f8c3f61a10a5744162760241528015f84fed51ed
[WebKit-https.git] / WebKitTools / ChangeLog
1 2010-11-17  Satish Sampath  <satish@chromium.org>
2
3         Reviewed by Jeremy Orlow.
4
5         Clear the speech input mock explicitly before each test.
6         https://bugs.webkit.org/show_bug.cgi?id=49660
7
8         * DumpRenderTree/chromium/LayoutTestController.cpp:
9         (LayoutTestController::setMockSpeechInputResult):
10         * DumpRenderTree/chromium/LayoutTestController.h:
11         * DumpRenderTree/chromium/WebViewHost.cpp:
12         (WebViewHost::speechInputController):
13         (WebViewHost::reset): Invoke speech input mock's clearResults method.
14         * DumpRenderTree/chromium/WebViewHost.h:
15         (WebViewHost::speechInputControllerMock): Moved mock from LayoutTestController.
16
17 2010-11-17  Alexander Pavlov  <apavlov@chromium.org>
18
19         Unreviewed, specify IRC nickname in committers.py.
20
21         * Scripts/webkitpy/common/config/committers.py:
22
23 2010-11-17  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
24
25         Reviewed by Tony Chang.
26
27         [NRWT] Make http locking similar to perl implementation
28         https://bugs.webkit.org/show_bug.cgi?id=49187
29
30         * Scripts/webkitpy/common/system/file_lock.py: Added.
31         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
32         * Scripts/webkitpy/layout_tests/port/http_lock.py:
33
34 2010-11-17  MORITA Hajime  <morrita@google.com>
35
36         Reviewed by Kent Tamura.
37
38         [Chromium][DRT] EventSender.contextClick() should aware spellchecking
39         https://bugs.webkit.org/show_bug.cgi?id=49366
40
41         - EvenSender: Checked WebContextMenuData.misspelledWord and added extra context menu entries
42           according to the spellchecker suggestion
43         - MockSpellCheck: add fillSuggestionList to provide fake suggestions.
44
45         test_expectations.txt will be changed after this change is ported to test_shell.
46         
47         * DumpRenderTree/chromium/EventSender.cpp:
48         (makeMenuItemStringsFor):
49         (EventSender::contextClick):
50         * DumpRenderTree/chromium/MockSpellCheck.cpp:
51         (MockSpellCheck::fillSuggestionList):
52         (MockSpellCheck::initializeIfNeeded):
53         * DumpRenderTree/chromium/MockSpellCheck.h:
54         * DumpRenderTree/chromium/WebViewHost.cpp:
55         (WebViewHost::mockSpellCheck):
56         * DumpRenderTree/chromium/WebViewHost.h:
57
58 2010-11-16  Dave Hyatt  <hyatt@apple.com>
59
60         Reviewed by Sam Weinig.
61
62         Make sure the pref for minimum font size is just explicitly set to 0 every time,
63         since the bots have 1 stuck in their plist otherwise.
64
65         * DumpRenderTree/mac/DumpRenderTree.mm:
66         (resetDefaultsToConsistentValues):
67
68 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
69
70         Reviewed by Tony Chang.
71
72         Rebaseline server: compute diffs client-side
73         https://bugs.webkit.org/show_bug.cgi?id=49640
74         
75         The image diff output from the DRT is pretty bad on some ports (at
76         least the Mac one), so it's better to compute diffs on the client by
77         using <canvas>.
78
79         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
80         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
81         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
82
83 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
84
85         Reviewed by Tony Chang.
86
87         Rebaseline server: display test results
88         https://bugs.webkit.org/show_bug.cgi?id=49626
89         
90         Adds basic result display to the rebaseline server. On the Python side
91         this involves:
92         - Parsing the unexpected_results.json into a dictionary.
93         - Serving it as JSON under /results.json.
94         (the JSON -> dict -> JSON transform isn't strictly necessary right now,
95         but I'll need to have access to the parsed results on the Python side
96         for follow-up changes).
97         
98         On the web UI side this adds:
99         - Markup for display image and text results (expected, actual, diff),
100           and JS for populating it.
101         - Markup for breaking down test results by failure type and directory,
102           and JS for populating it.
103
104         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
105         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
106         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
107         * Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js: Added.
108         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
109
110 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
111
112         Reviewed by Ojan Vafai.
113
114         See Chromium issue http://codereview.chromium.org/5133001/ - we
115         are modifying the build bots to pass in the builder name with
116         the "GPU" string appended instead of appending it in the code.
117
118         https://bugs.webkit.org/show_bug.cgi?id=49636
119
120         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
121
122 2010-11-16  Steve Falkenburg  <sfalken@apple.com>
123
124         Reviewed by Adam Roben.
125
126         Disable LTCG for Windows Release builds. Add new Release_LTCG configuration.
127         https://bugs.webkit.org/show_bug.cgi?id=49632
128
129         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
130         * DumpRenderTree/win/DumpRenderTree.vcproj:
131         * DumpRenderTree/win/ImageDiff.vcproj:
132         * MiniBrowser/MiniBrowser.vcproj:
133         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
134         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
135         * WebKitTestRunner/win/InjectedBundle.vcproj:
136         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
137         * WinLauncher/WinLauncher.vcproj:
138         * record-memory-win/record-memory-win.vcproj:
139
140 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
141
142         Reviewed by Ojan Vafai.
143
144         new-run-webkit-tests: rename TestInfo to TestInput, move image hash to work thread
145
146         Rename the TestInfo class to TestInput to be clearer about its
147         function, and move the checksum-reading code into dump_render_tree_thread
148         to avoid cross-thread access.
149
150         https://bugs.webkit.org/show_bug.cgi?id=49573
151
152         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
153         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
154         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
155
156 2010-11-16  Dave Hyatt  <hyatt@apple.com>
157
158         Fix bustage. Remove the minimum font size pref setting in DumpRenderTree.
159
160         * DumpRenderTree/mac/DumpRenderTree.mm:
161         (resetDefaultsToConsistentValues):
162
163 2010-11-16  Stephanie Lewis  <slewis@apple.com>
164
165         Reviewed by Geoff Garen.
166
167         <rdar://problem/8624267> Leak creating offscreen webview running fast/dom tests
168
169         Use shared DeviceOrientationProviderMock. The old allocation method created a new WebDeviceOrientationProviderMock 
170         for every WebView, and the WebKit API doesn't support that ownership model.
171
172         I also fixed a leak in setMockDeviceOrientation by adding a missing release.
173
174         * DumpRenderTree/mac/DumpRenderTree.mm:
175         (createWebViewAndOffscreenWindow):
176         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
177         (LayoutTestController::setMockDeviceOrientation):
178
179 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
180
181         Reviewed by Ojan Vafai.
182
183         new-run-webkit-tests: log the process id in --verbose mode
184
185         As part of the switch to multi-process mode, we should log the
186         process id when running w/ --verbose.
187
188         https://bugs.webkit.org/show_bug.cgi?id=49571
189
190         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
191
192 2010-11-16  Adam Roben  <aroben@apple.com>
193
194         Only trigger Windows builds when files that we actually use on Windows
195         are changed
196
197         A new Scheduler subclass, PlatformSpecificScheduler, has been added.
198         It uses the new webkitpy.common.config.build module to determine
199         whether a particular change should trigger a build on a particular
200         platform. The Windows builders have been switched to use a
201         PlatformSpecificScheduler.
202
203         The logic to determine whether or not a particular change should
204         trigger a build on a given platform has only been implemented/tested
205         for Windows. I tried to make it easy to add more platforms in the
206         future, but I don't have enough familiarity with all platforms to be
207         able to implement it for them.
208
209         Fixes <http://webkit.org/b/49407> Windows builders kick off builds for
210         lots irrelevant changes (e.g., rebaselining Chromium test results)
211
212         Reviewed by Eric Seidel.
213
214         * BuildSlaveSupport/build.webkit.org-config/config.json: Use a
215         PlatformSpecificScheduler for the Windows builders.
216
217         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
218         (PlatformSpecificScheduler.__init__): Added. Stores our platform, sets
219         up a ChangeFilter that filters to our branch and will call through to
220         our filter method, and calls up to our base class.
221         (PlatformSpecificScheduler.filter): Calls through to
222         build.should_build to find out whether we should trigger a build for
223         this change.
224
225         * Scripts/webkitpy/common/net/build.py: Added.
226         (_should_file_trigger_build): Uses a set of directories and regexp
227         patterns to determine whether the given file should trigger a build on
228         the given platform. As mentioned earlier, this has only been
229         implemented for Windows, though I did try to make some guesses about
230         other platforms.
231         (should_build): Returns true if any of the files should trigger a
232         build on the given platform.
233
234         * Scripts/webkitpy/common/net/build_unittest.py: Added.
235         (ShouldBuildTest.test_should_build): Does some basic testing to make
236         sure we're triggering builds for the right files. It only tests
237         Windows for now, though I tried to make some guesses about other
238         platforms.
239
240 2010-11-16  Simon Hausmann  <simon.hausmann@nokia.com>
241
242         Reviewed by Andreas Kling.
243
244         [Qt] Remove synchronous QWebPage::checkPermissions signal
245         https://bugs.webkit.org/show_bug.cgi?id=46810
246
247         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
248         (WebCore::WebPage::WebPage):
249         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
250         (LayoutTestController::grantDesktopNotificationPermission): When granting
251         permission, grant it directly on the QWebPage/Frame, that will remember it.
252         * QtTestBrowser/webpage.cpp:
253         (WebPage::WebPage):
254         * QtTestBrowser/webpage.h:
255
256
257 2010-11-16  Andras Becsi  <abecsi@inf.u-szeged.hu>
258
259         Reviewed by Csaba Osztrogonác.
260
261         [Qt] Add isPrinting and setPrinting methods to DRT's LayoutTestController.
262
263         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
264         (LayoutTestController::reset):
265         * DumpRenderTree/qt/LayoutTestControllerQt.h:
266         (LayoutTestController::isPrinting):
267         (LayoutTestController::setPrinting):
268
269 2010-11-16  Balazs Kelemen  <kbalazs@webkit.org>
270
271         Reviewed by Andreas Kling.
272
273         [Qt][WK2] Add a way to turn on tiled backing store in MiniBrowser
274         https://bugs.webkit.org/show_bug.cgi?id=49587
275
276         * MiniBrowser/qt/BrowserView.cpp:
277         (BrowserView::BrowserView): Added BackingStoreType
278         argument to ctor. Propagating it to the QGraphicsWKView.
279         * MiniBrowser/qt/BrowserView.h:
280         * MiniBrowser/qt/BrowserWindow.cpp:
281         (BrowserWindow::BrowserWindow):
282         (BrowserWindow::newWindow): Create the new BrowserWindow
283         with the same BackingStoreType that this instance has.
284         * MiniBrowser/qt/BrowserWindow.h: Added BackingStoreType member.
285         * MiniBrowser/qt/main.cpp:
286         (main): Use tiled backing store if got -tiled command line argument.
287         Avoid redundant copy of command line args.
288
289 2010-11-16  John Knottenbelt  <jknotten@chromium.org>
290
291         Reviewed by Jeremy Orlow.
292
293         Move DeviceOrientationClientMock from LayoutTestController to WebViewHost.
294         https://bugs.webkit.org/show_bug.cgi?id=48506
295
296         This ensures that there is one mock per page WebView / WebViewHost. This is a
297         design constraint of the DeviceOrientationClientMock because the m_controller
298         field can only store a pointer to a single instance of the corresponding
299         DeviceOrientationController.
300
301         Test: fast/dom/DeviceOrientation/no-page-cache.html
302
303         * DumpRenderTree/chromium/LayoutTestController.cpp:
304         (LayoutTestController::setMockDeviceOrientation):
305         (LayoutTestController::deviceOrientationClientMock):
306         * DumpRenderTree/chromium/LayoutTestController.h:
307         * DumpRenderTree/chromium/WebViewHost.cpp:
308         (WebViewHost::deviceOrientationClientMock):
309         (WebViewHost::deviceOrientationClient):
310         (WebViewHost::reset):
311         * DumpRenderTree/chromium/WebViewHost.h:
312
313 2010-11-15  Andras Becsi  <abecsi@webkit.org>
314
315         Reviewed by Andreas Kling.
316
317         [Qt][WK2] Avoid polling in WebKitTestRunner.
318         https://bugs.webkit.org/show_bug.cgi?id=49542
319
320         Make the WK2 testing session about 10-15% faster and decrease random flakiness
321         resulting from timing skew by removing the millisecond polling from TestControllerQt.
322
323         * WebKitTestRunner/TestController.cpp:
324         (WTR::TestController::didFinishLoadForFrame):
325         * WebKitTestRunner/TestController.h:
326         * WebKitTestRunner/TestInvocation.cpp:
327         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
328         * WebKitTestRunner/mac/TestControllerMac.mm:
329         (WTR::TestController::notifyDone): Add method stub.
330         * WebKitTestRunner/qt/TestControllerQt.cpp:
331         (WTR::TestControllerRunLoop::instance):
332         (WTR::TestControllerRunLoop::start):
333         (WTR::TestControllerRunLoop::stop):
334         (WTR::TestControllerRunLoop::TestControllerRunLoop):
335         (WTR::TestControllerRunLoop::timerEvent):
336         (WTR::TestController::notifyDone):
337         (WTR::TestController::platformRunUntil):
338         * WebKitTestRunner/win/TestControllerWin.cpp:
339         (WTR::TestController::notifyDone): Add method stub.
340
341 2010-11-15  Darin Adler  <darin@apple.com>
342
343         * Scripts/webkitpy/common/net/bugzilla: Added property svn:ignore.
344
345 2010-11-15  Benjamin Poulain  <benjamin.poulain@nokia.com>
346
347         Reviewed by Andreas Kling.
348
349         [Qt] [WK2] Add Ctrl+L shortcut to Qt Minibrowser
350         https://bugs.webkit.org/show_bug.cgi?id=49544
351
352         Add the shortcut to BrowserWindow.
353
354         * MiniBrowser/qt/BrowserWindow.cpp:
355         (BrowserWindow::BrowserWindow):
356         (BrowserWindow::openLocation):
357         * MiniBrowser/qt/BrowserWindow.h:
358
359 2010-11-14  David Levin  <levin@chromium.org>
360
361         Reviewed by Daniel Bates.
362
363         check-webkit-style should detect PassRefPtr usage in functions.
364         https://bugs.webkit.org/show_bug.cgi?id=49513
365
366         * Scripts/webkitpy/style/checkers/cpp.py:
367         (check_for_function_lengths): Revert a comment change that I
368          accidentally made in r71986.
369         (check_pass_ptr_usage): Added the code to do the check.
370         (process_line): Added the call to check_pass_ptr_usage.
371         (CppChecker): Added the new error category.
372         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
373         (CppStyleTestBase::perform_pass_ptr_check): Runs the new check for
374          testing purposes.
375         (PassPtrTest::*): The class/functions to unit test the new
376          functionality.
377
378 2010-11-14  David Levin  <levin@chromium.org>
379
380         Reviewed by Shinichiro Hamaji.
381
382         check-webkit-style function detection and the line count style checks should be separate.
383         https://bugs.webkit.org/show_bug.cgi?id=49512
384
385         * Scripts/webkitpy/style/checkers/cpp.py: Do the separation.
386         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Adjust the test to
387           call the detection function and fix line counts in two places now that
388           the code really only counts the lines in the body of the function.
389
390 2010-11-14  Andreas Kling  <kling@webkit.org>
391
392         Reviewed by Antonio Gomes.
393
394         http/tests/plugins tests print "Unhandled variable" to stderr under Qt
395         https://bugs.webkit.org/show_bug.cgi?id=33438
396
397         Stifle "Unhandled variable" warning to match what the other
398         TestNetscapePlugin does (NPP_GetValue in TestNetscapePlugIn/main.cpp
399         simply returns NPERR_GENERIC_ERROR for unhandled variables.)
400
401         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
402         (webkit_test_plugin_get_value):
403
404 2010-11-13  David Levin  <levin@chromium.org>
405
406         Reviewed by Eric Seidel.
407
408         check-webkit-style function detection crashes on functions in templates.
409         https://bugs.webkit.org/show_bug.cgi?id=49504
410
411         For example "template <bool shouldClose(const Element*)>".
412
413         * Scripts/webkitpy/style/checkers/cpp.py: If not function is detected, bail out.
414         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Modified the complex
415           function detection test to expose this issue.
416
417 2010-11-12  David Levin  <levin@chromium.org>
418
419         Reviewed by Shinichiro Hamaji.
420
421         check-webkit-style function detection doesn't detect indented functions declaractions.
422         https://bugs.webkit.org/show_bug.cgi?id=49446
423
424         Indented function declarations occur inside class definitions, so
425         they are a pretty common (and worth detecting).
426
427         * Scripts/webkitpy/style/checkers/cpp.py:
428           Changed regex to allow indentation.
429           Changed the function start detection to only happen when not in a
430           function.
431           Changed function end detection to work based on matching braces
432           instead of finding a close brace at the beginning of the line.
433           Fixed close_expression to do what it says when it doesn't find
434           the close.
435         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Indented function test.
436
437 2010-11-12  Daniel Bates  <dbates@rim.com>
438
439         Rubber-stamped by Nikolas Zimmermann.
440
441         Make do-file-rename rename files in the directory WebKit2.
442
443         This also makes the list of searched directories in do-file-rename
444         consistent with the list of searched directories in the script do-webcore-rename.
445
446         * Scripts/do-file-rename:
447
448 2010-11-12  Dirk Pranke  <dpranke@chromium.org>
449
450         Reviewed by James Robinson.
451
452         Attempt yet again to land the fix for bug 49360 (respecting
453         set-webkit-configuration). We need to handle the cases where
454         trying to run webkit-build-directory to find out where the
455         default configuration might be fails (that shows up on some
456         Chromium bots that apparently don't have perl installed).
457
458         https://bugs.webkit.org/show_bug.cgi?id=49360
459
460         * Scripts/webkitpy/layout_tests/port/config.py:
461         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Added.
462         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
463
464 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
465
466         Reviewed by Adam Barth.
467
468         webkit-patch land logs commit message twice
469         https://bugs.webkit.org/show_bug.cgi?id=49482
470         
471         Don't log the comment that's used by bugzilla.close_bug_as_fixed, since
472         that's just the commit message, which we now output following the
473         actual commit.
474
475         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
476
477 2010-11-12  Tony Chang  <tony@chromium.org>
478
479         Reviewed by Adam Barth.
480
481         run platform/chromium/plugins/script-object-invoke.html on all platforms
482         https://bugs.webkit.org/show_bug.cgi?id=49280
483
484         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
485         (testScriptObjectInvoke):
486         (pluginInvoke):
487
488 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
489
490         Reviewed by Adam Barth.
491
492         webkit-patch land-cowboy (and other commands) should display the committed SVN revision after landing a patch
493         https://bugs.webkit.org/show_bug.cgi?id=49471
494
495         Right now there's no indication of a successful commit.
496         
497         * Scripts/webkitpy/tool/commands/download_unittest.py:
498         * Scripts/webkitpy/tool/steps/commit.py:
499
500 2010-11-12  Kinuko Yasuda  <kinuko@chromium.org>
501
502         Reviewed by Ojan Vafai.
503
504         Include detailed test modifiers in results.json and enable incremental uploading for non-layout tests
505         https://bugs.webkit.org/show_bug.cgi?id=49354
506
507         Also moved/integrated the upload method from run_webkit_tests.py to json_results_generator.py.
508
509         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
510         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
511         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
512         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
513
514 2010-11-12  Tony Chang  <tony@chromium.org>
515
516         Reviewed by Kent Tamura.
517
518         [chromium] Load a blank page when closing DRT WebViewHosts
519         https://bugs.webkit.org/show_bug.cgi?id=49417
520
521         In test_shell, when we close a window, we first load about:blank
522         and run GC to fire the destruction logic of the page.  In test_shell,
523         this happens in ~TestShell.  In DRT, we manage virtual windows as
524         WebViewHosts, so we need to replicate this logic in ~WebViewHost.
525
526         Managing focus is a bit trickier, since normally you would get window
527         events to reset the focus to the right window.  We do it manually
528         instead.
529
530         This fixes 2 plugin tests on DRT.
531
532         * DumpRenderTree/chromium/TestShell.cpp:
533         (TestShell::~TestShell): Move window close logic to ~WebViewHost.
534         (TestShell::closeWindow): Move widget close() to WVH destructor and
535             manage focus manually.
536         (TestShell::closeRemainingWindows): Comment typo.
537         * DumpRenderTree/chromium/WebViewHost.cpp:
538         (WebViewHost::closeWidget):
539         (invokeCloseWidget):
540         (WebViewHost::closeWidgetSoon): Close the window asynchronously (matches
541             test_shell).
542         (WebViewHost::~WebViewHost):
543         * DumpRenderTree/chromium/WebViewHost.h:
544
545 2010-11-12  David Levin  <levin@chromium.org>
546
547         Reviewed by Adam Barth.
548
549         cpp.py needs some tidying.
550         https://bugs.webkit.org/show_bug.cgi?id=49443
551
552         * Scripts/webkitpy/style/checkers/cpp.py:
553           Fix parameter ordering for iteratively_replace_matches_with_char
554           to mirror other similar functions (and improved comments, etc.)
555           Also, use a python-ism (for/else) to get rid of a variable that
556           was there to determine if something had been found in the loop.
557
558 2010-11-12  Robert Kroeger  <rjkroege@chromium.org>
559
560         Reviewed by James Robinson.
561
562         [Chromium/DRT] Make EventSendingController send complete touch events.
563         https://bugs.webkit.org/show_bug.cgi?id=49285
564
565         * DumpRenderTree/chromium/EventSender.cpp:
566         (EventSender::addTouchPoint):
567         (EventSender::updateTouchPoint):
568         (EventSender::sendCurrentTouchEvent):
569
570 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
571
572         Unreviewed rollout of r71858.
573
574         Rollout out r71858 since it breaks new-run-webkit-httpd as used by the
575         NaCl tests.
576
577         * Scripts/webkitpy/layout_tests/port/config.py:
578         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Removed.
579         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
580
581 2010-11-12  Benjamin Poulain  <benjamin.poulain@nokia.com>
582
583         Reviewed by Kenneth Rohde Christiansen.
584
585         [Qt] TestController::runUntil doesn't honor the timeout parameter
586         https://bugs.webkit.org/show_bug.cgi?id=48941
587
588         Implement the missing timeout of WebKitTestRunner for Qt.
589
590         A QElapsedTimer has been added to RunUntilConditionLoop in order
591         to measure how long the loop has been running. When the timer
592         is bigger or equal than the timeout, the loop ends even if the
593         condition is not met.
594
595         * WebKitTestRunner/qt/TestControllerQt.cpp:
596         (WTR::RunUntilConditionLoop::start):
597         (WTR::RunUntilConditionLoop::run):
598         (WTR::RunUntilConditionLoop::timerEvent):
599         (WTR::TestController::platformRunUntil):
600
601 2010-11-11  David Levin  <levin@chromium.org>
602
603         Reviewed by Shinichiro Hamaji.
604
605         check-webkit-style function detection doesn't handle templates with spaces.
606         https://bugs.webkit.org/show_bug.cgi?id=49427
607
608         * Scripts/webkitpy/style/checkers/cpp.py: Stripped out templates when
609           finding the function name.
610         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test have a
611           template with a space.
612
613 2010-11-11  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
614
615         Reviewed by Shinichiro Hamaji.
616
617         Enable check-webkit-style on Qt files
618         https://bugs.webkit.org/show_bug.cgi?id=48322
619
620         Adding exception for _q_... private slots and
621         ..._data() methods in tests
622
623         * Scripts/webkitpy/style/checker.py: enable Qt dirs, add exceptions
624         * Scripts/webkitpy/style/checker_unittest.py: improve unittests
625
626 2010-11-11  David Levin  <levin@chromium.org>
627
628         Reviewed by Shinichiro Hamaji.
629
630         check-webkit-style function detection doesn't work for templates and destructors.
631         https://bugs.webkit.org/show_bug.cgi?id=49425
632
633         * Scripts/webkitpy/style/checkers/cpp.py: Changed regex for templates and destructors.
634         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test hit this issue.
635
636 2010-11-11  Joone Hur  <joone@kldp.org>
637
638         Reviewed by Antonio Gomes.
639
640         [GTK][DRT] Implement LayoutTestController::nodesFromRect
641         https://bugs.webkit.org/show_bug.cgi?id=46598
642
643         Support nodesFromRect in DRT
644
645         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
646         (LayoutTestController::nodesFromRect):
647
648 2010-11-11  MORITA Hajime  <morrita@google.com>
649
650         Reviewed by Kent Tamura.
651
652         editing/selection/context-menu-on-text.html fails on chromium
653         https://bugs.webkit.org/show_bug.cgi?id=45898
654
655         Enhance Chromium's EventSender::contextClick() to return
656         array of menu strings. The implementation incomplete because
657         Although actual context menu is implemented by the browser,
658         it reflects editability of underlying context thus makes
659         context-menu-on-text.html passed.
660         
661         * DumpRenderTree/chromium/EventSender.cpp:
662         (makeMenuItemStringsFor): Added.
663         (EventSender::contextClick): Now returns array of menu strings.
664         * DumpRenderTree/chromium/WebViewHost.cpp:
665         (WebViewHost::showContextMenu): Now storing given WebContextMenuData object.
666         (WebViewHost::clearContextMenuData): Added.
667         (WebViewHost::lastContextMenuData): Added.
668         * DumpRenderTree/chromium/WebViewHost.h:
669
670 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
671
672         Unreviewed, build fix.
673
674         fix build breakage caused by fix for bug 49380 (r71858).
675
676         config_standalone needs to ensure that the packages it needs are
677         in sys.path.
678
679         https://bugs.webkit.org/show_bug.cgi?id=49419
680
681         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
682
683 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
684
685         Reviewed by Adam Roben.
686
687         Fix NRWT to respect set-webkit-configuration again :(
688
689         This change fixes a typo in config.py that was causing the wrong
690         value to be read initially and us never actually looking into
691         the filesystem to get the default configuration.
692         
693         * Scripts/webkitpy/layout_tests/port/config.py:
694         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
695         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
696
697 2010-11-11  Eric Seidel  <eric@webkit.org>
698
699         Reviewed by Adam Barth.
700
701         Split out Bug, Attachment and CommitterValidator from bugzilla.py
702         https://bugs.webkit.org/show_bug.cgi?id=49403
703
704         CommitterValidator really had nothing to do with bugzilla.py.
705         I've put it next to committers.py for now since its rather tied to that file.
706
707         * Scripts/webkitpy/common/config/committervalidator.py: Added.
708         * Scripts/webkitpy/common/config/committervalidator_unittest.py: Added.
709         * Scripts/webkitpy/common/net/bugzilla/__init__.py:
710         * Scripts/webkitpy/common/net/bugzilla/attachment.py: Added.
711         * Scripts/webkitpy/common/net/bugzilla/bug.py: Added.
712         * Scripts/webkitpy/common/net/bugzilla/bug_unittest.py: Added.
713         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
714         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py:
715         * Scripts/webkitpy/tool/bot/feeders.py:
716         * Scripts/webkitpy/tool/commands/queues.py:
717
718 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
719
720         Reviewed by Eric Seidel.
721
722         fix missing 'pid' argument for http lockfile checking on win32
723         https://bugs.webkit.org/show_bug.cgi?id=49363
724
725         * Scripts/webkitpy/common/system/executive.py:
726
727 2010-11-11  Eric Seidel  <eric@webkit.org>
728
729         Rubber-stamped by Adam Barth.
730
731         Move bugzilla.py into its own module in preparation for splitting one-file-per-class
732         https://bugs.webkit.org/show_bug.cgi?id=49402
733
734         * Scripts/webkitpy/common/net/bugzilla/__init__.py: Added.
735         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla.py.
736         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla_unittest.py.
737
738 2010-11-10  Chris Guillory  <chris.guillory@google.com>
739
740         Reviewed by Chris Fleizach.
741
742         Use IAccessibleComparable to compare accessibility objects.
743         https://bugs.webkit.org/show_bug.cgi?id=49118
744
745         * DumpRenderTree/AccessibilityUIElement.cpp:
746         (AccessibilityUIElement::isEqual):
747         * DumpRenderTree/AccessibilityUIElement.h:
748         * DumpRenderTree/win/AccessibilityUIElementWin.cpp:
749         (comparableObject):
750         (AccessibilityUIElement::isEqual):
751
752 2010-11-10  Csaba Osztrogonác  <ossy@webkit.org>
753
754         Reviewed by David Hyatt.
755
756         HTML5 Ruby support should be mandatory feature
757         https://bugs.webkit.org/show_bug.cgi?id=49272
758
759         Remove Ruby as optional feature.
760
761         * Scripts/build-webkit:
762
763 2010-11-10  Adam Roben  <aroben@apple.com>
764
765         Print VC++ Express build logs in the same order that the projects are
766         built
767
768         Fixes <http://webkit.org/b/49326> print-vse-failure-logs prints logs
769         in an unhelpful order
770
771         Reviewed by Eric Seidel.
772
773         * Scripts/print-vse-failure-logs:
774         (PrintVisualStudioExpressLogs._build_order): Added. Uses
775         print-msvc-project-dependencies to figure out the order in which
776         projects are built.
777         (PrintVisualStudioExpressLogs._sort_buildlogs): Added. Sorts the logs
778         based on their build order and project name.
779         (PrintVisualStudioExpressLogs._obj_directory): Moved code to find the
780         scripts directory from here...
781         (PrintVisualStudioExpressLogs._scripts_directory): ...to here.
782         (PrintVisualStudioExpressLogs.main): Sort the logs before printing
783         them.
784
785 2010-11-10  Adam Roben  <aroben@apple.com>
786
787         Only print the interesting text from VC++ Express build logs
788
789         Fixes <http://webkit.org/b/49325> It's hard to find the interesting
790         output from print-vse-failure-logs
791
792         Reviewed by Eric Seidel.
793
794         * Scripts/print-vse-failure-logs: Removed unnecessary os.path import,
795         added newly-required imports. Removed a comment that was essentially a
796         FIXME, and is now fixed.
797         (PrintVisualStudioExpressLogs._relevant_text): Added. Uses
798         BeautifulSoup to extract the relevant text from the build log.
799         (PrintVisualStudioExpressLogs.main): Only print the relevant text.
800
801 2010-11-09  Mihai Parparita  <mihaip@chromium.org>
802
803         Reviewed by Tony Chang.
804
805         [Chromium] http/tests/security/XFrameOptions fail with Chromium DRT
806         https://bugs.webkit.org/show_bug.cgi?id=49286
807         
808         The tests in http/tests/security/XFrameOptions started to fail after
809         r71297, since it changed the WebViewHost::assignIdentifierToRequest
810         logic to always assign identifiers to requests, even if we're not going
811         to be dumping resource load callbacks at that point. These tests
812         only call LayoutTestController.dumpResourceLoadCallbacks after the
813         request is started, thus their expectations have "<unknown>" as the
814         identifier. Change WebViewHost to the previous behavior, which also
815         match the other ports'.
816
817         * DumpRenderTree/chromium/WebViewHost.cpp:
818         (WebViewHost::assignIdentifierToRequest):
819         (WebViewHost::removeIdentifierForRequest):
820
821 2010-11-08  Tony Chang  <tony@chromium.org>
822
823         Reviewed by Adam Barth.
824
825         run platform/chromium/plugins/return-npobject.html on all platforms
826         https://bugs.webkit.org/show_bug.cgi?id=49103
827
828         This tests that we can get an NPObject returned through a method on
829         an NPAPI Object.
830
831         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
832         (testCallbackReturn):
833         (pluginInvoke):
834
835 2010-11-09  Sheriff Bot  <webkit.review.bot@gmail.com>
836
837         Unreviewed, rolling out r71672.
838         http://trac.webkit.org/changeset/71672
839         https://bugs.webkit.org/show_bug.cgi?id=49276
840
841         It doesn't work on Chromium Windows bot (Requested by Ossy on
842         #webkit).
843
844         * Scripts/webkitpy/common/system/file_lock.py: Removed.
845         * Scripts/webkitpy/common/system/file_lock_unittest.py: Removed.
846         * Scripts/webkitpy/layout_tests/port/http_lock.py:
847
848 2010-11-09  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
849
850         Reviewed by Tony Chang.
851
852         [NRWT] Make http locking similar to perl implementation
853         https://bugs.webkit.org/show_bug.cgi?id=49187
854
855         * Scripts/webkitpy/common/system/file_lock.py: Added.
856         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
857         * Scripts/webkitpy/layout_tests/port/http_lock.py:
858
859 2010-11-09  James Kozianski  <koz@chromium.org>
860
861         Reviewed by Eric Seidel.
862
863         Clean up imports for webkit-patch.
864         https://bugs.webkit.org/show_bug.cgi?id=49083
865
866         Move some imports from commands into main.py where they are used.
867         Move command imports into the __init__ of the commands module.
868
869         * Scripts/webkitpy/tool/commands/__init__.py:
870         * Scripts/webkitpy/tool/commands/download.py:
871         * Scripts/webkitpy/tool/commands/queues.py:
872         * Scripts/webkitpy/tool/main.py:
873
874 2010-11-09  Yi Shen  <yi.4.shen@nokia.com>
875
876         Reviewed by Andreas Kling.
877
878         [Qt][QtTestBrowser] Don't erase incorrect url in the Url Bar
879         https://bugs.webkit.org/show_bug.cgi?id=49047
880
881         * QtTestBrowser/mainwindow.cpp:
882         (MainWindow::setAddressUrl):
883
884 2010-11-09  Andras Becsi  <abecsi@webkit.org>
885
886         Reviewed by Csaba Osztrogonác.
887
888         Improve old-run-webkit-tests --verbose to show DumpTool opening/closing
889         https://bugs.webkit.org/show_bug.cgi?id=49252
890
891         * Scripts/old-run-webkit-tests:
892
893 2010-11-09  Csaba Osztrogonác  <ossy@webkit.org>
894
895         Reviewed by Tor Arne Vestbø.
896
897         Output of old-run-webkit-tests is buggy in verbose mode
898         https://bugs.webkit.org/show_bug.cgi?id=49249
899
900         * Scripts/old-run-webkit-tests: Redundant outputs removed.
901
902 2010-11-08  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
903
904         Reviewed by Martin Robinson.
905
906         [GTK] Link with target name set does not work
907         https://bugs.webkit.org/show_bug.cgi?id=48865
908
909         When a new page is created with a name (target=myFrame), the new
910         mainFrame could not be found because they where not stored in the
911         same PageGroup. As PageGroup are not exposed externally so the
912         simpliest solution is to use a global page group name. This also fixes
913         issue with visited link coloration across pages. After this change the
914         private function webkit_web_view_set_group_name() was no longer used
915         so it was removed completly.
916
917         * DumpRenderTree/gtk/DumpRenderTree.cpp:
918         (createWebView):
919
920 2010-11-08  John Knottenbelt  <jknotten@chromium.org>
921
922         Reviewed by Kent Tamura.
923
924         WebViewHost::reset() uses placement new.
925         https://bugs.webkit.org/show_bug.cgi?id=49069
926
927         WebViewHost is using placement destruction / new to simulate a fresh
928         WebViewHost object at the same address. This is because the WebView remains
929         open across tests and maintains a pointer to the WebViewHost.
930         This change resets member variables explictly instead of the placement new dance.
931
932         * DumpRenderTree/chromium/WebViewHost.cpp:
933         (WebViewHost::WebViewHost):
934         (WebViewHost::reset):
935
936 2010-11-08  Dirk Pranke <dpranke@chromium.org>
937
938         Reviewed by Eric Seidel.
939
940         Fix the webkit_base_dir logic in webkitpy/layout_tests/port/config.py. 
941         It turns out that NRWT can't use the code in scm.find_checkout_root()
942         because the Chromium bots don't do full checkouts of the WebKit
943         tree; they only check out subdirectories like WebKitTools/Scripts.
944         Until we can figure out a better approach for this, I've
945         restored the base_dir-detecting code from NRWT, which works in
946         any directory tree, scm or no.
947
948         This also restores the files modified in r71475 and r71474.
949
950         https://bugs.webkit.org/show_bug.cgi?id=49151
951
952         * WebKitTools/Scripts/webkitpy/layout_tests/port/config.py
953         * WebKitTools/Scripts/webkitpy/layout_tests/port/config_unittest.py
954
955 2010-11-08  Eric Seidel  <eric@webkit.org>
956
957         Reviewed by Adam Barth.
958
959         webkit-patch failure-reason explodes if a build is missing
960         https://bugs.webkit.org/show_bug.cgi?id=49195
961
962         This is likely a recent regression.  Adding a unit test for this case.
963
964         * Scripts/webkitpy/common/net/buildbot.py:
965         * Scripts/webkitpy/common/net/buildbot_unittest.py:
966
967 2010-11-08  Dan Bernstein  <mitz@apple.com>
968
969         Reviewed by Darin Adler.
970
971         Ignore the system scrollbar setting.
972
973         * DumpRenderTree/mac/DumpRenderTree.mm:
974         (resetDefaultsToConsistentValues): Set the scrollbar default.
975
976 2010-11-05  Adam Roben  <aroben@apple.com>
977
978         Make webkitpy.common.system.executive_unittest pass when running under
979         Win32 Python
980
981         Fixes <http://webkit.org/b/49033>.
982
983         Reviewed by Dave Levin and Eric Seidel.
984
985         * Scripts/webkitpy/common/system/executive.py:
986         (Executive._run_command_with_teed_output): Pass the arguments through
987         encode_argument_if_needed rather than using Cygwin-specific code here.
988         (Executive.run_and_throw_if_fail): Use child_process_encoding to decode
989         the output.
990         (Executive.run_command): Use encode_argument_if_needed to encode the
991         arguments and child_process_encoding to decode the output.
992         (Executive._child_process_encoding): Returns the encoding that should be
993         used when communicating with child processes. On Windows we use mbcs,
994         which maps to the current code page. On all other platforms we use
995         UTF-8.
996         (Executive._should_encode_child_process_arguments): Returns True if
997         arguments to child processes need to be encoded. This is currently
998         only needed on Cygwin and Win32 Python 2.x.
999         (Executive._encode_argument_if_needed): Encode the argument using
1000         child_process_encoding if we need to encode arguments to child
1001         processes on this platform.
1002
1003         * Scripts/webkitpy/common/system/executive_unittest.py:
1004         (never_ending_command): Added. Returns arguments to run a command that
1005         will not quit until we kill it. On Windows we use wmic, on other
1006         platforms we use yes.
1007         (ExecutiveTest.test_run_command_with_unicode): Changed to expect the
1008         mbcs encoding to be used and for output from the child processes to
1009         have been roundtripped through encode/decode on Win32 Python. When
1010         UTF-8 is the encoding the roundtripping is undetectable, but with mbcs
1011         it's possible that some characters will not be able to be converted
1012         and will be replaced by question marks; the round-tripping allows us
1013         to expect this result.
1014
1015         (ExecutiveTest.test_kill_process):
1016         (ExecutiveTest.test_kill_all):
1017         Use never_ending_command instead of invoking "yes" directly. Expect an
1018         exit code of 1 when using Win32 Python, as that's what seems to happen.
1019
1020 2010-11-08  Adam Roben  <aroben@apple.com>
1021
1022         Roll out r71532
1023
1024         It broke the build for Cygwin 1.7 installs. Cygwin 1.7's default
1025         .bashrc unsets %TEMP%, which broke copy-tools.cmd.
1026
1027         * Scripts/webkitdirs.pm:
1028
1029 2010-11-08  Tony Chang  <tony@chromium.org>
1030
1031         Reviewed by Adam Barth.
1032
1033         run platform/chromium/plugins/nested-plugin-objects.html on all platforms
1034         https://bugs.webkit.org/show_bug.cgi?id=49094
1035
1036         This tests that objects created by plugins are proplery cleaned up.
1037
1038         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1039         (testPassTestObject):
1040         (pluginInvoke):
1041         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
1042         (testAllocate):
1043         (testDeallocate):
1044         (testGetProperty):
1045         (testConstruct):
1046
1047 2010-11-08  Adam Roben  <aroben@apple.com>
1048
1049         Mark Windows builds triggered from Perl as being non-interactive
1050
1051         This affects whether some of our scripts will show alerts vs. printing
1052         to the build log.
1053
1054         Fixes <http://webkit.org/b/49181> Windows build fail mysteriously when
1055         .vsprops files are updated
1056
1057         Reviewed by Steve Falkenburg.
1058
1059         * Scripts/webkitdirs.pm:
1060         (buildVisualStudioProject): Set WEBKIT_NONINTERACTIVE_BUILD to 1.
1061
1062 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1063
1064         Reviewed by Ojan Vafai.
1065
1066         Make http locking default in NRWT.
1067         https://bugs.webkit.org/show_bug.cgi?id=48053
1068
1069         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1070         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1071
1072 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1073
1074         Reviewed by Csaba Osztrogonác.
1075
1076         [NRWT] If the http lock fails we shouldn't do any locking
1077         https://bugs.webkit.org/show_bug.cgi?id=49164
1078
1079         If something goes wrong with the locking, the test should keep going.
1080
1081         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1082
1083 2010-11-08  Adam Roben  <aroben@apple.com>
1084
1085         Switch back to using kCGImageAlphaPremultipliedFirst when generating
1086         pixel dumps on Windows
1087
1088         I changed this behavior in r71418 thinking that it was required for
1089         getting plugins to show up in pixel dumps. But it doesn't seem to be
1090         necessary, and was making it impossible to compare new Windows pixel
1091         dumps with existing Windows or Mac pixel dumps (because ImageDiff won't
1092         compare an image with alpha to an image without alpha).
1093
1094         Fixes <http://webkit.org/b/49172> REGRESION (r71418): Can't compare
1095         new Windows pixel results to existing Windows or Mac results
1096
1097         Reviewed by Antti Koivisto.
1098
1099         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
1100         (createBitmapContextFromWebView): Replaced kCGImageAlphaNoneSkipFirst
1101         with kCGImageAlphaPremultipliedFirst.
1102
1103 2010-11-08  Csaba Osztrogonac  <ossy@webkit.org>
1104
1105         Unreviewed, rolling out r71466.
1106         http://trac.webkit.org/changeset/71466
1107         https://bugs.webkit.org/show_bug.cgi?id=48865
1108
1109         It broke layout tests on GTK bots.
1110
1111         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1112         (createWebView):
1113
1114 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1115
1116         Reviewed by Csaba Osztrogonác.
1117
1118         Enable running of Qt API tests on BuildBot
1119         https://bugs.webkit.org/show_bug.cgi?id=49004
1120
1121         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1122
1123 2010-11-08  Eric Seidel  <eric@webkit.org>
1124
1125         Reviewed by Adam Barth.
1126
1127         Add clean-review-queue command to remove closed bugs from the webkit.org/pending-review
1128         https://bugs.webkit.org/show_bug.cgi?id=49160
1129
1130         Bugzilla doesn't automatically remove r? when a bug gets closed.
1131         This script takes care of that for webkit.org.
1132
1133         * Scripts/webkitpy/common/net/bugzilla.py:
1134         * Scripts/webkitpy/tool/commands/upload.py:
1135
1136 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
1137
1138         Unreviewed, rolling out r71474.
1139         http://trac.webkit.org/changeset/71474
1140         https://bugs.webkit.org/show_bug.cgi?id=48280
1141
1142         breaks chromium webkit tests
1143         https://bugs.webkit.org/show_bug.cgi?id=49151
1144
1145         * Scripts/webkitpy/layout_tests/port/base.py:
1146         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
1147         * Scripts/webkitpy/layout_tests/port/config.py:
1148         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1149         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
1150         * Scripts/webkitpy/layout_tests/port/webkit.py:
1151         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1152         * Scripts/webkitpy/tool/mocktool.py:
1153
1154 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
1155
1156         Unreviewed, rolling out r71475.
1157         http://trac.webkit.org/changeset/71475
1158  
1159         breaks chromium webkit tests
1160         https://bugs.webkit.org/show_bug.cgi?id=49151
1161
1162         * Scripts/webkitpy/common/newstringio.py: Removed.
1163         * Scripts/webkitpy/common/newstringio_unittest.py: Removed.
1164         * Scripts/webkitpy/common/system/executive_mock.py: Removed.
1165         * Scripts/webkitpy/common/system/filesystem_mock.py: Removed.
1166         * Scripts/webkitpy/layout_tests/port/config_mock.py: Removed.
1167
1168 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1169
1170         Unreviewed, build fix.
1171
1172         Add files inexplicably not committed in r71474 as part of the
1173         fix for bug 48280.
1174
1175         * Scripts/webkitpy/common/newstringio.py: Added.
1176         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
1177         * Scripts/webkitpy/common/system/executive_mock.py: Added.
1178         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
1179         * Scripts/webkitpy/layout_tests/port/config_mock.py: Added.
1180
1181 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1182
1183         Reviewed by Eric Siedel.
1184
1185         new-run-webkit-tests: update port/base and port/webkit to use the
1186         new FileSystem and Config abstractions, pulling more logic out of
1187         the base Port classes into separate, mockable objects.
1188
1189         Also create a MockFileSystem object, a MockConfig object, move
1190         MockExecutive into common/system to be next to executive, and
1191         update the config object to use a FileSystem.
1192
1193         https://bugs.webkit.org/show_bug.cgi?id=48280
1194
1195         * Scripts/webkitpy/common/newstringio.py: Added.
1196         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
1197         * Scripts/webkitpy/common/system/executive_mock.py: Added.
1198         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
1199         * Scripts/webkitpy/layout_tests/port/base.py:
1200         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
1201         * Scripts/webkitpy/layout_tests/port/config.py:
1202         * Scripts/webkitpy/layout_tests/port/config_mock.py:
1203         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1204         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
1205         * Scripts/webkitpy/layout_tests/port/webkit.py:
1206         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1207         * Scripts/webkitpy/tool/mocktool.py:
1208
1209 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1210
1211         Unreviewed, build breakage.
1212
1213         Apparently I uploaded the wrong version of the file to fix 49122
1214         and neither Eric or I noticed - it was missing a dirname() call.
1215         Fixing ...
1216
1217         https://bugs.webkit.org/show_bug.cgi?id=49122
1218
1219         * Scripts/webkitpy/common/checkout/scm.py:
1220
1221 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1222
1223         Reviewed by Eric Seidel.
1224
1225         webkitpy/tool/* unittests change cwd and don't clean up properly
1226
1227         https://bugs.webkit.org/show_bug.cgi?id=49122
1228
1229         * Scripts/webkitpy/common/checkout/scm.py:
1230
1231 2010-11-05  Chris Marrin  <cmarrin@apple.com>
1232
1233         Reviewed by Simon Fraser.
1234
1235         suspendAnimations/resumeAnimations not present in WebKit2
1236         https://bugs.webkit.org/show_bug.cgi?id=49109
1237
1238         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
1239         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
1240         (WTR::LayoutTestController::suspendAnimations):
1241         (WTR::LayoutTestController::resumeAnimations):
1242         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
1243
1244 2010-11-05  Tony Chang  <tony@chromium.org>
1245
1246         Reviewed by David Levin.
1247
1248         cleanup style in TestNetscapePlugIn/PluginObject.cpp
1249         https://bugs.webkit.org/show_bug.cgi?id=49044
1250
1251         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1252         (getPluginClass):
1253         (pluginGetProperty):
1254         (pluginSetProperty):
1255         (variantToIdentifier):
1256         (testCallback):
1257         (getURL):
1258         (getURLNotify):
1259         (testInvokeDefault):
1260         (testGetProperty):
1261         (testDocumentOpen):
1262         (testWindowOpen):
1263         (pluginInvoke):
1264         (handleCallback):
1265         (notifyStream):
1266         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
1267
1268 2010-11-05  Eric Seidel  <eric@webkit.org>
1269
1270         Reviewed by David Levin.
1271
1272         Add QueueStatusServer/__init__.py so others can run the QueueStatusServer tests
1273         https://bugs.webkit.org/show_bug.cgi?id=49032
1274
1275         I wrote this file as part of bug 47847, but I forgot to commit it.
1276         No one else noticed it missing because test-webkitpy knows how
1277         to recover in the case where it can't import QueueStatusServer
1278         (which generally occurs due to not having installed the AppEngine SDK).
1279
1280         * QueueStatusServer/__init__.py: Added.
1281         * QueueStatusServer/model/workitems_unittest.py:
1282          - Remove a test which fails.  This was probably landed (by me)
1283            from my other machine, which since this __init__.py was missing
1284            I never noticed the failure and landed this invalid test.
1285            Sadly we can't really test remove_work_item as it depends
1286            on .key() working.  .key() will throw unless the object
1287            has already been saved it seems.
1288            This may be a mis-design in our remove_work_item implementation,
1289            but for now, just removing the test.
1290
1291 2010-11-04  Adam Roben  <aroben@apple.com>
1292
1293         Teach check-webkit-style about TestNetscapePlugIn
1294
1295         Fixes <http://webkit.org/b/49030> check-webkit-style is ignorant of
1296         TestNetscapePlugIn's coding conventions
1297
1298         Reviewed by John Sullivan.
1299
1300         * Scripts/webkitpy/style/checker.py: Lump TestNetscapePlugIn in with
1301         WebKitAPITest and TestWebKitAPI in having different include order and
1302         naming conventions than WebCore. Added some comments about why the
1303         exceptions exist.
1304
1305 2010-11-04  Adam Roben  <aroben@apple.com>
1306
1307         Add a test that shows that windowed plugins are able to paint
1308
1309         Somehow we never had a test for this before.
1310
1311         Fixes <http://webkit.org/b/49029> Should add a test that shows
1312         windowed plugins can paint
1313
1314         Reviewed by Jon Honeycutt.
1315
1316         * DumpRenderTree/TestNetscapePlugIn/Tests/win/DrawsGradient.cpp: Added.
1317         (DrawsGradient::DrawsGradient):
1318         (DrawsGradient::wndProc): We handle the WM_PAINT and WM_PRINTCLIENT messages.
1319
1320         (DrawsGradient::onPaint):
1321         (DrawsGradient::onPrintClient):
1322         These both just call through to paint.
1323
1324         (DrawsGradient::paint): Fills our client area with some gradients.
1325
1326         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Link
1327         against Msimg32.lib for ::GradientFill and added DrawsGradient.
1328
1329         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
1330         (createBitmapContextFromWebView): Use WM_PRINT instead of
1331         WM_PRINTCLIENT so that ::DefWindowProc will send
1332         WM_PRINT/WM_PRINTCLIENT messages to the WebView's child windows.
1333         Replaced kCGImageAlphaPremultipledFirst with kCGImageAlphaNoneSkipFirst
1334         because GDI doesn't support alpha and kCGImageBitmapOrder32Little with
1335         kCGImageBitmapOrder32Host because, who knows, maybe someday Windows
1336         will run on a big-endian platform.
1337
1338 2010-11-04  Adam Roben  <aroben@apple.com>
1339
1340         Extract much of NPNInvalidateRectInvalidatesWindow's code into a
1341         WindowedPluginTest base class
1342
1343         The base class takes care of subclassing the plugin's window so that a
1344         custom WNDPROC is called. This will make it easier to write tests that
1345         need to handle window messages.
1346
1347         Fixes <http://webkit.org/b/49028> It's hard to write a PluginTest with
1348         a custom WNDPROC
1349
1350         Reviewed by Jon Honeycutt.
1351
1352         * DumpRenderTree/TestNetscapePlugIn/Tests/win/NPNInvalidateRectInvalidatesWindow.cpp:
1353         (NPNInvalidateRectInvalidatesWindow::NPNInvalidateRectInvalidatesWindow):
1354         (NPNInvalidateRectInvalidatesWindow::NPP_SetWindow):
1355         (NPNInvalidateRectInvalidatesWindow::wndProc):
1356         (NPNInvalidateRectInvalidatesWindow::testInvalidateRect):
1357         Moved code from here to WindowedPluginTest. Changed to use window()
1358         instead of m_window.
1359
1360         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1361         Added WindowedPluginTest and added TestNetscapePlugIn/win to the
1362         include path. Also reordered the include path to make a little more
1363         sense and simplified the entry that added TestNetscapePlugIn itself to
1364         the include path.
1365
1366         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.cpp: Added.
1367         (WindowedPluginTest::WindowedPluginTest):
1368         (WindowedPluginTest::NPP_SetWindow):
1369         (WindowedPluginTest::staticWndProc):
1370         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.h: Added.
1371         (WindowedPluginTest::window):
1372         Code came from NPNInvalidateRectInvalidatesWindow.
1373
1374 2010-11-04  Adam Roben  <aroben@apple.com>
1375
1376         Add a plugin test to show that windowed plugins are clipped correctly
1377
1378         Fixes <http://webkit.org/b/49024> <rdar://problem/8487847> Windowed
1379         plugins aren't clipped in WebKit2 on Windows
1380
1381         Reviewed by Jon Honeycutt.
1382
1383         * DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowRegionIsSetToClipRect.cpp: Added.
1384         (WindowRegionIsSetToClipRect::WindowRegionIsSetToClipRect): Initialize members.
1385         (WindowRegionIsSetToClipRect::NPP_SetWindow): Check that our window
1386         region matches the clip rect we know we should have based on
1387         window-region-is-set-to-clip-rect.html, and check that our window class
1388         doesn't have the CS_PARENTDC style.
1389
1390         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1391         Added WindowRegionIsSetToClipRect.
1392
1393 2010-11-05  Alejandro G. Castro  <alex@igalia.com>
1394
1395         Reviewed by Martin Robinson.
1396
1397         [GTK] Avoid font hinting in the DRT
1398         https://bugs.webkit.org/show_bug.cgi?id=48548
1399
1400         Change the settings to avoid font hinting, it was causing
1401         different results depending on the distribution.
1402
1403         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1404         (initializeGtkFontSettings):
1405         * DumpRenderTree/gtk/fonts/fonts.conf:
1406
1407 2010-11-05  Eric Seidel  <eric@webkit.org>
1408
1409         Reviewed by Adam Barth.
1410
1411         build-webkit spams system.log by repeatedly checking for PBXProductDirectory in com.apple.Xcode
1412         https://bugs.webkit.org/show_bug.cgi?id=49051
1413
1414         This is a speculative fix.  The unit tests cover these methods, however
1415         I don't know if this will fully stop the system.log spam.
1416
1417         * Scripts/webkitpy/layout_tests/port/base.py:
1418         * Scripts/webkitpy/layout_tests/port/webkit.py:
1419
1420 2010-11-05  Eric Seidel  <eric@webkit.org>
1421
1422         Reviewed by Adam Barth.
1423
1424         Add basic support for showing bot id on /queue-status/ pages
1425         https://bugs.webkit.org/show_bug.cgi?id=49037
1426
1427         This support is really simple.  Eventually we'll want to
1428         show the bot id in the lock table too, but we don't have
1429         that information stored in the server yet.
1430
1431         * QueueStatusServer/handlers/queuestatus.py:
1432         * QueueStatusServer/handlers/queuestatus_unittest.py: Added.
1433         * QueueStatusServer/templates/includes/singlequeuestatus.html:
1434
1435 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
1436
1437         Reviewed by Adam Barth.
1438
1439         Create a filesystem wrapper that we can use to enforce
1440         particular conventions and use for mocking and dependency
1441         injection down the line.
1442
1443         https://bugs.webkit.org/show_bug.cgi?id=48144
1444
1445         * Scripts/webkitpy/common/system/filesystem.py: Added.
1446         * Scripts/webkitpy/common/system/filesystem_unittest.py: Added.
1447
1448 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1449
1450         Reviewed by Tony Chang.
1451
1452         run_webkit_tests_unittest fails under Python 2.5
1453         https://bugs.webkit.org/show_bug.cgi?id=49043
1454         
1455         Switch from itertools.chain.from_iterable (which was added in 2.6)
1456         to using itertools.chain directly.
1457
1458         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1459
1460 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1461
1462         Unreviewed fix to webkit-patch rebaseline-server so that it runs under
1463         Python 2.5 (it needs an import to use the with statement).
1464
1465         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
1466
1467 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
1468
1469         Reviewed by Adam Barth.
1470
1471         new-run-webkit-tests: split out webkit-specific configuration stuff into a new module
1472
1473         The current NRWT code has webkit-specific configuration code (like
1474         _script_path, default configuration, etc.) mixed in with
1475         layout-test-specific stuff in port/base. The configuration code
1476         should be split out into a separate module for easier mocking,
1477         testing, and isolation.
1478
1479         https://bugs.webkit.org/show_bug.cgi?id=48264
1480
1481         * Scripts/webkitpy/layout_tests/port/config.py: Added.
1482         * Scripts/webkitpy/layout_tests/port/config_unittest.py: Added.
1483
1484 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1485
1486         Reviewed by Tony Chang.
1487
1488         Rebaseline server: initial framework
1489         https://bugs.webkit.org/show_bug.cgi?id=48892
1490         
1491         Adds the basic framework for the rebaseline server (details at 
1492         http://webkit.org/b/47761). Includes the rebaseline-server webkit-patch
1493         command, which starts an HTTP server that can serve static files or
1494         invoke handler methods on a class.
1495
1496         * Scripts/webkitpy/tool/commands/__init__.py:
1497         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html: Added.
1498         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css: Added.
1499         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js: Added.
1500         * Scripts/webkitpy/tool/commands/rebaselineserver.py: Added.
1501
1502 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
1503
1504         Reviewed by James Robinson.
1505
1506         new-run-webkit-tests wasn't using DRT by default for
1507         --platform chromium-gpu
1508
1509         The default value was set to False instead of None, which meant
1510         that the platform specific logic wasn't firing to change the
1511         value to True (b/c we were afraid we'd be overriding the user
1512         preference).
1513
1514         https://bugs.webkit.org/show_bug.cgi?id=49038
1515
1516         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1517
1518 2010-11-04  Tony Chang  <tony@chromium.org>
1519
1520         Reviewed by Adam Barth.
1521
1522         make platform/chromium/plugins/multiple-plugins.html pass on all platforms
1523         https://bugs.webkit.org/show_bug.cgi?id=49026
1524
1525         Have the testObject.bar property return the string "bar".  This lets
1526         us run and pass platform/chromium/plugins/multiple-plugins.html on all
1527         platforms.
1528
1529         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
1530         (testGetProperty):
1531
1532 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1533
1534         Reviewed by Tony Chang.
1535
1536         NRWT doesn't have good test coverage for --run-chunk, --batch-size, --run-part, etc.
1537         https://bugs.webkit.org/show_bug.cgi?id=48878
1538         
1539         Add get_tests_run so that it's easy to see which tests get run (and with
1540         what batching) for a given flag combination. Flesh out the various
1541         test cases that have FIXMEs.
1542         
1543         Also fixes an off-by-one error (batch sizes were one larger than 
1544         expected) and makes --run-part also have wraparound behavior, like
1545         --run-chunk.
1546
1547         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1548         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1549         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1550
1551 2010-11-04  Erik Arvidsson  <arv@chromium.org>
1552
1553         Reviewed by Dimitri Glazkov.
1554
1555         Support box-sizing without the vendor prefix
1556         https://bugs.webkit.org/show_bug.cgi?id=36713
1557
1558         Based on patch by Peter Beverloo <peter@lvp-media.com>
1559
1560         * iExploder/htdocs/cssproperties.in: Change -webkit-box-sizing to box-sizing.
1561
1562 2010-11-04  Csaba Osztrogonác  <ossy@webkit.org>
1563
1564         Unreviewed rollout r71340, because it broke Chromium Windows bot.
1565
1566         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1567         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1568
1569 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1570
1571         Reviewed by Ojan Vafai.
1572
1573         Make http locking default in NRWT.
1574         https://bugs.webkit.org/show_bug.cgi?id=48053
1575
1576         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1577         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1578
1579 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1580
1581         Reviewed by Eric Seidel.
1582
1583         [NRWT] Clear invalid http locks on Windows platform as well
1584         https://bugs.webkit.org/show_bug.cgi?id=48515
1585
1586         * Scripts/webkitpy/common/system/executive.py:
1587         * Scripts/webkitpy/common/system/executive_unittest.py:
1588         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1589
1590 2010-11-02  Adam Roben  <aroben@apple.com>
1591
1592         Reduce our dependence on coreutils when running Python tests
1593
1594         This patch introduces versions of the standard echo and cat utilities
1595         implemented in Python. They are probably missing features of their
1596         coreutils equivalents, but they can do what's necessary for our Python
1597         tests. This is useful on Windows, where these utilities typically
1598         aren't available.
1599
1600         Fixes <http://webkit.org/b/48883> executive_unittest relies on echo
1601         and cat utilities from coreutils, which are not present on Windows
1602
1603         Reviewed by Eric Seidel.
1604
1605         * Scripts/webkitpy/common/system/executive_unittest.py: Changed to use
1606         our Python-based echo and cat.
1607
1608         * Scripts/webkitpy/common/system/fileutils.py: Added.
1609         (make_stdout_binary): On Windows, puts sys.stdout into binary mode so
1610         that \n won't be translated into \r\n. I couldn't think of a good way
1611         to test this directly without touching the filesystem, but it is tested
1612         indirectly by echo_unittest.
1613
1614         * Scripts/webkitpy/test/cat.py: Added.
1615         (command_arguments): Returns a list for invoking cat with the given arguments.
1616         (main): Acts like a simplified version of the coreutils cat utility.
1617
1618         * Scripts/webkitpy/test/cat_unittest.py: Added.
1619         (CatTest.assert_cat): Runs cat with the given input and ensures the
1620         output matches the input.
1621         (CatTest.test_basic): Performs a simple test of cat.
1622         (CatTest.test_no_newline): Tests what happens when the input string
1623         doesn't have a trailing newline.
1624         (CatTest.test_unicode): Tests passing a unicode string to cat.
1625         (CatTest.test_as_command): Tests running cat as a separate command.
1626
1627         * Scripts/webkitpy/test/echo.py: Added.
1628         (command_arguments): Returns a list for invoking echo with the given arguments.
1629         (main): Acts like a simplified version of the coreutils echo utility.
1630
1631         * Scripts/webkitpy/test/echo_unittest.py: Added.
1632         (EchoTest.test_basic): Performs a simple test of echo.
1633         (EchoTest.test_no_newline): Tests passing -n to echo to suppress the
1634         trailing newline.
1635         (EchoTest.test_unicode): Tests passing unicode and non-unicode strings
1636         to echo.
1637         (EchoTest.test_argument_order): Tests what happens when -n is not the
1638         first argument.
1639         (EchoTest.test_empty_arguments): Tests what happens when you pass [] to
1640         echo.main.
1641         (EchoTest.test_no_arguments): Tests what happens when you call
1642         echo.main with no arguments.
1643         (EchoTest.test_as_command): Tests running echo as a separate command.
1644
1645 2010-11-04  Renata Hodovan  <reni@inf.u-szeged.hu>
1646
1647         Unreviewed: Add myself to the list of Committers.
1648
1649         * Scripts/webkitpy/common/config/committers.py:
1650
1651 2010-11-04  Andreas Kling  <kling@webkit.org>
1652
1653         Reviewed by Kenneth Rohde Christiansen.
1654
1655         [Qt] Hook into QWebPage::scrollRequested for resizes-to-contents mode
1656
1657         In RTC mode the QGraphicsWebView item is the size of the contents,
1658         scrolling works a bit differently  (we need to react to scrollRequested.)
1659
1660         Normally QGraphicsView will replay the last mouse event when scrolling,
1661         so to prevent WebKit from getting confused by this we temporarily make
1662         the QGraphicsView non-interactive.
1663
1664         * QtTestBrowser/webview.cpp:
1665         (WebViewGraphicsBased::setPage):
1666         (WebViewGraphicsBased::scrollRequested):
1667         * QtTestBrowser/webview.h:
1668
1669 2010-11-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
1670
1671         Reviewed by Andreas Kling.
1672
1673         Use OS(MAC_OS_X) rather than PLATFORM(MAC) in TestControllerQt
1674
1675         PLATFORM(MAC) is not defined for the Qt port, as it refers to
1676         the Mac-port, not the Mac OS X operating system.
1677
1678         * WebKitTestRunner/qt/TestControllerQt.cpp:
1679         (WTR::isExistingLibrary):
1680
1681 2010-11-04  Leonid Ebril  <leonid.ebril@nokia.com>
1682
1683         Reviewed by Andreas Kling.
1684
1685         [Qt] Adding iphone user agent string the user agent list for QtTestBrowser
1686         https://bugs.webkit.org/show_bug.cgi?id=48869
1687
1688         * QtTestBrowser/useragentlist.txt:
1689
1690 2010-11-03  Adam Roben  <aroben@apple.com>
1691
1692         Always use uppercase drive names in strings returned by abspath_to_uri
1693
1694         Some versions of cygpath use lowercase drive letters while others use
1695         uppercase, which makes it hard to test the output of code that uses
1696         cygpath.
1697
1698         Fixes <http://webkit.org/b/48914> webkitpy.common.system.path_unittest
1699         fails with Cygwin 1.5
1700
1701         Reviewed by Eric Seidel.
1702
1703         * Scripts/webkitpy/common/system/path.py:
1704         (cygpath): Updated the docstring to indicate that only absolute paths
1705         should be passed for now (though relative paths will work fine).
1706         (_Cygpath.convert): Upper-case the first letter of the converted Windows path.
1707
1708 2010-11-03  George Guo  <George.Guo@Nokia.com>
1709
1710         Reviewed by Andreas Kling.
1711
1712         [Qt] QtTestBrowser : set mmp rule pageddata in Symbian
1713         https://bugs.webkit.org/show_bug.cgi?id=48767
1714     
1715         Paging is needd on Symbian devices to support benchmarks tests like
1716         dromaeo.com and Celtic Kane that need a lot of memory to run  
1717
1718         * QtTestBrowser/QtTestBrowser.pro:
1719
1720 2010-11-03  Jenn Braithwaite  <jennb@chromium.org>
1721
1722         Reviewed by Dmitry Titov.
1723
1724         Chromium: Update resource tracking when moving a frame between documents
1725         https://bugs.webkit.org/show_bug.cgi?id=48363
1726
1727         * DumpRenderTree/chromium/WebViewHost.cpp:
1728         (WebViewHost::assignIdentifierToRequest):
1729         Always put resource id in map so we can make assumptions about its
1730         presence.
1731         (WebViewHost::removeIdentifierForRequest):
1732         Added.
1733         (WebViewHost::didFinishResourceLoad):
1734         (WebViewHost::didFailResourceLoad):
1735         Check existence of resource id before removing from map.
1736         * DumpRenderTree/chromium/WebViewHost.h:
1737
1738 2010-11-03  Victor Wang  <victorw@chromium.org>
1739
1740         Reviewed by Adam Barth.
1741
1742         [Chromium] update buildbot names in chromium rebaseline tool.
1743
1744         https://bugs.webkit.org/show_bug.cgi?id=48881
1745
1746         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
1747
1748 2010-11-03  Darin Adler  <darin@apple.com>
1749
1750         Updated Xcode projects by opening them with Xcode 3.2.4.
1751         Updated svn:ignore for Xcode projects.
1752
1753         * MiniBrowser/MiniBrowser.xcodeproj: Added property svn:ignore.
1754         * TestWebKitAPI/TestWebKitAPI.xcodeproj: Added property svn:ignore.
1755         * WebKitLauncher/WebKitLauncher.xcodeproj: Modified property svn:ignore.
1756         * WebKitLauncher/WebKitLauncher.xcodeproj/project.pbxproj: Updated with Xcode 3.2.4.
1757
1758 2010-11-03  Adam Roben  <aroben@apple.com>
1759
1760         Stop waiting for messages from the web process after a timeout elapses
1761
1762         TestController::runUntil now takes a timeout parameter that specifies
1763         how long to wait for the condition to become true. WebKitTestRunner
1764         takes a --timeout flag that specifies how long the timeout should be.
1765         run-webkit-tests passes this flag to WebKitTestRunner so its timeout
1766         will be similar to run-webkit-tests.
1767
1768         Fixes <http://webkit.org/b/43047> <rdar://problem/8365833>
1769         WebKitTestRunner waits forever if the web process crashes
1770
1771         Reviewed by Darin Adler and Anders Carlsson.
1772
1773         * Scripts/old-run-webkit-tests:
1774         (top level): Moved the GuardMalloc timeout adjustment here from
1775         readFromDumpToolWithTimer.
1776         (openDumpTool): Make WTR use a timeout similar to but slightly shorter
1777         than the one that was specified on the command line.
1778
1779         * WebKitTestRunner/TestController.cpp:
1780         (WTR::TestController::TestController): Initialize our timeout values.
1781         (WTR::TestController::initialize): Parse the --timeout flag and use it
1782         to modify our timeout values.
1783         (WTR::TestController::resetStateToConsistentValues): Changed to use a
1784         short timeout while waiting for the web process to reset and to return
1785         a boolean indicating whether we were able to reset the web process.
1786         Uses a 5-second timeout while waiting for the process to be reset.
1787         (WTR::TestController::runTest): Changed to return a boolean indicating
1788         whether we were able to reset the web process (and thus run the test).
1789         (WTR::TestController::runUntil): Call through to platformRunUntil.
1790
1791         (WTR::TestController::runTestingServerLoop):
1792         (WTR::TestController::run):
1793         Changed to bail if any test can't be run. This will cause the process
1794         to exit. (Unfortunately this will make run-webkit-tests think we
1795         crashed; see <http://webkit.org/b/48943>.)
1796
1797         * WebKitTestRunner/TestController.h: Added platformRunUntil,
1798         m_longTimeout, and m_shortTimeout.
1799
1800         * WebKitTestRunner/TestInvocation.cpp:
1801         (WTR::TestInvocation::invoke): Use a short timeout when waiting for the
1802         initial response and a long timeout when waiting for the test to
1803         complete. Check whether runUntil timed out and print an error message
1804         if so.
1805
1806         * WebKitTestRunner/mac/TestControllerMac.mm:
1807         (WTR::TestController::platformRunUntil): Renamed from runUntil. Pass
1808         [NSDate distantPast] to -[NSRunLoop runMode:beforeDate:] so that we
1809         won't block waiting for the run loop. Only loop until the timeout
1810         elapses.
1811
1812         * WebKitTestRunner/qt/TestControllerQt.cpp:
1813         (WTR::TestController::platformRunUntil): Renamed from runUntil. Added a
1814         FIXME about honoring the timeout.
1815
1816         * WebKitTestRunner/win/TestControllerWin.cpp:
1817         (WTR::TestController::platformRunUntil): Renamed from runUntil. Use
1818         ::MsgWaitForMultipleObjectsEx to implement the timeout. Changed to use
1819         ::PeekMessageW so that we don't block waiting for messages to become
1820         available.
1821
1822 2010-11-03  Adam Roben  <aroben@apple.com>
1823
1824         Add a plugin test that evaluates JS after removing the plugin element
1825         from the document
1826
1827         This test replaces platform/win/plugins/plugin-delayed-destroy.html.
1828         That test was made to prevent a crash very similar to this one, but
1829         unfortunately tested only the mechanism that prevented the crash and
1830         not whether the crash itself was prevented. Since WebKit2 uses a
1831         different mechanism to prevent the crash, the test was failing even
1832         though WebKit2 was not vulnerable to the crash. This new test crashes
1833         if there is no mechanism in place to prevent it and passes in both
1834         WebKit1 and WebKit2.
1835
1836         Fixes <http://webkit.org/b/46711> <rdar://problem/8485903>
1837         platform/win/plugins/plugin-delayed-destroy.html fails in WebKit2
1838
1839         Reviewed by Anders Carlsson.
1840
1841         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
1842         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1843         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
1844         * GNUmakefile.am:
1845         Added new file.
1846
1847         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1848         (pluginDeallocate): Make sure we delete the PluginTest object. This
1849         prevents a leak and also allows us to test the crash.
1850
1851         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
1852         (PluginTest::executeScript): Made this into a non-static member
1853         function.
1854
1855         (PluginTest::waitUntilDone):
1856         (PluginTest::notifyDone):
1857         Updated for changes to executeScript.
1858
1859         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Added executeScript.
1860
1861         * DumpRenderTree/TestNetscapePlugIn/Tests/EvaluateJSAfterRemovingPluginElement.cpp: Added.
1862         (EvaluateJSAfterRemovingPluginElement::EvaluateJSAfterRemovingPluginElement):
1863         Initialize ourselves and tell the test harness to wait.
1864         (EvaluateJSAfterRemovingPluginElement::NPP_DestroyStream): Remove our
1865         plugin element from the document, then execute some JavaScript. If
1866         WebKit does not have appropriate mechanisms in place, we'll be
1867         destroyed inside the first call to executeScript and crash on the
1868         second call.
1869
1870 2010-11-02  Stephen White  <senorblanco@chromium.org>
1871
1872         Reviewed by Tony Chang.
1873
1874         [chromium] Fix LayoutTestController UMRs.
1875         https://bugs.webkit.org/show_bug.cgi?id=48872
1876
1877         * DumpRenderTree/chromium/LayoutTestController.cpp:
1878         (LayoutTestController::LayoutTestController):
1879
1880 2010-11-03  Kent Tamura  <tkent@chromium.org>
1881
1882         Reviewed by Dimitri Glazkov.
1883
1884         REGRESSION: rebaseline-chromium-webkit-tests uses non-zero tolerance for
1885         image dup detection
1886         https://bugs.webkit.org/show_bug.cgi?id=48744
1887
1888         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
1889          - Make a function for option parsing for ease of test
1890          - Set 0 to options.tolerance
1891         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
1892          - Add a test for this change
1893
1894 2010-11-02  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
1895
1896         Unreviewed: Add myself to the list of Committers.
1897
1898         * Scripts/webkitpy/common/config/committers.py:
1899
1900 2010-11-02  Anders Carlsson  <andersca@apple.com>
1901
1902         Fix build.
1903
1904         * TestWebKitAPI/PlatformUtilities.cpp:
1905         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
1906
1907 2010-11-02  Sam Weinig  <sam@webkit.org>
1908
1909         Reviewed by Anders Carlsson.
1910
1911         Add a way to send startup messages on the context which can be posted when a process launches
1912         <rdar://problem/8617928>
1913         https://bugs.webkit.org/show_bug.cgi?id=48838
1914
1915         * MiniBrowser/mac/WebBundle/WebBundleMain.m:
1916         (WKBundleInitialize):
1917         * TestWebKitAPI/InjectedBundleController.cpp:
1918         (TestWebKitAPI::InjectedBundleController::initialize):
1919         (TestWebKitAPI::InjectedBundleController::didReceiveMessage):
1920         * TestWebKitAPI/InjectedBundleController.h:
1921         * TestWebKitAPI/InjectedBundleMain.cpp:
1922         (WKBundleInitialize):
1923         * TestWebKitAPI/PlatformUtilities.cpp:
1924         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
1925         * WebKitTestRunner/InjectedBundle/InjectedBundleMain.cpp:
1926         (WKBundleInitialize):
1927         Update implementations of WKBundleInitialize to take an initial userData
1928         argument.  Change TestWebKitAPI to use the new initial userData to initialize
1929         each test's bundle.
1930
1931 2010-11-02  Benjamin Kalman  <kalman@google.com>
1932
1933         Reviewed by Ojan Vafai.
1934
1935         new-run-webkit-tests doesn't strip "LayoutTests/" from prefix, unlike old-run-webkit-tests
1936         https://bugs.webkit.org/show_bug.cgi?id=48794
1937
1938         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: Strip the "LayoutTests/" prefix from test argument paths.
1939
1940 2010-11-02  Adam Roben  <aroben@apple.com>
1941
1942         Skip webkitpy.layout_tests.run_webkit_tests_unittest.MainTest on Cygwin
1943         Python 2.5.x
1944
1945         It is known to hang on that version of Python. See
1946         <http://webkit.org/b/48614>.
1947
1948         Reviewed by Adam Barth.
1949
1950         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py: Use
1951         skip_if to skip MainTest on Cygwin Python 2.5.x.
1952
1953         * Scripts/webkitpy/test/skip.py: Added.
1954         (skip_if): If the passed-in condition is false, find all the test_*
1955         methods of the given class and replace them with a function that just
1956         logs that we're skipping these tests. This is loosely based on the
1957         unittest.skip_if decorator added in Python 3.1.
1958         (_skipped_method): Returns a function that just logs that the tests are
1959         being skipped. This is used in place of the actual test_* functions.
1960
1961         * Scripts/webkitpy/test/skip_unittest.py: Added.
1962         (SkipTest.setUp): Create a logger for ourselves and give it a handler
1963         that logs to a private stream.
1964         (SkipTest.tearDown): Reset the logger.
1965         (SkipTest.create_fixture_class): Returns a class that we use to test
1966         skip_if. It has a single test_* method, test_foo, that just calls
1967         through to a callback.
1968         (SkipTest.foo_callback): Record that test_foo was called.
1969         (SkipTest.test_skip_if_false): Pass skip_if a False condition and test
1970         that test_foo does get called.
1971         (SkipTest.test_skip_if_true): Pass skip_if a True condition and test
1972         that test_foo does not get called and the appropriate message gets
1973         logged.
1974
1975 2010-11-02  Adam Barth  <abarth@webkit.org>
1976
1977         Reviewed by Eric Seidel.
1978
1979         webkit-patch should tell check-webkit-style which files were changed so
1980         check-webkit-style doesn't have to stat the whole working copy again
1981         https://bugs.webkit.org/show_bug.cgi?id=48792
1982
1983         * Scripts/webkitpy/tool/mocktool.py:
1984         * Scripts/webkitpy/tool/steps/checkstyle.py:
1985
1986 2010-11-02  Robert Kroeger  <rjkroege@chromium.org>
1987
1988         Reviewed by James Robinson.
1989
1990         [Chromium/DRT] Make EventSendingController honour leapForward for touch events.
1991         https://bugs.webkit.org/show_bug.cgi?id=48777
1992
1993         * DumpRenderTree/chromium/EventSender.cpp:
1994         (EventSender::sendCurrentTouchEvent):
1995
1996 2010-11-02  Adam Roben  <aroben@apple.com>
1997
1998         Only track resource identifiers in DRT when dumpResourceLoadCallbacks
1999         is on
2000
2001         This reverts Windows to our pre-r71097 behavior. That patch made us
2002         track all resource identifiers, including the main resource, so the
2003         main resource's URL started appearing in test output instead of
2004         "<unknown>". Arguably having the main resource's URL is better, but all
2005         other platforms print "<unknown>" and we want to match.
2006
2007         Fixes <http://webkit.org/b/48837> <rdar://problem/8620351> REGRESSION
2008         (r71097): Many http tests failing on Windows
2009
2010         Reviewed by Anders Carlsson.
2011
2012         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
2013         (ResourceLoadDelegate::identifierForInitialRequest): Don't add the
2014         identifier to the URL map if we're not supposed to dump resource load
2015         callbacks.
2016         (ResourceLoadDelegate::removeIdentifierForRequest): Always remove the
2017         identifier from the URL map even if we're already "done". There's no
2018         point in keeping out-of-date identifiers around.
2019
2020 2010-11-01  Jenn Braithwaite  <jennb@chromium.org>
2021
2022         Reviewed by Adam Roben.
2023
2024         Windows: Update resource tracking when moving a frame between documents
2025         https://bugs.webkit.org/show_bug.cgi?id=48364
2026
2027         * DumpRenderTree/win/DumpRenderTree.cpp:
2028         (createWebViewAndOffscreenWindow):
2029         (main):
2030         Give each WebView its own ResourceLoadDelegate instance in order to
2031         make assertions about resource ids on a particular WebView.
2032         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
2033         (ResourceLoadDelegate::identifierForInitialRequest):
2034         Always add id to the map.
2035         (ResourceLoadDelegate::removeIdentifierForRequest):
2036         Added.
2037         (ResourceLoadDelegate::willSendRequest):
2038         (ResourceLoadDelegate::didReceiveAuthenticationChallenge):
2039         (ResourceLoadDelegate::didReceiveResponse):
2040         (ResourceLoadDelegate::didFinishLoadingFromDataSource):
2041         (ResourceLoadDelegate::didFailLoadingWithError):
2042         (ResourceLoadDelegate::descriptionSuitableForTestResult):
2043         Replace static descriptionSuitableForTestResult with static member function to access identifier map.
2044         * DumpRenderTree/win/ResourceLoadDelegate.h:
2045         (ResourceLoadDelegate::urlMap):
2046         Moved within class so that each WebView has its own id map.
2047
2048 2010-11-01  Ojan Vafai  <ojan@chromium.org>
2049
2050         Reviewed by Dimitri Glazkov.
2051
2052         remove debug code from run_webkit_tests.py
2053         https://bugs.webkit.org/show_bug.cgi?id=48800
2054
2055         Remove temporary debug code and make --master-name required
2056         if --test-results-server is set now that all clients set
2057         --master-name.
2058
2059         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2060
2061 2010-11-01  Tony Chang  <tony@chromium.org>
2062
2063         Reviewed by Kent Tamura.
2064
2065         add plugin event logging for linux
2066         https://bugs.webkit.org/show_bug.cgi?id=48779
2067
2068         This is taken from Chromium's fork of the layout test plugin:
2069         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
2070
2071         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
2072         (handleEventX11):
2073         (NPP_HandleEvent):
2074
2075 2010-11-01  Dirk Pranke  <dpranke@chromium.org>
2076
2077         Reviewed by James Robinson.
2078
2079         new-run-webkit-tests: use DRT, child-processes=1 for GPU tests by default
2080         https://bugs.webkit.org/show_bug.cgi?id=48790
2081
2082         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2083         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2084
2085 2010-11-01  Adam Barth  <abarth@webkit.org>
2086
2087         Reviewed by Eric Seidel.
2088
2089         Teach check-webkit-style how to accept a list of files to diff on the
2090         command line
2091         https://bugs.webkit.org/show_bug.cgi?id=48784
2092
2093         In a future patch, webkit-patch will use this option to improve
2094         performance.  I'm landing this in two pieces to avoid causing a version
2095         skew problem for the style-bot.
2096
2097         * Scripts/check-webkit-style:
2098         * Scripts/webkitpy/style/optparser.py:
2099         * Scripts/webkitpy/style/optparser_unittest.py:
2100         * Scripts/webkitpy/style_references.py:
2101
2102 2010-11-01  Anders Carlsson  <andersca@apple.com>
2103
2104         Reviewed by John Sullivan.
2105
2106         Tear down the related WebProcessProxy when a WebContext is deallocated
2107         https://bugs.webkit.org/show_bug.cgi?id=48769
2108
2109         * TestWebKitAPI/Tests/WebKit2/FailedLoad.cpp:
2110         (TestWebKitAPI::didFailProvisionalLoadWithErrorForFrame):
2111         We don't support empty URLs anymore, update test to expect a null URL instead.
2112
2113 2010-11-01  Søren Gjesse  <sgjesse@chromium.org>
2114
2115         Reviewed by Andreas Kling.
2116
2117         Fix warning when compiling the chromium port of DumpRenderShell
2118         with clang.
2119         https://bugs.webkit.org/show_bug.cgi?id=48414
2120
2121         * DumpRenderTree/chromium/TestShell.h:
2122         (TestShell::javaScriptFlagsForLoad):
2123
2124 2010-11-01  Adam Roben  <aroben@apple.com>
2125
2126         Fix typo from r71022
2127
2128         * BuildSlaveSupport/build.webkit.org-config/config.json:
2129
2130 2010-11-01  Adam Roben  <aroben@apple.com>
2131
2132         Trigger the Windows Release WebKit2 tests when a Release build
2133         finishes, not when a Debug build finishes
2134
2135         Fixes <http://webkit.org/b/48754> Windows Release WebKit2 tests are
2136         triggered at the wrong time
2137
2138         Reviewed by Sam Weinig.
2139
2140         * BuildSlaveSupport/build.webkit.org-config/config.json: Fixed
2141         triggerable name and trigger.
2142
2143 2010-11-01  Mario Sanchez Prada  <msanchez@igalia.com>
2144
2145         Unreviewed. Adding my IRC nickname to the list of committers.
2146
2147         * Scripts/webkitpy/common/config/committers.py:
2148
2149 2010-10-31  Robert Hogan  <robert@webkit.org>
2150
2151         Reviewed by Antonio Gomes.
2152
2153         [Qt] [Gtk] Plug-ins having upper case in mime type are failing to load
2154
2155         Qt and Gtk are case-sensitive when storing the declared mime-type
2156         of plugins. Since plugin mime-types are lowercased prior to searching
2157         for them in the plugin database, ensure they are loaded with the
2158         mime-type in lower case too.
2159
2160         Change the test netscape plugin to declare its mimetype in sentence
2161         case so that the correct behaviour is enforced.
2162
2163         https://bugs.webkit.org/show_bug.cgi?id=36815
2164
2165         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
2166         (NP_GetMIMEDescription):
2167
2168 2010-10-31  Robert Hogan  <robert@webkit.org>
2169
2170         Reviewed by Antonio Gomes.
2171
2172         [Qt] Support nodesFromRect in DRT
2173
2174         https://bugs.webkit.org/show_bug.cgi?id=48716
2175
2176         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2177         (LayoutTestController::nodesFromRect):
2178         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2179
2180 2010-10-30  Andreas Kling  <kling@webkit.org>
2181
2182         Reviewed by Antonio Gomes.
2183
2184         [Qt] QtTestBrowser: Switching view type moves the embedded inspector
2185         https://bugs.webkit.org/show_bug.cgi?id=48705
2186
2187         Reinsert the embedded inspector into the splitter after changing
2188         between QWebView/QGraphicsWebView.
2189
2190         * QtTestBrowser/launcherwindow.cpp:
2191         (LauncherWindow::init):
2192         (LauncherWindow::initializeView):
2193         * QtTestBrowser/webinspector.h:
2194         (WebInspector::WebInspector):
2195
2196 2010-10-28  Antonio Gomes  <agomes@rim.com>
2197
2198         Reviewed by Ojan Vafai.
2199
2200         Needs a "LinuxEditingBehavior", perhaps with a better name
2201         https://bugs.webkit.org/show_bug.cgi?id=36627
2202
2203         Adding support to Mac's, GTK+'s, Windows' and Chromium's LayoutTestController class to test the newly introduced Unix editing behavior.
2204
2205         * DumpRenderTree/chromium/LayoutTestController.cpp:
2206         (LayoutTestController::setEditingBehavior):
2207         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2208         (LayoutTestController::setEditingBehavior):
2209         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2210         (LayoutTestController::setEditingBehavior):
2211         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2212         (LayoutTestController::setEditingBehavior):
2213
2214 2010-10-29  Andreas Kling  <kling@webkit.org>
2215
2216         Reviewed by Kenneth Rohde Christiansen.
2217
2218         [Qt] QtTestBrowser: Add keyboard shortcut to toggle full screen (F11)
2219         https://bugs.webkit.org/show_bug.cgi?id=48695
2220
2221         * QtTestBrowser/launcherwindow.cpp:
2222         (LauncherWindow::createChrome):
2223
2224 2010-10-29  Andreas Kling  <kling@webkit.org>
2225
2226         Reviewed by Kenneth Rohde Christiansen.
2227
2228         [Qt] QtTestBrowser: Fix uninitialized read in FpsTimer
2229         https://bugs.webkit.org/show_bug.cgi?id=48675
2230
2231         FpsTimer::m_timer was never initialized and passed to QObject::killTimer()
2232         on startup with in -graphicsbased mode.
2233
2234         * QtTestBrowser/fpstimer.cpp:
2235         (FpsTimer::FpsTimer):
2236         (FpsTimer::numFrames):
2237         (FpsTimer::stop):
2238         (FpsTimer::timerEvent):
2239         * QtTestBrowser/fpstimer.h:
2240
2241 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
2242
2243         Reviewed by Tony Chang.
2244
2245         new-run-webkit-tests: change TestResults to be serializable
2246
2247         In preparation for changing new-run-webkit-tests from
2248         multithreaded to multiprocess, we need to make sure the data
2249         going between the threads is easily serialized over a socket.
2250
2251         This change adds serialization/pickling for the TestResults and
2252         TestFailure objects (using cPickle).
2253
2254         The TestFailure objects included a "has_wdiff" flag for Text
2255         results, but the flag wasn't being used, so I've removed it,
2256         simplifying the state to basically a set of enum objects with
2257         associated methods.
2258
2259         https://bugs.webkit.org/show_bug.cgi?id=48616
2260
2261         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2262         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
2263         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
2264         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
2265         * Scripts/webkitpy/layout_tests/layout_package/test_results.py: Added.
2266         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:Added. 
2267         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2268         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2269
2270 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
2271
2272         Reviewed by Ojan Vafai.
2273
2274         new-run-webkit-tests: need to provide separate dashboard results for GPU tests
2275         https://bugs.webkit.org/show_bug.cgi?id=48687
2276
2277         Modify the steps to upload the JSON files to the dashboards so
2278         that the GPU tests don't conflict with the regular tests on a
2279         port. We do this by modifying the --builder-name parameter to
2280         append " - GPU", which should cause the app to treat the results
2281         as a completely new builder. This is a little non-obvious, but
2282         keeps us from having to restructure the app.
2283
2284         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2285         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2286         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2287
2288 2010-10-29  Ojan Vafai  <ojan@chromium.org>
2289
2290         Reviewed by Tony Chang.
2291
2292         [chromium] add debug logging to help diagnose flakiness dashboard issues
2293         https://bugs.webkit.org/show_bug.cgi?id=48657
2294
2295         The appengine app thinks it's getting empty files uploaded. Add some logging
2296         to see if new-run-webkit-tests agrees.
2297         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2298
2299 2010-10-29  Dimitri Glazkov  <dglazkov@chromium.org>
2300
2301         Reviewed by Ojan Vafai.
2302
2303         [Chromium/DRT] Add master-name flag to new-run-webkit-tests.
2304         https://bugs.webkit.org/show_bug.cgi?id=48649
2305
2306         * BuildSlaveSupport/build.webkit.org-config/master.cfg: Added flag.
2307
2308 2010-10-29  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
2309
2310         Reviewed by Martin Robinson.
2311
2312         Enable popup window in GtkLauncher
2313         https://bugs.webkit.org/show_bug.cgi?id=48335
2314
2315         The GtkLauncher application does not support opening new window when
2316         a link with "target=_blank" is clicked or similar call to
2317         window.open(). Instead, GtkLauncher does nothing which breaks
2318         navigation of some websites.
2319
2320         * GtkLauncher/main.c:
2321         (activate_uri_entry_cb):
2322         (update_title):
2323         (link_hover_cb):
2324         (notify_title_cb):
2325         (notify_load_status_cb):
2326         (notify_progress_cb):
2327         (destroy_cb):
2328         (go_back_cb):
2329         (go_forward_cb):
2330         (create_web_view_cb):
2331         (web_view_ready_cb):
2332         (close_web_view_cb):
2333         (create_browser):
2334         (create_statusbar):
2335         (create_toolbar):
2336         (create_window):
2337         (main):
2338
2339 2010-10-29  Adam Roben  <aroben@apple.com>
2340
2341         Teach check-webkit-style about WebKit2's idiosyncracies
2342
2343         Fixes <http://webkit.org/b/48638> Style bot complains about a number
2344         of WebKit2 conventions
2345
2346         Reviewed by Anders Carlsson.
2347
2348         * Scripts/webkitpy/style/checker.py: Excluded some rules for various
2349         WebKit2-related files. Also updated the excluded rules for
2350         WebKitAPITest to match the current code.
2351
2352 2010-10-29  Csaba Osztrogonác  <ossy@webkit.org>
2353
2354         Reviewed by Adam Roben and David Kilzer.
2355
2356         Fix and cleanup of build systems
2357         https://bugs.webkit.org/show_bug.cgi?id=48342
2358
2359         * Scripts/build-webkit:
2360          - Remove unnecessary ENABLE_SANDBOX option.
2361          - Add ENABLE_FULLSCREEN_API option.
2362
2363 2010-10-28  Adam Roben  <aroben@apple.com>
2364
2365         Switch the Windows WebKit2 bot to the Release configuration
2366
2367         We only have one machine testing WebKit2 on Windows right now, and
2368         Debug is just too slow for it to keep up.
2369
2370         Fixes (hopefully!) <http://webkit.org/b/48615> Windows WebKit2 bot is
2371         always way behind
2372
2373         Reviewed by Jon Honeycutt.
2374
2375         * BuildSlaveSupport/build.webkit.org-config/config.json:
2376
2377 2010-10-29  Leandro Gracia Gil  <leandrogracia@google.com>
2378
2379         Reviewed by Jeremy Orlow.
2380
2381         Added a second parameter to setMockSpeechInputResult for
2382         the language used in speech input.
2383         https://bugs.webkit.org/show_bug.cgi?id=47089
2384
2385         * DumpRenderTree/LayoutTestController.cpp:
2386         (setMockSpeechInputResultCallback):
2387         * DumpRenderTree/LayoutTestController.h:
2388         * DumpRenderTree/chromium/LayoutTestController.cpp:
2389         (LayoutTestController::setMockSpeechInputResult):
2390         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2391         (LayoutTestController::setMockSpeechInputResult):
2392         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2393         (LayoutTestController::setMockSpeechInputResult):
2394         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2395         (LayoutTestController::setMockSpeechInputResult):
2396         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2397         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2398         (LayoutTestController::setMockSpeechInputResult):
2399         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2400         (LayoutTestController::setMockSpeechInputResult):
2401         (LayoutTestController::pageSizeAndMarginsInPixels):
2402
2403 2010-10-28  Ojan Vafai  <ojan@chromium.org>
2404
2405         Reviewed by Tony Chang.
2406
2407         [chromium] add a result-small.json file for the test dashboard
2408         https://bugs.webkit.org/show_bug.cgi?id=48547
2409
2410         Output both a results.json file and a results-small.json file.
2411         The dashboard will load results-small.json by default so it loads faster.
2412
2413         * TestResultServer/model/jsonresults.py:
2414         * TestResultServer/model/jsonresults_unittest.py:
2415         Added a bunch of sys.path hackery. Unforunately, this uses hardcoded
2416         paths. That obviously needs to be fixed, but at least this way it
2417         clearly documents what paths are necessary.
2418
2419 2010-10-28  Eric Seidel  <eric@webkit.org>
2420
2421         Reviewed by Adam Barth.
2422
2423         webkit-patch upload calls changed_files more often than it should
2424         https://bugs.webkit.org/show_bug.cgi?id=48567
2425
2426         Passing changed_files around everywhere isn't a very elegant solution
2427         but it's the one we have for the moment.  I think keeping an explicit
2428         cache on Checkout (or making StepState() a real class) is a better
2429         long-term option.
2430
2431         Previously bug_id_for_this_commit was calling changed_files and the
2432         result was never getting cached on the state.  Now we're explicitly
2433         caching the result on the state and passing that to the bug_id_for_this_commit call.
2434
2435         I looked into building unit tests for this.  Doing so would require
2436         using a real Checkout object with a MockSCM and overriding the appropriate
2437         calls on SCM to count how often we're stating the file system.
2438         That's a useful set of tests to build for a separate change.
2439
2440         * Scripts/webkitpy/common/checkout/api.py:
2441         * Scripts/webkitpy/tool/commands/download.py:
2442         * Scripts/webkitpy/tool/commands/upload.py:
2443         * Scripts/webkitpy/tool/mocktool.py:
2444
2445 2010-10-28  Eric Seidel  <eric@webkit.org>
2446
2447         Reviewed by Adam Barth.
2448
2449         Make suggest-reviewers slightly faster
2450         https://bugs.webkit.org/show_bug.cgi?id=48562
2451
2452         Add @memoized to one more common call.
2453
2454         * Scripts/webkitpy/common/checkout/api.py:
2455
2456 2010-10-28  Eric Seidel  <eric@webkit.org>
2457
2458         Reviewed by Adam Barth.
2459
2460         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is still very slow
2461         https://bugs.webkit.org/show_bug.cgi?id=48536
2462
2463         This patch makes it about 40 seconds faster, but it still
2464         takes 1:40.  This will require more refinement.
2465
2466         The suggested reviewers list appears to be the same.
2467
2468         I think the next step may be to have it stop the search after
2469         5 reviewers are found.  We never want to suggest 30 people.
2470
2471         * Scripts/webkitpy/common/checkout/scm.py:
2472          - Using --remove-empty to theoretically stop lookups past
2473            when a file is removed.  I'm not entirely clear that the option
2474            does what it says it does.  Example:
2475            git log --pretty=format:%H -5 --remove-empty -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
2476            returns only one commit
2477            vs.
2478            git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
2479            which returns 5.  I was not aware that wx files were ever removed from the repository?
2480
2481 2010-10-28  Kinuko Yasuda  <kinuko@chromium.org>
2482
2483         Reviewed by David Levin.
2484
2485         [Chromium] Support FileSystem in chromium DRT
2486         https://bugs.webkit.org/show_bug.cgi?id=47643
2487
2488         * DumpRenderTree/chromium/WebViewHost.cpp:
2489         (WebViewHost::openFileSystem): Added.
2490         * DumpRenderTree/chromium/WebViewHost.h:
2491         (WebViewHost::openFileSystem): Added.
2492
2493 2010-10-28  Sam Weinig  <sam@webkit.org>
2494
2495         Reviewed by Anders Carlsson.
2496
2497         WKURLRefs should be allowed to be null
2498         <rdar://problem/8575621>
2499         https://bugs.webkit.org/show_bug.cgi?id=48535
2500
2501         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
2502         * TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp:
2503         (TestWebKitAPI::didStartProvisionalLoadForFrame):
2504         (TestWebKitAPI::didCommitLoadForFrame):
2505         (TestWebKitAPI::didFinishLoadForFrame):
2506         Test that URLs are null pointers when unset.
2507
2508 2010-10-28  Chris Fleizach  <cfleizach@apple.com>
2509
2510         Reviewed by Adele Peterson.
2511
2512         AX: multi select group option does not handle setting of AXSelectedChildren correctly
2513         https://bugs.webkit.org/show_bug.cgi?id=48464
2514
2515         Add support for querying information about selected children to DRT, including:
2516            selectedChildAtIndex
2517            selectedChildrenCount
2518            setSelectedChild
2519
2520         * DumpRenderTree/AccessibilityUIElement.cpp:
2521         (selectedChildAtIndexCallback):
2522         (setSelectedChildCallback):
2523         (selectedChildrenCountCallback):
2524         (AccessibilityUIElement::setSelectedChild):
2525         (AccessibilityUIElement::selectedChildrenCount):
2526         (AccessibilityUIElement::selectedChildAtIndex):
2527         (AccessibilityUIElement::getJSClass):
2528         * DumpRenderTree/AccessibilityUIElement.h:
2529         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
2530         (AccessibilityUIElement::selectedChildAtIndex):
2531         (AccessibilityUIElement::selectedChildrenCount):
2532         (AccessibilityUIElement::setSelectedChild):
2533
2534 2010-10-27  Sam Weinig  <sam@webkit.org>
2535
2536         Reviewed by Anders Carlsson.
2537
2538         Add WebKit2 API for window feature getter/setters
2539         <rdar://problem/8590373>
2540         https://bugs.webkit.org/show_bug.cgi?id=48496
2541
2542         * MiniBrowser/mac/BrowserWindowController.m:
2543         (-[BrowserWindowController awakeFromNib]):
2544         * MiniBrowser/win/BrowserView.cpp:
2545         (BrowserView::create):
2546         * WebKitTestRunner/TestController.cpp:
2547         (WTR::createOtherPage):
2548         (WTR::TestController::initialize):
2549
2550 2010-10-28  Søren Gjesse  <sgjesse@chromium.org>
2551
2552         Reviewed by Tony Chang.
2553
2554         Added support for the DumpRenderTree flags --multiple-loads and --js-flags to the Python test runner.
2555         https://bugs.webkit.org/show_bug.cgi?id=48236
2556
2557         * Scripts/webkitpy/layout_tests/port/chromium.py:
2558         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2559
2560 2010-10-28  Sergio Villar Senin  <svillar@igalia.com>
2561
2562         Reviewed by Martin Robinson.
2563
2564         [GTK] WebKitWebFrame's load-status is not properly notified to the tests
2565         https://bugs.webkit.org/show_bug.cgi?id=48048
2566
2567         DRT now listens to WebKitWebFrame load-status signals for each
2568         created frame instead of the load-status signal of the
2569         WebKitWebView that only notifies about changes in the main frame.
2570
2571         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2572         (webViewLoadFinished):
2573         (webFrameLoadStatusNotified):
2574         (frameCreatedCallback):
2575         (createWebView):
2576         (main):
2577
2578 2010-10-28  Tony Chang  <tony@chromium.org>
2579
2580         Reviewed by Kent Tamura.
2581
2582         [chromium] fix textInputController.{selectedRange,markedRange}
2583         https://bugs.webkit.org/show_bug.cgi?id=48487
2584
2585         * DumpRenderTree/chromium/TextInputController.cpp:
2586         (TextInputController::markedRange): Return arrays of ints, rather than a string
2587         (TextInputController::selectedRange): Ditto.
2588
2589 2010-10-27  Ojan Vafai  <ojan@chromium.org>
2590
2591         Reviewed by Tony Chang.
2592
2593         [chromium] add a master-name flag to new-run-webkit-tests
2594         https://bugs.webkit.org/show_bug.cgi?id=48488
2595
2596         The test results server now allows adding a master name to
2597         the uploaded files. This lets us distinguish bots that have
2598         the same name, but are on different masters.
2599
2600         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2601
2602 2010-10-28  Adam Roben  <aroben@apple.com>
2603
2604         Don't append a newline to the test output if the frame has no document
2605         element in WebKitTestRunner
2606
2607         Fixes <http://webkit.org/b/48526> Extra trailing newline when running
2608         plugins/document-open.html in WebKitTestRunner
2609
2610         Reviewed by Anders Carlsson.
2611
2612         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2613         (WTR::toJS): Added. Turns a UTF-8 C string into a JSStringRef.
2614         (WTR::hasDocumentElement): Added. Uses the JSC API to figure out
2615         whether the frame has a document element.
2616         (WTR::dumpFrameText): Match DRT by bailing (rather than appending an
2617         empty string and a newline) if the frame has no document element.
2618
2619 2010-10-28  Adam Roben  <aroben@apple.com>
2620
2621         Skip npn-invalidate-rect-invalidates-window.html on headless XP
2622         machines
2623
2624         TestNetscapePlugIn never receives a WM_PAINT message on headless XP
2625         machines, so this test times out. Fixing the test is covered by
2626         <http://webkit.org/b/48333>.
2627
2628         * Scripts/old-run-webkit-tests: Skip
2629         npn-invalidate-rect-invalidates-window.html on Windows if accelerated
2630         compositing support is disabled, which likely means we're on a headless
2631         XP machine.
2632
2633 2010-10-28  Kimmo Kinnunen  <kimmok@iki.fi>
2634
2635         Adding myself as a committer.
2636
2637         * Scripts/webkitpy/common/config/committers.py:
2638
2639 2010-10-27  Eric Seidel  <eric@webkit.org>
2640
2641         Reviewed by Adam Barth.
2642
2643         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is very slow
2644         https://bugs.webkit.org/show_bug.cgi?id=48500
2645
2646         This doesn't fix the problem, but it makes things slightly better.
2647         Each git svn find-rev call takes about .25 seconds on my desktop.
2648         This patch uses a new memoized class to avoid those calls when possible.
2649
2650         The real slowness is still git log on some files, like:
2651         git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
2652         I'm not yet sure how to make the pathological git logs better.
2653
2654         * Scripts/webkitpy/common/checkout/scm.py:
2655         * Scripts/webkitpy/common/memoized.py: Added.
2656         * Scripts/webkitpy/common/memoized_unittest.py: Added.
2657
2658 2010-10-27  Eric Seidel  <eric@webkit.org>
2659
2660         Reviewed by Ojan Vafai.
2661
2662         EWS bots should not use --quiet when running build-webkit
2663         https://bugs.webkit.org/show_bug.cgi?id=48482
2664
2665         --quiet is only correct when the sub-process does the error reporting.
2666         In the case of _can_build() the parent process is reporting the error.
2667         We'd like the full build log at queues.webkit.org so someone can look
2668         at the log and understand why the EWS is failing to build trunk.
2669
2670         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2671         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
2672         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
2673         * Scripts/webkitpy/tool/mocktool.py:
2674
2675 2010-10-27  Brian Weinstein  <bweinstein@apple.com>
2676
2677         More Windows build fixage. Rename a variable that was named string.
2678
2679         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2680         (WTR::hasPrefix):
2681
2682 2010-10-27  Ojan Vafai  <ojan@chromium.org>
2683
2684         Reviewed by Dimitri Glazkov.
2685
2686         [chromium] Make the test results server store which master the bot is on
2687         https://bugs.webkit.org/show_bug.cgi?id=48478
2688
2689         The chromium bots recently changed so that there are multiple slaves with
2690         the same name on different masters. Up till now, the test results server
2691         assumed slave names were unique. Adds a master field to the file in order
2692         to distinguish.
2693
2694         Also, for files that currently lack a master or testtype, set them appropriately.
2695
2696         * TestResultServer/handlers/testfilehandler.py:
2697         * TestResultServer/index.yaml:
2698         * TestResultServer/model/jsonresults.py:
2699         * TestResultServer/model/testfile.py:
2700         * TestResultServer/templates/showfilelist.html:
2701         * TestResultServer/templates/uploadform.html:
2702
2703 2010-10-26  Darin Adler  <darin@apple.com>
2704
2705         Reviewed by Sam Weinig.
2706
2707         WebKitTestRunner needs to support layoutTestController.dumpBackForwardList
2708         https://bugs.webkit.org/show_bug.cgi?id=42322
2709         rdar://problem/8193631
2710
2711         WebKitTestRunner needs to support layoutTestController.clearBackForwardList
2712         https://bugs.webkit.org/show_bug.cgi?id=42333
2713         rdar://problem/8193643
2714
2715         * WebKitTestRunner/Configurations/InjectedBundle.xcconfig:
2716         Renamed the product to WebKitTestRunnerInjectedBundle to avoid
2717         name conflicts in the build directory.
2718
2719         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
2720         Added dumpBackForwardList and clearBackForwardList.
2721
2722         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
2723         (WTR::InjectedBundle::InjectedBundle): Removed initialization of m_mainPage.
2724         (WTR::InjectedBundle::didCreatePage): Removed unneeded underscore.
2725         (WTR::InjectedBundle::willDestroyPage): Ditto.
2726         (WTR::InjectedBundle::didReceiveMessage): Ditto.
2727         (WTR::InjectedBundle::initialize): Ditto.
2728         (WTR::InjectedBundle::didCreatePage): Changed code to use m_pages
2729         instead of m_mainPage and m_otherPages.
2730         (WTR::InjectedBundle::willDestroyPage): Ditto.
2731         (WTR::InjectedBundle::page): Ditto.
2732         (WTR::InjectedBundle::beginTesting): Ditto.
2733         (WTR::InjectedBundle::done): Ditto.
2734         (WTR::InjectedBundle::closeOtherPages): Ditto.
2735         (WTR::InjectedBundle::dumpBackForwardListsForAllPages): Added.
2736
2737         * WebKitTestRunner/InjectedBundle/InjectedBundle.h: Changed
2738         page function to longer be inline, and pageCount function
2739         to use m_pages. Added dumpBackForwardListsForAllPages, and
2740         removed some underscores. Replaced m_mainPage and m_otherPageas
2741         with m_pages.
2742
2743         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2744         (WTR::adoptWK): Added.
2745         (WTR::hasPrefix): Added.
2746         (WTR::InjectedBundlePage::reset): Added code to set up
2747         m_previousTestBackForwardListItem.
2748         (WTR::InjectedBundlePage::dump): Added code to call
2749         dumpBackForwardListsForAllPages.
2750         (WTR::compareByTargetName): Added.
2751         (WTR::dumpBackForwardListItem): Added.
2752         (WTR::InjectedBundlePage::dumpBackForwardList): Added.
2753
2754         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h: Added
2755         dumpBackForwardList and m_previousTestBackForwardListItem.
2756
2757         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
2758         (WTR::LayoutTestController::LayoutTestController): Initialize
2759         m_shouldDumpBackForwardListsForAllWindows to false.
2760         (WTR::LayoutTestController::clearBackForwardList): Added.
2761
2762         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
2763         Added dumpBackForwardList, clearBackForwardList,
2764         shouldDumpBackForwardListsForAllWindows, and
2765         m_shouldDumpBackForwardListsForAllWindows.
2766
2767         * WebKitTestRunner/TestController.cpp:
2768         (WTR::TestController::initialize): Set up
2769         didReceiveSynchronousMessageFromInjectedBundle.
2770         (WTR::TestController::didReceiveSynchronousMessageFromInjectedBundle):
2771         Added.
2772         * WebKitTestRunner/TestController.h: Ditto.
2773
2774         * WebKitTestRunner/TestInvocation.cpp:
2775         (WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):
2776         Added.
2777         * WebKitTestRunner/TestInvocation.h: Ditto.
2778
2779         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
2780         Renamed bundle to WebKitTestRunnerInjectedBundle (see above).
2781         * WebKitTestRunner/mac/TestControllerMac.mm:
2782         (WTR::TestController::initializeInjectedBundlePath): Ditto.
2783
2784 2010-10-27  Chris Rogers  <crogers@google.com>
2785
2786         Reviewed by Chris Marrin.
2787
2788         Add ENABLE_WEB_AUDIO feature enable flag (initially disabled) to build-webkit
2789         https://bugs.webkit.org/show_bug.cgi?id=48279
2790
2791         * Scripts/build-webkit:
2792
2793 2010-10-27  Eric Seidel  <eric@webkit.org>
2794
2795         Unreviewed.
2796
2797         build-webkit should collect Visual Studio Express logs and display them
2798         https://bugs.webkit.org/show_bug.cgi?id=39199
2799
2800         It turns out my previous patch wasn't actually working on the win-ews
2801         machine.  So I've fixed my mistakes from before.
2802
2803         * Scripts/build-webkit:
2804          - Windows VSE builds change the CWD while building.  Why?  Who knows.
2805         * Scripts/print-vse-failure-logs:
2806          - windows VSE builds don't use Debug/Release as I expected, so I've
2807            moved off of --configuration to --top-level and added the /obj
2808            optimization while I was there.
2809
2810 2010-10-27  Dimitri Glazkov  <dglazkov@chromium.org>
2811
2812         Unreviewed, rolling out r70674.
2813         http://trac.webkit.org/changeset/70674
2814         https://bugs.webkit.org/show_bug.cgi?id=48053
2815
2816         Broke Chromium Windows build.
2817
2818         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2819         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2820
2821 2010-10-25  Tony Chang  <tony@chromium.org>
2822
2823         Reviewed by Anders Carlsson.
2824
2825         compile TestNetscapePlugIn on chromium linux
2826         https://bugs.webkit.org/show_bug.cgi?id=48274
2827
2828         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
2829         (NP_Initialize): On Linux, plugin funcs are set in initialize.
2830         (NPP_New): Mark the plugin as windowless.
2831         (NPP_GetValue): Handle mime type values.
2832         (NP_GetMIMEDescription):
2833         (NP_GetValue):
2834         * DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npapi.h: define TRUE and FALSE, which are in webkit's npapi.h.
2835         * DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npfunctions.h: Pull in npapi.h to get TRUE/FALSE (matches webkit's npfunctions.h)
2836
2837 2010-10-27  Kenneth Rohde Christiansen  <kenneth@webkit.org>
2838
2839         Reviewed by Anders Carlsson.
2840
2841         Remove contentSizeChanged callbacks as it is no longer
2842         part of the public UIClient.
2843
2844         Make WKPageContentsSizeChangedCallback be a private API
2845         https://bugs.webkit.org/show_bug.cgi?id=48409
2846
2847         * MiniBrowser/mac/BrowserWindowController.m:
2848         (-[BrowserWindowController awakeFromNib]):
2849         * MiniBrowser/win/BrowserView.cpp:
2850         (BrowserView::create):
2851         * WebKitTestRunner/TestController.cpp:
2852         (WTR::createOtherPage):
2853         (WTR::TestController::initialize):
2854
2855 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2856
2857         Reviewed by Ojan Vafai.
2858
2859         Make http locking default in NRWT.
2860         https://bugs.webkit.org/show_bug.cgi?id=48053
2861
2862         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2863         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2864
2865 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2866
2867         Reviewed by Tony Chang.
2868
2869         [NRWT] Fix http lock on Windows platform
2870         https://bugs.webkit.org/show_bug.cgi?id=48321
2871
2872         * Scripts/webkitpy/layout_tests/port/http_lock.py:
2873
2874 2010-10-27  Satish Sampath  <satish@chromium.org>
2875
2876         Unreviewed, rolling out r70665.
2877         http://trac.webkit.org/changeset/70665
2878         https://bugs.webkit.org/show_bug.cgi?id=47089
2879
2880         Need to address Alexey's review comments.
2881
2882         * DumpRenderTree/LayoutTestController.cpp:
2883         (setMockSpeechInputResultCallback):
2884         * DumpRenderTree/LayoutTestController.h:
2885         * DumpRenderTree/chromium/LayoutTestController.cpp:
2886         (LayoutTestController::setMockSpeechInputResult):
2887         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2888         (LayoutTestController::setMockSpeechInputResult):
2889         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2890         (LayoutTestController::setMockSpeechInputResult):
2891         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2892         (LayoutTestController::setMockSpeechInputResult):
2893         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2894         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2895         (LayoutTestController::setMockSpeechInputResult):
2896         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2897         (LayoutTestController::setMockSpeechInputResult):
2898
2899 2010-10-27  Leandro Gracia Gil  <leandrogracia@google.com>
2900
2901         Reviewed by Jeremy Orlow.
2902
2903         Added a second parameter to setMockSpeechInputResult for
2904         the language used in speech input.
2905         https://bugs.webkit.org/show_bug.cgi?id=47089
2906
2907         * DumpRenderTree/LayoutTestController.cpp:
2908         (setMockSpeechInputResultCallback):
2909         * DumpRenderTree/LayoutTestController.h:
2910         * DumpRenderTree/chromium/LayoutTestController.cpp:
2911         (LayoutTestController::setMockSpeechInputResult):
2912         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2913         (LayoutTestController::setMockSpeechInputResult):
2914         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2915         (LayoutTestController::setMockSpeechInputResult):
2916         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2917         (LayoutTestController::setMockSpeechInputResult):
2918         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2919         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2920         (LayoutTestController::setMockSpeechInputResult):
2921         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2922         (LayoutTestController::setMockSpeechInputResult):
2923         (LayoutTestController::pageSizeAndMarginsInPixels):
2924
2925 2010-10-27  Adam Roben  <aroben@apple.com>
2926
2927         Reset TestNetscapePlugIn's NPP_GetValue pointer when
2928         NullNPPGetValuePointer finishes running
2929
2930         Fixes <http://webkit.org/b/48435> REGRESSION (r70655): Many plugins
2931         tests are failing on Qt
2932
2933         Reviewed by Anders Carlsson.
2934
2935         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
2936         (PluginTest::NPP_Destroy):
2937         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
2938         Added a do-nothing NPP_Destroy implementation.
2939
2940         * DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp:
2941         (NullNPPGetValuePointer::NullNPPGetValuePointer): Save the original
2942         NPP_GetValue pointer in m_originalNPPGetValuePointer so we can restore
2943         it later.
2944         (NullNPPGetValuePointer::NPP_Destroy): Added. Restores the original
2945         NPP_GetValue pointer so it can be used in other tests.
2946
2947         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
2948         (NPP_Destroy):
2949         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
2950         (webkit_test_plugin_destroy_instance):
2951         Call through to the PluginTest when NPP_Destroy is called.
2952
2953 2010-10-26  Brian Weinstein  <bweinstein@apple.com>
2954
2955         Reviewed by Adam Roben.
2956
2957         WebKit2 shouldn't try to send an empty user agent
2958         https://bugs.webkit.org/show_bug.cgi?id=48397
2959         
2960         Add a test that when we set our custom user agent to an empty string, we don't send
2961         an empty user agent.
2962
2963         * TestWebKitAPI/Tests/WebKit2/PreventEmptyUserAgent.cpp: Added.
2964         (TestWebKitAPI::didRunJavaScript): Make sure that the result of navigator.userAgent isn't empty.
2965         (TestWebKitAPI::TEST): Set our custom user agent to the empty string, and run navigator.userAgent.
2966         
2967         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Add the new file.
2968         * TestWebKitAPI/win/TestWebKitAPI.vcproj: Ditto.
2969
2970 2010-10-27  Adam Roben  <aroben@apple.com>
2971
2972         Check in file I forgot in r70653
2973
2974         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
2975         (NP_Initialize): Save the NPPluginFuncs struct the browser passed to us
2976         so we can be naughty and modify it later.
2977
2978 2010-10-27  Adam Roben  <aroben@apple.com>
2979
2980         Test that WebKit doesn't crash if the plugin passes 0 for its
2981         NPP_GetValue pointer
2982
2983         Test for <http://webkit.org/b/48433> Crash in
2984         NetscapePlugin::shouldLoadSrcURL when using Shockwave Director 10.3 in
2985         WebKit2 on Windows
2986
2987         Reviewed by Eric Carlson.
2988
2989         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
2990         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2991         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
2992         * GNUmakefile.am:
2993         Added NullNPPGetValuePointer.cpp.
2994
2995         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
2996         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
2997         Set up a place to store the NPPluginFuncs struct the browser passed to us.
2998
2999         * DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp: Added.
3000         (NullNPPGetValuePointer::NullNPPGetValuePointer): Null out the
3001         NPP_GetValue pointer we passed to the browser to simulate a plugin that
3002         doesn't implement NPP_GetValue.
3003         (NullNPPGetValuePointer::NPP_GetValue): Print an error message. If this
3004         function is called, it means that WebKit has changed in a way that
3005         makes this test invalid.
3006
3007         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
3008         (NP_GetEntryPoints): Save the NPPluginFuncs struct the browser passed
3009         to us so we can be naughty and modify it later.
3010
3011 2010-10-27  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
3012
3013         Reviewed by Andreas Kling.
3014
3015         Remove references to ancient QGVLauncher and QtLauncher
3016         https://bugs.webkit.org/show_bug.cgi?id=48430
3017
3018         QtTestBrowser substitutes both and has checks enabled.
3019
3020         * Scripts/webkitpy/style/checker.py: remove references
3021         * Scripts/webkitpy/style/checker_unittest.py: ditto
3022
3023 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3024
3025         Reviewed by Ojan Vafai.
3026
3027         [NRWT] Don't use image hash when it's no need in single test mode.
3028         https://bugs.webkit.org/show_bug.cgi?id=48326
3029
3030         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3031
3032 2010-10-27  David Kilzer  <ddkilzer@apple.com>
3033
3034         Fix leak of CFMutableDictionaryRef in createXMLStringFromWebArchiveData()
3035
3036         Reviewed by Adam Roben.
3037
3038         Follow-up fix for: <https://bugs.webkit.org/show_bug.cgi?id=48278>
3039
3040         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
3041         (createXMLStringFromWebArchiveData): Use RetainPtr<> to fix a
3042         leak introduced in r70613.
3043
3044 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3045
3046         Reviewed by Csaba Osztrogonác.
3047
3048         [NRWT] Add platform specific baseline search paths for Qt port
3049         https://bugs.webkit.org/show_bug.cgi?id=48428
3050
3051         * Scripts/webkitpy/layout_tests/port/qt.py:
3052
3053 2010-10-27  Adam Roben  <aroben@apple.com>
3054
3055         Catch exceptions when checking if we're inside a git working directory
3056
3057         Fixes <http://webkit.org/b/48420> REGRESSION (r70562): test-webkitpy
3058         fails on systems without git installed
3059
3060         Reviewed by Anders Carlsson.
3061
3062         * Scripts/webkitpy/common/net/credentials.py:
3063         (Credentials._credentials_from_git): Put the call to
3064         Git.in_working_directory inside the try/except since it, too, attempts
3065         to execute git and thus will throw on systems that don't have git
3066         installed.
3067
3068 2010-10-27  Nikolas Zimmermann  <nzimmermann@rim.com>
3069
3070         Reviewed by David Kilzer.
3071
3072         Convert DumpRenderTree webarchive code to CoreFoundation
3073         https://bugs.webkit.org/show_bug.cgi?id=48278
3074
3075         CFPropertyListCreateWithData and CFPropertyListCreateData are only available in 10.6+.
3076         Replace CFPropertyListCreateWithData by a combination of CFReadStreamCreateWithBytesNoCopy and CFPropertyListCreateFromStream.
3077         Replace CFPropertyListCreateData by CFPropertyListCreateXMLData.
3078
3079         These changes are wrapped in BUILDING_ON_LEOPARD, as the methods are deprecated on 10.6+.
3080
3081         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
3082         (createXMLStringFromWebArchiveData):
3083
3084 2010-10-26  Antonio Gomes  <agomes@rim.com>
3085
3086         Reviewed by Martin Robinson.
3087
3088         Remove the absolute path used to include DumpRenderTreeSupportGtk.h from LayoutTestController.h
3089
3090         It turns out this is rather unneeded since WebKitTools/GNUMakefile.am has WebKit/gtk/ in its include
3091         path, and then we can just do #include "WebCoreSupport/DumpRenderTreeSupportGtk.h"
3092
3093         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3094
3095 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3096
3097         Rename WebArchiveDumpSupport.mm to WebArchiveDumpSupport.cpp
3098
3099         Reviewed by Adam Roben.
3100
3101         Part 4 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3102
3103         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3104         Updated for file moves and renames.
3105         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.mm.
3106         * DumpRenderTree/cf/WebArchiveDumpSupport.h: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.h.
3107
3108 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3109
3110         Convert WebArchiveDumpSupport.mm from NS objects to CF types
3111
3112         Reviewed by Adam Roben.
3113
3114         Part 3 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3115
3116         In order to share WebArchive code between the Mac and Windows
3117         ports, the code in WebArchiveDumpSupport.mm was converted from
3118         Cocoa to C++ using CoreFoundation (CF) types.
3119
3120         * DumpRenderTree/mac/DumpRenderTree.mm:
3121         (dump): Renamed serializeWebArchiveToXML() to
3122         createXMLStringFromWebArchiveData() and added HardAutorelease()
3123         to prevent leaks.
3124         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
3125         (createXMLStringFromWebArchiveData): Renamed from
3126         serializeWebArchiveToXML().  Changed to use CF types.
3127         (createCFURLResponseFromResponseData): Renamed from
3128         unarchiveNSURLResponseFromResponseData().  Changed to use CF
3129         types for its parameter and return type.
3130         * DumpRenderTree/mac/WebArchiveDumpSupport.mm: Replaced use of
3131         NS objects with CF types.  It will be renamed to *.cpp in a
3132         future commit.
3133         (convertMIMEType): Changed to use case-insensitive string
3134         comparisons.
3135         (convertWebResourceDataToString):
3136         (normalizeHTTPResponseHeaderFields):
3137         (normalizeWebResourceURL):
3138         (convertWebResourceResponseToDictionary):
3139         (compareResourceURLs):
3140         (createXMLStringFromWebArchiveData):
3141         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
3142         (createCFURLResponseFromResponseData): Renamed from
3143         unarchiveNSURLResponseFromResponseData().  Changed to use CF
3144         types for its parameter and return type.
3145
3146 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3147
3148         Extract use of NSKeyedUnarchiver from WebArchiveDumpSupport.mm
3149
3150         Reviewed by Adam Roben.
3151
3152         Part 2 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3153
3154         There is no equivalent to NSKeyedUnarchiver in CoreFoundation,
3155         so extract it into a platform-specific source file.
3156
3157         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
3158         (unarchiveNSURLResponseFromResponseData): Added declaration.
3159         * DumpRenderTree/mac/WebArchiveDumpSupport.mm:
3160         (convertWebResourceResponseToDictionary): Extracted code to
3161         unarchiveNSURLResponseFromResponseData() in
3162         WebArchiveDumpSupportMac.mm.  Updated logic to return early if
3163         nil is returned from unarchiveNSURLResponseFromResponseData().
3164         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
3165         (unarchiveNSURLResponseFromResponseData): Added.  Extracted
3166         code from convertWebResourceResponseToDictionary() in
3167         WebArchiveDumpSupport.mm.
3168
3169 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3170
3171         Extract call to -[WebHTMLRepresentation supportedNonImageMIMETypes] from WebArchiveDumpSupport.mm
3172
3173         Reviewed by Adam Roben.
3174
3175         Part 1 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3176
3177         The call to -[WebHTMLRepresentation supportedNonImageMIMETypes]
3178         is not cross-platform between Mac and Windows, so extract it
3179         into a platform-specific source file.
3180
3181         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Added
3182         WebArchiveDumpSupportMac.mm to the project.
3183         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
3184         (supportedNonImageMIMETypes): Added declaration.
3185         * DumpRenderTree/mac/WebArchiveDumpSupport.mm:
3186         (convertWebResourceDataToString): Replaced call to
3187         -[WebHTMLRepresentation supportedNonImageMIMETypes] with
3188         supportedNonImageMIMETypes().
3189         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm: Added.
3190         (supportedNonImageMIMETypes): Added.  Extracted from
3191         WebArchiveDumpSupport.mm.
3192
3193 2010-10-26  Antonio Gomes  <agomes@rim.com>
3194
3195         Reviewed by Martin Robinson.
3196
3197         [GTK] Implement DumpRenderTreeSupportGtk (similarly to DumpRenderTreeSupportQt idea)
3198         https://bugs.webkit.org/show_bug.cgi?id=48199
3199
3200         Implements support to WebKitTabToLinksPreferenceKey through LayoutTestController::overridePreference.
3201         The corresponding DumpRenderTreeSupportGtk method is called in the DRT context only.
3202
3203         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3204         (resetDefaultsToConsistentValues):
3205         (createWebView):
3206         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3207         (LayoutTestController::overridePreference):
3208
3209 2010-10-26  Ariya Hidayat  <ariya@sencha.com>
3210
3211         Unreviewed, change the order of my emails for bugzilla autocompletion.
3212
3213         * Scripts/webkitpy/common/config/committers.py:
3214
3215 2010-10-26  Eric Seidel  <eric@webkit.org>
3216
3217         Reviewed by David Kilzer.
3218
3219         build-webkit should collect Visual Studio Express logs and display them
3220         https://bugs.webkit.org/show_bug.cgi?id=39199
3221
3222         * Scripts/build-webkit:
3223         * Scripts/print-vse-failure-logs: Added.
3224         * Scripts/webkitdirs.pm:
3225
3226 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3227
3228         <http://webkit.org/b/48224> build-webkit: add support for --meter-tag switch
3229
3230         Reviewed by Csaba Osztrogonác.
3231
3232         * Scripts/build-webkit: Added support for --meter-tag switch.
3233         It should have been added with r60820.
3234
3235 2010-10-26  Ademar de Souza Reis Jr.  <ademar.reis@openbossa.org>
3236
3237         Reviewed by Dumitru Daniliuc.
3238
3239         check-webkit-style fails on operator+=, operator-=, ... methods
3240         https://bugs.webkit.org/show_bug.cgi?id=48258
3241
3242         * Scripts/webkitpy/style/checkers/cpp.py: Added exceptions
3243         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added unit-tests
3244
3245 2010-10-26  Ojan Vafai  <ojan@chromium.org>
3246
3247         Reviewed by Tony Chang.
3248
3249         remove DEFER support from new-run-webkit-tests
3250         https://bugs.webkit.org/show_bug.cgi?id=48387
3251
3252         DEFER was needed when we were trying to ship Chrome beta.
3253         Now it's just extra complication.
3254
3255         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
3256         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
3257         * Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
3258         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
3259         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3260         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
3261
3262 2010-10-26  Eric Seidel  <eric@webkit.org>
3263
3264         Unreviewed.  Adding import with_statement to fix python 2.5.
3265
3266         Teach webkit-patch how to read credentials from the environment
3267         https://bugs.webkit.org/show_bug.cgi?id=48275
3268
3269         * Scripts/webkitpy/common/net/credentials_unittest.py:
3270          - import with_statement to fix python 2.5.
3271
3272 2010-10-26  Alexey Proskuryakov  <ap@apple.com>
3273
3274         Reviewed by Sam Weinig.
3275
3276         https://bugs.webkit.org/show_bug.cgi?id=48375
3277         <rdar://problem/8392724> Need delegate calls in PageLoaderClient to indicate if we have
3278         loaded insecure content
3279
3280         Updated for WebKit2 changes.
3281
3282         * MiniBrowser/mac/BrowserWindowController.m:
3283         (didDisplayInsecureContentForFrame):
3284         (didRunInsecureContentForFrame):
3285         (-[BrowserWindowController awakeFromNib]):
3286         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3287         (WTR::InjectedBundlePage::InjectedBundlePage):
3288         (WTR::InjectedBundlePage::didDisplayInsecureContentForFrame):
3289         (WTR::InjectedBundlePage::didRunInsecureContentForFrame):
3290         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
3291         * WebKitTestRunner/TestController.cpp:
3292         (WTR::TestController::initialize):
3293
3294 2010-10-26  Adam Roben  <aroben@apple.com>
3295
3296         Clear up confusion between "3D rendering" and "accelerated compositing"
3297
3298         WebKit has no software-based 3D rendering implementation, so 3D
3299         rendering can only be enabled when accelerated compositing is. But DRT
3300         was falsely reporting that 3D rendering was available on machines that
3301         don't support accelerated compositing, leading to much confusion.
3302
3303         Reviewed by Darin Adler.
3304
3305         Fixes <http://webkit.org/b/48370> REGRESSION (r70540): Many
3306         transforms/3d tests are failing on the XP bots
3307
3308         * DumpRenderTree/win/DumpRenderTree.cpp:
3309         (main): When ENABLE(3D_RENDERING) is turned on, only report that 3D
3310         rendering is available when acclerated compositing is also available.
3311
3312         * Scripts/old-run-webkit-tests: Skip tests that use the
3313         -webkit-transform-3d media query when 3D rendering is disabled,
3314         rather than when accelerated compositing is disabled.
3315
3316 2010-10-26  Eric Seidel  <eric@webkit.org>
3317
3318         Reviewed by Ojan Vafai.
3319
3320         Remove rietveld code now that it's unused
3321         https://bugs.webkit.org/show_bug.cgi?id=48359
3322
3323         Was cool that we added this support, but now that it's
3324         not used it makes little sense to keep it around.
3325         We can always restore this code from SVN if we need it.
3326
3327         * Scripts/webkitpy/common/config/__init__.py:
3328         * Scripts/webkitpy/common/net/bugzilla.py:
3329         * Scripts/webkitpy/common/net/bugzilla_unittest.py:
3330         * Scripts/webkitpy/common/net/rietveld.py: Removed.
3331         * Scripts/webkitpy/common/net/rietveld_unittest.py: Removed.
3332         * Scripts/webkitpy/thirdparty/__init__.py:
3333         * Scripts/webkitpy/tool/commands/download.py:
3334         * Scripts/webkitpy/tool/commands/download_unittest.py:
3335         * Scripts/webkitpy/tool/commands/queues.py:
3336         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3337         * Scripts/webkitpy/tool/main.py:
3338         * Scripts/webkitpy/tool/mocktool.py:
3339         * Scripts/webkitpy/tool/steps/__init__.py:
3340         * Scripts/webkitpy/tool/steps/postcodereview.py: Removed.
3341
3342 2010-10-26  Anders Carlsson  <andersca@apple.com>
3343
3344         Reviewed by Sam Weinig.
3345
3346         pageDidScroll callback should be on the UI process client rather than (or in addition to) the web process client
3347         https://bugs.webkit.org/show_bug.cgi?id=48366
3348         <rdar://problem/8595202>
3349
3350         * MiniBrowser/mac/BrowserWindowController.m:
3351         (-[BrowserWindowController awakeFromNib]):
3352         * WebKitTestRunner/TestController.cpp:
3353         (WTR::createOtherPage):
3354         (WTR::TestController::initialize):
3355
3356 2010-10-26  Eric Seidel  <eric@webkit.org>
3357
3358         Reviewed by Tony Chang.
3359
3360         Teach webkit-patch how to read credentials from the environment
3361         https://bugs.webkit.org/show_bug.cgi?id=48275
3362
3363         This makes it possible for svn users to have their bugzilla credentials
3364         stored in their environment instead of typing them every time.
3365
3366         We need this for making it easy to run the win-ews bot (which currently
3367         uses svn instead of git).
3368
3369         * Scripts/webkitpy/common/net/credentials.py:
3370         * Scripts/webkitpy/common/net/credentials_unittest.py:
3371
3372 2010-10-26  Kenichi Ishibashi  <bashi@google.com>
3373
3374         Reviewed by Kent Tamura.
3375
3376         Input Method inserts conversion candidates unexpectedly
3377         https://bugs.webkit.org/show_bug.cgi?id=46868
3378
3379         Adds setComposition() to TextInputController to make DRT emulate
3380         an input method behavior.
3381
3382         * DumpRenderTree/chromium/TextInputController.cpp:
3383         (TextInputController::TextInputController):
3384         (TextInputController::setComposition): Added.
3385         * DumpRenderTree/chromium/TextInputController.h:
3386
3387 2010-10-26  Simon Fraser  <simon.fraser@apple.com>
3388
3389         Reviewed by Sam Weinig.
3390
3391         Support layoutTestController.layerTreeAsText in WebKitTestRunner
3392         https://bugs.webkit.org/show_bug.cgi?id=42145
3393         
3394         Implement layerTreeAsText() in WebKitTestRunner.
3395
3396         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
3397         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
3398         (WTR::LayoutTestController::layerTreeAsText):
3399         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
3400
3401 2010-10-26  Adam Barth  <abarth@webkit.org>
3402
3403         Reviewed by Eric Seidel.
3404
3405         webkit-patch rollout produces incorrect patch when using svn move
3406         https://bugs.webkit.org/show_bug.cgi?id=48244
3407
3408         We need to flush our caches when we modify the working copy.
3409
3410         * Scripts/webkitpy/tool/steps/revertrevision.py:
3411
3412 2010-10-26  Eric Seidel  <eric@webkit.org>
3413
3414         Reviewed by Adam Barth.
3415
3416         commit-queue flaky test message can list the same author more than once
3417         https://bugs.webkit.org/show_bug.cgi?id=48268
3418
3419         tonikitoo reported to me over IRC this morning that he's seen
3420         the commit-queue report flaky tests with author lists like:
3421         "adam, adam and adam", suggesting we're not uniquing authors
3422         before writing the message.
3423
3424         I fixed the uniquing and added a bunch more unit testing.
3425
3426         * Scripts/webkitpy/tool/commands/queues.py:
3427         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3428         * Scripts/webkitpy/tool/mocktool.py:
3429
3430 2010-10-26  Adam Roben  <aroben@apple.com>
3431
3432         Pull in the FeatureDefines*.vsprops files when building DRT
3433
3434         This ensures that various ENABLE() macros will get set correctly.
3435         (Prior to r70320 we were picking up the ENABLE(3D_RENDERING) flag
3436         through wtf/Platform.h.)
3437
3438         Fixes <http://webkit.org/b/48343> REGRESSION (r70320): DumpRenderTree
3439         says 3D_RENDERING is disabled even when it is enabled
3440
3441         Reviewed by Ada Chan.
3442
3443         * DumpRenderTree/win/DumpRenderTree.vcproj: Added
3444         FeaturesDefines*.vsprops to all configurations.
3445
3446         * DumpRenderTree/win/DumpRenderTreeWin.h: Touched to force a rebuild.
3447
3448 2010-10-26  Søren Gjesse  <sgjesse@chromium.org>
3449
3450         Reviewed by Dimitri Glazkov.
3451
3452         Added options --multiple-loads and --js-flags to chromium DumpRenderTree. The option --multiple-loads=X
3453         is used to have DumpRenderTree load each test it runs X times. To be able to have more fine-grained control
3454         of how the JavaScript engine behaves for each load the flag --js-flags can specify a list of flag-sets like this
3455
3456           --js-flags="--xxx,--noxxx --yyy,--noyyy"
3457
3458         First load will run with --xxx, the second with --yyy and the third without any (the 'no' prefix is handled by
3459         V8 to turn off the flag).
3460
3461         The changes to the Python test runner will be in a separate change.
3462
3463
3464         * DumpRenderTree/chromium/DumpRenderTree.cpp:
3465         (runTest):
3466         (main):
3467         * DumpRenderTree/chromium/TestShell.cpp:
3468         (TestShell::TestShell):
3469         (TestShell::runFileTest):
3470         (TestShell::testFinished):
3471         * DumpRenderTree/chromium/TestShell.h:
3472         (TestShell::loadCount):
3473         (TestShell::setLoadCount):
3474         (TestShell::javaScriptFlagsForLoad):
3475         (TestShell::setJavaScriptFlags):
3476         (TestShell::setDumpWhenFinished):
3477
3478 2010-10-26  Adam Roben  <aroben@apple.com>
3479
3480         Skip more tests that depend on accelerated compositing when accelerated
3481         compositing is disabled
3482
3483         Fixes <http://webkit.org/b/48329> Some tests fail when accelerated
3484         compositing is disabled
3485
3486         Reviewed by John Sullivan.
3487
3488         * Scripts/old-run-webkit-tests: Added more tests to skip when
3489         accelerated compositing is disable. The tests all have output that
3490         changes depending on whether accelerated compositing is enabled.
3491
3492 2010-10-25  Eric Seidel  <eric@webkit.org>
3493
3494         Reviewed by Adam Barth.
3495
3496         webkit-patch should clean up svn locks when passed --force-clean
3497         https://bugs.webkit.org/show_bug.cgi?id=48269
3498
3499         * Scripts/webkitpy/common/checkout/scm.py:
3500
3501 2010-10-25  Anders Carlsson  <andersca@apple.com>
3502
3503         Reviewed by Adam Roben.
3504
3505         Add a pageDidScroll BundleUIClient callback
3506         https://bugs.webkit.org/show_bug.cgi?id=48260
3507         <rdar://problem/8531159>
3508
3509         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3510         (WTR::InjectedBundlePage::InjectedBundlePage):
3511
3512 2010-10-25  Johnny Ding  <jnd@chromium.org>
3513
3514         Reviewed by Tony Chang.
3515
3516         Dump the gesture status of frame in frame load callbacks in DumpRenderTree
3517         by adding a new method dumpUserGestureInFrameLoadCallbacks.
3518         Now only dump the gesture status in "DidStartProvisionalLoad" callback.
3519         https://bugs.webkit.org/show_bug.cgi?id=47849
3520
3521         * DumpRenderTree/LayoutTestController.cpp:
3522         (LayoutTestController::LayoutTestController):
3523         (dumpUserGestureInFrameLoadCallbacksCallback):
3524         (LayoutTestController::staticFunctions):
3525         * DumpRenderTree/LayoutTestController.h:
3526         (LayoutTestController::dumpUserGestureInFrameLoadCallbacks):
3527         (LayoutTestController::setDumpUserGestureInFrameLoadCallbacks):
3528         * DumpRenderTree/chromium/LayoutTestController.cpp:
3529         (LayoutTestController::LayoutTestController):
3530         (LayoutTestController::dumpUserGestureInFrameLoadCallbacks):
3531         (LayoutTestController::reset):
3532         * DumpRenderTree/chromium/LayoutTestController.h:
3533         (LayoutTestController::shouldDumpUserGestureInFrameLoadCallbacks):
3534         (LayoutTestController::setShouldDumpUserGestureInFrameLoadCallbacks):
3535         * DumpRenderTree/chromium/TestShell.h:
3536         (TestShell::shouldDumpUserGestureInFrameLoadCallbacks):
3537         * DumpRenderTree/chromium/WebViewHost.cpp:
3538         (WebViewHost::didStartProvisionalLoad):
3539         (WebViewHost::printFrameUserGestureStatus):
3540         * DumpRenderTree/chromium/WebViewHost.h:
3541         * DumpRenderTree/mac/FrameLoadDelegate.mm:
3542         (-[WebFrame _drt_printFrameUserGestureStatus]):
3543         (-[FrameLoadDelegate webView:didStartProvisionalLoadForFrame:]):
3544         (-[FrameLoadDelegate webView:didCommitLoadForFrame:]):
3545         (-[FrameLoadDelegate webView:didFailProvisionalLoadWithError:forFrame:]):
3546         (-[FrameLoadDelegate webView:didFinishLoadForFrame:]):
3547         (-[FrameLoadDelegate webView:didFailLoadWithError:forFrame:]):
3548         (-[FrameLoadDelegate webView:windowScriptObjectAvailable:]):
3549         (-[FrameLoadDelegate webView:didReceiveTitle:forFrame:]):
3550         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3551         (LayoutTestController::reset):
3552         (LayoutTestController::dumpUserGestureInFrameLoadCallbacks):
3553         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3554
3555 2010-10-25  Sheriff Bot  <webkit.review.bot@gmail.com>
3556
3557         Unreviewed, rolling out r70442.
3558         http://trac.webkit.org/changeset/70442
3559         https://bugs.webkit.org/show_bug.cgi?id=48248
3560
3561         http locking doesn't work on Windows (Requested by Ossy on
3562         #webkit).
3563
3564         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3565         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3566
3567 2010-10-25  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3568
3569         Reviewed by Ojan Vafai.
3570
3571         Make http locking default in NRWT.
3572         https://bugs.webkit.org/show_bug.cgi?id=48053
3573
3574         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3575         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3576
3577 2010-10-25  David Kilzer  <ddkilzer@apple.com>
3578
3579         <http://webkit.org/b/48185> build-webkit: add support for --progress-tag switch
3580         (Re-landing r70413 after Qt minimal buildfix: r70440.)
3581
3582         Reviewed by Kenneth Rohde Christiansen.
3583
3584         * Scripts/build-webkit: Added support for --progress-tag switch.
3585         It should have been added with r57051.
3586
3587 2010-10-24  Adam Barth  <abarth@webkit.org>
3588
3589         Reviewed by David Kilzer.
3590
3591         webkit-patch upload fails when the patch removes a file
3592         https://bugs.webkit.org/show_bug.cgi?id=48187
3593
3594         We need to use "--" to separate file names from the rest of the
3595         command.
3596
3597         * Scripts/webkitpy/common/checkout/scm.py:
3598
3599 2010-10-24  Sheriff Bot  <webkit.review.bot@gmail.com>
3600
3601         Unreviewed, rolling out r70413.
3602         http://trac.webkit.org/changeset/70413
3603         https://bugs.webkit.org/show_bug.cgi?id=48210
3604
3605         It broke Qt minimal build (Requested by Ossy on #webkit).
3606
3607         * Scripts/build-webkit:
3608
3609 2010-10-24  Yi Shen  <yi.4.shen@nokia.com>
3610
3611         Reviewed by Andreas Kling.
3612
3613         [Qt][QtTestBrowser] Toggle use of QGraphicsView messes up the menu
3614         https://bugs.webkit.org/show_bug.cgi?id=48141
3615
3616         * QtTestBrowser/launcherwindow.cpp:
3617         (LauncherWindow::init):
3618         (LauncherWindow::initializeView):
3619         (LauncherWindow::toggleWebView):
3620
3621 2010-10-24  Daniel Bates  <dbates@rim.com>
3622
3623         Reviewed by David Kilzer.
3624
3625         Fix Perl uninitialized warnings in VCSUtils::svnStatus()
3626         and VCSUtils::removeEOL().
3627         https://bugs.webkit.org/show_bug.cgi?id=48196
3628
3629         VCSUtils::svnStatus() concatenates the output of svn status with
3630         a new line character and svn status may return no output (say for
3631         a file that has not been added, deleted, or modified). We should
3632         only concatenate the output of svn status if there is some.
3633
3634         Also, VCSUtils::removeEOL() should ensure that its argument
3635         is initialized before performing a string substitution.
3636
3637         * Scripts/VCSUtils.pm:
3638           - Modified removeEOL() to return "" if its argument is undefined.
3639           - Exported removeEOL() so that it can be tested. 
3640         * Scripts/webkitperl/VCSUtils_unittest/removeEOL.pl: Added.
3641
3642 2010-10-24  David Kilzer  <ddkilzer@apple.com>
3643
3644         <http://webkit.org/b/48185> build-webkit: add support for --progress-tag switch
3645
3646         Reviewed by Kenneth Rohde Christiansen.
3647
3648         * Scripts/build-webkit: Added support for --progress-tag switch.
3649         It should have been added with r57051.
3650
3651 2010-10-24  David Kilzer  <ddkilzer@apple.com>
3652
3653         <http://webkit.org/b/48184> build-webkit: reformat support variables for better maintainability
3654
3655         Reviewed by Daniel Bates.
3656
3657         * Scripts/build-webkit: Alphabetized the @features array based
3658         on the name of the feature.  Reformatted the list of support
3659         variables so that they match the order of @features, and so that
3660         each has its own line.  This makes it easy to add new variables
3661         in the correct order.
3662
3663 2010-10-23  Eric Seidel  <eric@webkit.org>
3664
3665         Reviewed by Adam Barth.
3666
3667         EWS never removes invalid patch ids
3668         https://bugs.webkit.org/show_bug.cgi?id=48173
3669
3670         This is just sticking another finger in the dam.
3671         However this adds more unit testing which will help
3672         us make sure we're always releasing patches once we
3673         redesign the release_patch API and call these from
3674         a more central place.
3675
3676         * Scripts/webkitpy/tool/commands/queues.py:
3677         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3678         * Scripts/webkitpy/tool/mocktool.py:
3679          - Added the ability to request invalid patches.
3680            Log a warning message to make sure we don't ever have
3681            tests use invalid patch fetches by mistake.
3682
3683 2010-10-23  Dan Bernstein  <mitz@apple.com>
3684
3685         Build fix. Add stub implementations for required NSDraggingInfo methods.
3686
3687         * DumpRenderTree/mac/DumpRenderTreeDraggingInfo.mm:
3688         (-[DumpRenderTreeDraggingInfo draggingFormation]):
3689         (-[DumpRenderTreeDraggingInfo setDraggingFormation:]):
3690         (-[DumpRenderTreeDraggingInfo animatesToDestination]):
3691         (-[DumpRenderTreeDraggingInfo setAnimatesToDestination:]):
3692         (-[DumpRenderTreeDraggingInfo numberOfValidItemsForDrop]):
3693         (-[DumpRenderTreeDraggingInfo setNumberOfValidItemsForDrop:]):
3694         (-[DumpRenderTreeDraggingInfo enumerateDraggingItemsWithOptions:forView:classes:searchOptions:usingBlock:]):
3695
3696 2010-10-23  David Kilzer  <ddkilzer@apple.com>
3697
3698         <http://webkit.org/b/48186> Remove unneeded WebHTMLRepresentationInternal.h header
3699
3700         Reviewed by Sam Weinig.
3701
3702         The only method defined in WebHTMLRepresentationInternal.h is
3703         also defined in WebHTMLRepresentation.h, so use that instead.
3704
3705         * DumpRenderTree/mac/InternalHeaders/WebKit/WebHTMLRepresentationInternal.h: Removed.
3706         * DumpRenderTree/mac/WebArchiveDumpSupport.mm: Switched to use
3707         WebHTMLRepresentation.h.
3708
3709 2010-10-23  Sheriff Bot  <webkit.review.bot@gmail.com>
3710
3711         Unreviewed, rolling out r70367.
3712         http://trac.webkit.org/changeset/70367
3713         https://bugs.webkit.org/show_bug.cgi?id=48176
3714
3715         It made 8-10 tests crash on Qt bot (Requested by Ossy on
3716         #webkit).
3717
3718         * DumpRenderTree/LayoutTestController.cpp:
3719         (setCustomPolicyDelegateCallback):
3720         * DumpRenderTree/LayoutTestController.h:
3721         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3722         (LayoutTestController::setCustomPolicyDelegate):
3723         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3724         (LayoutTestController::setCustomPolicyDelegate):
3725         * DumpRenderTree/mac/PolicyDelegate.h:
3726         * DumpRenderTree/mac/PolicyDelegate.mm:
3727         (-[PolicyDelegate webView:decidePolicyForNavigationAction:request:frame:decisionListener:]):
3728         (-[PolicyDelegate setPermissive:]):
3729         (-[PolicyDelegate setControllerToNotifyDone:]):
3730         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3731         (LayoutTestController::setCustomPolicyDelegate):
3732         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3733         (LayoutTestController::setCustomPolicyDelegate):
3734
3735 2010-10-22  Eric Seidel  <eric@webkit.org>
3736
3737         Reviewed by Adam Barth.
3738
3739         EWS never releases patches which fail to apply
3740         https://bugs.webkit.org/show_bug.cgi?id=48171
3741
3742         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
3743         * Scripts/webkitpy/tool/commands/queues.py:
3744
3745 2010-10-22  Eric Seidel  <eric@webkit.org>
3746
3747         Unreviewed.
3748
3749         Make the EWS queues restart themselves more often.
3750         This matches the commit-queue.
3751
3752         * EWSTools/start-queue.sh:
3753
3754 2010-10-22  Eric Seidel  <eric@webkit.org>
3755
3756         Reviewed by Adam Barth.
3757
3758         commit-queue should not pass --quiet to subcommands now that the parent command does the reporting
3759         https://bugs.webkit.org/show_bug.cgi?id=48165
3760
3761         * Scripts/webkitpy/tool/bot/commitqueuetask.py:
3762         * Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
3763         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3764
3765 2010-10-22  Mario Sanchez Prada  <msanchez@igalia.com>
3766
3767         Adding myself as a committer.
3768
3769         * Scripts/webkitpy/common/config/committers.py:
3770
3771 2010-10-20  Anders Carlsson  <andersca@apple.com>
3772
3773         Reviewed by Adam Barth and Darin Adler.
3774
3775         Using the decidePolicyForMIMEType delegate message in an asynchronous manner does not work
3776         https://bugs.webkit.org/show_bug.cgi?id=48014
3777         <rdar://problem/8202716>
3778         
3779         * DumpRenderTree/LayoutTestController.cpp:
3780         (setCustomPolicyDelegateCallback):
3781         Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
3782
3783         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3784         (LayoutTestController::setCustomPolicyDelegate):
3785         Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
3786
3787         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3788         (LayoutTestController::setCustomPolicyDelegate):
3789         Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
3790
3791         * DumpRenderTree/mac/PolicyDelegate.h:
3792         * DumpRenderTree/mac/PolicyDelegate.mm:
3793         (-[PolicyDelegate webView:decidePolicyForMIMEType:request:frame:decisionListener:]):
3794         If _callIgnoreInDecidePolicyForMIMETypeAfterOneSecond is true, call [listener ignore] after one second.
3795
3796         (-[PolicyDelegate setCallIgnoreInDecidePolicyForMIMETypeAfterOneSecond:]):
3797         Update _callIgnoreInDecidePolicyForMIMETypeAfterOneSecond.
3798         
3799         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3800         (LayoutTestController::setCustomPolicyDelegate):
3801         Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
3802
3803         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3804         (LayoutTestController::setCustomPolicyDelegate):
3805         Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
3806
3807 2010-10-22  Sam Weinig  <sam@webkit.org>
3808
3809         Reviewed by Anders Carlsson.
3810
3811         Rename the InjectedBundle for TestWebKitAPI to InjectedBundleTestWebKitAPI.bundle
3812
3813         * TestWebKitAPI/Configurations/InjectedBundle.xcconfig:
3814         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
3815         * TestWebKitAPI/mac/PlatformUtilitiesMac.mm:
3816         (TestWebKitAPI::Util::createInjectedBundlePath):
3817
3818 2010-10-22  Dirk Pranke  <dpranke@chromium.org>
3819
3820         Reviewed by Eric Seidel.
3821
3822         test-webkitpy fails when passed -v (or any other option)
3823
3824         Fix port/base_unittest to pass an explicit list of arguments
3825         rather than accidentally picking up sys.argv.
3826
3827         https://bugs.webkit.org/show_bug.cgi?id=48071
3828
3829         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
3830
3831 2010-10-22  Brian Weinstein  <bweinstein@apple.com>
3832
3833         Windows build fix. Update the createNewPage callback to account for the new
3834         arguments.
3835
3836         * MiniBrowser/win/BrowserView.cpp:
3837         (createNewPage):
3838
3839 2010-10-22  Sam Weinig  <sam@webkit.org>
3840
3841         Reviewed by Anders Carlsson.
3842
3843         WebKit2 needs to pass the current event modifier flags when requesting a new window
3844         https://bugs.webkit.org/show_bug.cgi?id=48140
3845
3846         * MiniBrowser/mac/BrowserWindowController.m:
3847         (createNewPage):
3848         * WebKitTestRunner/TestController.cpp:
3849         (WTR::createOtherPage):
3850         Update for new signature for the WKPageCreateNewPageCallback.
3851
3852 2010-10-22  Eric Seidel  <eric@webkit.org>
3853
3854         Unreviewed.  Just adding a test case.
3855
3856         The style-queue was not recognizing new patches.  It turns
3857         out I had fixed this in an earlier change, but just not
3858         deployed to queues.webkit.org.  As part of investigating why it
3859         was broken, I wrote a test for my previous change which I'm now landing.
3860
3861         * QueueStatusServer/model/queues_unittest.py:
3862
3863 2010-10-22  Eric Seidel  <eric@webkit.org>
3864
3865         Reviewed by Adam Barth.
3866
3867         EWS does not need to process obsolete patches
3868         https://bugs.webkit.org/show_bug.cgi?id=48093
3869
3870         This was an easy change, but to test it I had to pipe
3871         real Attachment objects into the queue testing system.
3872         Doing so revealed a whole bunch of bugs in our unit tests,
3873         which I fixed as part of this patch.
3874
3875         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
3876          - This is the actual code change.  This will not reduce the
3877            backlog in the EWS queues much, but it will make rejections
3878            much quicker for obsolete patches or closed bugs.
3879         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
3880          - Test my new code.
3881          - Used a real attachment object and got rid of MockPatch
3882          - Shared code between the mac-ews and cr-mac-ews tests.
3883         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3884          - Can't use MockPatch anymore.
3885          - Removing MockPatch found more bugs here!
3886         * Scripts/webkitpy/tool/commands/queuestest.py:
3887         * Scripts/webkitpy/tool/commands/sheriffbot_unittest.py:
3888         * Scripts/webkitpy/tool/mocktool.py:
3889          - MockBugzilla should not be a "Mock" object.  Right now tool.bugs()
3890            is allowed, but wrong.  Making it not a Mock will make tool.bugs() correctly fail.
3891
3892 2010-10-22  Sheriff Bot  <webkit.review.bot@gmail.com>
3893
3894         Unreviewed, rolling out r70301.
3895         http://trac.webkit.org/changeset/70301
3896         https://bugs.webkit.org/show_bug.cgi?id=48126
3897
3898         "Lang attribute layout tests failing" (Requested by satish on
3899         #webkit).
3900
3901         * DumpRenderTree/LayoutTestController.cpp:
3902         (setMockSpeechInputResultCallback):
3903         * DumpRenderTree/LayoutTestController.h:
3904         * DumpRenderTree/chromium/LayoutTestController.cpp:
3905         (LayoutTestController::setMockSpeechInputResult):
3906         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3907         (LayoutTestController::setMockSpeechInputResult):
3908         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3909         (LayoutTestController::setMockSpeechInputResult):
3910         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3911         (LayoutTestController::setMockSpeechInputResult):
3912         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3913         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3914         (LayoutTestController::setMockSpeechInputResult):
3915         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3916         (LayoutTestController::setMockSpeechInputResult):
3917
3918 2010-10-22  Leandro Gracia Gil  <leandrogracia@google.com>
3919
3920         Reviewed by Jeremy Orlow.
3921
3922         Added a second parameter to setMockSpeechInputResult for
3923         the language used in speech input.
3924         https://bugs.webkit.org/show_bug.cgi?id=47089
3925
3926         * DumpRenderTree/LayoutTestController.cpp:
3927         (setMockSpeechInputResultCallback):
3928         * DumpRenderTree/LayoutTestController.h:
3929         * DumpRenderTree/chromium/LayoutTestController.cpp:
3930         (LayoutTestController::setMockSpeechInputResult):
3931         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3932         (LayoutTestController::setMockSpeechInputResult):
3933         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3934         (LayoutTestController::setMockSpeechInputResult):
3935         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3936         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3937         (LayoutTestController::setMockSpeechInputResult):
3938         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3939         (LayoutTestController::setMockSpeechInputResult):
3940         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3941         (LayoutTestController::setMockSpeechInputResult):
3942
3943 2010-10-21  MORITA Hajime  <morrita@google.com>
3944
3945         Reviewed by Kent Tamura.
3946
3947         [Win][DRT] should have LayoutTestController.hasSpellingMarker()
3948         https://bugs.webkit.org/show_bug.cgi?id=47885
3949
3950         - Implemented LayoutTestController.hasSpellingMarker(),
3951         - Added fake spellcheck implementation.
3952
3953         * DumpRenderTree/win/DumpRenderTree.cpp:
3954         (main):
3955         * DumpRenderTree/win/EditingDelegate.cpp:
3956         (indexOfFirstWordCharacter):
3957         (wordLength):
3958         (EditingDelegate::checkSpellingOfString):
3959         * DumpRenderTree/win/EditingDelegate.h:
3960         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3961         (LayoutTestController::hasSpellingMarker):
3962
3963 2010-10-21  Mihai Parparita  <mihaip@chromium.org>
3964
3965         Unreviewed. Re-enable test that was mistakenly disabled by r67974.
3966
3967         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
3968
3969 2010-10-21  Mihai Parparita  <mihaip@chromium.org>
3970
3971         Unreviewed fix for rebaseline-chromium-webkit-tests.
3972
3973         Port.diff_image no longer has a tolerance parameter. Also, use
3974         get_option('tolerance') which is safer if the options object doesn't
3975         define a tolerance attribute (it doesn't for the one used in
3976         rebaseline_chromium_webkit_tests).
3977
3978         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
3979         * Scripts/webkitpy/layout_tests/port/webkit.py:
3980         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
3981
3982 2010-10-21  Mihai Parparita  <mihaip@chromium.org>
3983
3984         Reviewed by Ojan Vafai.
3985
3986         Add support for --tolerance in NRWT
3987         https://bugs.webkit.org/show_bug.cgi?id=47959
3988         
3989         Add support for the --tolerance flag in NRWT. The Port.diff_image
3990         signature shouldn't need a tolerance parameter (it's not set per test),
3991         just have ports that use it (currently only WebKitPort) read it from 
3992         the options object.
3993
3994         * Scripts/webkitpy/layout_tests/port/chromium.py:
3995         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
3996         * Scripts/webkitpy/layout_tests/port/test.py:
3997         * Scripts/webkitpy/layout_tests/port/webkit.py:
3998         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3999         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
4000
4001 2010-10-21  Eric Seidel  <eric@webkit.org>
4002
4003         Reviewed by Adam Barth.
4004
4005         Add --suggest-reviewers option to upload to auto-suggest reviewers for your patch
4006         https://bugs.webkit.org/show_bug.cgi?id=48088
4007
4008         This is a first-pass.  Works, but we'll eventually
4009         turn this on by default, refine the suggestion algorithm
4010         and possibly move it to a different place in the upload step order.
4011
4012         * Scripts/webkitpy/common/checkout/api.py:
4013         * Scripts/webkitpy/tool/commands/download_unittest.py:
4014         * Scripts/webkitpy/tool/commands/upload.py:
4015         * Scripts/webkitpy/tool/commands/upload_unittest.py:
4016         * Scripts/webkitpy/tool/mocktool.py:
4017         * Scripts/webkitpy/tool/steps/__init__.py:
4018         * Scripts/webkitpy/tool/steps/options.py:
4019
4020 2010-10-21  Mario Sanchez Prada  <msanchez@igalia.com>
4021
4022         Reviewed by Chris Fleizach.
4023
4024         [GTK] Segfault while testing accessibility/iframe-bastardization.html
4025         https://bugs.webkit.org/show_bug.cgi?id=30123
4026
4027         Check m_element before actually using it to get the parent.
4028
4029         * DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
4030         (AccessibilityUIElement::parentElement): Add an extra check to
4031         make sure m_element points to a valid value before actually using
4032         it to get the parent. Also, assert ATK_IS_OBJECT(m_element).
4033
4034 2010-10-21  Anders Carlsson  <andersca@apple.com>
4035
4036         Reviewed by Adam Roben.
4037
4038         Crash evaluating JavaScript string that throws an exception
4039         https://bugs.webkit.org/show_bug.cgi?id=48092
4040         <rdar://problem/8487657>
4041
4042         Add a test that evaluates a JavaScript string that throws an exception and check that 
4043         the callback is called.
4044
4045         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
4046         * TestWebKitAPI/Tests/WebKit2/EvaluateJavaScript.cpp: Added.
4047         (TestWebKitAPI::didRunJavaScript):
4048         (TestWebKitAPI::WebKit2_EvaluateJavaScript):
4049         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
4050
4051 2010-10-21  Adam Roben  <aroben@apple.com>
4052
4053         Test that the plugin's HWND is invalidated when NPN_InvalidateRect is
4054         called
4055
4056         Test for <http://webkit.org/b/48086> <rdar://problem/8482944>
4057         Silverlight doesn't repaint in WebKit2
4058
4059         Reviewed by Anders Carlsson.
4060
4061         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
4062         (PluginTest::NPN_InvalidateRect): Added. Calls through to the browser.
4063         (executeScript): Added. Asks the browser to evaluate the script.
4064
4065         (PluginTest::waitUntilDone):
4066         (PluginTest::notifyDone):
4067         Added. Calls through to layoutTestController.
4068
4069         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Added
4070         * NPN_InvalidateRect and waitUntilDone/notifyDone.
4071
4072         * DumpRenderTree/TestNetscapePlugIn/Tests/win/NPNInvalidateRectInvalidatesWindow.cpp: Added.
4073         (TemporaryWindowMover::moveSucceeded):
4074         (TemporaryWindowMover::TemporaryWindowMover):
4075         (TemporaryWindowMover::~TemporaryWindowMover):
4076         This class moves a window on-screen and shows it, then moves it back and hides it.
4077
4078         (NPNInvalidateRectInvalidatesWindow::NPNInvalidateRectInvalidatesWindow):
4079         Initialize our members.
4080         (NPNInvalidateRectInvalidatesWindow::~NPNInvalidateRectInvalidatesWindow):
4081         Delete our window mover if it hasn't been deleted already.
4082         (NPNInvalidateRectInvalidatesWindow::NPP_SetWindow): Subclass the
4083         plugin HWND and move the test harness window on screen.
4084         (NPNInvalidateRectInvalidatesWindow::wndProc): Call through to onPaint
4085         when we get a WM_PAINT message.
4086         (NPNInvalidateRectInvalidatesWindow::onPaint): Do the test and tell
4087         LayoutTestController we're done.
4088         (NPNInvalidateRectInvalidatesWindow::testInvalidateRect): Validate
4089         ourselves, invalidate our lower-right quadrant via NPN_InvalidateRect,
4090         then check that our HWND's invalid region is the rect that we
4091         invalidated.
4092
4093         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
4094         * Added NPNInvalidateRectInvalidatesWindow.
4095
4096 2010-10-21  Daniel Bates  <dbates@rim.com>
4097
4098         Add Git-support to do-file-rename
4099         https://bugs.webkit.org/show_bug.cgi?id=48015
4100
4101         Fix tools. Export function scmMoveOrRenameFile so that it can be called from
4102         do-file-rename and do-webcore-rename. I inadvertently left this out of the patch.
4103
4104         * Scripts/VCSUtils.pm:
4105
4106 2010-10-21  Eric Seidel  <eric@webkit.org>
4107
4108         Reviewed by Adam Barth.
4109
4110         mac-ews is not properly releasing patches
4111         https://bugs.webkit.org/show_bug.cgi?id=48076
4112
4113         mac-ews overrides process_work_item, so it was not calling
4114         release_work_item like the default process_work_item would.
4115         To fix this I made all the status-reporting methods just
4116         release the patch.  I expect we'll iterate on this design further.
4117
4118         * Scripts/webkitpy/common/net/statusserver.py:
4119         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
4120         * Scripts/webkitpy/tool/commands/queues.py:
4121         * Scripts/webkitpy/tool/commands/queues_unittest.py:
4122
4123 2010-10-18  Martin Robinson  <mrobinson@igalia.com>
4124
4125         Reviewed by Gustavo Noronha Silva.
4126
4127         [GTK] mathml flaky tests after adding SystemFonts to the theme
4128         https://bugs.webkit.org/show_bug.cgi?id=47727
4129
4130         The Fontconfig setup in the GTK+ now specifically checks for and
4131         loads the STIX fonts for MathML tests.
4132
4133         * DumpRenderTree/gtk/DumpRenderTree.cpp:
4134         (initializeFonts): Add hooks for loading the STIX fonts specifically.
4135
4136 2010-10-20  Adam Roben  <aroben@apple.com>
4137
4138         Test that the UI client gets notified when WKView receives a WM_CLOSE
4139         message
4140
4141         Test for <http://webkit.org/b/48044> <rdar://problem/8488446> Pressing
4142         Ctrl+W when viewing a full-page PDF destroys the WKView but doesn't
4143         close its parent window
4144
4145         Reviewed by Jon Honeycutt.
4146
4147         * TestWebKitAPI/Tests/WebKit2/win/WMCloseCallsUIClientClose.cpp: Added.
4148         (TestWebKitAPI::close): Record that this function was called.
4149         (TestWebKitAPI::WebKit2_WMCloseCallsUIClientClose): Create a WKView,
4150         send it a WM_CLOSE message, and test that the UI client's close
4151         callback got called. Note that this will hang if the bug is
4152         reintroduced; <http://webkit.org/b/48043> covers making TestWebKitAPI
4153         able to handle hangs.
4154
4155         * TestWebKitAPI/win/TestWebKitAPI.vcproj: Added the new test.
4156
4157 2010-10-20  Adam Roben  <aroben@apple.com>
4158
4159         Make prepare-ChangeLog much faster when using git
4160
4161         This change also seems to make it detect renames better in some cases.
4162
4163         Fixes <http://webkit.org/b/48040> prepare-ChangeLog is slow when using
4164         git
4165
4166         Reviewed by David Kilzer.
4167
4168         * Scripts/prepare-ChangeLog:
4169         (statusCommand):
4170         (createPatchCommand):
4171         Use "-M -C" instead of "-C -C -M" to tell git-diff to detect renames
4172         and copies. The two "-C"s were making it read many more files than were
4173         necessary.
4174
4175 2010-10-21  Sam Weinig  <sam@webkit.org>
4176
4177         Reviewed by Adam Roben.
4178
4179         Null frame passed when running alert from UserScript run at document start
4180         <rdar://problem/8573809>
4181         https://bugs.webkit.org/show_bug.cgi?id=48036
4182
4183         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
4184         * TestWebKitAPI/InjectedBundleController.cpp:
4185         (TestWebKitAPI::InjectedBundleController::didReceiveMessage):
4186         (TestWebKitAPI::InjectedBundleController::initializeTestNamed):
4187         * TestWebKitAPI/InjectedBundleController.h:
4188         * TestWebKitAPI/InjectedBundleTest.h:
4189         (TestWebKitAPI::InjectedBundleTest::initialize):
4190         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
4191         * TestWebKitAPI/Tests/WebKit2/DocumentStartUserScriptAlertCrash.cpp: Added.
4192         (TestWebKitAPI::runJavaScriptAlert):
4193         (TestWebKitAPI::TEST):
4194         * TestWebKitAPI/Tests/WebKit2/DocumentStartUserScriptAlertCrash_Bundle.cpp: Added.
4195         (TestWebKitAPI::DocumentStartUserScriptAlertCrashTest::DocumentStartUserScriptAlertCrashTest):
4196         (TestWebKitAPI::DocumentStartUserScriptAlertCrashTest::initialize):
4197         Add test for invoking an alert during a UserScript run at document start.
4198
4199 2010-10-21  Andreas Kling  <kling@webkit.org>
4200
4201         Reviewed by Adam Roben.
4202
4203         TestNetscapePlugIn: Actually call NPP_SetWindow on Unix
4204
4205         We were just returning NPERR_NO_ERROR previously.
4206
4207         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp: