ac40e0b4a7200f68d9f17b4a505cbd6ad694833f
[WebKit-https.git] / WebKitTools / ChangeLog
1 2010-11-08  John Knottenbelt  <jknotten@chromium.org>
2
3         Reviewed by Kent Tamura.
4
5         WebViewHost::reset() uses placement new.
6         https://bugs.webkit.org/show_bug.cgi?id=49069
7
8         WebViewHost is using placement destruction / new to simulate a fresh
9         WebViewHost object at the same address. This is because the WebView remains
10         open across tests and maintains a pointer to the WebViewHost.
11         This change resets member variables explictly instead of the placement new dance.
12
13         * DumpRenderTree/chromium/WebViewHost.cpp:
14         (WebViewHost::WebViewHost):
15         (WebViewHost::reset):
16
17 2010-11-08  Dirk Pranke <dpranke@chromium.org>
18
19         Reviewed by Eric Seidel.
20
21         Fix the webkit_base_dir logic in webkitpy/layout_tests/port/config.py. 
22         It turns out that NRWT can't use the code in scm.find_checkout_root()
23         because the Chromium bots don't do full checkouts of the WebKit
24         tree; they only check out subdirectories like WebKitTools/Scripts.
25         Until we can figure out a better approach for this, I've
26         restored the base_dir-detecting code from NRWT, which works in
27         any directory tree, scm or no.
28
29         This also restores the files modified in r71475 and r71474.
30
31         https://bugs.webkit.org/show_bug.cgi?id=49151
32
33         * WebKitTools/Scripts/webkitpy/layout_tests/port/config.py
34         * WebKitTools/Scripts/webkitpy/layout_tests/port/config_unittest.py
35
36 2010-11-08  Eric Seidel  <eric@webkit.org>
37
38         Reviewed by Adam Barth.
39
40         webkit-patch failure-reason explodes if a build is missing
41         https://bugs.webkit.org/show_bug.cgi?id=49195
42
43         This is likely a recent regression.  Adding a unit test for this case.
44
45         * Scripts/webkitpy/common/net/buildbot.py:
46         * Scripts/webkitpy/common/net/buildbot_unittest.py:
47
48 2010-11-08  Dan Bernstein  <mitz@apple.com>
49
50         Reviewed by Darin Adler.
51
52         Ignore the system scrollbar setting.
53
54         * DumpRenderTree/mac/DumpRenderTree.mm:
55         (resetDefaultsToConsistentValues): Set the scrollbar default.
56
57 2010-11-05  Adam Roben  <aroben@apple.com>
58
59         Make webkitpy.common.system.executive_unittest pass when running under
60         Win32 Python
61
62         Fixes <http://webkit.org/b/49033>.
63
64         Reviewed by Dave Levin and Eric Seidel.
65
66         * Scripts/webkitpy/common/system/executive.py:
67         (Executive._run_command_with_teed_output): Pass the arguments through
68         encode_argument_if_needed rather than using Cygwin-specific code here.
69         (Executive.run_and_throw_if_fail): Use child_process_encoding to decode
70         the output.
71         (Executive.run_command): Use encode_argument_if_needed to encode the
72         arguments and child_process_encoding to decode the output.
73         (Executive._child_process_encoding): Returns the encoding that should be
74         used when communicating with child processes. On Windows we use mbcs,
75         which maps to the current code page. On all other platforms we use
76         UTF-8.
77         (Executive._should_encode_child_process_arguments): Returns True if
78         arguments to child processes need to be encoded. This is currently
79         only needed on Cygwin and Win32 Python 2.x.
80         (Executive._encode_argument_if_needed): Encode the argument using
81         child_process_encoding if we need to encode arguments to child
82         processes on this platform.
83
84         * Scripts/webkitpy/common/system/executive_unittest.py:
85         (never_ending_command): Added. Returns arguments to run a command that
86         will not quit until we kill it. On Windows we use wmic, on other
87         platforms we use yes.
88         (ExecutiveTest.test_run_command_with_unicode): Changed to expect the
89         mbcs encoding to be used and for output from the child processes to
90         have been roundtripped through encode/decode on Win32 Python. When
91         UTF-8 is the encoding the roundtripping is undetectable, but with mbcs
92         it's possible that some characters will not be able to be converted
93         and will be replaced by question marks; the round-tripping allows us
94         to expect this result.
95
96         (ExecutiveTest.test_kill_process):
97         (ExecutiveTest.test_kill_all):
98         Use never_ending_command instead of invoking "yes" directly. Expect an
99         exit code of 1 when using Win32 Python, as that's what seems to happen.
100
101 2010-11-08  Adam Roben  <aroben@apple.com>
102
103         Roll out r71532
104
105         It broke the build for Cygwin 1.7 installs. Cygwin 1.7's default
106         .bashrc unsets %TEMP%, which broke copy-tools.cmd.
107
108         * Scripts/webkitdirs.pm:
109
110 2010-11-08  Tony Chang  <tony@chromium.org>
111
112         Reviewed by Adam Barth.
113
114         run platform/chromium/plugins/nested-plugin-objects.html on all platforms
115         https://bugs.webkit.org/show_bug.cgi?id=49094
116
117         This tests that objects created by plugins are proplery cleaned up.
118
119         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
120         (testPassTestObject):
121         (pluginInvoke):
122         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
123         (testAllocate):
124         (testDeallocate):
125         (testGetProperty):
126         (testConstruct):
127
128 2010-11-08  Adam Roben  <aroben@apple.com>
129
130         Mark Windows builds triggered from Perl as being non-interactive
131
132         This affects whether some of our scripts will show alerts vs. printing
133         to the build log.
134
135         Fixes <http://webkit.org/b/49181> Windows build fail mysteriously when
136         .vsprops files are updated
137
138         Reviewed by Steve Falkenburg.
139
140         * Scripts/webkitdirs.pm:
141         (buildVisualStudioProject): Set WEBKIT_NONINTERACTIVE_BUILD to 1.
142
143 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
144
145         Reviewed by Ojan Vafai.
146
147         Make http locking default in NRWT.
148         https://bugs.webkit.org/show_bug.cgi?id=48053
149
150         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
151         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
152
153 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
154
155         Reviewed by Csaba Osztrogonác.
156
157         [NRWT] If the http lock fails we shouldn't do any locking
158         https://bugs.webkit.org/show_bug.cgi?id=49164
159
160         If something goes wrong with the locking, the test should keep going.
161
162         * Scripts/webkitpy/layout_tests/port/http_lock.py:
163
164 2010-11-08  Adam Roben  <aroben@apple.com>
165
166         Switch back to using kCGImageAlphaPremultipliedFirst when generating
167         pixel dumps on Windows
168
169         I changed this behavior in r71418 thinking that it was required for
170         getting plugins to show up in pixel dumps. But it doesn't seem to be
171         necessary, and was making it impossible to compare new Windows pixel
172         dumps with existing Windows or Mac pixel dumps (because ImageDiff won't
173         compare an image with alpha to an image without alpha).
174
175         Fixes <http://webkit.org/b/49172> REGRESION (r71418): Can't compare
176         new Windows pixel results to existing Windows or Mac results
177
178         Reviewed by Antti Koivisto.
179
180         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
181         (createBitmapContextFromWebView): Replaced kCGImageAlphaNoneSkipFirst
182         with kCGImageAlphaPremultipliedFirst.
183
184 2010-11-08  Csaba Osztrogonac  <ossy@webkit.org>
185
186         Unreviewed, rolling out r71466.
187         http://trac.webkit.org/changeset/71466
188         https://bugs.webkit.org/show_bug.cgi?id=48865
189
190         It broke layout tests on GTK bots.
191
192         * DumpRenderTree/gtk/DumpRenderTree.cpp:
193         (createWebView):
194
195 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
196
197         Reviewed by Csaba Osztrogonác.
198
199         Enable running of Qt API tests on BuildBot
200         https://bugs.webkit.org/show_bug.cgi?id=49004
201
202         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
203
204 2010-11-08  Eric Seidel  <eric@webkit.org>
205
206         Reviewed by Adam Barth.
207
208         Add clean-review-queue command to remove closed bugs from the webkit.org/pending-review
209         https://bugs.webkit.org/show_bug.cgi?id=49160
210
211         Bugzilla doesn't automatically remove r? when a bug gets closed.
212         This script takes care of that for webkit.org.
213
214         * Scripts/webkitpy/common/net/bugzilla.py:
215         * Scripts/webkitpy/tool/commands/upload.py:
216
217 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
218
219         Unreviewed, rolling out r71474.
220         http://trac.webkit.org/changeset/71474
221         https://bugs.webkit.org/show_bug.cgi?id=48280
222
223         breaks chromium webkit tests
224         https://bugs.webkit.org/show_bug.cgi?id=49151
225
226         * Scripts/webkitpy/layout_tests/port/base.py:
227         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
228         * Scripts/webkitpy/layout_tests/port/config.py:
229         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
230         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
231         * Scripts/webkitpy/layout_tests/port/webkit.py:
232         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
233         * Scripts/webkitpy/tool/mocktool.py:
234
235 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
236
237         Unreviewed, rolling out r71475.
238         http://trac.webkit.org/changeset/71475
239  
240         breaks chromium webkit tests
241         https://bugs.webkit.org/show_bug.cgi?id=49151
242
243         * Scripts/webkitpy/common/newstringio.py: Removed.
244         * Scripts/webkitpy/common/newstringio_unittest.py: Removed.
245         * Scripts/webkitpy/common/system/executive_mock.py: Removed.
246         * Scripts/webkitpy/common/system/filesystem_mock.py: Removed.
247         * Scripts/webkitpy/layout_tests/port/config_mock.py: Removed.
248
249 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
250
251         Unreviewed, build fix.
252
253         Add files inexplicably not committed in r71474 as part of the
254         fix for bug 48280.
255
256         * Scripts/webkitpy/common/newstringio.py: Added.
257         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
258         * Scripts/webkitpy/common/system/executive_mock.py: Added.
259         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
260         * Scripts/webkitpy/layout_tests/port/config_mock.py: Added.
261
262 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
263
264         Reviewed by Eric Siedel.
265
266         new-run-webkit-tests: update port/base and port/webkit to use the
267         new FileSystem and Config abstractions, pulling more logic out of
268         the base Port classes into separate, mockable objects.
269
270         Also create a MockFileSystem object, a MockConfig object, move
271         MockExecutive into common/system to be next to executive, and
272         update the config object to use a FileSystem.
273
274         https://bugs.webkit.org/show_bug.cgi?id=48280
275
276         * Scripts/webkitpy/common/newstringio.py: Added.
277         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
278         * Scripts/webkitpy/common/system/executive_mock.py: Added.
279         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
280         * Scripts/webkitpy/layout_tests/port/base.py:
281         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
282         * Scripts/webkitpy/layout_tests/port/config.py:
283         * Scripts/webkitpy/layout_tests/port/config_mock.py:
284         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
285         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
286         * Scripts/webkitpy/layout_tests/port/webkit.py:
287         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
288         * Scripts/webkitpy/tool/mocktool.py:
289
290 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
291
292         Unreviewed, build breakage.
293
294         Apparently I uploaded the wrong version of the file to fix 49122
295         and neither Eric or I noticed - it was missing a dirname() call.
296         Fixing ...
297
298         https://bugs.webkit.org/show_bug.cgi?id=49122
299
300         * Scripts/webkitpy/common/checkout/scm.py:
301
302 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
303
304         Reviewed by Eric Seidel.
305
306         webkitpy/tool/* unittests change cwd and don't clean up properly
307
308         https://bugs.webkit.org/show_bug.cgi?id=49122
309
310         * Scripts/webkitpy/common/checkout/scm.py:
311
312 2010-11-05  Chris Marrin  <cmarrin@apple.com>
313
314         Reviewed by Simon Fraser.
315
316         suspendAnimations/resumeAnimations not present in WebKit2
317         https://bugs.webkit.org/show_bug.cgi?id=49109
318
319         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
320         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
321         (WTR::LayoutTestController::suspendAnimations):
322         (WTR::LayoutTestController::resumeAnimations):
323         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
324
325 2010-11-05  Tony Chang  <tony@chromium.org>
326
327         Reviewed by David Levin.
328
329         cleanup style in TestNetscapePlugIn/PluginObject.cpp
330         https://bugs.webkit.org/show_bug.cgi?id=49044
331
332         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
333         (getPluginClass):
334         (pluginGetProperty):
335         (pluginSetProperty):
336         (variantToIdentifier):
337         (testCallback):
338         (getURL):
339         (getURLNotify):
340         (testInvokeDefault):
341         (testGetProperty):
342         (testDocumentOpen):
343         (testWindowOpen):
344         (pluginInvoke):
345         (handleCallback):
346         (notifyStream):
347         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
348
349 2010-11-05  Eric Seidel  <eric@webkit.org>
350
351         Reviewed by David Levin.
352
353         Add QueueStatusServer/__init__.py so others can run the QueueStatusServer tests
354         https://bugs.webkit.org/show_bug.cgi?id=49032
355
356         I wrote this file as part of bug 47847, but I forgot to commit it.
357         No one else noticed it missing because test-webkitpy knows how
358         to recover in the case where it can't import QueueStatusServer
359         (which generally occurs due to not having installed the AppEngine SDK).
360
361         * QueueStatusServer/__init__.py: Added.
362         * QueueStatusServer/model/workitems_unittest.py:
363          - Remove a test which fails.  This was probably landed (by me)
364            from my other machine, which since this __init__.py was missing
365            I never noticed the failure and landed this invalid test.
366            Sadly we can't really test remove_work_item as it depends
367            on .key() working.  .key() will throw unless the object
368            has already been saved it seems.
369            This may be a mis-design in our remove_work_item implementation,
370            but for now, just removing the test.
371
372 2010-11-04  Adam Roben  <aroben@apple.com>
373
374         Teach check-webkit-style about TestNetscapePlugIn
375
376         Fixes <http://webkit.org/b/49030> check-webkit-style is ignorant of
377         TestNetscapePlugIn's coding conventions
378
379         Reviewed by John Sullivan.
380
381         * Scripts/webkitpy/style/checker.py: Lump TestNetscapePlugIn in with
382         WebKitAPITest and TestWebKitAPI in having different include order and
383         naming conventions than WebCore. Added some comments about why the
384         exceptions exist.
385
386 2010-11-04  Adam Roben  <aroben@apple.com>
387
388         Add a test that shows that windowed plugins are able to paint
389
390         Somehow we never had a test for this before.
391
392         Fixes <http://webkit.org/b/49029> Should add a test that shows
393         windowed plugins can paint
394
395         Reviewed by Jon Honeycutt.
396
397         * DumpRenderTree/TestNetscapePlugIn/Tests/win/DrawsGradient.cpp: Added.
398         (DrawsGradient::DrawsGradient):
399         (DrawsGradient::wndProc): We handle the WM_PAINT and WM_PRINTCLIENT messages.
400
401         (DrawsGradient::onPaint):
402         (DrawsGradient::onPrintClient):
403         These both just call through to paint.
404
405         (DrawsGradient::paint): Fills our client area with some gradients.
406
407         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Link
408         against Msimg32.lib for ::GradientFill and added DrawsGradient.
409
410         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
411         (createBitmapContextFromWebView): Use WM_PRINT instead of
412         WM_PRINTCLIENT so that ::DefWindowProc will send
413         WM_PRINT/WM_PRINTCLIENT messages to the WebView's child windows.
414         Replaced kCGImageAlphaPremultipledFirst with kCGImageAlphaNoneSkipFirst
415         because GDI doesn't support alpha and kCGImageBitmapOrder32Little with
416         kCGImageBitmapOrder32Host because, who knows, maybe someday Windows
417         will run on a big-endian platform.
418
419 2010-11-04  Adam Roben  <aroben@apple.com>
420
421         Extract much of NPNInvalidateRectInvalidatesWindow's code into a
422         WindowedPluginTest base class
423
424         The base class takes care of subclassing the plugin's window so that a
425         custom WNDPROC is called. This will make it easier to write tests that
426         need to handle window messages.
427
428         Fixes <http://webkit.org/b/49028> It's hard to write a PluginTest with
429         a custom WNDPROC
430
431         Reviewed by Jon Honeycutt.
432
433         * DumpRenderTree/TestNetscapePlugIn/Tests/win/NPNInvalidateRectInvalidatesWindow.cpp:
434         (NPNInvalidateRectInvalidatesWindow::NPNInvalidateRectInvalidatesWindow):
435         (NPNInvalidateRectInvalidatesWindow::NPP_SetWindow):
436         (NPNInvalidateRectInvalidatesWindow::wndProc):
437         (NPNInvalidateRectInvalidatesWindow::testInvalidateRect):
438         Moved code from here to WindowedPluginTest. Changed to use window()
439         instead of m_window.
440
441         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
442         Added WindowedPluginTest and added TestNetscapePlugIn/win to the
443         include path. Also reordered the include path to make a little more
444         sense and simplified the entry that added TestNetscapePlugIn itself to
445         the include path.
446
447         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.cpp: Added.
448         (WindowedPluginTest::WindowedPluginTest):
449         (WindowedPluginTest::NPP_SetWindow):
450         (WindowedPluginTest::staticWndProc):
451         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.h: Added.
452         (WindowedPluginTest::window):
453         Code came from NPNInvalidateRectInvalidatesWindow.
454
455 2010-11-04  Adam Roben  <aroben@apple.com>
456
457         Add a plugin test to show that windowed plugins are clipped correctly
458
459         Fixes <http://webkit.org/b/49024> <rdar://problem/8487847> Windowed
460         plugins aren't clipped in WebKit2 on Windows
461
462         Reviewed by Jon Honeycutt.
463
464         * DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowRegionIsSetToClipRect.cpp: Added.
465         (WindowRegionIsSetToClipRect::WindowRegionIsSetToClipRect): Initialize members.
466         (WindowRegionIsSetToClipRect::NPP_SetWindow): Check that our window
467         region matches the clip rect we know we should have based on
468         window-region-is-set-to-clip-rect.html, and check that our window class
469         doesn't have the CS_PARENTDC style.
470
471         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
472         Added WindowRegionIsSetToClipRect.
473
474 2010-11-05  Alejandro G. Castro  <alex@igalia.com>
475
476         Reviewed by Martin Robinson.
477
478         [GTK] Avoid font hinting in the DRT
479         https://bugs.webkit.org/show_bug.cgi?id=48548
480
481         Change the settings to avoid font hinting, it was causing
482         different results depending on the distribution.
483
484         * DumpRenderTree/gtk/DumpRenderTree.cpp:
485         (initializeGtkFontSettings):
486         * DumpRenderTree/gtk/fonts/fonts.conf:
487
488 2010-11-05  Eric Seidel  <eric@webkit.org>
489
490         Reviewed by Adam Barth.
491
492         build-webkit spams system.log by repeatedly checking for PBXProductDirectory in com.apple.Xcode
493         https://bugs.webkit.org/show_bug.cgi?id=49051
494
495         This is a speculative fix.  The unit tests cover these methods, however
496         I don't know if this will fully stop the system.log spam.
497
498         * Scripts/webkitpy/layout_tests/port/base.py:
499         * Scripts/webkitpy/layout_tests/port/webkit.py:
500
501 2010-11-05  Eric Seidel  <eric@webkit.org>
502
503         Reviewed by Adam Barth.
504
505         Add basic support for showing bot id on /queue-status/ pages
506         https://bugs.webkit.org/show_bug.cgi?id=49037
507
508         This support is really simple.  Eventually we'll want to
509         show the bot id in the lock table too, but we don't have
510         that information stored in the server yet.
511
512         * QueueStatusServer/handlers/queuestatus.py:
513         * QueueStatusServer/handlers/queuestatus_unittest.py: Added.
514         * QueueStatusServer/templates/includes/singlequeuestatus.html:
515
516 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
517
518         Reviewed by Adam Barth.
519
520         Create a filesystem wrapper that we can use to enforce
521         particular conventions and use for mocking and dependency
522         injection down the line.
523
524         https://bugs.webkit.org/show_bug.cgi?id=48144
525
526         * Scripts/webkitpy/common/system/filesystem.py: Added.
527         * Scripts/webkitpy/common/system/filesystem_unittest.py: Added.
528
529 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
530
531         Reviewed by Tony Chang.
532
533         run_webkit_tests_unittest fails under Python 2.5
534         https://bugs.webkit.org/show_bug.cgi?id=49043
535         
536         Switch from itertools.chain.from_iterable (which was added in 2.6)
537         to using itertools.chain directly.
538
539         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
540
541 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
542
543         Unreviewed fix to webkit-patch rebaseline-server so that it runs under
544         Python 2.5 (it needs an import to use the with statement).
545
546         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
547
548 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
549
550         Reviewed by Adam Barth.
551
552         new-run-webkit-tests: split out webkit-specific configuration stuff into a new module
553
554         The current NRWT code has webkit-specific configuration code (like
555         _script_path, default configuration, etc.) mixed in with
556         layout-test-specific stuff in port/base. The configuration code
557         should be split out into a separate module for easier mocking,
558         testing, and isolation.
559
560         https://bugs.webkit.org/show_bug.cgi?id=48264
561
562         * Scripts/webkitpy/layout_tests/port/config.py: Added.
563         * Scripts/webkitpy/layout_tests/port/config_unittest.py: Added.
564
565 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
566
567         Reviewed by Tony Chang.
568
569         Rebaseline server: initial framework
570         https://bugs.webkit.org/show_bug.cgi?id=48892
571         
572         Adds the basic framework for the rebaseline server (details at 
573         http://webkit.org/b/47761). Includes the rebaseline-server webkit-patch
574         command, which starts an HTTP server that can serve static files or
575         invoke handler methods on a class.
576
577         * Scripts/webkitpy/tool/commands/__init__.py:
578         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html: Added.
579         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css: Added.
580         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js: Added.
581         * Scripts/webkitpy/tool/commands/rebaselineserver.py: Added.
582
583 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
584
585         Reviewed by James Robinson.
586
587         new-run-webkit-tests wasn't using DRT by default for
588         --platform chromium-gpu
589
590         The default value was set to False instead of None, which meant
591         that the platform specific logic wasn't firing to change the
592         value to True (b/c we were afraid we'd be overriding the user
593         preference).
594
595         https://bugs.webkit.org/show_bug.cgi?id=49038
596
597         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
598
599 2010-11-04  Tony Chang  <tony@chromium.org>
600
601         Reviewed by Adam Barth.
602
603         make platform/chromium/plugins/multiple-plugins.html pass on all platforms
604         https://bugs.webkit.org/show_bug.cgi?id=49026
605
606         Have the testObject.bar property return the string "bar".  This lets
607         us run and pass platform/chromium/plugins/multiple-plugins.html on all
608         platforms.
609
610         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
611         (testGetProperty):
612
613 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
614
615         Reviewed by Tony Chang.
616
617         NRWT doesn't have good test coverage for --run-chunk, --batch-size, --run-part, etc.
618         https://bugs.webkit.org/show_bug.cgi?id=48878
619         
620         Add get_tests_run so that it's easy to see which tests get run (and with
621         what batching) for a given flag combination. Flesh out the various
622         test cases that have FIXMEs.
623         
624         Also fixes an off-by-one error (batch sizes were one larger than 
625         expected) and makes --run-part also have wraparound behavior, like
626         --run-chunk.
627
628         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
629         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
630         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
631
632 2010-11-04  Erik Arvidsson  <arv@chromium.org>
633
634         Reviewed by Dimitri Glazkov.
635
636         Support box-sizing without the vendor prefix
637         https://bugs.webkit.org/show_bug.cgi?id=36713
638
639         Based on patch by Peter Beverloo <peter@lvp-media.com>
640
641         * iExploder/htdocs/cssproperties.in: Change -webkit-box-sizing to box-sizing.
642
643 2010-11-04  Csaba Osztrogonác  <ossy@webkit.org>
644
645         Unreviewed rollout r71340, because it broke Chromium Windows bot.
646
647         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
648         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
649
650 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
651
652         Reviewed by Ojan Vafai.
653
654         Make http locking default in NRWT.
655         https://bugs.webkit.org/show_bug.cgi?id=48053
656
657         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
658         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
659
660 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
661
662         Reviewed by Eric Seidel.
663
664         [NRWT] Clear invalid http locks on Windows platform as well
665         https://bugs.webkit.org/show_bug.cgi?id=48515
666
667         * Scripts/webkitpy/common/system/executive.py:
668         * Scripts/webkitpy/common/system/executive_unittest.py:
669         * Scripts/webkitpy/layout_tests/port/http_lock.py:
670
671 2010-11-02  Adam Roben  <aroben@apple.com>
672
673         Reduce our dependence on coreutils when running Python tests
674
675         This patch introduces versions of the standard echo and cat utilities
676         implemented in Python. They are probably missing features of their
677         coreutils equivalents, but they can do what's necessary for our Python
678         tests. This is useful on Windows, where these utilities typically
679         aren't available.
680
681         Fixes <http://webkit.org/b/48883> executive_unittest relies on echo
682         and cat utilities from coreutils, which are not present on Windows
683
684         Reviewed by Eric Seidel.
685
686         * Scripts/webkitpy/common/system/executive_unittest.py: Changed to use
687         our Python-based echo and cat.
688
689         * Scripts/webkitpy/common/system/fileutils.py: Added.
690         (make_stdout_binary): On Windows, puts sys.stdout into binary mode so
691         that \n won't be translated into \r\n. I couldn't think of a good way
692         to test this directly without touching the filesystem, but it is tested
693         indirectly by echo_unittest.
694
695         * Scripts/webkitpy/test/cat.py: Added.
696         (command_arguments): Returns a list for invoking cat with the given arguments.
697         (main): Acts like a simplified version of the coreutils cat utility.
698
699         * Scripts/webkitpy/test/cat_unittest.py: Added.
700         (CatTest.assert_cat): Runs cat with the given input and ensures the
701         output matches the input.
702         (CatTest.test_basic): Performs a simple test of cat.
703         (CatTest.test_no_newline): Tests what happens when the input string
704         doesn't have a trailing newline.
705         (CatTest.test_unicode): Tests passing a unicode string to cat.
706         (CatTest.test_as_command): Tests running cat as a separate command.
707
708         * Scripts/webkitpy/test/echo.py: Added.
709         (command_arguments): Returns a list for invoking echo with the given arguments.
710         (main): Acts like a simplified version of the coreutils echo utility.
711
712         * Scripts/webkitpy/test/echo_unittest.py: Added.
713         (EchoTest.test_basic): Performs a simple test of echo.
714         (EchoTest.test_no_newline): Tests passing -n to echo to suppress the
715         trailing newline.
716         (EchoTest.test_unicode): Tests passing unicode and non-unicode strings
717         to echo.
718         (EchoTest.test_argument_order): Tests what happens when -n is not the
719         first argument.
720         (EchoTest.test_empty_arguments): Tests what happens when you pass [] to
721         echo.main.
722         (EchoTest.test_no_arguments): Tests what happens when you call
723         echo.main with no arguments.
724         (EchoTest.test_as_command): Tests running echo as a separate command.
725
726 2010-11-04  Renata Hodovan  <reni@inf.u-szeged.hu>
727
728         Unreviewed: Add myself to the list of Committers.
729
730         * Scripts/webkitpy/common/config/committers.py:
731
732 2010-11-04  Andreas Kling  <kling@webkit.org>
733
734         Reviewed by Kenneth Rohde Christiansen.
735
736         [Qt] Hook into QWebPage::scrollRequested for resizes-to-contents mode
737
738         In RTC mode the QGraphicsWebView item is the size of the contents,
739         scrolling works a bit differently  (we need to react to scrollRequested.)
740
741         Normally QGraphicsView will replay the last mouse event when scrolling,
742         so to prevent WebKit from getting confused by this we temporarily make
743         the QGraphicsView non-interactive.
744
745         * QtTestBrowser/webview.cpp:
746         (WebViewGraphicsBased::setPage):
747         (WebViewGraphicsBased::scrollRequested):
748         * QtTestBrowser/webview.h:
749
750 2010-11-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
751
752         Reviewed by Andreas Kling.
753
754         Use OS(MAC_OS_X) rather than PLATFORM(MAC) in TestControllerQt
755
756         PLATFORM(MAC) is not defined for the Qt port, as it refers to
757         the Mac-port, not the Mac OS X operating system.
758
759         * WebKitTestRunner/qt/TestControllerQt.cpp:
760         (WTR::isExistingLibrary):
761
762 2010-11-04  Leonid Ebril  <leonid.ebril@nokia.com>
763
764         Reviewed by Andreas Kling.
765
766         [Qt] Adding iphone user agent string the user agent list for QtTestBrowser
767         https://bugs.webkit.org/show_bug.cgi?id=48869
768
769         * QtTestBrowser/useragentlist.txt:
770
771 2010-11-03  Adam Roben  <aroben@apple.com>
772
773         Always use uppercase drive names in strings returned by abspath_to_uri
774
775         Some versions of cygpath use lowercase drive letters while others use
776         uppercase, which makes it hard to test the output of code that uses
777         cygpath.
778
779         Fixes <http://webkit.org/b/48914> webkitpy.common.system.path_unittest
780         fails with Cygwin 1.5
781
782         Reviewed by Eric Seidel.
783
784         * Scripts/webkitpy/common/system/path.py:
785         (cygpath): Updated the docstring to indicate that only absolute paths
786         should be passed for now (though relative paths will work fine).
787         (_Cygpath.convert): Upper-case the first letter of the converted Windows path.
788
789 2010-11-03  George Guo  <George.Guo@Nokia.com>
790
791         Reviewed by Andreas Kling.
792
793         [Qt] QtTestBrowser : set mmp rule pageddata in Symbian
794         https://bugs.webkit.org/show_bug.cgi?id=48767
795     
796         Paging is needd on Symbian devices to support benchmarks tests like
797         dromaeo.com and Celtic Kane that need a lot of memory to run  
798
799         * QtTestBrowser/QtTestBrowser.pro:
800
801 2010-11-03  Jenn Braithwaite  <jennb@chromium.org>
802
803         Reviewed by Dmitry Titov.
804
805         Chromium: Update resource tracking when moving a frame between documents
806         https://bugs.webkit.org/show_bug.cgi?id=48363
807
808         * DumpRenderTree/chromium/WebViewHost.cpp:
809         (WebViewHost::assignIdentifierToRequest):
810         Always put resource id in map so we can make assumptions about its
811         presence.
812         (WebViewHost::removeIdentifierForRequest):
813         Added.
814         (WebViewHost::didFinishResourceLoad):
815         (WebViewHost::didFailResourceLoad):
816         Check existence of resource id before removing from map.
817         * DumpRenderTree/chromium/WebViewHost.h:
818
819 2010-11-03  Victor Wang  <victorw@chromium.org>
820
821         Reviewed by Adam Barth.
822
823         [Chromium] update buildbot names in chromium rebaseline tool.
824
825         https://bugs.webkit.org/show_bug.cgi?id=48881
826
827         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
828
829 2010-11-03  Darin Adler  <darin@apple.com>
830
831         Updated Xcode projects by opening them with Xcode 3.2.4.
832         Updated svn:ignore for Xcode projects.
833
834         * MiniBrowser/MiniBrowser.xcodeproj: Added property svn:ignore.
835         * TestWebKitAPI/TestWebKitAPI.xcodeproj: Added property svn:ignore.
836         * WebKitLauncher/WebKitLauncher.xcodeproj: Modified property svn:ignore.
837         * WebKitLauncher/WebKitLauncher.xcodeproj/project.pbxproj: Updated with Xcode 3.2.4.
838
839 2010-11-03  Adam Roben  <aroben@apple.com>
840
841         Stop waiting for messages from the web process after a timeout elapses
842
843         TestController::runUntil now takes a timeout parameter that specifies
844         how long to wait for the condition to become true. WebKitTestRunner
845         takes a --timeout flag that specifies how long the timeout should be.
846         run-webkit-tests passes this flag to WebKitTestRunner so its timeout
847         will be similar to run-webkit-tests.
848
849         Fixes <http://webkit.org/b/43047> <rdar://problem/8365833>
850         WebKitTestRunner waits forever if the web process crashes
851
852         Reviewed by Darin Adler and Anders Carlsson.
853
854         * Scripts/old-run-webkit-tests:
855         (top level): Moved the GuardMalloc timeout adjustment here from
856         readFromDumpToolWithTimer.
857         (openDumpTool): Make WTR use a timeout similar to but slightly shorter
858         than the one that was specified on the command line.
859
860         * WebKitTestRunner/TestController.cpp:
861         (WTR::TestController::TestController): Initialize our timeout values.
862         (WTR::TestController::initialize): Parse the --timeout flag and use it
863         to modify our timeout values.
864         (WTR::TestController::resetStateToConsistentValues): Changed to use a
865         short timeout while waiting for the web process to reset and to return
866         a boolean indicating whether we were able to reset the web process.
867         Uses a 5-second timeout while waiting for the process to be reset.
868         (WTR::TestController::runTest): Changed to return a boolean indicating
869         whether we were able to reset the web process (and thus run the test).
870         (WTR::TestController::runUntil): Call through to platformRunUntil.
871
872         (WTR::TestController::runTestingServerLoop):
873         (WTR::TestController::run):
874         Changed to bail if any test can't be run. This will cause the process
875         to exit. (Unfortunately this will make run-webkit-tests think we
876         crashed; see <http://webkit.org/b/48943>.)
877
878         * WebKitTestRunner/TestController.h: Added platformRunUntil,
879         m_longTimeout, and m_shortTimeout.
880
881         * WebKitTestRunner/TestInvocation.cpp:
882         (WTR::TestInvocation::invoke): Use a short timeout when waiting for the
883         initial response and a long timeout when waiting for the test to
884         complete. Check whether runUntil timed out and print an error message
885         if so.
886
887         * WebKitTestRunner/mac/TestControllerMac.mm:
888         (WTR::TestController::platformRunUntil): Renamed from runUntil. Pass
889         [NSDate distantPast] to -[NSRunLoop runMode:beforeDate:] so that we
890         won't block waiting for the run loop. Only loop until the timeout
891         elapses.
892
893         * WebKitTestRunner/qt/TestControllerQt.cpp:
894         (WTR::TestController::platformRunUntil): Renamed from runUntil. Added a
895         FIXME about honoring the timeout.
896
897         * WebKitTestRunner/win/TestControllerWin.cpp:
898         (WTR::TestController::platformRunUntil): Renamed from runUntil. Use
899         ::MsgWaitForMultipleObjectsEx to implement the timeout. Changed to use
900         ::PeekMessageW so that we don't block waiting for messages to become
901         available.
902
903 2010-11-03  Adam Roben  <aroben@apple.com>
904
905         Add a plugin test that evaluates JS after removing the plugin element
906         from the document
907
908         This test replaces platform/win/plugins/plugin-delayed-destroy.html.
909         That test was made to prevent a crash very similar to this one, but
910         unfortunately tested only the mechanism that prevented the crash and
911         not whether the crash itself was prevented. Since WebKit2 uses a
912         different mechanism to prevent the crash, the test was failing even
913         though WebKit2 was not vulnerable to the crash. This new test crashes
914         if there is no mechanism in place to prevent it and passes in both
915         WebKit1 and WebKit2.
916
917         Fixes <http://webkit.org/b/46711> <rdar://problem/8485903>
918         platform/win/plugins/plugin-delayed-destroy.html fails in WebKit2
919
920         Reviewed by Anders Carlsson.
921
922         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
923         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
924         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
925         * GNUmakefile.am:
926         Added new file.
927
928         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
929         (pluginDeallocate): Make sure we delete the PluginTest object. This
930         prevents a leak and also allows us to test the crash.
931
932         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
933         (PluginTest::executeScript): Made this into a non-static member
934         function.
935
936         (PluginTest::waitUntilDone):
937         (PluginTest::notifyDone):
938         Updated for changes to executeScript.
939
940         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Added executeScript.
941
942         * DumpRenderTree/TestNetscapePlugIn/Tests/EvaluateJSAfterRemovingPluginElement.cpp: Added.
943         (EvaluateJSAfterRemovingPluginElement::EvaluateJSAfterRemovingPluginElement):
944         Initialize ourselves and tell the test harness to wait.
945         (EvaluateJSAfterRemovingPluginElement::NPP_DestroyStream): Remove our
946         plugin element from the document, then execute some JavaScript. If
947         WebKit does not have appropriate mechanisms in place, we'll be
948         destroyed inside the first call to executeScript and crash on the
949         second call.
950
951 2010-11-02  Stephen White  <senorblanco@chromium.org>
952
953         Reviewed by Tony Chang.
954
955         [chromium] Fix LayoutTestController UMRs.
956         https://bugs.webkit.org/show_bug.cgi?id=48872
957
958         * DumpRenderTree/chromium/LayoutTestController.cpp:
959         (LayoutTestController::LayoutTestController):
960
961 2010-11-03  Kent Tamura  <tkent@chromium.org>
962
963         Reviewed by Dimitri Glazkov.
964
965         REGRESSION: rebaseline-chromium-webkit-tests uses non-zero tolerance for
966         image dup detection
967         https://bugs.webkit.org/show_bug.cgi?id=48744
968
969         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
970          - Make a function for option parsing for ease of test
971          - Set 0 to options.tolerance
972         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
973          - Add a test for this change
974
975 2010-11-02  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
976
977         Unreviewed: Add myself to the list of Committers.
978
979         * Scripts/webkitpy/common/config/committers.py:
980
981 2010-11-02  Anders Carlsson  <andersca@apple.com>
982
983         Fix build.
984
985         * TestWebKitAPI/PlatformUtilities.cpp:
986         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
987
988 2010-11-02  Sam Weinig  <sam@webkit.org>
989
990         Reviewed by Anders Carlsson.
991
992         Add a way to send startup messages on the context which can be posted when a process launches
993         <rdar://problem/8617928>
994         https://bugs.webkit.org/show_bug.cgi?id=48838
995
996         * MiniBrowser/mac/WebBundle/WebBundleMain.m:
997         (WKBundleInitialize):
998         * TestWebKitAPI/InjectedBundleController.cpp:
999         (TestWebKitAPI::InjectedBundleController::initialize):
1000         (TestWebKitAPI::InjectedBundleController::didReceiveMessage):
1001         * TestWebKitAPI/InjectedBundleController.h:
1002         * TestWebKitAPI/InjectedBundleMain.cpp:
1003         (WKBundleInitialize):
1004         * TestWebKitAPI/PlatformUtilities.cpp:
1005         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
1006         * WebKitTestRunner/InjectedBundle/InjectedBundleMain.cpp:
1007         (WKBundleInitialize):
1008         Update implementations of WKBundleInitialize to take an initial userData
1009         argument.  Change TestWebKitAPI to use the new initial userData to initialize
1010         each test's bundle.
1011
1012 2010-11-02  Benjamin Kalman  <kalman@google.com>
1013
1014         Reviewed by Ojan Vafai.
1015
1016         new-run-webkit-tests doesn't strip "LayoutTests/" from prefix, unlike old-run-webkit-tests
1017         https://bugs.webkit.org/show_bug.cgi?id=48794
1018
1019         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: Strip the "LayoutTests/" prefix from test argument paths.
1020
1021 2010-11-02  Adam Roben  <aroben@apple.com>
1022
1023         Skip webkitpy.layout_tests.run_webkit_tests_unittest.MainTest on Cygwin
1024         Python 2.5.x
1025
1026         It is known to hang on that version of Python. See
1027         <http://webkit.org/b/48614>.
1028
1029         Reviewed by Adam Barth.
1030
1031         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py: Use
1032         skip_if to skip MainTest on Cygwin Python 2.5.x.
1033
1034         * Scripts/webkitpy/test/skip.py: Added.
1035         (skip_if): If the passed-in condition is false, find all the test_*
1036         methods of the given class and replace them with a function that just
1037         logs that we're skipping these tests. This is loosely based on the
1038         unittest.skip_if decorator added in Python 3.1.
1039         (_skipped_method): Returns a function that just logs that the tests are
1040         being skipped. This is used in place of the actual test_* functions.
1041
1042         * Scripts/webkitpy/test/skip_unittest.py: Added.
1043         (SkipTest.setUp): Create a logger for ourselves and give it a handler
1044         that logs to a private stream.
1045         (SkipTest.tearDown): Reset the logger.
1046         (SkipTest.create_fixture_class): Returns a class that we use to test
1047         skip_if. It has a single test_* method, test_foo, that just calls
1048         through to a callback.
1049         (SkipTest.foo_callback): Record that test_foo was called.
1050         (SkipTest.test_skip_if_false): Pass skip_if a False condition and test
1051         that test_foo does get called.
1052         (SkipTest.test_skip_if_true): Pass skip_if a True condition and test
1053         that test_foo does not get called and the appropriate message gets
1054         logged.
1055
1056 2010-11-02  Adam Barth  <abarth@webkit.org>
1057
1058         Reviewed by Eric Seidel.
1059
1060         webkit-patch should tell check-webkit-style which files were changed so
1061         check-webkit-style doesn't have to stat the whole working copy again
1062         https://bugs.webkit.org/show_bug.cgi?id=48792
1063
1064         * Scripts/webkitpy/tool/mocktool.py:
1065         * Scripts/webkitpy/tool/steps/checkstyle.py:
1066
1067 2010-11-02  Robert Kroeger  <rjkroege@chromium.org>
1068
1069         Reviewed by James Robinson.
1070
1071         [Chromium/DRT] Make EventSendingController honour leapForward for touch events.
1072         https://bugs.webkit.org/show_bug.cgi?id=48777
1073
1074         * DumpRenderTree/chromium/EventSender.cpp:
1075         (EventSender::sendCurrentTouchEvent):
1076
1077 2010-11-02  Adam Roben  <aroben@apple.com>
1078
1079         Only track resource identifiers in DRT when dumpResourceLoadCallbacks
1080         is on
1081
1082         This reverts Windows to our pre-r71097 behavior. That patch made us
1083         track all resource identifiers, including the main resource, so the
1084         main resource's URL started appearing in test output instead of
1085         "<unknown>". Arguably having the main resource's URL is better, but all
1086         other platforms print "<unknown>" and we want to match.
1087
1088         Fixes <http://webkit.org/b/48837> <rdar://problem/8620351> REGRESSION
1089         (r71097): Many http tests failing on Windows
1090
1091         Reviewed by Anders Carlsson.
1092
1093         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
1094         (ResourceLoadDelegate::identifierForInitialRequest): Don't add the
1095         identifier to the URL map if we're not supposed to dump resource load
1096         callbacks.
1097         (ResourceLoadDelegate::removeIdentifierForRequest): Always remove the
1098         identifier from the URL map even if we're already "done". There's no
1099         point in keeping out-of-date identifiers around.
1100
1101 2010-11-01  Jenn Braithwaite  <jennb@chromium.org>
1102
1103         Reviewed by Adam Roben.
1104
1105         Windows: Update resource tracking when moving a frame between documents
1106         https://bugs.webkit.org/show_bug.cgi?id=48364
1107
1108         * DumpRenderTree/win/DumpRenderTree.cpp:
1109         (createWebViewAndOffscreenWindow):
1110         (main):
1111         Give each WebView its own ResourceLoadDelegate instance in order to
1112         make assertions about resource ids on a particular WebView.
1113         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
1114         (ResourceLoadDelegate::identifierForInitialRequest):
1115         Always add id to the map.
1116         (ResourceLoadDelegate::removeIdentifierForRequest):
1117         Added.
1118         (ResourceLoadDelegate::willSendRequest):
1119         (ResourceLoadDelegate::didReceiveAuthenticationChallenge):
1120         (ResourceLoadDelegate::didReceiveResponse):
1121         (ResourceLoadDelegate::didFinishLoadingFromDataSource):
1122         (ResourceLoadDelegate::didFailLoadingWithError):
1123         (ResourceLoadDelegate::descriptionSuitableForTestResult):
1124         Replace static descriptionSuitableForTestResult with static member function to access identifier map.
1125         * DumpRenderTree/win/ResourceLoadDelegate.h:
1126         (ResourceLoadDelegate::urlMap):
1127         Moved within class so that each WebView has its own id map.
1128
1129 2010-11-01  Ojan Vafai  <ojan@chromium.org>
1130
1131         Reviewed by Dimitri Glazkov.
1132
1133         remove debug code from run_webkit_tests.py
1134         https://bugs.webkit.org/show_bug.cgi?id=48800
1135
1136         Remove temporary debug code and make --master-name required
1137         if --test-results-server is set now that all clients set
1138         --master-name.
1139
1140         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1141
1142 2010-11-01  Tony Chang  <tony@chromium.org>
1143
1144         Reviewed by Kent Tamura.
1145
1146         add plugin event logging for linux
1147         https://bugs.webkit.org/show_bug.cgi?id=48779
1148
1149         This is taken from Chromium's fork of the layout test plugin:
1150         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
1151
1152         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
1153         (handleEventX11):
1154         (NPP_HandleEvent):
1155
1156 2010-11-01  Dirk Pranke  <dpranke@chromium.org>
1157
1158         Reviewed by James Robinson.
1159
1160         new-run-webkit-tests: use DRT, child-processes=1 for GPU tests by default
1161         https://bugs.webkit.org/show_bug.cgi?id=48790
1162
1163         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
1164         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
1165
1166 2010-11-01  Adam Barth  <abarth@webkit.org>
1167
1168         Reviewed by Eric Seidel.
1169
1170         Teach check-webkit-style how to accept a list of files to diff on the
1171         command line
1172         https://bugs.webkit.org/show_bug.cgi?id=48784
1173
1174         In a future patch, webkit-patch will use this option to improve
1175         performance.  I'm landing this in two pieces to avoid causing a version
1176         skew problem for the style-bot.
1177
1178         * Scripts/check-webkit-style:
1179         * Scripts/webkitpy/style/optparser.py:
1180         * Scripts/webkitpy/style/optparser_unittest.py:
1181         * Scripts/webkitpy/style_references.py:
1182
1183 2010-11-01  Anders Carlsson  <andersca@apple.com>
1184
1185         Reviewed by John Sullivan.
1186
1187         Tear down the related WebProcessProxy when a WebContext is deallocated
1188         https://bugs.webkit.org/show_bug.cgi?id=48769
1189
1190         * TestWebKitAPI/Tests/WebKit2/FailedLoad.cpp:
1191         (TestWebKitAPI::didFailProvisionalLoadWithErrorForFrame):
1192         We don't support empty URLs anymore, update test to expect a null URL instead.
1193
1194 2010-11-01  Søren Gjesse  <sgjesse@chromium.org>
1195
1196         Reviewed by Andreas Kling.
1197
1198         Fix warning when compiling the chromium port of DumpRenderShell
1199         with clang.
1200         https://bugs.webkit.org/show_bug.cgi?id=48414
1201
1202         * DumpRenderTree/chromium/TestShell.h:
1203         (TestShell::javaScriptFlagsForLoad):
1204
1205 2010-11-01  Adam Roben  <aroben@apple.com>
1206
1207         Fix typo from r71022
1208
1209         * BuildSlaveSupport/build.webkit.org-config/config.json:
1210
1211 2010-11-01  Adam Roben  <aroben@apple.com>
1212
1213         Trigger the Windows Release WebKit2 tests when a Release build
1214         finishes, not when a Debug build finishes
1215
1216         Fixes <http://webkit.org/b/48754> Windows Release WebKit2 tests are
1217         triggered at the wrong time
1218
1219         Reviewed by Sam Weinig.
1220
1221         * BuildSlaveSupport/build.webkit.org-config/config.json: Fixed
1222         triggerable name and trigger.
1223
1224 2010-11-01  Mario Sanchez Prada  <msanchez@igalia.com>
1225
1226         Unreviewed. Adding my IRC nickname to the list of committers.
1227
1228         * Scripts/webkitpy/common/config/committers.py:
1229
1230 2010-10-31  Robert Hogan  <robert@webkit.org>
1231
1232         Reviewed by Antonio Gomes.
1233
1234         [Qt] [Gtk] Plug-ins having upper case in mime type are failing to load
1235
1236         Qt and Gtk are case-sensitive when storing the declared mime-type
1237         of plugins. Since plugin mime-types are lowercased prior to searching
1238         for them in the plugin database, ensure they are loaded with the
1239         mime-type in lower case too.
1240
1241         Change the test netscape plugin to declare its mimetype in sentence
1242         case so that the correct behaviour is enforced.
1243
1244         https://bugs.webkit.org/show_bug.cgi?id=36815
1245
1246         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
1247         (NP_GetMIMEDescription):
1248
1249 2010-10-31  Robert Hogan  <robert@webkit.org>
1250
1251         Reviewed by Antonio Gomes.
1252
1253         [Qt] Support nodesFromRect in DRT
1254
1255         https://bugs.webkit.org/show_bug.cgi?id=48716
1256
1257         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1258         (LayoutTestController::nodesFromRect):
1259         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1260
1261 2010-10-30  Andreas Kling  <kling@webkit.org>
1262
1263         Reviewed by Antonio Gomes.
1264
1265         [Qt] QtTestBrowser: Switching view type moves the embedded inspector
1266         https://bugs.webkit.org/show_bug.cgi?id=48705
1267
1268         Reinsert the embedded inspector into the splitter after changing
1269         between QWebView/QGraphicsWebView.
1270
1271         * QtTestBrowser/launcherwindow.cpp:
1272         (LauncherWindow::init):
1273         (LauncherWindow::initializeView):
1274         * QtTestBrowser/webinspector.h:
1275         (WebInspector::WebInspector):
1276
1277 2010-10-28  Antonio Gomes  <agomes@rim.com>
1278
1279         Reviewed by Ojan Vafai.
1280
1281         Needs a "LinuxEditingBehavior", perhaps with a better name
1282         https://bugs.webkit.org/show_bug.cgi?id=36627
1283
1284         Adding support to Mac's, GTK+'s, Windows' and Chromium's LayoutTestController class to test the newly introduced Unix editing behavior.
1285
1286         * DumpRenderTree/chromium/LayoutTestController.cpp:
1287         (LayoutTestController::setEditingBehavior):
1288         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1289         (LayoutTestController::setEditingBehavior):
1290         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1291         (LayoutTestController::setEditingBehavior):
1292         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1293         (LayoutTestController::setEditingBehavior):
1294
1295 2010-10-29  Andreas Kling  <kling@webkit.org>
1296
1297         Reviewed by Kenneth Rohde Christiansen.
1298
1299         [Qt] QtTestBrowser: Add keyboard shortcut to toggle full screen (F11)
1300         https://bugs.webkit.org/show_bug.cgi?id=48695
1301
1302         * QtTestBrowser/launcherwindow.cpp:
1303         (LauncherWindow::createChrome):
1304
1305 2010-10-29  Andreas Kling  <kling@webkit.org>
1306
1307         Reviewed by Kenneth Rohde Christiansen.
1308
1309         [Qt] QtTestBrowser: Fix uninitialized read in FpsTimer
1310         https://bugs.webkit.org/show_bug.cgi?id=48675
1311
1312         FpsTimer::m_timer was never initialized and passed to QObject::killTimer()
1313         on startup with in -graphicsbased mode.
1314
1315         * QtTestBrowser/fpstimer.cpp:
1316         (FpsTimer::FpsTimer):
1317         (FpsTimer::numFrames):
1318         (FpsTimer::stop):
1319         (FpsTimer::timerEvent):
1320         * QtTestBrowser/fpstimer.h:
1321
1322 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
1323
1324         Reviewed by Tony Chang.
1325
1326         new-run-webkit-tests: change TestResults to be serializable
1327
1328         In preparation for changing new-run-webkit-tests from
1329         multithreaded to multiprocess, we need to make sure the data
1330         going between the threads is easily serialized over a socket.
1331
1332         This change adds serialization/pickling for the TestResults and
1333         TestFailure objects (using cPickle).
1334
1335         The TestFailure objects included a "has_wdiff" flag for Text
1336         results, but the flag wasn't being used, so I've removed it,
1337         simplifying the state to basically a set of enum objects with
1338         associated methods.
1339
1340         https://bugs.webkit.org/show_bug.cgi?id=48616
1341
1342         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1343         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
1344         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
1345         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
1346         * Scripts/webkitpy/layout_tests/layout_package/test_results.py: Added.
1347         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:Added. 
1348         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1349         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
1350
1351 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
1352
1353         Reviewed by Ojan Vafai.
1354
1355         new-run-webkit-tests: need to provide separate dashboard results for GPU tests
1356         https://bugs.webkit.org/show_bug.cgi?id=48687
1357
1358         Modify the steps to upload the JSON files to the dashboards so
1359         that the GPU tests don't conflict with the regular tests on a
1360         port. We do this by modifying the --builder-name parameter to
1361         append " - GPU", which should cause the app to treat the results
1362         as a completely new builder. This is a little non-obvious, but
1363         keeps us from having to restructure the app.
1364
1365         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
1366         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
1367         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1368
1369 2010-10-29  Ojan Vafai  <ojan@chromium.org>
1370
1371         Reviewed by Tony Chang.
1372
1373         [chromium] add debug logging to help diagnose flakiness dashboard issues
1374         https://bugs.webkit.org/show_bug.cgi?id=48657
1375
1376         The appengine app thinks it's getting empty files uploaded. Add some logging
1377         to see if new-run-webkit-tests agrees.
1378         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1379
1380 2010-10-29  Dimitri Glazkov  <dglazkov@chromium.org>
1381
1382         Reviewed by Ojan Vafai.
1383
1384         [Chromium/DRT] Add master-name flag to new-run-webkit-tests.
1385         https://bugs.webkit.org/show_bug.cgi?id=48649
1386
1387         * BuildSlaveSupport/build.webkit.org-config/master.cfg: Added flag.
1388
1389 2010-10-29  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
1390
1391         Reviewed by Martin Robinson.
1392
1393         Enable popup window in GtkLauncher
1394         https://bugs.webkit.org/show_bug.cgi?id=48335
1395
1396         The GtkLauncher application does not support opening new window when
1397         a link with "target=_blank" is clicked or similar call to
1398         window.open(). Instead, GtkLauncher does nothing which breaks
1399         navigation of some websites.
1400
1401         * GtkLauncher/main.c:
1402         (activate_uri_entry_cb):
1403         (update_title):
1404         (link_hover_cb):
1405         (notify_title_cb):
1406         (notify_load_status_cb):
1407         (notify_progress_cb):
1408         (destroy_cb):
1409         (go_back_cb):
1410         (go_forward_cb):
1411         (create_web_view_cb):
1412         (web_view_ready_cb):
1413         (close_web_view_cb):
1414         (create_browser):
1415         (create_statusbar):
1416         (create_toolbar):
1417         (create_window):
1418         (main):
1419
1420 2010-10-29  Adam Roben  <aroben@apple.com>
1421
1422         Teach check-webkit-style about WebKit2's idiosyncracies
1423
1424         Fixes <http://webkit.org/b/48638> Style bot complains about a number
1425         of WebKit2 conventions
1426
1427         Reviewed by Anders Carlsson.
1428
1429         * Scripts/webkitpy/style/checker.py: Excluded some rules for various
1430         WebKit2-related files. Also updated the excluded rules for
1431         WebKitAPITest to match the current code.
1432
1433 2010-10-29  Csaba Osztrogonác  <ossy@webkit.org>
1434
1435         Reviewed by Adam Roben and David Kilzer.
1436
1437         Fix and cleanup of build systems
1438         https://bugs.webkit.org/show_bug.cgi?id=48342
1439
1440         * Scripts/build-webkit:
1441          - Remove unnecessary ENABLE_SANDBOX option.
1442          - Add ENABLE_FULLSCREEN_API option.
1443
1444 2010-10-28  Adam Roben  <aroben@apple.com>
1445
1446         Switch the Windows WebKit2 bot to the Release configuration
1447
1448         We only have one machine testing WebKit2 on Windows right now, and
1449         Debug is just too slow for it to keep up.
1450
1451         Fixes (hopefully!) <http://webkit.org/b/48615> Windows WebKit2 bot is
1452         always way behind
1453
1454         Reviewed by Jon Honeycutt.
1455
1456         * BuildSlaveSupport/build.webkit.org-config/config.json:
1457
1458 2010-10-29  Leandro Gracia Gil  <leandrogracia@google.com>
1459
1460         Reviewed by Jeremy Orlow.
1461
1462         Added a second parameter to setMockSpeechInputResult for
1463         the language used in speech input.
1464         https://bugs.webkit.org/show_bug.cgi?id=47089
1465
1466         * DumpRenderTree/LayoutTestController.cpp:
1467         (setMockSpeechInputResultCallback):
1468         * DumpRenderTree/LayoutTestController.h:
1469         * DumpRenderTree/chromium/LayoutTestController.cpp:
1470         (LayoutTestController::setMockSpeechInputResult):
1471         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1472         (LayoutTestController::setMockSpeechInputResult):
1473         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1474         (LayoutTestController::setMockSpeechInputResult):
1475         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1476         (LayoutTestController::setMockSpeechInputResult):
1477         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1478         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1479         (LayoutTestController::setMockSpeechInputResult):
1480         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
1481         (LayoutTestController::setMockSpeechInputResult):
1482         (LayoutTestController::pageSizeAndMarginsInPixels):
1483
1484 2010-10-28  Ojan Vafai  <ojan@chromium.org>
1485
1486         Reviewed by Tony Chang.
1487
1488         [chromium] add a result-small.json file for the test dashboard
1489         https://bugs.webkit.org/show_bug.cgi?id=48547
1490
1491         Output both a results.json file and a results-small.json file.
1492         The dashboard will load results-small.json by default so it loads faster.
1493
1494         * TestResultServer/model/jsonresults.py:
1495         * TestResultServer/model/jsonresults_unittest.py:
1496         Added a bunch of sys.path hackery. Unforunately, this uses hardcoded
1497         paths. That obviously needs to be fixed, but at least this way it
1498         clearly documents what paths are necessary.
1499
1500 2010-10-28  Eric Seidel  <eric@webkit.org>
1501
1502         Reviewed by Adam Barth.
1503
1504         webkit-patch upload calls changed_files more often than it should
1505         https://bugs.webkit.org/show_bug.cgi?id=48567
1506
1507         Passing changed_files around everywhere isn't a very elegant solution
1508         but it's the one we have for the moment.  I think keeping an explicit
1509         cache on Checkout (or making StepState() a real class) is a better
1510         long-term option.
1511
1512         Previously bug_id_for_this_commit was calling changed_files and the
1513         result was never getting cached on the state.  Now we're explicitly
1514         caching the result on the state and passing that to the bug_id_for_this_commit call.
1515
1516         I looked into building unit tests for this.  Doing so would require
1517         using a real Checkout object with a MockSCM and overriding the appropriate
1518         calls on SCM to count how often we're stating the file system.
1519         That's a useful set of tests to build for a separate change.
1520
1521         * Scripts/webkitpy/common/checkout/api.py:
1522         * Scripts/webkitpy/tool/commands/download.py:
1523         * Scripts/webkitpy/tool/commands/upload.py:
1524         * Scripts/webkitpy/tool/mocktool.py:
1525
1526 2010-10-28  Eric Seidel  <eric@webkit.org>
1527
1528         Reviewed by Adam Barth.
1529
1530         Make suggest-reviewers slightly faster
1531         https://bugs.webkit.org/show_bug.cgi?id=48562
1532
1533         Add @memoized to one more common call.
1534
1535         * Scripts/webkitpy/common/checkout/api.py:
1536
1537 2010-10-28  Eric Seidel  <eric@webkit.org>
1538
1539         Reviewed by Adam Barth.
1540
1541         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is still very slow
1542         https://bugs.webkit.org/show_bug.cgi?id=48536
1543
1544         This patch makes it about 40 seconds faster, but it still
1545         takes 1:40.  This will require more refinement.
1546
1547         The suggested reviewers list appears to be the same.
1548
1549         I think the next step may be to have it stop the search after
1550         5 reviewers are found.  We never want to suggest 30 people.
1551
1552         * Scripts/webkitpy/common/checkout/scm.py:
1553          - Using --remove-empty to theoretically stop lookups past
1554            when a file is removed.  I'm not entirely clear that the option
1555            does what it says it does.  Example:
1556            git log --pretty=format:%H -5 --remove-empty -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
1557            returns only one commit
1558            vs.
1559            git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
1560            which returns 5.  I was not aware that wx files were ever removed from the repository?
1561
1562 2010-10-28  Kinuko Yasuda  <kinuko@chromium.org>
1563
1564         Reviewed by David Levin.
1565
1566         [Chromium] Support FileSystem in chromium DRT
1567         https://bugs.webkit.org/show_bug.cgi?id=47643
1568
1569         * DumpRenderTree/chromium/WebViewHost.cpp:
1570         (WebViewHost::openFileSystem): Added.
1571         * DumpRenderTree/chromium/WebViewHost.h:
1572         (WebViewHost::openFileSystem): Added.
1573
1574 2010-10-28  Sam Weinig  <sam@webkit.org>
1575
1576         Reviewed by Anders Carlsson.
1577
1578         WKURLRefs should be allowed to be null
1579         <rdar://problem/8575621>
1580         https://bugs.webkit.org/show_bug.cgi?id=48535
1581
1582         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
1583         * TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp:
1584         (TestWebKitAPI::didStartProvisionalLoadForFrame):
1585         (TestWebKitAPI::didCommitLoadForFrame):
1586         (TestWebKitAPI::didFinishLoadForFrame):
1587         Test that URLs are null pointers when unset.
1588
1589 2010-10-28  Chris Fleizach  <cfleizach@apple.com>
1590
1591         Reviewed by Adele Peterson.
1592
1593         AX: multi select group option does not handle setting of AXSelectedChildren correctly
1594         https://bugs.webkit.org/show_bug.cgi?id=48464
1595
1596         Add support for querying information about selected children to DRT, including:
1597            selectedChildAtIndex
1598            selectedChildrenCount
1599            setSelectedChild
1600
1601         * DumpRenderTree/AccessibilityUIElement.cpp:
1602         (selectedChildAtIndexCallback):
1603         (setSelectedChildCallback):
1604         (selectedChildrenCountCallback):
1605         (AccessibilityUIElement::setSelectedChild):
1606         (AccessibilityUIElement::selectedChildrenCount):
1607         (AccessibilityUIElement::selectedChildAtIndex):
1608         (AccessibilityUIElement::getJSClass):
1609         * DumpRenderTree/AccessibilityUIElement.h:
1610         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
1611         (AccessibilityUIElement::selectedChildAtIndex):
1612         (AccessibilityUIElement::selectedChildrenCount):
1613         (AccessibilityUIElement::setSelectedChild):
1614
1615 2010-10-27  Sam Weinig  <sam@webkit.org>
1616
1617         Reviewed by Anders Carlsson.
1618
1619         Add WebKit2 API for window feature getter/setters
1620         <rdar://problem/8590373>
1621         https://bugs.webkit.org/show_bug.cgi?id=48496
1622
1623         * MiniBrowser/mac/BrowserWindowController.m:
1624         (-[BrowserWindowController awakeFromNib]):
1625         * MiniBrowser/win/BrowserView.cpp:
1626         (BrowserView::create):
1627         * WebKitTestRunner/TestController.cpp:
1628         (WTR::createOtherPage):
1629         (WTR::TestController::initialize):
1630
1631 2010-10-28  Søren Gjesse  <sgjesse@chromium.org>
1632
1633         Reviewed by Tony Chang.
1634
1635         Added support for the DumpRenderTree flags --multiple-loads and --js-flags to the Python test runner.
1636         https://bugs.webkit.org/show_bug.cgi?id=48236
1637
1638         * Scripts/webkitpy/layout_tests/port/chromium.py:
1639         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1640
1641 2010-10-28  Sergio Villar Senin  <svillar@igalia.com>
1642
1643         Reviewed by Martin Robinson.
1644
1645         [GTK] WebKitWebFrame's load-status is not properly notified to the tests
1646         https://bugs.webkit.org/show_bug.cgi?id=48048
1647
1648         DRT now listens to WebKitWebFrame load-status signals for each
1649         created frame instead of the load-status signal of the
1650         WebKitWebView that only notifies about changes in the main frame.
1651
1652         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1653         (webViewLoadFinished):
1654         (webFrameLoadStatusNotified):
1655         (frameCreatedCallback):
1656         (createWebView):
1657         (main):
1658
1659 2010-10-28  Tony Chang  <tony@chromium.org>
1660
1661         Reviewed by Kent Tamura.
1662
1663         [chromium] fix textInputController.{selectedRange,markedRange}
1664         https://bugs.webkit.org/show_bug.cgi?id=48487
1665
1666         * DumpRenderTree/chromium/TextInputController.cpp:
1667         (TextInputController::markedRange): Return arrays of ints, rather than a string
1668         (TextInputController::selectedRange): Ditto.
1669
1670 2010-10-27  Ojan Vafai  <ojan@chromium.org>
1671
1672         Reviewed by Tony Chang.
1673
1674         [chromium] add a master-name flag to new-run-webkit-tests
1675         https://bugs.webkit.org/show_bug.cgi?id=48488
1676
1677         The test results server now allows adding a master name to
1678         the uploaded files. This lets us distinguish bots that have
1679         the same name, but are on different masters.
1680
1681         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1682
1683 2010-10-28  Adam Roben  <aroben@apple.com>
1684
1685         Don't append a newline to the test output if the frame has no document
1686         element in WebKitTestRunner
1687
1688         Fixes <http://webkit.org/b/48526> Extra trailing newline when running
1689         plugins/document-open.html in WebKitTestRunner
1690
1691         Reviewed by Anders Carlsson.
1692
1693         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1694         (WTR::toJS): Added. Turns a UTF-8 C string into a JSStringRef.
1695         (WTR::hasDocumentElement): Added. Uses the JSC API to figure out
1696         whether the frame has a document element.
1697         (WTR::dumpFrameText): Match DRT by bailing (rather than appending an
1698         empty string and a newline) if the frame has no document element.
1699
1700 2010-10-28  Adam Roben  <aroben@apple.com>
1701
1702         Skip npn-invalidate-rect-invalidates-window.html on headless XP
1703         machines
1704
1705         TestNetscapePlugIn never receives a WM_PAINT message on headless XP
1706         machines, so this test times out. Fixing the test is covered by
1707         <http://webkit.org/b/48333>.
1708
1709         * Scripts/old-run-webkit-tests: Skip
1710         npn-invalidate-rect-invalidates-window.html on Windows if accelerated
1711         compositing support is disabled, which likely means we're on a headless
1712         XP machine.
1713
1714 2010-10-28  Kimmo Kinnunen  <kimmok@iki.fi>
1715
1716         Adding myself as a committer.
1717
1718         * Scripts/webkitpy/common/config/committers.py:
1719
1720 2010-10-27  Eric Seidel  <eric@webkit.org>
1721
1722         Reviewed by Adam Barth.
1723
1724         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is very slow
1725         https://bugs.webkit.org/show_bug.cgi?id=48500
1726
1727         This doesn't fix the problem, but it makes things slightly better.
1728         Each git svn find-rev call takes about .25 seconds on my desktop.
1729         This patch uses a new memoized class to avoid those calls when possible.
1730
1731         The real slowness is still git log on some files, like:
1732         git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
1733         I'm not yet sure how to make the pathological git logs better.
1734
1735         * Scripts/webkitpy/common/checkout/scm.py:
1736         * Scripts/webkitpy/common/memoized.py: Added.
1737         * Scripts/webkitpy/common/memoized_unittest.py: Added.
1738
1739 2010-10-27  Eric Seidel  <eric@webkit.org>
1740
1741         Reviewed by Ojan Vafai.
1742
1743         EWS bots should not use --quiet when running build-webkit
1744         https://bugs.webkit.org/show_bug.cgi?id=48482
1745
1746         --quiet is only correct when the sub-process does the error reporting.
1747         In the case of _can_build() the parent process is reporting the error.
1748         We'd like the full build log at queues.webkit.org so someone can look
1749         at the log and understand why the EWS is failing to build trunk.
1750
1751         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
1752         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
1753         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
1754         * Scripts/webkitpy/tool/mocktool.py:
1755
1756 2010-10-27  Brian Weinstein  <bweinstein@apple.com>
1757
1758         More Windows build fixage. Rename a variable that was named string.
1759
1760         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1761         (WTR::hasPrefix):
1762
1763 2010-10-27  Ojan Vafai  <ojan@chromium.org>
1764
1765         Reviewed by Dimitri Glazkov.
1766
1767         [chromium] Make the test results server store which master the bot is on
1768         https://bugs.webkit.org/show_bug.cgi?id=48478
1769
1770         The chromium bots recently changed so that there are multiple slaves with
1771         the same name on different masters. Up till now, the test results server
1772         assumed slave names were unique. Adds a master field to the file in order
1773         to distinguish.
1774
1775         Also, for files that currently lack a master or testtype, set them appropriately.
1776
1777         * TestResultServer/handlers/testfilehandler.py:
1778         * TestResultServer/index.yaml:
1779         * TestResultServer/model/jsonresults.py:
1780         * TestResultServer/model/testfile.py:
1781         * TestResultServer/templates/showfilelist.html:
1782         * TestResultServer/templates/uploadform.html:
1783
1784 2010-10-26  Darin Adler  <darin@apple.com>
1785
1786         Reviewed by Sam Weinig.
1787
1788         WebKitTestRunner needs to support layoutTestController.dumpBackForwardList
1789         https://bugs.webkit.org/show_bug.cgi?id=42322
1790         rdar://problem/8193631
1791
1792         WebKitTestRunner needs to support layoutTestController.clearBackForwardList
1793         https://bugs.webkit.org/show_bug.cgi?id=42333
1794         rdar://problem/8193643
1795
1796         * WebKitTestRunner/Configurations/InjectedBundle.xcconfig:
1797         Renamed the product to WebKitTestRunnerInjectedBundle to avoid
1798         name conflicts in the build directory.
1799
1800         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
1801         Added dumpBackForwardList and clearBackForwardList.
1802
1803         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
1804         (WTR::InjectedBundle::InjectedBundle): Removed initialization of m_mainPage.
1805         (WTR::InjectedBundle::didCreatePage): Removed unneeded underscore.
1806         (WTR::InjectedBundle::willDestroyPage): Ditto.
1807         (WTR::InjectedBundle::didReceiveMessage): Ditto.
1808         (WTR::InjectedBundle::initialize): Ditto.
1809         (WTR::InjectedBundle::didCreatePage): Changed code to use m_pages
1810         instead of m_mainPage and m_otherPages.
1811         (WTR::InjectedBundle::willDestroyPage): Ditto.
1812         (WTR::InjectedBundle::page): Ditto.
1813         (WTR::InjectedBundle::beginTesting): Ditto.
1814         (WTR::InjectedBundle::done): Ditto.
1815         (WTR::InjectedBundle::closeOtherPages): Ditto.
1816         (WTR::InjectedBundle::dumpBackForwardListsForAllPages): Added.
1817
1818         * WebKitTestRunner/InjectedBundle/InjectedBundle.h: Changed
1819         page function to longer be inline, and pageCount function
1820         to use m_pages. Added dumpBackForwardListsForAllPages, and
1821         removed some underscores. Replaced m_mainPage and m_otherPageas
1822         with m_pages.
1823
1824         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
1825         (WTR::adoptWK): Added.
1826         (WTR::hasPrefix): Added.
1827         (WTR::InjectedBundlePage::reset): Added code to set up
1828         m_previousTestBackForwardListItem.
1829         (WTR::InjectedBundlePage::dump): Added code to call
1830         dumpBackForwardListsForAllPages.
1831         (WTR::compareByTargetName): Added.
1832         (WTR::dumpBackForwardListItem): Added.
1833         (WTR::InjectedBundlePage::dumpBackForwardList): Added.
1834
1835         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h: Added
1836         dumpBackForwardList and m_previousTestBackForwardListItem.
1837
1838         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
1839         (WTR::LayoutTestController::LayoutTestController): Initialize
1840         m_shouldDumpBackForwardListsForAllWindows to false.
1841         (WTR::LayoutTestController::clearBackForwardList): Added.
1842
1843         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
1844         Added dumpBackForwardList, clearBackForwardList,
1845         shouldDumpBackForwardListsForAllWindows, and
1846         m_shouldDumpBackForwardListsForAllWindows.
1847
1848         * WebKitTestRunner/TestController.cpp:
1849         (WTR::TestController::initialize): Set up
1850         didReceiveSynchronousMessageFromInjectedBundle.
1851         (WTR::TestController::didReceiveSynchronousMessageFromInjectedBundle):
1852         Added.
1853         * WebKitTestRunner/TestController.h: Ditto.
1854
1855         * WebKitTestRunner/TestInvocation.cpp:
1856         (WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):
1857         Added.
1858         * WebKitTestRunner/TestInvocation.h: Ditto.
1859
1860         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
1861         Renamed bundle to WebKitTestRunnerInjectedBundle (see above).
1862         * WebKitTestRunner/mac/TestControllerMac.mm:
1863         (WTR::TestController::initializeInjectedBundlePath): Ditto.
1864
1865 2010-10-27  Chris Rogers  <crogers@google.com>
1866
1867         Reviewed by Chris Marrin.
1868
1869         Add ENABLE_WEB_AUDIO feature enable flag (initially disabled) to build-webkit
1870         https://bugs.webkit.org/show_bug.cgi?id=48279
1871
1872         * Scripts/build-webkit:
1873
1874 2010-10-27  Eric Seidel  <eric@webkit.org>
1875
1876         Unreviewed.
1877
1878         build-webkit should collect Visual Studio Express logs and display them
1879         https://bugs.webkit.org/show_bug.cgi?id=39199
1880
1881         It turns out my previous patch wasn't actually working on the win-ews
1882         machine.  So I've fixed my mistakes from before.
1883
1884         * Scripts/build-webkit:
1885          - Windows VSE builds change the CWD while building.  Why?  Who knows.
1886         * Scripts/print-vse-failure-logs:
1887          - windows VSE builds don't use Debug/Release as I expected, so I've
1888            moved off of --configuration to --top-level and added the /obj
1889            optimization while I was there.
1890
1891 2010-10-27  Dimitri Glazkov  <dglazkov@chromium.org>
1892
1893         Unreviewed, rolling out r70674.
1894         http://trac.webkit.org/changeset/70674
1895         https://bugs.webkit.org/show_bug.cgi?id=48053
1896
1897         Broke Chromium Windows build.
1898
1899         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1900         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1901
1902 2010-10-25  Tony Chang  <tony@chromium.org>
1903
1904         Reviewed by Anders Carlsson.
1905
1906         compile TestNetscapePlugIn on chromium linux
1907         https://bugs.webkit.org/show_bug.cgi?id=48274
1908
1909         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
1910         (NP_Initialize): On Linux, plugin funcs are set in initialize.
1911         (NPP_New): Mark the plugin as windowless.
1912         (NPP_GetValue): Handle mime type values.
1913         (NP_GetMIMEDescription):
1914         (NP_GetValue):
1915         * DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npapi.h: define TRUE and FALSE, which are in webkit's npapi.h.
1916         * DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npfunctions.h: Pull in npapi.h to get TRUE/FALSE (matches webkit's npfunctions.h)
1917
1918 2010-10-27  Kenneth Rohde Christiansen  <kenneth@webkit.org>
1919
1920         Reviewed by Anders Carlsson.
1921
1922         Remove contentSizeChanged callbacks as it is no longer
1923         part of the public UIClient.
1924
1925         Make WKPageContentsSizeChangedCallback be a private API
1926         https://bugs.webkit.org/show_bug.cgi?id=48409
1927
1928         * MiniBrowser/mac/BrowserWindowController.m:
1929         (-[BrowserWindowController awakeFromNib]):
1930         * MiniBrowser/win/BrowserView.cpp:
1931         (BrowserView::create):
1932         * WebKitTestRunner/TestController.cpp:
1933         (WTR::createOtherPage):
1934         (WTR::TestController::initialize):
1935
1936 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1937
1938         Reviewed by Ojan Vafai.
1939
1940         Make http locking default in NRWT.
1941         https://bugs.webkit.org/show_bug.cgi?id=48053
1942
1943         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1944         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1945
1946 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1947
1948         Reviewed by Tony Chang.
1949
1950         [NRWT] Fix http lock on Windows platform
1951         https://bugs.webkit.org/show_bug.cgi?id=48321
1952
1953         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1954
1955 2010-10-27  Satish Sampath  <satish@chromium.org>
1956
1957         Unreviewed, rolling out r70665.
1958         http://trac.webkit.org/changeset/70665
1959         https://bugs.webkit.org/show_bug.cgi?id=47089
1960
1961         Need to address Alexey's review comments.
1962
1963         * DumpRenderTree/LayoutTestController.cpp:
1964         (setMockSpeechInputResultCallback):
1965         * DumpRenderTree/LayoutTestController.h:
1966         * DumpRenderTree/chromium/LayoutTestController.cpp:
1967         (LayoutTestController::setMockSpeechInputResult):
1968         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1969         (LayoutTestController::setMockSpeechInputResult):
1970         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1971         (LayoutTestController::setMockSpeechInputResult):
1972         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1973         (LayoutTestController::setMockSpeechInputResult):
1974         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1975         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1976         (LayoutTestController::setMockSpeechInputResult):
1977         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
1978         (LayoutTestController::setMockSpeechInputResult):
1979
1980 2010-10-27  Leandro Gracia Gil  <leandrogracia@google.com>
1981
1982         Reviewed by Jeremy Orlow.
1983
1984         Added a second parameter to setMockSpeechInputResult for
1985         the language used in speech input.
1986         https://bugs.webkit.org/show_bug.cgi?id=47089
1987
1988         * DumpRenderTree/LayoutTestController.cpp:
1989         (setMockSpeechInputResultCallback):
1990         * DumpRenderTree/LayoutTestController.h:
1991         * DumpRenderTree/chromium/LayoutTestController.cpp:
1992         (LayoutTestController::setMockSpeechInputResult):
1993         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1994         (LayoutTestController::setMockSpeechInputResult):
1995         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1996         (LayoutTestController::setMockSpeechInputResult):
1997         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1998         (LayoutTestController::setMockSpeechInputResult):
1999         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2000         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2001         (LayoutTestController::setMockSpeechInputResult):
2002         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2003         (LayoutTestController::setMockSpeechInputResult):
2004         (LayoutTestController::pageSizeAndMarginsInPixels):
2005
2006 2010-10-27  Adam Roben  <aroben@apple.com>
2007
2008         Reset TestNetscapePlugIn's NPP_GetValue pointer when
2009         NullNPPGetValuePointer finishes running
2010
2011         Fixes <http://webkit.org/b/48435> REGRESSION (r70655): Many plugins
2012         tests are failing on Qt
2013
2014         Reviewed by Anders Carlsson.
2015
2016         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
2017         (PluginTest::NPP_Destroy):
2018         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
2019         Added a do-nothing NPP_Destroy implementation.
2020
2021         * DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp:
2022         (NullNPPGetValuePointer::NullNPPGetValuePointer): Save the original
2023         NPP_GetValue pointer in m_originalNPPGetValuePointer so we can restore
2024         it later.
2025         (NullNPPGetValuePointer::NPP_Destroy): Added. Restores the original
2026         NPP_GetValue pointer so it can be used in other tests.
2027
2028         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
2029         (NPP_Destroy):
2030         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
2031         (webkit_test_plugin_destroy_instance):
2032         Call through to the PluginTest when NPP_Destroy is called.
2033
2034 2010-10-26  Brian Weinstein  <bweinstein@apple.com>
2035
2036         Reviewed by Adam Roben.
2037
2038         WebKit2 shouldn't try to send an empty user agent
2039         https://bugs.webkit.org/show_bug.cgi?id=48397
2040         
2041         Add a test that when we set our custom user agent to an empty string, we don't send
2042         an empty user agent.
2043
2044         * TestWebKitAPI/Tests/WebKit2/PreventEmptyUserAgent.cpp: Added.
2045         (TestWebKitAPI::didRunJavaScript): Make sure that the result of navigator.userAgent isn't empty.
2046         (TestWebKitAPI::TEST): Set our custom user agent to the empty string, and run navigator.userAgent.
2047         
2048         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Add the new file.
2049         * TestWebKitAPI/win/TestWebKitAPI.vcproj: Ditto.
2050
2051 2010-10-27  Adam Roben  <aroben@apple.com>
2052
2053         Check in file I forgot in r70653
2054
2055         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
2056         (NP_Initialize): Save the NPPluginFuncs struct the browser passed to us
2057         so we can be naughty and modify it later.
2058
2059 2010-10-27  Adam Roben  <aroben@apple.com>
2060
2061         Test that WebKit doesn't crash if the plugin passes 0 for its
2062         NPP_GetValue pointer
2063
2064         Test for <http://webkit.org/b/48433> Crash in
2065         NetscapePlugin::shouldLoadSrcURL when using Shockwave Director 10.3 in
2066         WebKit2 on Windows
2067
2068         Reviewed by Eric Carlson.
2069
2070         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
2071         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
2072         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
2073         * GNUmakefile.am:
2074         Added NullNPPGetValuePointer.cpp.
2075
2076         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
2077         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
2078         Set up a place to store the NPPluginFuncs struct the browser passed to us.
2079
2080         * DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp: Added.
2081         (NullNPPGetValuePointer::NullNPPGetValuePointer): Null out the
2082         NPP_GetValue pointer we passed to the browser to simulate a plugin that
2083         doesn't implement NPP_GetValue.
2084         (NullNPPGetValuePointer::NPP_GetValue): Print an error message. If this
2085         function is called, it means that WebKit has changed in a way that
2086         makes this test invalid.
2087
2088         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
2089         (NP_GetEntryPoints): Save the NPPluginFuncs struct the browser passed
2090         to us so we can be naughty and modify it later.
2091
2092 2010-10-27  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
2093
2094         Reviewed by Andreas Kling.
2095
2096         Remove references to ancient QGVLauncher and QtLauncher
2097         https://bugs.webkit.org/show_bug.cgi?id=48430
2098
2099         QtTestBrowser substitutes both and has checks enabled.
2100
2101         * Scripts/webkitpy/style/checker.py: remove references
2102         * Scripts/webkitpy/style/checker_unittest.py: ditto
2103
2104 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2105
2106         Reviewed by Ojan Vafai.
2107
2108         [NRWT] Don't use image hash when it's no need in single test mode.
2109         https://bugs.webkit.org/show_bug.cgi?id=48326
2110
2111         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2112
2113 2010-10-27  David Kilzer  <ddkilzer@apple.com>
2114
2115         Fix leak of CFMutableDictionaryRef in createXMLStringFromWebArchiveData()
2116
2117         Reviewed by Adam Roben.
2118
2119         Follow-up fix for: <https://bugs.webkit.org/show_bug.cgi?id=48278>
2120
2121         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
2122         (createXMLStringFromWebArchiveData): Use RetainPtr<> to fix a
2123         leak introduced in r70613.
2124
2125 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2126
2127         Reviewed by Csaba Osztrogonác.
2128
2129         [NRWT] Add platform specific baseline search paths for Qt port
2130         https://bugs.webkit.org/show_bug.cgi?id=48428
2131
2132         * Scripts/webkitpy/layout_tests/port/qt.py:
2133
2134 2010-10-27  Adam Roben  <aroben@apple.com>
2135
2136         Catch exceptions when checking if we're inside a git working directory
2137
2138         Fixes <http://webkit.org/b/48420> REGRESSION (r70562): test-webkitpy
2139         fails on systems without git installed
2140
2141         Reviewed by Anders Carlsson.
2142
2143         * Scripts/webkitpy/common/net/credentials.py:
2144         (Credentials._credentials_from_git): Put the call to
2145         Git.in_working_directory inside the try/except since it, too, attempts
2146         to execute git and thus will throw on systems that don't have git
2147         installed.
2148
2149 2010-10-27  Nikolas Zimmermann  <nzimmermann@rim.com>
2150
2151         Reviewed by David Kilzer.
2152
2153         Convert DumpRenderTree webarchive code to CoreFoundation
2154         https://bugs.webkit.org/show_bug.cgi?id=48278
2155
2156         CFPropertyListCreateWithData and CFPropertyListCreateData are only available in 10.6+.
2157         Replace CFPropertyListCreateWithData by a combination of CFReadStreamCreateWithBytesNoCopy and CFPropertyListCreateFromStream.
2158         Replace CFPropertyListCreateData by CFPropertyListCreateXMLData.
2159
2160         These changes are wrapped in BUILDING_ON_LEOPARD, as the methods are deprecated on 10.6+.
2161
2162         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
2163         (createXMLStringFromWebArchiveData):
2164
2165 2010-10-26  Antonio Gomes  <agomes@rim.com>
2166
2167         Reviewed by Martin Robinson.
2168
2169         Remove the absolute path used to include DumpRenderTreeSupportGtk.h from LayoutTestController.h
2170
2171         It turns out this is rather unneeded since WebKitTools/GNUMakefile.am has WebKit/gtk/ in its include
2172         path, and then we can just do #include "WebCoreSupport/DumpRenderTreeSupportGtk.h"
2173
2174         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2175
2176 2010-10-26  David Kilzer  <ddkilzer@apple.com>
2177
2178         Rename WebArchiveDumpSupport.mm to WebArchiveDumpSupport.cpp
2179
2180         Reviewed by Adam Roben.
2181
2182         Part 4 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
2183
2184         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
2185         Updated for file moves and renames.
2186         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.mm.
2187         * DumpRenderTree/cf/WebArchiveDumpSupport.h: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.h.
2188
2189 2010-10-26  David Kilzer  <ddkilzer@apple.com>
2190
2191         Convert WebArchiveDumpSupport.mm from NS objects to CF types
2192
2193         Reviewed by Adam Roben.
2194
2195         Part 3 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
2196
2197         In order to share WebArchive code between the Mac and Windows
2198         ports, the code in WebArchiveDumpSupport.mm was converted from
2199         Cocoa to C++ using CoreFoundation (CF) types.
2200
2201         * DumpRenderTree/mac/DumpRenderTree.mm:
2202         (dump): Renamed serializeWebArchiveToXML() to
2203         createXMLStringFromWebArchiveData() and added HardAutorelease()
2204         to prevent leaks.
2205         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
2206         (createXMLStringFromWebArchiveData): Renamed from
2207         serializeWebArchiveToXML().  Changed to use CF types.
2208         (createCFURLResponseFromResponseData): Renamed from
2209         unarchiveNSURLResponseFromResponseData().  Changed to use CF
2210         types for its parameter and return type.
2211         * DumpRenderTree/mac/WebArchiveDumpSupport.mm: Replaced use of
2212         NS objects with CF types.  It will be renamed to *.cpp in a
2213         future commit.
2214         (convertMIMEType): Changed to use case-insensitive string
2215         comparisons.
2216         (convertWebResourceDataToString):
2217         (normalizeHTTPResponseHeaderFields):
2218         (normalizeWebResourceURL):
2219         (convertWebResourceResponseToDictionary):
2220         (compareResourceURLs):
2221         (createXMLStringFromWebArchiveData):
2222         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
2223         (createCFURLResponseFromResponseData): Renamed from
2224         unarchiveNSURLResponseFromResponseData().  Changed to use CF
2225         types for its parameter and return type.
2226
2227 2010-10-26  David Kilzer  <ddkilzer@apple.com>
2228
2229         Extract use of NSKeyedUnarchiver from WebArchiveDumpSupport.mm
2230
2231         Reviewed by Adam Roben.
2232
2233         Part 2 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
2234
2235         There is no equivalent to NSKeyedUnarchiver in CoreFoundation,
2236         so extract it into a platform-specific source file.
2237
2238         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
2239         (unarchiveNSURLResponseFromResponseData): Added declaration.
2240         * DumpRenderTree/mac/WebArchiveDumpSupport.mm:
2241         (convertWebResourceResponseToDictionary): Extracted code to
2242         unarchiveNSURLResponseFromResponseData() in
2243         WebArchiveDumpSupportMac.mm.  Updated logic to return early if
2244         nil is returned from unarchiveNSURLResponseFromResponseData().
2245         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
2246         (unarchiveNSURLResponseFromResponseData): Added.  Extracted
2247         code from convertWebResourceResponseToDictionary() in
2248         WebArchiveDumpSupport.mm.
2249
2250 2010-10-26  David Kilzer  <ddkilzer@apple.com>
2251
2252         Extract call to -[WebHTMLRepresentation supportedNonImageMIMETypes] from WebArchiveDumpSupport.mm
2253
2254         Reviewed by Adam Roben.
2255
2256         Part 1 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
2257
2258         The call to -[WebHTMLRepresentation supportedNonImageMIMETypes]
2259         is not cross-platform between Mac and Windows, so extract it
2260         into a platform-specific source file.
2261
2262         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Added
2263         WebArchiveDumpSupportMac.mm to the project.
2264         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
2265         (supportedNonImageMIMETypes): Added declaration.
2266         * DumpRenderTree/mac/WebArchiveDumpSupport.mm:
2267         (convertWebResourceDataToString): Replaced call to
2268         -[WebHTMLRepresentation supportedNonImageMIMETypes] with
2269         supportedNonImageMIMETypes().
2270         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm: Added.
2271         (supportedNonImageMIMETypes): Added.  Extracted from
2272         WebArchiveDumpSupport.mm.
2273
2274 2010-10-26  Antonio Gomes  <agomes@rim.com>
2275
2276         Reviewed by Martin Robinson.
2277
2278         [GTK] Implement DumpRenderTreeSupportGtk (similarly to DumpRenderTreeSupportQt idea)
2279         https://bugs.webkit.org/show_bug.cgi?id=48199
2280
2281         Implements support to WebKitTabToLinksPreferenceKey through LayoutTestController::overridePreference.
2282         The corresponding DumpRenderTreeSupportGtk method is called in the DRT context only.
2283
2284         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2285         (resetDefaultsToConsistentValues):
2286         (createWebView):
2287         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2288         (LayoutTestController::overridePreference):
2289
2290 2010-10-26  Ariya Hidayat  <ariya@sencha.com>
2291
2292         Unreviewed, change the order of my emails for bugzilla autocompletion.
2293
2294         * Scripts/webkitpy/common/config/committers.py:
2295
2296 2010-10-26  Eric Seidel  <eric@webkit.org>
2297
2298         Reviewed by David Kilzer.
2299
2300         build-webkit should collect Visual Studio Express logs and display them
2301         https://bugs.webkit.org/show_bug.cgi?id=39199
2302
2303         * Scripts/build-webkit:
2304         * Scripts/print-vse-failure-logs: Added.
2305         * Scripts/webkitdirs.pm:
2306
2307 2010-10-26  David Kilzer  <ddkilzer@apple.com>
2308
2309         <http://webkit.org/b/48224> build-webkit: add support for --meter-tag switch
2310
2311         Reviewed by Csaba Osztrogonác.
2312
2313         * Scripts/build-webkit: Added support for --meter-tag switch.
2314         It should have been added with r60820.
2315
2316 2010-10-26  Ademar de Souza Reis Jr.  <ademar.reis@openbossa.org>
2317
2318         Reviewed by Dumitru Daniliuc.
2319
2320         check-webkit-style fails on operator+=, operator-=, ... methods
2321         https://bugs.webkit.org/show_bug.cgi?id=48258
2322
2323         * Scripts/webkitpy/style/checkers/cpp.py: Added exceptions
2324         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added unit-tests
2325
2326 2010-10-26  Ojan Vafai  <ojan@chromium.org>
2327
2328         Reviewed by Tony Chang.
2329
2330         remove DEFER support from new-run-webkit-tests
2331         https://bugs.webkit.org/show_bug.cgi?id=48387
2332
2333         DEFER was needed when we were trying to ship Chrome beta.
2334         Now it's just extra complication.
2335
2336         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
2337         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
2338         * Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
2339         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
2340         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2341         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
2342
2343 2010-10-26  Eric Seidel  <eric@webkit.org>
2344
2345         Unreviewed.  Adding import with_statement to fix python 2.5.
2346
2347         Teach webkit-patch how to read credentials from the environment
2348         https://bugs.webkit.org/show_bug.cgi?id=48275
2349
2350         * Scripts/webkitpy/common/net/credentials_unittest.py:
2351          - import with_statement to fix python 2.5.
2352
2353 2010-10-26  Alexey Proskuryakov  <ap@apple.com>
2354
2355         Reviewed by Sam Weinig.
2356
2357         https://bugs.webkit.org/show_bug.cgi?id=48375
2358         <rdar://problem/8392724> Need delegate calls in PageLoaderClient to indicate if we have
2359         loaded insecure content
2360
2361         Updated for WebKit2 changes.
2362
2363         * MiniBrowser/mac/BrowserWindowController.m:
2364         (didDisplayInsecureContentForFrame):
2365         (didRunInsecureContentForFrame):
2366         (-[BrowserWindowController awakeFromNib]):
2367         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2368         (WTR::InjectedBundlePage::InjectedBundlePage):
2369         (WTR::InjectedBundlePage::didDisplayInsecureContentForFrame):
2370         (WTR::InjectedBundlePage::didRunInsecureContentForFrame):
2371         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
2372         * WebKitTestRunner/TestController.cpp:
2373         (WTR::TestController::initialize):
2374
2375 2010-10-26  Adam Roben  <aroben@apple.com>
2376
2377         Clear up confusion between "3D rendering" and "accelerated compositing"
2378
2379         WebKit has no software-based 3D rendering implementation, so 3D
2380         rendering can only be enabled when accelerated compositing is. But DRT
2381         was falsely reporting that 3D rendering was available on machines that
2382         don't support accelerated compositing, leading to much confusion.
2383
2384         Reviewed by Darin Adler.
2385
2386         Fixes <http://webkit.org/b/48370> REGRESSION (r70540): Many
2387         transforms/3d tests are failing on the XP bots
2388
2389         * DumpRenderTree/win/DumpRenderTree.cpp:
2390         (main): When ENABLE(3D_RENDERING) is turned on, only report that 3D
2391         rendering is available when acclerated compositing is also available.
2392
2393         * Scripts/old-run-webkit-tests: Skip tests that use the
2394         -webkit-transform-3d media query when 3D rendering is disabled,
2395         rather than when accelerated compositing is disabled.
2396
2397 2010-10-26  Eric Seidel  <eric@webkit.org>
2398
2399         Reviewed by Ojan Vafai.
2400
2401         Remove rietveld code now that it's unused
2402         https://bugs.webkit.org/show_bug.cgi?id=48359
2403
2404         Was cool that we added this support, but now that it's
2405         not used it makes little sense to keep it around.
2406         We can always restore this code from SVN if we need it.
2407
2408         * Scripts/webkitpy/common/config/__init__.py:
2409         * Scripts/webkitpy/common/net/bugzilla.py:
2410         * Scripts/webkitpy/common/net/bugzilla_unittest.py:
2411         * Scripts/webkitpy/common/net/rietveld.py: Removed.
2412         * Scripts/webkitpy/common/net/rietveld_unittest.py: Removed.
2413         * Scripts/webkitpy/thirdparty/__init__.py:
2414         * Scripts/webkitpy/tool/commands/download.py:
2415         * Scripts/webkitpy/tool/commands/download_unittest.py:
2416         * Scripts/webkitpy/tool/commands/queues.py:
2417         * Scripts/webkitpy/tool/commands/queues_unittest.py:
2418         * Scripts/webkitpy/tool/main.py:
2419         * Scripts/webkitpy/tool/mocktool.py:
2420         * Scripts/webkitpy/tool/steps/__init__.py:
2421         * Scripts/webkitpy/tool/steps/postcodereview.py: Removed.
2422
2423 2010-10-26  Anders Carlsson  <andersca@apple.com>
2424
2425         Reviewed by Sam Weinig.
2426
2427         pageDidScroll callback should be on the UI process client rather than (or in addition to) the web process client
2428         https://bugs.webkit.org/show_bug.cgi?id=48366
2429         <rdar://problem/8595202>
2430
2431         * MiniBrowser/mac/BrowserWindowController.m:
2432         (-[BrowserWindowController awakeFromNib]):
2433         * WebKitTestRunner/TestController.cpp:
2434         (WTR::createOtherPage):
2435         (WTR::TestController::initialize):
2436
2437 2010-10-26  Eric Seidel  <eric@webkit.org>
2438
2439         Reviewed by Tony Chang.
2440
2441         Teach webkit-patch how to read credentials from the environment
2442         https://bugs.webkit.org/show_bug.cgi?id=48275
2443
2444         This makes it possible for svn users to have their bugzilla credentials
2445         stored in their environment instead of typing them every time.
2446
2447         We need this for making it easy to run the win-ews bot (which currently
2448         uses svn instead of git).
2449
2450         * Scripts/webkitpy/common/net/credentials.py:
2451         * Scripts/webkitpy/common/net/credentials_unittest.py:
2452
2453 2010-10-26  Kenichi Ishibashi  <bashi@google.com>
2454
2455         Reviewed by Kent Tamura.
2456
2457         Input Method inserts conversion candidates unexpectedly
2458         https://bugs.webkit.org/show_bug.cgi?id=46868
2459
2460         Adds setComposition() to TextInputController to make DRT emulate
2461         an input method behavior.
2462
2463         * DumpRenderTree/chromium/TextInputController.cpp:
2464         (TextInputController::TextInputController):
2465         (TextInputController::setComposition): Added.
2466         * DumpRenderTree/chromium/TextInputController.h:
2467
2468 2010-10-26  Simon Fraser  <simon.fraser@apple.com>
2469
2470         Reviewed by Sam Weinig.
2471
2472         Support layoutTestController.layerTreeAsText in WebKitTestRunner
2473         https://bugs.webkit.org/show_bug.cgi?id=42145
2474         
2475         Implement layerTreeAsText() in WebKitTestRunner.
2476
2477         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
2478         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
2479         (WTR::LayoutTestController::layerTreeAsText):
2480         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
2481
2482 2010-10-26  Adam Barth  <abarth@webkit.org>
2483
2484         Reviewed by Eric Seidel.
2485
2486         webkit-patch rollout produces incorrect patch when using svn move
2487         https://bugs.webkit.org/show_bug.cgi?id=48244
2488
2489         We need to flush our caches when we modify the working copy.
2490
2491         * Scripts/webkitpy/tool/steps/revertrevision.py:
2492
2493 2010-10-26  Eric Seidel  <eric@webkit.org>
2494
2495         Reviewed by Adam Barth.
2496
2497         commit-queue flaky test message can list the same author more than once
2498         https://bugs.webkit.org/show_bug.cgi?id=48268
2499
2500         tonikitoo reported to me over IRC this morning that he's seen
2501         the commit-queue report flaky tests with author lists like:
2502         "adam, adam and adam", suggesting we're not uniquing authors
2503         before writing the message.
2504
2505         I fixed the uniquing and added a bunch more unit testing.
2506
2507         * Scripts/webkitpy/tool/commands/queues.py:
2508         * Scripts/webkitpy/tool/commands/queues_unittest.py:
2509         * Scripts/webkitpy/tool/mocktool.py:
2510
2511 2010-10-26  Adam Roben  <aroben@apple.com>
2512
2513         Pull in the FeatureDefines*.vsprops files when building DRT
2514
2515         This ensures that various ENABLE() macros will get set correctly.
2516         (Prior to r70320 we were picking up the ENABLE(3D_RENDERING) flag
2517         through wtf/Platform.h.)
2518
2519         Fixes <http://webkit.org/b/48343> REGRESSION (r70320): DumpRenderTree
2520         says 3D_RENDERING is disabled even when it is enabled
2521
2522         Reviewed by Ada Chan.
2523
2524         * DumpRenderTree/win/DumpRenderTree.vcproj: Added
2525         FeaturesDefines*.vsprops to all configurations.
2526
2527         * DumpRenderTree/win/DumpRenderTreeWin.h: Touched to force a rebuild.
2528
2529 2010-10-26  Søren Gjesse  <sgjesse@chromium.org>
2530
2531         Reviewed by Dimitri Glazkov.
2532
2533         Added options --multiple-loads and --js-flags to chromium DumpRenderTree. The option --multiple-loads=X
2534         is used to have DumpRenderTree load each test it runs X times. To be able to have more fine-grained control
2535         of how the JavaScript engine behaves for each load the flag --js-flags can specify a list of flag-sets like this
2536
2537           --js-flags="--xxx,--noxxx --yyy,--noyyy"
2538
2539         First load will run with --xxx, the second with --yyy and the third without any (the 'no' prefix is handled by
2540         V8 to turn off the flag).
2541
2542         The changes to the Python test runner will be in a separate change.
2543
2544
2545         * DumpRenderTree/chromium/DumpRenderTree.cpp:
2546         (runTest):
2547         (main):
2548         * DumpRenderTree/chromium/TestShell.cpp:
2549         (TestShell::TestShell):
2550         (TestShell::runFileTest):
2551         (TestShell::testFinished):
2552         * DumpRenderTree/chromium/TestShell.h:
2553         (TestShell::loadCount):
2554         (TestShell::setLoadCount):
2555         (TestShell::javaScriptFlagsForLoad):
2556         (TestShell::setJavaScriptFlags):
2557         (TestShell::setDumpWhenFinished):
2558
2559 2010-10-26  Adam Roben  <aroben@apple.com>
2560
2561         Skip more tests that depend on accelerated compositing when accelerated
2562         compositing is disabled
2563
2564         Fixes <http://webkit.org/b/48329> Some tests fail when accelerated
2565         compositing is disabled
2566
2567         Reviewed by John Sullivan.
2568
2569         * Scripts/old-run-webkit-tests: Added more tests to skip when
2570         accelerated compositing is disable. The tests all have output that
2571         changes depending on whether accelerated compositing is enabled.
2572
2573 2010-10-25  Eric Seidel  <eric@webkit.org>
2574
2575         Reviewed by Adam Barth.
2576
2577         webkit-patch should clean up svn locks when passed --force-clean
2578         https://bugs.webkit.org/show_bug.cgi?id=48269
2579
2580         * Scripts/webkitpy/common/checkout/scm.py:
2581
2582 2010-10-25  Anders Carlsson  <andersca@apple.com>
2583
2584         Reviewed by Adam Roben.
2585
2586         Add a pageDidScroll BundleUIClient callback
2587         https://bugs.webkit.org/show_bug.cgi?id=48260
2588         <rdar://problem/8531159>
2589
2590         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2591         (WTR::InjectedBundlePage::InjectedBundlePage):
2592
2593 2010-10-25  Johnny Ding  <jnd@chromium.org>
2594
2595         Reviewed by Tony Chang.
2596
2597         Dump the gesture status of frame in frame load callbacks in DumpRenderTree
2598         by adding a new method dumpUserGestureInFrameLoadCallbacks.
2599         Now only dump the gesture status in "DidStartProvisionalLoad" callback.
2600         https://bugs.webkit.org/show_bug.cgi?id=47849
2601
2602         * DumpRenderTree/LayoutTestController.cpp:
2603         (LayoutTestController::LayoutTestController):
2604         (dumpUserGestureInFrameLoadCallbacksCallback):
2605         (LayoutTestController::staticFunctions):
2606         * DumpRenderTree/LayoutTestController.h:
2607         (LayoutTestController::dumpUserGestureInFrameLoadCallbacks):
2608         (LayoutTestController::setDumpUserGestureInFrameLoadCallbacks):
2609         * DumpRenderTree/chromium/LayoutTestController.cpp:
2610         (LayoutTestController::LayoutTestController):
2611         (LayoutTestController::dumpUserGestureInFrameLoadCallbacks):
2612         (LayoutTestController::reset):
2613         * DumpRenderTree/chromium/LayoutTestController.h:
2614         (LayoutTestController::shouldDumpUserGestureInFrameLoadCallbacks):
2615         (LayoutTestController::setShouldDumpUserGestureInFrameLoadCallbacks):
2616         * DumpRenderTree/chromium/TestShell.h:
2617         (TestShell::shouldDumpUserGestureInFrameLoadCallbacks):
2618         * DumpRenderTree/chromium/WebViewHost.cpp:
2619         (WebViewHost::didStartProvisionalLoad):
2620         (WebViewHost::printFrameUserGestureStatus):
2621         * DumpRenderTree/chromium/WebViewHost.h:
2622         * DumpRenderTree/mac/FrameLoadDelegate.mm:
2623         (-[WebFrame _drt_printFrameUserGestureStatus]):
2624         (-[FrameLoadDelegate webView:didStartProvisionalLoadForFrame:]):
2625         (-[FrameLoadDelegate webView:didCommitLoadForFrame:]):
2626         (-[FrameLoadDelegate webView:didFailProvisionalLoadWithError:forFrame:]):
2627         (-[FrameLoadDelegate webView:didFinishLoadForFrame:]):
2628         (-[FrameLoadDelegate webView:didFailLoadWithError:forFrame:]):
2629         (-[FrameLoadDelegate webView:windowScriptObjectAvailable:]):
2630         (-[FrameLoadDelegate webView:didReceiveTitle:forFrame:]):
2631         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2632         (LayoutTestController::reset):
2633         (LayoutTestController::dumpUserGestureInFrameLoadCallbacks):
2634         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2635
2636 2010-10-25  Sheriff Bot  <webkit.review.bot@gmail.com>
2637
2638         Unreviewed, rolling out r70442.
2639         http://trac.webkit.org/changeset/70442
2640         https://bugs.webkit.org/show_bug.cgi?id=48248
2641
2642         http locking doesn't work on Windows (Requested by Ossy on
2643         #webkit).
2644
2645         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2646         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2647
2648 2010-10-25  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2649
2650         Reviewed by Ojan Vafai.
2651
2652         Make http locking default in NRWT.
2653         https://bugs.webkit.org/show_bug.cgi?id=48053
2654
2655         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2656         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2657
2658 2010-10-25  David Kilzer  <ddkilzer@apple.com>
2659
2660         <http://webkit.org/b/48185> build-webkit: add support for --progress-tag switch
2661         (Re-landing r70413 after Qt minimal buildfix: r70440.)
2662
2663         Reviewed by Kenneth Rohde Christiansen.
2664
2665         * Scripts/build-webkit: Added support for --progress-tag switch.
2666         It should have been added with r57051.
2667
2668 2010-10-24  Adam Barth  <abarth@webkit.org>
2669
2670         Reviewed by David Kilzer.
2671
2672         webkit-patch upload fails when the patch removes a file
2673         https://bugs.webkit.org/show_bug.cgi?id=48187
2674
2675         We need to use "--" to separate file names from the rest of the
2676         command.
2677
2678         * Scripts/webkitpy/common/checkout/scm.py:
2679
2680 2010-10-24  Sheriff Bot  <webkit.review.bot@gmail.com>
2681
2682         Unreviewed, rolling out r70413.
2683         http://trac.webkit.org/changeset/70413
2684         https://bugs.webkit.org/show_bug.cgi?id=48210
2685
2686         It broke Qt minimal build (Requested by Ossy on #webkit).
2687
2688         * Scripts/build-webkit:
2689
2690 2010-10-24  Yi Shen  <yi.4.shen@nokia.com>
2691
2692         Reviewed by Andreas Kling.
2693
2694         [Qt][QtTestBrowser] Toggle use of QGraphicsView messes up the menu
2695         https://bugs.webkit.org/show_bug.cgi?id=48141
2696
2697         * QtTestBrowser/launcherwindow.cpp:
2698         (LauncherWindow::init):
2699         (LauncherWindow::initializeView):
2700         (LauncherWindow::toggleWebView):
2701
2702 2010-10-24  Daniel Bates  <dbates@rim.com>
2703
2704         Reviewed by David Kilzer.
2705
2706         Fix Perl uninitialized warnings in VCSUtils::svnStatus()
2707         and VCSUtils::removeEOL().
2708         https://bugs.webkit.org/show_bug.cgi?id=48196
2709
2710         VCSUtils::svnStatus() concatenates the output of svn status with
2711         a new line character and svn status may return no output (say for
2712         a file that has not been added, deleted, or modified). We should
2713         only concatenate the output of svn status if there is some.
2714
2715         Also, VCSUtils::removeEOL() should ensure that its argument
2716         is initialized before performing a string substitution.
2717
2718         * Scripts/VCSUtils.pm:
2719           - Modified removeEOL() to return "" if its argument is undefined.
2720           - Exported removeEOL() so that it can be tested. 
2721         * Scripts/webkitperl/VCSUtils_unittest/removeEOL.pl: Added.
2722
2723 2010-10-24  David Kilzer  <ddkilzer@apple.com>
2724
2725         <http://webkit.org/b/48185> build-webkit: add support for --progress-tag switch
2726
2727         Reviewed by Kenneth Rohde Christiansen.
2728
2729         * Scripts/build-webkit: Added support for --progress-tag switch.
2730         It should have been added with r57051.
2731
2732 2010-10-24  David Kilzer  <ddkilzer@apple.com>
2733
2734         <http://webkit.org/b/48184> build-webkit: reformat support variables for better maintainability
2735
2736         Reviewed by Daniel Bates.
2737
2738         * Scripts/build-webkit: Alphabetized the @features array based
2739         on the name of the feature.  Reformatted the list of support
2740         variables so that they match the order of @features, and so that
2741         each has its own line.  This makes it easy to add new variables
2742         in the correct order.
2743
2744 2010-10-23  Eric Seidel  <eric@webkit.org>
2745
2746         Reviewed by Adam Barth.
2747
2748         EWS never removes invalid patch ids
2749         https://bugs.webkit.org/show_bug.cgi?id=48173
2750
2751         This is just sticking another finger in the dam.
2752         However this adds more unit testing which will help
2753         us make sure we're always releasing patches once we
2754         redesign the release_patch API and call these from
2755         a more central place.
2756
2757         * Scripts/webkitpy/tool/commands/queues.py:
2758         * Scripts/webkitpy/tool/commands/queues_unittest.py:
2759         * Scripts/webkitpy/tool/mocktool.py:
2760          - Added the ability to request invalid patches.
2761            Log a warning message to make sure we don't ever have
2762            tests use invalid patch fetches by mistake.
2763
2764 2010-10-23  Dan Bernstein  <mitz@apple.com>
2765
2766         Build fix. Add stub implementations for required NSDraggingInfo methods.
2767
2768         * DumpRenderTree/mac/DumpRenderTreeDraggingInfo.mm:
2769         (-[DumpRenderTreeDraggingInfo draggingFormation]):
2770         (-[DumpRenderTreeDraggingInfo setDraggingFormation:]):
2771         (-[DumpRenderTreeDraggingInfo animatesToDestination]):
2772         (-[DumpRenderTreeDraggingInfo setAnimatesToDestination:]):
2773         (-[DumpRenderTreeDraggingInfo numberOfValidItemsForDrop]):
2774         (-[DumpRenderTreeDraggingInfo setNumberOfValidItemsForDrop:]):
2775         (-[DumpRenderTreeDraggingInfo enumerateDraggingItemsWithOptions:forView:classes:searchOptions:usingBlock:]):
2776
2777 2010-10-23  David Kilzer  <ddkilzer@apple.com>
2778
2779         <http://webkit.org/b/48186> Remove unneeded WebHTMLRepresentationInternal.h header
2780
2781         Reviewed by Sam Weinig.
2782
2783         The only method defined in WebHTMLRepresentationInternal.h is
2784         also defined in WebHTMLRepresentation.h, so use that instead.
2785
2786         * DumpRenderTree/mac/InternalHeaders/WebKit/WebHTMLRepresentationInternal.h: Removed.
2787         * DumpRenderTree/mac/WebArchiveDumpSupport.mm: Switched to use
2788         WebHTMLRepresentation.h.
2789
2790 2010-10-23  Sheriff Bot  <webkit.review.bot@gmail.com>
2791
2792         Unreviewed, rolling out r70367.
2793         http://trac.webkit.org/changeset/70367
2794         https://bugs.webkit.org/show_bug.cgi?id=48176
2795
2796         It made 8-10 tests crash on Qt bot (Requested by Ossy on
2797         #webkit).
2798
2799         * DumpRenderTree/LayoutTestController.cpp:
2800         (setCustomPolicyDelegateCallback):
2801         * DumpRenderTree/LayoutTestController.h:
2802         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2803         (LayoutTestController::setCustomPolicyDelegate):
2804         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2805         (LayoutTestController::setCustomPolicyDelegate):
2806         * DumpRenderTree/mac/PolicyDelegate.h:
2807         * DumpRenderTree/mac/PolicyDelegate.mm:
2808         (-[PolicyDelegate webView:decidePolicyForNavigationAction:request:frame:decisionListener:]):
2809         (-[PolicyDelegate setPermissive:]):
2810         (-[PolicyDelegate setControllerToNotifyDone:]):
2811         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2812         (LayoutTestController::setCustomPolicyDelegate):
2813         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2814         (LayoutTestController::setCustomPolicyDelegate):
2815
2816 2010-10-22  Eric Seidel  <eric@webkit.org>
2817
2818         Reviewed by Adam Barth.
2819
2820         EWS never releases patches which fail to apply
2821         https://bugs.webkit.org/show_bug.cgi?id=48171
2822
2823         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
2824         * Scripts/webkitpy/tool/commands/queues.py:
2825
2826 2010-10-22  Eric Seidel  <eric@webkit.org>
2827
2828         Unreviewed.
2829
2830         Make the EWS queues restart themselves more often.
2831         This matches the commit-queue.
2832
2833         * EWSTools/start-queue.sh:
2834
2835 2010-10-22  Eric Seidel  <eric@webkit.org>
2836
2837         Reviewed by Adam Barth.
2838
2839         commit-queue should not pass --quiet to subcommands now that the parent command does the reporting
2840         https://bugs.webkit.org/show_bug.cgi?id=48165
2841
2842         * Scripts/webkitpy/tool/bot/commitqueuetask.py:
2843         * Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
2844         * Scripts/webkitpy/tool/commands/queues_unittest.py:
2845
2846 2010-10-22  Mario Sanchez Prada  <msanchez@igalia.com>
2847
2848         Adding myself as a committer.
2849
2850         * Scripts/webkitpy/common/config/committers.py:
2851
2852 2010-10-20  Anders Carlsson  <andersca@apple.com>
2853
2854         Reviewed by Adam Barth and Darin Adler.
2855
2856         Using the decidePolicyForMIMEType delegate message in an asynchronous manner does not work
2857         https://bugs.webkit.org/show_bug.cgi?id=48014
2858         <rdar://problem/8202716>
2859         
2860         * DumpRenderTree/LayoutTestController.cpp:
2861         (setCustomPolicyDelegateCallback):
2862         Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
2863
2864         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2865         (LayoutTestController::setCustomPolicyDelegate):
2866         Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
2867
2868         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2869         (LayoutTestController::setCustomPolicyDelegate):
2870         Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
2871
2872         * DumpRenderTree/mac/PolicyDelegate.h:
2873         * DumpRenderTree/mac/PolicyDelegate.mm:
2874         (-[PolicyDelegate webView:decidePolicyForMIMEType:request:frame:decisionListener:]):
2875         If _callIgnoreInDecidePolicyForMIMETypeAfterOneSecond is true, call [listener ignore] after one second.
2876
2877         (-[PolicyDelegate setCallIgnoreInDecidePolicyForMIMETypeAfterOneSecond:]):
2878         Update _callIgnoreInDecidePolicyForMIMETypeAfterOneSecond.
2879         
2880         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2881         (LayoutTestController::setCustomPolicyDelegate):
2882         Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
2883
2884         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2885         (LayoutTestController::setCustomPolicyDelegate):
2886         Add callIgnoreInDecidePolicyForMIMETypeAfterOneSecond argument.
2887
2888 2010-10-22  Sam Weinig  <sam@webkit.org>
2889
2890         Reviewed by Anders Carlsson.
2891
2892         Rename the InjectedBundle for TestWebKitAPI to InjectedBundleTestWebKitAPI.bundle
2893
2894         * TestWebKitAPI/Configurations/InjectedBundle.xcconfig:
2895         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
2896         * TestWebKitAPI/mac/PlatformUtilitiesMac.mm:
2897         (TestWebKitAPI::Util::createInjectedBundlePath):
2898
2899 2010-10-22  Dirk Pranke  <dpranke@chromium.org>
2900
2901         Reviewed by Eric Seidel.
2902
2903         test-webkitpy fails when passed -v (or any other option)
2904
2905         Fix port/base_unittest to pass an explicit list of arguments
2906         rather than accidentally picking up sys.argv.
2907
2908         https://bugs.webkit.org/show_bug.cgi?id=48071
2909
2910         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2911
2912 2010-10-22  Brian Weinstein  <bweinstein@apple.com>
2913
2914         Windows build fix. Update the createNewPage callback to account for the new
2915         arguments.
2916
2917         * MiniBrowser/win/BrowserView.cpp:
2918         (createNewPage):
2919
2920 2010-10-22  Sam Weinig  <sam@webkit.org>
2921
2922         Reviewed by Anders Carlsson.
2923
2924         WebKit2 needs to pass the current event modifier flags when requesting a new window
2925         https://bugs.webkit.org/show_bug.cgi?id=48140
2926
2927         * MiniBrowser/mac/BrowserWindowController.m:
2928         (createNewPage):
2929         * WebKitTestRunner/TestController.cpp:
2930         (WTR::createOtherPage):
2931         Update for new signature for the WKPageCreateNewPageCallback.
2932
2933 2010-10-22  Eric Seidel  <eric@webkit.org>
2934
2935         Unreviewed.  Just adding a test case.
2936
2937         The style-queue was not recognizing new patches.  It turns
2938         out I had fixed this in an earlier change, but just not
2939         deployed to queues.webkit.org.  As part of investigating why it
2940         was broken, I wrote a test for my previous change which I'm now landing.
2941
2942         * QueueStatusServer/model/queues_unittest.py:
2943
2944 2010-10-22  Eric Seidel  <eric@webkit.org>
2945
2946         Reviewed by Adam Barth.
2947
2948         EWS does not need to process obsolete patches
2949         https://bugs.webkit.org/show_bug.cgi?id=48093
2950
2951         This was an easy change, but to test it I had to pipe
2952         real Attachment objects into the queue testing system.
2953         Doing so revealed a whole bunch of bugs in our unit tests,
2954         which I fixed as part of this patch.
2955
2956         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
2957          - This is the actual code change.  This will not reduce the
2958            backlog in the EWS queues much, but it will make rejections
2959            much quicker for obsolete patches or closed bugs.
2960         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
2961          - Test my new code.
2962          - Used a real attachment object and got rid of MockPatch
2963          - Shared code between the mac-ews and cr-mac-ews tests.
2964         * Scripts/webkitpy/tool/commands/queues_unittest.py:
2965          - Can't use MockPatch anymore.
2966          - Removing MockPatch found more bugs here!
2967         * Scripts/webkitpy/tool/commands/queuestest.py:
2968         * Scripts/webkitpy/tool/commands/sheriffbot_unittest.py:
2969         * Scripts/webkitpy/tool/mocktool.py:
2970          - MockBugzilla should not be a "Mock" object.  Right now tool.bugs()
2971            is allowed, but wrong.  Making it not a Mock will make tool.bugs() correctly fail.
2972
2973 2010-10-22  Sheriff Bot  <webkit.review.bot@gmail.com>
2974
2975         Unreviewed, rolling out r70301.
2976         http://trac.webkit.org/changeset/70301
2977         https://bugs.webkit.org/show_bug.cgi?id=48126
2978
2979         "Lang attribute layout tests failing" (Requested by satish on
2980         #webkit).
2981
2982         * DumpRenderTree/LayoutTestController.cpp:
2983         (setMockSpeechInputResultCallback):
2984         * DumpRenderTree/LayoutTestController.h:
2985         * DumpRenderTree/chromium/LayoutTestController.cpp:
2986         (LayoutTestController::setMockSpeechInputResult):
2987         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2988         (LayoutTestController::setMockSpeechInputResult):
2989         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2990         (LayoutTestController::setMockSpeechInputResult):
2991         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2992         (LayoutTestController::setMockSpeechInputResult):
2993         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2994         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2995         (LayoutTestController::setMockSpeechInputResult):
2996         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2997         (LayoutTestController::setMockSpeechInputResult):
2998
2999 2010-10-22  Leandro Gracia Gil  <leandrogracia@google.com>
3000
3001         Reviewed by Jeremy Orlow.
3002
3003         Added a second parameter to setMockSpeechInputResult for
3004         the language used in speech input.
3005         https://bugs.webkit.org/show_bug.cgi?id=47089
3006
3007         * DumpRenderTree/LayoutTestController.cpp:
3008         (setMockSpeechInputResultCallback):
3009         * DumpRenderTree/LayoutTestController.h:
3010         * DumpRenderTree/chromium/LayoutTestController.cpp:
3011         (LayoutTestController::setMockSpeechInputResult):
3012         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3013         (LayoutTestController::setMockSpeechInputResult):
3014         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3015         (LayoutTestController::setMockSpeechInputResult):
3016         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3017         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3018         (LayoutTestController::setMockSpeechInputResult):
3019         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3020         (LayoutTestController::setMockSpeechInputResult):
3021         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3022         (LayoutTestController::setMockSpeechInputResult):
3023
3024 2010-10-21  MORITA Hajime  <morrita@google.com>
3025
3026         Reviewed by Kent Tamura.
3027
3028         [Win][DRT] should have LayoutTestController.hasSpellingMarker()
3029         https://bugs.webkit.org/show_bug.cgi?id=47885
3030
3031         - Implemented LayoutTestController.hasSpellingMarker(),
3032         - Added fake spellcheck implementation.
3033
3034         * DumpRenderTree/win/DumpRenderTree.cpp:
3035         (main):
3036         * DumpRenderTree/win/EditingDelegate.cpp:
3037         (indexOfFirstWordCharacter):
3038         (wordLength):
3039         (EditingDelegate::checkSpellingOfString):
3040         * DumpRenderTree/win/EditingDelegate.h:
3041         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3042         (LayoutTestController::hasSpellingMarker):
3043
3044 2010-10-21  Mihai Parparita  <mihaip@chromium.org>
3045
3046         Unreviewed. Re-enable test that was mistakenly disabled by r67974.
3047
3048         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
3049
3050 2010-10-21  Mihai Parparita  <mihaip@chromium.org>
3051
3052         Unreviewed fix for rebaseline-chromium-webkit-tests.
3053
3054         Port.diff_image no longer has a tolerance parameter. Also, use
3055         get_option('tolerance') which is safer if the options object doesn't
3056         define a tolerance attribute (it doesn't for the one used in
3057         rebaseline_chromium_webkit_tests).
3058
3059         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
3060         * Scripts/webkitpy/layout_tests/port/webkit.py:
3061         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
3062
3063 2010-10-21  Mihai Parparita  <mihaip@chromium.org>
3064
3065         Reviewed by Ojan Vafai.
3066
3067         Add support for --tolerance in NRWT
3068         https://bugs.webkit.org/show_bug.cgi?id=47959
3069         
3070         Add support for the --tolerance flag in NRWT. The Port.diff_image
3071         signature shouldn't need a tolerance parameter (it's not set per test),
3072         just have ports that use it (currently only WebKitPort) read it from 
3073         the options object.
3074
3075         * Scripts/webkitpy/layout_tests/port/chromium.py:
3076         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
3077         * Scripts/webkitpy/layout_tests/port/test.py:
3078         * Scripts/webkitpy/layout_tests/port/webkit.py:
3079         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3080         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
3081
3082 2010-10-21  Eric Seidel  <eric@webkit.org>
3083
3084         Reviewed by Adam Barth.
3085
3086         Add --suggest-reviewers option to upload to auto-suggest reviewers for your patch
3087         https://bugs.webkit.org/show_bug.cgi?id=48088
3088
3089         This is a first-pass.  Works, but we'll eventually
3090         turn this on by default, refine the suggestion algorithm
3091         and possibly move it to a different place in the upload step order.
3092
3093         * Scripts/webkitpy/common/checkout/api.py:
3094         * Scripts/webkitpy/tool/commands/download_unittest.py:
3095         * Scripts/webkitpy/tool/commands/upload.py:
3096         * Scripts/webkitpy/tool/commands/upload_unittest.py:
3097         * Scripts/webkitpy/tool/mocktool.py:
3098         * Scripts/webkitpy/tool/steps/__init__.py:
3099         * Scripts/webkitpy/tool/steps/options.py:
3100
3101 2010-10-21  Mario Sanchez Prada  <msanchez@igalia.com>
3102
3103         Reviewed by Chris Fleizach.
3104
3105         [GTK] Segfault while testing accessibility/iframe-bastardization.html
3106         https://bugs.webkit.org/show_bug.cgi?id=30123
3107
3108         Check m_element before actually using it to get the parent.
3109
3110         * DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
3111         (AccessibilityUIElement::parentElement): Add an extra check to
3112         make sure m_element points to a valid value before actually using
3113         it to get the parent. Also, assert ATK_IS_OBJECT(m_element).
3114
3115 2010-10-21  Anders Carlsson  <andersca@apple.com>
3116
3117         Reviewed by Adam Roben.
3118
3119         Crash evaluating JavaScript string that throws an exception
3120         https://bugs.webkit.org/show_bug.cgi?id=48092
3121         <rdar://problem/8487657>
3122
3123         Add a test that evaluates a JavaScript string that throws an exception and check that 
3124         the callback is called.
3125
3126         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
3127         * TestWebKitAPI/Tests/WebKit2/EvaluateJavaScript.cpp: Added.
3128         (TestWebKitAPI::didRunJavaScript):
3129         (TestWebKitAPI::WebKit2_EvaluateJavaScript):
3130         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
3131
3132 2010-10-21  Adam Roben  <aroben@apple.com>
3133
3134         Test that the plugin's HWND is invalidated when NPN_InvalidateRect is
3135         called
3136
3137         Test for <http://webkit.org/b/48086> <rdar://problem/8482944>
3138         Silverlight doesn't repaint in WebKit2
3139
3140         Reviewed by Anders Carlsson.
3141
3142         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
3143         (PluginTest::NPN_InvalidateRect): Added. Calls through to the browser.
3144         (executeScript): Added. Asks the browser to evaluate the script.
3145
3146         (PluginTest::waitUntilDone):
3147         (PluginTest::notifyDone):
3148         Added. Calls through to layoutTestController.
3149
3150         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Added
3151         * NPN_InvalidateRect and waitUntilDone/notifyDone.
3152
3153         * DumpRenderTree/TestNetscapePlugIn/Tests/win/NPNInvalidateRectInvalidatesWindow.cpp: Added.
3154         (TemporaryWindowMover::moveSucceeded):
3155         (TemporaryWindowMover::TemporaryWindowMover):
3156         (TemporaryWindowMover::~TemporaryWindowMover):
3157         This class moves a window on-screen and shows it, then moves it back and hides it.
3158
3159         (NPNInvalidateRectInvalidatesWindow::NPNInvalidateRectInvalidatesWindow):
3160         Initialize our members.
3161         (NPNInvalidateRectInvalidatesWindow::~NPNInvalidateRectInvalidatesWindow):
3162         Delete our window mover if it hasn't been deleted already.
3163         (NPNInvalidateRectInvalidatesWindow::NPP_SetWindow): Subclass the
3164         plugin HWND and move the test harness window on screen.
3165         (NPNInvalidateRectInvalidatesWindow::wndProc): Call through to onPaint
3166         when we get a WM_PAINT message.
3167         (NPNInvalidateRectInvalidatesWindow::onPaint): Do the test and tell
3168         LayoutTestController we're done.
3169         (NPNInvalidateRectInvalidatesWindow::testInvalidateRect): Validate
3170         ourselves, invalidate our lower-right quadrant via NPN_InvalidateRect,
3171         then check that our HWND's invalid region is the rect that we
3172         invalidated.
3173
3174         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
3175         * Added NPNInvalidateRectInvalidatesWindow.
3176
3177 2010-10-21  Daniel Bates  <dbates@rim.com>
3178
3179         Add Git-support to do-file-rename
3180         https://bugs.webkit.org/show_bug.cgi?id=48015
3181
3182         Fix tools. Export function scmMoveOrRenameFile so that it can be called from
3183         do-file-rename and do-webcore-rename. I inadvertently left this out of the patch.
3184
3185         * Scripts/VCSUtils.pm:
3186
3187 2010-10-21  Eric Seidel  <eric@webkit.org>
3188
3189         Reviewed by Adam Barth.
3190
3191         mac-ews is not properly releasing patches
3192         https://bugs.webkit.org/show_bug.cgi?id=48076
3193
3194         mac-ews overrides process_work_item, so it was not calling
3195         release_work_item like the default process_work_item would.
3196         To fix this I made all the status-reporting methods just
3197         release the patch.  I expect we'll iterate on this design further.
3198
3199         * Scripts/webkitpy/common/net/statusserver.py:
3200         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
3201         * Scripts/webkitpy/tool/commands/queues.py:
3202         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3203
3204 2010-10-18  Martin Robinson  <mrobinson@igalia.com>
3205
3206         Reviewed by Gustavo Noronha Silva.
3207
3208         [GTK] mathml flaky tests after adding SystemFonts to the theme
3209         https://bugs.webkit.org/show_bug.cgi?id=47727
3210
3211         The Fontconfig setup in the GTK+ now specifically checks for and
3212         loads the STIX fonts for MathML tests.
3213
3214         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3215         (initializeFonts): Add hooks for loading the STIX fonts specifically.
3216
3217 2010-10-20  Adam Roben  <aroben@apple.com>
3218
3219         Test that the UI client gets notified when WKView receives a WM_CLOSE
3220         message
3221
3222         Test for <http://webkit.org/b/48044> <rdar://problem/8488446> Pressing
3223         Ctrl+W when viewing a full-page PDF destroys the WKView but doesn't
3224         close its parent window
3225
3226         Reviewed by Jon Honeycutt.
3227
3228         * TestWebKitAPI/Tests/WebKit2/win/WMCloseCallsUIClientClose.cpp: Added.
3229         (TestWebKitAPI::close): Record that this function was called.
3230         (TestWebKitAPI::WebKit2_WMCloseCallsUIClientClose): Create a WKView,
3231         send it a WM_CLOSE message, and test that the UI client's close
3232         callback got called. Note that this will hang if the bug is
3233         reintroduced; <http://webkit.org/b/48043> covers making TestWebKitAPI
3234         able to handle hangs.
3235
3236         * TestWebKitAPI/win/TestWebKitAPI.vcproj: Added the new test.
3237
3238 2010-10-20  Adam Roben  <aroben@apple.com>
3239
3240         Make prepare-ChangeLog much faster when using git
3241
3242         This change also seems to make it detect renames better in some cases.
3243
3244         Fixes <http://webkit.org/b/48040> prepare-ChangeLog is slow when using
3245         git
3246
3247         Reviewed by David Kilzer.
3248
3249         * Scripts/prepare-ChangeLog:
3250         (statusCommand):
3251         (createPatchCommand):
3252         Use "-M -C" instead of "-C -C -M" to tell git-diff to detect renames
3253         and copies. The two "-C"s were making it read many more files than were
3254         necessary.
3255
3256 2010-10-21  Sam Weinig  <sam@webkit.org>
3257
3258         Reviewed by Adam Roben.
3259
3260         Null frame passed when running alert from UserScript run at document start
3261         <rdar://problem/8573809>
3262         https://bugs.webkit.org/show_bug.cgi?id=48036
3263
3264         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3265         * TestWebKitAPI/InjectedBundleController.cpp:
3266         (TestWebKitAPI::InjectedBundleController::didReceiveMessage):
3267         (TestWebKitAPI::InjectedBundleController::initializeTestNamed):
3268         * TestWebKitAPI/InjectedBundleController.h:
3269         * TestWebKitAPI/InjectedBundleTest.h:
3270         (TestWebKitAPI::InjectedBundleTest::initialize):
3271         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
3272         * TestWebKitAPI/Tests/WebKit2/DocumentStartUserScriptAlertCrash.cpp: Added.
3273         (TestWebKitAPI::runJavaScriptAlert):
3274         (TestWebKitAPI::TEST):
3275         * TestWebKitAPI/Tests/WebKit2/DocumentStartUserScriptAlertCrash_Bundle.cpp: Added.
3276         (TestWebKitAPI::DocumentStartUserScriptAlertCrashTest::DocumentStartUserScriptAlertCrashTest):
3277         (TestWebKitAPI::DocumentStartUserScriptAlertCrashTest::initialize):
3278         Add test for invoking an alert during a UserScript run at document start.
3279
3280 2010-10-21  Andreas Kling  <kling@webkit.org>
3281
3282         Reviewed by Adam Roben.
3283
3284         TestNetscapePlugIn: Actually call NPP_SetWindow on Unix
3285
3286         We were just returning NPERR_NO_ERROR previously.
3287
3288         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
3289         (webkit_test_plugin_set_window):
3290
3291 2010-10-21  David Kilzer  <ddkilzer@apple.com>
3292
3293         <http://webkit.org/b/48051> Move macros from DumpRenderTreeMac.h to config.h
3294
3295         Reviewed by Adam Roben.
3296
3297         * DumpRenderTree/config.h: Moved macros to here...
3298         * DumpRenderTree/mac/DumpRenderTreeMac.h: ...from here.
3299         Addresses a FIXME comment from r28419.
3300
3301 2010-10-21  David Kilzer  <ddkilzer@apple.com>
3302
3303         <http://webkit.org/b/48047> Fix warnings found by check-Xcode-source-file-types
3304
3305         Reviewed by Adam Roben.
3306
3307         Fixes the following warnings:
3308
3309             WARNING: Incorrect file type 'sourcecode.cpp.objcpp' for file 'PixelDumpSupport.cpp'.
3310             WARNING: Incorrect file type 'sourcecode.cpp.objcpp' for file 'LayoutTestController.cpp'.
3311             WARNING: Incorrect file type 'sourcecode.cpp.objcpp' for file 'WorkQueue.cpp'.
3312             WARNING: Incorrect file type 'sourcecode.cpp.objcpp' for file 'cg/PixelDumpSupportCG.cpp'.
3313             WARNING: Incorrect file type 'sourcecode.cpp.objcpp' for file 'GCController.cpp'.
3314             5 issues found for WebKitTools/DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj.
3315
3316         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3317         Removed explicit file type for *.cpp files above to make them
3318         match other C++ source files.  This exposed the fact that
3319         DumpRenderTreeMac.h was not safe to include in C++ source, which
3320         caused PixelDumpSupport.cpp and PixelDumpSupportCG.cpp to fail
3321         to build.
3322         * DumpRenderTree/PixelDumpSupport.cpp: Adjusted header order.
3323         * DumpRenderTree/mac/DumpRenderTreeMac.h: Made safe to include
3324         in C++ source files.  Included CoreFoundation/CoreFoundation.h
3325         to make sure all CF types were defined.  Removed CFStringRef
3326         typedef.
3327
3328 2010-10-21  Adam Roben  <aroben@apple.com>
3329
3330         Attempt to fix plugins/pass-different-npp-struct.html on GTK.
3331
3332         See <http://webkit.org/b/47690>.
3333
3334         * GNUmakefile.am: Added PassDifferentNPPStruct.cpp.
3335
3336 2010-10-21  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3337
3338         Reviewed by Csaba Osztrogonác.
3339
3340         [NRWT] Set ImageDiff path on Qt port
3341         https://bugs.webkit.org/show_bug.cgi?id=48052
3342
3343         * Scripts/webkitpy/layout_tests/port/qt.py:
3344
3345 2010-10-21  Adam Barth  <abarth@webkit.org>
3346
3347         Reviewed by Eric Seidel.
3348
3349         Don't CC authors of flaky tests
3350         https://bugs.webkit.org/show_bug.cgi?id=48038
3351
3352         Some authors found this too annoying.  We'll look for another way to
3353         close the flaky test loop.
3354
3355         * Scripts/webkitpy/tool/commands/queues.py:
3356         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3357
3358 2010-10-21  Eric Seidel  <eric@webkit.org>
3359
3360         Reviewed by Adam Barth.
3361
3362         commit-queue should run run-webkit-tests with --no-new-test-results
3363         https://bugs.webkit.org/show_bug.cgi?id=47998
3364
3365         * Scripts/webkitpy/tool/steps/runtests.py:
3366         * Scripts/webkitpy/tool/steps/steps_unittest.py:
3367
3368 2010-10-21  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3369
3370         Reviewed by Ojan Vafai.
3371
3372         [NRWT] Get child process number from an environment variable
3373         https://bugs.webkit.org/show_bug.cgi?id=47981
3374
3375         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3376
3377 2010-10-20  David Kilzer  <ddkilzer@apple.com>
3378
3379         <http://webkit.org/b/47754> New script to verify explicit source file types in Xcode project files
3380
3381         Reviewed by Darin Adler.
3382
3383         The script parses an Xcode project file and makes sure the file
3384         extension matches the explicit file type set for all source
3385         files.  Note that the majority of source files will have their
3386         type set by Xcode, so the script doesn't check them since there
3387         is no need to second-guess Xcode.
3388
3389         * Scripts/check-Xcode-source-file-types: Added.  Code borrowed
3390         heavily from sort-Xcode-project-file.
3391
3392 2010-10-20  Adam Roben  <aroben@apple.com>
3393         
3394         Windows build fix
3395
3396         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp: Added missing #include.
3397
3398 2010-10-20  Adam Roben  <aroben@apple.com>
3399
3400         Qt test fix
3401
3402         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro: Added
3403         PassDifferentNPPStruct.cpp.
3404
3405 2010-10-20  Sam Weinig  <sam@webkit.org>
3406
3407         Reviewed by Adam Roben.
3408
3409         Add ability to test injected bundle API using TestWebKitAPI
3410         https://bugs.webkit.org/show_bug.cgi?id=48027
3411
3412         * TestWebKitAPI/InjectedBundleController.cpp: Added.
3413         * TestWebKitAPI/InjectedBundleController.h: Added.
3414         Shared main object for bundle functionality.
3415
3416         * TestWebKitAPI/InjectedBundleMain.cpp: Added.
3417         Bundle entry point.
3418
3419         * TestWebKitAPI/InjectedBundleTest.h: Added.
3420         Base class for which the bundle portion of a test derives from. 
3421
3422         * TestWebKitAPI/PlatformUtilities.h:
3423         * TestWebKitAPI/PlatformUtilities.cpp: Added.
3424         * TestWebKitAPI/mac/PlatformUtilitiesMac.mm:
3425         * TestWebKitAPI/win/PlatformUtilitiesWin.cpp:
3426         (TestWebKitAPI::Util::createInjectedBundlePath):
3427         (TestWebKitAPI::Util::createURLForResource):
3428         (TestWebKitAPI::Util::URLForNonExistentResource):
3429         Add helper to create a context with the shared injected bundle,
3430         and send the initial message to set up the test.
3431
3432         * TestWebKitAPI/Configurations/InjectedBundle.xcconfig: Added.
3433         * TestWebKitAPI/InjectedBundle-Info.plist: Added.
3434         Add mac configuration files.
3435
3436         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
3437         Add the new files.
3438
3439         * TestWebKitAPI/Tests/WebKit2/InjectedBundleBasic.cpp: Added.
3440         * TestWebKitAPI/Tests/WebKit2/InjectedBundleBasic_Bundle.cpp: Added.
3441         Add a simple initial bundle test.
3442
3443 2010-10-20  Eric Seidel  <eric@webkit.org>
3444
3445         Unreviewed.  Fixing /release-patch url used by the EWS bots.
3446
3447         EWS should test patches with r+
3448         https://bugs.webkit.org/show_bug.cgi?id=35460
3449
3450         * Scripts/webkitpy/common/net/statusserver.py:
3451          - I changed the URL during development, and 404s are
3452          intentionally silenced during release_patch.
3453
3454 2010-10-14  Adam Roben  <aroben@apple.com>
3455
3456         Test that passing a different NPP struct back to the browser doesn't
3457         cause an assertion failure
3458
3459         Test for <http://webkit.org/b/47690> <rdar://problem/8553020>
3460         Assertion failure in NetscapePlugin::fromNPP when using Shockwave in
3461         WebKit2
3462
3463         Reviewed by John Sullivan.
3464
3465         * DumpRenderTree/TestNetscapePlugIn/Tests/PassDifferentNPPStruct.cpp: Copied from WebKitTools/DumpRenderTree/TestNetscapePlugIn/Tests/win/GetValueNetscapeWindow.cpp.
3466         (PassDifferentNPPStruct::PassDifferentNPPStruct): Initialize our
3467         members.
3468         (PassDifferentNPPStruct::NPP_SetWindow): Pass a different NPP to the
3469         browser than the one it gave us in NPP_New and see if it works.
3470
3471         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3472         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
3473         Added PassDifferentNPPStruct.
3474
3475 2010-10-19  Adam Roben  <aroben@apple.com>
3476
3477         Gently nudge old-run-webkit-tests toward working with Win32 Perl
3478
3479         This makes old-run-webkit-tests able to build DRT and find all the
3480         tests to run. It even invokes DRT and passes it the list of tests. But
3481         DRT ends up hung blocking on I/O.
3482
3483         Fixes <http://webkit.org/b/47961> Get old-run-webkit-tests mostly
3484         working with Win32 Perl
3485
3486         Reviewed by David Kilzer.
3487
3488         * Scripts/old-run-webkit-tests:
3489           - Use File::Spec instead of manually concatenating paths
3490           - Use dirname instead of manually stripping off the base name
3491           - Use isCygwin/isWindows/isAppleWinWebKit more judiciously
3492           - Explicitly invoke Perl when running Perl scripts
3493           - Quote paths when using them in regular expressions to allow them
3494             to include characters that have special meanings in regular
3495             expressions
3496
3497         * Scripts/run-webkit-tests: Use File::Spec instead of manually
3498         concatenating paths.
3499
3500         * Scripts/webkitdirs.pm:
3501           - Remove the unused $windowsTmpPath variable
3502           - Use isCygwin/isWindows/isAppleWinWebKit more judiciously
3503           - Only pass paths to cygpath when using Cygwin Perl
3504           - Only use pdevenv when using Cygwin Perl, for now
3505
3506 2010-10-20  Daniel Bates  <dbates@rim.com>
3507
3508         Reviewed by Martin Robinson.
3509
3510         Add Git-support to do-file-rename
3511         https://bugs.webkit.org/show_bug.cgi?id=48015
3512
3513         Also, abstracts the SCM move/rename functionality in do-file-rename and
3514         do-webcore-rename into a common function VCSUtils::scmMoveOrRenameFile().
3515
3516         Currently, do-file-rename is hard coded to assume the SCM is Subversion.
3517         Instead, we should abstract the rename logic to be SCM-independent. This
3518         will allow us to add Git support as well move such functionality into
3519         our SCM library VCSUtils, where it can be shared by do-webcore-rename.
3520
3521         * Scripts/VCSUtils.pm:
3522           - Added function scmMoveOrRenameFile.
3523         * Scripts/do-file-rename: Modified to call VCSUtils::scmMoveOrRenameFile().
3524         * Scripts/do-webcore-rename: Ditto.
3525
3526 2010-10-20  Adam Roben  <aroben@apple.com>
3527
3528         Fix old-run-webkit-tests when there's a space in the path to DRT
3529
3530         Reviewed by Jon Honeycutt.
3531
3532         * Scripts/old-run-webkit-tests: Quote the path to DRT before executing
3533         it.
3534
3535 2010-10-20  Kenneth Russell  <kbr@google.com>
3536
3537         Reviewed by Dimitri Glazkov.
3538
3539         Regression in chromium_gpu_unittests after r70175
3540         https://bugs.webkit.org/show_bug.cgi?id=48008
3541
3542         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
3543
3544 2010-10-20  Kenneth Russell  <kbr@google.com>
3545
3546         Reviewed by James Robinson.
3547
3548         chromium_gpu port of new-run-webkit-tests must do Linux -> Win expectations fallback
3549         https://bugs.webkit.org/show_bug.cgi?id=48005
3550
3551         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
3552
3553 2010-10-20  Adam Barth  <abarth@webkit.org>
3554
3555         Reviewed by Eric Seidel.
3556
3557         webkit-patch doesn't get along with git rm
3558         https://bugs.webkit.org/show_bug.cgi?id=47940
3559
3560         Turns out we need to pass "--" to tell git this is a path.
3561
3562         * Scripts/webkitpy/common/checkout/scm.py:
3563
3564 2010-10-20  Robert Hogan  <robert@webkit.org>
3565
3566         Reviewed by Antonio Gomes.
3567
3568         [Qt] Fix layoutTestController.getJsObjectCount
3569
3570         Qt Bridge doesn't know size_t so pass result as unsigned int.
3571
3572         Unskip fast/dom/gc-10.html
3573
3574         https://bugs.webkit.org/show_bug.cgi?id=47931
3575
3576         * DumpRenderTree/qt/GCControllerQt.cpp:
3577         (GCController::getJSObjectCount):
3578         * DumpRenderTree/qt/GCControllerQt.h:
3579
3580 2010-10-20  Sheriff Bot  <webkit.review.bot@gmail.com>
3581
3582         Unreviewed, rolling out r70149.
3583         http://trac.webkit.org/changeset/70149
3584         https://bugs.webkit.org/show_bug.cgi?id=47989
3585
3586         "Build breaks in mac and win" (Requested by satish on
3587         #webkit).
3588
3589         * DumpRenderTree/LayoutTestController.cpp:
3590         (setMockSpeechInputResultCallback):
3591         * DumpRenderTree/LayoutTestController.h:
3592         * DumpRenderTree/chromium/LayoutTestController.cpp:
3593         (LayoutTestController::setMockSpeechInputResult):
3594         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3595         (LayoutTestController::setMockSpeechInputResult):
3596
3597 2010-10-20  Leandro Gracia Gil  <leandrogracia@google.com>
3598
3599         Reviewed by Jeremy Orlow.
3600
3601         Added a second parameter to setMockSpeechInputResult for
3602         the language used in speech input.
3603         https://bugs.webkit.org/show_bug.cgi?id=47089
3604
3605         * DumpRenderTree/LayoutTestController.cpp:
3606         (setMockSpeechInputResultCallback):
3607         * DumpRenderTree/LayoutTestController.h:
3608         * DumpRenderTree/chromium/LayoutTestController.cpp:
3609         (LayoutTestController::setMockSpeechInputResult):
3610         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3611         (LayoutTestController::setMockSpeechInputResult):
3612
3613 2010-10-19  Adam Roben  <aroben@apple.com>
3614
3615         Teach update-webkit-support-libs about the new versioning of
3616         WebKitSupportLibrary
3617
3618         Fixes <http://webkit.org/b/47915> update-webkit-support-libs should
3619         check version numbers instead of modification times
3620
3621         Reviewed by Sam Weinig.
3622
3623         * Scripts/update-webkit-support-libs: Fetch the expected version
3624         number from developer.apple.com, then compare with the version number
3625         of the extracted library and of the zipped library to see if anything
3626         needs to be done. Removed code that tracked the modified timestamp of
3627         the library, as it is no longer needed.
3628
3629 2010-10-19  Eric Seidel  <eric@webkit.org>
3630
3631         Reviewed by Adam Barth.
3632
3633         EWS should test patches with r+
3634         https://bugs.webkit.org/show_bug.cgi?id=35460
3635
3636         * QueueStatusServer/handlers/submittoews.py:
3637         * QueueStatusServer/model/queues.py:
3638         * Scripts/webkitpy/tool/bot/patchcollection.py: Removed.
3639         * Scripts/webkitpy/tool/bot/patchcollection_unittest.py: Removed.
3640         * Scripts/webkitpy/tool/commands/queues.py:
3641
3642 2010-10-19  Eric Seidel  <eric@webkit.org>
3643
3644         Reviewed by Adam Barth.
3645
3646         Teach feeder-queue how to feed the EWS bots
3647         https://bugs.webkit.org/show_bug.cgi?id=47943
3648
3649         queues.webkit.org already knew how to accept EWS submissions
3650         via /submit-to-ews.  This teaches the feeder queue how to post
3651         to that page with any new r? patches it sees.
3652
3653         * QueueStatusServer/model/activeworkitems_unitest.py: Added.
3654          - More unit testing is always a good thing.
3655         * Scripts/webkitpy/common/net/bugzilla.py:
3656         * Scripts/webkitpy/common/net/statusserver.py:
3657         * Scripts/webkitpy/tool/bot/feeders.py:
3658         * Scripts/webkitpy/tool/commands/queues.py:
3659         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3660         * Scripts/webkitpy/tool/mocktool.py:
3661
3662 2010-10-19  Eric Seidel  <eric@webkit.org>
3663
3664         Unreviewed, just fixing test-webkitpy.  I'm really on a roll today.
3665
3666         commit-queue gets stuck when release-patch returns 404
3667         https://bugs.webkit.org/show_bug.cgi?id=47935
3668
3669         Fix test-webkitpy and unittest NetworkTransaction.
3670
3671         * Scripts/webkitpy/common/net/networktransaction.py:
3672         * Scripts/webkitpy/common/net/networktransaction_unittest.py:
3673
3674 2010-10-19  Eric Seidel  <eric@webkit.org>
3675
3676         Unreviewed.
3677
3678         commit-queue gets stuck when release-patch returns 404
3679         https://bugs.webkit.org/show_bug.cgi?id=47935
3680
3681         Turns out ClientForm gets upset if passed an int() instead of a string type.
3682         Yay for untestable code.
3683
3684         * Scripts/webkitpy/common/net/statusserver.py:
3685
3686 2010-10-19  Eric Seidel  <eric@webkit.org>
3687
3688         Unreviewed.  Will get Adam's commentary after his meeting
3689         for now this gets the commit-cluster back running.
3690
3691         commit-queue gets stuck when release-patch returns 404
3692         https://bugs.webkit.org/show_bug.cgi?id=47935
3693
3694         I taught NetworkTransaction the basics of 404 handling.
3695         We'll want to go back and teach it how to handle urllib2 404's too
3696         and then deploy it to the places that want it.
3697
3698         * QueueStatusServer/handlers/releasepatch.py:
3699         * Scripts/webkitpy/common/net/buildbot.py:
3700         * Scripts/webkitpy/common/net/networktransaction.py:
3701         * Scripts/webkitpy/common/net/statusserver.py:
3702
3703 2010-10-19  Eric Seidel  <eric@webkit.org>
3704
3705         Unreviewed.  Fixing typos in my previous commit.
3706
3707         Make patch release explicit and not a magic part of "retry" status
3708         https://bugs.webkit.org/show_bug.cgi?id=47909
3709
3710         All of these typos again due to our inability to unit
3711         test much of this code.  I added one unit test where
3712         possible.  activeworkitems_unittest.py will be in a separate patch.
3713
3714         * QueueStatusServer/handlers/releasepatch.py:
3715         * QueueStatusServer/main.py:
3716         * QueueStatusServer/model/activeworkitems.py:
3717         * QueueStatusServer/model/workitems.py:
3718         * QueueStatusServer/model/workitems_unittest.py:
3719         * QueueStatusServer/templates/releasepatch.html:
3720         * Scripts/webkitpy/common/net/statusserver.py:
3721
3722 2010-10-19  Tony Chang  <tony@chromium.org>
3723
3724         Reviewed by Kent Tamura.
3725
3726         [chromium] Use webkit's TestNetscapePlugIn in DRT mac
3727         https://bugs.webkit.org/show_bug.cgi?id=47850
3728
3729         * DumpRenderTree/chromium/TestNetscapePlugIn/Info.plist: Added. Forked
3730             because we're going to name our plugin WebKitTestNetscapePlugIn
3731             temporarily until the chromium forked plugin goes away.
3732
3733 2010-10-19  Eric Seidel  <eric@webkit.org>
3734
3735         Reviewed by Adam Barth.
3736
3737         Make patch release explicit and not a magic part of "retry" status
3738         https://bugs.webkit.org/show_bug.cgi?id=47909
3739
3740         This moves us another step closer to running r+ patches on the EWS bots.
3741         Currently all bots just spam /update-work-items with their list of current
3742         work items.  queues.webkit.org uses that data for display.  As part of making
3743         the EWS run r+ patches, we're moving the official list of patches-to-process
3744         into the server, and feeding them out to bots one at a time.  We need to be
3745         able to remove patches from the queues one at a time instead of just spamming
3746         /update-work-items with a new complete list.  That's what this patch adds.
3747
3748         * QueueStatusServer/handlers/nextpatch.py:
3749         * QueueStatusServer/handlers/queuestatus.py:
3750         * QueueStatusServer/handlers/releasepatch.py: Added.
3751         * QueueStatusServer/handlers/statusbubble_unittest.py:
3752          - Fix a typo causing test failure.  This was not caught by the bots
3753            because they don't have AppEngineLauncher installed and thus don't run
3754            the QueueStatusServer tests.
3755         * QueueStatusServer/handlers/updatestatus.py:
3756         * QueueStatusServer/model/activeworkitems.py:
3757         * QueueStatusServer/templates/releasepatch.html: Added.
3758         * Scripts/webkitpy/common/net/statusserver.py:
3759         * Scripts/webkitpy/tool/commands/queues.py:
3760         * Scripts/webkitpy/tool/mocktool.py:
3761
3762 2010-10-19  Eric Seidel  <eric@webkit.org>
3763
3764         Reviewed by Adam Barth.
3765
3766         cr-mac bubble has caused status bubbles to wrap
3767         https://bugs.webkit.org/show_bug.cgi?id=47928
3768
3769         We now have too many EWSes to fit in the bugs.webkit.org
3770         status-bubble iframe when more than a couple EWS builds are pending.
3771         To fix this I've reduced the space taken up by queue position,
3772         and also moved cr-mac to the end of the list (since it's going to be
3773         triple-digits for a while).
3774
3775         * QueueStatusServer/model/queues.py:
3776         * QueueStatusServer/templates/statusbubble.html:
3777
3778 2010-10-19  Kenneth Russell  <kbr@google.com>
3779
3780         Reviewed by David Levin.
3781
3782         chromium_gpu port of new-run-webkit-tests must search chromium-gpu directory for expectations
3783         https://bugs.webkit.org/show_bug.cgi?id=47874
3784
3785         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
3786         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
3787
3788 2010-10-19  Tony Chang  <tony@chromium.org>
3789
3790         Reviewed by Kent Tamura.
3791
3792         [chromium] DumpRenderTree shouldn't put '.' in include path
3793         https://bugs.webkit.org/show_bug.cgi?id=47877
3794
3795         Fix include paths.
3796
3797         * DumpRenderTree/chromium/AccessibilityController.cpp:
3798         * DumpRenderTree/chromium/AccessibilityUIElement.cpp:
3799         * DumpRenderTree/chromium/AccessibilityUIElement.h:
3800         * DumpRenderTree/chromium/CppBoundClass.cpp:
3801         * DumpRenderTree/chromium/CppVariant.cpp:
3802         * DumpRenderTree/chromium/CppVariant.h:
3803         * DumpRenderTree/chromium/DRTDevToolsAgent.cpp:
3804         * DumpRenderTree/chromium/DRTDevToolsAgent.h:
3805         * DumpRenderTree/chromium/DRTDevToolsCallArgs.h:
3806         * DumpRenderTree/chromium/DRTDevToolsClient.cpp:
3807         * DumpRenderTree/chromium/DRTDevToolsClient.h:
3808         * DumpRenderTree/chromium/EventSender.cpp:
3809         * DumpRenderTree/chromium/EventSender.h:
3810         * DumpRenderTree/chromium/LayoutTestController.cpp:
3811         * DumpRenderTree/chromium/LayoutTestController.h:
3812         * DumpRenderTree/chromium/MockSpellCheck.cpp:
3813         * DumpRenderTree/chromium/NotificationPresenter.cpp:
3814         * DumpRenderTree/chromium/NotificationPresenter.h:
3815         * DumpRenderTree/chromium/PlainTextController.cpp:
3816         * DumpRenderTree/chromium/Task.cpp:
3817         * DumpRenderTree/chromium/TestNavigationController.h:
3818         * DumpRenderTree/chromium/TestShell.cpp:
3819         * DumpRenderTree/chromium/TestWebWorker.h:
3820         * DumpRenderTree/chromium/TextInputController.cpp:
3821         * DumpRenderTree/chromium/WebPreferences.cpp:
3822         * DumpRenderTree/chromium/WebPreferences.h:
3823         * DumpRenderTree/chromium/WebThemeEngineDRT.cpp:
3824         * DumpRenderTree/chromium/WebThemeEngineDRT.h:
3825         * DumpRenderTree/chromium/WebViewHost.cpp:
3826         * DumpRenderTree/chromium/WebViewHost.h:
3827
3828 2010-10-19  Sam Weinig  <sam@webkit.org>
3829
3830         Reviewed by Dan Bernstein.
3831
3832         Need to include WKErrorRef in the WKPageDidFail... functions
3833         https://bugs.webkit.org/show_bug.cgi?id=47871
3834
3835         Update tools for new parameter in failure callbacks.
3836
3837         * MiniBrowser/mac/BrowserWindowController.m:
3838         (didFailProvisionalLoadWithErrorForFrame):
3839         (didFailLoadWithErrorForFrame):
3840         (-[BrowserWindowController updateProvisionalURLForFrame:]):
3841         * TestWebKitAPI/Tests/WebKit2/FailedLoad.cpp:
3842         (TestWebKitAPI::didFailProvisionalLoadWithErrorForFrame):
3843         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3844         (WTR::InjectedBundlePage::didFailProvisionalLoadWithErrorForFrame):
3845         (WTR::InjectedBundlePage::didFailLoadWithErrorForFrame):
3846         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
3847
3848 2010-10-19  Luiz Agostini  <luiz.agostini@openbossa.org>
3849
3850         Reviewed by Antti Koivisto.
3851
3852         [Qt] WebKit2 MacOS build fix
3853         https://bugs.webkit.org/show_bug.cgi?id=47897
3854
3855         Qt WebKit2 MacOS build fix.
3856
3857         * WebKitTestRunner/PlatformWebView.h:
3858         * WebKitTestRunner/qt/TestControllerQt.cpp:
3859
3860 2010-10-19  Adam Barth  <abarth@webkit.org>
3861
3862         Reviewed by Eric Seidel.
3863
3864         webkit-patch stats the filesystem too many times
3865         https://bugs.webkit.org/show_bug.cgi?id=47883
3866
3867         This patch attempts to cache the list of changed files more agressively
3868         and to use that list to compute the diff instead of stating the file
3869         system again.
3870
3871         * Scripts/webkitpy/common/checkout/api.py:
3872         * Scripts/webkitpy/common/checkout/scm.py:
3873         * Scripts/webkitpy/tool/mocktool.py:
3874         * Scripts/webkitpy/tool/steps/abstractstep.py:
3875         * Scripts/webkitpy/tool/steps/editchangelog.py:
3876         * Scripts/webkitpy/tool/steps/preparechangelog.py:
3877
3878 2010-10-19  David Kilzer  <ddkilzer@apple.com>
3879
3880         <http://webkit.org/b/47741> Make sort-Xcode-project-file a little more friendly
3881
3882         Reviewed by Darin Adler.
3883
3884         * Scripts/sort-Xcode-project-file:
3885         - Don't print an error message about missing arguments when
3886           -h|--help is used.
3887         - Allow Xcode project files to be specified as Project.xcodeproj
3888           instead of Project.xcodeproj/project.pbxproj.
3889
3890 2010-10-18  Jocelyn Turcotte  <jocelyn.turcotte@nokia.com>
3891
3892         Reviewed by Andreas Kling.
3893
3894         [Qt] QtTestBrowser: Prevent calling load() directly from loadFinished() in robot mode.
3895         https://bugs.webkit.org/show_bug.cgi?id=47809
3896
3897         Connecting a call to load from the loadFinished signal can cause
3898         re-entrance crashes in WebCore. This patch uses a timer to do so,
3899         also giving some time to subsequent frames to finish loading.
3900
3901         * QtTestBrowser/urlloader.cpp:
3902         (UrlLoader::UrlLoader):
3903         (UrlLoader::loadNext):
3904         (UrlLoader::checkIfFinished):
3905         (UrlLoader::frameLoadStarted):
3906         (UrlLoader::frameLoadFinished):
3907         * QtTestBrowser/urlloader.h:
3908
3909 2010-10-19  Sergio Villar Senín  <svillar@igalia.com>
3910
3911         Reviewed by Xan Lopez.
3912
3913         [GTK] http/history tests are failing
3914         https://bugs.webkit.org/show_bug.cgi?id=36173
3915
3916         Clear the history each time a test is run. Return the actual
3917         history item count when calling
3918         LayoutTestController::webHistoryItemCount
3919
3920         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3921         (resetDefaultsToConsistentValues):
3922         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3923         (LayoutTestController::webHistoryItemCount):
3924
3925 2010-10-18  Eric Seidel  <eric@webkit.org>
3926
3927         Reviewed by Adam Barth.
3928
3929         Add Chromium Mac EWS to the list of queues at queues.webkit.org
3930         https://bugs.webkit.org/show_bug.cgi?id=47878
3931
3932         * QueueStatusServer/model/queues.py:
3933
3934 2010-10-18  Eric Seidel  <eric@webkit.org>
3935
3936         Reviewed by Adam Barth.
3937
3938         queues.webkit.org/next_patch is always 404
3939         https://bugs.webkit.org/show_bug.cgi?id=47881
3940
3941         With the addition of the Queue class, I changed most of the
3942         code to lookup WorkItems using get_or_insert with a key_name
3943         instead of WorkItems.all().filter(queue_name=).
3944         Because the new get_or_insert code uses an explicit key_name
3945         (which is obviously different from the previously autogenerated
3946         ones), there were new WorkItem records created for each queue.
3947         However, some parts of the code still use WorkItems.all().filter,
3948         thus some parts were getting the new record and some parts the old record.
3949
3950         The same basic bug was occurring with ActiveWorkItems, because I
3951         changed the key_name for that class as well.
3952
3953         To fix this I've moved more of the code over to using Queue.*work_items.
3954         I've also enabled the datastore_admin (new in GAE 1.3.8) so that
3955         we can go delete the old WorkItems records.
3956         I also changed remote_api to use the new builtin: syntax (also added in GAE 1.3.8).
3957
3958         * QueueStatusServer/app.yaml:
3959         * QueueStatusServer/handlers/queuestatus.py:
3960         * QueueStatusServer/handlers/recentstatus.py:
3961         * QueueStatusServer/handlers/updatestatus.py:
3962         * QueueStatusServer/handlers/updateworkitems.py:
3963
3964 2010-10-18  Eric Seidel  <eric@webkit.org>
3965
3966         Reviewed by Adam Barth.
3967
3968         Correct a bunch of typos in QueueStatusServer
3969         https://bugs.webkit.org/show_bug.cgi?id=47880
3970
3971         These are all due to our complete lack of unit testing in QueueStatusServer.
3972         I added a couple unit tests to cover a few of these fixes, but most of these
3973         are still only caught by actually running the application.
3974
3975         * QueueStatusServer/handlers/nextpatch.py:
3976         * QueueStatusServer/handlers/statusbubble.py:
3977         * QueueStatusServer/handlers/statusbubble_unittest.py: Added.
3978         * QueueStatusServer/handlers/updateworkitems.py:
3979         * QueueStatusServer/model/attachment.py:
3980         * QueueStatusServer/model/queuepropertymixin.py:
3981         * QueueStatusServer/model/queuepropertymixin_unittest.py:
3982         * QueueStatusServer/model/workitems.py:
3983         * QueueStatusServer/model/workitems_unittest.py: Added.
3984
3985 2010-10-18  Adam Barth  <abarth@webkit.org>
3986
3987         Disable this test because it's failing on the bots and the authors
3988         aren't around to fix it.
3989
3990         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
3991
3992 2010-10-18  Adam Barth  <abarth@webkit.org>
3993
3994         Reviewed by Eric Seidel.
3995
3996         CC authors of flaky tests when the commit-queue hits a flaky test
3997         https://bugs.webkit.org/show_bug.cgi?id=47872
3998
3999         * Scripts/webkitpy/common/checkout/api.py:
4000         * Scripts/webkitpy/common/net/layouttestresults.py:
4001         * Scripts/webkitpy/tool/commands/queues.py:
4002         * Scripts/webkitpy/tool/commands/queues_unittest.py:
4003         * Scripts/webkitpy/tool/mocktool.py:
4004
4005 2010-10-18  Eric Seidel  <eric@webkit.org>
4006
4007         Reviewed by Adam Barth.
4008
4009         Make it possible to run a chromium-mac-ews builder
4010         https://bugs.webkit.org/show_bug.cgi?id=47876
4011
4012         Since we can't run Mac OS X in a VM, we need to only run committer patches.
4013         There was a multiple inheritance problem which was holding this patch back,
4014         but I decided to just ignore the problem and go with a functional hack for now.
4015
4016         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
4017         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
4018
4019 2010-10-18  Kenneth Russell  <kbr@google.com>
4020
4021         Reviewed by Tony Chang.
4022
4023         Regression in run_webkit_tests_unittest from r70017
4024         https://bugs.webkit.org/show_bug.cgi?id=47875
4025
4026         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
4027
4028 2010-10-18  Adam Barth  <abarth@webkit.org>
4029
4030         Reviewed by Eric Seidel.
4031
4032         Implement webkit-patch suggest-reviewers
4033         https://bugs.webkit.org/show_bug.cgi?id=47866
4034
4035         * Scripts/webkitpy/common/checkout/api.py:
4036             - The main logic.  We look at the last five changes to each
4037               modified (non-ChangeLog) file and collect up the reviewers of
4038               those changes as well as the authors of those changes who are
4039               reviewers.
4040         * Scripts/webkitpy/common/checkout/api_unittest.py:
4041             - Test the logic with some fun mocks.
4042         * Scripts/webkitpy/common/checkout/scm.py:
4043             - Fix a bug when you have local git commits.
4044         * Scripts/webkitpy/common/checkout/scm_unittest.py:
4045             - Test that the bug is fixed.
4046         * Scripts/webkitpy/tool/commands/queries.py:
4047             - Add the query.
4048
4049 2010-10-18  Kenneth Russell  <kbr@google.com>
4050
4051         Reviewed by Eric Seidel.
4052
4053         new-run-webkit-tests produces corrupt PNG baselines on Windows
4054         https://bugs.webkit.org/show_bug.cgi?id=47867
4055
4056         * Scripts/webkitpy/layout_tests/port/base.py:
4057
4058 2010-10-18  Eric Seidel  <eric@webkit.org>
4059
4060         Reviewed by Adam Barth.
4061
4062         Make it possible to submit patches to the EWS bots
4063         https://bugs.webkit.org/show_bug.cgi?id=47869
4064
4065         * QueueStatusServer/handlers/nextpatch.py:
4066          - Move more logic into Queue, so that it can be shared with SubmitToEWS.
4067         * QueueStatusServer/handlers/queuestatus.py:
4068          - Fix two typos from a previous commit.
4069         * QueueStatusServer/handlers/submittoews.py: Added.
4070         * QueueStatusServer/handlers/updatestatus.py:
4071          - Use the new is_retry_request method to share this (hacky) code with SubmitToEWS
4072         * QueueStatusServer/main.py:
4073          - Add /submit-to-ews
4074         * QueueStatusServer/model/queuepropertymixin.py:
4075          - Fix circular imports caused by adding Queue.work_items()
4076         * QueueStatusServer/model/queues.py:
4077          - Add work_items() and active_work_items()
4078         * QueueStatusServer/model/queuestatus.py:
4079         * QueueStatusServer/model/workitems.py:
4080          - Add transaction-safe add/remove methods.
4081         * QueueStatusServer/templates/submittoews.html: Added.
4082
4083 2010-10-18  Eric Seidel  <eric@webkit.org>
4084
4085         Reviewed by Adam Barth.
4086
4087         Deploy Queue class in more places throughout QueueStatusServer
4088         https://bugs.webkit.org/show_bug.cgi?id=47855
4089
4090         I also caught two typos from the previous change.  Unfortunately
4091         I don't yet know how to unittest request handlers yet.
4092
4093         * QueueStatusServer/handlers/dashboard.py:
4094         * QueueStatusServer/handlers/statusbubble.py:
4095         * QueueStatusServer/handlers/updateworkitems.py:
4096         * QueueStatusServer/model/activeworkitems.py:
4097         * QueueStatusServer/model/attachment.py:
4098         * QueueStatusServer/model/queuepropertymixin.py: Added.
4099         * QueueStatusServer/model/queuepropertymixin_unittest.py: Added.
4100         * QueueStatusServer/model/queuestatus.py:
4101         * QueueStatusServer/model/workitems.py:
4102
4103 2010-10-18  Adam Barth  <abarth@webkit.org>
4104
4105         Reviewed by Eric Seidel.
4106
4107         scm.py should be able tell us what revisions made changes to a given file
4108         https://bugs.webkit.org/show_bug.cgi?id=47863
4109
4110         Look again, your SCM.py can now log files.
4111
4112         * Scripts/webkitpy/common/checkout/scm.py:
4113         * Scripts/webkitpy/common/checkout/scm_unittest.py:
4114
4115 2010-10-18  Dirk Pranke  <dpranke@chromium.org>
4116
4117         Reviewed by Eric Siedel.
4118
4119         new-run-webkit-tests: clean up the options-parsing code in the port
4120         classes.
4121         
4122         This change modifies the Port interface to have a get_option() and
4123         set_option_default() method for accessing the options argument
4124         passed to the constructor. If the constructor is not passed an
4125         options argument, we default to a MockOptions() argument from
4126         mocktool, which has the same semantics we want.
4127
4128         Note that there is a disadvantage to port.get_option('foo') over
4129         port._options.foo, which is that you lose some of the checking
4130         for whether 'foo' is set (typos result in the default value, not
4131         an exception being raised. This is desired in this case, since the
4132         Port class is not allowed to assume that options does have any
4133         particular values set, and so this change ensures that all of
4134         the subclasses are following the same, intended, logic.
4135
4136         Arguably this is the wrong semantics to have, and the Port
4137         classes should be able to assume a default set of
4138         attributes/arguments, but that change will need to wait for a
4139         different CL where we can modify new-run-webkit-tests to pull a
4140         list of arguments from the port factory routines.
4141
4142         Also, add unit tests for webkitpy.tool.mocktool.MockOptions .
4143
4144         https://bugs.webkit.org/show_bug.cgi?id=47510
4145
4146         * Scripts/webkitpy/layout_tests/port/base.py:
4147         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
4148         * Scripts/webkitpy/layout_tests/port/chromium.py:
4149         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
4150         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
4151         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
4152         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
4153         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
4154         * Scripts/webkitpy/layout_tests/port/dryrun.py:
4155         * Scripts/webkitpy/layout_tests/port/factory_unittest.py:
4156         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
4157         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
4158         * Scripts/webkitpy/layout_tests/port/test.py:
4159         * Scripts/webkitpy/layout_tests/port/webkit.py:
4160         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
4161         * Scripts/webkitpy/tool/mocktool_unittest.py: Added.
4162
4163 2010-10-18  Dirk Pranke  <dpranke@chromium.org>
4164
4165         Reviewed by Eric Seidel.
4166
4167         Re-submit a revised version of r69638 - enabling new-run-webkit-tests
4168         under cygwin. The initial version had a bug in base:uri_to_test_name
4169         that was causing tests to fail. This version corrects that bug, but
4170         also makes the code safer by calling cygpath more reliably, and
4171         leaving a long-running cygpath process open.
4172
4173         This patch also corrects a couple of minor bugs in http_lock_unittest,
4174         chromium_unittest, and dedpulicate_tests_unittest that showed up
4175         while testing this.
4176
4177         https://bugs.webkit.org/show_bug.cgi?id=47220
4178
4179         * Scripts/webkitpy/common/system/path.py:
4180         * Scripts/webkitpy/common/system/path_unittest.py:
4181         * Scripts/webkitpy/layout_tests/deduplicate_tests_unittest.py:
4182         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
4183         * Scripts/webkitpy/layout_tests/port/base.py:
4184         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
4185         * Scripts/webkitpy/layout_tests/port/chromium.py:
4186         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
4187         * Scripts/webkitpy/layout_tests/port/http_lock_unittest.py:
4188         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
4189
4190 2010-10-18  Eric Seidel  <eric@webkit.org>
4191
4192         Reviewed by Adam Barth.
4193
4194         Add Queue class and add minimal unittesting of QueueStatusServer code
4195         https://bugs.webkit.org/show_bug.cgi?id=47847
4196
4197         * QueueStatusServer/handlers/dashboard.py:
4198         * QueueStatusServer/handlers/queuestatus.py:
4199         * QueueStatusServer/handlers/recentstatus.py: