkitpy/style/checker.py: Lump TestNetscapePlugIn in with
1402         WebKitAPITest and TestWebKitAPI in having different include order and
1403         naming conventions than WebCore. Added some comments about why the
1404         exceptions exist.
1405
1406 2010-11-04  Adam Roben  <aroben@apple.com>
1407
1408         Add a test that shows that windowed plugins are able to paint
1409
1410         Somehow we never had a test for this before.
1411
1412         Fixes <http://webkit.org/b/49029> Should add a test that shows
1413         windowed plugins can paint
1414
1415         Reviewed by Jon Honeycutt.
1416
1417         * DumpRenderTree/TestNetscapePlugIn/Tests/win/DrawsGradient.cpp: Added.
1418         (DrawsGradient::DrawsGradient):
1419         (DrawsGradient::wndProc): We handle the WM_PAINT and WM_PRINTCLIENT messages.
1420
1421         (DrawsGradient::onPaint):
1422         (DrawsGradient::onPrintClient):
1423         These both just call through to paint.
1424
1425         (DrawsGradient::paint): Fills our client area with some gradients.
1426
1427         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Link
1428         against Msimg32.lib for ::GradientFill and added DrawsGradient.
1429
1430         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
1431         (createBitmapContextFromWebView): Use WM_PRINT instead of
1432         WM_PRINTCLIENT so that ::DefWindowProc will send
1433         WM_PRINT/WM_PRINTCLIENT messages to the WebView's child windows.
1434         Replaced kCGImageAlphaPremultipledFirst with kCGImageAlphaNoneSkipFirst
1435         because GDI doesn't support alpha and kCGImageBitmapOrder32Little with
1436         kCGImageBitmapOrder32Host because, who knows, maybe someday Windows
1437         will run on a big-endian platform.
1438
1439 2010-11-04  Adam Roben  <aroben@apple.com>
1440
1441         Extract much of NPNInvalidateRectInvalidatesWindow's code into a
1442         WindowedPluginTest base class
1443
1444         The base class takes care of subclassing the plugin's window so that a
1445         custom WNDPROC is called. This will make it easier to write tests that
1446         need to handle window messages.
1447
1448         Fixes <http://webkit.org/b/49028> It's hard to write a PluginTest with
1449         a custom WNDPROC
1450
1451         Reviewed by Jon Honeycutt.
1452
1453         * DumpRenderTree/TestNetscapePlugIn/Tests/win/NPNInvalidateRectInvalidatesWindow.cpp:
1454         (NPNInvalidateRectInvalidatesWindow::NPNInvalidateRectInvalidatesWindow):
1455         (NPNInvalidateRectInvalidatesWindow::NPP_SetWindow):
1456         (NPNInvalidateRectInvalidatesWindow::wndProc):
1457         (NPNInvalidateRectInvalidatesWindow::testInvalidateRect):
1458         Moved code from here to WindowedPluginTest. Changed to use window()
1459         instead of m_window.
1460
1461         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1462         Added WindowedPluginTest and added TestNetscapePlugIn/win to the
1463         include path. Also reordered the include path to make a little more
1464         sense and simplified the entry that added TestNetscapePlugIn itself to
1465         the include path.
1466
1467         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.cpp: Added.
1468         (WindowedPluginTest::WindowedPluginTest):
1469         (WindowedPluginTest::NPP_SetWindow):
1470         (WindowedPluginTest::staticWndProc):
1471         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.h: Added.
1472         (WindowedPluginTest::window):
1473         Code came from NPNInvalidateRectInvalidatesWindow.
1474
1475 2010-11-04  Adam Roben  <aroben@apple.com>
1476
1477         Add a plugin test to show that windowed plugins are clipped correctly
1478
1479         Fixes <http://webkit.org/b/49024> <rdar://problem/8487847> Windowed
1480         plugins aren't clipped in WebKit2 on Windows
1481
1482         Reviewed by Jon Honeycutt.
1483
1484         * DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowRegionIsSetToClipRect.cpp: Added.
1485         (WindowRegionIsSetToClipRect::WindowRegionIsSetToClipRect): Initialize members.
1486         (WindowRegionIsSetToClipRect::NPP_SetWindow): Check that our window
1487         region matches the clip rect we know we should have based on
1488         window-region-is-set-to-clip-rect.html, and check that our window class
1489         doesn't have the CS_PARENTDC style.
1490
1491         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1492         Added WindowRegionIsSetToClipRect.
1493
1494 2010-11-05  Alejandro G. Castro  <alex@igalia.com>
1495
1496         Reviewed by Martin Robinson.
1497
1498         [GTK] Avoid font hinting in the DRT
1499         https://bugs.webkit.org/show_bug.cgi?id=48548
1500
1501         Change the settings to avoid font hinting, it was causing
1502         different results depending on the distribution.
1503
1504         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1505         (initializeGtkFontSettings):
1506         * DumpRenderTree/gtk/fonts/fonts.conf:
1507
1508 2010-11-05  Eric Seidel  <eric@webkit.org>
1509
1510         Reviewed by Adam Barth.
1511
1512         build-webkit spams system.log by repeatedly checking for PBXProductDirectory in com.apple.Xcode
1513         https://bugs.webkit.org/show_bug.cgi?id=49051
1514
1515         This is a speculative fix.  The unit tests cover these methods, however
1516         I don't know if this will fully stop the system.log spam.
1517
1518         * Scripts/webkitpy/layout_tests/port/base.py:
1519         * Scripts/webkitpy/layout_tests/port/webkit.py:
1520
1521 2010-11-05  Eric Seidel  <eric@webkit.org>
1522
1523         Reviewed by Adam Barth.
1524
1525         Add basic support for showing bot id on /queue-status/ pages
1526         https://bugs.webkit.org/show_bug.cgi?id=49037
1527
1528         This support is really simple.  Eventually we'll want to
1529         show the bot id in the lock table too, but we don't have
1530         that information stored in the server yet.
1531
1532         * QueueStatusServer/handlers/queuestatus.py:
1533         * QueueStatusServer/handlers/queuestatus_unittest.py: Added.
1534         * QueueStatusServer/templates/includes/singlequeuestatus.html:
1535
1536 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
1537
1538         Reviewed by Adam Barth.
1539
1540         Create a filesystem wrapper that we can use to enforce
1541         particular conventions and use for mocking and dependency
1542         injection down the line.
1543
1544         https://bugs.webkit.org/show_bug.cgi?id=48144
1545
1546         * Scripts/webkitpy/common/system/filesystem.py: Added.
1547         * Scripts/webkitpy/common/system/filesystem_unittest.py: Added.
1548
1549 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1550
1551         Reviewed by Tony Chang.
1552
1553         run_webkit_tests_unittest fails under Python 2.5
1554         https://bugs.webkit.org/show_bug.cgi?id=49043
1555         
1556         Switch from itertools.chain.from_iterable (which was added in 2.6)
1557         to using itertools.chain directly.
1558
1559         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1560
1561 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1562
1563         Unreviewed fix to webkit-patch rebaseline-server so that it runs under
1564         Python 2.5 (it needs an import to use the with statement).
1565
1566         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
1567
1568 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
1569
1570         Reviewed by Adam Barth.
1571
1572         new-run-webkit-tests: split out webkit-specific configuration stuff into a new module
1573
1574         The current NRWT code has webkit-specific configuration code (like
1575         _script_path, default configuration, etc.) mixed in with
1576         layout-test-specific stuff in port/base. The configuration code
1577         should be split out into a separate module for easier mocking,
1578         testing, and isolation.
1579
1580         https://bugs.webkit.org/show_bug.cgi?id=48264
1581
1582         * Scripts/webkitpy/layout_tests/port/config.py: Added.
1583         * Scripts/webkitpy/layout_tests/port/config_unittest.py: Added.
1584
1585 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1586
1587         Reviewed by Tony Chang.
1588
1589         Rebaseline server: initial framework
1590         https://bugs.webkit.org/show_bug.cgi?id=48892
1591         
1592         Adds the basic framework for the rebaseline server (details at 
1593         http://webkit.org/b/47761). Includes the rebaseline-server webkit-patch
1594         command, which starts an HTTP server that can serve static files or
1595         invoke handler methods on a class.
1596
1597         * Scripts/webkitpy/tool/commands/__init__.py:
1598         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html: Added.
1599         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css: Added.
1600         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js: Added.
1601         * Scripts/webkitpy/tool/commands/rebaselineserver.py: Added.
1602
1603 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
1604
1605         Reviewed by James Robinson.
1606
1607         new-run-webkit-tests wasn't using DRT by default for
1608         --platform chromium-gpu
1609
1610         The default value was set to False instead of None, which meant
1611         that the platform specific logic wasn't firing to change the
1612         value to True (b/c we were afraid we'd be overriding the user
1613         preference).
1614
1615         https://bugs.webkit.org/show_bug.cgi?id=49038
1616
1617         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1618
1619 2010-11-04  Tony Chang  <tony@chromium.org>
1620
1621         Reviewed by Adam Barth.
1622
1623         make platform/chromium/plugins/multiple-plugins.html pass on all platforms
1624         https://bugs.webkit.org/show_bug.cgi?id=49026
1625
1626         Have the testObject.bar property return the string "bar".  This lets
1627         us run and pass platform/chromium/plugins/multiple-plugins.html on all
1628         platforms.
1629
1630         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
1631         (testGetProperty):
1632
1633 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1634
1635         Reviewed by Tony Chang.
1636
1637         NRWT doesn't have good test coverage for --run-chunk, --batch-size, --run-part, etc.
1638         https://bugs.webkit.org/show_bug.cgi?id=48878
1639         
1640         Add get_tests_run so that it's easy to see which tests get run (and with
1641         what batching) for a given flag combination. Flesh out the various
1642         test cases that have FIXMEs.
1643         
1644         Also fixes an off-by-one error (batch sizes were one larger than 
1645         expected) and makes --run-part also have wraparound behavior, like
1646         --run-chunk.
1647
1648         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1649         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1650         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1651
1652 2010-11-04  Erik Arvidsson  <arv@chromium.org>
1653
1654         Reviewed by Dimitri Glazkov.
1655
1656         Support box-sizing without the vendor prefix
1657         https://bugs.webkit.org/show_bug.cgi?id=36713
1658
1659         Based on patch by Peter Beverloo <peter@lvp-media.com>
1660
1661         * iExploder/htdocs/cssproperties.in: Change -webkit-box-sizing to box-sizing.
1662
1663 2010-11-04  Csaba Osztrogonác  <ossy@webkit.org>
1664
1665         Unreviewed rollout r71340, because it broke Chromium Windows bot.
1666
1667         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1668         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1669
1670 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1671
1672         Reviewed by Ojan Vafai.
1673
1674         Make http locking default in NRWT.
1675         https://bugs.webkit.org/show_bug.cgi?id=48053
1676
1677         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1678         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1679
1680 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1681
1682         Reviewed by Eric Seidel.
1683
1684         [NRWT] Clear invalid http locks on Windows platform as well
1685         https://bugs.webkit.org/show_bug.cgi?id=48515
1686
1687         * Scripts/webkitpy/common/system/executive.py:
1688         * Scripts/webkitpy/common/system/executive_unittest.py:
1689         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1690
1691 2010-11-02  Adam Roben  <aroben@apple.com>
1692
1693         Reduce our dependence on coreutils when running Python tests
1694
1695         This patch introduces versions of the standard echo and cat utilities
1696         implemented in Python. They are probably missing features of their
1697         coreutils equivalents, but they can do what's necessary for our Python
1698         tests. This is useful on Windows, where these utilities typically
1699         aren't available.
1700
1701         Fixes <http://webkit.org/b/48883> executive_unittest relies on echo
1702         and cat utilities from coreutils, which are not present on Windows
1703
1704         Reviewed by Eric Seidel.
1705
1706         * Scripts/webkitpy/common/system/executive_unittest.py: Changed to use
1707         our Python-based echo and cat.
1708
1709         * Scripts/webkitpy/common/system/fileutils.py: Added.
1710         (make_stdout_binary): On Windows, puts sys.stdout into binary mode so
1711         that \n won't be translated into \r\n. I couldn't think of a good way
1712         to test this directly without touching the filesystem, but it is tested
1713         indirectly by echo_unittest.
1714
1715         * Scripts/webkitpy/test/cat.py: Added.
1716         (command_arguments): Returns a list for invoking cat with the given arguments.
1717         (main): Acts like a simplified version of the coreutils cat utility.
1718
1719         * Scripts/webkitpy/test/cat_unittest.py: Added.
1720         (CatTest.assert_cat): Runs cat with the given input and ensures the
1721         output matches the input.
1722         (CatTest.test_basic): Performs a simple test of cat.
1723         (CatTest.test_no_newline): Tests what happens when the input string
1724         doesn't have a trailing newline.
1725         (CatTest.test_unicode): Tests passing a unicode string to cat.
1726         (CatTest.test_as_command): Tests running cat as a separate command.
1727
1728         * Scripts/webkitpy/test/echo.py: Added.
1729         (command_arguments): Returns a list for invoking echo with the given arguments.
1730         (main): Acts like a simplified version of the coreutils echo utility.
1731
1732         * Scripts/webkitpy/test/echo_unittest.py: Added.
1733         (EchoTest.test_basic): Performs a simple test of echo.
1734         (EchoTest.test_no_newline): Tests passing -n to echo to suppress the
1735         trailing newline.
1736         (EchoTest.test_unicode): Tests passing unicode and non-unicode strings
1737         to echo.
1738         (EchoTest.test_argument_order): Tests what happens when -n is not the
1739         first argument.
1740         (EchoTest.test_empty_arguments): Tests what happens when you pass [] to
1741         echo.main.
1742         (EchoTest.test_no_arguments): Tests what happens when you call
1743         echo.main with no arguments.
1744         (EchoTest.test_as_command): Tests running echo as a separate command.
1745
1746 2010-11-04  Renata Hodovan  <reni@inf.u-szeged.hu>
1747
1748         Unreviewed: Add myself to the list of Committers.
1749
1750         * Scripts/webkitpy/common/config/committers.py:
1751
1752 2010-11-04  Andreas Kling  <kling@webkit.org>
1753
1754         Reviewed by Kenneth Rohde Christiansen.
1755
1756         [Qt] Hook into QWebPage::scrollRequested for resizes-to-contents mode
1757
1758         In RTC mode the QGraphicsWebView item is the size of the contents,
1759         scrolling works a bit differently  (we need to react to scrollRequested.)
1760
1761         Normally QGraphicsView will replay the last mouse event when scrolling,
1762         so to prevent WebKit from getting confused by this we temporarily make
1763         the QGraphicsView non-interactive.
1764
1765         * QtTestBrowser/webview.cpp:
1766         (WebViewGraphicsBased::setPage):
1767         (WebViewGraphicsBased::scrollRequested):
1768         * QtTestBrowser/webview.h:
1769
1770 2010-11-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
1771
1772         Reviewed by Andreas Kling.
1773
1774         Use OS(MAC_OS_X) rather than PLATFORM(MAC) in TestControllerQt
1775
1776         PLATFORM(MAC) is not defined for the Qt port, as it refers to
1777         the Mac-port, not the Mac OS X operating system.
1778
1779         * WebKitTestRunner/qt/TestControllerQt.cpp:
1780         (WTR::isExistingLibrary):
1781
1782 2010-11-04  Leonid Ebril  <leonid.ebril@nokia.com>
1783
1784         Reviewed by Andreas Kling.
1785
1786         [Qt] Adding iphone user agent string the user agent list for QtTestBrowser
1787         https://bugs.webkit.org/show_bug.cgi?id=48869
1788
1789         * QtTestBrowser/useragentlist.txt:
1790
1791 2010-11-03  Adam Roben  <aroben@apple.com>
1792
1793         Always use uppercase drive names in strings returned by abspath_to_uri
1794
1795         Some versions of cygpath use lowercase drive letters while others use
1796         uppercase, which makes it hard to test the output of code that uses
1797         cygpath.
1798
1799         Fixes <http://webkit.org/b/48914> webkitpy.common.system.path_unittest
1800         fails with Cygwin 1.5
1801
1802         Reviewed by Eric Seidel.
1803
1804         * Scripts/webkitpy/common/system/path.py:
1805         (cygpath): Updated the docstring to indicate that only absolute paths
1806         should be passed for now (though relative paths will work fine).
1807         (_Cygpath.convert): Upper-case the first letter of the converted Windows path.
1808
1809 2010-11-03  George Guo  <George.Guo@Nokia.com>
1810
1811         Reviewed by Andreas Kling.
1812
1813         [Qt] QtTestBrowser : set mmp rule pageddata in Symbian
1814         https://bugs.webkit.org/show_bug.cgi?id=48767
1815     
1816         Paging is needd on Symbian devices to support benchmarks tests like
1817         dromaeo.com and Celtic Kane that need a lot of memory to run  
1818
1819         * QtTestBrowser/QtTestBrowser.pro:
1820
1821 2010-11-03  Jenn Braithwaite  <jennb@chromium.org>
1822
1823         Reviewed by Dmitry Titov.
1824
1825         Chromium: Update resource tracking when moving a frame between documents
1826         https://bugs.webkit.org/show_bug.cgi?id=48363
1827
1828         * DumpRenderTree/chromium/WebViewHost.cpp:
1829         (WebViewHost::assignIdentifierToRequest):
1830         Always put resource id in map so we can make assumptions about its
1831         presence.
1832         (WebViewHost::removeIdentifierForRequest):
1833         Added.
1834         (WebViewHost::didFinishResourceLoad):
1835         (WebViewHost::didFailResourceLoad):
1836         Check existence of resource id before removing from map.
1837         * DumpRenderTree/chromium/WebViewHost.h:
1838
1839 2010-11-03  Victor Wang  <victorw@chromium.org>
1840
1841         Reviewed by Adam Barth.
1842
1843         [Chromium] update buildbot names in chromium rebaseline tool.
1844
1845         https://bugs.webkit.org/show_bug.cgi?id=48881
1846
1847         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
1848
1849 2010-11-03  Darin Adler  <darin@apple.com>
1850
1851         Updated Xcode projects by opening them with Xcode 3.2.4.
1852         Updated svn:ignore for Xcode projects.
1853
1854         * MiniBrowser/MiniBrowser.xcodeproj: Added property svn:ignore.
1855         * TestWebKitAPI/TestWebKitAPI.xcodeproj: Added property svn:ignore.
1856         * WebKitLauncher/WebKitLauncher.xcodeproj: Modified property svn:ignore.
1857         * WebKitLauncher/WebKitLauncher.xcodeproj/project.pbxproj: Updated with Xcode 3.2.4.
1858
1859 2010-11-03  Adam Roben  <aroben@apple.com>
1860
1861         Stop waiting for messages from the web process after a timeout elapses
1862
1863         TestController::runUntil now takes a timeout parameter that specifies
1864         how long to wait for the condition to become true. WebKitTestRunner
1865         takes a --timeout flag that specifies how long the timeout should be.
1866         run-webkit-tests passes this flag to WebKitTestRunner so its timeout
1867         will be similar to run-webkit-tests.
1868
1869         Fixes <http://webkit.org/b/43047> <rdar://problem/8365833>
1870         WebKitTestRunner waits forever if the web process crashes
1871
1872         Reviewed by Darin Adler and Anders Carlsson.
1873
1874         * Scripts/old-run-webkit-tests:
1875         (top level): Moved the GuardMalloc timeout adjustment here from
1876         readFromDumpToolWithTimer.
1877         (openDumpTool): Make WTR use a timeout similar to but slightly shorter
1878         than the one that was specified on the command line.
1879
1880         * WebKitTestRunner/TestController.cpp:
1881         (WTR::TestController::TestController): Initialize our timeout values.
1882         (WTR::TestController::initialize): Parse the --timeout flag and use it
1883         to modify our timeout values.
1884         (WTR::TestController::resetStateToConsistentValues): Changed to use a
1885         short timeout while waiting for the web process to reset and to return
1886         a boolean indicating whether we were able to reset the web process.
1887         Uses a 5-second timeout while waiting for the process to be reset.
1888         (WTR::TestController::runTest): Changed to return a boolean indicating
1889         whether we were able to reset the web process (and thus run the test).
1890         (WTR::TestController::runUntil): Call through to platformRunUntil.
1891
1892         (WTR::TestController::runTestingServerLoop):
1893         (WTR::TestController::run):
1894         Changed to bail if any test can't be run. This will cause the process
1895         to exit. (Unfortunately this will make run-webkit-tests think we
1896         crashed; see <http://webkit.org/b/48943>.)
1897
1898         * WebKitTestRunner/TestController.h: Added platformRunUntil,
1899         m_longTimeout, and m_shortTimeout.
1900
1901         * WebKitTestRunner/TestInvocation.cpp:
1902         (WTR::TestInvocation::invoke): Use a short timeout when waiting for the
1903         initial response and a long timeout when waiting for the test to
1904         complete. Check whether runUntil timed out and print an error message
1905         if so.
1906
1907         * WebKitTestRunner/mac/TestControllerMac.mm:
1908         (WTR::TestController::platformRunUntil): Renamed from runUntil. Pass
1909         [NSDate distantPast] to -[NSRunLoop runMode:beforeDate:] so that we
1910         won't block waiting for the run loop. Only loop until the timeout
1911         elapses.
1912
1913         * WebKitTestRunner/qt/TestControllerQt.cpp:
1914         (WTR::TestController::platformRunUntil): Renamed from runUntil. Added a
1915         FIXME about honoring the timeout.
1916
1917         * WebKitTestRunner/win/TestControllerWin.cpp:
1918         (WTR::TestController::platformRunUntil): Renamed from runUntil. Use
1919         ::MsgWaitForMultipleObjectsEx to implement the timeout. Changed to use
1920         ::PeekMessageW so that we don't block waiting for messages to become
1921         available.
1922
1923 2010-11-03  Adam Roben  <aroben@apple.com>
1924
1925         Add a plugin test that evaluates JS after removing the plugin element
1926         from the document
1927
1928         This test replaces platform/win/plugins/plugin-delayed-destroy.html.
1929         That test was made to prevent a crash very similar to this one, but
1930         unfortunately tested only the mechanism that prevented the crash and
1931         not whether the crash itself was prevented. Since WebKit2 uses a
1932         different mechanism to prevent the crash, the test was failing even
1933         though WebKit2 was not vulnerable to the crash. This new test crashes
1934         if there is no mechanism in place to prevent it and passes in both
1935         WebKit1 and WebKit2.
1936
1937         Fixes <http://webkit.org/b/46711> <rdar://problem/8485903>
1938         platform/win/plugins/plugin-delayed-destroy.html fails in WebKit2
1939
1940         Reviewed by Anders Carlsson.
1941
1942         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
1943         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1944         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
1945         * GNUmakefile.am:
1946         Added new file.
1947
1948         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1949         (pluginDeallocate): Make sure we delete the PluginTest object. This
1950         prevents a leak and also allows us to test the crash.
1951
1952         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
1953         (PluginTest::executeScript): Made this into a non-static member
1954         function.
1955
1956         (PluginTest::waitUntilDone):
1957         (PluginTest::notifyDone):
1958         Updated for changes to executeScript.
1959
1960         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Added executeScript.
1961
1962         * DumpRenderTree/TestNetscapePlugIn/Tests/EvaluateJSAfterRemovingPluginElement.cpp: Added.
1963         (EvaluateJSAfterRemovingPluginElement::EvaluateJSAfterRemovingPluginElement):
1964         Initialize ourselves and tell the test harness to wait.
1965         (EvaluateJSAfterRemovingPluginElement::NPP_DestroyStream): Remove our
1966         plugin element from the document, then execute some JavaScript. If
1967         WebKit does not have appropriate mechanisms in place, we'll be
1968         destroyed inside the first call to executeScript and crash on the
1969         second call.
1970
1971 2010-11-02  Stephen White  <senorblanco@chromium.org>
1972
1973         Reviewed by Tony Chang.
1974
1975         [chromium] Fix LayoutTestController UMRs.
1976         https://bugs.webkit.org/show_bug.cgi?id=48872
1977
1978         * DumpRenderTree/chromium/LayoutTestController.cpp:
1979         (LayoutTestController::LayoutTestController):
1980
1981 2010-11-03  Kent Tamura  <tkent@chromium.org>
1982
1983         Reviewed by Dimitri Glazkov.
1984
1985         REGRESSION: rebaseline-chromium-webkit-tests uses non-zero tolerance for
1986         image dup detection
1987         https://bugs.webkit.org/show_bug.cgi?id=48744
1988
1989         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
1990          - Make a function for option parsing for ease of test
1991          - Set 0 to options.tolerance
1992         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
1993          - Add a test for this change
1994
1995 2010-11-02  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
1996
1997         Unreviewed: Add myself to the list of Committers.
1998
1999         * Scripts/webkitpy/common/config/committers.py:
2000
2001 2010-11-02  Anders Carlsson  <andersca@apple.com>
2002
2003         Fix build.
2004
2005         * TestWebKitAPI/PlatformUtilities.cpp:
2006         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
2007
2008 2010-11-02  Sam Weinig  <sam@webkit.org>
2009
2010         Reviewed by Anders Carlsson.
2011
2012         Add a way to send startup messages on the context which can be posted when a process launches
2013         <rdar://problem/8617928>
2014         https://bugs.webkit.org/show_bug.cgi?id=48838
2015
2016         * MiniBrowser/mac/WebBundle/WebBundleMain.m:
2017         (WKBundleInitialize):
2018         * TestWebKitAPI/InjectedBundleController.cpp:
2019         (TestWebKitAPI::InjectedBundleController::initialize):
2020         (TestWebKitAPI::InjectedBundleController::didReceiveMessage):
2021         * TestWebKitAPI/InjectedBundleController.h:
2022         * TestWebKitAPI/InjectedBundleMain.cpp:
2023         (WKBundleInitialize):
2024         * TestWebKitAPI/PlatformUtilities.cpp:
2025         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
2026         * WebKitTestRunner/InjectedBundle/InjectedBundleMain.cpp:
2027         (WKBundleInitialize):
2028         Update implementations of WKBundleInitialize to take an initial userData
2029         argument.  Change TestWebKitAPI to use the new initial userData to initialize
2030         each test's bundle.
2031
2032 2010-11-02  Benjamin Kalman  <kalman@google.com>
2033
2034         Reviewed by Ojan Vafai.
2035
2036         new-run-webkit-tests doesn't strip "LayoutTests/" from prefix, unlike old-run-webkit-tests
2037         https://bugs.webkit.org/show_bug.cgi?id=48794
2038
2039         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: Strip the "LayoutTests/" prefix from test argument paths.
2040
2041 2010-11-02  Adam Roben  <aroben@apple.com>
2042
2043         Skip webkitpy.layout_tests.run_webkit_tests_unittest.MainTest on Cygwin
2044         Python 2.5.x
2045
2046         It is known to hang on that version of Python. See
2047         <http://webkit.org/b/48614>.
2048
2049         Reviewed by Adam Barth.
2050
2051         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py: Use
2052         skip_if to skip MainTest on Cygwin Python 2.5.x.
2053
2054         * Scripts/webkitpy/test/skip.py: Added.
2055         (skip_if): If the passed-in condition is false, find all the test_*
2056         methods of the given class and replace them with a function that just
2057         logs that we're skipping these tests. This is loosely based on the
2058         unittest.skip_if decorator added in Python 3.1.
2059         (_skipped_method): Returns a function that just logs that the tests are
2060         being skipped. This is used in place of the actual test_* functions.
2061
2062         * Scripts/webkitpy/test/skip_unittest.py: Added.
2063         (SkipTest.setUp): Create a logger for ourselves and give it a handler
2064         that logs to a private stream.
2065         (SkipTest.tearDown): Reset the logger.
2066         (SkipTest.create_fixture_class): Returns a class that we use to test
2067         skip_if. It has a single test_* method, test_foo, that just calls
2068         through to a callback.
2069         (SkipTest.foo_callback): Record that test_foo was called.
2070         (SkipTest.test_skip_if_false): Pass skip_if a False condition and test
2071         that test_foo does get called.
2072         (SkipTest.test_skip_if_true): Pass skip_if a True condition and test
2073         that test_foo does not get called and the appropriate message gets
2074         logged.
2075
2076 2010-11-02  Adam Barth  <abarth@webkit.org>
2077
2078         Reviewed by Eric Seidel.
2079
2080         webkit-patch should tell check-webkit-style which files were changed so
2081         check-webkit-style doesn't have to stat the whole working copy again
2082         https://bugs.webkit.org/show_bug.cgi?id=48792
2083
2084         * Scripts/webkitpy/tool/mocktool.py:
2085         * Scripts/webkitpy/tool/steps/checkstyle.py:
2086
2087 2010-11-02  Robert Kroeger  <rjkroege@chromium.org>
2088
2089         Reviewed by James Robinson.
2090
2091         [Chromium/DRT] Make EventSendingController honour leapForward for touch events.
2092         https://bugs.webkit.org/show_bug.cgi?id=48777
2093
2094         * DumpRenderTree/chromium/EventSender.cpp:
2095         (EventSender::sendCurrentTouchEvent):
2096
2097 2010-11-02  Adam Roben  <aroben@apple.com>
2098
2099         Only track resource identifiers in DRT when dumpResourceLoadCallbacks
2100         is on
2101
2102         This reverts Windows to our pre-r71097 behavior. That patch made us
2103         track all resource identifiers, including the main resource, so the
2104         main resource's URL started appearing in test output instead of
2105         "<unknown>". Arguably having the main resource's URL is better, but all
2106         other platforms print "<unknown>" and we want to match.
2107
2108         Fixes <http://webkit.org/b/48837> <rdar://problem/8620351> REGRESSION
2109         (r71097): Many http tests failing on Windows
2110
2111         Reviewed by Anders Carlsson.
2112
2113         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
2114         (ResourceLoadDelegate::identifierForInitialRequest): Don't add the
2115         identifier to the URL map if we're not supposed to dump resource load
2116         callbacks.
2117         (ResourceLoadDelegate::removeIdentifierForRequest): Always remove the
2118         identifier from the URL map even if we're already "done". There's no
2119         point in keeping out-of-date identifiers around.
2120
2121 2010-11-01  Jenn Braithwaite  <jennb@chromium.org>
2122
2123         Reviewed by Adam Roben.
2124
2125         Windows: Update resource tracking when moving a frame between documents
2126         https://bugs.webkit.org/show_bug.cgi?id=48364
2127
2128         * DumpRenderTree/win/DumpRenderTree.cpp:
2129         (createWebViewAndOffscreenWindow):
2130         (main):
2131         Give each WebView its own ResourceLoadDelegate instance in order to
2132         make assertions about resource ids on a particular WebView.
2133         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
2134         (ResourceLoadDelegate::identifierForInitialRequest):
2135         Always add id to the map.
2136         (ResourceLoadDelegate::removeIdentifierForRequest):
2137         Added.
2138         (ResourceLoadDelegate::willSendRequest):
2139         (ResourceLoadDelegate::didReceiveAuthenticationChallenge):
2140         (ResourceLoadDelegate::didReceiveResponse):
2141         (ResourceLoadDelegate::didFinishLoadingFromDataSource):
2142         (ResourceLoadDelegate::didFailLoadingWithError):
2143         (ResourceLoadDelegate::descriptionSuitableForTestResult):
2144         Replace static descriptionSuitableForTestResult with static member function to access identifier map.
2145         * DumpRenderTree/win/ResourceLoadDelegate.h:
2146         (ResourceLoadDelegate::urlMap):
2147         Moved within class so that each WebView has its own id map.
2148
2149 2010-11-01  Ojan Vafai  <ojan@chromium.org>
2150
2151         Reviewed by Dimitri Glazkov.
2152
2153         remove debug code from run_webkit_tests.py
2154         https://bugs.webkit.org/show_bug.cgi?id=48800
2155
2156         Remove temporary debug code and make --master-name required
2157         if --test-results-server is set now that all clients set
2158         --master-name.
2159
2160         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2161
2162 2010-11-01  Tony Chang  <tony@chromium.org>
2163
2164         Reviewed by Kent Tamura.
2165
2166         add plugin event logging for linux
2167         https://bugs.webkit.org/show_bug.cgi?id=48779
2168
2169         This is taken from Chromium's fork of the layout test plugin:
2170         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
2171
2172         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
2173         (handleEventX11):
2174         (NPP_HandleEvent):
2175
2176 2010-11-01  Dirk Pranke  <dpranke@chromium.org>
2177
2178         Reviewed by James Robinson.
2179
2180         new-run-webkit-tests: use DRT, child-processes=1 for GPU tests by default
2181         https://bugs.webkit.org/show_bug.cgi?id=48790
2182
2183         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2184         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2185
2186 2010-11-01  Adam Barth  <abarth@webkit.org>
2187
2188         Reviewed by Eric Seidel.
2189
2190         Teach check-webkit-style how to accept a list of files to diff on the
2191         command line
2192         https://bugs.webkit.org/show_bug.cgi?id=48784
2193
2194         In a future patch, webkit-patch will use this option to improve
2195         performance.  I'm landing this in two pieces to avoid causing a version
2196         skew problem for the style-bot.
2197
2198         * Scripts/check-webkit-style:
2199         * Scripts/webkitpy/style/optparser.py:
2200         * Scripts/webkitpy/style/optparser_unittest.py:
2201         * Scripts/webkitpy/style_references.py:
2202
2203 2010-11-01  Anders Carlsson  <andersca@apple.com>
2204
2205         Reviewed by John Sullivan.
2206
2207         Tear down the related WebProcessProxy when a WebContext is deallocated
2208         https://bugs.webkit.org/show_bug.cgi?id=48769
2209
2210         * TestWebKitAPI/Tests/WebKit2/FailedLoad.cpp:
2211         (TestWebKitAPI::didFailProvisionalLoadWithErrorForFrame):
2212         We don't support empty URLs anymore, update test to expect a null URL instead.
2213
2214 2010-11-01  Søren Gjesse  <sgjesse@chromium.org>
2215
2216         Reviewed by Andreas Kling.
2217
2218         Fix warning when compiling the chromium port of DumpRenderShell
2219         with clang.
2220         https://bugs.webkit.org/show_bug.cgi?id=48414
2221
2222         * DumpRenderTree/chromium/TestShell.h:
2223         (TestShell::javaScriptFlagsForLoad):
2224
2225 2010-11-01  Adam Roben  <aroben@apple.com>
2226
2227         Fix typo from r71022
2228
2229         * BuildSlaveSupport/build.webkit.org-config/config.json:
2230
2231 2010-11-01  Adam Roben  <aroben@apple.com>
2232
2233         Trigger the Windows Release WebKit2 tests when a Release build
2234         finishes, not when a Debug build finishes
2235
2236         Fixes <http://webkit.org/b/48754> Windows Release WebKit2 tests are
2237         triggered at the wrong time
2238
2239         Reviewed by Sam Weinig.
2240
2241         * BuildSlaveSupport/build.webkit.org-config/config.json: Fixed
2242         triggerable name and trigger.
2243
2244 2010-11-01  Mario Sanchez Prada  <msanchez@igalia.com>
2245
2246         Unreviewed. Adding my IRC nickname to the list of committers.
2247
2248         * Scripts/webkitpy/common/config/committers.py:
2249
2250 2010-10-31  Robert Hogan  <robert@webkit.org>
2251
2252         Reviewed by Antonio Gomes.
2253
2254         [Qt] [Gtk] Plug-ins having upper case in mime type are failing to load
2255
2256         Qt and Gtk are case-sensitive when storing the declared mime-type
2257         of plugins. Since plugin mime-types are lowercased prior to searching
2258         for them in the plugin database, ensure they are loaded with the
2259         mime-type in lower case too.
2260
2261         Change the test netscape plugin to declare its mimetype in sentence
2262         case so that the correct behaviour is enforced.
2263
2264         https://bugs.webkit.org/show_bug.cgi?id=36815
2265
2266         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
2267         (NP_GetMIMEDescription):
2268
2269 2010-10-31  Robert Hogan  <robert@webkit.org>
2270
2271         Reviewed by Antonio Gomes.
2272
2273         [Qt] Support nodesFromRect in DRT
2274
2275         https://bugs.webkit.org/show_bug.cgi?id=48716
2276
2277         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2278         (LayoutTestController::nodesFromRect):
2279         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2280
2281 2010-10-30  Andreas Kling  <kling@webkit.org>
2282
2283         Reviewed by Antonio Gomes.
2284
2285         [Qt] QtTestBrowser: Switching view type moves the embedded inspector
2286         https://bugs.webkit.org/show_bug.cgi?id=48705
2287
2288         Reinsert the embedded inspector into the splitter after changing
2289         between QWebView/QGraphicsWebView.
2290
2291         * QtTestBrowser/launcherwindow.cpp:
2292         (LauncherWindow::init):
2293         (LauncherWindow::initializeView):
2294         * QtTestBrowser/webinspector.h:
2295         (WebInspector::WebInspector):
2296
2297 2010-10-28  Antonio Gomes  <agomes@rim.com>
2298
2299         Reviewed by Ojan Vafai.
2300
2301         Needs a "LinuxEditingBehavior", perhaps with a better name
2302         https://bugs.webkit.org/show_bug.cgi?id=36627
2303
2304         Adding support to Mac's, GTK+'s, Windows' and Chromium's LayoutTestController class to test the newly introduced Unix editing behavior.
2305
2306         * DumpRenderTree/chromium/LayoutTestController.cpp:
2307         (LayoutTestController::setEditingBehavior):
2308         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2309         (LayoutTestController::setEditingBehavior):
2310         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2311         (LayoutTestController::setEditingBehavior):
2312         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2313         (LayoutTestController::setEditingBehavior):
2314
2315 2010-10-29  Andreas Kling  <kling@webkit.org>
2316
2317         Reviewed by Kenneth Rohde Christiansen.
2318
2319         [Qt] QtTestBrowser: Add keyboard shortcut to toggle full screen (F11)
2320         https://bugs.webkit.org/show_bug.cgi?id=48695
2321
2322         * QtTestBrowser/launcherwindow.cpp:
2323         (LauncherWindow::createChrome):
2324
2325 2010-10-29  Andreas Kling  <kling@webkit.org>
2326
2327         Reviewed by Kenneth Rohde Christiansen.
2328
2329         [Qt] QtTestBrowser: Fix uninitialized read in FpsTimer
2330         https://bugs.webkit.org/show_bug.cgi?id=48675
2331
2332         FpsTimer::m_timer was never initialized and passed to QObject::killTimer()
2333         on startup with in -graphicsbased mode.
2334
2335         * QtTestBrowser/fpstimer.cpp:
2336         (FpsTimer::FpsTimer):
2337         (FpsTimer::numFrames):
2338         (FpsTimer::stop):
2339         (FpsTimer::timerEvent):
2340         * QtTestBrowser/fpstimer.h:
2341
2342 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
2343
2344         Reviewed by Tony Chang.
2345
2346         new-run-webkit-tests: change TestResults to be serializable
2347
2348         In preparation for changing new-run-webkit-tests from
2349         multithreaded to multiprocess, we need to make sure the data
2350         going between the threads is easily serialized over a socket.
2351
2352         This change adds serialization/pickling for the TestResults and
2353         TestFailure objects (using cPickle).
2354
2355         The TestFailure objects included a "has_wdiff" flag for Text
2356         results, but the flag wasn't being used, so I've removed it,
2357         simplifying the state to basically a set of enum objects with
2358         associated methods.
2359
2360         https://bugs.webkit.org/show_bug.cgi?id=48616
2361
2362         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2363         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
2364         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
2365         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
2366         * Scripts/webkitpy/layout_tests/layout_package/test_results.py: Added.
2367         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:Added. 
2368         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2369         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2370
2371 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
2372
2373         Reviewed by Ojan Vafai.
2374
2375         new-run-webkit-tests: need to provide separate dashboard results for GPU tests
2376         https://bugs.webkit.org/show_bug.cgi?id=48687
2377
2378         Modify the steps to upload the JSON files to the dashboards so
2379         that the GPU tests don't conflict with the regular tests on a
2380         port. We do this by modifying the --builder-name parameter to
2381         append " - GPU", which should cause the app to treat the results
2382         as a completely new builder. This is a little non-obvious, but
2383         keeps us from having to restructure the app.
2384
2385         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2386         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2387         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2388
2389 2010-10-29  Ojan Vafai  <ojan@chromium.org>
2390
2391         Reviewed by Tony Chang.
2392
2393         [chromium] add debug logging to help diagnose flakiness dashboard issues
2394         https://bugs.webkit.org/show_bug.cgi?id=48657
2395
2396         The appengine app thinks it's getting empty files uploaded. Add some logging
2397         to see if new-run-webkit-tests agrees.
2398         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2399
2400 2010-10-29  Dimitri Glazkov  <dglazkov@chromium.org>
2401
2402         Reviewed by Ojan Vafai.
2403
2404         [Chromium/DRT] Add master-name flag to new-run-webkit-tests.
2405         https://bugs.webkit.org/show_bug.cgi?id=48649
2406
2407         * BuildSlaveSupport/build.webkit.org-config/master.cfg: Added flag.
2408
2409 2010-10-29  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
2410
2411         Reviewed by Martin Robinson.
2412
2413         Enable popup window in GtkLauncher
2414         https://bugs.webkit.org/show_bug.cgi?id=48335
2415
2416         The GtkLauncher application does not support opening new window when
2417         a link with "target=_blank" is clicked or similar call to
2418         window.open(). Instead, GtkLauncher does nothing which breaks
2419         navigation of some websites.
2420
2421         * GtkLauncher/main.c:
2422         (activate_uri_entry_cb):
2423         (update_title):
2424         (link_hover_cb):
2425         (notify_title_cb):
2426         (notify_load_status_cb):
2427         (notify_progress_cb):
2428         (destroy_cb):
2429         (go_back_cb):
2430         (go_forward_cb):
2431         (create_web_view_cb):
2432         (web_view_ready_cb):
2433         (close_web_view_cb):
2434         (create_browser):
2435         (create_statusbar):
2436         (create_toolbar):
2437         (create_window):
2438         (main):
2439
2440 2010-10-29  Adam Roben  <aroben@apple.com>
2441
2442         Teach check-webkit-style about WebKit2's idiosyncracies
2443
2444         Fixes <http://webkit.org/b/48638> Style bot complains about a number
2445         of WebKit2 conventions
2446
2447         Reviewed by Anders Carlsson.
2448
2449         * Scripts/webkitpy/style/checker.py: Excluded some rules for various
2450         WebKit2-related files. Also updated the excluded rules for
2451         WebKitAPITest to match the current code.
2452
2453 2010-10-29  Csaba Osztrogonác  <ossy@webkit.org>
2454
2455         Reviewed by Adam Roben and David Kilzer.
2456
2457         Fix and cleanup of build systems
2458         https://bugs.webkit.org/show_bug.cgi?id=48342
2459
2460         * Scripts/build-webkit:
2461          - Remove unnecessary ENABLE_SANDBOX option.
2462          - Add ENABLE_FULLSCREEN_API option.
2463
2464 2010-10-28  Adam Roben  <aroben@apple.com>
2465
2466         Switch the Windows WebKit2 bot to the Release configuration
2467
2468         We only have one machine testing WebKit2 on Windows right now, and
2469         Debug is just too slow for it to keep up.
2470
2471         Fixes (hopefully!) <http://webkit.org/b/48615> Windows WebKit2 bot is
2472         always way behind
2473
2474         Reviewed by Jon Honeycutt.
2475
2476         * BuildSlaveSupport/build.webkit.org-config/config.json:
2477
2478 2010-10-29  Leandro Gracia Gil  <leandrogracia@google.com>
2479
2480         Reviewed by Jeremy Orlow.
2481
2482         Added a second parameter to setMockSpeechInputResult for
2483         the language used in speech input.
2484         https://bugs.webkit.org/show_bug.cgi?id=47089
2485
2486         * DumpRenderTree/LayoutTestController.cpp:
2487         (setMockSpeechInputResultCallback):
2488         * DumpRenderTree/LayoutTestController.h:
2489         * DumpRenderTree/chromium/LayoutTestController.cpp:
2490         (LayoutTestController::setMockSpeechInputResult):
2491         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2492         (LayoutTestController::setMockSpeechInputResult):
2493         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2494         (LayoutTestController::setMockSpeechInputResult):
2495         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2496         (LayoutTestController::setMockSpeechInputResult):
2497         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2498         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2499         (LayoutTestController::setMockSpeechInputResult):
2500         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2501         (LayoutTestController::setMockSpeechInputResult):
2502         (LayoutTestController::pageSizeAndMarginsInPixels):
2503
2504 2010-10-28  Ojan Vafai  <ojan@chromium.org>
2505
2506         Reviewed by Tony Chang.
2507
2508         [chromium] add a result-small.json file for the test dashboard
2509         https://bugs.webkit.org/show_bug.cgi?id=48547
2510
2511         Output both a results.json file and a results-small.json file.
2512         The dashboard will load results-small.json by default so it loads faster.
2513
2514         * TestResultServer/model/jsonresults.py:
2515         * TestResultServer/model/jsonresults_unittest.py:
2516         Added a bunch of sys.path hackery. Unforunately, this uses hardcoded
2517         paths. That obviously needs to be fixed, but at least this way it
2518         clearly documents what paths are necessary.
2519
2520 2010-10-28  Eric Seidel  <eric@webkit.org>
2521
2522         Reviewed by Adam Barth.
2523
2524         webkit-patch upload calls changed_files more often than it should
2525         https://bugs.webkit.org/show_bug.cgi?id=48567
2526
2527         Passing changed_files around everywhere isn't a very elegant solution
2528         but it's the one we have for the moment.  I think keeping an explicit
2529         cache on Checkout (or making StepState() a real class) is a better
2530         long-term option.
2531
2532         Previously bug_id_for_this_commit was calling changed_files and the
2533         result was never getting cached on the state.  Now we're explicitly
2534         caching the result on the state and passing that to the bug_id_for_this_commit call.
2535
2536         I looked into building unit tests for this.  Doing so would require
2537         using a real Checkout object with a MockSCM and overriding the appropriate
2538         calls on SCM to count how often we're stating the file system.
2539         That's a useful set of tests to build for a separate change.
2540
2541         * Scripts/webkitpy/common/checkout/api.py:
2542         * Scripts/webkitpy/tool/commands/download.py:
2543         * Scripts/webkitpy/tool/commands/upload.py:
2544         * Scripts/webkitpy/tool/mocktool.py:
2545
2546 2010-10-28  Eric Seidel  <eric@webkit.org>
2547
2548         Reviewed by Adam Barth.
2549
2550         Make suggest-reviewers slightly faster
2551         https://bugs.webkit.org/show_bug.cgi?id=48562
2552
2553         Add @memoized to one more common call.
2554
2555         * Scripts/webkitpy/common/checkout/api.py:
2556
2557 2010-10-28  Eric Seidel  <eric@webkit.org>
2558
2559         Reviewed by Adam Barth.
2560
2561         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is still very slow
2562         https://bugs.webkit.org/show_bug.cgi?id=48536
2563
2564         This patch makes it about 40 seconds faster, but it still
2565         takes 1:40.  This will require more refinement.
2566
2567         The suggested reviewers list appears to be the same.
2568
2569         I think the next step may be to have it stop the search after
2570         5 reviewers are found.  We never want to suggest 30 people.
2571
2572         * Scripts/webkitpy/common/checkout/scm.py:
2573          - Using --remove-empty to theoretically stop lookups past
2574            when a file is removed.  I'm not entirely clear that the option
2575            does what it says it does.  Example:
2576            git log --pretty=format:%H -5 --remove-empty -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
2577            returns only one commit
2578            vs.
2579            git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
2580            which returns 5.  I was not aware that wx files were ever removed from the repository?
2581
2582 2010-10-28  Kinuko Yasuda  <kinuko@chromium.org>
2583
2584         Reviewed by David Levin.
2585
2586         [Chromium] Support FileSystem in chromium DRT
2587         https://bugs.webkit.org/show_bug.cgi?id=47643
2588
2589         * DumpRenderTree/chromium/WebViewHost.cpp:
2590         (WebViewHost::openFileSystem): Added.
2591         * DumpRenderTree/chromium/WebViewHost.h:
2592         (WebViewHost::openFileSystem): Added.
2593
2594 2010-10-28  Sam Weinig  <sam@webkit.org>
2595
2596         Reviewed by Anders Carlsson.
2597
2598         WKURLRefs should be allowed to be null
2599         <rdar://problem/8575621>
2600         https://bugs.webkit.org/show_bug.cgi?id=48535
2601
2602         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
2603         * TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp:
2604         (TestWebKitAPI::didStartProvisionalLoadForFrame):
2605         (TestWebKitAPI::didCommitLoadForFrame):
2606         (TestWebKitAPI::didFinishLoadForFrame):
2607         Test that URLs are null pointers when unset.
2608
2609 2010-10-28  Chris Fleizach  <cfleizach@apple.com>
2610
2611         Reviewed by Adele Peterson.
2612
2613         AX: multi select group option does not handle setting of AXSelectedChildren correctly
2614         https://bugs.webkit.org/show_bug.cgi?id=48464
2615
2616         Add support for querying information about selected children to DRT, including:
2617            selectedChildAtIndex
2618            selectedChildrenCount
2619            setSelectedChild
2620
2621         * DumpRenderTree/AccessibilityUIElement.cpp:
2622         (selectedChildAtIndexCallback):
2623         (setSelectedChildCallback):
2624         (selectedChildrenCountCallback):
2625         (AccessibilityUIElement::setSelectedChild):
2626         (AccessibilityUIElement::selectedChildrenCount):
2627         (AccessibilityUIElement::selectedChildAtIndex):
2628         (AccessibilityUIElement::getJSClass):
2629         * DumpRenderTree/AccessibilityUIElement.h:
2630         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
2631         (AccessibilityUIElement::selectedChildAtIndex):
2632         (AccessibilityUIElement::selectedChildrenCount):
2633         (AccessibilityUIElement::setSelectedChild):
2634
2635 2010-10-27  Sam Weinig  <sam@webkit.org>
2636
2637         Reviewed by Anders Carlsson.
2638
2639         Add WebKit2 API for window feature getter/setters
2640         <rdar://problem/8590373>
2641         https://bugs.webkit.org/show_bug.cgi?id=48496
2642
2643         * MiniBrowser/mac/BrowserWindowController.m:
2644         (-[BrowserWindowController awakeFromNib]):
2645         * MiniBrowser/win/BrowserView.cpp:
2646         (BrowserView::create):
2647         * WebKitTestRunner/TestController.cpp:
2648         (WTR::createOtherPage):
2649         (WTR::TestController::initialize):
2650
2651 2010-10-28  Søren Gjesse  <sgjesse@chromium.org>
2652
2653         Reviewed by Tony Chang.
2654
2655         Added support for the DumpRenderTree flags --multiple-loads and --js-flags to the Python test runner.
2656         https://bugs.webkit.org/show_bug.cgi?id=48236
2657
2658         * Scripts/webkitpy/layout_tests/port/chromium.py:
2659         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2660
2661 2010-10-28  Sergio Villar Senin  <svillar@igalia.com>
2662
2663         Reviewed by Martin Robinson.
2664
2665         [GTK] WebKitWebFrame's load-status is not properly notified to the tests
2666         https://bugs.webkit.org/show_bug.cgi?id=48048
2667
2668         DRT now listens to WebKitWebFrame load-status signals for each
2669         created frame instead of the load-status signal of the
2670         WebKitWebView that only notifies about changes in the main frame.
2671
2672         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2673         (webViewLoadFinished):
2674         (webFrameLoadStatusNotified):
2675         (frameCreatedCallback):
2676         (createWebView):
2677         (main):
2678
2679 2010-10-28  Tony Chang  <tony@chromium.org>
2680
2681         Reviewed by Kent Tamura.
2682
2683         [chromium] fix textInputController.{selectedRange,markedRange}
2684         https://bugs.webkit.org/show_bug.cgi?id=48487
2685
2686         * DumpRenderTree/chromium/TextInputController.cpp:
2687         (TextInputController::markedRange): Return arrays of ints, rather than a string
2688         (TextInputController::selectedRange): Ditto.
2689
2690 2010-10-27  Ojan Vafai  <ojan@chromium.org>
2691
2692         Reviewed by Tony Chang.
2693
2694         [chromium] add a master-name flag to new-run-webkit-tests
2695         https://bugs.webkit.org/show_bug.cgi?id=48488
2696
2697         The test results server now allows adding a master name to
2698         the uploaded files. This lets us distinguish bots that have
2699         the same name, but are on different masters.
2700
2701         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2702
2703 2010-10-28  Adam Roben  <aroben@apple.com>
2704
2705         Don't append a newline to the test output if the frame has no document
2706         element in WebKitTestRunner
2707
2708         Fixes <http://webkit.org/b/48526> Extra trailing newline when running
2709         plugins/document-open.html in WebKitTestRunner
2710
2711         Reviewed by Anders Carlsson.
2712
2713         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2714         (WTR::toJS): Added. Turns a UTF-8 C string into a JSStringRef.
2715         (WTR::hasDocumentElement): Added. Uses the JSC API to figure out
2716         whether the frame has a document element.
2717         (WTR::dumpFrameText): Match DRT by bailing (rather than appending an
2718         empty string and a newline) if the frame has no document element.
2719
2720 2010-10-28  Adam Roben  <aroben@apple.com>
2721
2722         Skip npn-invalidate-rect-invalidates-window.html on headless XP
2723         machines
2724
2725         TestNetscapePlugIn never receives a WM_PAINT message on headless XP
2726         machines, so this test times out. Fixing the test is covered by
2727         <http://webkit.org/b/48333>.
2728
2729         * Scripts/old-run-webkit-tests: Skip
2730         npn-invalidate-rect-invalidates-window.html on Windows if accelerated
2731         compositing support is disabled, which likely means we're on a headless
2732         XP machine.
2733
2734 2010-10-28  Kimmo Kinnunen  <kimmok@iki.fi>
2735
2736         Adding myself as a committer.
2737
2738         * Scripts/webkitpy/common/config/committers.py:
2739
2740 2010-10-27  Eric Seidel  <eric@webkit.org>
2741
2742         Reviewed by Adam Barth.
2743
2744         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is very slow
2745         https://bugs.webkit.org/show_bug.cgi?id=48500
2746
2747         This doesn't fix the problem, but it makes things slightly better.
2748         Each git svn find-rev call takes about .25 seconds on my desktop.
2749         This patch uses a new memoized class to avoid those calls when possible.
2750
2751         The real slowness is still git log on some files, like:
2752         git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
2753         I'm not yet sure how to make the pathological git logs better.
2754
2755         * Scripts/webkitpy/common/checkout/scm.py:
2756         * Scripts/webkitpy/common/memoized.py: Added.
2757         * Scripts/webkitpy/common/memoized_unittest.py: Added.
2758
2759 2010-10-27  Eric Seidel  <eric@webkit.org>
2760
2761         Reviewed by Ojan Vafai.
2762
2763         EWS bots should not use --quiet when running build-webkit
2764         https://bugs.webkit.org/show_bug.cgi?id=48482
2765
2766         --quiet is only correct when the sub-process does the error reporting.
2767         In the case of _can_build() the parent process is reporting the error.
2768         We'd like the full build log at queues.webkit.org so someone can look
2769         at the log and understand why the EWS is failing to build trunk.
2770
2771         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2772         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
2773         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
2774         * Scripts/webkitpy/tool/mocktool.py:
2775
2776 2010-10-27  Brian Weinstein  <bweinstein@apple.com>
2777
2778         More Windows build fixage. Rename a variable that was named string.
2779
2780         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2781         (WTR::hasPrefix):
2782
2783 2010-10-27  Ojan Vafai  <ojan@chromium.org>
2784
2785         Reviewed by Dimitri Glazkov.
2786
2787         [chromium] Make the test results server store which master the bot is on
2788         https://bugs.webkit.org/show_bug.cgi?id=48478
2789
2790         The chromium bots recently changed so that there are multiple slaves with
2791         the same name on different masters. Up till now, the test results server
2792         assumed slave names were unique. Adds a master field to the file in order
2793         to distinguish.
2794
2795         Also, for files that currently lack a master or testtype, set them appropriately.
2796
2797         * TestResultServer/handlers/testfilehandler.py:
2798         * TestResultServer/index.yaml:
2799         * TestResultServer/model/jsonresults.py:
2800         * TestResultServer/model/testfile.py:
2801         * TestResultServer/templates/showfilelist.html:
2802         * TestResultServer/templates/uploadform.html:
2803
2804 2010-10-26  Darin Adler  <darin@apple.com>
2805
2806         Reviewed by Sam Weinig.
2807
2808         WebKitTestRunner needs to support layoutTestController.dumpBackForwardList
2809         https://bugs.webkit.org/show_bug.cgi?id=42322
2810         rdar://problem/8193631
2811
2812         WebKitTestRunner needs to support layoutTestController.clearBackForwardList
2813         https://bugs.webkit.org/show_bug.cgi?id=42333
2814         rdar://problem/8193643
2815
2816         * WebKitTestRunner/Configurations/InjectedBundle.xcconfig:
2817         Renamed the product to WebKitTestRunnerInjectedBundle to avoid
2818         name conflicts in the build directory.
2819
2820         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
2821         Added dumpBackForwardList and clearBackForwardList.
2822
2823         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
2824         (WTR::InjectedBundle::InjectedBundle): Removed initialization of m_mainPage.
2825         (WTR::InjectedBundle::didCreatePage): Removed unneeded underscore.
2826         (WTR::InjectedBundle::willDestroyPage): Ditto.
2827         (WTR::InjectedBundle::didReceiveMessage): Ditto.
2828         (WTR::InjectedBundle::initialize): Ditto.
2829         (WTR::InjectedBundle::didCreatePage): Changed code to use m_pages
2830         instead of m_mainPage and m_otherPages.
2831         (WTR::InjectedBundle::willDestroyPage): Ditto.
2832         (WTR::InjectedBundle::page): Ditto.
2833         (WTR::InjectedBundle::beginTesting): Ditto.
2834         (WTR::InjectedBundle::done): Ditto.
2835         (WTR::InjectedBundle::closeOtherPages): Ditto.
2836         (WTR::InjectedBundle::dumpBackForwardListsForAllPages): Added.
2837
2838         * WebKitTestRunner/InjectedBundle/InjectedBundle.h: Changed
2839         page function to longer be inline, and pageCount function
2840         to use m_pages. Added dumpBackForwardListsForAllPages, and
2841         removed some underscores. Replaced m_mainPage and m_otherPageas
2842         with m_pages.
2843
2844         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2845         (WTR::adoptWK): Added.
2846         (WTR::hasPrefix): Added.
2847         (WTR::InjectedBundlePage::reset): Added code to set up
2848         m_previousTestBackForwardListItem.
2849         (WTR::InjectedBundlePage::dump): Added code to call
2850         dumpBackForwardListsForAllPages.
2851         (WTR::compareByTargetName): Added.
2852         (WTR::dumpBackForwardListItem): Added.
2853         (WTR::InjectedBundlePage::dumpBackForwardList): Added.
2854
2855         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h: Added
2856         dumpBackForwardList and m_previousTestBackForwardListItem.
2857
2858         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
2859         (WTR::LayoutTestController::LayoutTestController): Initialize
2860         m_shouldDumpBackForwardListsForAllWindows to false.
2861         (WTR::LayoutTestController::clearBackForwardList): Added.
2862
2863         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
2864         Added dumpBackForwardList, clearBackForwardList,
2865         shouldDumpBackForwardListsForAllWindows, and
2866         m_shouldDumpBackForwardListsForAllWindows.
2867
2868         * WebKitTestRunner/TestController.cpp:
2869         (WTR::TestController::initialize): Set up
2870         didReceiveSynchronousMessageFromInjectedBundle.
2871         (WTR::TestController::didReceiveSynchronousMessageFromInjectedBundle):
2872         Added.
2873         * WebKitTestRunner/TestController.h: Ditto.
2874
2875         * WebKitTestRunner/TestInvocation.cpp:
2876         (WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):
2877         Added.
2878         * WebKitTestRunner/TestInvocation.h: Ditto.
2879
2880         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
2881         Renamed bundle to WebKitTestRunnerInjectedBundle (see above).
2882         * WebKitTestRunner/mac/TestControllerMac.mm:
2883         (WTR::TestController::initializeInjectedBundlePath): Ditto.
2884
2885 2010-10-27  Chris Rogers  <crogers@google.com>
2886
2887         Reviewed by Chris Marrin.
2888
2889         Add ENABLE_WEB_AUDIO feature enable flag (initially disabled) to build-webkit
2890         https://bugs.webkit.org/show_bug.cgi?id=48279
2891
2892         * Scripts/build-webkit:
2893
2894 2010-10-27  Eric Seidel  <eric@webkit.org>
2895
2896         Unreviewed.
2897
2898         build-webkit should collect Visual Studio Express logs and display them
2899         https://bugs.webkit.org/show_bug.cgi?id=39199
2900
2901         It turns out my previous patch wasn't actually working on the win-ews
2902         machine.  So I've fixed my mistakes from before.
2903
2904         * Scripts/build-webkit:
2905          - Windows VSE builds change the CWD while building.  Why?  Who knows.
2906         * Scripts/print-vse-failure-logs:
2907          - windows VSE builds don't use Debug/Release as I expected, so I've
2908            moved off of --configuration to --top-level and added the /obj
2909            optimization while I was there.
2910
2911 2010-10-27  Dimitri Glazkov  <dglazkov@chromium.org>
2912
2913         Unreviewed, rolling out r70674.
2914         http://trac.webkit.org/changeset/70674
2915         https://bugs.webkit.org/show_bug.cgi?id=48053
2916
2917         Broke Chromium Windows build.
2918
2919         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2920         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2921
2922 2010-10-25  Tony Chang  <tony@chromium.org>
2923
2924         Reviewed by Anders Carlsson.
2925
2926         compile TestNetscapePlugIn on chromium linux
2927         https://bugs.webkit.org/show_bug.cgi?id=48274
2928
2929         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
2930         (NP_Initialize): On Linux, plugin funcs are set in initialize.
2931         (NPP_New): Mark the plugin as windowless.
2932         (NPP_GetValue): Handle mime type values.
2933         (NP_GetMIMEDescription):
2934         (NP_GetValue):
2935         * DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npapi.h: define TRUE and FALSE, which are in webkit's npapi.h.
2936         * DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npfunctions.h: Pull in npapi.h to get TRUE/FALSE (matches webkit's npfunctions.h)
2937
2938 2010-10-27  Kenneth Rohde Christiansen  <kenneth@webkit.org>
2939
2940         Reviewed by Anders Carlsson.
2941
2942         Remove contentSizeChanged callbacks as it is no longer
2943         part of the public UIClient.
2944
2945         Make WKPageContentsSizeChangedCallback be a private API
2946         https://bugs.webkit.org/show_bug.cgi?id=48409
2947
2948         * MiniBrowser/mac/BrowserWindowController.m:
2949         (-[BrowserWindowController awakeFromNib]):
2950         * MiniBrowser/win/BrowserView.cpp:
2951         (BrowserView::create):
2952         * WebKitTestRunner/TestController.cpp:
2953         (WTR::createOtherPage):
2954         (WTR::TestController::initialize):
2955
2956 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2957
2958         Reviewed by Ojan Vafai.
2959
2960         Make http locking default in NRWT.
2961         https://bugs.webkit.org/show_bug.cgi?id=48053
2962
2963         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2964         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2965
2966 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2967
2968         Reviewed by Tony Chang.
2969
2970         [NRWT] Fix http lock on Windows platform
2971         https://bugs.webkit.org/show_bug.cgi?id=48321
2972
2973         * Scripts/webkitpy/layout_tests/port/http_lock.py:
2974
2975 2010-10-27  Satish Sampath  <satish@chromium.org>
2976
2977         Unreviewed, rolling out r70665.
2978         http://trac.webkit.org/changeset/70665
2979         https://bugs.webkit.org/show_bug.cgi?id=47089
2980
2981         Need to address Alexey's review comments.
2982
2983         * DumpRenderTree/LayoutTestController.cpp:
2984         (setMockSpeechInputResultCallback):
2985         * DumpRenderTree/LayoutTestController.h:
2986         * DumpRenderTree/chromium/LayoutTestController.cpp:
2987         (LayoutTestController::setMockSpeechInputResult):
2988         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2989         (LayoutTestController::setMockSpeechInputResult):
2990         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2991         (LayoutTestController::setMockSpeechInputResult):
2992         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2993         (LayoutTestController::setMockSpeechInputResult):
2994         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2995         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2996         (LayoutTestController::setMockSpeechInputResult):
2997         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2998         (LayoutTestController::setMockSpeechInputResult):
2999
3000 2010-10-27  Leandro Gracia Gil  <leandrogracia@google.com>
3001
3002         Reviewed by Jeremy Orlow.
3003
3004         Added a second parameter to setMockSpeechInputResult for
3005         the language used in speech input.
3006         https://bugs.webkit.org/show_bug.cgi?id=47089
3007
3008         * DumpRenderTree/LayoutTestController.cpp:
3009         (setMockSpeechInputResultCallback):
3010         * DumpRenderTree/LayoutTestController.h:
3011         * DumpRenderTree/chromium/LayoutTestController.cpp:
3012         (LayoutTestController::setMockSpeechInputResult):
3013         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3014         (LayoutTestController::setMockSpeechInputResult):
3015         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3016         (LayoutTestController::setMockSpeechInputResult):
3017         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3018         (LayoutTestController::setMockSpeechInputResult):
3019         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3020         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3021         (LayoutTestController::setMockSpeechInputResult):
3022         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3023         (LayoutTestController::setMockSpeechInputResult):
3024         (LayoutTestController::pageSizeAndMarginsInPixels):
3025
3026 2010-10-27  Adam Roben  <aroben@apple.com>
3027
3028         Reset TestNetscapePlugIn's NPP_GetValue pointer when
3029         NullNPPGetValuePointer finishes running
3030
3031         Fixes <http://webkit.org/b/48435> REGRESSION (r70655): Many plugins
3032         tests are failing on Qt
3033
3034         Reviewed by Anders Carlsson.
3035
3036         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
3037         (PluginTest::NPP_Destroy):
3038         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
3039         Added a do-nothing NPP_Destroy implementation.
3040
3041         * DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp:
3042         (NullNPPGetValuePointer::NullNPPGetValuePointer): Save the original
3043         NPP_GetValue pointer in m_originalNPPGetValuePointer so we can restore
3044         it later.
3045         (NullNPPGetValuePointer::NPP_Destroy): Added. Restores the original
3046         NPP_GetValue pointer so it can be used in other tests.
3047
3048         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
3049         (NPP_Destroy):
3050         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
3051         (webkit_test_plugin_destroy_instance):
3052         Call through to the PluginTest when NPP_Destroy is called.
3053
3054 2010-10-26  Brian Weinstein  <bweinstein@apple.com>
3055
3056         Reviewed by Adam Roben.
3057
3058         WebKit2 shouldn't try to send an empty user agent
3059         https://bugs.webkit.org/show_bug.cgi?id=48397
3060         
3061         Add a test that when we set our custom user agent to an empty string, we don't send
3062         an empty user agent.
3063
3064         * TestWebKitAPI/Tests/WebKit2/PreventEmptyUserAgent.cpp: Added.
3065         (TestWebKitAPI::didRunJavaScript): Make sure that the result of navigator.userAgent isn't empty.
3066         (TestWebKitAPI::TEST): Set our custom user agent to the empty string, and run navigator.userAgent.
3067         
3068         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Add the new file.
3069         * TestWebKitAPI/win/TestWebKitAPI.vcproj: Ditto.
3070
3071 2010-10-27  Adam Roben  <aroben@apple.com>
3072
3073         Check in file I forgot in r70653
3074
3075         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
3076         (NP_Initialize): Save the NPPluginFuncs struct the browser passed to us
3077         so we can be naughty and modify it later.
3078
3079 2010-10-27  Adam Roben  <aroben@apple.com>
3080
3081         Test that WebKit doesn't crash if the plugin passes 0 for its
3082         NPP_GetValue pointer
3083
3084         Test for <http://webkit.org/b/48433> Crash in
3085         NetscapePlugin::shouldLoadSrcURL when using Shockwave Director 10.3 in
3086         WebKit2 on Windows
3087
3088         Reviewed by Eric Carlson.
3089
3090         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3091         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
3092         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
3093         * GNUmakefile.am:
3094         Added NullNPPGetValuePointer.cpp.
3095
3096         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
3097         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
3098         Set up a place to store the NPPluginFuncs struct the browser passed to us.
3099
3100         * DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp: Added.
3101         (NullNPPGetValuePointer::NullNPPGetValuePointer): Null out the
3102         NPP_GetValue pointer we passed to the browser to simulate a plugin that
3103         doesn't implement NPP_GetValue.
3104         (NullNPPGetValuePointer::NPP_GetValue): Print an error message. If this
3105         function is called, it means that WebKit has changed in a way that
3106         makes this test invalid.
3107
3108         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
3109         (NP_GetEntryPoints): Save the NPPluginFuncs struct the browser passed
3110         to us so we can be naughty and modify it later.
3111
3112 2010-10-27  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
3113
3114         Reviewed by Andreas Kling.
3115
3116         Remove references to ancient QGVLauncher and QtLauncher
3117         https://bugs.webkit.org/show_bug.cgi?id=48430
3118
3119         QtTestBrowser substitutes both and has checks enabled.
3120
3121         * Scripts/webkitpy/style/checker.py: remove references
3122         * Scripts/webkitpy/style/checker_unittest.py: ditto
3123
3124 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3125
3126         Reviewed by Ojan Vafai.
3127
3128         [NRWT] Don't use image hash when it's no need in single test mode.
3129         https://bugs.webkit.org/show_bug.cgi?id=48326
3130
3131         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3132
3133 2010-10-27  David Kilzer  <ddkilzer@apple.com>
3134
3135         Fix leak of CFMutableDictionaryRef in createXMLStringFromWebArchiveData()
3136
3137         Reviewed by Adam Roben.
3138
3139         Follow-up fix for: <https://bugs.webkit.org/show_bug.cgi?id=48278>
3140
3141         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
3142         (createXMLStringFromWebArchiveData): Use RetainPtr<> to fix a
3143         leak introduced in r70613.
3144
3145 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3146
3147         Reviewed by Csaba Osztrogonác.
3148
3149         [NRWT] Add platform specific baseline search paths for Qt port
3150         https://bugs.webkit.org/show_bug.cgi?id=48428
3151
3152         * Scripts/webkitpy/layout_tests/port/qt.py:
3153
3154 2010-10-27  Adam Roben  <aroben@apple.com>
3155
3156         Catch exceptions when checking if we're inside a git working directory
3157
3158         Fixes <http://webkit.org/b/48420> REGRESSION (r70562): test-webkitpy
3159         fails on systems without git installed
3160
3161         Reviewed by Anders Carlsson.
3162
3163         * Scripts/webkitpy/common/net/credentials.py:
3164         (Credentials._credentials_from_git): Put the call to
3165         Git.in_working_directory inside the try/except since it, too, attempts
3166         to execute git and thus will throw on systems that don't have git
3167         installed.
3168
3169 2010-10-27  Nikolas Zimmermann  <nzimmermann@rim.com>
3170
3171         Reviewed by David Kilzer.
3172
3173         Convert DumpRenderTree webarchive code to CoreFoundation
3174         https://bugs.webkit.org/show_bug.cgi?id=48278
3175
3176         CFPropertyListCreateWithData and CFPropertyListCreateData are only available in 10.6+.
3177         Replace CFPropertyListCreateWithData by a combination of CFReadStreamCreateWithBytesNoCopy and CFPropertyListCreateFromStream.
3178         Replace CFPropertyListCreateData by CFPropertyListCreateXMLData.
3179
3180         These changes are wrapped in BUILDING_ON_LEOPARD, as the methods are deprecated on 10.6+.
3181
3182         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
3183         (createXMLStringFromWebArchiveData):
3184
3185 2010-10-26  Antonio Gomes  <agomes@rim.com>
3186
3187         Reviewed by Martin Robinson.
3188
3189         Remove the absolute path used to include DumpRenderTreeSupportGtk.h from LayoutTestController.h
3190
3191         It turns out this is rather unneeded since WebKitTools/GNUMakefile.am has WebKit/gtk/ in its include
3192         path, and then we can just do #include "WebCoreSupport/DumpRenderTreeSupportGtk.h"
3193
3194         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3195
3196 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3197
3198         Rename WebArchiveDumpSupport.mm to WebArchiveDumpSupport.cpp
3199
3200         Reviewed by Adam Roben.
3201
3202         Part 4 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3203
3204         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3205         Updated for file moves and renames.
3206         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.mm.
3207         * DumpRenderTree/cf/WebArchiveDumpSupport.h: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.h.
3208
3209 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3210
3211         Convert WebArchiveDumpSupport.mm from NS objects to CF types
3212
3213         Reviewed by Adam Roben.
3214
3215         Part 3 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3216
3217         In order to share WebArchive code between the Mac and Windows
3218         ports, the code in WebArchiveDumpSupport.mm was converted from
3219         Cocoa to C++ using CoreFoundation (CF) types.
3220
3221         * DumpRenderTree/mac/DumpRenderTree.mm:
3222         (dump): Renamed serializeWebArchiveToXML() to
3223         createXMLStringFromWebArchiveData() and added HardAutorelease()
3224         to prevent leaks.
3225         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
3226         (createXMLStringFromWebArchiveData): Renamed from
3227         serializeWebArchiveToXML().  Changed to use CF types.
3228         (createCFURLResponseFromResponseData): Renamed from
3229         unarchiveNSURLResponseFromResponseData().  Changed to use CF
3230         types for its parameter and return type.
3231         * DumpRenderTree/mac/WebArchiveDumpSupport.mm: Replaced use of
3232         NS objects with CF types.  It will be renamed to *.cpp in a
3233         future commit.
3234         (convertMIMEType): Changed to use case-insensitive string
3235         comparisons.
3236         (convertWebResourceDataToString):
3237         (normalizeHTTPResponseHeaderFields):
3238         (normalizeWebResourceURL):
3239         (convertWebResourceResponseToDictionary):
3240         (compareResourceURLs):
3241         (createXMLStringFromWebArchiveData):
3242         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
3243         (createCFURLResponseFromResponseData): Renamed from
3244         unarchiveNSURLResponseFromResponseData().  Changed to use CF
3245         types for its parameter and return type.
3246
3247 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3248
3249         Extract use of NSKeyedUnarchiver from WebArchiveDumpSupport.mm
3250
3251         Reviewed by Adam Roben.
3252
3253         Part 2 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3254
3255         There is no equivalent to NSKeyedUnarchiver in CoreFoundation,
3256         so extract it into a platform-specific source file.
3257
3258         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
3259         (unarchiveNSURLResponseFromResponseData): Added declaration.
3260         * DumpRenderTree/mac/WebArchiveDumpSupport.mm:
3261         (convertWebResourceResponseToDictionary): Extracted code to
3262         unarchiveNSURLResponseFromResponseData() in
3263         WebArchiveDumpSupportMac.mm.  Updated logic to return early if
3264         nil is returned from unarchiveNSURLResponseFromResponseData().
3265         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
3266         (unarchiveNSURLResponseFromResponseData): Added.  Extracted
3267         code from convertWebResourceResponseToDictionary() in
3268         WebArchiveDumpSupport.mm.
3269
3270 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3271
3272         Extract call to -[WebHTMLRepresentation supportedNonImageMIMETypes] from WebArchiveDumpSupport.mm
3273
3274         Reviewed by Adam Roben.
3275
3276         Part 1 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3277
3278         The call to -[WebHTMLRepresentation supportedNonImageMIMETypes]
3279         is not cross-platform between Mac and Windows, so extract it
3280         into a platform-specific source file.
3281
3282         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Added
3283         WebArchiveDumpSupportMac.mm to the project.
3284         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
3285         (supportedNonImageMIMETypes): Added declaration.
3286         * DumpRenderTree/mac/WebArchiveDumpSupport.mm:
3287         (convertWebResourceDataToString): Replaced call to
3288         -[WebHTMLRepresentation supportedNonImageMIMETypes] with
3289         supportedNonImageMIMETypes().
3290         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm: Added.
3291         (supportedNonImageMIMETypes): Added.  Extracted from
3292         WebArchiveDumpSupport.mm.
3293
3294 2010-10-26  Antonio Gomes  <agomes@rim.com>
3295
3296         Reviewed by Martin Robinson.
3297
3298         [GTK] Implement DumpRenderTreeSupportGtk (similarly to DumpRenderTreeSupportQt idea)
3299         https://bugs.webkit.org/show_bug.cgi?id=48199
3300
3301         Implements support to WebKitTabToLinksPreferenceKey through LayoutTestController::overridePreference.
3302         The corresponding DumpRenderTreeSupportGtk method is called in the DRT context only.
3303
3304         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3305         (resetDefaultsToConsistentValues):
3306         (createWebView):
3307         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3308         (LayoutTestController::overridePreference):
3309
3310 2010-10-26  Ariya Hidayat  <ariya@sencha.com>
3311
3312         Unreviewed, change the order of my emails for bugzilla autocompletion.
3313
3314         * Scripts/webkitpy/common/config/committers.py:
3315
3316 2010-10-26  Eric Seidel  <eric@webkit.org>
3317
3318         Reviewed by David Kilzer.
3319
3320         build-webkit should collect Visual Studio Express logs and display them
3321         https://bugs.webkit.org/show_bug.cgi?id=39199
3322
3323         * Scripts/build-webkit:
3324         * Scripts/print-vse-failure-logs: Added.
3325         * Scripts/webkitdirs.pm:
3326
3327 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3328
3329         <http://webkit.org/b/48224> build-webkit: add support for --meter-tag switch
3330
3331         Reviewed by Csaba Osztrogonác.
3332
3333         * Scripts/build-webkit: Added support for --meter-tag switch.
3334         It should have been added with r60820.
3335
3336 2010-10-26  Ademar de Souza Reis Jr.  <ademar.reis@openbossa.org>
3337
3338         Reviewed by Dumitru Daniliuc.
3339
3340         check-webkit-style fails on operator+=, operator-=, ... methods
3341         https://bugs.webkit.org/show_bug.cgi?id=48258
3342
3343         * Scripts/webkitpy/style/checkers/cpp.py: Added exceptions
3344         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added unit-tests
3345
3346 2010-10-26  Ojan Vafai  <ojan@chromium.org>
3347
3348         Reviewed by Tony Chang.
3349
3350         remove DEFER support from new-run-webkit-tests
3351         https://bugs.webkit.org/show_bug.cgi?id=48387
3352
3353         DEFER was needed when we were trying to ship Chrome beta.
3354         Now it's just extra complication.
3355
3356         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
3357         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
3358         * Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
3359         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
3360         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3361         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
3362
3363 2010-10-26  Eric Seidel  <eric@webkit.org>
3364
3365         Unreviewed.  Adding import with_statement to fix python 2.5.
3366
3367         Teach webkit-patch how to read credentials from the environment
3368         https://bugs.webkit.org/show_bug.cgi?id=48275
3369
3370         * Scripts/webkitpy/common/net/credentials_unittest.py:
3371          - import with_statement to fix python 2.5.
3372
3373 2010-10-26  Alexey Proskuryakov  <ap@apple.com>
3374
3375         Reviewed by Sam Weinig.
3376
3377         https://bugs.webkit.org/show_bug.cgi?id=48375
3378         <rdar://problem/8392724> Need delegate calls in PageLoaderClient to indicate if we have
3379         loaded insecure content
3380
3381         Updated for WebKit2 changes.
3382
3383         * MiniBrowser/mac/BrowserWindowController.m:
3384         (didDisplayInsecureContentForFrame):
3385         (didRunInsecureContentForFrame):
3386         (-[BrowserWindowController awakeFromNib]):
3387         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3388         (WTR::InjectedBundlePage::InjectedBundlePage):
3389         (WTR::InjectedBundlePage::didDisplayInsecureContentForFrame):
3390         (WTR::InjectedBundlePage::didRunInsecureContentForFrame):
3391         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
3392         * WebKitTestRunner/TestController.cpp:
3393         (WTR::TestController::initialize):
3394
3395 2010-10-26  Adam Roben  <aroben@apple.com>
3396
3397         Clear up confusion between "3D rendering" and "accelerated compositing"
3398
3399         WebKit has no software-based 3D rendering implementation, so 3D
3400         rendering can only be enabled when accelerated compositing is. But DRT
3401         was falsely reporting that 3D rendering was available on machines that
3402         don't support accelerated compositing, leading to much confusion.
3403
3404         Reviewed by Darin Adler.
3405
3406         Fixes <http://webkit.org/b/48370> REGRESSION (r70540): Many
3407         transforms/3d tests are failing on the XP bots
3408
3409         * DumpRenderTree/win/DumpRenderTree.cpp:
3410         (main): When ENABLE(3D_RENDERING) is turned on, only report that 3D
3411         rendering is available when acclerated compositing is also available.
3412
3413         * Scripts/old-run-webkit-tests: Skip tests that use the
3414         -webkit-transform-3d media query when 3D rendering is disabled,
3415         rather than when accelerated compositing is disabled.
3416
3417 2010-10-26  Eric Seidel  <eric@webkit.org>
3418
3419         Reviewed by Ojan Vafai.
3420
3421         Remove rietveld code now that it's unused
3422         https://bugs.webkit.org/show_bug.cgi?id=48359
3423
3424         Was cool that we added this support, but now that it's
3425         not used it makes little sense to keep it around.
3426         We can always restore this code from SVN if we need it.
3427
3428         * Scripts/webkitpy/common/config/__init__.py:
3429         * Scripts/webkitpy/common/net/bugzilla.py:
3430         * Scripts/webkitpy/common/net/bugzilla_unittest.py:
3431         * Scripts/webkitpy/common/net/rietveld.py: Removed.
3432         * Scripts/webkitpy/common/net/rietveld_unittest.py: Removed.
3433         * Scripts/webkitpy/thirdparty/__init__.py:
3434         * Scripts/webkitpy/tool/commands/download.py:
3435         * Scripts/webkitpy/tool/commands/download_unittest.py:
3436         * Scripts/webkitpy/tool/commands/queues.py:
3437         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3438         * Scripts/webkitpy/tool/main.py:
3439         * Scripts/webkitpy/tool/mocktool.py:
3440         * Scripts/webkitpy/tool/steps/__init__.py:
3441         * Scripts/webkitpy/tool/steps/postcodereview.py: Removed.
3442
3443 2010-10-26  Anders Carlsson  <andersca@apple.com>
3444
3445         Reviewed by Sam Weinig.
3446
3447         pageDidScroll callback should be on the UI process client rather than (or in addition to) the web process client
3448         https://bugs.webkit.org/show_bug.cgi?id=48366
3449         <rdar://problem/8595202>
3450
3451         * MiniBrowser/mac/BrowserWindowController.m:
3452         (-[BrowserWindowController awakeFromNib]):
3453         * WebKitTestRunner/TestController.cpp:
3454         (WTR::createOtherPage):
3455         (WTR::TestController::initialize):
3456
3457 2010-10-26  Eric Seidel  <eric@webkit.org>
3458
3459         Reviewed by Tony Chang.
3460
3461         Teach webkit-patch how to read credentials from the environment
3462         https://bugs.webkit.org/show_bug.cgi?id=48275
3463
3464         This makes it possible for svn users to have their bugzilla credentials
3465         stored in their environment instead of typing them every time.
3466
3467         We need this for making it easy to run the win-ews bot (which currently
3468         uses svn instead of git).
3469
3470         * Scripts/webkitpy/common/net/credentials.py:
3471         * Scripts/webkitpy/common/net/credentials_unittest.py:
3472
3473 2010-10-26  Kenichi Ishibashi  <bashi@google.com>
3474
3475         Reviewed by Kent Tamura.
3476
3477         Input Method inserts conversion candidates unexpectedly
3478         https://bugs.webkit.org/show_bug.cgi?id=46868
3479
3480         Adds setComposition() to TextInputController to make DRT emulate
3481         an input method behavior.
3482
3483         * DumpRenderTree/chromium/TextInputController.cpp:
3484         (TextInputController::TextInputController):
3485         (TextInputController::setComposition): Added.
3486         * DumpRenderTree/chromium/TextInputController.h:
3487
3488 2010-10-26  Simon Fraser  <simon.fraser@apple.com>
3489
3490         Reviewed by Sam Weinig.
3491
3492         Support layoutTestController.layerTreeAsText in WebKitTestRunner
3493         https://bugs.webkit.org/show_bug.cgi?id=42145
3494         
3495         Implement layerTreeAsText() in WebKitTestRunner.
3496
3497         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
3498         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
3499         (WTR::LayoutTestController::layerTreeAsText):
3500         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
3501
3502 2010-10-26  Adam Barth  <abarth@webkit.org>
3503
3504         Reviewed by Eric Seidel.
3505
3506         webkit-patch rollout produces incorrect patch when using svn move
3507         https://bugs.webkit.org/show_bug.cgi?id=48244
3508
3509         We need to flush our caches when we modify the working copy.
3510
3511         * Scripts/webkitpy/tool/steps/revertrevision.py:
3512
3513 2010-10-26  Eric Seidel  <eric@webkit.org>
3514
3515         Reviewed by Adam Barth.
3516
3517         commit-queue flaky test message can list the same author more than once
3518         https://bugs.webkit.org/show_bug.cgi?id=48268
3519
3520         tonikitoo reported to me over IRC this morning that he's seen
3521         the commit-queue report flaky tests with author lists like:
3522         "adam, adam and adam", suggesting we're not uniquing authors
3523         before writing the message.
3524
3525         I fixed the uniquing and added a bunch more unit testing.
3526
3527         * Scripts/webkitpy/tool/commands/queues.py:
3528         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3529         * Scripts/webkitpy/tool/mocktool.py:
3530
3531 2010-10-26  Adam Roben  <aroben@apple.com>
3532
3533         Pull in the FeatureDefines*.vsprops files when building DRT
3534
3535         This ensures that various ENABLE() macros will get set correctly.
3536         (Prior to r70320 we were picking up the ENABLE(3D_RENDERING) flag
3537         through wtf/Platform.h.)
3538
3539         Fixes <http://webkit.org/b/48343> REGRESSION (r70320): DumpRenderTree
3540         says 3D_RENDERING is disabled even when it is enabled
3541
3542         Reviewed by Ada Chan.
3543
3544         * DumpRenderTree/win/DumpRenderTree.vcproj: Added
3545         FeaturesDefines*.vsprops to all configurations.
3546
3547         * DumpRenderTree/win/DumpRenderTreeWin.h: Touched to force a rebuild.
3548
3549 2010-10-26  Søren Gjesse  <sgjesse@chromium.org>
3550
3551         Reviewed by Dimitri Glazkov.
3552
3553         Added options --multiple-loads and --js-flags to chromium DumpRenderTree. The option --multiple-loads=X
3554         is used to have DumpRenderTree load each test it runs X times. To be able to have more fine-grained control
3555         of how the JavaScript engine behaves for each load the flag --js-flags can specify a list of flag-sets like this
3556
3557           --js-flags="--xxx,--noxxx --yyy,--noyyy"
3558
3559         First load will run with --xxx, the second with --yyy and the third without any (the 'no' prefix is handled by
3560         V8 to turn off the flag).
3561
3562         The changes to the Python test runner will be in a separate change.
3563
3564
3565         * DumpRenderTree/chromium/DumpRenderTree.cpp:
3566         (runTest):
3567         (main):
3568         * DumpRenderTree/chromium/TestShell.cpp:
3569         (TestShell::TestShell):
3570         (TestShell::runFileTest):
3571         (TestShell::testFinished):
3572         * DumpRenderTree/chromium/TestShell.h: