2010-11-08 Nicolas Dufresne <nicolas.dufresne@collabora.co.uk>
[WebKit-https.git] / WebKitTools / ChangeLog
1 2010-11-08  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
2
3         Reviewed by Martin Robinson.
4
5         [GTK] Link with target name set does not work
6         https://bugs.webkit.org/show_bug.cgi?id=48865
7
8         When a new page is created with a name (target=myFrame), the new
9         mainFrame could not be found because they where not stored in the
10         same PageGroup. As PageGroup are not exposed externally so the
11         simpliest solution is to use a global page group name. This also fixes
12         issue with visited link coloration across pages. After this change the
13         private function webkit_web_view_set_group_name() was no longer used
14         so it was removed completly.
15
16         * DumpRenderTree/gtk/DumpRenderTree.cpp:
17         (createWebView):
18
19 2010-11-08  John Knottenbelt  <jknotten@chromium.org>
20
21         Reviewed by Kent Tamura.
22
23         WebViewHost::reset() uses placement new.
24         https://bugs.webkit.org/show_bug.cgi?id=49069
25
26         WebViewHost is using placement destruction / new to simulate a fresh
27         WebViewHost object at the same address. This is because the WebView remains
28         open across tests and maintains a pointer to the WebViewHost.
29         This change resets member variables explictly instead of the placement new dance.
30
31         * DumpRenderTree/chromium/WebViewHost.cpp:
32         (WebViewHost::WebViewHost):
33         (WebViewHost::reset):
34
35 2010-11-08  Dirk Pranke <dpranke@chromium.org>
36
37         Reviewed by Eric Seidel.
38
39         Fix the webkit_base_dir logic in webkitpy/layout_tests/port/config.py. 
40         It turns out that NRWT can't use the code in scm.find_checkout_root()
41         because the Chromium bots don't do full checkouts of the WebKit
42         tree; they only check out subdirectories like WebKitTools/Scripts.
43         Until we can figure out a better approach for this, I've
44         restored the base_dir-detecting code from NRWT, which works in
45         any directory tree, scm or no.
46
47         This also restores the files modified in r71475 and r71474.
48
49         https://bugs.webkit.org/show_bug.cgi?id=49151
50
51         * WebKitTools/Scripts/webkitpy/layout_tests/port/config.py
52         * WebKitTools/Scripts/webkitpy/layout_tests/port/config_unittest.py
53
54 2010-11-08  Eric Seidel  <eric@webkit.org>
55
56         Reviewed by Adam Barth.
57
58         webkit-patch failure-reason explodes if a build is missing
59         https://bugs.webkit.org/show_bug.cgi?id=49195
60
61         This is likely a recent regression.  Adding a unit test for this case.
62
63         * Scripts/webkitpy/common/net/buildbot.py:
64         * Scripts/webkitpy/common/net/buildbot_unittest.py:
65
66 2010-11-08  Dan Bernstein  <mitz@apple.com>
67
68         Reviewed by Darin Adler.
69
70         Ignore the system scrollbar setting.
71
72         * DumpRenderTree/mac/DumpRenderTree.mm:
73         (resetDefaultsToConsistentValues): Set the scrollbar default.
74
75 2010-11-05  Adam Roben  <aroben@apple.com>
76
77         Make webkitpy.common.system.executive_unittest pass when running under
78         Win32 Python
79
80         Fixes <http://webkit.org/b/49033>.
81
82         Reviewed by Dave Levin and Eric Seidel.
83
84         * Scripts/webkitpy/common/system/executive.py:
85         (Executive._run_command_with_teed_output): Pass the arguments through
86         encode_argument_if_needed rather than using Cygwin-specific code here.
87         (Executive.run_and_throw_if_fail): Use child_process_encoding to decode
88         the output.
89         (Executive.run_command): Use encode_argument_if_needed to encode the
90         arguments and child_process_encoding to decode the output.
91         (Executive._child_process_encoding): Returns the encoding that should be
92         used when communicating with child processes. On Windows we use mbcs,
93         which maps to the current code page. On all other platforms we use
94         UTF-8.
95         (Executive._should_encode_child_process_arguments): Returns True if
96         arguments to child processes need to be encoded. This is currently
97         only needed on Cygwin and Win32 Python 2.x.
98         (Executive._encode_argument_if_needed): Encode the argument using
99         child_process_encoding if we need to encode arguments to child
100         processes on this platform.
101
102         * Scripts/webkitpy/common/system/executive_unittest.py:
103         (never_ending_command): Added. Returns arguments to run a command that
104         will not quit until we kill it. On Windows we use wmic, on other
105         platforms we use yes.
106         (ExecutiveTest.test_run_command_with_unicode): Changed to expect the
107         mbcs encoding to be used and for output from the child processes to
108         have been roundtripped through encode/decode on Win32 Python. When
109         UTF-8 is the encoding the roundtripping is undetectable, but with mbcs
110         it's possible that some characters will not be able to be converted
111         and will be replaced by question marks; the round-tripping allows us
112         to expect this result.
113
114         (ExecutiveTest.test_kill_process):
115         (ExecutiveTest.test_kill_all):
116         Use never_ending_command instead of invoking "yes" directly. Expect an
117         exit code of 1 when using Win32 Python, as that's what seems to happen.
118
119 2010-11-08  Adam Roben  <aroben@apple.com>
120
121         Roll out r71532
122
123         It broke the build for Cygwin 1.7 installs. Cygwin 1.7's default
124         .bashrc unsets %TEMP%, which broke copy-tools.cmd.
125
126         * Scripts/webkitdirs.pm:
127
128 2010-11-08  Tony Chang  <tony@chromium.org>
129
130         Reviewed by Adam Barth.
131
132         run platform/chromium/plugins/nested-plugin-objects.html on all platforms
133         https://bugs.webkit.org/show_bug.cgi?id=49094
134
135         This tests that objects created by plugins are proplery cleaned up.
136
137         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
138         (testPassTestObject):
139         (pluginInvoke):
140         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
141         (testAllocate):
142         (testDeallocate):
143         (testGetProperty):
144         (testConstruct):
145
146 2010-11-08  Adam Roben  <aroben@apple.com>
147
148         Mark Windows builds triggered from Perl as being non-interactive
149
150         This affects whether some of our scripts will show alerts vs. printing
151         to the build log.
152
153         Fixes <http://webkit.org/b/49181> Windows build fail mysteriously when
154         .vsprops files are updated
155
156         Reviewed by Steve Falkenburg.
157
158         * Scripts/webkitdirs.pm:
159         (buildVisualStudioProject): Set WEBKIT_NONINTERACTIVE_BUILD to 1.
160
161 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
162
163         Reviewed by Ojan Vafai.
164
165         Make http locking default in NRWT.
166         https://bugs.webkit.org/show_bug.cgi?id=48053
167
168         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
169         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
170
171 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
172
173         Reviewed by Csaba Osztrogonác.
174
175         [NRWT] If the http lock fails we shouldn't do any locking
176         https://bugs.webkit.org/show_bug.cgi?id=49164
177
178         If something goes wrong with the locking, the test should keep going.
179
180         * Scripts/webkitpy/layout_tests/port/http_lock.py:
181
182 2010-11-08  Adam Roben  <aroben@apple.com>
183
184         Switch back to using kCGImageAlphaPremultipliedFirst when generating
185         pixel dumps on Windows
186
187         I changed this behavior in r71418 thinking that it was required for
188         getting plugins to show up in pixel dumps. But it doesn't seem to be
189         necessary, and was making it impossible to compare new Windows pixel
190         dumps with existing Windows or Mac pixel dumps (because ImageDiff won't
191         compare an image with alpha to an image without alpha).
192
193         Fixes <http://webkit.org/b/49172> REGRESION (r71418): Can't compare
194         new Windows pixel results to existing Windows or Mac results
195
196         Reviewed by Antti Koivisto.
197
198         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
199         (createBitmapContextFromWebView): Replaced kCGImageAlphaNoneSkipFirst
200         with kCGImageAlphaPremultipliedFirst.
201
202 2010-11-08  Csaba Osztrogonac  <ossy@webkit.org>
203
204         Unreviewed, rolling out r71466.
205         http://trac.webkit.org/changeset/71466
206         https://bugs.webkit.org/show_bug.cgi?id=48865
207
208         It broke layout tests on GTK bots.
209
210         * DumpRenderTree/gtk/DumpRenderTree.cpp:
211         (createWebView):
212
213 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
214
215         Reviewed by Csaba Osztrogonác.
216
217         Enable running of Qt API tests on BuildBot
218         https://bugs.webkit.org/show_bug.cgi?id=49004
219
220         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
221
222 2010-11-08  Eric Seidel  <eric@webkit.org>
223
224         Reviewed by Adam Barth.
225
226         Add clean-review-queue command to remove closed bugs from the webkit.org/pending-review
227         https://bugs.webkit.org/show_bug.cgi?id=49160
228
229         Bugzilla doesn't automatically remove r? when a bug gets closed.
230         This script takes care of that for webkit.org.
231
232         * Scripts/webkitpy/common/net/bugzilla.py:
233         * Scripts/webkitpy/tool/commands/upload.py:
234
235 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
236
237         Unreviewed, rolling out r71474.
238         http://trac.webkit.org/changeset/71474
239         https://bugs.webkit.org/show_bug.cgi?id=48280
240
241         breaks chromium webkit tests
242         https://bugs.webkit.org/show_bug.cgi?id=49151
243
244         * Scripts/webkitpy/layout_tests/port/base.py:
245         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
246         * Scripts/webkitpy/layout_tests/port/config.py:
247         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
248         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
249         * Scripts/webkitpy/layout_tests/port/webkit.py:
250         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
251         * Scripts/webkitpy/tool/mocktool.py:
252
253 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
254
255         Unreviewed, rolling out r71475.
256         http://trac.webkit.org/changeset/71475
257  
258         breaks chromium webkit tests
259         https://bugs.webkit.org/show_bug.cgi?id=49151
260
261         * Scripts/webkitpy/common/newstringio.py: Removed.
262         * Scripts/webkitpy/common/newstringio_unittest.py: Removed.
263         * Scripts/webkitpy/common/system/executive_mock.py: Removed.
264         * Scripts/webkitpy/common/system/filesystem_mock.py: Removed.
265         * Scripts/webkitpy/layout_tests/port/config_mock.py: Removed.
266
267 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
268
269         Unreviewed, build fix.
270
271         Add files inexplicably not committed in r71474 as part of the
272         fix for bug 48280.
273
274         * Scripts/webkitpy/common/newstringio.py: Added.
275         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
276         * Scripts/webkitpy/common/system/executive_mock.py: Added.
277         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
278         * Scripts/webkitpy/layout_tests/port/config_mock.py: Added.
279
280 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
281
282         Reviewed by Eric Siedel.
283
284         new-run-webkit-tests: update port/base and port/webkit to use the
285         new FileSystem and Config abstractions, pulling more logic out of
286         the base Port classes into separate, mockable objects.
287
288         Also create a MockFileSystem object, a MockConfig object, move
289         MockExecutive into common/system to be next to executive, and
290         update the config object to use a FileSystem.
291
292         https://bugs.webkit.org/show_bug.cgi?id=48280
293
294         * Scripts/webkitpy/common/newstringio.py: Added.
295         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
296         * Scripts/webkitpy/common/system/executive_mock.py: Added.
297         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
298         * Scripts/webkitpy/layout_tests/port/base.py:
299         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
300         * Scripts/webkitpy/layout_tests/port/config.py:
301         * Scripts/webkitpy/layout_tests/port/config_mock.py:
302         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
303         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
304         * Scripts/webkitpy/layout_tests/port/webkit.py:
305         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
306         * Scripts/webkitpy/tool/mocktool.py:
307
308 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
309
310         Unreviewed, build breakage.
311
312         Apparently I uploaded the wrong version of the file to fix 49122
313         and neither Eric or I noticed - it was missing a dirname() call.
314         Fixing ...
315
316         https://bugs.webkit.org/show_bug.cgi?id=49122
317
318         * Scripts/webkitpy/common/checkout/scm.py:
319
320 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
321
322         Reviewed by Eric Seidel.
323
324         webkitpy/tool/* unittests change cwd and don't clean up properly
325
326         https://bugs.webkit.org/show_bug.cgi?id=49122
327
328         * Scripts/webkitpy/common/checkout/scm.py:
329
330 2010-11-05  Chris Marrin  <cmarrin@apple.com>
331
332         Reviewed by Simon Fraser.
333
334         suspendAnimations/resumeAnimations not present in WebKit2
335         https://bugs.webkit.org/show_bug.cgi?id=49109
336
337         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
338         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
339         (WTR::LayoutTestController::suspendAnimations):
340         (WTR::LayoutTestController::resumeAnimations):
341         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
342
343 2010-11-05  Tony Chang  <tony@chromium.org>
344
345         Reviewed by David Levin.
346
347         cleanup style in TestNetscapePlugIn/PluginObject.cpp
348         https://bugs.webkit.org/show_bug.cgi?id=49044
349
350         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
351         (getPluginClass):
352         (pluginGetProperty):
353         (pluginSetProperty):
354         (variantToIdentifier):
355         (testCallback):
356         (getURL):
357         (getURLNotify):
358         (testInvokeDefault):
359         (testGetProperty):
360         (testDocumentOpen):
361         (testWindowOpen):
362         (pluginInvoke):
363         (handleCallback):
364         (notifyStream):
365         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
366
367 2010-11-05  Eric Seidel  <eric@webkit.org>
368
369         Reviewed by David Levin.
370
371         Add QueueStatusServer/__init__.py so others can run the QueueStatusServer tests
372         https://bugs.webkit.org/show_bug.cgi?id=49032
373
374         I wrote this file as part of bug 47847, but I forgot to commit it.
375         No one else noticed it missing because test-webkitpy knows how
376         to recover in the case where it can't import QueueStatusServer
377         (which generally occurs due to not having installed the AppEngine SDK).
378
379         * QueueStatusServer/__init__.py: Added.
380         * QueueStatusServer/model/workitems_unittest.py:
381          - Remove a test which fails.  This was probably landed (by me)
382            from my other machine, which since this __init__.py was missing
383            I never noticed the failure and landed this invalid test.
384            Sadly we can't really test remove_work_item as it depends
385            on .key() working.  .key() will throw unless the object
386            has already been saved it seems.
387            This may be a mis-design in our remove_work_item implementation,
388            but for now, just removing the test.
389
390 2010-11-04  Adam Roben  <aroben@apple.com>
391
392         Teach check-webkit-style about TestNetscapePlugIn
393
394         Fixes <http://webkit.org/b/49030> check-webkit-style is ignorant of
395         TestNetscapePlugIn's coding conventions
396
397         Reviewed by John Sullivan.
398
399         * Scripts/webkitpy/style/checker.py: Lump TestNetscapePlugIn in with
400         WebKitAPITest and TestWebKitAPI in having different include order and
401         naming conventions than WebCore. Added some comments about why the
402         exceptions exist.
403
404 2010-11-04  Adam Roben  <aroben@apple.com>
405
406         Add a test that shows that windowed plugins are able to paint
407
408         Somehow we never had a test for this before.
409
410         Fixes <http://webkit.org/b/49029> Should add a test that shows
411         windowed plugins can paint
412
413         Reviewed by Jon Honeycutt.
414
415         * DumpRenderTree/TestNetscapePlugIn/Tests/win/DrawsGradient.cpp: Added.
416         (DrawsGradient::DrawsGradient):
417         (DrawsGradient::wndProc): We handle the WM_PAINT and WM_PRINTCLIENT messages.
418
419         (DrawsGradient::onPaint):
420         (DrawsGradient::onPrintClient):
421         These both just call through to paint.
422
423         (DrawsGradient::paint): Fills our client area with some gradients.
424
425         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Link
426         against Msimg32.lib for ::GradientFill and added DrawsGradient.
427
428         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
429         (createBitmapContextFromWebView): Use WM_PRINT instead of
430         WM_PRINTCLIENT so that ::DefWindowProc will send
431         WM_PRINT/WM_PRINTCLIENT messages to the WebView's child windows.
432         Replaced kCGImageAlphaPremultipledFirst with kCGImageAlphaNoneSkipFirst
433         because GDI doesn't support alpha and kCGImageBitmapOrder32Little with
434         kCGImageBitmapOrder32Host because, who knows, maybe someday Windows
435         will run on a big-endian platform.
436
437 2010-11-04  Adam Roben  <aroben@apple.com>
438
439         Extract much of NPNInvalidateRectInvalidatesWindow's code into a
440         WindowedPluginTest base class
441
442         The base class takes care of subclassing the plugin's window so that a
443         custom WNDPROC is called. This will make it easier to write tests that
444         need to handle window messages.
445
446         Fixes <http://webkit.org/b/49028> It's hard to write a PluginTest with
447         a custom WNDPROC
448
449         Reviewed by Jon Honeycutt.
450
451         * DumpRenderTree/TestNetscapePlugIn/Tests/win/NPNInvalidateRectInvalidatesWindow.cpp:
452         (NPNInvalidateRectInvalidatesWindow::NPNInvalidateRectInvalidatesWindow):
453         (NPNInvalidateRectInvalidatesWindow::NPP_SetWindow):
454         (NPNInvalidateRectInvalidatesWindow::wndProc):
455         (NPNInvalidateRectInvalidatesWindow::testInvalidateRect):
456         Moved code from here to WindowedPluginTest. Changed to use window()
457         instead of m_window.
458
459         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
460         Added WindowedPluginTest and added TestNetscapePlugIn/win to the
461         include path. Also reordered the include path to make a little more
462         sense and simplified the entry that added TestNetscapePlugIn itself to
463         the include path.
464
465         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.cpp: Added.
466         (WindowedPluginTest::WindowedPluginTest):
467         (WindowedPluginTest::NPP_SetWindow):
468         (WindowedPluginTest::staticWndProc):
469         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.h: Added.
470         (WindowedPluginTest::window):
471         Code came from NPNInvalidateRectInvalidatesWindow.
472
473 2010-11-04  Adam Roben  <aroben@apple.com>
474
475         Add a plugin test to show that windowed plugins are clipped correctly
476
477         Fixes <http://webkit.org/b/49024> <rdar://problem/8487847> Windowed
478         plugins aren't clipped in WebKit2 on Windows
479
480         Reviewed by Jon Honeycutt.
481
482         * DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowRegionIsSetToClipRect.cpp: Added.
483         (WindowRegionIsSetToClipRect::WindowRegionIsSetToClipRect): Initialize members.
484         (WindowRegionIsSetToClipRect::NPP_SetWindow): Check that our window
485         region matches the clip rect we know we should have based on
486         window-region-is-set-to-clip-rect.html, and check that our window class
487         doesn't have the CS_PARENTDC style.
488
489         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
490         Added WindowRegionIsSetToClipRect.
491
492 2010-11-05  Alejandro G. Castro  <alex@igalia.com>
493
494         Reviewed by Martin Robinson.
495
496         [GTK] Avoid font hinting in the DRT
497         https://bugs.webkit.org/show_bug.cgi?id=48548
498
499         Change the settings to avoid font hinting, it was causing
500         different results depending on the distribution.
501
502         * DumpRenderTree/gtk/DumpRenderTree.cpp:
503         (initializeGtkFontSettings):
504         * DumpRenderTree/gtk/fonts/fonts.conf:
505
506 2010-11-05  Eric Seidel  <eric@webkit.org>
507
508         Reviewed by Adam Barth.
509
510         build-webkit spams system.log by repeatedly checking for PBXProductDirectory in com.apple.Xcode
511         https://bugs.webkit.org/show_bug.cgi?id=49051
512
513         This is a speculative fix.  The unit tests cover these methods, however
514         I don't know if this will fully stop the system.log spam.
515
516         * Scripts/webkitpy/layout_tests/port/base.py:
517         * Scripts/webkitpy/layout_tests/port/webkit.py:
518
519 2010-11-05  Eric Seidel  <eric@webkit.org>
520
521         Reviewed by Adam Barth.
522
523         Add basic support for showing bot id on /queue-status/ pages
524         https://bugs.webkit.org/show_bug.cgi?id=49037
525
526         This support is really simple.  Eventually we'll want to
527         show the bot id in the lock table too, but we don't have
528         that information stored in the server yet.
529
530         * QueueStatusServer/handlers/queuestatus.py:
531         * QueueStatusServer/handlers/queuestatus_unittest.py: Added.
532         * QueueStatusServer/templates/includes/singlequeuestatus.html:
533
534 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
535
536         Reviewed by Adam Barth.
537
538         Create a filesystem wrapper that we can use to enforce
539         particular conventions and use for mocking and dependency
540         injection down the line.
541
542         https://bugs.webkit.org/show_bug.cgi?id=48144
543
544         * Scripts/webkitpy/common/system/filesystem.py: Added.
545         * Scripts/webkitpy/common/system/filesystem_unittest.py: Added.
546
547 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
548
549         Reviewed by Tony Chang.
550
551         run_webkit_tests_unittest fails under Python 2.5
552         https://bugs.webkit.org/show_bug.cgi?id=49043
553         
554         Switch from itertools.chain.from_iterable (which was added in 2.6)
555         to using itertools.chain directly.
556
557         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
558
559 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
560
561         Unreviewed fix to webkit-patch rebaseline-server so that it runs under
562         Python 2.5 (it needs an import to use the with statement).
563
564         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
565
566 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
567
568         Reviewed by Adam Barth.
569
570         new-run-webkit-tests: split out webkit-specific configuration stuff into a new module
571
572         The current NRWT code has webkit-specific configuration code (like
573         _script_path, default configuration, etc.) mixed in with
574         layout-test-specific stuff in port/base. The configuration code
575         should be split out into a separate module for easier mocking,
576         testing, and isolation.
577
578         https://bugs.webkit.org/show_bug.cgi?id=48264
579
580         * Scripts/webkitpy/layout_tests/port/config.py: Added.
581         * Scripts/webkitpy/layout_tests/port/config_unittest.py: Added.
582
583 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
584
585         Reviewed by Tony Chang.
586
587         Rebaseline server: initial framework
588         https://bugs.webkit.org/show_bug.cgi?id=48892
589         
590         Adds the basic framework for the rebaseline server (details at 
591         http://webkit.org/b/47761). Includes the rebaseline-server webkit-patch
592         command, which starts an HTTP server that can serve static files or
593         invoke handler methods on a class.
594
595         * Scripts/webkitpy/tool/commands/__init__.py:
596         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html: Added.
597         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css: Added.
598         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js: Added.
599         * Scripts/webkitpy/tool/commands/rebaselineserver.py: Added.
600
601 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
602
603         Reviewed by James Robinson.
604
605         new-run-webkit-tests wasn't using DRT by default for
606         --platform chromium-gpu
607
608         The default value was set to False instead of None, which meant
609         that the platform specific logic wasn't firing to change the
610         value to True (b/c we were afraid we'd be overriding the user
611         preference).
612
613         https://bugs.webkit.org/show_bug.cgi?id=49038
614
615         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
616
617 2010-11-04  Tony Chang  <tony@chromium.org>
618
619         Reviewed by Adam Barth.
620
621         make platform/chromium/plugins/multiple-plugins.html pass on all platforms
622         https://bugs.webkit.org/show_bug.cgi?id=49026
623
624         Have the testObject.bar property return the string "bar".  This lets
625         us run and pass platform/chromium/plugins/multiple-plugins.html on all
626         platforms.
627
628         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
629         (testGetProperty):
630
631 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
632
633         Reviewed by Tony Chang.
634
635         NRWT doesn't have good test coverage for --run-chunk, --batch-size, --run-part, etc.
636         https://bugs.webkit.org/show_bug.cgi?id=48878
637         
638         Add get_tests_run so that it's easy to see which tests get run (and with
639         what batching) for a given flag combination. Flesh out the various
640         test cases that have FIXMEs.
641         
642         Also fixes an off-by-one error (batch sizes were one larger than 
643         expected) and makes --run-part also have wraparound behavior, like
644         --run-chunk.
645
646         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
647         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
648         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
649
650 2010-11-04  Erik Arvidsson  <arv@chromium.org>
651
652         Reviewed by Dimitri Glazkov.
653
654         Support box-sizing without the vendor prefix
655         https://bugs.webkit.org/show_bug.cgi?id=36713
656
657         Based on patch by Peter Beverloo <peter@lvp-media.com>
658
659         * iExploder/htdocs/cssproperties.in: Change -webkit-box-sizing to box-sizing.
660
661 2010-11-04  Csaba Osztrogonác  <ossy@webkit.org>
662
663         Unreviewed rollout r71340, because it broke Chromium Windows bot.
664
665         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
666         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
667
668 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
669
670         Reviewed by Ojan Vafai.
671
672         Make http locking default in NRWT.
673         https://bugs.webkit.org/show_bug.cgi?id=48053
674
675         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
676         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
677
678 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
679
680         Reviewed by Eric Seidel.
681
682         [NRWT] Clear invalid http locks on Windows platform as well
683         https://bugs.webkit.org/show_bug.cgi?id=48515
684
685         * Scripts/webkitpy/common/system/executive.py:
686         * Scripts/webkitpy/common/system/executive_unittest.py:
687         * Scripts/webkitpy/layout_tests/port/http_lock.py:
688
689 2010-11-02  Adam Roben  <aroben@apple.com>
690
691         Reduce our dependence on coreutils when running Python tests
692
693         This patch introduces versions of the standard echo and cat utilities
694         implemented in Python. They are probably missing features of their
695         coreutils equivalents, but they can do what's necessary for our Python
696         tests. This is useful on Windows, where these utilities typically
697         aren't available.
698
699         Fixes <http://webkit.org/b/48883> executive_unittest relies on echo
700         and cat utilities from coreutils, which are not present on Windows
701
702         Reviewed by Eric Seidel.
703
704         * Scripts/webkitpy/common/system/executive_unittest.py: Changed to use
705         our Python-based echo and cat.
706
707         * Scripts/webkitpy/common/system/fileutils.py: Added.
708         (make_stdout_binary): On Windows, puts sys.stdout into binary mode so
709         that \n won't be translated into \r\n. I couldn't think of a good way
710         to test this directly without touching the filesystem, but it is tested
711         indirectly by echo_unittest.
712
713         * Scripts/webkitpy/test/cat.py: Added.
714         (command_arguments): Returns a list for invoking cat with the given arguments.
715         (main): Acts like a simplified version of the coreutils cat utility.
716
717         * Scripts/webkitpy/test/cat_unittest.py: Added.
718         (CatTest.assert_cat): Runs cat with the given input and ensures the
719         output matches the input.
720         (CatTest.test_basic): Performs a simple test of cat.
721         (CatTest.test_no_newline): Tests what happens when the input string
722         doesn't have a trailing newline.
723         (CatTest.test_unicode): Tests passing a unicode string to cat.
724         (CatTest.test_as_command): Tests running cat as a separate command.
725
726         * Scripts/webkitpy/test/echo.py: Added.
727         (command_arguments): Returns a list for invoking echo with the given arguments.
728         (main): Acts like a simplified version of the coreutils echo utility.
729
730         * Scripts/webkitpy/test/echo_unittest.py: Added.
731         (EchoTest.test_basic): Performs a simple test of echo.
732         (EchoTest.test_no_newline): Tests passing -n to echo to suppress the
733         trailing newline.
734         (EchoTest.test_unicode): Tests passing unicode and non-unicode strings
735         to echo.
736         (EchoTest.test_argument_order): Tests what happens when -n is not the
737         first argument.
738         (EchoTest.test_empty_arguments): Tests what happens when you pass [] to
739         echo.main.
740         (EchoTest.test_no_arguments): Tests what happens when you call
741         echo.main with no arguments.
742         (EchoTest.test_as_command): Tests running echo as a separate command.
743
744 2010-11-04  Renata Hodovan  <reni@inf.u-szeged.hu>
745
746         Unreviewed: Add myself to the list of Committers.
747
748         * Scripts/webkitpy/common/config/committers.py:
749
750 2010-11-04  Andreas Kling  <kling@webkit.org>
751
752         Reviewed by Kenneth Rohde Christiansen.
753
754         [Qt] Hook into QWebPage::scrollRequested for resizes-to-contents mode
755
756         In RTC mode the QGraphicsWebView item is the size of the contents,
757         scrolling works a bit differently  (we need to react to scrollRequested.)
758
759         Normally QGraphicsView will replay the last mouse event when scrolling,
760         so to prevent WebKit from getting confused by this we temporarily make
761         the QGraphicsView non-interactive.
762
763         * QtTestBrowser/webview.cpp:
764         (WebViewGraphicsBased::setPage):
765         (WebViewGraphicsBased::scrollRequested):
766         * QtTestBrowser/webview.h:
767
768 2010-11-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
769
770         Reviewed by Andreas Kling.
771
772         Use OS(MAC_OS_X) rather than PLATFORM(MAC) in TestControllerQt
773
774         PLATFORM(MAC) is not defined for the Qt port, as it refers to
775         the Mac-port, not the Mac OS X operating system.
776
777         * WebKitTestRunner/qt/TestControllerQt.cpp:
778         (WTR::isExistingLibrary):
779
780 2010-11-04  Leonid Ebril  <leonid.ebril@nokia.com>
781
782         Reviewed by Andreas Kling.
783
784         [Qt] Adding iphone user agent string the user agent list for QtTestBrowser
785         https://bugs.webkit.org/show_bug.cgi?id=48869
786
787         * QtTestBrowser/useragentlist.txt:
788
789 2010-11-03  Adam Roben  <aroben@apple.com>
790
791         Always use uppercase drive names in strings returned by abspath_to_uri
792
793         Some versions of cygpath use lowercase drive letters while others use
794         uppercase, which makes it hard to test the output of code that uses
795         cygpath.
796
797         Fixes <http://webkit.org/b/48914> webkitpy.common.system.path_unittest
798         fails with Cygwin 1.5
799
800         Reviewed by Eric Seidel.
801
802         * Scripts/webkitpy/common/system/path.py:
803         (cygpath): Updated the docstring to indicate that only absolute paths
804         should be passed for now (though relative paths will work fine).
805         (_Cygpath.convert): Upper-case the first letter of the converted Windows path.
806
807 2010-11-03  George Guo  <George.Guo@Nokia.com>
808
809         Reviewed by Andreas Kling.
810
811         [Qt] QtTestBrowser : set mmp rule pageddata in Symbian
812         https://bugs.webkit.org/show_bug.cgi?id=48767
813     
814         Paging is needd on Symbian devices to support benchmarks tests like
815         dromaeo.com and Celtic Kane that need a lot of memory to run  
816
817         * QtTestBrowser/QtTestBrowser.pro:
818
819 2010-11-03  Jenn Braithwaite  <jennb@chromium.org>
820
821         Reviewed by Dmitry Titov.
822
823         Chromium: Update resource tracking when moving a frame between documents
824         https://bugs.webkit.org/show_bug.cgi?id=48363
825
826         * DumpRenderTree/chromium/WebViewHost.cpp:
827         (WebViewHost::assignIdentifierToRequest):
828         Always put resource id in map so we can make assumptions about its
829         presence.
830         (WebViewHost::removeIdentifierForRequest):
831         Added.
832         (WebViewHost::didFinishResourceLoad):
833         (WebViewHost::didFailResourceLoad):
834         Check existence of resource id before removing from map.
835         * DumpRenderTree/chromium/WebViewHost.h:
836
837 2010-11-03  Victor Wang  <victorw@chromium.org>
838
839         Reviewed by Adam Barth.
840
841         [Chromium] update buildbot names in chromium rebaseline tool.
842
843         https://bugs.webkit.org/show_bug.cgi?id=48881
844
845         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
846
847 2010-11-03  Darin Adler  <darin@apple.com>
848
849         Updated Xcode projects by opening them with Xcode 3.2.4.
850         Updated svn:ignore for Xcode projects.
851
852         * MiniBrowser/MiniBrowser.xcodeproj: Added property svn:ignore.
853         * TestWebKitAPI/TestWebKitAPI.xcodeproj: Added property svn:ignore.
854         * WebKitLauncher/WebKitLauncher.xcodeproj: Modified property svn:ignore.
855         * WebKitLauncher/WebKitLauncher.xcodeproj/project.pbxproj: Updated with Xcode 3.2.4.
856
857 2010-11-03  Adam Roben  <aroben@apple.com>
858
859         Stop waiting for messages from the web process after a timeout elapses
860
861         TestController::runUntil now takes a timeout parameter that specifies
862         how long to wait for the condition to become true. WebKitTestRunner
863         takes a --timeout flag that specifies how long the timeout should be.
864         run-webkit-tests passes this flag to WebKitTestRunner so its timeout
865         will be similar to run-webkit-tests.
866
867         Fixes <http://webkit.org/b/43047> <rdar://problem/8365833>
868         WebKitTestRunner waits forever if the web process crashes
869
870         Reviewed by Darin Adler and Anders Carlsson.
871
872         * Scripts/old-run-webkit-tests:
873         (top level): Moved the GuardMalloc timeout adjustment here from
874         readFromDumpToolWithTimer.
875         (openDumpTool): Make WTR use a timeout similar to but slightly shorter
876         than the one that was specified on the command line.
877
878         * WebKitTestRunner/TestController.cpp:
879         (WTR::TestController::TestController): Initialize our timeout values.
880         (WTR::TestController::initialize): Parse the --timeout flag and use it
881         to modify our timeout values.
882         (WTR::TestController::resetStateToConsistentValues): Changed to use a
883         short timeout while waiting for the web process to reset and to return
884         a boolean indicating whether we were able to reset the web process.
885         Uses a 5-second timeout while waiting for the process to be reset.
886         (WTR::TestController::runTest): Changed to return a boolean indicating
887         whether we were able to reset the web process (and thus run the test).
888         (WTR::TestController::runUntil): Call through to platformRunUntil.
889
890         (WTR::TestController::runTestingServerLoop):
891         (WTR::TestController::run):
892         Changed to bail if any test can't be run. This will cause the process
893         to exit. (Unfortunately this will make run-webkit-tests think we
894         crashed; see <http://webkit.org/b/48943>.)
895
896         * WebKitTestRunner/TestController.h: Added platformRunUntil,
897         m_longTimeout, and m_shortTimeout.
898
899         * WebKitTestRunner/TestInvocation.cpp:
900         (WTR::TestInvocation::invoke): Use a short timeout when waiting for the
901         initial response and a long timeout when waiting for the test to
902         complete. Check whether runUntil timed out and print an error message
903         if so.
904
905         * WebKitTestRunner/mac/TestControllerMac.mm:
906         (WTR::TestController::platformRunUntil): Renamed from runUntil. Pass
907         [NSDate distantPast] to -[NSRunLoop runMode:beforeDate:] so that we
908         won't block waiting for the run loop. Only loop until the timeout
909         elapses.
910
911         * WebKitTestRunner/qt/TestControllerQt.cpp:
912         (WTR::TestController::platformRunUntil): Renamed from runUntil. Added a
913         FIXME about honoring the timeout.
914
915         * WebKitTestRunner/win/TestControllerWin.cpp:
916         (WTR::TestController::platformRunUntil): Renamed from runUntil. Use
917         ::MsgWaitForMultipleObjectsEx to implement the timeout. Changed to use
918         ::PeekMessageW so that we don't block waiting for messages to become
919         available.
920
921 2010-11-03  Adam Roben  <aroben@apple.com>
922
923         Add a plugin test that evaluates JS after removing the plugin element
924         from the document
925
926         This test replaces platform/win/plugins/plugin-delayed-destroy.html.
927         That test was made to prevent a crash very similar to this one, but
928         unfortunately tested only the mechanism that prevented the crash and
929         not whether the crash itself was prevented. Since WebKit2 uses a
930         different mechanism to prevent the crash, the test was failing even
931         though WebKit2 was not vulnerable to the crash. This new test crashes
932         if there is no mechanism in place to prevent it and passes in both
933         WebKit1 and WebKit2.
934
935         Fixes <http://webkit.org/b/46711> <rdar://problem/8485903>
936         platform/win/plugins/plugin-delayed-destroy.html fails in WebKit2
937
938         Reviewed by Anders Carlsson.
939
940         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
941         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
942         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
943         * GNUmakefile.am:
944         Added new file.
945
946         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
947         (pluginDeallocate): Make sure we delete the PluginTest object. This
948         prevents a leak and also allows us to test the crash.
949
950         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
951         (PluginTest::executeScript): Made this into a non-static member
952         function.
953
954         (PluginTest::waitUntilDone):
955         (PluginTest::notifyDone):
956         Updated for changes to executeScript.
957
958         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Added executeScript.
959
960         * DumpRenderTree/TestNetscapePlugIn/Tests/EvaluateJSAfterRemovingPluginElement.cpp: Added.
961         (EvaluateJSAfterRemovingPluginElement::EvaluateJSAfterRemovingPluginElement):
962         Initialize ourselves and tell the test harness to wait.
963         (EvaluateJSAfterRemovingPluginElement::NPP_DestroyStream): Remove our
964         plugin element from the document, then execute some JavaScript. If
965         WebKit does not have appropriate mechanisms in place, we'll be
966         destroyed inside the first call to executeScript and crash on the
967         second call.
968
969 2010-11-02  Stephen White  <senorblanco@chromium.org>
970
971         Reviewed by Tony Chang.
972
973         [chromium] Fix LayoutTestController UMRs.
974         https://bugs.webkit.org/show_bug.cgi?id=48872
975
976         * DumpRenderTree/chromium/LayoutTestController.cpp:
977         (LayoutTestController::LayoutTestController):
978
979 2010-11-03  Kent Tamura  <tkent@chromium.org>
980
981         Reviewed by Dimitri Glazkov.
982
983         REGRESSION: rebaseline-chromium-webkit-tests uses non-zero tolerance for
984         image dup detection
985         https://bugs.webkit.org/show_bug.cgi?id=48744
986
987         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
988          - Make a function for option parsing for ease of test
989          - Set 0 to options.tolerance
990         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
991          - Add a test for this change
992
993 2010-11-02  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
994
995         Unreviewed: Add myself to the list of Committers.
996
997         * Scripts/webkitpy/common/config/committers.py:
998
999 2010-11-02  Anders Carlsson  <andersca@apple.com>
1000
1001         Fix build.
1002
1003         * TestWebKitAPI/PlatformUtilities.cpp:
1004         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
1005
1006 2010-11-02  Sam Weinig  <sam@webkit.org>
1007
1008         Reviewed by Anders Carlsson.
1009
1010         Add a way to send startup messages on the context which can be posted when a process launches
1011         <rdar://problem/8617928>
1012         https://bugs.webkit.org/show_bug.cgi?id=48838
1013
1014         * MiniBrowser/mac/WebBundle/WebBundleMain.m:
1015         (WKBundleInitialize):
1016         * TestWebKitAPI/InjectedBundleController.cpp:
1017         (TestWebKitAPI::InjectedBundleController::initialize):
1018         (TestWebKitAPI::InjectedBundleController::didReceiveMessage):
1019         * TestWebKitAPI/InjectedBundleController.h:
1020         * TestWebKitAPI/InjectedBundleMain.cpp:
1021         (WKBundleInitialize):
1022         * TestWebKitAPI/PlatformUtilities.cpp:
1023         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
1024         * WebKitTestRunner/InjectedBundle/InjectedBundleMain.cpp:
1025         (WKBundleInitialize):
1026         Update implementations of WKBundleInitialize to take an initial userData
1027         argument.  Change TestWebKitAPI to use the new initial userData to initialize
1028         each test's bundle.
1029
1030 2010-11-02  Benjamin Kalman  <kalman@google.com>
1031
1032         Reviewed by Ojan Vafai.
1033
1034         new-run-webkit-tests doesn't strip "LayoutTests/" from prefix, unlike old-run-webkit-tests
1035         https://bugs.webkit.org/show_bug.cgi?id=48794
1036
1037         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: Strip the "LayoutTests/" prefix from test argument paths.
1038
1039 2010-11-02  Adam Roben  <aroben@apple.com>
1040
1041         Skip webkitpy.layout_tests.run_webkit_tests_unittest.MainTest on Cygwin
1042         Python 2.5.x
1043
1044         It is known to hang on that version of Python. See
1045         <http://webkit.org/b/48614>.
1046
1047         Reviewed by Adam Barth.
1048
1049         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py: Use
1050         skip_if to skip MainTest on Cygwin Python 2.5.x.
1051
1052         * Scripts/webkitpy/test/skip.py: Added.
1053         (skip_if): If the passed-in condition is false, find all the test_*
1054         methods of the given class and replace them with a function that just
1055         logs that we're skipping these tests. This is loosely based on the
1056         unittest.skip_if decorator added in Python 3.1.
1057         (_skipped_method): Returns a function that just logs that the tests are
1058         being skipped. This is used in place of the actual test_* functions.
1059
1060         * Scripts/webkitpy/test/skip_unittest.py: Added.
1061         (SkipTest.setUp): Create a logger for ourselves and give it a handler
1062         that logs to a private stream.
1063         (SkipTest.tearDown): Reset the logger.
1064         (SkipTest.create_fixture_class): Returns a class that we use to test
1065         skip_if. It has a single test_* method, test_foo, that just calls
1066         through to a callback.
1067         (SkipTest.foo_callback): Record that test_foo was called.
1068         (SkipTest.test_skip_if_false): Pass skip_if a False condition and test
1069         that test_foo does get called.
1070         (SkipTest.test_skip_if_true): Pass skip_if a True condition and test
1071         that test_foo does not get called and the appropriate message gets
1072         logged.
1073
1074 2010-11-02  Adam Barth  <abarth@webkit.org>
1075
1076         Reviewed by Eric Seidel.
1077
1078         webkit-patch should tell check-webkit-style which files were changed so
1079         check-webkit-style doesn't have to stat the whole working copy again
1080         https://bugs.webkit.org/show_bug.cgi?id=48792
1081
1082         * Scripts/webkitpy/tool/mocktool.py:
1083         * Scripts/webkitpy/tool/steps/checkstyle.py:
1084
1085 2010-11-02  Robert Kroeger  <rjkroege@chromium.org>
1086
1087         Reviewed by James Robinson.
1088
1089         [Chromium/DRT] Make EventSendingController honour leapForward for touch events.
1090         https://bugs.webkit.org/show_bug.cgi?id=48777
1091
1092         * DumpRenderTree/chromium/EventSender.cpp:
1093         (EventSender::sendCurrentTouchEvent):
1094
1095 2010-11-02  Adam Roben  <aroben@apple.com>
1096
1097         Only track resource identifiers in DRT when dumpResourceLoadCallbacks
1098         is on
1099
1100         This reverts Windows to our pre-r71097 behavior. That patch made us
1101         track all resource identifiers, including the main resource, so the
1102         main resource's URL started appearing in test output instead of
1103         "<unknown>". Arguably having the main resource's URL is better, but all
1104         other platforms print "<unknown>" and we want to match.
1105
1106         Fixes <http://webkit.org/b/48837> <rdar://problem/8620351> REGRESSION
1107         (r71097): Many http tests failing on Windows
1108
1109         Reviewed by Anders Carlsson.
1110
1111         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
1112         (ResourceLoadDelegate::identifierForInitialRequest): Don't add the
1113         identifier to the URL map if we're not supposed to dump resource load
1114         callbacks.
1115         (ResourceLoadDelegate::removeIdentifierForRequest): Always remove the
1116         identifier from the URL map even if we're already "done". There's no
1117         point in keeping out-of-date identifiers around.
1118
1119 2010-11-01  Jenn Braithwaite  <jennb@chromium.org>
1120
1121         Reviewed by Adam Roben.
1122
1123         Windows: Update resource tracking when moving a frame between documents
1124         https://bugs.webkit.org/show_bug.cgi?id=48364
1125
1126         * DumpRenderTree/win/DumpRenderTree.cpp:
1127         (createWebViewAndOffscreenWindow):
1128         (main):
1129         Give each WebView its own ResourceLoadDelegate instance in order to
1130         make assertions about resource ids on a particular WebView.
1131         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
1132         (ResourceLoadDelegate::identifierForInitialRequest):
1133         Always add id to the map.
1134         (ResourceLoadDelegate::removeIdentifierForRequest):
1135         Added.
1136         (ResourceLoadDelegate::willSendRequest):
1137         (ResourceLoadDelegate::didReceiveAuthenticationChallenge):
1138         (ResourceLoadDelegate::didReceiveResponse):
1139         (ResourceLoadDelegate::didFinishLoadingFromDataSource):
1140         (ResourceLoadDelegate::didFailLoadingWithError):
1141         (ResourceLoadDelegate::descriptionSuitableForTestResult):
1142         Replace static descriptionSuitableForTestResult with static member function to access identifier map.
1143         * DumpRenderTree/win/ResourceLoadDelegate.h:
1144         (ResourceLoadDelegate::urlMap):
1145         Moved within class so that each WebView has its own id map.
1146
1147 2010-11-01  Ojan Vafai  <ojan@chromium.org>
1148
1149         Reviewed by Dimitri Glazkov.
1150
1151         remove debug code from run_webkit_tests.py
1152         https://bugs.webkit.org/show_bug.cgi?id=48800
1153
1154         Remove temporary debug code and make --master-name required
1155         if --test-results-server is set now that all clients set
1156         --master-name.
1157
1158         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1159
1160 2010-11-01  Tony Chang  <tony@chromium.org>
1161
1162         Reviewed by Kent Tamura.
1163
1164         add plugin event logging for linux
1165         https://bugs.webkit.org/show_bug.cgi?id=48779
1166
1167         This is taken from Chromium's fork of the layout test plugin:
1168         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
1169
1170         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
1171         (handleEventX11):
1172         (NPP_HandleEvent):
1173
1174 2010-11-01  Dirk Pranke  <dpranke@chromium.org>
1175
1176         Reviewed by James Robinson.
1177
1178         new-run-webkit-tests: use DRT, child-processes=1 for GPU tests by default
1179         https://bugs.webkit.org/show_bug.cgi?id=48790
1180
1181         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
1182         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
1183
1184 2010-11-01  Adam Barth  <abarth@webkit.org>
1185
1186         Reviewed by Eric Seidel.
1187
1188         Teach check-webkit-style how to accept a list of files to diff on the
1189         command line
1190         https://bugs.webkit.org/show_bug.cgi?id=48784
1191
1192         In a future patch, webkit-patch will use this option to improve
1193         performance.  I'm landing this in two pieces to avoid causing a version
1194         skew problem for the style-bot.
1195
1196         * Scripts/check-webkit-style:
1197         * Scripts/webkitpy/style/optparser.py:
1198         * Scripts/webkitpy/style/optparser_unittest.py:
1199         * Scripts/webkitpy/style_references.py:
1200
1201 2010-11-01  Anders Carlsson  <andersca@apple.com>
1202
1203         Reviewed by John Sullivan.
1204
1205         Tear down the related WebProcessProxy when a WebContext is deallocated
1206         https://bugs.webkit.org/show_bug.cgi?id=48769
1207
1208         * TestWebKitAPI/Tests/WebKit2/FailedLoad.cpp:
1209         (TestWebKitAPI::didFailProvisionalLoadWithErrorForFrame):
1210         We don't support empty URLs anymore, update test to expect a null URL instead.
1211
1212 2010-11-01  Søren Gjesse  <sgjesse@chromium.org>
1213
1214         Reviewed by Andreas Kling.
1215
1216         Fix warning when compiling the chromium port of DumpRenderShell
1217         with clang.
1218         https://bugs.webkit.org/show_bug.cgi?id=48414
1219
1220         * DumpRenderTree/chromium/TestShell.h:
1221         (TestShell::javaScriptFlagsForLoad):
1222
1223 2010-11-01  Adam Roben  <aroben@apple.com>
1224
1225         Fix typo from r71022
1226
1227         * BuildSlaveSupport/build.webkit.org-config/config.json:
1228
1229 2010-11-01  Adam Roben  <aroben@apple.com>
1230
1231         Trigger the Windows Release WebKit2 tests when a Release build
1232         finishes, not when a Debug build finishes
1233
1234         Fixes <http://webkit.org/b/48754> Windows Release WebKit2 tests are
1235         triggered at the wrong time
1236
1237         Reviewed by Sam Weinig.
1238
1239         * BuildSlaveSupport/build.webkit.org-config/config.json: Fixed
1240         triggerable name and trigger.
1241
1242 2010-11-01  Mario Sanchez Prada  <msanchez@igalia.com>
1243
1244         Unreviewed. Adding my IRC nickname to the list of committers.
1245
1246         * Scripts/webkitpy/common/config/committers.py:
1247
1248 2010-10-31  Robert Hogan  <robert@webkit.org>
1249
1250         Reviewed by Antonio Gomes.
1251
1252         [Qt] [Gtk] Plug-ins having upper case in mime type are failing to load
1253
1254         Qt and Gtk are case-sensitive when storing the declared mime-type
1255         of plugins. Since plugin mime-types are lowercased prior to searching
1256         for them in the plugin database, ensure they are loaded with the
1257         mime-type in lower case too.
1258
1259         Change the test netscape plugin to declare its mimetype in sentence
1260         case so that the correct behaviour is enforced.
1261
1262         https://bugs.webkit.org/show_bug.cgi?id=36815
1263
1264         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
1265         (NP_GetMIMEDescription):
1266
1267 2010-10-31  Robert Hogan  <robert@webkit.org>
1268
1269         Reviewed by Antonio Gomes.
1270
1271         [Qt] Support nodesFromRect in DRT
1272
1273         https://bugs.webkit.org/show_bug.cgi?id=48716
1274
1275         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1276         (LayoutTestController::nodesFromRect):
1277         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1278
1279 2010-10-30  Andreas Kling  <kling@webkit.org>
1280
1281         Reviewed by Antonio Gomes.
1282
1283         [Qt] QtTestBrowser: Switching view type moves the embedded inspector
1284         https://bugs.webkit.org/show_bug.cgi?id=48705
1285
1286         Reinsert the embedded inspector into the splitter after changing
1287         between QWebView/QGraphicsWebView.
1288
1289         * QtTestBrowser/launcherwindow.cpp:
1290         (LauncherWindow::init):
1291         (LauncherWindow::initializeView):
1292         * QtTestBrowser/webinspector.h:
1293         (WebInspector::WebInspector):
1294
1295 2010-10-28  Antonio Gomes  <agomes@rim.com>
1296
1297         Reviewed by Ojan Vafai.
1298
1299         Needs a "LinuxEditingBehavior", perhaps with a better name
1300         https://bugs.webkit.org/show_bug.cgi?id=36627
1301
1302         Adding support to Mac's, GTK+'s, Windows' and Chromium's LayoutTestController class to test the newly introduced Unix editing behavior.
1303
1304         * DumpRenderTree/chromium/LayoutTestController.cpp:
1305         (LayoutTestController::setEditingBehavior):
1306         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1307         (LayoutTestController::setEditingBehavior):
1308         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1309         (LayoutTestController::setEditingBehavior):
1310         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1311         (LayoutTestController::setEditingBehavior):
1312
1313 2010-10-29  Andreas Kling  <kling@webkit.org>
1314
1315         Reviewed by Kenneth Rohde Christiansen.
1316
1317         [Qt] QtTestBrowser: Add keyboard shortcut to toggle full screen (F11)
1318         https://bugs.webkit.org/show_bug.cgi?id=48695
1319
1320         * QtTestBrowser/launcherwindow.cpp:
1321         (LauncherWindow::createChrome):
1322
1323 2010-10-29  Andreas Kling  <kling@webkit.org>
1324
1325         Reviewed by Kenneth Rohde Christiansen.
1326
1327         [Qt] QtTestBrowser: Fix uninitialized read in FpsTimer
1328         https://bugs.webkit.org/show_bug.cgi?id=48675
1329
1330         FpsTimer::m_timer was never initialized and passed to QObject::killTimer()
1331         on startup with in -graphicsbased mode.
1332
1333         * QtTestBrowser/fpstimer.cpp:
1334         (FpsTimer::FpsTimer):
1335         (FpsTimer::numFrames):
1336         (FpsTimer::stop):
1337         (FpsTimer::timerEvent):
1338         * QtTestBrowser/fpstimer.h:
1339
1340 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
1341
1342         Reviewed by Tony Chang.
1343
1344         new-run-webkit-tests: change TestResults to be serializable
1345
1346         In preparation for changing new-run-webkit-tests from
1347         multithreaded to multiprocess, we need to make sure the data
1348         going between the threads is easily serialized over a socket.
1349
1350         This change adds serialization/pickling for the TestResults and
1351         TestFailure objects (using cPickle).
1352
1353         The TestFailure objects included a "has_wdiff" flag for Text
1354         results, but the flag wasn't being used, so I've removed it,
1355         simplifying the state to basically a set of enum objects with
1356         associated methods.
1357
1358         https://bugs.webkit.org/show_bug.cgi?id=48616
1359
1360         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1361         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
1362         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
1363         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
1364         * Scripts/webkitpy/layout_tests/layout_package/test_results.py: Added.
1365         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:Added. 
1366         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1367         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
1368
1369 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
1370
1371         Reviewed by Ojan Vafai.
1372
1373         new-run-webkit-tests: need to provide separate dashboard results for GPU tests
1374         https://bugs.webkit.org/show_bug.cgi?id=48687
1375
1376         Modify the steps to upload the JSON files to the dashboards so
1377         that the GPU tests don't conflict with the regular tests on a
1378         port. We do this by modifying the --builder-name parameter to
1379         append " - GPU", which should cause the app to treat the results
1380         as a completely new builder. This is a little non-obvious, but
1381         keeps us from having to restructure the app.
1382
1383         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
1384         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
1385         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1386
1387 2010-10-29  Ojan Vafai  <ojan@chromium.org>
1388
1389         Reviewed by Tony Chang.
1390
1391         [chromium] add debug logging to help diagnose flakiness dashboard issues
1392         https://bugs.webkit.org/show_bug.cgi?id=48657
1393
1394         The appengine app thinks it's getting empty files uploaded. Add some logging
1395         to see if new-run-webkit-tests agrees.
1396         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1397
1398 2010-10-29  Dimitri Glazkov  <dglazkov@chromium.org>
1399
1400         Reviewed by Ojan Vafai.
1401
1402         [Chromium/DRT] Add master-name flag to new-run-webkit-tests.
1403         https://bugs.webkit.org/show_bug.cgi?id=48649
1404
1405         * BuildSlaveSupport/build.webkit.org-config/master.cfg: Added flag.
1406
1407 2010-10-29  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
1408
1409         Reviewed by Martin Robinson.
1410
1411         Enable popup window in GtkLauncher
1412         https://bugs.webkit.org/show_bug.cgi?id=48335
1413
1414         The GtkLauncher application does not support opening new window when
1415         a link with "target=_blank" is clicked or similar call to
1416         window.open(). Instead, GtkLauncher does nothing which breaks
1417         navigation of some websites.
1418
1419         * GtkLauncher/main.c:
1420         (activate_uri_entry_cb):
1421         (update_title):
1422         (link_hover_cb):
1423         (notify_title_cb):
1424         (notify_load_status_cb):
1425         (notify_progress_cb):
1426         (destroy_cb):
1427         (go_back_cb):
1428         (go_forward_cb):
1429         (create_web_view_cb):
1430         (web_view_ready_cb):
1431         (close_web_view_cb):
1432         (create_browser):
1433         (create_statusbar):
1434         (create_toolbar):
1435         (create_window):
1436         (main):
1437
1438 2010-10-29  Adam Roben  <aroben@apple.com>
1439
1440         Teach check-webkit-style about WebKit2's idiosyncracies
1441
1442         Fixes <http://webkit.org/b/48638> Style bot complains about a number
1443         of WebKit2 conventions
1444
1445         Reviewed by Anders Carlsson.
1446
1447         * Scripts/webkitpy/style/checker.py: Excluded some rules for various
1448         WebKit2-related files. Also updated the excluded rules for
1449         WebKitAPITest to match the current code.
1450
1451 2010-10-29  Csaba Osztrogonác  <ossy@webkit.org>
1452
1453         Reviewed by Adam Roben and David Kilzer.
1454
1455         Fix and cleanup of build systems
1456         https://bugs.webkit.org/show_bug.cgi?id=48342
1457
1458         * Scripts/build-webkit:
1459          - Remove unnecessary ENABLE_SANDBOX option.
1460          - Add ENABLE_FULLSCREEN_API option.
1461
1462 2010-10-28  Adam Roben  <aroben@apple.com>
1463
1464         Switch the Windows WebKit2 bot to the Release configuration
1465
1466         We only have one machine testing WebKit2 on Windows right now, and
1467         Debug is just too slow for it to keep up.
1468
1469         Fixes (hopefully!) <http://webkit.org/b/48615> Windows WebKit2 bot is
1470         always way behind
1471
1472         Reviewed by Jon Honeycutt.
1473
1474         * BuildSlaveSupport/build.webkit.org-config/config.json:
1475
1476 2010-10-29  Leandro Gracia Gil  <leandrogracia@google.com>
1477
1478         Reviewed by Jeremy Orlow.
1479
1480         Added a second parameter to setMockSpeechInputResult for
1481         the language used in speech input.
1482         https://bugs.webkit.org/show_bug.cgi?id=47089
1483
1484         * DumpRenderTree/LayoutTestController.cpp:
1485         (setMockSpeechInputResultCallback):
1486         * DumpRenderTree/LayoutTestController.h:
1487         * DumpRenderTree/chromium/LayoutTestController.cpp:
1488         (LayoutTestController::setMockSpeechInputResult):
1489         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1490         (LayoutTestController::setMockSpeechInputResult):
1491         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1492         (LayoutTestController::setMockSpeechInputResult):
1493         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1494         (LayoutTestController::setMockSpeechInputResult):
1495         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1496         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1497         (LayoutTestController::setMockSpeechInputResult):
1498         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
1499         (LayoutTestController::setMockSpeechInputResult):
1500         (LayoutTestController::pageSizeAndMarginsInPixels):
1501
1502 2010-10-28  Ojan Vafai  <ojan@chromium.org>
1503
1504         Reviewed by Tony Chang.
1505
1506         [chromium] add a result-small.json file for the test dashboard
1507         https://bugs.webkit.org/show_bug.cgi?id=48547
1508
1509         Output both a results.json file and a results-small.json file.
1510         The dashboard will load results-small.json by default so it loads faster.
1511
1512         * TestResultServer/model/jsonresults.py:
1513         * TestResultServer/model/jsonresults_unittest.py:
1514         Added a bunch of sys.path hackery. Unforunately, this uses hardcoded
1515         paths. That obviously needs to be fixed, but at least this way it
1516         clearly documents what paths are necessary.
1517
1518 2010-10-28  Eric Seidel  <eric@webkit.org>
1519
1520         Reviewed by Adam Barth.
1521
1522         webkit-patch upload calls changed_files more often than it should
1523         https://bugs.webkit.org/show_bug.cgi?id=48567
1524
1525         Passing changed_files around everywhere isn't a very elegant solution
1526         but it's the one we have for the moment.  I think keeping an explicit
1527         cache on Checkout (or making StepState() a real class) is a better
1528         long-term option.
1529
1530         Previously bug_id_for_this_commit was calling changed_files and the
1531         result was never getting cached on the state.  Now we're explicitly
1532         caching the result on the state and passing that to the bug_id_for_this_commit call.
1533
1534         I looked into building unit tests for this.  Doing so would require
1535         using a real Checkout object with a MockSCM and overriding the appropriate
1536         calls on SCM to count how often we're stating the file system.
1537         That's a useful set of tests to build for a separate change.
1538
1539         * Scripts/webkitpy/common/checkout/api.py:
1540         * Scripts/webkitpy/tool/commands/download.py:
1541         * Scripts/webkitpy/tool/commands/upload.py:
1542         * Scripts/webkitpy/tool/mocktool.py:
1543
1544 2010-10-28  Eric Seidel  <eric@webkit.org>
1545
1546         Reviewed by Adam Barth.
1547
1548         Make suggest-reviewers slightly faster
1549         https://bugs.webkit.org/show_bug.cgi?id=48562
1550
1551         Add @memoized to one more common call.
1552
1553         * Scripts/webkitpy/common/checkout/api.py:
1554
1555 2010-10-28  Eric Seidel  <eric@webkit.org>
1556
1557         Reviewed by Adam Barth.
1558
1559         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is still very slow
1560         https://bugs.webkit.org/show_bug.cgi?id=48536
1561
1562         This patch makes it about 40 seconds faster, but it still
1563         takes 1:40.  This will require more refinement.
1564
1565         The suggested reviewers list appears to be the same.
1566
1567         I think the next step may be to have it stop the search after
1568         5 reviewers are found.  We never want to suggest 30 people.
1569
1570         * Scripts/webkitpy/common/checkout/scm.py:
1571          - Using --remove-empty to theoretically stop lookups past
1572            when a file is removed.  I'm not entirely clear that the option
1573            does what it says it does.  Example:
1574            git log --pretty=format:%H -5 --remove-empty -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
1575            returns only one commit
1576            vs.
1577            git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
1578            which returns 5.  I was not aware that wx files were ever removed from the repository?
1579
1580 2010-10-28  Kinuko Yasuda  <kinuko@chromium.org>
1581
1582         Reviewed by David Levin.
1583
1584         [Chromium] Support FileSystem in chromium DRT
1585         https://bugs.webkit.org/show_bug.cgi?id=47643
1586
1587         * DumpRenderTree/chromium/WebViewHost.cpp:
1588         (WebViewHost::openFileSystem): Added.
1589         * DumpRenderTree/chromium/WebViewHost.h:
1590         (WebViewHost::openFileSystem): Added.
1591
1592 2010-10-28  Sam Weinig  <sam@webkit.org>
1593
1594         Reviewed by Anders Carlsson.
1595
1596         WKURLRefs should be allowed to be null
1597         <rdar://problem/8575621>
1598         https://bugs.webkit.org/show_bug.cgi?id=48535
1599
1600         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
1601         * TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp:
1602         (TestWebKitAPI::didStartProvisionalLoadForFrame):
1603         (TestWebKitAPI::didCommitLoadForFrame):
1604         (TestWebKitAPI::didFinishLoadForFrame):
1605         Test that URLs are null pointers when unset.
1606
1607 2010-10-28  Chris Fleizach  <cfleizach@apple.com>
1608
1609         Reviewed by Adele Peterson.
1610
1611         AX: multi select group option does not handle setting of AXSelectedChildren correctly
1612         https://bugs.webkit.org/show_bug.cgi?id=48464
1613
1614         Add support for querying information about selected children to DRT, including:
1615            selectedChildAtIndex
1616            selectedChildrenCount
1617            setSelectedChild
1618
1619         * DumpRenderTree/AccessibilityUIElement.cpp:
1620         (selectedChildAtIndexCallback):
1621         (setSelectedChildCallback):
1622         (selectedChildrenCountCallback):
1623         (AccessibilityUIElement::setSelectedChild):
1624         (AccessibilityUIElement::selectedChildrenCount):
1625         (AccessibilityUIElement::selectedChildAtIndex):
1626         (AccessibilityUIElement::getJSClass):
1627         * DumpRenderTree/AccessibilityUIElement.h:
1628         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
1629         (AccessibilityUIElement::selectedChildAtIndex):
1630         (AccessibilityUIElement::selectedChildrenCount):
1631         (AccessibilityUIElement::setSelectedChild):
1632
1633 2010-10-27  Sam Weinig  <sam@webkit.org>
1634
1635         Reviewed by Anders Carlsson.
1636
1637         Add WebKit2 API for window feature getter/setters
1638         <rdar://problem/8590373>
1639         https://bugs.webkit.org/show_bug.cgi?id=48496
1640
1641         * MiniBrowser/mac/BrowserWindowController.m:
1642         (-[BrowserWindowController awakeFromNib]):
1643         * MiniBrowser/win/BrowserView.cpp:
1644         (BrowserView::create):
1645         * WebKitTestRunner/TestController.cpp:
1646         (WTR::createOtherPage):
1647         (WTR::TestController::initialize):
1648
1649 2010-10-28  Søren Gjesse  <sgjesse@chromium.org>
1650
1651         Reviewed by Tony Chang.
1652
1653         Added support for the DumpRenderTree flags --multiple-loads and --js-flags to the Python test runner.
1654         https://bugs.webkit.org/show_bug.cgi?id=48236
1655
1656         * Scripts/webkitpy/layout_tests/port/chromium.py:
1657         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1658
1659 2010-10-28  Sergio Villar Senin  <svillar@igalia.com>
1660
1661         Reviewed by Martin Robinson.
1662
1663         [GTK] WebKitWebFrame's load-status is not properly notified to the tests
1664         https://bugs.webkit.org/show_bug.cgi?id=48048
1665
1666         DRT now listens to WebKitWebFrame load-status signals for each
1667         created frame instead of the load-status signal of the
1668         WebKitWebView that only notifies about changes in the main frame.
1669
1670         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1671         (webViewLoadFinished):
1672         (webFrameLoadStatusNotified):
1673         (frameCreatedCallback):
1674         (createWebView):
1675         (main):
1676
1677 2010-10-28  Tony Chang  <tony@chromium.org>
1678
1679         Reviewed by Kent Tamura.
1680
1681         [chromium] fix textInputController.{selectedRange,markedRange}
1682         https://bugs.webkit.org/show_bug.cgi?id=48487
1683
1684         * DumpRenderTree/chromium/TextInputController.cpp:
1685         (TextInputController::markedRange): Return arrays of ints, rather than a string
1686         (TextInputController::selectedRange): Ditto.
1687
1688 2010-10-27  Ojan Vafai  <ojan@chromium.org>
1689
1690         Reviewed by Tony Chang.
1691
1692         [chromium] add a master-name flag to new-run-webkit-tests
1693         https://bugs.webkit.org/show_bug.cgi?id=48488
1694
1695         The test results server now allows adding a master name to
1696         the uploaded files. This lets us distinguish bots that have
1697         the same name, but are on different masters.
1698
1699         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1700
1701 2010-10-28  Adam Roben  <aroben@apple.com>
1702
1703         Don't append a newline to the test output if the frame has no document
1704         element in WebKitTestRunner
1705
1706         Fixes <http://webkit.org/b/48526> Extra trailing newline when running
1707         plugins/document-open.html in WebKitTestRunner
1708
1709         Reviewed by Anders Carlsson.
1710
1711         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1712         (WTR::toJS): Added. Turns a UTF-8 C string into a JSStringRef.
1713         (WTR::hasDocumentElement): Added. Uses the JSC API to figure out
1714         whether the frame has a document element.
1715         (WTR::dumpFrameText): Match DRT by bailing (rather than appending an
1716         empty string and a newline) if the frame has no document element.
1717
1718 2010-10-28  Adam Roben  <aroben@apple.com>
1719
1720         Skip npn-invalidate-rect-invalidates-window.html on headless XP
1721         machines
1722
1723         TestNetscapePlugIn never receives a WM_PAINT message on headless XP
1724         machines, so this test times out. Fixing the test is covered by
1725         <http://webkit.org/b/48333>.
1726
1727         * Scripts/old-run-webkit-tests: Skip
1728         npn-invalidate-rect-invalidates-window.html on Windows if accelerated
1729         compositing support is disabled, which likely means we're on a headless
1730         XP machine.
1731
1732 2010-10-28  Kimmo Kinnunen  <kimmok@iki.fi>
1733
1734         Adding myself as a committer.
1735
1736         * Scripts/webkitpy/common/config/committers.py:
1737
1738 2010-10-27  Eric Seidel  <eric@webkit.org>
1739
1740         Reviewed by Adam Barth.
1741
1742         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is very slow
1743         https://bugs.webkit.org/show_bug.cgi?id=48500
1744
1745         This doesn't fix the problem, but it makes things slightly better.
1746         Each git svn find-rev call takes about .25 seconds on my desktop.
1747         This patch uses a new memoized class to avoid those calls when possible.
1748
1749         The real slowness is still git log on some files, like:
1750         git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
1751         I'm not yet sure how to make the pathological git logs better.
1752
1753         * Scripts/webkitpy/common/checkout/scm.py:
1754         * Scripts/webkitpy/common/memoized.py: Added.
1755         * Scripts/webkitpy/common/memoized_unittest.py: Added.
1756
1757 2010-10-27  Eric Seidel  <eric@webkit.org>
1758
1759         Reviewed by Ojan Vafai.
1760
1761         EWS bots should not use --quiet when running build-webkit
1762         https://bugs.webkit.org/show_bug.cgi?id=48482
1763
1764         --quiet is only correct when the sub-process does the error reporting.
1765         In the case of _can_build() the parent process is reporting the error.
1766         We'd like the full build log at queues.webkit.org so someone can look
1767         at the log and understand why the EWS is failing to build trunk.
1768
1769         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1770         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
1771         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
1772         * Scripts/webkitpy/tool/mocktool.py:
1773
1774 2010-10-27  Brian Weinstein  <bweinstein@apple.com>
1775
1776         More Windows build fixage. Rename a variable that was named string.
1777
1778         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1779         (WTR::hasPrefix):
1780
1781 2010-10-27  Ojan Vafai  <ojan@chromium.org>
1782
1783         Reviewed by Dimitri Glazkov.
1784
1785         [chromium] Make the test results server store which master the bot is on
1786         https://bugs.webkit.org/show_bug.cgi?id=48478
1787
1788         The chromium bots recently changed so that there are multiple slaves with
1789         the same name on different masters. Up till now, the test results server
1790         assumed slave names were unique. Adds a master field to the file in order
1791         to distinguish.
1792
1793         Also, for files that currently lack a master or testtype, set them appropriately.
1794
1795         * TestResultServer/handlers/testfilehandler.py:
1796         * TestResultServer/index.yaml:
1797         * TestResultServer/model/jsonresults.py:
1798         * TestResultServer/model/testfile.py:
1799         * TestResultServer/templates/showfilelist.html:
1800         * TestResultServer/templates/uploadform.html:
1801
1802 2010-10-26  Darin Adler  <darin@apple.com>
1803
1804         Reviewed by Sam Weinig.
1805
1806         WebKitTestRunner needs to support layoutTestController.dumpBackForwardList
1807         https://bugs.webkit.org/show_bug.cgi?id=42322
1808         rdar://problem/8193631
1809
1810         WebKitTestRunner needs to support layoutTestController.clearBackForwardList
1811         https://bugs.webkit.org/show_bug.cgi?id=42333
1812         rdar://problem/8193643
1813
1814         * WebKitTestRunner/Configurations/InjectedBundle.xcconfig:
1815         Renamed the product to WebKitTestRunnerInjectedBundle to avoid
1816         name conflicts in the build directory.
1817
1818         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
1819         Added dumpBackForwardList and clearBackForwardList.
1820
1821         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
1822         (WTR::InjectedBundle::InjectedBundle): Removed initialization of m_mainPage.
1823         (WTR::InjectedBundle::didCreatePage): Removed unneeded underscore.
1824         (WTR::InjectedBundle::willDestroyPage): Ditto.
1825         (WTR::InjectedBundle::didReceiveMessage): Ditto.
1826         (WTR::InjectedBundle::initialize): Ditto.
1827         (WTR::InjectedBundle::didCreatePage): Changed code to use m_pages
1828         instead of m_mainPage and m_otherPages.
1829         (WTR::InjectedBundle::willDestroyPage): Ditto.
1830         (WTR::InjectedBundle::page): Ditto.
1831         (WTR::InjectedBundle::beginTesting): Ditto.
1832         (WTR::InjectedBundle::done): Ditto.
1833         (WTR::InjectedBundle::closeOtherPages): Ditto.
1834         (WTR::InjectedBundle::dumpBackForwardListsForAllPages): Added.
1835
1836         * WebKitTestRunner/InjectedBundle/InjectedBundle.h: Changed
1837         page function to longer be inline, and pageCount function
1838         to use m_pages. Added dumpBackForwardListsForAllPages, and
1839         removed some underscores. Replaced m_mainPage and m_otherPageas
1840         with m_pages.
1841
1842         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1843         (WTR::adoptWK): Added.
1844         (WTR::hasPrefix): Added.
1845         (WTR::InjectedBundlePage::reset): Added code to set up
1846         m_previousTestBackForwardListItem.
1847         (WTR::InjectedBundlePage::dump): Added code to call
1848         dumpBackForwardListsForAllPages.
1849         (WTR::compareByTargetName): Added.
1850         (WTR::dumpBackForwardListItem): Added.
1851         (WTR::InjectedBundlePage::dumpBackForwardList): Added.
1852
1853         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h: Added
1854         dumpBackForwardList and m_previousTestBackForwardListItem.
1855
1856         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
1857         (WTR::LayoutTestController::LayoutTestController): Initialize
1858         m_shouldDumpBackForwardListsForAllWindows to false.
1859         (WTR::LayoutTestController::clearBackForwardList): Added.
1860
1861         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
1862         Added dumpBackForwardList, clearBackForwardList,
1863         shouldDumpBackForwardListsForAllWindows, and
1864         m_shouldDumpBackForwardListsForAllWindows.
1865
1866         * WebKitTestRunner/TestController.cpp:
1867         (WTR::TestController::initialize): Set up
1868         didReceiveSynchronousMessageFromInjectedBundle.
1869         (WTR::TestController::didReceiveSynchronousMessageFromInjectedBundle):
1870         Added.
1871         * WebKitTestRunner/TestController.h: Ditto.
1872
1873         * WebKitTestRunner/TestInvocation.cpp:
1874         (WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):
1875         Added.
1876         * WebKitTestRunner/TestInvocation.h: Ditto.
1877
1878         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
1879         Renamed bundle to WebKitTestRunnerInjectedBundle (see above).
1880         * WebKitTestRunner/mac/TestControllerMac.mm:
1881         (WTR::TestController::initializeInjectedBundlePath): Ditto.
1882
1883 2010-10-27  Chris Rogers  <crogers@google.com>
1884
1885         Reviewed by Chris Marrin.
1886
1887         Add ENABLE_WEB_AUDIO feature enable flag (initially disabled) to build-webkit
1888         https://bugs.webkit.org/show_bug.cgi?id=48279
1889
1890         * Scripts/build-webkit:
1891
1892 2010-10-27  Eric Seidel  <eric@webkit.org>
1893
1894         Unreviewed.
1895
1896         build-webkit should collect Visual Studio Express logs and display them
1897         https://bugs.webkit.org/show_bug.cgi?id=39199
1898
1899         It turns out my previous patch wasn't actually working on the win-ews
1900         machine.  So I've fixed my mistakes from before.
1901
1902         * Scripts/build-webkit:
1903          - Windows VSE builds change the CWD while building.  Why?  Who knows.
1904         * Scripts/print-vse-failure-logs:
1905          - windows VSE builds don't use Debug/Release as I expected, so I've
1906            moved off of --configuration to --top-level and added the /obj
1907            optimization while I was there.
1908
1909 2010-10-27  Dimitri Glazkov  <dglazkov@chromium.org>
1910
1911         Unreviewed, rolling out r70674.
1912         http://trac.webkit.org/changeset/70674
1913         https://bugs.webkit.org/show_bug.cgi?id=48053
1914
1915         Broke Chromium Windows build.
1916
1917         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1918         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1919
1920 2010-10-25  Tony Chang  <tony@chromium.org>
1921
1922         Reviewed by Anders Carlsson.
1923
1924         compile TestNetscapePlugIn on chromium linux
1925         https://bugs.webkit.org/show_bug.cgi?id=48274
1926
1927         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
1928         (NP_Initialize): On Linux, plugin funcs are set in initialize.
1929         (NPP_New): Mark the plugin as windowless.
1930         (NPP_GetValue): Handle mime type values.
1931         (NP_GetMIMEDescription):
1932         (NP_GetValue):
1933         * DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npapi.h: define TRUE and FALSE, which are in webkit's npapi.h.
1934         * DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npfunctions.h: Pull in npapi.h to get TRUE/FALSE (matches webkit's npfunctions.h)
1935
1936 2010-10-27  Kenneth Rohde Christiansen  <kenneth@webkit.org>
1937
1938         Reviewed by Anders Carlsson.
1939
1940         Remove contentSizeChanged callbacks as it is no longer
1941         part of the public UIClient.
1942
1943         Make WKPageContentsSizeChangedCallback be a private API
1944         https://bugs.webkit.org/show_bug.cgi?id=48409
1945
1946         * MiniBrowser/mac/BrowserWindowController.m:
1947         (-[BrowserWindowController awakeFromNib]):
1948         * MiniBrowser/win/BrowserView.cpp:
1949         (BrowserView::create):
1950         * WebKitTestRunner/TestController.cpp:
1951         (WTR::createOtherPage):
1952         (WTR::TestController::initialize):
1953
1954 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1955
1956         Reviewed by Ojan Vafai.
1957
1958         Make http locking default in NRWT.
1959         https://bugs.webkit.org/show_bug.cgi?id=48053
1960
1961         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1962         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1963
1964 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1965
1966         Reviewed by Tony Chang.
1967
1968         [NRWT] Fix http lock on Windows platform
1969         https://bugs.webkit.org/show_bug.cgi?id=48321
1970
1971         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1972
1973 2010-10-27  Satish Sampath  <satish@chromium.org>
1974
1975         Unreviewed, rolling out r70665.
1976         http://trac.webkit.org/changeset/70665
1977         https://bugs.webkit.org/show_bug.cgi?id=47089
1978
1979         Need to address Alexey's review comments.
1980
1981         * DumpRenderTree/LayoutTestController.cpp:
1982         (setMockSpeechInputResultCallback):
1983         * DumpRenderTree/LayoutTestController.h:
1984         * DumpRenderTree/chromium/LayoutTestController.cpp:
1985         (LayoutTestController::setMockSpeechInputResult):
1986         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1987         (LayoutTestController::setMockSpeechInputResult):
1988         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1989         (LayoutTestController::setMockSpeechInputResult):
1990         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1991         (LayoutTestController::setMockSpeechInputResult):
1992         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1993         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1994         (LayoutTestController::setMockSpeechInputResult):
1995         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
1996         (LayoutTestController::setMockSpeechInputResult):
1997
1998 2010-10-27  Leandro Gracia Gil  <leandrogracia@google.com>
1999
2000         Reviewed by Jeremy Orlow.
2001
2002         Added a second parameter to setMockSpeechInputResult for
2003         the language used in speech input.
2004         https://bugs.webkit.org/show_bug.cgi?id=47089
2005
2006         * DumpRenderTree/LayoutTestController.cpp:
2007         (setMockSpeechInputResultCallback):
2008         * DumpRenderTree/LayoutTestController.h:
2009         * DumpRenderTree/chromium/LayoutTestController.cpp:
2010         (LayoutTestController::setMockSpeechInputResult):
2011         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2012         (LayoutTestController::setMockSpeechInputResult):
2013         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2014         (LayoutTestController::setMockSpeechInputResult):
2015         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2016         (LayoutTestController::setMockSpeechInputResult):
2017         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2018         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2019         (LayoutTestController::setMockSpeechInputResult):
2020         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2021         (LayoutTestController::setMockSpeechInputResult):
2022         (LayoutTestController::pageSizeAndMarginsInPixels):
2023
2024 2010-10-27  Adam Roben  <aroben@apple.com>
2025
2026         Reset TestNetscapePlugIn's NPP_GetValue pointer when
2027         NullNPPGetValuePointer finishes running
2028
2029         Fixes <http://webkit.org/b/48435> REGRESSION (r70655): Many plugins
2030         tests are failing on Qt
2031
2032         Reviewed by Anders Carlsson.
2033
2034         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
2035         (PluginTest::NPP_Destroy):
2036         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
2037         Added a do-nothing NPP_Destroy implementation.
2038
2039         * DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp:
2040         (NullNPPGetValuePointer::NullNPPGetValuePointer): Save the original
2041         NPP_GetValue pointer in m_originalNPPGetValuePointer so we can restore
2042         it later.
2043         (NullNPPGetValuePointer::NPP_Destroy): Added. Restores the original
2044         NPP_GetValue pointer so it can be used in other tests.
2045
2046         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
2047         (NPP_Destroy):
2048         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
2049         (webkit_test_plugin_destroy_instance):
2050         Call through to the PluginTest when NPP_Destroy is called.
2051
2052 2010-10-26  Brian Weinstein  <bweinstein@apple.com>
2053
2054         Reviewed by Adam Roben.
2055
2056         WebKit2 shouldn't try to send an empty user agent
2057         https://bugs.webkit.org/show_bug.cgi?id=48397
2058         
2059         Add a test that when we set our custom user agent to an empty string, we don't send
2060         an empty user agent.
2061
2062         * TestWebKitAPI/Tests/WebKit2/PreventEmptyUserAgent.cpp: Added.
2063         (TestWebKitAPI::didRunJavaScript): Make sure that the result of navigator.userAgent isn't empty.
2064         (TestWebKitAPI::TEST): Set our custom user agent to the empty string, and run navigator.userAgent.
2065         
2066         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Add the new file.
2067         * TestWebKitAPI/win/TestWebKitAPI.vcproj: Ditto.
2068
2069 2010-10-27  Adam Roben  <aroben@apple.com>
2070
2071         Check in file I forgot in r70653
2072
2073         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
2074         (NP_Initialize): Save the NPPluginFuncs struct the browser passed to us
2075         so we can be naughty and modify it later.
2076
2077 2010-10-27  Adam Roben  <aroben@apple.com>
2078
2079         Test that WebKit doesn't crash if the plugin passes 0 for its
2080         NPP_GetValue pointer
2081
2082         Test for <http://webkit.org/b/48433> Crash in
2083         NetscapePlugin::shouldLoadSrcURL when using Shockwave Director 10.3 in
2084         WebKit2 on Windows
2085
2086         Reviewed by Eric Carlson.
2087
2088         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
2089         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2090         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
2091         * GNUmakefile.am:
2092         Added NullNPPGetValuePointer.cpp.
2093
2094         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
2095         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
2096         Set up a place to store the NPPluginFuncs struct the browser passed to us.
2097
2098         * DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp: Added.
2099         (NullNPPGetValuePointer::NullNPPGetValuePointer): Null out the
2100         NPP_GetValue pointer we passed to the browser to simulate a plugin that
2101         doesn't implement NPP_GetValue.
2102         (NullNPPGetValuePointer::NPP_GetValue): Print an error message. If this
2103         function is called, it means that WebKit has changed in a way that
2104         makes this test invalid.
2105
2106         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
2107         (NP_GetEntryPoints): Save the NPPluginFuncs struct the browser passed
2108         to us so we can be naughty and modify it later.
2109
2110 2010-10-27  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
2111
2112         Reviewed by Andreas Kling.
2113
2114         Remove references to ancient QGVLauncher and QtLauncher
2115         https://bugs.webkit.org/show_bug.cgi?id=48430
2116
2117         QtTestBrowser substitutes both and has checks enabled.
2118
2119         * Scripts/webkitpy/style/checker.py: remove references
2120         * Scripts/webkitpy/style/checker_unittest.py: ditto
2121
2122 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2123
2124         Reviewed by Ojan Vafai.
2125
2126         [NRWT] Don't use image hash when it's no need in single test mode.
2127         https://bugs.webkit.org/show_bug.cgi?id=48326
2128
2129         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2130
2131 2010-10-27  David Kilzer  <ddkilzer@apple.com>
2132
2133         Fix leak of CFMutableDictionaryRef in createXMLStringFromWebArchiveData()
2134
2135         Reviewed by Adam Roben.
2136
2137         Follow-up fix for: <https://bugs.webkit.org/show_bug.cgi?id=48278>
2138
2139         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
2140         (createXMLStringFromWebArchiveData): Use RetainPtr<> to fix a
2141         leak introduced in r70613.
2142
2143 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2144
2145         Reviewed by Csaba Osztrogonác.
2146
2147         [NRWT] Add platform specific baseline search paths for Qt port
2148         https://bugs.webkit.org/show_bug.cgi?id=48428
2149
2150         * Scripts/webkitpy/layout_tests/port/qt.py:
2151
2152 2010-10-27  Adam Roben  <aroben@apple.com>
2153
2154         Catch exceptions when checking if we're inside a git working directory
2155
2156         Fixes <http://webkit.org/b/48420> REGRESSION (r70562): test-webkitpy
2157         fails on systems without git installed
2158
2159         Reviewed by Anders Carlsson.
2160
2161         * Scripts/webkitpy/common/net/credentials.py:
2162         (Credentials._credentials_from_git): Put the call to
2163         Git.in_working_directory inside the try/except since it, too, attempts
2164         to execute git and thus will throw on systems that don't have git
2165         installed.
2166
2167 2010-10-27  Nikolas Zimmermann  <nzimmermann@rim.com>
2168
2169         Reviewed by David Kilzer.
2170
2171         Convert DumpRenderTree webarchive code to CoreFoundation
2172         https://bugs.webkit.org/show_bug.cgi?id=48278
2173
2174         CFPropertyListCreateWithData and CFPropertyListCreateData are only available in 10.6+.
2175         Replace CFPropertyListCreateWithData by a combination of CFReadStreamCreateWithBytesNoCopy and CFPropertyListCreateFromStream.
2176         Replace CFPropertyListCreateData by CFPropertyListCreateXMLData.
2177
2178         These changes are wrapped in BUILDING_ON_LEOPARD, as the methods are deprecated on 10.6+.
2179
2180         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
2181         (createXMLStringFromWebArchiveData):
2182
2183 2010-10-26  Antonio Gomes  <agomes@rim.com>
2184
2185         Reviewed by Martin Robinson.
2186
2187         Remove the absolute path used to include DumpRenderTreeSupportGtk.h from LayoutTestController.h
2188
2189         It turns out this is rather unneeded since WebKitTools/GNUMakefile.am has WebKit/gtk/ in its include
2190         path, and then we can just do #include "WebCoreSupport/DumpRenderTreeSupportGtk.h"
2191
2192         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2193
2194 2010-10-26  David Kilzer  <ddkilzer@apple.com>
2195
2196         Rename WebArchiveDumpSupport.mm to WebArchiveDumpSupport.cpp
2197
2198         Reviewed by Adam Roben.
2199
2200         Part 4 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
2201
2202         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
2203         Updated for file moves and renames.
2204         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.mm.
2205         * DumpRenderTree/cf/WebArchiveDumpSupport.h: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.h.
2206
2207 2010-10-26  David Kilzer  <ddkilzer@apple.com>
2208
2209         Convert WebArchiveDumpSupport.mm from NS objects to CF types
2210
2211         Reviewed by Adam Roben.
2212
2213         Part 3 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
2214
2215         In order to share WebArchive code between the Mac and Windows
2216         ports, the code in WebArchiveDumpSupport.mm was converted from
2217         Cocoa to C++ using CoreFoundation (CF) types.
2218
2219         * DumpRenderTree/mac/DumpRenderTree.mm:
2220         (dump): Renamed serializeWebArchiveToXML() to
2221         createXMLStringFromWebArchiveData() and added HardAutorelease()
2222         to prevent leaks.
2223         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
2224         (createXMLStringFromWebArchiveData): Renamed from
2225         serializeWebArchiveToXML().  Changed to use CF types.
2226         (createCFURLResponseFromResponseData): Renamed from
2227         unarchiveNSURLResponseFromResponseData().  Changed to use CF
2228         types for its parameter and return type.
2229         * DumpRenderTree/mac/WebArchiveDumpSupport.mm: Replaced use of
2230         NS objects with CF types.  It will be renamed to *.cpp in a
2231         future commit.
2232         (convertMIMEType): Changed to use case-insensitive string
2233         comparisons.
2234         (convertWebResourceDataToString):
2235         (normalizeHTTPResponseHeaderFields):
2236         (normalizeWebResourceURL):
2237         (convertWebResourceResponseToDictionary):
2238         (compareResourceURLs):
2239         (createXMLStringFromWebArchiveData):
2240         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
2241         (createCFURLResponseFromResponseData): Renamed from
2242         unarchiveNSURLResponseFromResponseData().  Changed to use CF
2243         types for its parameter and return type.
2244
2245 2010-10-26  David Kilzer  <ddkilzer@apple.com>
2246
2247         Extract use of NSKeyedUnarchiver from WebArchiveDumpSupport.mm
2248
2249         Reviewed by Adam Roben.
2250
2251         Part 2 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
2252
2253         There is no equivalent to NSKeyedUnarchiver in CoreFoundation,
2254         so extract it into a platform-specific source file.
2255
2256         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
2257         (unarchiveNSURLResponseFromResponseData): Added declaration.
2258         * DumpRenderTree/mac/WebArchiveDumpSupport.mm:
2259         (convertWebResourceResponseToDictionary): Extracted code to
2260         unarchiveNSURLResponseFromResponseData() in
2261         WebArchiveDumpSupportMac.mm.  Updated logic to return early if
2262         nil is returned from unarchiveNSURLResponseFromResponseData().
2263         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
2264         (unarchiveNSURLResponseFromResponseData): Added.  Extracted
2265         code from convertWebResourceResponseToDictionary() in
2266         WebArchiveDumpSupport.mm.
2267
2268 2010-10-26  David Kilzer  <ddkilzer@apple.com>
2269
2270         Extract call to -[WebHTMLRepresentation supportedNonImageMIMETypes] from WebArchiveDumpSupport.mm
2271
2272         Reviewed by Adam Roben.
2273
2274         Part 1 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
2275
2276         The call to -[WebHTMLRepresentation supportedNonImageMIMETypes]
2277         is not cross-platform between Mac and Windows, so extract it
2278         into a platform-specific source file.
2279
2280         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Added
2281         WebArchiveDumpSupportMac.mm to the project.
2282         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
2283         (supportedNonImageMIMETypes): Added declaration.
2284         * DumpRenderTree/mac/WebArchiveDumpSupport.mm:
2285         (convertWebResourceDataToString): Replaced call to
2286         -[WebHTMLRepresentation supportedNonImageMIMETypes] with
2287         supportedNonImageMIMETypes().
2288         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm: Added.
2289         (supportedNonImageMIMETypes): Added.  Extracted from
2290         WebArchiveDumpSupport.mm.
2291
2292 2010-10-26  Antonio Gomes  <agomes@rim.com>
2293
2294         Reviewed by Martin Robinson.
2295
2296         [GTK] Implement DumpRenderTreeSupportGtk (similarly to DumpRenderTreeSupportQt idea)
2297         https://bugs.webkit.org/show_bug.cgi?id=48199
2298
2299         Implements support to WebKitTabToLinksPreferenceKey through LayoutTestController::overridePreference.
2300         The corresponding DumpRenderTreeSupportGtk method is called in the DRT context only.
2301
2302         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2303         (resetDefaultsToConsistentValues):
2304         (createWebView):
2305         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2306         (LayoutTestController::overridePreference):
2307
2308 2010-10-26  Ariya Hidayat  <ariya@sencha.com>
2309
2310         Unreviewed, change the order of my emails for bugzilla autocompletion.
2311
2312         * Scripts/webkitpy/common/config/committers.py:
2313
2314 2010-10-26  Eric Seidel  <eric@webkit.org>
2315
2316         Reviewed by David Kilzer.
2317
2318         build-webkit should collect Visual Studio Express logs and display them
2319         https://bugs.webkit.org/show_bug.cgi?id=39199
2320
2321         * Scripts/build-webkit:
2322         * Scripts/print-vse-failure-logs: Added.
2323         * Scripts/webkitdirs.pm:
2324
2325 2010-10-26  David Kilzer  <ddkilzer@apple.com>
2326
2327         <http://webkit.org/b/48224> build-webkit: add support for --meter-tag switch
2328
2329         Reviewed by Csaba Osztrogonác.
2330
2331         * Scripts/build-webkit: Added support for --meter-tag switch.
2332         It should have been added with r60820.
2333
2334 2010-10-26  Ademar de Souza Reis Jr.  <ademar.reis@openbossa.org>
2335
2336         Reviewed by Dumitru Daniliuc.
2337
2338         check-webkit-style fails on operator+=, operator-=, ... methods
2339         https://bugs.webkit.org/show_bug.cgi?id=48258
2340
2341         * Scripts/webkitpy/style/checkers/cpp.py: Added exceptions
2342         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added unit-tests
2343
2344 2010-10-26  Ojan Vafai  <ojan@chromium.org>
2345
2346         Reviewed by Tony Chang.
2347
2348         remove DEFER support from new-run-webkit-tests
2349         https://bugs.webkit.org/show_bug.cgi?id=48387
2350
2351         DEFER was needed when we were trying to ship Chrome beta.
2352         Now it's just extra complication.
2353
2354         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
2355         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
2356         * Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
2357         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2358         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2359         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2360
2361 2010-10-26  Eric Seidel  <eric@webkit.org>
2362
2363         Unreviewed.  Adding import with_statement to fix python 2.5.
2364
2365         Teach webkit-patch how to read credentials from the environment
2366         https://bugs.webkit.org/show_bug.cgi?id=48275
2367
2368         * Scripts/webkitpy/common/net/credentials_unittest.py:
2369          - import with_statement to fix python 2.5.
2370
2371 2010-10-26  Alexey Proskuryakov  <ap@apple.com>
2372
2373         Reviewed by Sam Weinig.
2374
2375         https://bugs.webkit.org/show_bug.cgi?id=48375
2376         <rdar://problem/8392724> Need delegate calls in PageLoaderClient to indicate if we have
2377         loaded insecure content
2378
2379         Updated for WebKit2 changes.
2380
2381         * MiniBrowser/mac/BrowserWindowController.m:
2382         (didDisplayInsecureContentForFrame):
2383         (didRunInsecureContentForFrame):
2384         (-[BrowserWindowController awakeFromNib]):
2385         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2386         (WTR::InjectedBundlePage::InjectedBundlePage):
2387         (WTR::InjectedBundlePage::didDisplayInsecureContentForFrame):
2388         (WTR::InjectedBundlePage::didRunInsecureContentForFrame):
2389         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
2390         * WebKitTestRunner/TestController.cpp:
2391         (WTR::TestController::initialize):
2392
2393 2010-10-26  Adam Roben  <aroben@apple.com>
2394
2395         Clear up confusion between "3D rendering" and "accelerated compositing"
2396
2397         WebKit has no software-based 3D rendering implementation, so 3D
2398         rendering can only be enabled when accelerated compositing is. But DRT
2399         was falsely reporting that 3D rendering was available on machines that
2400         don't support accelerated compositing, leading to much confusion.
2401
2402         Reviewed by Darin Adler.
2403
2404         Fixes <http://webkit.org/b/48370> REGRESSION (r70540): Many
2405         transforms/3d tests are failing on the XP bots
2406
2407         * DumpRenderTree/win/DumpRenderTree.cpp:
2408         (main): When ENABLE(3D_RENDERING) is turned on, only report that 3D
2409         rendering is available when acclerated compositing is also available.
2410
2411         * Scripts/old-run-webkit-tests: Skip tests that use the
2412         -webkit-transform-3d media query when 3D rendering is disabled,
2413         rather than when accelerated compositing is disabled.
2414
2415 2010-10-26  Eric Seidel  <eric@webkit.org>
2416
2417         Reviewed by Ojan Vafai.
2418
2419         Remove rietveld code now that it's unused
2420         https://bugs.webkit.org/show_bug.cgi?id=48359
2421
2422         Was cool that we added this support, but now that it's
2423         not used it makes little sense to keep it around.
2424         We can always restore this code from SVN if we need it.
2425
2426         * Scripts/webkitpy/common/config/__init__.py:
2427         * Scripts/webkitpy/common/net/bugzilla.py:
2428         * Scripts/webkitpy/common/net/bugzilla_unittest.py:
2429         * Scripts/webkitpy/common/net/rietveld.py: Removed.
2430         * Scripts/webkitpy/common/net/rietveld_unittest.py: Removed.
2431         * Scripts/webkitpy/thirdparty/__init__.py:
2432         * Scripts/webkitpy/tool/commands/download.py:
2433         * Scripts/webkitpy/tool/commands/download_unittest.py:
2434         * Scripts/webkitpy/tool/commands/queues.py:
2435         * Scripts/webkitpy/tool/commands/queues_unittest.py:
2436         * Scripts/webkitpy/tool/main.py:
2437         * Scripts/webkitpy/tool/mocktool.py:
2438         * Scripts/webkitpy/tool/steps/__init__.py:
2439         * Scripts/webkitpy/tool/steps/postcodereview.py: Removed.
2440
2441 2010-10-26  Anders Carlsson  <andersca@apple.com>
2442
2443         Reviewed by Sam Weinig.
2444
2445         pageDidScroll callback should be on the UI process client rather than (or in addition to) the web process client
2446         https://bugs.webkit.org/show_bug.cgi?id=48366
2447         <rdar://problem/8595202>
2448
2449         * MiniBrowser/mac/BrowserWindowController.m:
2450         (-[BrowserWindowController awakeFromNib]):
2451         * WebKitTestRunner/TestController.cpp:
2452         (WTR::createOtherPage):
2453         (WTR::TestController::initialize):
2454
2455 2010-10-26  Eric Seidel  <eric@webkit.org>
2456
2457         Reviewed by Tony Chang.
2458
2459         Teach webkit-patch how to read credentials from the environment
2460         https://bugs.webkit.org/show_bug.cgi?id=48275
2461
2462         This makes it possible for svn users to have their bugzilla credentials
2463         stored in their environment instead of typing them every time.
2464
2465         We need this for making it easy to run the win-ews bot (which currently
2466         uses svn instead of git).
2467
2468         * Scripts/webkitpy/common/net/credentials.py:
2469         * Scripts/webkitpy/common/net/credentials_unittest.py:
2470
2471 2010-10-26  Kenichi Ishibashi  <bashi@google.com>
2472
2473         Reviewed by Kent Tamura.
2474
2475         Input Method inserts conversion candidates unexpectedly
2476         https://bugs.webkit.org/show_bug.cgi?id=46868
2477
2478         Adds setComposition() to TextInputController to make DRT emulate
2479         an input method behavior.
2480
2481         * DumpRenderTree/chromium/TextInputController.cpp:
2482         (TextInputController::TextInputController):
2483         (TextInputController::setComposition): Added.
2484         * DumpRenderTree/chromium/TextInputController.h:
2485
2486 2010-10-26  Simon Fraser  <simon.fraser@apple.com>
2487
2488         Reviewed by Sam Weinig.
2489
2490         Support layoutTestController.layerTreeAsText in WebKitTestRunner
2491         https://bugs.webkit.org/show_bug.cgi?id=42145
2492         
2493         Implement layerTreeAsText() in WebKitTestRunner.
2494
2495         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
2496         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
2497         (WTR::LayoutTestController::layerTreeAsText):
2498         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
2499
2500 2010-10-26  Adam Barth  <abarth@webkit.org>
2501
2502         Reviewed by Eric Seidel.
2503
2504         webkit-patch rollout produces incorrect patch when using svn move
2505         https://bugs.webkit.org/show_bug.cgi?id=48244
2506
2507         We need to flush our caches when we modify the working copy.
2508
2509         * Scripts/webkitpy/tool/steps/revertrevision.py:
2510
2511 2010-10-26  Eric Seidel  <eric@webkit.org>
2512
2513         Reviewed by Adam Barth.
2514
2515         commit-queue flaky test message can list the same author more than once
2516         https://bugs.webkit.org/show_bug.cgi?id=48268
2517
2518         tonikitoo reported to me over IRC this morning that he's seen
2519         the commit-queue report flaky tests with author lists like:
2520         "adam, adam and adam", suggesting we're not uniquing authors
2521         before writing the message.
2522
2523         I fixed the uniquing and added a bunch more unit testing.
2524
2525         * Scripts/webkitpy/tool/commands/queues.py:
2526         * Scripts/webkitpy/tool/commands/queues_unittest.py:
2527         * Scripts/webkitpy/tool/mocktool.py:
2528
2529 2010-10-26  Adam Roben  <aroben@apple.com>
2530
2531         Pull in the FeatureDefines*.vsprops files when building DRT
2532
2533         This ensures that various ENABLE() macros will get set correctly.
2534         (Prior to r70320 we were picking up the ENABLE(3D_RENDERING) flag
2535         through wtf/Platform.h.)
2536
2537         Fixes <http://webkit.org/b/48343> REGRESSION (r70320): DumpRenderTree
2538         says 3D_RENDERING is disabled even when it is enabled
2539
2540         Reviewed by Ada Chan.
2541
2542         * DumpRenderTree/win/DumpRenderTree.vcproj: Added
2543         FeaturesDefines*.vsprops to all configurations.
2544
2545         * DumpRenderTree/win/DumpRenderTreeWin.h: Touched to force a rebuild.
2546
2547 2010-10-26  Søren Gjesse  <sgjesse@chromium.org>
2548
2549         Reviewed by Dimitri Glazkov.
2550
2551         Added options --multiple-loads and --js-flags to chromium DumpRenderTree. The option --multiple-loads=X
2552         is used to have DumpRenderTree load each test it runs X times. To be able to have more fine-grained control
2553         of how the JavaScript engine behaves for each load the flag --js-flags can specify a list of flag-sets like this
2554
2555           --js-flags="--xxx,--noxxx --yyy,--noyyy"
2556
2557         First load will run with --xxx, the second with --yyy and the third without any (the 'no' prefix is handled by
2558         V8 to turn off the flag).
2559
2560         The changes to the Python test runner will be in a separate change.
2561
2562
2563         * DumpRenderTree/chromium/DumpRenderTree.cpp:
2564         (runTest):
2565         (main):
2566         * DumpRenderTree/chromium/TestShell.cpp:
2567         (TestShell::TestShell):
2568         (TestShell::runFileTest):
2569         (TestShell::testFinished):
2570         * DumpRenderTree/chromium/TestShell.h:
2571         (TestShell::loadCount):
2572         (TestShell::setLoadCount):
2573         (TestShell::javaScriptFlagsForLoad):
2574         (TestShell::setJavaScriptFlags):
2575         (TestShell::setDumpWhenFinished):
2576
2577 2010-10-26  Adam Roben  <aroben@apple.com>
2578
2579         Skip more tests that depend on accelerated compositing when accelerated
2580         compositing is disabled
2581
2582         Fixes <http://webkit.org/b/48329> Some tests fail when accelerated
2583         compositing is disabled
2584
2585         Reviewed by John Sullivan.
2586
2587         * Scripts/old-run-webkit-tests: Added more tests to skip when
2588         accelerated compositing is disable. The tests all have output that
2589         changes depending on whether accelerated compositing is enabled.
2590
2591 2010-10-25  Eric Seidel  <eric@webkit.org>
2592
2593         Reviewed by Adam Barth.
2594
2595         webkit-patch should clean up svn locks when passed --force-clean
2596         https://bugs.webkit.org/show_bug.cgi?id=48269
2597
2598         * Scripts/webkitpy/common/checkout/scm.py:
2599
2600 2010-10-25  Anders Carlsson  <andersca@apple.com>
2601
2602         Reviewed by Adam Roben.
2603
2604         Add a pageDidScroll BundleUIClient callback
2605         https://bugs.webkit.org/show_bug.cgi?id=48260
2606         <rdar://problem/8531159>
2607
2608         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2609         (WTR::InjectedBundlePage::InjectedBundlePage):
2610
2611 2010-10-25  Johnny Ding  <jnd@chromium.org>
2612
2613         Reviewed by Tony Chang.
2614
2615         Dump the gesture status of frame in frame load callbacks in DumpRenderTree
2616         by adding a new method dumpUserGestureInFrameLoadCallbacks.
2617         Now only dump the gesture status in "DidStartProvisionalLoad" callback.
2618         https://bugs.webkit.org/show_bug.cgi?id=47849
2619
2620         * DumpRenderTree/LayoutTestController.cpp:
2621         (LayoutTestController::LayoutTestController):
2622         (dumpUserGestureInFrameLoadCallbacksCallback):
2623         (LayoutTestController::staticFunctions):
2624         * DumpRenderTree/LayoutTestController.h:
2625         (LayoutTestController::dumpUserGestureInFrameLoadCallbacks):
2626         (LayoutTestController::setDumpUserGestureInFrameLoadCallbacks):
2627         * DumpRenderTree/chromium/LayoutTestController.cpp:
2628         (LayoutTestController::LayoutTestController):
2629         (LayoutTestController::dumpUserGestureInFrameLoadCallbacks):
2630         (LayoutTestController::reset):
2631         * DumpRenderTree/chromium/LayoutTestController.h:
2632         (LayoutTestController::shouldDumpUserGestureInFrameLoadCallbacks):
2633         (LayoutTestController::setShouldDumpUserGestureInFrameLoadCallbacks):
2634         * DumpRenderTree/chromium/TestShell.h:
2635         (TestShell::shouldDumpUserGestureInFrameLoadCallbacks):
2636         * DumpRenderTree/chromium/WebViewHost.cpp:
2637         (WebViewHost::didStartProvisionalLoad):
2638         (WebViewHost::printFrameUserGestureStatus):
2639         * DumpRenderTree/chromium/WebViewHost.h:
2640         * DumpRenderTree/mac/FrameLoadDelegate.mm:
2641         (-[WebFrame _drt_printFrameUserGestureStatus]):
2642         (-[FrameLoadDelegate webView:didStartProvisionalLoadForFrame:]):
2643         (-[FrameLoadDelegate webView:didCommitLoadForFrame:]):
2644         (-[FrameLoadDelegate webView:didFailProvisionalLoadWithError:forFrame:]):
2645         (-[FrameLoadDelegate webView:didFinishLoadForFrame:]):
2646         (-[FrameLoadDelegate webView:didFailLoadWithError:forFrame:]):
2647         (-[FrameLoadDelegate webView:windowScriptObjectAvailable:]):
2648         (-[FrameLoadDelegate webView:didReceiveTitle:forFrame:]):
2649         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2650         (LayoutTestController::reset):
2651         (LayoutTestController::dumpUserGestureInFrameLoadCallbacks):
2652         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2653
2654 2010-10-25  Sheriff Bot  <webkit.review.bot@gmail.com>
2655
2656         Unreviewed, rolling out r70442.
2657         http://trac.webkit.org/changeset/70442
2658         https://bugs.webkit.org/show_bug.cgi?id=48248
2659
2660         http locking doesn't work on Windows (Requested by Ossy on
2661         #webkit).
2662
2663         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2664         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2665
2666 2010-10-25  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2667
2668         Reviewed by Ojan Vafai.
2669
2670         Make http locking default in NRWT.
2671         https://bugs.webkit.org/show_bug.cgi?id=48053
2672
2673         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2674         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2675
2676 2010-10-25  David Kilzer  <ddkilzer@apple.com>
2677
2678         <http://webkit.org/b/48185> build-webkit: add support for --progress-tag switch
2679         (Re-landing r70413 after Qt minimal buildfix: r70440.)
2680
2681         Reviewed by Kenneth Rohde Christiansen.
2682
2683         * Scripts/build-webkit: Added support for --progress-tag switch.
2684         It should have been added with r57051.
2685
2686 2010-10-24  Adam Barth  <abarth@webkit.org>
2687
2688         Reviewed by David Kilzer.
2689
2690         webkit-patch upload fails when the patch removes a file
2691         https://bugs.webkit.org/show_bug.cgi?id=48187
2692
2693         We need to use "--" to separate file names from the rest of the
2694         command.
2695
2696         * Scripts/webkitpy/common/checkout/scm.py:
2697
2698 2010-10-24  Sheriff Bot  <webkit.review.bot@gmail.com>
2699
2700         Unreviewed, rolling out r70413.
2701         http://trac.webkit.org/changeset/70413
2702         https://bugs.webkit.org/show_bug.cgi?id=48210
2703
2704         It broke Qt minimal build (Requested by Ossy on #webkit).
2705
2706         * Scripts/build-webkit:
2707
2708 2010-10-24  Yi Shen  <yi.4.shen@nokia.com>
2709
2710         Reviewed by Andreas Kling.
2711
2712         [Qt][QtTestBrowser] Toggle use of QGraphicsView messes up the menu
2713         https://bugs.webkit.org/show_bug.cgi?id=48141
2714
2715         * QtTestBrowser/launcherwindow.cpp:
2716         (LauncherWindow::init):
2717         (LauncherWindow::initializeView):
2718         (LauncherWindow::toggleWebView):
2719
2720 2010-10-24  Daniel Bates  <dbates@rim.com>
2721
2722         Reviewed by David Kilzer.
2723
2724         Fix Perl uninitialized warnings in VCSUtils::svnStatus()
2725         and VCSUtils::removeEOL().
2726         https://bugs.webkit.org/show_bug.cgi?id=48196
2727
2728         VCSUtils::svnStatus() concatenates the output of svn status with
2729         a new line character and svn status may return no output (say for
2730         a file that has not been added, deleted, or modified). We should
2731         only concatenate the output of svn status if there is some.
2732
2733         Also, VCSUtils::removeEOL() should ensure that its argument
2734         is initialized before performing a string substitution.
2735
2736         * Scripts/VCSUtils.pm:
2737           - Modified removeEOL() to return "" if its argument is undefined.
2738           - Exported removeEOL() so that it can be tested. 
2739         * Scripts/webkitperl/VCSUtils_unittest/removeEOL.pl: Added.
2740
2741 2010-10-24  David Kilzer  <ddkilzer@apple.com>
2742
2743         <http://webkit.org/b/48185> build-webkit: add support for --progress-tag switch
2744
2745         Reviewed by Kenneth Rohde Christiansen.
2746
2747         * Scripts/build-webkit: Added support for --progress-tag switch.
2748         It should have been added with r57051.
2749
2750 2010-10-24  David Kilzer  <ddkilzer@apple.com>
2751
2752         <http://webkit.org/b/48184> build-webkit: reformat support variables for better maintainability
2753
2754         Reviewed by Daniel Bates.
2755
2756         * Scripts/build-webkit: Alphabetized the @features array based
2757         on the name of the feature.  Reformatted the list of support
2758         variables so that they match the order of @features, and so that
2759         each has its own line.  This makes it easy to add new variables
2760         in the correct order.
2761
2762 2010-10-23  Eric Seidel  <eric@webkit.org>
2763
2764         Reviewed by Adam Barth.
2765
2766         EWS never removes invalid patch ids
2767         https://bugs.webkit.org/show_bug.cgi?id=48173
2768
2769         This is just sticking another finger in the dam.
2770         However this adds more unit testing which will help
2771         us make sure we're always releasing patches once we
2772         redesign the release_patch API and call these from
2773         a more central place.
2774
2775         * Scripts/webkitpy/tool/commands/queues.py:
2776         * Scripts/webkitpy/tool/commands/queues_unittest.py:
2777         * Scripts/webkitpy/tool/mocktool.py:
2778          - Added the ability to request invalid patches.
2779            Log a warning message to make sure we don't ever have
2780            tests use invalid patch fetches by mistake.
2781
2782 2010-10-23  Dan Bernstein  <mitz@apple.com>
2783
2784         Build fix. Add stub implementations for required NSDraggingInfo methods.
2785
2786         * DumpRenderTree/mac/DumpRenderTreeDraggingInfo.mm:
2787         (-[DumpRenderTreeDraggingInfo draggingFormation]):
2788         (-[DumpRenderTreeDraggingInfo setDraggingFormation:]):
2789         (-[DumpRenderTreeDraggingInfo animatesToDestination]):
2790         (-[DumpRenderTreeDraggingInfo setAnimatesToDestination:]):
2791         (-[DumpRenderTreeDraggingInfo numberOfValidItemsForDrop]):
2792         (-[DumpRenderTreeDraggingInfo setNumberOfValidItemsForDrop:]):
2793         (-[DumpRenderTreeDraggingInfo enumerateDraggingItemsWithOptions:forView:classes:searchOptions:usingBlock:]):
2794
2795 2010-10-23  David Kilzer  <ddkilzer@apple.com>
2796
2797         <http://webkit.org/b/48186> Remove unneeded WebHTMLRepresentationInternal.h header
2798
2799         Reviewed by Sam Weinig.
2800
2801         The only method defined in WebHTMLRepresentationInternal.h is
2802         also defined in WebHTMLRepresentation.h, so use that instead.
2803
2804         * DumpRenderTree/mac/InternalHeaders/WebKit/WebHTMLRepresentationInternal.h: Removed.
2805         * DumpRenderTree/mac/WebArchiveDumpSupport.mm: Switched to use
2806         WebHTMLRepresentation.h.
2807
2808 2010-10-23  Sheriff Bot  <webkit.review.bot@gmail.com>
2809
2810         Unreviewed, rolling out r70367.
2811         http://trac.webkit.org/changeset/70367
2812         https://bugs.webkit.org/show_bug.cgi?id=48176
2813
2814         It made 8-10 tests crash on Qt bot (Requested by Ossy on
2815         #webkit).
2816
2817         * DumpRenderTree/LayoutTestController.cpp:
2818         (setCustomPolicyDelegateCallback):
2819         * DumpRenderTree/LayoutTestController.h:
2820         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2821         (LayoutTestController::setCustomPolicyDelegate):
2822         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2823         (LayoutTestController::setCustomPolicyDelegate):
2824         * DumpRenderTree/mac/PolicyDelegate.h:
2825         * DumpRenderTree/mac/PolicyDelegate.mm:
2826         (-[PolicyDelegate webView:decidePolicyForNavigationAction:request:frame:decisionListener:]):
2827         (-[PolicyDelegate setPermissive:]):
2828         (-[PolicyDelegate setControllerToNotifyDone:]):
2829         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2830         (LayoutTestController::setCustomPolicyDelegate):
2831         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2832         (LayoutTestController::setCustomPolicyDelegate):
2833
2834 2010-10-22  Eric Seidel  <eric@webkit.org>
2835
2836         Reviewed by Adam Barth.
2837
2838         EWS never releases patches which fail to apply
2839         https://bugs.webkit.org/show_bug.cgi?id=48171
2840
2841         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
2842         * Scripts/webkitpy/tool/commands/queues.py:
2843
2844 2010-10-22  Eric Seidel  <eric@webkit.org>
2845
2846         Unreviewed.
2847
2848         Make the EWS queues restart themselves more often.
2849         This matches the commit-queue.
2850
2851         * EWSTools/start-queue.sh:
2852
2853 2010-10-22  Eric Seidel  <eric@webkit.org>
2854
2855         Reviewed by Adam Barth.
2856
2857         commit-queue should not pass --quiet to subcommands now that the parent command does the reporting
2858         https://bugs.webkit.org/show_bug.cgi?id=48165
2859
2860         * Scripts/webkitpy/tool/bot/commitqueuetask.py:
2861         * Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
2862         * Scripts/webkitpy/tool/commands/queues_unittest.py:
2863
2864 2010-10-22  Mario Sanchez Prada  <msanchez@igalia.com>
2865
2866         Adding myself as a committer.
2867
2868         * Scripts/webkitpy/common/config/committers.py:
2869
2870 2010-10-20  Anders Carlsson  <andersca@apple.com>
2871
2872         Reviewed by Adam Barth and Darin Adler.
2873
2874         Using the decidePolicyForMIMEType delegate message in an asynchronous manner does not work
2875         https://bugs.webkit.org/show_bug.cgi?id=48014
2876         <rdar://problem/8202716>
2877         
2878         * DumpRenderTree/LayoutTestController.cpp:
2879         (setCustomPolicyDelegateCallback):
2880         Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
2881
2882         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2883         (LayoutTestController::setCustomPolicyDelegate):
2884         Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
2885
2886         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2887         (LayoutTestController::setCustomPolicyDelegate):
2888         Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
2889
2890         * DumpRenderTree/mac/PolicyDelegate.h:
2891         * DumpRenderTree/mac/PolicyDelegate.mm:
2892         (-[PolicyDelegate webView:decidePolicyForMIMEType:request:frame:decisionListener:]):
2893         If _callIgnoreInDecidePolicyForMIMETypeAfterOneSecond is true, call [listener ignore] after one second.
2894
2895         (-[PolicyDelegate setCallIgnoreInDecidePolicyForMIMETypeAfterOneSecond:]):
2896         Update _callIgnoreInDecidePolicyForMIMETypeAfterOneSecond.
2897         
2898         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2899         (LayoutTestController::setCustomPolicyDelegate):
2900         Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
2901
2902         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2903         (LayoutTestController::setCustomPolicyDelegate):
2904         Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
2905
2906 2010-10-22  Sam Weinig  <sam@webkit.org>
2907
2908         Reviewed by Anders Carlsson.
2909
2910         Rename the InjectedBundle for TestWebKitAPI to InjectedBundleTestWebKitAPI.bundle
2911
2912         * TestWebKitAPI/Configurations/InjectedBundle.xcconfig:
2913         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
2914         * TestWebKitAPI/mac/PlatformUtilitiesMac.mm:
2915         (TestWebKitAPI::Util::createInjectedBundlePath):
2916
2917 2010-10-22  Dirk Pranke  <dpranke@chromium.org>
2918
2919         Reviewed by Eric Seidel.
2920
2921         test-webkitpy fails when passed -v (or any other option)
2922
2923         Fix port/base_unittest to pass an explicit list of arguments
2924         rather than accidentally picking up sys.argv.
2925
2926         https://bugs.webkit.org/show_bug.cgi?id=48071
2927
2928         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2929
2930 2010-10-22  Brian Weinstein  <bweinstein@apple.com>
2931
2932         Windows build fix. Update the createNewPage callback to account for the new
2933         arguments.
2934
2935         * MiniBrowser/win/BrowserView.cpp:
2936         (createNewPage):
2937
2938 2010-10-22  Sam Weinig  <sam@webkit.org>
2939
2940         Reviewed by Anders Carlsson.
2941
2942         WebKit2 needs to pass the current event modifier flags when requesting a new window
2943         https://bugs.webkit.org/show_bug.cgi?id=48140
2944
2945         * MiniBrowser/mac/BrowserWindowController.m:
2946         (createNewPage):
2947         * WebKitTestRunner/TestController.cpp:
2948         (WTR::createOtherPage):
2949         Update for new signature for the WKPageCreateNewPageCallback.
2950
2951 2010-10-22  Eric Seidel  <eric@webkit.org>
2952
2953         Unreviewed.  Just adding a test case.
2954
2955         The style-queue was not recognizing new patches.  It turns
2956         out I had fixed this in an earlier change, but just not
2957         deployed to queues.webkit.org.  As part of investigating why it
2958         was broken, I wrote a test for my previous change which I'm now landing.
2959
2960         * QueueStatusServer/model/queues_unittest.py:
2961
2962 2010-10-22  Eric Seidel  <eric@webkit.org>
2963
2964         Reviewed by Adam Barth.
2965
2966         EWS does not need to process obsolete patches
2967         https://bugs.webkit.org/show_bug.cgi?id=48093
2968
2969         This was an easy change, but to test it I had to pipe
2970         real Attachment objects into the queue testing system.
2971         Doing so revealed a whole bunch of bugs in our unit tests,
2972         which I fixed as part of this patch.
2973
2974         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
2975          - This is the actual code change.  This will not reduce the
2976            backlog in the EWS queues much, but it will make rejections
2977            much quicker for obsolete patches or closed bugs.
2978         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
2979          - Test my new code.
2980          - Used a real attachment object and got rid of MockPatch
2981          - Shared code between the mac-ews and cr-mac-ews tests.
2982         * Scripts/webkitpy/tool/commands/queues_unittest.py:
2983          - Can't use MockPatch anymore.
2984          - Removing MockPatch found more bugs here!
2985         * Scripts/webkitpy/tool/commands/queuestest.py:
2986         * Scripts/webkitpy/tool/commands/sheriffbot_unittest.py:
2987         * Scripts/webkitpy/tool/mocktool.py:
2988          - MockBugzilla should not be a "Mock" object.  Right now tool.bugs()
2989            is allowed, but wrong.  Making it not a Mock will make tool.bugs() correctly fail.
2990
2991 2010-10-22  Sheriff Bot  <webkit.review.bot@gmail.com>
2992
2993         Unreviewed, rolling out r70301.
2994         http://trac.webkit.org/changeset/70301
2995         https://bugs.webkit.org/show_bug.cgi?id=48126
2996
2997         "Lang attribute layout tests failing" (Requested by satish on
2998         #webkit).
2999
3000         * DumpRenderTree/LayoutTestController.cpp:
3001         (setMockSpeechInputResultCallback):
3002         * DumpRenderTree/LayoutTestController.h:
3003         * DumpRenderTree/chromium/LayoutTestController.cpp:
3004         (LayoutTestController::setMockSpeechInputResult):
3005         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3006         (LayoutTestController::setMockSpeechInputResult):
3007         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3008         (LayoutTestController::setMockSpeechInputResult):
3009         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3010         (LayoutTestController::setMockSpeechInputResult):
3011         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3012         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3013         (LayoutTestController::setMockSpeechInputResult):
3014         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3015         (LayoutTestController::setMockSpeechInputResult):
3016
3017 2010-10-22  Leandro Gracia Gil  <leandrogracia@google.com>
3018
3019         Reviewed by Jeremy Orlow.
3020
3021         Added a second parameter to setMockSpeechInputResult for
3022         the language used in speech input.
3023         https://bugs.webkit.org/show_bug.cgi?id=47089
3024
3025         * DumpRenderTree/LayoutTestController.cpp:
3026         (setMockSpeechInputResultCallback):
3027         * DumpRenderTree/LayoutTestController.h:
3028         * DumpRenderTree/chromium/LayoutTestController.cpp:
3029         (LayoutTestController::setMockSpeechInputResult):
3030         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3031         (LayoutTestController::setMockSpeechInputResult):
3032         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3033         (LayoutTestController::setMockSpeechInputResult):
3034         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3035         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3036         (LayoutTestController::setMockSpeechInputResult):
3037         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3038         (LayoutTestController::setMockSpeechInputResult):
3039         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3040         (LayoutTestController::setMockSpeechInputResult):
3041
3042 2010-10-21  MORITA Hajime  <morrita@google.com>
3043
3044         Reviewed by Kent Tamura.
3045
3046         [Win][DRT] should have LayoutTestController.hasSpellingMarker()
3047         https://bugs.webkit.org/show_bug.cgi?id=47885
3048
3049         - Implemented LayoutTestController.hasSpellingMarker(),
3050         - Added fake spellcheck implementation.
3051
3052         * DumpRenderTree/win/DumpRenderTree.cpp:
3053         (main):
3054         * DumpRenderTree/win/EditingDelegate.cpp:
3055         (indexOfFirstWordCharacter):
3056         (wordLength):
3057         (EditingDelegate::checkSpellingOfString):
3058         * DumpRenderTree/win/EditingDelegate.h:
3059         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3060         (LayoutTestController::hasSpellingMarker):
3061
3062 2010-10-21  Mihai Parparita  <mihaip@chromium.org>
3063
3064         Unreviewed. Re-enable test that was mistakenly disabled by r67974.
3065
3066         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
3067
3068 2010-10-21  Mihai Parparita  <mihaip@chromium.org>
3069
3070         Unreviewed fix for rebaseline-chromium-webkit-tests.
3071
3072         Port.diff_image no longer has a tolerance parameter. Also, use
3073         get_option('tolerance') which is safer if the options object doesn't
3074         define a tolerance attribute (it doesn't for the one used in
3075         rebaseline_chromium_webkit_tests).
3076
3077         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
3078         * Scripts/webkitpy/layout_tests/port/webkit.py:
3079         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
3080
3081 2010-10-21  Mihai Parparita  <mihaip@chromium.org>
3082
3083         Reviewed by Ojan Vafai.
3084
3085         Add support for --tolerance in NRWT
3086         https://bugs.webkit.org/show_bug.cgi?id=47959
3087         
3088         Add support for the --tolerance flag in NRWT. The Port.diff_image
3089         signature shouldn't need a tolerance parameter (it's not set per test),
3090         just have ports that use it (currently only WebKitPort) read it from 
3091         the options object.
3092
3093         * Scripts/webkitpy/layout_tests/port/chromium.py:
3094         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
3095         * Scripts/webkitpy/layout_tests/port/test.py:
3096         * Scripts/webkitpy/layout_tests/port/webkit.py:
3097         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3098         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
3099
3100 2010-10-21  Eric Seidel  <eric@webkit.org>
3101
3102         Reviewed by Adam Barth.
3103
3104         Add --suggest-reviewers option to upload to auto-suggest reviewers for your patch
3105         https://bugs.webkit.org/show_bug.cgi?id=48088
3106
3107         This is a first-pass.  Works, but we'll eventually
3108         turn this on by default, refine the suggestion algorithm
3109         and possibly move it to a different place in the upload step order.
3110
3111         * Scripts/webkitpy/common/checkout/api.py:
3112         * Scripts/webkitpy/tool/commands/download_unittest.py:
3113         * Scripts/webkitpy/tool/commands/upload.py:
3114         * Scripts/webkitpy/tool/commands/upload_unittest.py:
3115         * Scripts/webkitpy/tool/mocktool.py:
3116         * Scripts/webkitpy/tool/steps/__init__.py:
3117         * Scripts/webkitpy/tool/steps/options.py:
3118
3119 2010-10-21  Mario Sanchez Prada  <msanchez@igalia.com>
3120
3121         Reviewed by Chris Fleizach.
3122
3123         [GTK] Segfault while testing accessibility/iframe-bastardization.html
3124         https://bugs.webkit.org/show_bug.cgi?id=30123
3125
3126         Check m_element before actually using it to get the parent.
3127
3128         * DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
3129         (AccessibilityUIElement::parentElement): Add an extra check to
3130         make sure m_element points to a valid value before actually using
3131         it to get the parent. Also, assert ATK_IS_OBJECT(m_element).
3132
3133 2010-10-21  Anders Carlsson  <andersca@apple.com>
3134
3135         Reviewed by Adam Roben.
3136
3137         Crash evaluating JavaScript string that throws an exception
3138         https://bugs.webkit.org/show_bug.cgi?id=48092
3139         <rdar://problem/8487657>
3140
3141         Add a test that evaluates a JavaScript string that throws an exception and check that 
3142         the callback is called.
3143
3144         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
3145         * TestWebKitAPI/Tests/WebKit2/EvaluateJavaScript.cpp: Added.
3146         (TestWebKitAPI::didRunJavaScript):
3147         (TestWebKitAPI::WebKit2_EvaluateJavaScript):
3148         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
3149
3150 2010-10-21  Adam Roben  <aroben@apple.com>
3151
3152         Test that the plugin's HWND is invalidated when NPN_InvalidateRect is
3153         called
3154
3155         Test for <http://webkit.org/b/48086> <rdar://problem/8482944>
3156         Silverlight doesn't repaint in WebKit2
3157
3158         Reviewed by Anders Carlsson.
3159
3160         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
3161         (PluginTest::NPN_InvalidateRect): Added. Calls through to the browser.
3162         (executeScript): Added. Asks the browser to evaluate the script.
3163
3164         (PluginTest::waitUntilDone):
3165         (PluginTest::notifyDone):
3166         Added. Calls through to layoutTestController.
3167
3168         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Added
3169         * NPN_InvalidateRect and waitUntilDone/notifyDone.
3170
3171         * DumpRenderTree/TestNetscapePlugIn/Tests/win/NPNInvalidateRectInvalidatesWindow.cpp: Added.
3172         (TemporaryWindowMover::moveSucceeded):
3173         (TemporaryWindowMover::TemporaryWindowMover):
3174         (TemporaryWindowMover::~TemporaryWindowMover):
3175         This class moves a window on-screen and shows it, then moves it back and hides it.
3176
3177         (NPNInvalidateRectInvalidatesWindow::NPNInvalidateRectInvalidatesWindow):
3178         Initialize our members.
3179         (NPNInvalidateRectInvalidatesWindow::~NPNInvalidateRectInvalidatesWindow):
3180         Delete our window mover if it hasn't been deleted already.
3181         (NPNInvalidateRectInvalidatesWindow::NPP_SetWindow): Subclass the
3182         plugin HWND and move the test harness window on screen.
3183         (NPNInvalidateRectInvalidatesWindow::wndProc): Call through to onPaint
3184         when we get a WM_PAINT message.
3185         (NPNInvalidateRectInvalidatesWindow::onPaint): Do the test and tell
3186         LayoutTestController we're done.
3187         (NPNInvalidateRectInvalidatesWindow::testInvalidateRect): Validate
3188         ourselves, invalidate our lower-right quadrant via NPN_InvalidateRect,
3189         then check that our HWND's invalid region is the rect that we
3190         invalidated.
3191
3192         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
3193         * Added NPNInvalidateRectInvalidatesWindow.
3194
3195 2010-10-21  Daniel Bates  <dbates@rim.com>
3196
3197         Add Git-support to do-file-rename
3198         https://bugs.webkit.org/show_bug.cgi?id=48015
3199
3200         Fix tools. Export function scmMoveOrRenameFile so that it can be called from
3201         do-file-rename and do-webcore-rename. I inadvertently left this out of the patch.
3202
3203         * Scripts/VCSUtils.pm:
3204
3205 2010-10-21  Eric Seidel  <eric@webkit.org>
3206
3207         Reviewed by Adam Barth.
3208
3209         mac-ews is not properly releasing patches
3210         https://bugs.webkit.org/show_bug.cgi?id=48076
3211
3212         mac-ews overrides process_work_item, so it was not calling
3213         release_work_item like the default process_work_item would.
3214         To fix this I made all the status-reporting methods just
3215         release the patch.  I expect we'll iterate on this design further.
3216
3217         * Scripts/webkitpy/common/net/statusserver.py:
3218         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
3219         * Scripts/webkitpy/tool/commands/queues.py:
3220         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3221
3222 2010-10-18  Martin Robinson  <mrobinson@igalia.com>
3223
3224         Reviewed by Gustavo Noronha Silva.
3225
3226         [GTK] mathml flaky tests after adding SystemFonts to the theme
3227         https://bugs.webkit.org/show_bug.cgi?id=47727
3228
3229         The Fontconfig setup in the GTK+ now specifically checks for and
3230         loads the STIX fonts for MathML tests.
3231
3232         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3233         (initializeFonts): Add hooks for loading the STIX fonts specifically.
3234
3235 2010-10-20  Adam Roben  <aroben@apple.com>
3236
3237         Test that the UI client gets notified when WKView receives a WM_CLOSE
3238         message
3239
3240         Test for <http://webkit.org/b/48044> <rdar://problem/8488446> Pressing
3241         Ctrl+W when viewing a full-page PDF destroys the WKView but doesn't
3242         close its parent window
3243
3244         Reviewed by Jon Honeycutt.
3245
3246         * TestWebKitAPI/Tests/WebKit2/win/WMCloseCallsUIClientClose.cpp: Added.
3247         (TestWebKitAPI::close): Record that this function was called.
3248         (TestWebKitAPI::WebKit2_WMCloseCallsUIClientClose): Create a WKView,
3249         send it a WM_CLOSE message, and test that the UI client's close
3250         callback got called. Note that this will hang if the bug is
3251         reintroduced; <http://webkit.org/b/48043> covers making TestWebKitAPI
3252         able to handle hangs.
3253
3254         * TestWebKitAPI/win/TestWebKitAPI.vcproj: Added the new test.
3255
3256 2010-10-20  Adam Roben  <aroben@apple.com>
3257
3258         Make prepare-ChangeLog much faster when using git
3259
3260         This change also seems to make it detect renames better in some cases.
3261
3262         Fixes <http://webkit.org/b/48040> prepare-ChangeLog is slow when using
3263         git
3264
3265         Reviewed by David Kilzer.
3266
3267         * Scripts/prepare-ChangeLog:
3268         (statusCommand):
3269         (createPatchCommand):
3270         Use "-M -C" instead of "-C -C -M" to tell git-diff to detect renames
3271         and copies. The two "-C"s were making it read many more files than were
3272         necessary.
3273
3274 2010-10-21  Sam Weinig  <sam@webkit.org>
3275
3276         Reviewed by Adam Roben.
3277
3278         Null frame passed when running alert from UserScript run at document start
3279         <rdar://problem/8573809>
3280         https://bugs.webkit.org/show_bug.cgi?id=48036
3281
3282         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3283         * TestWebKitAPI/InjectedBundleController.cpp:
3284         (TestWebKitAPI::InjectedBundleController::didReceiveMessage):
3285         (TestWebKitAPI::InjectedBundleController::initializeTestNamed):
3286         * TestWebKitAPI/InjectedBundleController.h:
3287         * TestWebKitAPI/InjectedBundleTest.h:
3288         (TestWebKitAPI::InjectedBundleTest::initialize):
3289         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
3290         * TestWebKitAPI/Tests/WebKit2/DocumentStartUserScriptAlertCrash.cpp: Added.
3291         (TestWebKitAPI::runJavaScriptAlert):
3292         (TestWebKitAPI::TEST):
3293         * TestWebKitAPI/Tests/WebKit2/DocumentStartUserScriptAlertCrash_Bundle.cpp: Added.
3294         (TestWebKitAPI::DocumentStartUserScriptAlertCrashTest::DocumentStartUserScriptAlertCrashTest):
3295         (TestWebKitAPI::DocumentStartUserScriptAlertCrashTest::initialize):
3296         Add test for invoking an alert during a UserScript run at document start.
3297
3298 2010-10-21  Andreas Kling  <kling@webkit.org>
3299
3300         Reviewed by Adam Roben.
3301
3302         TestNetscapePlugIn: Actually call NPP_SetWindow on Unix
3303
3304         We were just returning NPERR_NO_ERROR previously.
3305
3306         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
3307         (webkit_test_plugin_set_window):
3308
3309 2010-10-21  David Kilzer  <ddkilzer@apple.com>
3310
3311         <http://webkit.org/b/48051> Move macros from DumpRenderTreeMac.h to config.h
3312
3313         Reviewed by Adam Roben.
3314
3315         * DumpRenderTree/config.h: Moved macros to here...
3316         * DumpRenderTree/mac/DumpRenderTreeMac.h: ...from here.
3317         Addresses a FIXME comment from r28419.
3318
3319 2010-10-21  David Kilzer  <ddkilzer@apple.com>
3320
3321         <http://webkit.org/b/48047> Fix warnings found by check-Xcode-source-file-types
3322
3323         Reviewed by Adam Roben.
3324
3325         Fixes the following warnings:
3326
3327             WARNING: Incorrect file type 'sourcecode.cpp.objcpp' for file 'PixelDumpSupport.cpp'.
3328             WARNING: Incorrect file type 'sourcecode.cpp.objcpp' for file 'LayoutTestController.cpp'.
3329             WARNING: Incorrect file type 'sourcecode.cpp.objcpp' for file 'WorkQueue.cpp'.
3330             WARNING: Incorrect file type 'sourcecode.cpp.objcpp' for file 'cg/PixelDumpSupportCG.cpp'.
3331             WARNING: Incorrect file type 'sourcecode.cpp.objcpp' for file 'GCController.cpp'.
3332             5 issues found for WebKitTools/DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj.
3333
3334         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3335         Removed explicit file type for *.cpp files above to make them
3336         match other C++ source files.  This exposed the fact that
3337         DumpRenderTreeMac.h was not safe to include in C++ source, which
3338         caused PixelDumpSupport.cpp and PixelDumpSupportCG.cpp to fail
3339         to build.
3340         * DumpRenderTree/PixelDumpSupport.cpp: Adjusted header order.
3341         * DumpRenderTree/mac/DumpRenderTreeMac.h: Made safe to include
3342         in C++ source files.  Included CoreFoundation/CoreFoundation.h
3343         to make sure all CF types were defined.  Removed CFStringRef
3344         typedef.
3345
3346 2010-10-21  Adam Roben  <aroben@apple.com>
3347
3348         Attempt to fix plugins/pass-different-npp-struct.html on GTK.
3349
3350         See <http://webkit.org/b/47690>.
3351
3352         * GNUmakefile.am: Added PassDifferentNPPStruct.cpp.
3353
3354 2010-10-21  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3355
3356         Reviewed by Csaba Osztrogonác.
3357
3358         [NRWT] Set ImageDiff path on Qt port
3359         https://bugs.webkit.org/show_bug.cgi?id=48052
3360
3361         * Scripts/webkitpy/layout_tests/port/qt.py:
3362
3363 2010-10-21  Adam Barth  <abarth@webkit.org>
3364
3365         Reviewed by Eric Seidel.
3366
3367         Don't CC authors of flaky tests
3368         https://bugs.webkit.org/show_bug.cgi?id=48038
3369
3370         Some authors found this too annoying.  We'll look for another way to
3371         close the flaky test loop.
3372
3373         * Scripts/webkitpy/tool/commands/queues.py:
3374         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3375
3376 2010-10-21  Eric Seidel  <eric@webkit.org>
3377
3378         Reviewed by Adam Barth.
3379
3380         commit-queue should run run-webkit-tests with --no-new-test-results
3381         https://bugs.webkit.org/show_bug.cgi?id=47998
3382
3383         * Scripts/webkitpy/tool/steps/runtests.py:
3384         * Scripts/webkitpy/tool/steps/steps_unittest.py:
3385
3386 2010-10-21  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3387
3388         Reviewed by Ojan Vafai.
3389
3390         [NRWT] Get child process number from an environment variable
3391         https://bugs.webkit.org/show_bug.cgi?id=47981
3392
3393         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3394
3395 2010-10-20  David Kilzer  <ddkilzer@apple.com>
3396
3397         <http://webkit.org/b/47754> New script to verify explicit source file types in Xcode project files
3398
3399         Reviewed by Darin Adler.
3400
3401         The script parses an Xcode project file and makes sure the file
3402         extension matches the explicit file type set for all source
3403         files.  Note that the majority of source files will have their
3404         type set by Xcode, so the script doesn't check them since there
3405         is no need to second-guess Xcode.
3406
3407         * Scripts/check-Xcode-source-file-types: Added.  Code borrowed
3408         heavily from sort-Xcode-project-file.
3409
3410 2010-10-20  Adam Roben  <aroben@apple.com>
3411         
3412         Windows build fix
3413
3414         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp: Added missing #include.
3415
3416 2010-10-20  Adam Roben  <aroben@apple.com>
3417
3418         Qt test fix
3419
3420         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro: Added
3421         PassDifferentNPPStruct.cpp.
3422
3423 2010-10-20  Sam Weinig  <sam@webkit.org>
3424
3425         Reviewed by Adam Roben.
3426
3427         Add ability to test injected bundle API using TestWebKitAPI
3428         https://bugs.webkit.org/show_bug.cgi?id=48027
3429
3430         * TestWebKitAPI/InjectedBundleController.cpp: Added.
3431         * TestWebKitAPI/InjectedBundleController.h: Added.
3432         Shared main object for bundle functionality.
3433
3434         * TestWebKitAPI/InjectedBundleMain.cpp: Added.
3435         Bundle entry point.
3436
3437         * TestWebKitAPI/InjectedBundleTest.h: Added.
3438         Base class for which the bundle portion of a test derives from. 
3439
3440         * TestWebKitAPI/PlatformUtilities.h:
3441         * TestWebKitAPI/PlatformUtilities.cpp: Added.
3442         * TestWebKitAPI/mac/PlatformUtilitiesMac.mm:
3443         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp:
3444         (TestWebKitAPI::Util::createInjectedBundlePath):
3445         (TestWebKitAPI::Util::createURLForResource):
3446         (TestWebKitAPI::Util::URLForNonExistentResource):
3447         Add helper to create a context with the shared injected bundle,
3448         and send the initial message to set up the test.
3449
3450         * TestWebKitAPI/Configurations/InjectedBundle.xcconfig: Added.
3451         * TestWebKitAPI/InjectedBundle-Info.plist: Added.
3452         Add mac configuration files.
3453
3454         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
3455         Add the new files.
3456
3457         * TestWebKitAPI/Tests/WebKit2/InjectedBundleBasic.cpp: Added.
3458         * TestWebKitAPI/Tests/WebKit2/InjectedBundleBasic_Bundle.cpp: Added.
3459         Add a simple initial bundle test.
3460
3461 2010-10-20  Eric Seidel  <eric@webkit.org>
3462
3463         Unreviewed.  Fixing /release-patch url used by the EWS bots.
3464
3465         EWS should test patches with r+
3466         https://bugs.webkit.org/show_bug.cgi?id=35460
3467
3468         * Scripts/webkitpy/common/net/statusserver.py:
3469          - I changed the URL during development, and 404s are
3470          intentionally silenced during release_patch.
3471
3472 2010-10-14  Adam Roben  <aroben@apple.com>
3473
3474         Test that passing a different NPP struct back to the browser doesn't
3475         cause an assertion failure
3476
3477         Test for <http://webkit.org/b/47690> <rdar://problem/8553020>
3478         Assertion failure in NetscapePlugin::fromNPP when using Shockwave in
3479         WebKit2
3480
3481         Reviewed by John Sullivan.
3482
3483         * DumpRenderTree/TestNetscapePlugIn/Tests/PassDifferentNPPStruct.cpp: Copied from WebKitTools/DumpRenderTree/TestNetscapePlugIn/Tests/win/GetValueNetscapeWindow.cpp.
3484         (PassDifferentNPPStruct::PassDifferentNPPStruct): Initialize our
3485         members.
3486         (PassDifferentNPPStruct::NPP_SetWindow): Pass a different NPP to the
3487         browser than the one it gave us in NPP_New and see if it works.
3488
3489         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3490         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
3491         Added PassDifferentNPPStruct.
3492
3493 2010-10-19  Adam Roben  <aroben@apple.com>
3494
3495         Gently nudge old-run-webkit-tests toward working with Win32 Perl
3496
3497         This makes old-run-webkit-tests able to build DRT and find all the
3498         tests to run. It even invokes DRT and passes it the list of tests. But
3499         DRT ends up hung blocking on I/O.
3500
3501         Fixes <http://webkit.org/b/47961> Get old-run-webkit-tests mostly
3502         working with Win32 Perl
3503
3504         Reviewed by David Kilzer.
3505
3506         * Scripts/old-run-webkit-tests:
3507           - Use File::Spec instead of manually concatenating paths
3508           - Use dirname instead of manually stripping off the base name
3509           - Use isCygwin/isWindows/isAppleWinWebKit more judiciously
3510           - Explicitly invoke Perl when running Perl scripts
3511           - Quote paths when using them in regular expressions to allow them
3512             to include characters that have special meanings in regular
3513             expressions
3514
3515         * Scripts/run-webkit-tests: Use File::Spec instead of manually
3516         concatenating paths.
3517
3518         * Scripts/webkitdirs.pm:
3519           - Remove the unused $windowsTmpPath variable
3520           - Use isCygwin/isWindows/isAppleWinWebKit more judiciously
3521           - Only pass paths to cygpath when using Cygwin Perl
3522           - Only use pdevenv when using Cygwin Perl, for now
3523
3524 2010-10-20  Daniel Bates  <dbates@rim.com>
3525
3526         Reviewed by Martin Robinson.
3527
3528         Add Git-support to do-file-rename
3529         https://bugs.webkit.org/show_bug.cgi?id=48015
3530
3531         Also, abstracts the SCM move/rename functionality in do-file-rename and
3532         do-webcore-rename into a common function VCSUtils::scmMoveOrRenameFile().
3533
3534         Currently, do-file-rename is hard coded to assume the SCM is Subversion.
3535         Instead, we should abstract the rename logic to be SCM-independent. This
3536         will allow us to add Git support as well move such functionality into
3537         our SCM library VCSUtils, where it can be shared by do-webcore-rename.
3538
3539         * Scripts/VCSUtils.pm:
3540           - Added function scmMoveOrRenameFile.
3541         * Scripts/do-file-rename: Modified to call VCSUtils::scmMoveOrRenameFile().
3542         * Scripts/do-webcore-rename: Ditto.
3543
3544 2010-10-20  Adam Roben  <aroben@apple.com>
3545
3546         Fix old-run-webkit-tests when there's a space in the path to DRT
3547
3548         Reviewed by Jon Honeycutt.
3549
3550         * Scripts/old-run-webkit-tests: Quote the path to DRT before executing
3551         it.
3552
3553 2010-10-20  Kenneth Russell  <kbr@google.com>
3554
3555         Reviewed by Dimitri Glazkov.
3556
3557         Regression in chromium_gpu_unittests after r70175
3558         https://bugs.webkit.org/show_bug.cgi?id=48008
3559
3560         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
3561
3562 2010-10-20  Kenneth Russell  <kbr@google.com>
3563
3564         Reviewed by James Robinson.
3565
3566         chromium_gpu port of new-run-webkit-tests must do Linux -> Win expectations fallback
3567         https://bugs.webkit.org/show_bug.cgi?id=48005
3568
3569         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py: