9533d39ac9d788ca154d30bc867824dcde8540d3
[WebKit-https.git] / WebKitTools / ChangeLog
1 2010-04-28  Fumitoshi Ukai  <ukai@chromium.org>
2
3         Unreviewed build fix.
4
5         * DumpRenderTree/chromium/TestShell.cpp:
6         (TestShell::dumpImage): format '%u' expects type 'unsigned int', but argument 2 has type 'size_t'.
7         * DumpRenderTree/chromium/WebViewHost.cpp:
8         (WebViewHost::willSendRequest): too few arguments for format.
9
10 2010-04-27  Shinichiro Hamaji  <hamaji@chromium.org>
11
12         Reviewed by Darin Adler and Eric Seidel.
13
14         Add layoutTestController.setPrinting()
15         https://bugs.webkit.org/show_bug.cgi?id=37203
16
17         * DumpRenderTree/LayoutTestController.cpp:
18         (LayoutTestController::LayoutTestController):
19         (setPrintingCallback):
20         (LayoutTestController::staticFunctions):
21         * DumpRenderTree/LayoutTestController.h:
22         (LayoutTestController::isPrinting):
23         (LayoutTestController::setIsPrinting):
24         * DumpRenderTree/mac/DumpRenderTree.mm:
25         (dump):
26
27 2010-04-27  Michael Nordman  <michaeln@google.com>
28
29         Reviewed by Dmitry Titov.
30
31         [Chromium] Add two things to the webkit API to support appcaches in workers.
32         1) WebURLRequest TargetTypes for worker and shared worker main resources.
33         2) Factory method on class WebCommonWorkerClient to
34            createApplicationCacheHost() for the associated worker.
35
36         https://bugs.webkit.org/show_bug.cgi?id=38147
37
38         * DumpRenderTree/chromium/TestWebWorker.h add a stub impl of the factory method 
39         (TestWebWorker::createApplicationCacheHost):
40
41 2010-04-27  Adam Barth  <abarth@webkit.org>
42
43         Reviewed by Eric Seidel.
44
45         run-bindings-tests doesn't work in an SVN checkout
46         https://bugs.webkit.org/show_bug.cgi?id=38225
47
48         Previously detect_scm_system needed an absolute path for SVN.  Now we
49         accept a relative path.
50
51         * Scripts/webkitpy/common/checkout/scm.py:
52         * Scripts/webkitpy/common/checkout/scm_unittest.py:
53
54 2010-04-27  Sam Weinig  <sam@webkit.org>
55
56         Reviewed by Stephanie Lewis.
57
58         Always build WebKit2 when building on SnowLeopard and later. 
59
60         * Scripts/build-webkit:
61
62 2010-04-27  Sam Weinig  <sam@webkit.org>
63
64         Reviewed by Maciej Stachowiak.
65
66         Fix for https://bugs.webkit.org/show_bug.cgi?id=38238
67         Allow both WebKit and WebKit2 to link to the same WebCore.framework
68
69         * Scripts/build-webkit: Remove the UMBRELLA_LDFLAGS override when building
70         WebCore for WebKit2, it is no longer necessary.
71
72 2010-04-27  James Robinson  <jamesr@chromium.org>
73
74         Reviewed by David Levin.
75
76         Fix a typo in chromium.py that causes NRWT to fail in --verbose
77         mode in a Chromium checkout.
78         https://bugs.webkit.org/show_bug.cgi?id=38234
79
80         * Scripts/webkitpy/layout_tests/port/chromium.py:
81
82 2010-04-27  Darin Fisher  <darin@chromium.org>
83
84         Reviewed by Dimitri Glazkov.
85
86         [Chromium] Remove deprecated form of didChangeLocationWithinPage
87         https://bugs.webkit.org/show_bug.cgi?id=38178
88
89         Switch over to implementing didNavigateWithinPage.
90
91         * DumpRenderTree/chromium/WebViewHost.cpp:
92         (TestWebViewDelegate::didNavigateWithinPage):
93         (WebViewHost::didChangeLocationWithinPage):
94         * DumpRenderTree/chromium/WebViewHost.h:
95
96 2010-04-27  Evan Martin  <evan@chromium.org>
97
98         Unreviewed.
99
100         Adding myself to commmitters list.
101
102         * Scripts/webkitpy/common/config/committers.py:
103
104 2010-04-27  Adam Barth  <abarth@webkit.org>
105
106         Reviewed by Darin Adler.
107
108         run-bindings-tests should use --reset-results instead of --overwrite
109         https://bugs.webkit.org/show_bug.cgi?id=38200
110
111         As requested by Ojan.
112
113         * Scripts/run-bindings-tests:
114
115 2010-04-27  Adam Barth  <abarth@webkit.org>
116
117         Reviewed by Eric Seidel.
118
119         REGRESSION(r58261): webkit-patch edit-changelogs is broken
120         https://bugs.webkit.org/show_bug.cgi?id=38204
121
122         In 58261, we added code to abstract step that interrogates the squash
123         and git_commit options, but it doesn't declare that it uses those
124         options.  That means any command that doesn't happen to declare those
125         options might be broken if it uses the cached_lookup mechanism.
126
127         * Scripts/webkitpy/tool/steps/abstractstep.py:
128         * Scripts/webkitpy/tool/steps/applypatch.py:
129         * Scripts/webkitpy/tool/steps/applypatchwithlocalcommit.py:
130         * Scripts/webkitpy/tool/steps/build.py:
131         * Scripts/webkitpy/tool/steps/checkstyle.py:
132         * Scripts/webkitpy/tool/steps/cleanworkingdirectory.py:
133         * Scripts/webkitpy/tool/steps/closebug.py:
134         * Scripts/webkitpy/tool/steps/closebugforlanddiff.py:
135         * Scripts/webkitpy/tool/steps/commit.py:
136         * Scripts/webkitpy/tool/steps/confirmdiff.py:
137         * Scripts/webkitpy/tool/steps/createbug.py:
138         * Scripts/webkitpy/tool/steps/ensurebuildersaregreen.py:
139         * Scripts/webkitpy/tool/steps/ensurelocalcommitifneeded.py:
140         * Scripts/webkitpy/tool/steps/obsoletepatches.py:
141         * Scripts/webkitpy/tool/steps/postcodereview.py:
142         * Scripts/webkitpy/tool/steps/postdiff.py:
143         * Scripts/webkitpy/tool/steps/preparechangelog.py:
144         * Scripts/webkitpy/tool/steps/runtests.py:
145         * Scripts/webkitpy/tool/steps/update.py:
146         * Scripts/webkitpy/tool/steps/updatechangelogswithreviewer.py:
147         * Scripts/webkitpy/tool/steps/validatereviewer.py:
148
149 2010-04-27  Eric Seidel  <eric@webkit.org>
150
151         Reviewed by David Levin.
152
153         prepare-ChangeLog --bug fails on some CYGWIN installs due to missing certs
154         https://bugs.webkit.org/show_bug.cgi?id=38212
155
156         * Scripts/prepare-ChangeLog:
157          - Pass --insecure to curl to work around CYGWIN missing certs.
158
159 2010-04-27  Sam Weinig  <sam@webkit.org>
160
161         Rubber-stamped by Beth Dakin.
162
163         It is no longer necessary to set ENABLE_EXPERIMENTAL_SINGLE_VIEW_MODE
164         or WTF_USE_WEB_THREAD. Remove them.
165
166         * Scripts/build-webkit:
167
168 2010-04-27  Chris Jerdonek  <cjerdonek@webkit.org>
169
170         Reviewed by David Levin.
171
172         Changed Chris Jerdonek from committer to reviewer.
173
174         * Scripts/webkitpy/common/config/committers.py:
175
176 2010-04-27  Benjamin Poulain  <benjamin.poulain@nokia.com>
177
178         Reviewed by Kenneth Rohde Christiansen.
179
180         [Qt] QGraphicsWebView: Arrow keys scroll the graphics-view instead of the web-page
181         https://bugs.webkit.org/show_bug.cgi?id=35834
182
183         The scene should always have the size of the web view otherwhise it is
184         possible to scroll the graphics view.
185
186         * QtLauncher/webview.cpp:
187         (WebViewGraphicsBased::resizeEvent):
188
189 2010-04-27  Diego Gonzalez  <diegohcg@webkit.org>
190
191         Unreviewed.
192
193         Adding myself to committers.py
194
195         * Scripts/webkitpy/common/config/committers.py:
196
197 2010-04-27  Eric Seidel  <eric@webkit.org>
198
199         Reviewed by Adam Barth.
200
201         [chromium] new-run-webkit-tests hangs on Chromium Bots (OS X and Linux)
202         https://bugs.webkit.org/show_bug.cgi?id=37987
203
204         After further research, I believe the hang is caused by:
205         http://bugs.python.org/issue2320
206         Basically Popen() is not reentrant.
207         The workaround is to pass close_fds=True to Popen() on Mac/Linux.
208
209         I fixed our main Popen wrapper "Executive.run_command" to use close_fds=True
210         when appropriate.
211
212         I audited all places we call Popen() and either moved them to run_command
213         or left a FIXME that they are not thread safe.  A few places I added the
214         close_fds workaround there and left an explanitory note.
215
216         * Scripts/webkitpy/common/checkout/scm_unittest.py:
217          - Added note that this Popen use is not threadsafe.
218         * Scripts/webkitpy/common/system/executive.py:
219          - Fixed our Executive.run_* to workaround python bug 2320.
220         * Scripts/webkitpy/common/system/user.py:
221          _ Added note that this Popen use is not threadsafe.
222         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py: ditto.
223         * Scripts/webkitpy/layout_tests/port/apache_http_server.py: ditto.
224         * Scripts/webkitpy/layout_tests/port/base.py:
225          - Change wdiff back to using run_command now that we believe it
226            to be threadsafe.
227         * Scripts/webkitpy/layout_tests/port/chromium.py:
228          - Fix to use Executive in places.
229          - Pass self._executive down to the Driver for easier unit testing.
230         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
231          - Re-factor to use a _kill_all method.
232          - Made the _kill_all method use run_command to be threadsafe.
233         * Scripts/webkitpy/layout_tests/port/http_server.py:
234          - Add FIXME about using Executive.
235         * Scripts/webkitpy/layout_tests/port/server_process.py:
236          - Use Executive to be threadsafe.
237         * Scripts/webkitpy/layout_tests/port/webkit.py:
238          - Pass self._executive down to the Driver.
239         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
240          - Add note about Popen not being threadsafe.
241         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
242          - Move one caller to run_command add notes about moving others.
243
244 2010-04-27  Adam Barth  <abarth@webkit.org>
245
246         Reviewed by Maciej Stachowiak.
247
248         REGRESSION(r58261): webkit-patch upload does not work in an SVN checkout.
249         https://bugs.webkit.org/show_bug.cgi?id=38186
250
251         Unfortunately, we don't have a good way of testing this change because
252         our test coverage of the scm.py API is poor...
253
254         * Scripts/webkitpy/common/checkout/scm.py:
255         * Scripts/webkitpy/tool/steps/preparechangelog.py:
256
257 2010-04-26  Eric Seidel  <eric@webkit.org>
258
259         Reviewed by Adam Barth.
260
261         Remove unused code from text_diff.py
262         https://bugs.webkit.org/show_bug.cgi?id=38170
263
264         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
265          - Remove is_render_tree_dump which appears unused.
266
267 2010-04-26  Adam Barth  <abarth@webkit.org>
268
269         Reviewed by Eric Seidel.
270
271         webkit-patch pretty-diff is broken
272         https://bugs.webkit.org/show_bug.cgi?id=38172
273
274         We need to register for these options because they're used when we look
275         up the diff.
276
277         * Scripts/webkitpy/tool/steps/confirmdiff.py:
278
279 2010-04-26  Adam Barth  <abarth@webkit.org>
280
281         Reviewed by Eric Seidel.
282
283         Add ObjC and GObject to run-bindings-test
284         https://bugs.webkit.org/show_bug.cgi?id=38168
285
286         * Scripts/run-bindings-tests:
287
288 2010-04-26  Adam Barth  <abarth@webkit.org>
289
290         Reviewed by Eric Seidel.
291
292         Add testing infrastructure for JSC bindings generator
293         https://bugs.webkit.org/show_bug.cgi?id=38167
294
295         Add support for testing more than one bindings.  Also, converted the
296         script to PEP8 style.
297
298         * Scripts/run-bindings-tests:
299
300 2010-04-26  Eric Seidel  <eric@webkit.org>
301
302         Reviewed by Adam Barth.
303
304         Refactor results.html generation out into a new method and test it
305         https://bugs.webkit.org/show_bug.cgi?id=38164
306
307         Hopefully this results in no change in functionality.
308
309         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
310         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
311
312 2010-04-26  Adam Barth  <abarth@webkit.org>
313
314         Unreviewed, rolling out r58265.
315         http://trac.webkit.org/changeset/58265
316         https://bugs.webkit.org/show_bug.cgi?id=38021
317
318         This change prevents me from uploading patches.  It also breaks sheriff-bot.
319
320         * Scripts/webkitpy/tool/steps/options.py:
321
322 2010-04-26  Eric Seidel  <eric@webkit.org>
323
324         Reviewed by Adam Barth.
325
326         [chromium] new-run-webkit-tests hangs on Chromium Bots (OS X and Linux)
327         https://bugs.webkit.org/show_bug.cgi?id=37987
328
329         Rolled out:
330         http://trac.webkit.org/changeset/58062
331         http://trac.webkit.org/changeset/58060
332         http://trac.webkit.org/changeset/58059
333         http://trac.webkit.org/changeset/58055
334         http://trac.webkit.org/changeset/58054
335         and parts of:
336         http://trac.webkit.org/changeset/58050
337
338         I also wrote some new comments and a tiny amount of new
339         code to help make ChromiumDriver.run_test easier to read.
340
341         In order to unit-test my new code, I had to change ChromiumDriver
342         to not automatically start itself when created.  That ended up
343         being a lot of plumbing, but is hopefully easier to understand now.
344
345         There are no tests for the (restored) wdiff code.  wdiff does not
346         exist on all systems, so for now we will assume it worked since
347         it is just old code being reverted.
348
349          * Scripts/webkitpy/layout_tests/driver_test.py:
350           - Use create_driver instead of start_driver, and be sure to call .stop()
351          * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
352           - Use create_driver instead of start_driver
353          * Scripts/webkitpy/layout_tests/port/base.py:
354           - Added a comment to explain that diffs are binary files.
355           - Various patch reverts relating to wdiff
356           - Add Driver._command_wrapper to share code between WebKitDriver and ChromiumDriver.
357           - Made _command_wrapper use shlex.split to get rid of the FIXME.
358          * Scripts/webkitpy/layout_tests/port/base_unittest.py: Added.
359           - test the new _command_wrapper
360          * Scripts/webkitpy/layout_tests/port/chromium.py:
361          - Use _command_wrapper to get rid of a bunch of ugly code.
362          - Make __init__ stop auto-starting.
363          - Rename create_driver to start_driver.
364          - Added _write_command_and_read_line to make it possible to
365            put a FIXME next to read_line() w/o having to put it in two places.
366          - Moved test_shell command building into _test_shell_command and tested it.
367          - Fix comments to say test_shell since ChromiumDriver is test_shell only.
368          * Scripts/webkitpy/layout_tests/port/chromium_unittest.py: Added.
369           - Test the new test_shell_command method.
370          * Scripts/webkitpy/layout_tests/port/dryrun.py:
371           - Rename create_driver to start_driver.
372          * Scripts/webkitpy/layout_tests/port/test.py:
373           - Rename create_driver to start_driver.
374          * Scripts/webkitpy/layout_tests/port/webkit.py:
375           - Rename create_driver to start_driver.
376           - Treat output as binary arrays.
377          * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
378           - Treat diff files as binary.
379          * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
380           - Treat diff files as binary.
381
382 2010-04-26  Adam Barth  <abarth@webkit.org>
383
384         Reviewed by Eric Seidel.
385
386         Move bindings test directory into the scripts directory
387         https://bugs.webkit.org/show_bug.cgi?id=38161
388
389         Change script to point to the new location of these data files.
390
391         * Scripts/run-bindings-tests:
392
393 2010-04-26  Adam Barth  <abarth@webkit.org>
394
395         Unreviewed.  Mark run-bindings-tests executable so we can execute it.
396
397         * Scripts/run-bindings-tests:
398
399 2010-04-26  Tony Chang  <tony@chromium.org>
400
401         Reviewed by David Levin.
402
403         [chromium] build-webkit --chromium should build release by default
404         https://bugs.webkit.org/show_bug.cgi?id=38028
405
406         * Scripts/build-dumprendertree: Also should build release by default
407         * Scripts/build-webkit: Make sure to pass command line args through
408         * Scripts/webkitdirs.pm: Build the right configuration
409
410 2010-04-26  Ojan Vafai  <ojan@chromium.org>
411
412         Reviewed by Eric Seidel.
413
414         complete rietveld bugzilla integration
415         https://bugs.webkit.org/show_bug.cgi?id=38021
416
417         Makes --fancy-review the default. All this means is that the patch will
418         be uploaded to rietveld in addition to bugs.webkit.org. 
419
420         * Scripts/webkitpy/tool/steps/options.py:
421
422 2010-04-26  Chris Jerdonek  <cjerdonek@webkit.org>
423
424         Reviewed by Shinichiro Hamaji.
425
426         For check-webkit-style, renamed the style/processors/ directory
427         to style/checkers/.
428
429         https://bugs.webkit.org/show_bug.cgi?id=38122
430
431         * Scripts/webkitpy/style/checker.py:
432           - Updated import statements.
433
434         * Scripts/webkitpy/style/checker_unittest.py:
435           - Updated import statements.
436
437         * Scripts/webkitpy/style/checkers: Copied from WebKitTools/Scripts/webkitpy/style/processors.
438         * Scripts/webkitpy/style/processors: Removed.
439         * Scripts/webkitpy/style/processors/__init__.py: Removed.
440         * Scripts/webkitpy/style/processors/common.py: Removed.
441         * Scripts/webkitpy/style/processors/common_unittest.py: Removed.
442         * Scripts/webkitpy/style/processors/cpp.py: Removed.
443         * Scripts/webkitpy/style/processors/cpp_unittest.py: Removed.
444         * Scripts/webkitpy/style/processors/python.py: Removed.
445         * Scripts/webkitpy/style/processors/python_unittest.py: Removed.
446         * Scripts/webkitpy/style/processors/python_unittest_input.py: Removed.
447         * Scripts/webkitpy/style/processors/text.py: Removed.
448         * Scripts/webkitpy/style/processors/text_unittest.py: Removed.
449
450 2010-04-06  Ojan Vafai  <ojan@chromium.org>
451
452         Reviewed by Adam Barth.
453
454         Include git commits in the diff for webkit-patch upload/land.
455         https://bugs.webkit.org/show_bug.cgi?id=36394
456
457         Adds --squash, --no-squash and --git-commit.
458
459         --git-commit will use a specific local commit for land/upload.
460         If a commit-range is specified, then that range is treated as
461         a single squashed commit.
462
463         --squash will squash all local changes including working copy changes
464         into a single patch.
465
466         --no-squash is the legacy behavior (upload only considers the working copy,
467         land commits the working copy and then each local commit separately to SVN)
468
469         If neither is specified, then an informative error is raised if there is
470         more than one local commit or when there are local commit(s) and working
471         copy changes.
472
473         If the webkit-patch.squash git config parameter is set, then
474         that will be respected instead of raising an error.
475
476         * Scripts/check-webkit-style:
477         * Scripts/webkitpy/common/checkout/api.py:
478         * Scripts/webkitpy/common/checkout/api_unittest.py:
479         * Scripts/webkitpy/common/checkout/scm.py:
480         * Scripts/webkitpy/common/checkout/scm_unittest.py:
481         * Scripts/webkitpy/style/optparser.py:
482         --git-since is removed and --git-commit no longer implies commit_id..
483         Instead, it treats that individual commit, but also supports commit ranges
484         (e.g. commit_id..) as arguments.
485         * Scripts/webkitpy/style/optparser_unittest.py:
486         * Scripts/webkitpy/style_references.py:
487         * Scripts/webkitpy/tool/commands/download.py:
488         * Scripts/webkitpy/tool/commands/upload.py:
489         * Scripts/webkitpy/tool/main.py:
490         * Scripts/webkitpy/tool/mocktool.py:
491         * Scripts/webkitpy/tool/steps/abstractstep.py:
492         * Scripts/webkitpy/tool/steps/applypatchwithlocalcommit.py:
493         * Scripts/webkitpy/tool/steps/checkstyle.py:
494         * Scripts/webkitpy/tool/steps/commit.py:
495         * Scripts/webkitpy/tool/steps/options.py:
496         * Scripts/webkitpy/tool/steps/postdiff.py:
497         * Scripts/webkitpy/tool/steps/preparechangelog.py:
498         * Scripts/webkitpy/tool/steps/preparechangelogforrevert.py:
499         * Scripts/webkitpy/tool/steps/updatechangelogswithreviewer.py:
500         * Scripts/webkitpy/tool/steps/validatereviewer.py:
501
502 2010-04-26  Chris Jerdonek  <cjerdonek@webkit.org>
503
504         Reviewed by Shinichiro Hamaji.
505
506         Deleted the StyleChecker-related classes that are no longer
507         being used.
508
509         https://bugs.webkit.org/show_bug.cgi?id=38118
510
511         * Scripts/webkitpy/style/checker.py:
512           - Deleted the DeprecatedStyleChecker class.
513
514         * Scripts/webkitpy/style/checker_unittest.py:
515           - Deleted the StyleCheckerTest, StyleCheckerCheckFileBase,
516             StyleCheckerCheckFileTest, and StyleCheckerCheckPathsTest classes.
517
518 2010-04-26  Chris Jerdonek  <cjerdonek@webkit.org>
519
520         Reviewed by Shinichiro Hamaji.
521
522         Changed the StyleChecker class to use the new TextFileReader class.
523
524         https://bugs.webkit.org/show_bug.cgi?id=37066
525
526         This revision separates the code responsible for reading and iterating
527         over text files from the rest of check-webkit-style.
528
529         * Scripts/check-webkit-style:
530           - Changed the script to use the new StyleProcessor and
531             TextFileReader classes instead of the StyleChecker class.
532
533         * Scripts/webkitpy/style/checker.py:
534           - Added a FIXME to rename many of uses of the word "processor" to
535             "checker".  We did this to clarify the difference between
536             ProcessorBase instances passed to the TextFileReader and
537             classes that process and check lines for style.
538           - Added a FIXME to remove FileType.NONE as a possible return value
539             of ProcessorDispatcher._file_type().  This will better consolidate
540             the logic of which files should be skipped.
541           - Added a FIXME to delete the StyleChecker class.
542           - Added the StyleProcessor class which implements ProcessorBase.
543             This class is responsible for processing lines to check style
544             (but not for reading files).  For each file, this class creates
545             creates both a carriage-return checker and a format-specific
546             style checker (e.g. one of C++, Python, etc).
547           - Updated the PatchChecker class to use a TextFileReader instance
548             instead of a StyleChecker.
549
550         * Scripts/webkitpy/style/checker_unittest.py:
551           - Added the StyleProcessor_EndToEndTest class to do "end-to-end"
552             tests of the StyleProcessor class.
553           - Added the StyleProcessor_CodeCoverageTest to test the
554             StyleProcessor class with more complete code coverage.
555             Unlike the StyleProcessor_EndToEndTest class, this class makes
556             heavy use of mock parameters.
557           - Added FIXME's to delete the unit test classes that are no
558             longer needed.
559           - Updated the PatchCheckerTest class to use a MockTextFileReader
560             instead of a MockStyleChecker.
561
562         * Scripts/webkitpy/style/filereader.py:
563           - Updated the TextFileReader class to use the preferred logic
564             of checking file existence at the beginning of the process_file()
565             method instead of in the except block, per
566             https://bugs.webkit.org/show_bug.cgi?id=37122
567
568         * Scripts/webkitpy/style/filereader_unittest.py:
569           - In the TextFileReaderTest class:
570             - Moved the test_process_file__should_not_process() method.
571             - Added a test_process_file__file_stdin() method to test
572               the file path "-".
573
574 2010-04-20  Robert Hogan  <robert@webkit.org>
575
576         Reviewed by Simon Hausmann.
577
578         [Qt] Add more support for textInputController
579
580         Add support for selectedRange(), setMarkedText(), insertText(),
581         and firstRectForCharacterRange().
582
583         Unskip tests:
584
585         fast/forms/input-maxlength-ime-preedit.html
586         fast/forms/input-maxlength-ime-completed.html
587         fast/text/international/thai-cursor-position.html
588         fast/events/ime-composition-events-001.html
589         editing/selection/5825350-1.html
590         editing/selection/5825350-2.html
591         editing/selection/mixed-editability-10.html
592
593         https://bugs.webkit.org/show_bug.cgi?id=35702
594
595         * DumpRenderTree/qt/TextInputControllerQt.cpp:
596         (TextInputController::setMarkedText):
597         (TextInputController::insertText):
598         (TextInputController::selectedRange):
599         (TextInputController::firstRectForCharacterRange):
600         * DumpRenderTree/qt/TextInputControllerQt.h:
601
602 2010-04-23  Eric Seidel  <eric@webkit.org>
603
604         Reviewed by Adam Barth.
605
606         check-webkit-style complains about non-utf8 data in layout test result
607         https://bugs.webkit.org/show_bug.cgi?id=38027
608
609         The problem was we were assuming patch files/diff output as utf-8.
610         Turns out they're not.  We have to treat them as binary data because
611         a single patch may have multiple text files in it with conflicting encodings!
612
613         * Scripts/webkitpy/common/checkout/api.py:
614          - contents_at_revision returns a byte array, so decode it to unicode
615            before passing it to parse_latest_entry_from_file
616         * Scripts/webkitpy/common/checkout/api_unittest.py:
617          - Update our mock mock_contents_at_revision to match the encoding
618            semantics of the real one.
619         * Scripts/webkitpy/common/checkout/scm.py:
620          - Be careful not to decode output which may contain file contents
621            (like diff, cat or show) as the encoding for that content is unknown.
622         * Scripts/webkitpy/common/checkout/scm_unittest.py:
623          - Update our tests to use both latin1 and utf-8 encoded data.
624         * Scripts/webkitpy/common/net/bugzilla.py:
625          - _fill_attachment_form should not assume unicode data.  Callers
626            may wish to attach other types of files to bugs.
627         * Scripts/webkitpy/common/prettypatch.py:
628          - Diffs are byte arrays, deal with them as such.
629         * Scripts/webkitpy/common/prettypatch_unittest.py:
630          - Test to make sure we handle diffs with multiple conflicting encodings.
631         * Scripts/webkitpy/common/system/executive_unittest.py:
632          - Make sure that our unicode support does not break our
633            byte array input support for run_command.
634
635 2010-04-23  Sam Weinig  <sam@webkit.org>
636
637         Reviewed by David Levin.
638
639         Fix for https://bugs.webkit.org/show_bug.cgi?id=38060
640         Split up Threading.h
641
642         Add necessary forwarding headers.
643
644         * DumpRenderTree/ForwardingHeaders/wtf/Atomics.h: Added.
645         * DumpRenderTree/ForwardingHeaders/wtf/ThreadSafeShared.h: Added.
646         * DumpRenderTree/ForwardingHeaders/wtf/ThreadingPrimitives.h: Added.
647
648 2010-04-23  Xiaomei Ji  <xji@chromium.org>
649
650         No need to review.
651
652         Add xji as committer.
653
654         * Scripts/webkitpy/common/config/committers.py:
655
656 2010-04-23  Kevin Ollivier  <kevino@theolliviers.com>
657
658         [wx] Build fix, add platform stub for new LayoutTestController method.
659
660         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
661         (LayoutTestController::markerTextForListItem):
662
663 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
664
665         Reviewed by Adam Barth.
666
667         new-run-webkit-tests: add a "--retry-failures" flag and a
668         "--no-retry-failures" flag (the former is the default). Also, rename
669         "--print-unexpected-results" and "--retry-unexpected-results" to
670         "--print-last-failures" and "--retry-last-failures" because the
671         retry flag was confusing. The new flag names aren't great, but
672         hopefully they're less confusing.
673
674         https://bugs.webkit.org/show_bug.cgi?id=37838
675
676         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
677
678 2010-04-22  Sam Weinig  <sam@webkit.org>
679
680         Reviewed by Anders Carlsson.
681
682         Fix for https://bugs.webkit.org/show_bug.cgi?id=38022
683         Move isMainThread predicate function to MainThread.h
684
685         Added forwarding header for MainThread.h
686
687         * DumpRenderTree/ForwardingHeaders/wtf/MainThread.h: Added.
688
689 2010-04-23  Laszlo Gombos  <laszlo.1.gombos@nokia.com>
690
691         Reviewed by Kenneth Rohde Christiansen.
692
693         [Qt] Disable Netscape plugin support for minimal configuration
694         https://bugs.webkit.org/show_bug.cgi?id=38026
695
696         Pass the minimal configuration option to Qt build system 
697         as part of the CONFIG variable.
698
699         * Scripts/build-webkit:
700
701 2010-04-22  Fumitoshi Ukai  <ukai@chromium.org>
702
703         Unreviewed fix.
704
705         * Scripts/webkitpy/layout_tests/port/websocket_server.py: fix NameError: global name 'f' is not defined.
706
707 2010-04-22  James Robinson  <jamesr@chromium.org>
708
709         Reviewed by Adam Barth.
710
711         Revert 58077 and follow-ups.  It's broken.
712         https://bugs.webkit.org/show_bug.cgi?id=37664
713
714         * Scripts/new-run-webkit-httpd:
715         * Scripts/new-run-webkit-websocketserver:
716         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
717         * Scripts/webkitpy/layout_tests/port/http_server.py:
718         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
719
720 2010-04-22  Yaar Schnitman  <yaar@chromium.org>
721
722         Reviewed by Adam Barth.
723
724         Integrate v8 testing utility with webkit tests
725         https://bugs.webkit.org/show_bug.cgi?id=37731
726
727         * Scripts/run-bindings-tests: Added.
728
729 2010-04-22  Eric Seidel  <eric@webkit.org>
730
731         Reviewed by Dimitri Glazkov.
732
733         new-run-webkit-tests --verbose shows ever-increasing #EOF lines
734         https://bugs.webkit.org/show_bug.cgi?id=37794
735
736         * Scripts/webkitpy/layout_tests/port/webkit.py:
737          - Remove the assert() since our ServerProcess code does not always
738            seem to be reading the full stderr output (or we're not waiting for it to).
739
740 2010-04-22  Eric Seidel  <eric@webkit.org>
741
742         Reviewed by Adam Barth.
743
744         Add code to help debug new-run-webkit-test hangs on the Chromium bots
745         https://bugs.webkit.org/show_bug.cgi?id=38011
746
747         I can see no reasonable way to test this change.
748         Stubbing out sys._current_frames() and traceback.extract_stack
749         seems folly.  Dumping real data would have line number
750         (and possibly other call stack) variance between runs.
751
752         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
753          - Add _dump_thread_states and _dump_thread_states_if_necessary
754            to have our main thread dump the states of all threads every
755            60 seconds when running in verbose mode.
756          - Better document what is going on in our main loop.
757
758 2010-04-22  Sam Weinig  <sam@webkit.org>
759
760         Reviewed by Anders Carlsson.
761
762         Fix for https://bugs.webkit.org/show_bug.cgi?id=38002
763         Add rudimentary statistics gathering for WebKit2
764
765         * MiniBrowser/MiniBrowser.xcodeproj/project.pbxproj:
766         * MiniBrowser/mac/AppDelegate.h:
767         * MiniBrowser/mac/AppDelegate.m:
768         (-[BrowserAppDelegate init]):
769         (-[BrowserAppDelegate newWindow:]):
770         (-[BrowserAppDelegate getCurrentPageNamespace]):
771         (-[BrowserAppDelegate _setProcessModel:]):
772         (-[BrowserAppDelegate showStatisticsWindow:]):
773         (-[BrowserAppDelegate applicationWillTerminate:]):
774         * MiniBrowser/mac/BrowserStatisticsWindow.xib: Added.
775         * MiniBrowser/mac/BrowserStatisticsWindowController.h: Added.
776         * MiniBrowser/mac/BrowserStatisticsWindowController.m: Added.
777         (-[BrowserStatisticsWindowController initWithThreadedWKContextRef:processWKContextRef:]):
778         (-[BrowserStatisticsWindowController dealloc]):
779         (-[BrowserStatisticsWindowController windowDidLoad]):
780         (-[BrowserStatisticsWindowController refreshStatistics:]):
781         * MiniBrowser/mac/BrowserWindowController.h:
782         * MiniBrowser/mac/English.lproj/MainMenu.xib:
783         * MiniBrowser/mac/MiniBrowser_Prefix.pch:
784
785 2010-04-22  Dave Moore  <davemoore@chromium.org>
786
787         Reviewed by Dimitri Glazkov.
788
789         Add test support for icon changes.
790
791         https://bugs.webkit.org/show_bug.cgi?id=33812
792
793         * DumpRenderTree/LayoutTestController.cpp:
794         (LayoutTestController::LayoutTestController):
795         (dumpIconChangesCallback):
796         (LayoutTestController::staticFunctions):
797         * DumpRenderTree/LayoutTestController.h:
798         (LayoutTestController::dumpIconChanges):
799         (LayoutTestController::setDumpIconChanges):
800         * DumpRenderTree/win/FrameLoadDelegate.cpp:
801         (FrameLoadDelegate::didChangeIcons):
802         * DumpRenderTree/win/FrameLoadDelegate.h:
803         * WinLauncher/WinLauncher.h:
804         (WinLauncherWebHost::didChangeIcons):
805
806 2010-04-22  Eric Seidel  <eric@webkit.org>
807
808         Unreviewed.  Script fix.  Will ask dpranke to look tomorrow.
809
810         new-run-webkit-tests --verbose shows ever-increasing #EOF lines
811         https://bugs.webkit.org/show_bug.cgi?id=37794
812
813         The bots are seeing cases where .error is sometimes empty.
814         Lets make the code not crash in that case for now.
815
816         * Scripts/webkitpy/layout_tests/port/webkit.py:
817
818 2010-04-22  Fumitoshi Ukai  <ukai@chromium.org>
819
820         Unreviewed fix.
821
822         * Scripts/webkitpy/layout_tests/port/websocket_server.py: fix NameError: global name 'f' is not defined.
823
824 2010-04-22  Fumitoshi Ukai  <ukai@chromium.org>
825
826         Unreviewed fix.
827
828         * Scripts/new-run-webkit-websocketserver: PyWebSocket no longer takes keyword argument 'register_cygwin'.
829
830 2010-04-22  Eric Seidel  <eric@webkit.org>
831
832         Unreviewed.  Restore the Leopard bots to using
833         old-run-webkit-tests for now.
834
835         * Scripts/run-webkit-tests:
836
837 2010-04-22  Fumitoshi Ukai  <ukai@chromium.org>
838
839         Unreviewed fix.
840
841         * Scripts/webkitpy/layout_tests/port/http_server.py: fix NameError: global name 'env' is not defined.
842
843 2010-04-22  Fumitoshi Ukai  <ukai@chromium.org>
844
845         Unreviewed fix.
846
847         * Scripts/webkitpy/layout_tests/port/websocket_server.py: fix NameError: global name 'env' is not defined.
848
849 2010-04-22  Fumitoshi Ukai  <ukai@chromium.org>
850
851         Reviewed by Adam Barth.
852
853         Chromium: Add --chromium option to new-run-webkit-websocketserver
854         https://bugs.webkit.org/show_bug.cgi?id=37664
855
856         Move cygwin setup logic in chromium_win.py.
857
858         * Scripts/new-run-webkit-httpd: remove register_cygwin parameter to pass http_server.
859         * Scripts/new-run-webkit-websocketserver: add --chromium flag
860         * Scripts/webkitpy/layout_tests/port/chromium_win.py: setup for cygwin
861         * Scripts/webkitpy/layout_tests/port/http_server.py: remove cygwin setup logic
862         * Scripts/webkitpy/layout_tests/port/websocket_server.py: remove cygwin setup logic
863
864 2010-04-22  Eric Seidel  <eric@webkit.org>
865
866         Reviewed by Adam Barth.
867
868         new-run-webkit-tests --verbose shows ever-increasing #EOF lines
869         https://bugs.webkit.org/show_bug.cgi?id=37794
870
871         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
872          - Fix the log message to explain that this is stderr output, not test output.
873         * Scripts/webkitpy/layout_tests/port/webkit.py:
874          - Be sure to reset the server_process.error after reading (seems like the wrong
875            place to do this, but at least this fixes the bug and dpranke and I can talk
876            about better designs later).
877          - Also remove the #EOF from the stderr output before returning it.
878
879 2010-04-22  Eric Seidel  <eric@webkit.org>
880
881         Unreviewed.  Fixing new-run-webkit-tests on the bots.
882
883         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
884          - Fixing _compat_shim_option to take variable keyword args.
885          - Turns out that nargs + callback is ignored unless type is specified.
886            So I added the type so that the option was properly consumed.
887            This was why new-run-webkit-tests couldn't find any tests, it
888            was looking for them under "20" because that was the argument
889            which should have been consumed by --exit-after-n-failures.
890
891 2010-04-22  Tony Chang  <tony@chromium.org>
892
893         Not reviewed, build fix for chromium Windows.
894
895         [chromium] fix ImageDiff compile on windows
896         https://bugs.webkit.org/show_bug.cgi?id=37979
897
898         * DumpRenderTree/chromium/ImageDiff.cpp:
899
900 2010-04-22  Adam Barth  <abarth@webkit.org>
901
902         Unreviewed, rolling out r58069.
903         http://trac.webkit.org/changeset/58069
904         https://bugs.webkit.org/show_bug.cgi?id=27751
905
906         Broke compile on Windows.
907
908         * DumpRenderTree/LayoutTestController.cpp:
909         (LayoutTestController::staticFunctions):
910         * DumpRenderTree/LayoutTestController.h:
911         * DumpRenderTree/chromium/LayoutTestController.cpp:
912         (LayoutTestController::LayoutTestController):
913         (LayoutTestController::overridePreference):
914         * DumpRenderTree/chromium/LayoutTestController.h:
915         * DumpRenderTree/chromium/TestShell.cpp:
916         (TestShell::resetWebSettings):
917         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
918         * DumpRenderTree/mac/DumpRenderTree.mm:
919         (resetDefaultsToConsistentValues):
920         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
921         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
922         (WebCore::WebPage::WebPage):
923         (WebCore::WebPage::resetSettings):
924         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
925         * DumpRenderTree/qt/LayoutTestControllerQt.h:
926         * DumpRenderTree/win/DumpRenderTree.cpp:
927         (resetDefaultsToConsistentValues):
928         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
929         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
930
931 2010-04-22  Eric Seidel  <eric@webkit.org>
932
933         Unreviewed.  Just adding logging.
934
935         Adding logging to help debug why the Leopard Bot
936         can't find any tests to run.
937
938         * Scripts/webkitpy/layout_tests/layout_package/test_files.py:
939         * Scripts/webkitpy/layout_tests/port/base.py:
940         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
941          - Make it a fatal error to have no tests to run.
942
943 2010-04-22  Abhishek Arya  <inferno@chromium.org>
944
945         Reviewed by Adam Barth.
946
947         Add support for controlling clipboard access from javascript.
948         Clipboard access from javascript is enabled in test framework.
949         https://bugs.webkit.org/show_bug.cgi?id=27751
950
951         * DumpRenderTree/LayoutTestController.cpp:
952         (setJavaScriptCanAccessClipboardCallback):
953         (LayoutTestController::staticFunctions):
954         * DumpRenderTree/LayoutTestController.h:
955         * DumpRenderTree/chromium/LayoutTestController.cpp:
956         (LayoutTestController::LayoutTestController):
957         (LayoutTestController::setJavaScriptCanAccessClipboard):
958         (LayoutTestController::overridePreference):
959         * DumpRenderTree/chromium/LayoutTestController.h:
960         * DumpRenderTree/chromium/TestShell.cpp:
961         (TestShell::resetWebSettings):
962         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
963         (LayoutTestController::setJavaScriptCanAccessClipboard):
964         * DumpRenderTree/mac/DumpRenderTree.mm:
965         (resetDefaultsToConsistentValues):
966         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
967         (LayoutTestController::setJavaScriptCanAccessClipboard):
968         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
969         (WebCore::WebPage::WebPage):
970         (WebCore::WebPage::resetSettings):
971         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
972         (LayoutTestController::setJavaScriptCanAccessClipboard):
973         * DumpRenderTree/qt/LayoutTestControllerQt.h:
974         * DumpRenderTree/win/DumpRenderTree.cpp:
975         (resetDefaultsToConsistentValues):
976         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
977         (LayoutTestController::setJavaScriptCanAccessClipboard):
978         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
979         (LayoutTestController::setJavaScriptCanAccessClipboard):
980
981 2010-04-21  Adam Barth  <abarth@webkit.org>
982
983         Reviewed by Eric Seidel.
984
985         webkit-patch can't fetch attachments on security bugs
986         https://bugs.webkit.org/show_bug.cgi?id=37975
987
988         Instead of calling CURL, we just need to use our Mechanize object,
989         which understand bugs.webkit.org authentication.
990
991         * Scripts/webkitpy/common/checkout/api.py:
992         * Scripts/webkitpy/common/net/bugzilla.py:
993
994 2010-04-21  Eric Seidel  <eric@webkit.org>
995
996         Unreviewed.  Temporary commit, will roll-out before morning.
997
998         Turning on new-run-webkit-tests for the Leopard build bot
999         for testing of the harness.
1000         Users should not noctice (except for the results.html difference).
1001
1002         * Scripts/run-webkit-tests:
1003
1004 2010-04-21  Eric Seidel  <eric@webkit.org>
1005
1006         Unreviewed.  Fix typo in my previous fix attempt.
1007
1008         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1009         https://bugs.webkit.org/show_bug.cgi?id=37765
1010
1011         More on-the-bot debugging, sigh.  I wish I had a local build.
1012
1013         * Scripts/webkitpy/layout_tests/port/base.py:
1014
1015 2010-04-21  Eric Seidel  <eric@webkit.org>
1016
1017         Unreviewed.  Fix typo in my previous fix attempt.
1018
1019         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1020         https://bugs.webkit.org/show_bug.cgi?id=37765
1021
1022         * Scripts/webkitpy/layout_tests/port/base.py:
1023          - Add self. to class variable access.
1024
1025 2010-04-21  Eric Seidel  <eric@webkit.org>
1026
1027         Unreviewed.  Another attempt to fix NRWT for chromium.
1028
1029         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1030         https://bugs.webkit.org/show_bug.cgi?id=37765
1031
1032         * Scripts/webkitpy/layout_tests/port/base.py:
1033          - wdiff_text was returning a byte array instead of a
1034            unicode string.  The simple fix was to just decode
1035            the result.  However, seeing so much duplicated code
1036            with Executive made me cry, so I re-wrote the function
1037            to be more like pretty_patch_text and use run_command
1038            (which already knows how to handle unicode).
1039
1040 2010-04-21  Adam Barth  <abarth@webkit.org>
1041
1042         Reviewed by Eric Seidel.
1043
1044         Windows tests buildbots are too slow to be core builders
1045         https://bugs.webkit.org/show_bug.cgi?id=37970
1046
1047         It's 10:45.  The Windows test bots are still hours behind.  They're too
1048         slow to be core builders.  When they get fast, we can add them back.
1049
1050         * Scripts/webkitpy/common/net/buildbot.py:
1051         * Scripts/webkitpy/common/net/buildbot_unittest.py:
1052
1053 2010-04-21  Eric Seidel  <eric@webkit.org>
1054
1055         Unreviewed.  Attempt one more time to fix NRWT for chromium.
1056
1057         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1058         https://bugs.webkit.org/show_bug.cgi?id=37765
1059
1060         I'm debugging in the blind because I don't have a chromium
1061         build on this laptop.
1062
1063         * Scripts/webkitpy/layout_tests/port/chromium.py:
1064          - Apply the previous fix to a second caller.
1065
1066 2010-04-21  Eric Seidel  <eric@webkit.org>
1067
1068         Unreviewed.  Attempt one more time to fix NRWT for chromium.
1069
1070         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1071         https://bugs.webkit.org/show_bug.cgi?id=37765
1072
1073         * Scripts/webkitpy/layout_tests/port/chromium.py:
1074          - Fix handling of test_shell output so that we always
1075            decode it as utf-8.
1076            Unlike DumpRenderTree test_shell does not ever return
1077            pixel data.  It spits out the pixel dumps in a separate
1078            file, thus all output over stdout is utf-8 text.
1079
1080 2010-04-21  Eric Seidel  <eric@webkit.org>
1081
1082         Unreviewed.  Attempt one more time to fix NRWT for chromium.
1083
1084         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1085         https://bugs.webkit.org/show_bug.cgi?id=37765
1086
1087         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
1088          - Pass encoding to _save_baseline_data and write_output_files
1089         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
1090          - Make _save_baseline_data and write_output_files take an encoding.
1091         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
1092         - Pass encoding to _save_baseline_data and write_output_files
1093
1094 2010-04-21  No'am Rosenthal  <noam.rosenthal@nokia.com>
1095
1096         Reviewed by Kenneth Rohde Christiansen.
1097
1098         [Qt] QtLauncher: make FPS measurement accurate
1099         https://bugs.webkit.org/show_bug.cgi?id=37934
1100
1101         Instead of counting paints, which are not interchangeable with frames that
1102         the user sees, we now set an arbitrary timer for FPS measurements. The idea is
1103         that if the main thread is delayed for any reason, that timer would be delayed
1104         as well.
1105
1106         * QtLauncher/QtLauncher.pro:
1107         * QtLauncher/fpstimer.cpp: Added.
1108         (FpsTimer::FpsTimer):
1109         (FpsTimer::numFrames):
1110         (FpsTimer::start):
1111         (FpsTimer::stop):
1112         (FpsTimer::timerEvent):
1113         * QtLauncher/fpstimer.h: Added.
1114         * QtLauncher/webview.cpp:
1115         (WebViewGraphicsBased::setFrameRateMeasurementEnabled):
1116         (WebViewGraphicsBased::updateFrameRate):
1117         (WebViewGraphicsBased::paintEvent):
1118         * QtLauncher/webview.h:
1119
1120 2010-04-21  Eric Seidel  <eric@webkit.org>
1121
1122         Unreviewed, just adding missing ":" in python file.
1123
1124         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1125         https://bugs.webkit.org/show_bug.cgi?id=37765
1126
1127         new-run-webkit-tests --chromium was borked.
1128
1129         * Scripts/webkitpy/layout_tests/port/chromium.py:
1130
1131 2010-04-21  Eric Seidel  <eric@webkit.org>
1132
1133         Reviewed by Adam Barth.
1134
1135         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1136         https://bugs.webkit.org/show_bug.cgi?id=37765
1137
1138         Third time is the charm.  I've fixed all the
1139         new-run-webkit-tests regressions from previous attempts.
1140
1141         I fixed the queue to not ignore Tor as a reviwer in r57531,
1142         but instead it throws an exception every time his name is in a patch.
1143
1144         This fixes our Executive.run_command code to work around a Popen
1145         bug http://bugs.python.org/issue5290 whereby python versions before 2.6
1146         do not correctly handle unicode objects as input or output to
1147         Popen.communicate.
1148
1149         Following the advice of:
1150         http://farmdev.com/talks/unicode/
1151         I have changed all of webkitpy to use unicode() objects as strings
1152         instead of str objects (which in Python 3 are renamed "bytes").
1153
1154         String literals were left as "foo" instead of converting to u"foo"
1155         as u"foo" is only required if the string has a non-ascii code point.
1156         Python is smart about comparing str() and unicode() values and will
1157         log an error to the console if the comparison is ever invalid.
1158
1159         Executive.run* now correctly accept and return unicode() objects.
1160         I attempted to fix all the places that we call .write() to make sure we
1161         encode any unicode() objects into utf-8.
1162
1163         I removed several uses of StringIO.  StringIO should generally always be
1164         passed a unicode() value.
1165
1166         Likewise I replaced most calls to open() with codecs.open().
1167         codecs.open() matches Python 3 open semantics in requiring an encoding
1168         parameter.  Files opened with codecs.open() with a unicode-compatible
1169         encoding will vend unicode() objects from their read() calls, like how
1170         StringIO created with a unicode() object will do.
1171
1172         I also deployed "with" statements wider (now that the project has
1173         settled on Python 2.5) to close a bunch of file descriptor leaks.
1174
1175         * Scripts/webkitpy/common/checkout/api_unittest.py:
1176          - Read/write utf-8 files instead of ascii.
1177          - Update the tests to use test for proper unicode() handling.
1178         * Scripts/webkitpy/common/checkout/changelog.py:
1179          - Document that parse_latest_entry_from_file expects
1180            file-like objects which return unicode strings.
1181         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
1182          - Use unicode() strings instead of str() byte arrays.
1183          - Deploy "with" to close file descriptor leaks.
1184         * Scripts/webkitpy/common/checkout/commitinfo.py:
1185          - Remove unneeded import.
1186         * Scripts/webkitpy/common/checkout/scm.py:
1187          - Remove use of str().
1188         * Scripts/webkitpy/common/checkout/scm_unittest.py:
1189          - Read/write utf-8 files and use unicode() strings in testing.
1190         * Scripts/webkitpy/common/config/committers.py:
1191          - Use \u instead of \x to make slightly clearer what we're doing.
1192         * Scripts/webkitpy/common/net/bugzilla.py:
1193          - Add a new _string_contents() method and explain why
1194             we have to call unicode() on the result of soup.string
1195             and why it's safe to do so w/o needing to pass a codec name.
1196           - Remove the (unused) support for passing a file object to add_patch_to_bug().
1197         * Scripts/webkitpy/common/net/buildbot.py:
1198          - Use unicode() instead of str() when needing to coax a
1199             NavigableString object into a unicode() object.
1200         * Scripts/webkitpy/common/net/buildbot_unittest.py:
1201          - Add a test which contains a unicode builder name.
1202         * Scripts/webkitpy/common/net/statusserver.py:
1203          - Remove use of str()
1204         * Scripts/webkitpy/common/prettypatch.py:
1205          - Write out the patch file as utf-8.
1206         * Scripts/webkitpy/common/system/autoinstall.py:
1207          - Write out files with a explicit encodings.
1208          - Deploy "with" to close file descriptor leaks.
1209         * Scripts/webkitpy/common/system/deprecated_logging.py:
1210          - Write out log files as utf-8.
1211         * Scripts/webkitpy/common/system/executive.py:
1212          - Make run* properly take and return unicode() objects.
1213          - Cleaned up input handling in run_command a little by adding
1214            a _compute_input() method which can return early instead of having
1215            such a long/cluttered if-block.
1216         * Scripts/webkitpy/common/system/executive_unittest.py:
1217          - Added a unit test to make sure we don't break Tor again!
1218         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1219          - Write out the test list as utf-8.
1220         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1221          - Write out json files as utf-8.
1222         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1223          - Deploy "with" to close file descriptor leaks.
1224         * Scripts/webkitpy/layout_tests/port/chromium.py: Add Executive.py FIXME.
1225         * Scripts/webkitpy/layout_tests/port/chromium_linux.py: ditto.
1226         * Scripts/webkitpy/layout_tests/port/chromium_mac.py: ditto.
1227         * Scripts/webkitpy/layout_tests/port/gtk.py: ditto.
1228         * Scripts/webkitpy/layout_tests/port/mac.py: ditto.
1229         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1230          - Make the skipped file parsing test unicode.
1231         * Scripts/webkitpy/layout_tests/port/qt.py: Add Executive.py FIXME.
1232         * Scripts/webkitpy/layout_tests/port/server_process.py: ditto.
1233         * Scripts/webkitpy/layout_tests/port/webkit.py:
1234          - Deploy "with" to close file descriptor leaks.
1235         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
1236          - Make explicit the encodings of log files and pid files.
1237         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
1238          - Make encodings explicit and deploy "with".
1239         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: ditto.
1240         * Scripts/webkitpy/layout_tests/test_types/image_diff.py: ditto.
1241         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py: ditto.
1242         * Scripts/webkitpy/layout_tests/test_types/text_diff.py: ditto.
1243         * Scripts/webkitpy/style/filereader_unittest.py: ditto.
1244         * Scripts/webkitpy/thirdparty/__init__.py: ditto.
1245         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
1246          - Removed extra import.
1247         * Scripts/webkitpy/tool/commands/queues.py:
1248          - No need to map args to strings now that run_command does.
1249          - Update test results to match args changes.
1250          - Document our global argument hacks.
1251         * Scripts/webkitpy/tool/commands/upload.py:
1252          - Pass the diff directly to add_patch_to_bug instead of creating a StringIO file wrapper.
1253         * Scripts/webkitpy/tool/mocktool.py:
1254          - Rename add_patch_to_bug argument to match bugzilla.py
1255         * Scripts/webkitpy/tool/steps/abstractstep.py:
1256          - Executive.run_* now require lists instead of strings.
1257            The lack of this change was what broke webkit-patch
1258            for svn users the first time this was landed.
1259         * Scripts/webkitpy/tool/steps/postdiff.py:
1260          - Pass the diff directly to add_patch_to_bug instead of creating a StringIO file wrapper.
1261         * Scripts/webkitpy/tool/steps/postdiffforcommit.py: ditto
1262         * Scripts/webkitpy/tool/steps/postdiffforrevert.py: ditto
1263         * Scripts/webkitpy/tool/steps/steps_unittest.py:
1264          - Fixed spurious logging seen when running test-webkitpy
1265
1266 2010-04-21  Chris Fleizach  <cfleizach@apple.com>
1267
1268         Reviewed by Alexey Proskuryakov.
1269
1270         aria-liveregion-notifications.html fails on leopard release bot
1271         https://bugs.webkit.org/show_bug.cgi?id=37112
1272
1273         Change the way that notifications are listened for by forcing clients
1274         to call a remove listener as well to match the add listener. DRT will
1275         assert if those are not done in the correct order. 
1276
1277         * DumpRenderTree/AccessibilityUIElement.cpp:
1278         (removeNotificationListenerCallback):
1279         (AccessibilityUIElement::getJSClass):
1280         * DumpRenderTree/AccessibilityUIElement.h:
1281         * DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
1282         (AccessibilityUIElement::removeNotificationListener):
1283         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
1284         (-[AccessibilityNotificationHandler initWithPlatformElement:]):
1285         (-[AccessibilityNotificationHandler dealloc]):
1286         (-[AccessibilityNotificationHandler _notificationReceived:]):
1287         (-[AccessibilityNotificationHandler setCallback:]):
1288         (AccessibilityUIElement::AccessibilityUIElement):
1289         (AccessibilityUIElement::~AccessibilityUIElement):
1290         (AccessibilityUIElement::addNotificationListener):
1291         (AccessibilityUIElement::removeNotificationListener):
1292         * DumpRenderTree/win/AccessibilityUIElementWin.cpp:
1293         (AccessibilityUIElement::removeNotificationListener):
1294
1295 2010-04-21  Anders Carlsson  <andersca@apple.com>
1296
1297         Reviewed by Sam Weinig.
1298
1299         Add debug menu items to show/hide the Web View.
1300         https://bugs.webkit.org/show_bug.cgi?id=37958
1301
1302         * MiniBrowser/mac/BrowserWindowController.h:
1303         * MiniBrowser/mac/BrowserWindowController.m:
1304         (-[BrowserWindowController showHideWebView:]):
1305         (-[BrowserWindowController removeReinsertWebView:]):
1306         (-[BrowserWindowController validateMenuItem:]):
1307         * MiniBrowser/mac/English.lproj/MainMenu.xib:
1308
1309 2010-04-21  Eric Seidel  <eric@webkit.org>
1310
1311         Unreviewed.  Rolling out unicode() changes as they broke NRWT for chromium.
1312         Rolling out:
1313         http://trac.webkit.org/changeset/58014
1314         http://trac.webkit.org/changeset/58016
1315         http://trac.webkit.org/changeset/58020
1316
1317         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbo
1318         https://bugs.webkit.org/show_bug.cgi?id=37765
1319
1320         * Scripts/webkitpy/common/checkout/api_unittest.py:
1321         * Scripts/webkitpy/common/checkout/changelog.py:
1322         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
1323         * Scripts/webkitpy/common/checkout/commitinfo.py:
1324         * Scripts/webkitpy/common/checkout/scm.py:
1325         * Scripts/webkitpy/common/checkout/scm_unittest.py:
1326         * Scripts/webkitpy/common/config/committers.py:
1327         * Scripts/webkitpy/common/net/bugzilla.py:
1328         * Scripts/webkitpy/common/net/buildbot.py:
1329         * Scripts/webkitpy/common/net/buildbot_unittest.py:
1330         * Scripts/webkitpy/common/net/statusserver.py:
1331         * Scripts/webkitpy/common/prettypatch.py:
1332         * Scripts/webkitpy/common/system/autoinstall.py:
1333         * Scripts/webkitpy/common/system/deprecated_logging.py:
1334         * Scripts/webkitpy/common/system/executive.py:
1335         * Scripts/webkitpy/common/system/executive_unittest.py:
1336         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1337         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1338         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1339         * Scripts/webkitpy/layout_tests/port/chromium.py:
1340         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1341         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1342         * Scripts/webkitpy/layout_tests/port/gtk.py:
1343         * Scripts/webkitpy/layout_tests/port/mac.py:
1344         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1345         * Scripts/webkitpy/layout_tests/port/qt.py:
1346         * Scripts/webkitpy/layout_tests/port/server_process.py:
1347         * Scripts/webkitpy/layout_tests/port/webkit.py:
1348         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
1349         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
1350         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1351         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
1352         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
1353         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
1354         * Scripts/webkitpy/style/filereader_unittest.py:
1355         * Scripts/webkitpy/thirdparty/__init__.py:
1356         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
1357         * Scripts/webkitpy/tool/commands/queues.py:
1358         * Scripts/webkitpy/tool/commands/queues_unittest.py:
1359         * Scripts/webkitpy/tool/commands/upload.py:
1360         * Scripts/webkitpy/tool/mocktool.py:
1361         * Scripts/webkitpy/tool/steps/abstractstep.py:
1362         * Scripts/webkitpy/tool/steps/postdiff.py:
1363         * Scripts/webkitpy/tool/steps/postdiffforcommit.py:
1364         * Scripts/webkitpy/tool/steps/postdiffforrevert.py:
1365         * Scripts/webkitpy/tool/steps/steps_unittest.py:
1366
1367 2010-04-21  Eric Seidel  <eric@webkit.org>
1368
1369         Unreviewed, fixing NRWT for real this time.
1370
1371         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1372         https://bugs.webkit.org/show_bug.cgi?id=37765
1373
1374         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
1375          - Add a hack to fix new-run-webkit-tests
1376            my understanding of codecs.open(encoding=None)
1377            must have been wrong.
1378
1379 2010-04-21  Eric Seidel  <eric@webkit.org>
1380
1381         Unreviewed, just fixing exception seen on builders.
1382
1383         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1384         https://bugs.webkit.org/show_bug.cgi?id=37765
1385
1386         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
1387          - Pass and encoding to _write_into_file_at_path
1388
1389 2010-04-21  Eric Seidel  <eric@webkit.org>
1390
1391         Reviewed by Adam Barth.
1392
1393         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1394         https://bugs.webkit.org/show_bug.cgi?id=37765
1395
1396         I fixed the queue to not ignore Tor as a reviwer in r57531,
1397         but instead it throws an exception every time his name is in a patch.
1398
1399         This fixes our Executive.run_command code to work around a Popen
1400         bug http://bugs.python.org/issue5290 whereby python versions before 2.6
1401         do not correctly handle unicode objects as input or output to
1402         Popen.communicate.
1403
1404         Following the advice of:
1405         http://farmdev.com/talks/unicode/
1406         I have changed all of webkitpy to use unicode() objects as strings
1407         instead of str objects (which in Python 3 are renamed "bytes").
1408
1409         String literals were left as "foo" instead of converting to u"foo"
1410         as u"foo" is only required if the string has a non-ascii code point.
1411         Python is smart about comparing str() and unicode() values and will
1412         log an error to the console if the comparison is ever invalid.
1413
1414         Executive.run* now correctly accept and return unicode() objects.
1415         I attempted to fix all the places that we call .write() to make sure we
1416         encode any unicode() objects into utf-8.
1417
1418         I removed several uses of StringIO.  StringIO should generally always be
1419         passed a unicode() value.
1420
1421         Likewise I replaced most calls to open() with codecs.open().
1422         codecs.open() matches Python 3 open semantics in requiring an encoding
1423         parameter.  Files opened with codecs.open() with a unicode-compatible
1424         encoding will vend unicode() objects from their read() calls, like how
1425         StringIO created with a unicode() object will do.
1426
1427         I also deployed "with" statements wider (now that the project has
1428         settled on Python 2.5) to close a bunch of file descriptor leaks.
1429
1430         * Scripts/webkitpy/common/checkout/api_unittest.py:
1431          - Read/write utf-8 files instead of ascii.
1432          - Update the tests to use test for proper unicode() handling.
1433         * Scripts/webkitpy/common/checkout/changelog.py:
1434          - Document that parse_latest_entry_from_file expects
1435            file-like objects which return unicode strings.
1436         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
1437          - Use unicode() strings instead of str() byte arrays.
1438          - Deploy "with" to close file descriptor leaks.
1439         * Scripts/webkitpy/common/checkout/commitinfo.py:
1440          - Remove unneeded import.
1441         * Scripts/webkitpy/common/checkout/scm.py:
1442          - Remove use of str().
1443         * Scripts/webkitpy/common/checkout/scm_unittest.py:
1444          - Read/write utf-8 files and use unicode() strings in testing.
1445         * Scripts/webkitpy/common/config/committers.py:
1446          - Use \u instead of \x to make slightly clearer what we're doing.
1447         * Scripts/webkitpy/common/net/bugzilla.py:
1448          - Add a new _string_contents() method and explain why
1449             we have to call unicode() on the result of soup.string
1450             and why it's safe to do so w/o needing to pass a codec name.
1451           - Remove the (unused) support for passing a file object to add_patch_to_bug().
1452         * Scripts/webkitpy/common/net/buildbot.py:
1453          - Use unicode() instead of str() when needing to coax a
1454             NavigableString object into a unicode() object.
1455         * Scripts/webkitpy/common/net/buildbot_unittest.py:
1456          - Add a test which contains a unicode builder name.
1457         * Scripts/webkitpy/common/net/statusserver.py:
1458          - Remove use of str()
1459         * Scripts/webkitpy/common/prettypatch.py:
1460          - Write out the patch file as utf-8.
1461         * Scripts/webkitpy/common/system/autoinstall.py:
1462          - Write out files with a explicit encodings.
1463          - Deploy "with" to close file descriptor leaks.
1464         * Scripts/webkitpy/common/system/deprecated_logging.py:
1465          - Write out log files as utf-8.
1466         * Scripts/webkitpy/common/system/executive.py:
1467          - Make run* properly take and return unicode() objects.
1468          - Cleaned up input handling in run_command a little by adding
1469            a _compute_input() method which can return early instead of having
1470            such a long/cluttered if-block.
1471         * Scripts/webkitpy/common/system/executive_unittest.py:
1472          - Added a unit test to make sure we don't break Tor again!
1473         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1474          - Write out the test list as utf-8.
1475         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1476          - Write out json files as utf-8.
1477         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1478          - Deploy "with" to close file descriptor leaks.
1479         * Scripts/webkitpy/layout_tests/port/chromium.py: Add Executive.py FIXME.
1480         * Scripts/webkitpy/layout_tests/port/chromium_linux.py: ditto.
1481         * Scripts/webkitpy/layout_tests/port/chromium_mac.py: ditto.
1482         * Scripts/webkitpy/layout_tests/port/gtk.py: ditto.
1483         * Scripts/webkitpy/layout_tests/port/mac.py: ditto.
1484         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1485          - Make the skipped file parsing test unicode.
1486         * Scripts/webkitpy/layout_tests/port/qt.py: Add Executive.py FIXME.
1487         * Scripts/webkitpy/layout_tests/port/server_process.py: ditto.
1488         * Scripts/webkitpy/layout_tests/port/webkit.py:
1489          - Deploy "with" to close file descriptor leaks.
1490         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
1491          - Make explicit the encodings of log files and pid files.
1492         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
1493          - Make encodings explicit and deploy "with".
1494         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: ditto.
1495         * Scripts/webkitpy/layout_tests/test_types/image_diff.py: ditto.
1496         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py: ditto.
1497         * Scripts/webkitpy/layout_tests/test_types/text_diff.py: ditto.
1498         * Scripts/webkitpy/style/filereader_unittest.py: ditto.
1499         * Scripts/webkitpy/thirdparty/__init__.py: ditto.
1500         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
1501          - Removed extra import.
1502         * Scripts/webkitpy/tool/commands/queues.py:
1503          - No need to map args to strings now that run_command does.
1504          - Update test results to match args changes.
1505          - Document our global argument hacks.
1506         * Scripts/webkitpy/tool/commands/upload.py:
1507          - Pass the diff directly to add_patch_to_bug instead of creating a StringIO file wrapper.
1508         * Scripts/webkitpy/tool/mocktool.py:
1509          - Rename add_patch_to_bug argument to match bugzilla.py
1510         * Scripts/webkitpy/tool/steps/abstractstep.py:
1511          - Executive.run_* now require lists instead of strings.
1512            The lack of this change was what broke webkit-patch
1513            for svn users the first time this was landed.
1514         * Scripts/webkitpy/tool/steps/postdiff.py:
1515          - Pass the diff directly to add_patch_to_bug instead of creating a StringIO file wrapper.
1516         * Scripts/webkitpy/tool/steps/postdiffforcommit.py: ditto
1517         * Scripts/webkitpy/tool/steps/postdiffforrevert.py: ditto
1518         * Scripts/webkitpy/tool/steps/steps_unittest.py:
1519          - Fixed spurious logging seen when running test-webkitpy
1520
1521 2010-04-21  Kinuko Yasuda  <kinuko@chromium.org>
1522
1523         Unreviewed.
1524
1525         Add myself in committers.py.
1526
1527         * Scripts/webkitpy/common/config/committers.py:
1528
1529 2010-04-21  Dirk Pranke  <dpranke@chromium.org>
1530
1531         Reviewed by Dimitri Glazkov.
1532
1533         new-run-webkit-tests: fix a bug in the Chromium port where we would
1534         try to talk to a crashed test_shell and raise exceptions that weren't
1535         being caught.
1536
1537         https://bugs.webkit.org/show_bug.cgi?id=37941
1538
1539         * Scripts/webkitpy/layout_tests/port/chromium.py:
1540
1541 2010-04-21  Laszlo Gombos  <laszlo.1.gombos@nokia.com>
1542
1543         Unreviewed.
1544
1545         [Qt] [Symbian] Build fix.
1546
1547         Work around a Qt quirk. Some versions of Symbian port Qt
1548         QFontDatabase::removeAllApplicationFonts symbol is not available.
1549
1550         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1551         (WebCore::DumpRenderTree::open):
1552
1553 2010-04-21  Alexey Proskuryakov  <ap@apple.com>
1554
1555         Unreviewed.
1556
1557         https://bugs.webkit.org/show_bug.cgi?id=37933
1558         <rdar://problem/7719540> XMLHttpRequest.withCredentials should be better enforced.
1559
1560         Adding stub implementation of authenticateSession(). Depending on platform loader behavior,
1561         a real implementation may or may not be necessary for the one test that currently uses it
1562         to pass.
1563
1564         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1565         (LayoutTestController::authenticateSession):
1566         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1567
1568 2010-04-21  Alexey Proskuryakov  <ap@apple.com>
1569
1570         Reviewed by Brady Eidson.
1571
1572         https://bugs.webkit.org/show_bug.cgi?id=37933
1573         <rdar://problem/7719540> XMLHttpRequest.withCredentials should be better enforced.
1574
1575         Adding authenticateSession() method that adds credentials to per-process credential storage
1576         (for platforms that even have such). No Windows implementation, because writing another
1577         loader for DRT is painful.
1578
1579         * DumpRenderTree/LayoutTestController.cpp:
1580         (authenticateSessionCallback):
1581         (LayoutTestController::staticFunctions):
1582         * DumpRenderTree/LayoutTestController.h:
1583         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1584         (-[SynchronousLoader dealloc]):
1585         (-[SynchronousLoader connectionShouldUseCredentialStorage:]):
1586         (-[SynchronousLoader connection:didReceiveAuthenticationChallenge:]):
1587         (-[SynchronousLoader connection:didFailWithError:]):
1588         (-[SynchronousLoader connectionDidFinishLoading:]):
1589         (+[SynchronousLoader makeRequest:withUsername:password:]):
1590         (LayoutTestController::authenticateSession):
1591
1592         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1593         (LayoutTestController::authenticateSession):
1594         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
1595         (LayoutTestController::authenticateSession):
1596         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1597         (LayoutTestController::authenticateSession):
1598         Stub implementations.
1599
1600 2010-04-21  Kent Tamura  <tkent@chromium.org>
1601
1602         Reviewed by Dimitri Glazkov.
1603
1604         [DRT/Chromium] Make DRT compilable in Chromium tree
1605         https://bugs.webkit.org/show_bug.cgi?id=37923
1606
1607         We need to use different GYPs in a case of WebKit-only checkout
1608         and a case of whole Chromium checkout because the relative paths
1609         from webkit/ to WebKit/chromium/features.gypi are different in
1610         these cases and we can't use 'conditions' for 'includes' in GYPs.
1611
1612         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
1613
1614 2010-04-21  Jakub Wieczorek  <jwieczorek@webkit.org>
1615
1616         Reviewed by Darin Adler.
1617
1618         List item markers are not always updated after changes in the DOM.
1619         https://bugs.webkit.org/show_bug.cgi?id=37060
1620
1621         * DumpRenderTree/LayoutTestController.cpp:
1622         (markerTextForListItemCallback): A function that returns the marker text for a given list item.
1623         (LayoutTestController::staticFunctions):
1624         * DumpRenderTree/LayoutTestController.h:
1625         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1626         (LayoutTestController::markerTextForListItem): Implement it in the GTK port.
1627         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1628         (LayoutTestController::markerTextForListItem): Add a stub.
1629         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1630         (LayoutTestController::LayoutTestController):
1631         (LayoutTestController::markerTextForListItem): Implement it in the Qt port.
1632         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1633         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1634         (LayoutTestController::markerTextForListItem): Add a stub.
1635
1636 2010-04-21  Adam Roben  <aroben@apple.com>
1637
1638         Exclude leaks in CGGradientCreateWithColorComponents from
1639         run-webkit-tests leaks output
1640
1641         The leak in CG is covered by <rdar://problem/7888492>.
1642
1643         Fixes <http://webkit.org/b/37927>.
1644
1645         Reviewed by Eric Carlson.
1646
1647         * Scripts/old-run-webkit-tests:
1648         (sub countAndPrintLeaks): Exclude leaks in
1649         CGGradientCreateWithColorComponents on certain OSs.
1650
1651 2010-04-21  Kent Tamura  <tkent@chromium.org>
1652
1653         Reviewed by Shinichiro Hamaji.
1654
1655         [DRT/Chromium] Import MockSpellCheck from Chromium
1656         https://bugs.webkit.org/show_bug.cgi?id=37910
1657
1658         Import webkit/tools/test_shell/mock_spellcheck.{cc,h} rev.37241 of Chromium.
1659
1660         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
1661           Add ICU explicitly because WTFString.h includes ICU headers.
1662           Add MockSpellCheck.{cpp,h}.
1663         * DumpRenderTree/chromium/MockSpellCheck.cpp: Added.
1664         * DumpRenderTree/chromium/MockSpellCheck.h: Added.
1665         * DumpRenderTree/chromium/WebViewHost.cpp:
1666         (WebViewHost::spellCheck):
1667         * DumpRenderTree/chromium/WebViewHost.h:
1668         * DumpRenderTree/chromium/config.h: Define JS_EXPORTDATA, which is used
1669           by wtf/text/AtomicString.h included by wtf/text/WTFString.h.
1670
1671 2010-04-21  Adam Barth  <abarth@webkit.org>
1672
1673         Reviewed by Eric Seidel.
1674
1675         Add webkit-patch pretty-diff
1676         https://bugs.webkit.org/show_bug.cgi?id=37892
1677
1678         This is slightly lame because it asks you whether the diff is correct,
1679         but it's a starting point.
1680
1681         * Scripts/webkitpy/tool/commands/__init__.py:
1682         * Scripts/webkitpy/tool/commands/prettydiff.py: Added.
1683         * Scripts/webkitpy/tool/main.py:
1684
1685 2010-04-21  Adam Barth  <abarth@webkit.org>
1686
1687         Reviewed by Darin Adler.
1688
1689         Remove mention of non-existant --no-build option
1690         https://bugs.webkit.org/show_bug.cgi?id=37893
1691
1692         The option doesn't exist!
1693
1694         * Scripts/webkitpy/tool/commands/stepsequence.py:
1695
1696 2010-04-21  Balazs Kelemen  <kb@inf.u-szeged.hu>
1697
1698         Reviewed by Kenneth Rohde Christiansen.
1699
1700         [Qt] Interrupting JavaScript is cumbersome when you use QtLaucher for testing or profiling.
1701         https://bugs.webkit.org/show_bug.cgi?id=37198
1702
1703         * QtLauncher/main.cpp:
1704         (LauncherWindow::toggleInterruptingJavaScriptEnabled):
1705         (LauncherWindow::newWindow):
1706         (LauncherWindow::cloneWindow):
1707         (LauncherWindow::createChrome):
1708         (main):
1709         * QtLauncher/webpage.cpp:
1710         (WebPage::WebPage):
1711         (WebPage::shouldInterruptJavaScript):
1712         * QtLauncher/webpage.h:
1713         (WebPage::setInterruptingJavaScriptEnabled):
1714
1715 2010-04-21  Eric Seidel  <eric@webkit.org>
1716
1717         Unreviewed, rolling out r57963.
1718         http://trac.webkit.org/changeset/57963
1719         https://bugs.webkit.org/show_bug.cgi?id=37759
1720
1721         Three tests started crashing on the Qt bot.
1722
1723         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1724         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1725
1726 2010-04-21  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
1727
1728         Reviewed by Eric Seidel.
1729
1730         Make new-run-webkit-tests work for the Qt port
1731         https://bugs.webkit.org/show_bug.cgi?id=37588
1732
1733         * Scripts/webkitpy/layout_tests/port/qt.py:
1734
1735 2010-04-21  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
1736
1737         Reviewed by Eric Seidel.
1738
1739         new-run-webkit-tests: try to detect alternate apache path
1740         https://bugs.webkit.org/show_bug.cgi?id=37587
1741
1742         _check_port_build() also needs to return true in the
1743         base implementation to not fail the check_build step.
1744
1745         * Scripts/webkitpy/layout_tests/port/webkit.py:
1746
1747 2010-04-21  Yi Shen  <yi.4.shen@nokia.com>
1748
1749         Reviewed by Simon Hausmann.
1750
1751         [Qt] Add LayoutTestController interface: computedStyleIncludingVisitedInfo
1752         https://bugs.webkit.org/show_bug.cgi?id=37759
1753
1754         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1755         (LayoutTestController::computedStyleIncludingVisitedInfo):
1756         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1757
1758 2010-04-21  Eric Seidel  <eric@webkit.org>
1759
1760         Unreviewed, test fix only.
1761
1762         new-run-webkit-tests: implement a --log trace message to be able to display detailed output of an individual test run
1763         https://bugs.webkit.org/show_bug.cgi?id=37726
1764
1765         This change seems to have broken a test.
1766         Attempting to handle the case where we don't have any
1767         timing information.  Dirk may have to correct this change.
1768
1769         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1770
1771 2010-04-19  Eric Seidel  <eric@webkit.org>
1772
1773         Reviewed by Adam Barth.
1774
1775         new-run-webkit-tests has much higher startup latency than run-webkit-tests
1776         https://bugs.webkit.org/show_bug.cgi?id=37643
1777
1778         I got rid of the -expected.checksum reads during startup.
1779         This makes startup noticably better on my laptop.
1780
1781         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1782          - Use image_hash() instead of .image_hash now that expected.checksum
1783            file reads are done lazily.
1784         * Scripts/webkitpy/layout_tests/port/http_server_base.py:
1785          - Add debug logging for this sleep call.
1786            In my testing I never saw this sleep() hit.
1787         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
1788          - Sleep a shorter interval to make websocket server
1789            startup more responsive.  On my machine startup was
1790            taking around 1 second.
1791          - Remove the unconditional .5s delay on startup.
1792         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1793          - Make image_hash file reads done lazily in a new image_hash() function.
1794          - Add a "Starting testing ..." meter update after DRT threads have
1795            been started, but before we get updates from the first one.
1796          - Rename variable "t" to a full english name to match WebKit style.
1797
1798 2010-04-20  Daniel Bates  <dbates@rim.com>
1799
1800         Reviewed by Eric Seidel.
1801
1802         https://bugs.webkit.org/show_bug.cgi?id=37748
1803
1804         Make Sheriffbot more inspirational.
1805
1806         * Scripts/webkitpy/common/net/bugzilla.py:
1807         * Scripts/webkitpy/common/net/bugzilla_unittest.py:
1808         * Scripts/webkitpy/tool/bot/irc_command.py:
1809         * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:
1810         * Scripts/webkitpy/tool/mocktool.py:
1811
1812 2010-04-20  Kent Tamura  <tkent@chromium.org>
1813
1814         Reviewed by Dimitri Glazkov.
1815
1816         [DRT/Chromium] Import Chromium image_diff as ImageDiff
1817         https://bugs.webkit.org/show_bug.cgi?id=37790
1818
1819         ImageDiff.cpp is based on tools/imagediff/image_diff.cc r41911 of Chromium.
1820
1821         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
1822         * DumpRenderTree/chromium/ImageDiff.cpp: Added.
1823         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1824
1825 2010-04-20  Dirk Pranke <dpranke@chromium.org>
1826
1827         This patch to new-run-webkit-tests adds a --log 'trace' option
1828         that prints out detailed info about a given test as it executes
1829         (where the baselines are, what the expectation is, what we got,
1830         how long it took).
1831
1832         https://bugs.webkit.org/show_bug.cgi?id=37726
1833
1834         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1835           - use the newly exposed TestResult class and implement
1836             --log trace
1837         * Scripts/webkitpy/layout_tests/layout_package/dump_render_thread.py:
1838           - rename TestStats to TestResult and make it more public, resulting
1839             in cleaner code
1840         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
1841           - add expectation_to_string() as a separate callable function
1842
1843 2010-04-20  Eric Seidel  <eric@webkit.org>
1844
1845         Unreviewed, rolling out r57907.
1846         http://trac.webkit.org/changeset/57907
1847         https://bugs.webkit.org/show_bug.cgi?id=37765
1848
1849         Appears to have broken MacEWS and possibly webkit-patch upload
1850         for svn users.  Needs further investigation.
1851
1852         * Scripts/webkitpy/common/checkout/api.py:
1853         * Scripts/webkitpy/common/checkout/api_unittest.py:
1854         * Scripts/webkitpy/common/checkout/changelog.py:
1855         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
1856         * Scripts/webkitpy/common/checkout/commitinfo.py:
1857         * Scripts/webkitpy/common/checkout/scm.py:
1858         * Scripts/webkitpy/common/checkout/scm_unittest.py:
1859         * Scripts/webkitpy/common/config/committers.py:
1860         * Scripts/webkitpy/common/net/bugzilla.py:
1861         * Scripts/webkitpy/common/net/buildbot.py:
1862         * Scripts/webkitpy/common/net/buildbot_unittest.py:
1863         * Scripts/webkitpy/common/net/statusserver.py:
1864         * Scripts/webkitpy/common/prettypatch.py:
1865         * Scripts/webkitpy/common/system/autoinstall.py:
1866         * Scripts/webkitpy/common/system/deprecated_logging.py:
1867         * Scripts/webkitpy/common/system/executive.py:
1868         * Scripts/webkitpy/common/system/executive_unittest.py:
1869         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1870         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1871         * Scripts/webkitpy/layout_tests/layout_package/metered_stream.py:
1872         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1873         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
1874         * Scripts/webkitpy/tool/commands/queues.py:
1875         * Scripts/webkitpy/tool/commands/queues_unittest.py:
1876         * Scripts/webkitpy/tool/commands/upload.py:
1877         * Scripts/webkitpy/tool/mocktool.py:
1878         * Scripts/webkitpy/tool/steps/abstractstep.py:
1879         * Scripts/webkitpy/tool/steps/postdiff.py:
1880         * Scripts/webkitpy/tool/steps/postdiffforcommit.py:
1881         * Scripts/webkitpy/tool/steps/postdiffforrevert.py:
1882         * Scripts/webkitpy/tool/steps/steps_unittest.py:
1883
1884 2010-04-20  Nate Chapin  <japhet@chromium.org>
1885
1886         Unreviewed.
1887
1888         Update my irc handle in committers.py (natechapin -> japhet).
1889
1890         * Scripts/webkitpy/common/config/committers.py:
1891
1892 2010-04-20  Eric Seidel  <eric@webkit.org>
1893
1894         Reviewed by Adam Barth.
1895
1896         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1897         https://bugs.webkit.org/show_bug.cgi?id=37765
1898
1899         I fixed the queue to not ignore Tor as a reviwer in r57531,
1900         but instead it throws an exception every time his name is in a patch.
1901
1902         This fixes our Executive.run_command code to work around a Popen
1903         bug http://bugs.python.org/issue5290 whereby python versions before 2.6
1904         do not correctly handle unicode objects as input or output to
1905         Popen.communicate.
1906
1907         Following the advice of:
1908         http://farmdev.com/talks/unicode/
1909         I'm attempting to take the python unicode plunge and use unicode()
1910         objects as strings instead of str() objects everywhere in webkitpy.
1911
1912         We do not have to use u"" instead of "" because u"a" == "a" as expected
1913         in Python.  Python will generate a warning to the console in cases where
1914         a unicode() == str() operation cannot be performed.
1915
1916         I also cleaned up the input handling in run_command a little by adding
1917         a new _compute_input() method which can return early instead of having
1918         such a long/cluttered if-block.
1919
1920         Executive.run* now correctly accept and return unicode() objects.
1921         I attempted to fix all the places that we call .write() to make sure we
1922         encode any unicode() objects into utf-8.
1923
1924         All places which use StringIO need to be sure to pass StringIO a
1925         pre-encoded byte-array (str object) instead of unicode so that
1926         clients which read from the StringIO don't have encoding exceptions.
1927         To make this easier, I removed the patch_file_object support from
1928         add_patch_to_bug, and changed the 4 places which previously used
1929         StringIO to create a fake patch file.
1930
1931         I attempted to document any places where we are not correctly converting
1932         to/from bytes (str() objects) to strings (unicode() objects).
1933
1934         * Scripts/webkitpy/common/checkout/api_unittest.py:
1935          - Read/write utf-8 files instead of ascii.
1936          - Update the tests to use test for proper unicode() handling.
1937         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
1938          - Use unicode() strings instead of str() byte arrays.
1939         * Scripts/webkitpy/common/checkout/scm.py:
1940          - Remove use of str().
1941         * Scripts/webkitpy/common/checkout/scm_unittest.py:
1942          - Read/write utf-8 files and use unicode() strings in testing.
1943         * Scripts/webkitpy/common/config/committers.py:
1944          - Use \u instead of \x to make slightly clearer what we're doing.
1945         * Scripts/webkitpy/common/net/bugzilla.py:
1946          - Add a new _string_contents() method and explain why
1947            we have to call unicode() on the result of soup.string
1948            and why it's safe to do so w/o needing to pass a codec name.
1949          - Remove the (unused) support for passing a file object to add_patch_to_bug().
1950         * Scripts/webkitpy/common/net/buildbot.py:
1951          - Use unicode() instead of str() when needing to coax a
1952            NavigableString object into a unicode() object.
1953         * Scripts/webkitpy/common/net/statusserver.py:
1954          - Remove use of str()
1955         * Scripts/webkitpy/common/prettypatch.py:
1956          - Write out the patch file as utf-8.
1957         * Scripts/webkitpy/common/system/autoinstall.py:
1958          - Add a FIXME about encoding.
1959         * Scripts/webkitpy/common/system/deprecated_logging.py:
1960          - Document that tee() works on bytes, not strings.
1961         * Scripts/webkitpy/common/system/executive.py:
1962          - Make run* properly take and return unicode() objects.
1963         * Scripts/webkitpy/common/system/executive_unittest.py:
1964          - Added a unit test to make sure we don't break Tor again!
1965         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1966          - Write out the test list as utf-8.
1967         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1968          - Write out json files as utf-8.
1969         * Scripts/webkitpy/layout_tests/layout_package/metered_stream.py:
1970          - Add FIXME about encoding handling.
1971         * Scripts/webkitpy/tool/commands/upload.py:
1972          - Pass the diff directly to add_patch_to_bug instead of creating a StringIO file wrapper.
1973         * Scripts/webkitpy/tool/mocktool.py:
1974          - Rename add_patch_to_bug argument to match bugzilla.py
1975         * Scripts/webkitpy/tool/steps/postdiff.py:
1976          - Pass the diff directly to add_patch_to_bug instead of creating a StringIO file wrapper.
1977         * Scripts/webkitpy/tool/steps/postdiffforcommit.py: ditto.
1978         * Scripts/webkitpy/tool/steps/postdiffforrevert.py: ditto.
1979         * Scripts/webkitpy/tool/steps/steps_unittest.py:
1980          - Fixed spurious logging seen when running test-webkitpy
1981
1982 2010-04-20  Chris Jerdonek  <cjerdonek@webkit.org>
1983
1984         Reviewed by Shinichiro Hamaji.
1985
1986         For check-webkit-style, implemented __eq__() and __ne__() (the
1987         built-in equality and inequality methods) for the
1988         DefaultStyleErrorHandler class.
1989
1990         https://bugs.webkit.org/show_bug.cgi?id=37850
1991
1992         This will facilitate unit-testing for a subsequent patch,
1993         namely for https://bugs.webkit.org/show_bug.cgi?id=37850
1994
1995         * Scripts/webkitpy/style/error_handlers.py:
1996           - Added __eq__() and __ne__() to the DefaultStyleErrorHandler
1997             class.
1998
1999         * Scripts/webkitpy/style/error_handlers_unittest.py:
2000           - Added unit tests for __eq__() and __ne__().
2001           - Also included a minor clean-up refactoring of combining the
2002             StyleErrorHandlerTestBase class (which has not needed to
2003             be separate due to previous changes) into the
2004             DefaultStyleErrorHandlerTest class.
2005
2006 2010-04-20  Jakub Wieczorek  <jwieczorek@webkit.org>
2007
2008         Unreviewed.
2009
2010         Add my IRC nick to the committers.py list.
2011
2012         * Scripts/webkitpy/common/config/committers.py:
2013
2014 2010-04-20  Kim Grönholm  <kim.gronholm@nomovok.com>
2015
2016         Reviewed by Simon Hausmann.
2017
2018         [Qt] Multitouch mocking in QtLauncher doesn't work with QGraphicsView
2019         https://bugs.webkit.org/show_bug.cgi?id=37816
2020         
2021         Fix multi-touch mocking in QtLauncher when using QGraphicsView.
2022         Test: https://bug-32434-attachments.webkit.org/attachment.cgi?id=44955
2023
2024         * QtLauncher/main.cpp:
2025         (LauncherWindow::eventFilter):
2026         (LauncherWindow::initializeView):
2027
2028 2010-04-20  MORITA Hajime <morrita@google.com>
2029
2030         Unreviewed, add myself to the committers list.
2031
2032         * Scripts/webkitpy/common/config/committers.py:
2033
2034 2010-04-20  Kent Tamura  <tkent@chromium.org>
2035
2036         Reviewed by Shinichiro Hamaji.
2037
2038         [DRT/Chromium] Fix some unexpected results of editing
2039         https://bugs.webkit.org/show_bug.cgi?id=37843
2040
2041         This change fixes about 70 unexpected results.
2042         The original test_webview_delegate.cc doesn't have this bug.
2043         The bug was introduced when I ported it to WebKit tree.
2044
2045         * DumpRenderTree/chromium/WebViewHost.cpp:
2046         (printRangeDescription): Replace the latter startContainer() with endContainer().
2047
2048 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2049
2050         Unreviewed, build fix.
2051
2052         Turn off some unit tests for now - the new-run-webkit-tests dryrun
2053         tests for chromium won't work if you don't have a chromium checkout.
2054
2055         https://bugs.webkit.org/show_bug.cgi?id=37841
2056
2057         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2058
2059 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2060
2061         Unreviewed, build fix.
2062
2063         (Re-)add dryrun.py; this was renamed from passing.py in the previous
2064         CL but apparently somehow didn't get checked in.
2065
2066         https://bugs.webkit.org/show_bug.cgi?id=37841
2067
2068         * Scripts/webkitpy/layout_tests/port/dryrun.py: Added.
2069
2070 2010-04-19  Shinichiro Hamaji  <hamaji@chromium.org>
2071
2072         Reviewed by David Levin.
2073
2074         check-webkit-style: exits when encountering a deleted file
2075         https://bugs.webkit.org/show_bug.cgi?id=37122
2076
2077         This reverts the quick fix done by r57119 and makes check_patch
2078         not call check_file for deleted files.
2079
2080         Also this change fixes the behavior for "-", which should mean
2081         stdin.  Before this change, the style checker just ignored "-"
2082         with a warning message.
2083
2084         * Scripts/webkitpy/style/checker.py:
2085         * Scripts/webkitpy/style/checker_unittest.py:
2086
2087 2010-04-19  Daniel Bates  <dbates@rim.com>
2088
2089         No review, rolling out 57868.
2090         http://trac.webkit.org/changeset/57868
2091         https://bugs.webkit.org/show_bug.cgi?id=37748
2092
2093         Sheriffbot wasn't very inspirational after this patch.
2094         Instead, he was silent when you said hi :-(. Rolling
2095         out this patch so that I can debug/test this some more.
2096
2097         * Scripts/webkitpy/common/net/bugzilla.py:
2098         * Scripts/webkitpy/common/net/bugzilla_unittest.py:
2099         * Scripts/webkitpy/tool/bot/irc_command.py:
2100         * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:
2101         * Scripts/webkitpy/tool/mocktool.py:
2102
2103 2010-04-19  Daniel Bates  <dbates@rim.com>
2104
2105         Reviewed by Adam Barth.
2106
2107         https://bugs.webkit.org/show_bug.cgi?id=37748
2108
2109         Make Sheriffbot more inspirational.
2110
2111         * Scripts/webkitpy/common/net/bugzilla.py:
2112         * Scripts/webkitpy/common/net/bugzilla_unittest.py:
2113         * Scripts/webkitpy/tool/bot/irc_command.py:
2114         * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:
2115         * Scripts/webkitpy/tool/mocktool.py:
2116
2117 2010-04-19  Kevin Ollivier  <kevino@theolliviers.com>
2118
2119         [wx] Build fix, add missing header.
2120
2121         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2122
2123 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2124  
2125          Reviewed by Adam Barth.
2126  
2127          new-run-webkit-tests - repurpose the "Passing" port as "Dryrun" port
2128          that can be used to test platforms other than the one you are running
2129          on. This can be useful for checking baselines and testing code
2130          coverage.
2131  
2132          Note that running the code on the "wrong" port requires each
2133          port-specific implementation to actually not require any
2134          platform-specific python code (e.g., the chromium-win port must
2135          test for the existence of windows functions before calling them).
2136  
2137          https://bugs.webkit.org/show_bug.cgi?id=37782
2138  
2139          * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2140          * Scripts/webkitpy/layout_tests/port/dryrun.py: Renamed from WebKitTools/Scripts/webkitpy/layout_tests/port/passing.py.
2141          * Scripts/webkitpy/layout_tests/port/factory.py:
2142          * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2143
2144 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2145
2146         Reviewed by Eric Seidel.
2147
2148         new-run-webkit-tests: add --build (default) and --no-build
2149         options to make that step optional. This flag modifies what happens
2150         in port.check_build().
2151
2152         https://bugs.webkit.org/show_bug.cgi?id=37786
2153
2154         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2155         * Scripts/webkitpy/layout_tests/port/chromium.py:
2156         * Scripts/webkitpy/layout_tests/port/webkit.py:
2157
2158 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2159
2160         Unreviewed, build fix.
2161
2162         new-run-webkit-tests - fix a typo in r57480 that caused us to stop
2163         logging the actual list of unexpected results.
2164
2165         https://bugs.webkit.org/show_bug.cgi?id=37831
2166
2167         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2168
2169 2010-04-19  Dirk Pranke <dpranke@chromium.org>
2170
2171         Reviewed by Adam Barth.
2172
2173         From a patch by Tor Arne Vestbo  <tor.arne.vestbo@nokia.com>
2174
2175         new-run-webkit-tests: make the retry step more explicit
2176         https://bugs.webkit.org/show_bug.cgi?id=37606
2177
2178         It might be confusing to see the test and percent counters
2179         reset without any notice of what's going on, so we make the
2180         message that a retry-run is started explicit.
2181
2182         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2183
2184 2010-04-19  Sam Weinig  <weinig@apple.com>
2185
2186         Reviewed by Anders Carlsson.
2187
2188         Add support for opening a new window (File->New Window) to
2189         Windows MiniBrowser. Accelerator doesn't work.
2190
2191         * MiniBrowser/win/BrowserView.cpp:
2192         (createNewPage): Use BrowserView::create.
2193         * MiniBrowser/win/BrowserWindow.cpp:
2194         (BrowserWindow::onCommand): Respond to ID_FILE_NEW_WINDOW
2195         by creating a new window.
2196         * MiniBrowser/win/BrowserWindow.h:
2197         (BrowserWindow::create): Added. Don't allow creating
2198         BrowserWindows on the stack by making constructor
2199         private and exposing the create function.
2200         * MiniBrowser/win/MiniBrowser.cpp:
2201         (MiniBrowser::createNewWindow): Move new window creation
2202         logic here.
2203         * MiniBrowser/win/MiniBrowser.h:
2204         * MiniBrowser/win/MiniBrowser.rc: 
2205         * MiniBrowser/win/main.cpp:
2206         (_tWinMain): Use the new MiniBrowser::createNewWindow().
2207
2208 2010-04-19  Chris Fleizach  <cfleizach@apple.com>
2209
2210         Reviewed by Beth Dakin.
2211
2212         AX: aria-haspopup needs to be exposed
2213         https://bugs.webkit.org/show_bug.cgi?id=37808
2214
2215         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
2216         (AccessibilityUIElement::hasPopup):
2217
2218 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2219
2220         Reviewed by Eric Seidel.
2221
2222         Submit a better workaround for r57806 than the one in r57831 - log
2223         an error and exit if you try to run new-run-webkit-tests with --use-drt
2224         on Windows.
2225
2226         https://bugs.webkit.org/show_bug.cgi?id=37822
2227
2228         * Scripts/webkitpy/layout_tests/port/chromium.py:
2229
2230 2010-04-19  Jesus Sanchez-Palencia  <jesus@webkit.org>
2231
2232         Unreviewed.
2233
2234         Just adding myself as a committer.
2235
2236         * Scripts/webkitpy/common/config/committers.py:
2237
2238 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2239
2240         Reviewed by Eric Seidel.
2241
2242         Add slightly better logging to the websocket python wrapper script,
2243         including a --verbose flag for debug output.
2244
2245         https://bugs.webkit.org/show_bug.cgi?id=37233
2246
2247         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
2248         * Scripts/new-run-webkit-websocketserver:
2249
2250 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2251
2252         Reviewed by Eric Seidel.
2253
2254         new-run-webkit-tests: add a way (--print-unexpected-results) to
2255         (re-)print out the tests that produced unexpected results in the
2256         last run. Also add a way (--retry-unexpected-results) to
2257         automatically re-test them.
2258
2259         https://bugs.webkit.org/show_bug.cgi?id=37783
2260
2261         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2262
2263 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2264
2265         Reviewed by Eric Seidel.
2266
2267         Restructure the logging in new-run-webkit-tests so that many of log
2268         messages that were logged to the MeteredStream also get logged in
2269         --verbose mode.
2270
2271         https://bugs.webkit.org/show_bug.cgi?id=37780
2272
2273         * Scripts/webkitpy/layout_tests/layout_package/metered_stream.py:
2274         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2275
2276 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2277
2278         Reviewed by Eric Seidel.
2279
2280         new-run-webkit-tests: add 'summary' and 'unexpected-results' options
2281         to the --log flag. Also add a 'progress' flag to enable the regular
2282         progress bar (as opposed to 'detailed-progress', which enables the
2283         dots). Also add a 'nothing' flag to allow you to be explicit that
2284         you don't want any logging on the command line.
2285
2286         The default is
2287         '--log detailed-progress,summary,unexpected,unexpected-results'
2288
2289         (The default logging is unchanged by this patch, this just makes things
2290         properly configurable).
2291
2292         Note that 'nothing' doesn't work properly yet; I need a couple other
2293         patches to land to avoid rewriting things several different ways.
2294
2295         https://bugs.webkit.org/show_bug.cgi?id=37785
2296
2297         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2298
2299 2010-04-19  Kenneth Rohde Christiansen  <kenneth@webkit.org>
2300
2301         Reviewed by Simon Hausmann.
2302
2303         Make the URL change on committed load.
2304
2305         * QtLauncher/mainwindow.cpp:
2306         (MainWindow::buildUI):
2307         (MainWindow::setAddressUrl):
2308         * QtLauncher/mainwindow.h:
2309
2310 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2311
2312         Reviewed by Eric Seidel.
2313
2314         Fix a typo in the rebaselining tool that causes us to use "debug"
2315         instead of "Debug" in a directory path, which fails on platforms with
2316         case-sensitive filesystems.
2317
2318         https://bugs.webkit.org/show_bug.cgi?id=37819
2319
2320         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2321
2322 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2323
2324         Unreviewed, build fix.
2325
2326         Fix breakage of chromium-win canary bots caused by r57806. That patch
2327         introduced the option of using Chrome's new port of DumpRenderTree,
2328         but unfortunately that port relies on the webkit.py class
2329         implementation which uses non-blocking I/O that isn't available on
2330         Windows. This patch turns off that option and doesn't import the
2331         class if we're running on Windows.
2332
2333         https://bugs.webkit.org/show_bug.cgi?id=37817
2334
2335         * Scripts/webkitpy/layout_tests/port/chromium.py:
2336
2337 2010-04-19  James Robinson  <jamesr@chromium.org>
2338
2339         Reviewed by abarth.
2340
2341         Fix a typo
2342
2343         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2344
2345 2010-04-19  Adam Roben  <aroben@apple.com>
2346
2347         Fix run-webkit-tests when there are spaces in the path
2348
2349         Fixes <http://webkit.org/b/37809>
2350
2351         Reviewed by Adam Barth.
2352
2353         * Scripts/run-webkit-tests: Use an "indirect object" to specify the
2354         path to the harness to exec(). According to perldoc, this usage will
2355         prohibit perl from parsing the arguments to exec() via the shell,
2356         which would incorrectly split paths with spaces in them, etc.
2357
2358 2010-04-19  Dimitri Glazkov  <dglazkov@chromium.org>
2359
2360         Reviewed by Adam Roben.
2361
2362         WinLauncher.h should use LF line-endings and use native line-endings style.
2363         https://bugs.webkit.org/show_bug.cgi?id=37807
2364
2365         * WinLauncher/WinLauncher.h: Added property svn:eol-style, converted to LF line-endings.
2366
2367 2010-04-19  Kent Tamura  <tkent@chromium.org>
2368
2369         Reviewed by Shinichiro Hamaji.
2370
2371         [DRT/Chromium] Fix a test initialization problem
2372         https://bugs.webkit.org/show_bug.cgi?id=37791
2373
2374         * DumpRenderTree/chromium/DumpRenderTree.cpp:
2375         (runTest): Call resetTestController() before runFileTest(). Some
2376           controllers initialize their fields in reset() and not in their
2377           constructors.
2378
2379 2010-04-19  Kent Tamura  <tkent@chromium.org>
2380
2381         Reviewed by Eric Seidel.
2382
2383         [Chromium] new-run-webkit-tests should use WebKitDriver for --use-drt
2384         https://bugs.webkit.org/show_bug.cgi?id=37793
2385
2386         We need to use WebKitDriver instead of ChromiumDriver for Chromium
2387         DRT because its interface is different from test_shell.
2388
2389         Chromium DRT has no UI.  So we can't use it to show test results.
2390
2391         * Scripts/webkitpy/layout_tests/port/chromium.py:
2392
2393 2010-04-18  Eric Seidel  <eric@webkit.org>
2394
2395         Reviewed by Adam Roben.
2396
2397         Make failure-reason more forgiving
2398         https://bugs.webkit.org/show_bug.cgi?id=37525
2399
2400         Removed search_limit, which wasn't very useful anyway.
2401         Added a log about the long load time loading from the builders.
2402         Prompt the user for what revision to start walking from (makes it easy to restart upon failure).
2403
2404         * Scripts/webkitpy/common/net/buildbot.py:
2405         * Scripts/webkitpy/tool/commands/queries.py:
2406
2407 2010-04-18  Chris Jerdonek  <cjerdonek@webkit.org>
2408
2409         Reviewed by Shinichiro Hamaji.
2410
2411         Created a class for check-webkit-style that encapsulates iterating
2412         over text files and reading them.
2413
2414         https://bugs.webkit.org/show_bug.cgi?id=37754
2415
2416         This revision is an intermediate step towards separating our
2417         style-checking code from the logic of iterating over files and
2418         reading them.
2419
2420         * Scripts/webkitpy/common/system/logtesting.py:
2421           - Added a logMessages() method to the LoggingTestCase class.
2422             This method provides unit tests with access to the raw list
2423             of log messages in case the tester needs to do something more
2424             than simply assert the list of existing messages.
2425
2426         * Scripts/webkitpy/style/checker.py:
2427           - Added a ProcessorBase class that processors of lists of lines
2428             should eventually inherit from.
2429           - Also added a FIXME to use the ProcessorBase class and the
2430             TextFileReader class added below.
2431
2432         * Scripts/webkitpy/style/filereader.py: Added.
2433           - Created a TextFileReader class that encapsulates reading
2434             and iterating over text files.
2435
2436         * Scripts/webkitpy/style/filereader_unittest.py: Added.
2437           - Added a TextFileReaderTest class to unit-test the
2438             new TextFileReader class.
2439
2440 2010-04-15  Tony Chang  <tony@chromium.org>
2441
2442         Reviewed by Adam Barth.
2443
2444         [chromium] new-run-webkit-tests should be able to use chromium DRT
2445         https://bugs.webkit.org/show_bug.cgi?id=37645
2446
2447         Make sure that the lack of a chromium checkout doesn't cause the script to
2448         fail.
2449
2450         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
2451         * Scripts/webkitpy/layout_tests/port/chromium.py:
2452         * Scripts/webkitpy/layout_tests/port/chromium_mac.py: Fix up a few paths
2453             to be relative to an upstream output dir.
2454         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: Add --use-drt flag.
2455
2456 2010-04-18  Eric Seidel  <eric@webkit.org>
2457
2458         Unreviewed, fixing the Qt bot.
2459
2460         Add a layer of indirection when calling run-webkit-tests to allow testing new-run-webkit-tests on various platforms
2461         https://bugs.webkit.org/show_bug.cgi?id=37632
2462
2463         * Scripts/run-webkit-tests:
2464          - Exit non-zero of launching the harness fails.
2465
2466 2010-04-18  Eric Seidel  <eric@webkit.org>
2467
2468         Unreviewed, fixing the Qt bot.
2469
2470         Add a layer of indirection when calling run-webkit-tests to allow testing new-run-webkit-tests on various platforms
2471         https://bugs.webkit.org/show_bug.cgi?id=37632
2472
2473         * Scripts/run-webkit-tests:
2474          - Fix the wrapper to work for users who do not
2475            have WebKitTools/Scripts in their path.
2476
2477 2010-04-14  Eric Seidel  <eric@webkit.org>
2478
2479         Reviewed by Adam Barth.
2480
2481         Add a layer of indirection when calling run-webkit-tests to
2482         allow testing new-run-webkit-tests on various platforms.
2483         https://bugs.webkit.org/show_bug.cgi?id=37632
2484
2485         This will let us test and fix bugs in new-run-webkit-tests
2486         without needing to restart the buildbot master between tests.
2487
2488         Currently this change leaves run-webkit-tests as-is, but once
2489         its landed we will easily be able to turn on/off
2490         new-run-webkit-tests for various ports/configurations.
2491
2492         I will send a note out to webkit-dev about how we will
2493         be using this launcher script to test on the bots.
2494
2495         * Scripts/old-run-webkit-tests: Copied from WebKitTools/Scripts/run-webkit-tests.
2496         * Scripts/run-webkit-tests:
2497          - A new script which decides whether to run new- or old-
2498            run-webkit-tests based on the platform.
2499         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2500          - Add some dummy argument handling for arguments which
2501            old-run-webkit-tests supports but new-run-webkit-tests
2502            does not yet.
2503
2504 2010-04-18  Eric Seidel  <eric@webkit.org>
2505
2506         Reviewed by Adam Barth.
2507
2508         Add Gtk bots to the list of "core builders" (builders which stop the commit-queue when they turn red)
2509         https://bugs.webkit.org/show_bug.cgi?id=33295
2510
2511         The Gtk builders have been green every time I've looked
2512         at them in the last 5 days or so.  It would appear webkit
2513         is now keeping them green and we should update the core
2514         builder list to match reality.
2515
2516         * Scripts/webkitpy/common/net/buildbot.py:
2517         * Scripts/webkitpy/common/net/buildbot_unittest.py:
2518
2519 2010-04-18  Robert Hogan  <robert@webkit.org>
2520
2521         Reviewed by Simon Hausmann.
2522
2523         [Qt] Add support for LayoutTestController commands:
2524                    setSmartInsertDeleteEnabled
2525                    setSelectTrailingWhitespaceEnabled
2526                    execCommand
2527                    isCommandEnabled
2528
2529         https://bugs.webkit.org/show_bug.cgi?id=35844
2530
2531         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
2532         (WebCore::WebPage::resetSettings):
2533         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2534         (LayoutTestController::setSmartInsertDeleteEnabled):
2535         (LayoutTestController::setSelectTrailingWhitespaceEnabled):
2536         (LayoutTestController::execCommand):
2537         (LayoutTestController::isCommandEnabled):
2538         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2539
2540 2010-04-17  Adam Barth  <abarth@webkit.org>
2541
2542         Reviewed by Eric Seidel.
2543
2544         Remove steps_references and commands_references
2545         https://bugs.webkit.org/show_bug.cgi?id=37758
2546
2547         We tried using the mumble_references convention to manage our
2548         dependencies, but it doesn't seem to be providing much value for the
2549         steps and commands module because these modules are small pieces of the
2550         larger tool module.  In this patch, I've removed the references file
2551         for these modules.
2552  
2553         I've left the style_references file for the style module because that
2554         module seems better isolated from the rest of webkitpy and the
2555         style_references file appears to be providing some value.
2556
2557         * Scripts/webkitpy/tool/commands/commandtest.py:
2558         * Scripts/webkitpy/tool/commands/download_unittest.py:
2559         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
2560         * Scripts/webkitpy/tool/commands/queries_unittest.py:
2561         * Scripts/webkitpy/tool/commands/queues_unittest.py:
2562         * Scripts/webkitpy/tool/commands/queuestest.py:
2563         * Scripts/webkitpy/tool/commands/upload.py:
2564         * Scripts/webkitpy/tool/commands/upload_unittest.py:
2565         * Scripts/webkitpy/tool/commands_references.py: Removed.
2566         * Scripts/webkitpy/tool/steps/closebugforlanddiff_unittest.py:
2567         * Scripts/webkitpy/tool/steps/steps_unittest.py:
2568         * Scripts/webkitpy/tool/steps/updatechangelogswithreview_unittest.py:
2569         * Scripts/webkitpy/tool/steps/validatereviewer_unittest.py:
2570         * Scripts/webkitpy/tool/steps_references.py: Removed.
2571
2572 2010-04-17  Eric Seidel  <eric@webkit.org>
2573
2574         Reviewed by Adam Barth.
2575
2576         WebKit needs a Chromium Mac EWS Builder
2577         https://bugs.webkit.org/show_bug.cgi?id=37742
2578
2579         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
2580
2581 2010-04-17  Adam Barth  <abarth@webkit.org>
2582
2583         Fix expected results for unit test broken by
2584         http://trac.webkit.org/changeset/57772
2585
2586         * Scripts/webkitpy/common/net/buildbot_unittest.py:
2587
2588 2010-04-17  Sam Weinig  <weinig@apple.com>
2589
2590         Reviewed by Adam Roben.
2591
2592         Teach windows MiniBrowser how to work with window.open()
2593         and targeted links.
2594
2595         * MiniBrowser/win/BrowserView.cpp:
2596         (createNewPage): Create a new BrowserWindow and return its page.
2597         (showPage): Show the page.
2598         (closePage): Empty implementation.
2599         (runJavaScriptAlert): Empty implementation.
2600         (BrowserView::create): Register a UIClient.
2601         * MiniBrowser/win/BrowserView.h:
2602         (BrowserView::webView): Added.
2603         Change create to take a BrowserWindow instead of an HWND.
2604
2605         * MiniBrowser/win/BrowserWindow.cpp:
2606         (BrowserWindow::wndProc): Respond to WM_NCDESTROY.
2607         (BrowserWindow::goToURL): Added. Forwards to BrowserView.
2608         (BrowserWindow::onCreate): Don't always go to the default
2609         url. Let the caller do this.
2610         (BrowserWindow::onNCDestroy): Delete the window.
2611         * MiniBrowser/win/BrowserWindow.h:
2612         (BrowserWindow::view): Added.
2613         (BrowserWindow::window): Added.
2614
2615         * MiniBrowser/win/main.cpp:
2616         (_tWinMain):
2617         Go to the default URL for the initial page. Allocate the initial
2618         window on the heap for correctness.
2619
2620 2010-04-16  Adam Roben  <aroben@apple.com>
2621
2622         Add the Windows Debug (Test) builder to the list of core builders
2623
2624         It's been green for a few days now, and all the known Windows
2625         flakiness is Release-only.
2626
2627         Rubber-stamped by Mark Rowe.
2628
2629         * Scripts/webkitpy/common/net/buildbot.py:
2630         (BuildBot.core_builder_name_regexps): Added a regular expression to
2631         match the "Windows Debug (Test)" builder.
2632
2633 2010-04-16  Sam Weinig  <sam@webkit.org>
2634
2635         Reviewed by Anders Carlsson.
2636
2637         Fix window.open() and targeted links.
2638
2639         * MiniBrowser/mac/BrowserWindowController.m:
2640         (_createNewPage): Use the correct initializer to and load the window.
2641
2642 2010-04-16  Adam Roben  <aroben@apple.com>
2643
2644         Fix links to layout test results from build status pages
2645
2646         Reviewed by Mark Rowe.
2647
2648         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
2649         (ExtractTestResults.finished): Prepend "/" on the URL of the test
2650         results page so that it is treated as an absolute URL.
2651
2652 2010-04-16  Tony Chang  <tony@chromium.org>
2653
2654         Reviewed by Dimitri Glazkov.
2655
2656         [chromium] build DRT on Linux
2657         https://bugs.webkit.org/show_bug.cgi?id=37690
2658
2659         * Scripts/build-dumprendertree: Add support for win and linux
2660
2661 2010-04-16  Sam Weinig  <weinig@apple.com>
2662
2663         Reviewed by Adam Roben.
2664
2665         Use the threaded process model for MiniBrowser if holding down
2666         the shift key on startup.
2667
2668         * MiniBrowser/win/BrowserView.cpp:
2669         (BrowserView::create): 
2670
2671 2010-04-15  Tony Chang  <tony@chromium.org>
2672
2673         Reviewed by Dimitri Glazkov.
2674
2675         [chromium] Add TestShellGtk.cpp so we can link on Linux
2676         https://bugs.webkit.org/show_bug.cgi?id=37561
2677
2678         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp: Add new file and
2679             add platform file exceptions.
2680         * DumpRenderTree/chromium/TestShellGtk.cpp: Added.
2681         (AlarmHandler):
2682         (TestShell::waitTestFinished):
2683
2684 2010-04-15  Tony Chang  <tony@chromium.org>
2685
2686         Reviewed by Dimitri Glazkov.
2687
2688         build DRT on chromium mac
2689         https://bugs.webkit.org/show_bug.cgi?id=37639
2690
2691         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
2692         * Scripts/build-dumprendertree: enable build-dumprendertree --chromium
2693
2694 2010-04-15  Kent Tamura  <tkent@chromium.org>
2695
2696         Reviewed by Dimitri Glazkov.
2697
2698         [DRT/Chromium] Add LayoutTestHelper for Mac
2699         https://bugs.webkit.org/show_bug.cgi?id=37668
2700
2701         LayouTestHelper.mm is based on webkit/tools/test_shell/mac/layout_test_helper.mm
2702         of Chromium.
2703
2704         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
2705         * DumpRenderTree/chromium/LayoutTestHelper.mm: Added.
2706
2707 2010-04-15  Alexey Proskuryakov  <ap@apple.com>
2708
2709         Reviewed by Geoff Garen.
2710
2711         https://bugs.webkit.org/show_bug.cgi?id=37494
2712         <rdar://problem/7857060> REGRESSION (r57340): fast/events/mouse-drag-from-frame-to-other-frame.html
2713         fails on Windows
2714
2715         * DumpRenderTree/win/EventSender.cpp: (makeEventSender):
2716         * DumpRenderTree/win/EventSender.h:
2717         Tell EventSender if it's being created for a top frame.
2718         
2719         * DumpRenderTree/win/FrameLoadDelegate.cpp:
2720         (FrameLoadDelegate::didClearWindowObjectForFrameInStandardWorld): We only want to reset
2721         EventSender machinery when a new test is loaded, not when an iframe (or just its global
2722         object) is created.
2723
2724 2010-04-15  Adam Roben  <aroben@apple.com>
2725
2726         Fix Windows WebKit2 build.
2727
2728         * MiniBrowser/win/MiniBrowser.cpp:
2729
2730 2010-04-15  Ojan Vafai  <ojan@chromium.org>
2731
2732         Reviewed by Adam Barth.
2733
2734         Include codereview issue number in patch description
2735         https://bugs.webkit.org/show_bug.cgi?id=37677
2736
2737         This lets us know which rietveld issue this patch is tied to.
2738
2739         Also, make it so that --fancy-review overrides --no-review.
2740
2741         * Scripts/webkitpy/tool/steps/postcodereview.py:
2742         * Scripts/webkitpy/tool/steps/postdiff.py:
2743
2744 2010-04-15  Adam Roben  <aroben@apple.com>
2745
2746         Make --exit-after-n-failures work when all tests are timing out or crashing
2747
2748         Fixes <http://webkit.org/b/37679>.
2749
2750         Reviewed by Jon Honeycutt.
2751
2752         * Scripts/run-webkit-tests:
2753         (top level): When a test crashes or times out, break out of the main loop if
2754         stopRunningTestsEarlyIfNeeded returns true. Moved some code from the bottom of the main loop
2755         from here...
2756         (stopRunningTestsEarlyIfNeeded): ...to here.
2757
2758 2010-04-15  Sam Weinig  <sam@webkit.org>
2759
2760         Reviewed by Anders Carlsson.
2761
2762         Add WebHistoryClient support.
2763         https://bugs.webkit.org/show_bug.cgi?id=37671
2764
2765         Add HistoryClient logging.
2766
2767         * MiniBrowser/mac/BrowserWindowController.m:
2768         (_didNavigateWithNavigationData):
2769         (_didPerformClientRedirect):
2770         (_didPerformServerRedirect):
2771         (_didUpdateHistoryTitle):
2772         (-[BrowserWindowController awakeFromNib]):
2773
2774 2010-04-15  Anders Carlsson  <andersca@apple.com>
2775
2776         Reviewed by Sam Weinig.
2777
2778         Start the mini browser in threaded mode if shift is pressed during startup.
2779         https://bugs.webkit.org/show_bug.cgi?id=37670
2780
2781         * MiniBrowser/mac/AppDelegate.m:
2782         (-[BrowserAppDelegate init]):
2783
2784 2010-04-15  Chris Fleizach  <cfleizach@apple.com>
2785
2786         Reviewed by Beth Dakin.
2787
2788         AXHelp is being appended from ancestors incorrectly
2789         https://bugs.webkit.org/show_bug.cgi?id=37659
2790
2791         * DumpRenderTree/AccessibilityUIElement.cpp:
2792         (getHelpTextCallback):
2793         (AccessibilityUIElement::getJSClass):
2794         * DumpRenderTree/AccessibilityUIElement.h:
2795         * DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
2796         (AccessibilityUIElement::helpText):
2797         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
2798         (AccessibilityUIElement::helpText):
2799         * DumpRenderTree/win/AccessibilityUIElementWin.cpp:
2800         (AccessibilityUIElement::helpText):
2801
2802 2010-04-15  Dirk Pranke  <dpranke@chromium.org>
2803
2804         Reviewed by Adam Barth.
2805
2806         Re-format run-webkit-tests to fit in 80-columns for PEP-8 compliance.
2807         (broken by r57463 and r57381, at least). I've also filed bug 37477
2808         to fix check-webkit-style to catch these things.
2809
2810         https://bugs.webkit.org/show_bug.cgi?id=38586
2811
2812         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2813
2814 2010-04-15  Dirk Pranke  <dpranke@chromium.org>
2815
2816         Reviewed by Dimitri Glazkov.
2817
2818         Add a way to override the user-visible name for the test binary since
2819         some ports don't call it DumpRenderTree (e.g., Chromium Win uses
2820         test_shell, Chromium Mac uses TestShell) by adding a driver_name()
2821         method to the Port interface.
2822
2823         https://bugs.webkit.org/show_bug.cgi?id=37631
2824
2825         * Scripts/webkitpy/layout_tests/port/base.py:
2826         * Scripts/webkitpy/layout_tests/port/chromium.py:
2827         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
2828         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2829
2830 2010-04-14  Anders Carlsson  <andersca@apple.com>
2831
2832         Reviewed by Adam Roben.
2833
2834         Add "Force Repaint" to debug menu.
2835         https://bugs.webkit.org/show_bug.cgi?id=37627
2836
2837         * MiniBrowser/mac/BrowserWindowController.h:
2838         * MiniBrowser/mac/BrowserWindowController.m:
2839         (-[BrowserWindowController forceRepaint:]):
2840         * MiniBrowser/mac/English.lproj/MainMenu.xib:
2841
2842 2010-04-15  Sam Weinig  <sam@webkit.org>
2843
2844         Reviewed by Adam Roben.
2845
2846         Add debug-minibrowser script.
2847
2848         * Scripts/debug-minibrowser: Copied from Scripts/run-minibrowser.
2849         * Scripts/webkitdirs.pm:
2850
2851 2010-04-15  Roland Steiner  <rolandsteiner@chromium.org>
2852  
2853         Reviewed by Dimitri Glazkov.
2854
2855         Bug 37636 - [DRT/Chromium] Implement DRT/Chromium for Windows
2856         https://bugs.webkit.org/show_bug.cgi?id=37636
2857  
2858         Second patch: add Windows-specific implementation parts
2859         of TestShell.
2860
2861         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
2862         * DumpRenderTree/chromium/TestShell.h:
2863         (TestShell::finishedEvent):
2864         * DumpRenderTree/chromium/TestShellWin.cpp:
2865         (watchDogThread):
2866         (TestShell::waitTestFinished):
2867
2868 2010-04-15  Roland Steiner  <rolandsteiner@chromium.org>
2869
2870         Reviewed by Dimitri Glazkov.
2871
2872         Bug 37636 - [DRT/Chromium] Implement DRT/Chromium for Windows
2873         https://bugs.webkit.org/show_bug.cgi?id=37636
2874
2875         First patch: fix compiler errors.
2876
2877         * DumpRenderTree/chromium/EventSender.cpp:
2878         (EventSender::reset):
2879         (EventSender::dispatchMessage):
2880         * DumpRenderTree/chromium/LayoutTestController.cpp:
2881         (LayoutTestController::pathToLocalResource):
2882         * DumpRenderTree/chromium/TestWebWorker.h:
2883         * DumpRenderTree/chromium/TextInputController.cpp:
2884         * DumpRenderTree/chromium/WebViewHost.h:
2885         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
2886
2887 2010-04-14  Luiz Agostini  <luiz.agostini@openbossa.org>
2888
2889         Reviewed by Simon Hausmann.
2890
2891         [Qt] Moving setViewMode from DumpRenderTreeSupportQt to qwebpage.cpp
2892         https://bugs.webkit.org/show_bug.cgi?id=37622
2893
2894         Method qt_wrt_setViewMode was removed from qwebpage.cpp by mistake in r57433
2895         (bug 35844). Moving it back.
2896
2897         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2898         (LayoutTestController::setViewModeMediaFeature):
2899
2900 2010-04-15  Dirk Pranke  <dpranke@chromium.org>
2901
2902         Reviewed by Adam Barth.
2903
2904         Add some very minimal unit tests for new-run-webkit-tests. This should
2905         be enough to catch egregious brokenness like syntax errors and import
2906         declaration issues.
2907
2908         https://bugs.webkit.org/show_bug.cgi?id=37432
2909
2910         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2911         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py: Added.
2912
2913 2010-04-14  Brian Weinstein  <bweinstein@apple.com>
2914
2915         Reviewed by Adam Roben.
2916
2917         Add a way for the buildbot to kill any old processes that are running. This
2918         is useful because the Windows bots can get in states where a process remains
2919         running (httpd.exe, DumpRenderTree.exe), which causes the bots to get in a red
2920         state, and the processes must be killed manually.
2921
2922         * BuildSlaveSupport/build.webkit.org-config/master.cfg: If we are on Windows, kill
2923             the old processes that might be running.
2924         * BuildSlaveSupport/win/kill-old-processes: Added.
2925
2926 2010-04-14  Sam Weinig  <sam@webkit.org>
2927
2928         Reviewed by Anders Carlsson.
2929
2930         Tear down WebKit more completely on window closing and application
2931         termination. We still don't block application termination for pending
2932         close, but this is a step in the right direction.
2933
2934         * MiniBrowser/mac/AppDelegate.m:
2935         (-[BrowserAppDelegate applicationWillTerminate:]):
2936         * MiniBrowser/mac/BrowserWindowController.h:
2937         * MiniBrowser/mac/BrowserWindowController.m:
2938         (-[BrowserWindowController dealloc]):
2939         (-[BrowserWindowController windowWillClose:]):
2940         (-[BrowserWindowController applicationTerminating]):
2941
2942 2010-04-14  Adam Barth  <abarth@webkit.org>
2943
2944         Reviewed by Eric Seidel.
2945
2946         Use pretty patch for confirming webkit-patch diffs
2947         https://bugs.webkit.org/show_bug.cgi?id=37489
2948
2949         * Scripts/webkitpy/common/prettypatch.py: Added.
2950         * Scripts/webkitpy/tool/commands/download_unittest.py:
2951         * Scripts/webkitpy/tool/commands/upload_unittest.py:
2952         * Scripts/webkitpy/tool/mocktool.py:
2953         * Scripts/webkitpy/tool/steps/confirmdiff.py:
2954
2955 2010-04-14  Eric Seidel  <eric@webkit.org>
2956
2957         Reviewed by Adam Barth.
2958
2959         Teach webkit-patch how to handle revisions missing ChangeLogs
2960         https://bugs.webkit.org/show_bug.cgi?id=37519
2961
2962         Make commit_info_for_revision return None when revision
2963         is missing a ChangeLog.  Previously we would throw an array index
2964         exception.
2965         Teach callers how to handle None.
2966
2967         * Scripts/webkitpy/common/checkout/api.py:
2968         * Scripts/webkitpy/common/checkout/api_unittest.py:
2969         * Scripts/webkitpy/tool/commands/download.py:
2970         * Scripts/webkitpy/tool/commands/queries.py:
2971         * Scripts/webkitpy/tool/commands/sheriffbot.py:
2972
2973 2010-04-13  Eric Seidel  <eric@webkit.org>
2974
2975         Reviewed by Adam Roben.
2976
2977         Add Qt Bot to the list of "core builders" (builders which block the commit-queue when red)
2978         https://bugs.webkit.org/show_bug.cgi?id=33297
2979
2980         This is an experiment.  The bots have been green for
2981         a while.  We'll see if adding them under sheriff-bot protection
2982         will keep them green.
2983
2984         * Scripts/webkitpy/common/net/buildbot.py:
2985         * Scripts/webkitpy/common/net/buildbot_unittest.py:
2986
2987 2010-04-13  Eric Seidel  <eric@webkit.org>
2988
2989         Unreviewed, just fixing a constant in the Rietveld unit test.
2990
2991         * Scripts/webkitpy/common/net/rietveld_unittest.py
2992
2993 2010-04-13  Eric Seidel  <eric@webkit.org>
2994
2995         Unreviewed, just adding missing Mock to fix python tests.
2996
2997         * Scripts/webkitpy/tool/mocktool.py:
2998
2999 2010-04-13  Ojan Vafai  <ojan@chromium.org>
3000
3001         Reviewed by David Levin.
3002
3003         Add experimental prototype Rietveld integration to webkit-patch upload
3004         https://bugs.webkit.org/show_bug.cgi?id=37418
3005
3006         This patch adds bare-bones integration with Rietveld for code reviews.
3007         The behavior is hidden behind the --fancy-review command line flag.
3008         Currently, there's no support for uploading more than one patch per
3009         issue (which is a nice feature of Rietveld).  The plan is to play with
3010         this for a bit and see if it's useful.
3011
3012         Modified from Adam's original patch to autoinstall the rietveld upload script.
3013
3014         * Scripts/webkitpy/common/config/__init__.py:
3015         * Scripts/webkitpy/common/net/rietveld.py: Added.
3016         * Scripts/webkitpy/common/net/rietveld_unitttest.py: Added.
3017         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3018         * Scripts/webkitpy/tool/commands/upload.py:
3019         * Scripts/webkitpy/tool/commands/upload_unittest.py:
3020         * Scripts/webkitpy/tool/main.py:
3021         * Scripts/webkitpy/tool/mocktool.py:
3022         * Scripts/webkitpy/tool/steps/__init__.py:
3023         * Scripts/webkitpy/tool/steps/options.py:
3024         * Scripts/webkitpy/tool/steps/postcodereview.py: Added.
3025         * Scripts/webkitpy/tool/steps/postdiff.py:
3026
3027 2010-04-13  Sam Weinig  <sam@webkit.org>
3028
3029         Rubber-stamped by Mark Rowe.
3030
3031         Add Makefile to MiniBrowser.
3032
3033         * MiniBrowser/Makefile: Added.
3034
3035 2010-04-13  Eric Seidel  <eric@webkit.org>
3036
3037         Unreviewed, just adding a sanity check.
3038
3039         Add check to make sure commit-queue can never commit too short a message
3040         https://bugs.webkit.org/show_bug.cgi?id=37528
3041
3042         The commit-queue made bogus messages here:
3043         http://trac.webkit.org/changeset/57532
3044         http://trac.webkit.org/changeset/57534
3045
3046         This was a regression caused by adding unicode parsing for
3047         our ChangeLog files.  Popen does not seem to play nice with
3048         unicode strings.
3049
3050         I'm also adding an "assert" to make sure short ChangeLogs never happen again.
3051
3052         * Scripts/webkitpy/common/system/executive.py:
3053          - Cast input to strings before passing to POpen
3054         * Scripts/webkitpy/tool/steps/commit.py:
3055          - Validate that commit messages are not to short.
3056
3057 2010-04-13  Adam Roben  <aroben@apple.com>
3058
3059         Robustify new-run-webkit-tests against paths with spaces in them
3060
3061         Reviewed by Eric Seidel.
3062
3063         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
3064         (LayoutTestApacheHttpd.__init__): Quote all paths that we pass to
3065         Apache to ensure that paths with spaces in them are interpreted
3066         correctly.
3067
3068 2010-04-13  Csaba Osztrogonác  <ossy@webkit.org>
3069
3070         Unreviewed buildfix after r57537.
3071
3072         * DumpRenderTree/qt/LayoutTestControllerQt.h: Declaration of removeOriginAccessWhitelistEntry() added.
3073
3074 2010-04-13  Sam Weinig  <sam@webkit.org>
3075
3076         Reviewed by Maciej Stachowiak.
3077
3078         Make building new webkit API and MiniBrowser a little easier.
3079
3080         * Scripts/build-webkit: Make building with --webkit2 build the 
3081         MiniBrowser as well and tell you how to use it.
3082         * Scripts/run-minibrowser: Copied from Scripts/run-safari.
3083         * Scripts/webkitdirs.pm: Add runMiniBrowser function.
3084
3085 2010-04-12  Timothy Hatcher  <timothy@apple.com>
3086
3087         SecurityOrigin needs a way to remove individual OriginAccessEntries
3088         https://bugs.webkit.org/show_bug.cgi?id=37449
3089
3090         Reviewed by Dave Hyatt.
3091
3092         * DumpRenderTree/LayoutTestController.cpp:
3093         (removeOriginAccessWhitelistEntryCallback): Added. Call LayoutTestController::removeOriginAccessWhitelistEntry.
3094         (LayoutTestController::staticFunctions): Added removeOriginAccessWhitelistEntry.
3095         * DumpRenderTree/LayoutTestController.h:
3096         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3097         (LayoutTestController::removeOriginAccessWhitelistEntry): Added. FIXME to implement.
3098         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3099         (LayoutTestController::removeOriginAccessWhitelistEntry): Added.
3100         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3101         (LayoutTestController::removeOriginAccessWhitelistEntry): Added. FIXME to implement.
3102         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3103         (LayoutTestController::removeOriginAccessWhitelistEntry): Added.
3104         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3105         (LayoutTestController::whiteListAccessFromOrigin): FIXME to implement.
3106         (LayoutTestController::removeOriginAccessWhitelistEntry): Added. FIXME to implement.
3107
3108 2010-04-13  Timothy Hatcher  <timothy@apple.com>
3109
3110         Rename SecurityOrigin::whiteListAccessFromOrigin to addOriginAccessWhitelistEntry.
3111         And LayoutTestController.whiteListAccessFromOrigin to addOriginAccessWhitelistEntry.
3112         And SecurityOrigin::resetOriginAccessWhiteLists to resetOriginAccessWhitelists.
3113
3114         SecurityOrigin needs a way to remove individual OriginAccessEntries
3115         https://bugs.webkit.org/show_bug.cgi?id=37449
3116
3117         Reviewed by Dave Hyatt.
3118
3119         * DumpRenderTree/LayoutTestController.cpp:
3120         (addOriginAccessWhitelistEntryCallback):
3121         (LayoutTestController::staticFunctions):
3122         * DumpRenderTree/LayoutTestController.h:
3123         * DumpRenderTree/chromium/LayoutTestController.cpp:
3124         (LayoutTestController::LayoutTestController):
3125         (LayoutTestController::addOriginAccessWhitelistEntry):
3126         * DumpRenderTree/chromium/LayoutTestController.h:
3127         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3128         (LayoutTestController::addOriginAccessWhitelistEntry):
3129         * DumpRenderTree/mac/DumpRenderTree.mm:
3130         (resetWebViewToConsistentStateBeforeTesting):
3131         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3132         (LayoutTestController::addOriginAccessWhitelistEntry):
3133         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3134         (LayoutTestController::addOriginAccessWhitelistEntry):
3135         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3136         * DumpRenderTree/win/DumpRenderTree.cpp:
3137         (resetWebViewToConsistentStateBeforeTesting):
3138         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3139         (LayoutTestController::addOriginAccessWhitelistEntry):
3140         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3141         (LayoutTestController::addOriginAccessWhitelistEntry):
3142
3143 2010-04-13  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3144
3145         Reviewed by Eric Seidel.
3146
3147         Remove duplicate function for new-run-webkit-tests
3148         https://bugs.webkit.org/show_bug.cgi?id=37517
3149
3150         The version() function was already implemented.
3151
3152         * Scripts/webkitpy/layout_tests/port/webkit.py:
3153
3154 2010-04-13  Eric Seidel  <eric@webkit.org>
3155
3156         Reviewed by David Levin.
3157
3158         the commit-queue hates Tor Arne Vestbø
3159         https://bugs.webkit.org/show_bug.cgi?id=37511
3160
3161         We were failing to read reviewers out of ChangeLogs
3162         when the reviewer has unicode characters in his/her name.
3163         I fixed this by explicitly decoding from utf8 every time we
3164         read in a ChangeLog file (they are always UTF8).
3165
3166         * Scripts/webkitpy/common/checkout/changelog.py:
3167         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
3168
3169 2010-04-13  Adam Roben  <aroben@apple.com>
3170
3171         Fix run-webkit-tests on Windows with spaces in the path
3172
3173         Fixes <http://webkit.org/b/37509>.
3174
3175         Reviewed by Steve Falkenburg.
3176
3177         * Scripts/run-webkit-tests:
3178         (convertPathUsingCygpath): Remove spaces from the path before passing
3179         them to cygpath, then add them back in after conversion, as some
3180         versions of cygpath seem to convert spaces into newlines.
3181
3182 2010-04-13  Adam Barth  <abarth@webkit.org>
3183
3184         Unreviewed, but approved by Dumitru Daniliuc.  (This patch is intended
3185         to fix the downstream Chromium build bots.  Hopefully it will work!)
3186
3187         Add a driver script for the new websocket server
3188         https://bugs.webkit.org/show_bug.cgi?id=37495
3189
3190         websocket_server.py can't be run directly because its a module and not
3191         a standalone script.  This used to work by accident because it didn't
3192         depend on any other modules.
3193
3194         * Scripts/new-run-webkit-websocketserver: Added.
3195         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
3196
3197 2010-04-12  Eric Seidel  <eric@webkit.org>
3198
3199         Reviewed by Adam Barth.
3200
3201         Make new-run-webkit-test PrettyPatch failure reporting more awesome
3202         https://bugs.webkit.org/show_bug.cgi?id=37487
3203
3204         I also fixed an Executive/executive typo.
3205
3206         * Scripts/webkitpy/layout_tests/port/base.py:
3207
3208 2010-04-12  Eric Seidel  <eric@webkit.org>
3209
3210         Reviewed by Adam Barth.
3211
3212         new-run-webkit-tests should only build java support files on Mac
3213         https://bugs.webkit.org/show_bug.cgi?id=37482
3214
3215         Only the mac needs java support files, so I pushed _build_java
3216         down into the Mac port using a new hook "_check_port_build".
3217         In the process I noticed a bunch of code which could be shared
3218         between all ports and thus got rid of _tests_for_disabled_features
3219         and version() copy/paste between all webkit ports.
3220         I also made check_build only bother to check for ImageDiff if we're
3221         using pixel tests.
3222
3223         * Scripts/webkitpy/layout_tests/port/gtk.py:
3224         * Scripts/webkitpy/layout_tests/port/mac.py:
3225         * Scripts/webkitpy/layout_tests/port/qt.py:
3226         * Scripts/webkitpy/layout_tests/port/webkit.py:
3227         * Scripts/webkitpy/layout_tests/port/win.py:
3228
3229 2010-04-12  Dumitru Daniliuc  <dumi@chromium.org>
3230
3231         Unreviewed, trying to make scripts work on machines without
3232         Ruby...
3233
3234         * Scripts/webkitpy/layout_tests/port/base.py:
3235
3236 2010-04-12  Tor Arne Vestbø <tor.arne.vestbo@nokia.com>
3237
3238         Reviewed by Eric Seidel.
3239
3240         Add stub files for running new-run-webkit-tests for the Qt port
3241
3242         * Scripts/webkitpy/layout_tests/port/factory.py:
3243         * Scripts/webkitpy/layout_tests/port/qt.py: Added.
3244
3245 2010-04-12  Dirk Pranke  <dpranke@chromium.org>
3246
3247         Reviewed by Eric Seidel.
3248
3249         Modify run_webkit_tests.py to not call sys.exit() at the end of test
3250         run; doing so makes it more difficult to embed the routine for,
3251         among other things, unit tests. We push the exit calling up into
3252         new-run-webkit-tests.
3253
3254         https://bugs.webkit.org/show_bug.cgi?id=37464
3255
3256         * Scripts/new-run-webkit-tests:
3257         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3258
3259 2010-04-12  Eric Seidel  <eric@webkit.org>
3260
3261         Unreviewed.
3262
3263         Add stub Gtk implementation for new-run-webkit-tests.
3264
3265         * Scripts/webkitpy/layout_tests/port/factory.py:
3266         * Scripts/webkitpy/layout_tests/port/gtk.py: Added.
3267
3268 2010-04-12  Eric Seidel  <eric@webkit.org>
3269
3270         Reviewed by Adam Barth.
3271
3272         new-run-webkit-tests fails with exception on systems missing ruby
3273         https://bugs.webkit.org/show_bug.cgi?id=37441
3274
3275         * Scripts/webkitpy/layout_tests/port/base.py:
3276          - Catch failures similar to how wdiff code path does.
3277          - After one failure, stop trying.
3278
3279 2010-04-12  Tony Chang  <tony@chromium.org>
3280
3281         Reviewed by Eric Seidel.
3282
3283         Fix how we import simplejson based on how it's used in this file.
3284         This fixes exceptions raised when trying to write the simplejson output.
3285
3286         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3287
3288 2010-04-11  Chris Jerdonek  <cjerdonek@webkit.org>
3289
3290         Reviewed by Shinichiro Hamaji.
3291
3292         Refactored check-webkit-style so that the StyleChecker class
3293         has no dependencies on patch-related concepts.
3294
3295         https://bugs.webkit.org/show_bug.cgi?id=37065
3296
3297         This patch is an intermediate step towards making the StyleChecker
3298         class a generalized file processor that can do arbitary operations
3299         on the files corresponding to a list of paths.  This patch
3300         also simplifies the unit-testing of patch-checking code.
3301
3302         * Scripts/check-webkit-style:
3303           - Updated to use the new PatchChecker class.
3304
3305         * Scripts/webkitpy/style/checker.py:
3306           - Refactored the StyleChecker.check_patch() method into the
3307             check() method of a new PatchChecker class.
3308
3309         * Scripts/webkitpy/style/checker_unittest.py:
3310           - Refactored the unit tests as necessary, changing the
3311             StyleCheckerCheckPatchTest class to a PatchCheckerTest class.
3312
3313 2010-04-11  Adam Barth  <abarth@webkit.org>
3314
3315         Unreviewed.
3316
3317         Fix new-run-webkit-tests regressions cased by Eric's option parsing patch
3318         https://bugs.webkit.org/show_bug.cgi?id=37430
3319
3320         We need some basic unit testing of this script, or we're going to keep
3321         breaking it like this.  Added missing namespace qualifiers and
3322         propagated renaming of an option.
3323
3324         * Scripts/webkitpy/common/config/ports.py:
3325         * Scripts/webkitpy/layout_tests/driver_test.py:
3326         * Scripts/webkitpy/layout_tests/port/base.py:
3327         * Scripts/webkitpy/layout_tests/port/mac.py:
3328         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3329
3330 2010-04-11  Eric Seidel  <eric@webkit.org>
3331
3332         Reviewed by Adam Barth.
3333
3334         Break new-run-webkit-tests options into groups for easier re-use and possible relocation
3335         https://bugs.webkit.org/show_bug.cgi?id=37408
3336
3337         new-run-webkit-tests currently has one huge function for
3338         dealing with all options-parsing. 
3339         This patch is a first attempt at trying to split that large
3340         function down into smaller (hopefully more readable?) chunks
3341         dealing with the different areas of options.
3342         For example, it would make sense to move configuration
3343         options off into some module which deals with the vagries of
3344         WebKit's configuration system.  It would also make sense to move
3345         Chromium options off onto the Chromium port object (where they are used).
3346         It may make sense to move results.json options over to the results.json code.
3347         This change is a first iteration, and we will certainly need more
3348         refinement to this code over time.  Hopefully I didn't make things
3349         harder to read here.
3350
3351         * Scripts/webkitpy/layout_tests/driver_test.py:
3352         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3353
3354 2010-04-11  Adam Barth  <abarth@webkit.org>
3355
3356         Reviewed by Eric Seidel.
3357
3358         create-rollout copy needs to be updated to reflect removal of --no-build
3359         https://bugs.webkit.org/show_bug.cgi?id=37425
3360
3361         Removed --no-build and --no-test from the instructions because these
3362         don't exist anymore.
3363
3364         * Scripts/webkitpy/tool/commands/download_unittest.py:
3365             - Updated the expected results to reflect the new copy.
3366         * Scripts/webkitpy/tool/steps/postdiffforrevert.py:
3367
3368 2010-04-11  Sheriff Bot  <webkit.review.bot@gmail.com>
3369
3370         Unreviewed, rolling out r57460.
3371         http://trac.webkit.org/changeset/57460
3372         https://bugs.webkit.org/show_bug.cgi?id=37424
3373
3374         broke chromium builders (Requested by tony^work on #webkit).
3375
3376         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
3377
3378 2010-04-11  Tony Chang  <tony@chromium.org>
3379
3380         Reviewed by Darin Fisher.
3381
3382         [chromium] update chromium DEPS for upstream compile
3383         https://bugs.webkit.org/show_bug.cgi?id=36578
3384
3385         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp: Small fix to avoid a circular dependency between
3386           WebKit.gyp and webkit.gyp.
3387
3388 2010-04-11  Joseph Pecoraro  <joepeck@webkit.org>
3389
3390         Reviewed by Eric Seidel.
3391
3392         svn-apply errors out when removing directories in git
3393         https://bugs.webkit.org/show_bug.cgi?id=34871
3394
3395         * Scripts/svn-apply:
3396         (isDirectoryEmptyForRemoval): early break if the directory doesn't exist
3397         (scmRemove): have git ignore unmatched files
3398
3399 2010-04-11  Daniel Bates  <dbates@rim.com>
3400
3401         No review, rolling out 57440.
3402         http://trac.webkit.org/changeset/57440
3403         https://bugs.webkit.org/show_bug.cgi?id=27204
3404
3405         Did not handle Git patches that included both file and property
3406         changes to the same file. Rolling this change out while I look
3407         into this.
3408
3409         * Scripts/VCSUtils.pm:
3410         * Scripts/svn-apply:
3411         * Scripts/svn-unapply:
3412         * Scripts/webkitperl/VCSUtils_unittest/appendSVNExecutableBitChangeToPatch.pl: Removed.
3413         * Scripts/webkitperl/VCSUtils_unittest/parseGitFileMode.pl: Removed.
3414         * Scripts/webkitperl/VCSUtils_unittest/parseStartOfPatchOrPropertyChangeAndEndOfPropertyChange.pl: Removed.
3415
3416 2010-04-11  Adam Barth  <abarth@webkit.org>
3417
3418         Reviewed by Eric Seidel.
3419
3420         Fix typo in log path for AbstractQueue
3421         https://bugs.webkit.org/show_bug.cgi?id=37414
3422
3423         * Scripts/webkitpy/tool/commands/queues.py:
3424         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3425
3426 2010-04-11  Adam Barth  <abarth@webkit.org>
3427
3428         Unreviewed.  Add missing import statement.
3429
3430         * Scripts/webkitpy/common/system/executive.py:
3431
3432 2010-04-11  Robert Hogan  <robert@webkit.org>
3433
3434         Reviewed by Simon Hausmann.
3435
3436         [Qt] Add setWillSendRequestReturnsNull and setWillSendRequestClearHeader
3437
3438         https://bugs.webkit.org/show_bug.cgi?id=37410
3439
3440         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3441         (LayoutTestController::reset):
3442         (LayoutTestController::setWillSendRequestReturnsNull):
3443         (LayoutTestController::setWillSendRequestClearHeader):
3444         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3445
3446 2010-04-11  Csaba Osztrogonác  <ossy@webkit.org>
3447
3448         [Qt] Unreviewed buildfix for --debug build after r57433.
3449
3450         Refactor Qt DRT support in QtWebKit
3451         https://bugs.webkit.org/show_bug.cgi?id=35844
3452
3453         * QtLauncher/main.cpp: qt_drt_garbageCollector_collect(); renamed to DumpRenderTreeSupportQt::garbageCollectorCollect();
3454         (launcherMain):
3455
3456 2010-04-11  Joseph Pecoraro  <joepeck@webkit.org>
3457
3458         Reviewed by Mark Rowe.
3459
3460         Make commit-log-editor Rubber-stamp aware. And other minor cleanups.
3461         https://bugs.webkit.org/show_bug.cgi?id=37407
3462
3463         * Scripts/commit-log-editor:
3464
3465 2010-04-11  Eric Seidel  <eric@webkit.org>
3466
3467         Reviewed by Adam Barth.
3468
3469         Add PrettyPatch links to new-run-webkit-tests output
3470         https://bugs.webkit.org/show_bug.cgi?id=37406
3471
3472         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3473          - We're leaking a file handle here, add a FIXME.
3474         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
3475          - Add pretty diff links.
3476         * Scripts/webkitpy/layout_tests/port/base.py:
3477          - Add support for generating pretty diffs using PrettyPatch.
3478         * Scripts/webkitpy/layout_tests/port/webkit.py:
3479          - We're leaking another file handle here, another FIXME.
3480         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
3481          - Update write_output_files signature.
3482         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
3483          - Remove unused arguments from write_output_files.
3484          - Add support for dumping pretty diffs to write_output_files.
3485          - Fix a bunch of file descriptor leaks in this file.
3486         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
3487          - Update write_output_files signature.
3488
3489 2010-04-10  Adam Barth  <abarth@webkit.org>
3490
3491         Reviewed by Eric Seidel.
3492
3493         kill_process is copy/pasted in five places
3494         https://bugs.webkit.org/show_bug.cgi?id=37405
3495
3496         We shouldn't replicate the kill_process logic in every port.  Instead,
3497         we should move the process interaction to Executive.
3498
3499         Dirk mentioned that he wanted this abstraction to make it easier to
3500         mock things out for testing.  It turns out this function is only used
3501         in one place where it can't be used as a mock point for testing because
3502         the corresponding create process actually creates a real process.  In
3503         the long term, we should indirect both these calls through a non-static
3504         Executive as a mock point.  However, we should wait on that until we
3505         actually want to write the test.
3506
3507         * Scripts/webkitpy/layout_tests/port/base.py:
3508         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
3509         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
3510         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
3511         * Scripts/webkitpy/layout_tests/port/mac.py:
3512         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
3513         * Scripts/webkitpy/layout_tests/port/win.py:
3514
3515 2010-04-10  Adam Barth  <abarth@webkit.org>
3516
3517         Reviewed by Eric Seidel.
3518
3519         new-run-webkit-tests shouldn't alter its path to import packages
3520         https://bugs.webkit.org/show_bug.cgi?id=37404
3521
3522         * Scripts/new-run-webkit-tests:
3523         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
3524         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
3525         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
3526         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3527         * Scripts/webkitpy/layout_tests/test_types/fuzzy_image_diff.py:
3528         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
3529         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
3530         * Scripts/webkitpy/thirdparty/simplejson/decoder.py:
3531
3532 2010-04-10  Eric Seidel  <eric@webkit.org>
3533
3534         Reviewed by Adam Barth.
3535
3536         new-run-webkit-tests should store results to a directory under the build tree
3537         https://bugs.webkit.org/show_bug.cgi?id=37380
3538
3539         * Scripts/webkitpy/layout_tests/port/webkit.py:
3540
3541 2010-04-10  Daniel Bates  <dbates@rim.com>
3542
3543         Reviewed by Eric Seidel.
3544
3545         https://bugs.webkit.org/show_bug.cgi?id=27204
3546
3547         Implement support for changing the executable bit of a file.
3548         The executable bit is among the most changed file properties.
3549         Future support can include other property changes.
3550
3551         Currently, if a patch changes the executable bit of a file
3552         it is not respected by svn-apply or svn-unapply. Since the
3553         commit-queue bot uses these tools as part of its workflow,
3554         such patches cannot be committed by it. That is, such patches
3555         need to be committed by hand. Instead, we should add support
3556         for the executable bit so that such patches can be committed
3557         by the commit-queue bot.
3558
3559         * Scripts/VCSUtils.pm: Also change reference to Apple Computer, Inc.
3560         in copyright to Apple, Inc.
3561         * Scripts/svn-apply:
3562         * Scripts/svn-unapply:
3563         * Scripts/webkitperl/VCSUtils_unittest/appendSVNExecutableBitChangeToPatch.pl: Added.
3564         * Scripts/webkitperl/VCSUtils_unittest/parseGitFileMode.pl: Added.
3565         * Scripts/webkitperl/VCSUtils_unittest/parseStartOfPatchOrPropertyChangeAndEndOfPropertyChange.pl: Added.
3566
3567 2010-04-10  Eric Seidel  <eric@webkit.org>
3568
3569         Reviewed by Adam Barth.
3570
3571         WinEWS bot fails to svn update because scm.clean_working_directory leaves files around
3572         https://bugs.webkit.org/show_bug.cgi?id=37401
3573
3574         The Git-based bots don't have this trouble because
3575         Git.clean_working_directory fully removes files that were
3576         marked as "add".  SVN.clean_working_directory previously just
3577         called "svn revert" which would leave added files in the
3578         working directory untracked.  This patch makes
3579         SVN.clean_working_directory function more like
3580         Git.clean_working_directory by removing added files after revert.
3581
3582         * Scripts/webkitpy/common/checkout/scm.py:
3583          - Add SCM.absolute_path for easy conversion between
3584            repository-relative paths and absolute paths.
3585          - Add SCM.add and SCM.added_files
3586          - Make SVN.clean_working_directory remove any added_files after svn revert.
3587          - The new unit tests found a bug in Git.status_command, change to use git diff --name-status instead.
3588         * Scripts/webkitpy/common/checkout/scm_unittest.py:
3589          - Add tests for added code.
3590
3591 2010-04-10  Adam Barth  <abarth@webkit.org>
3592
3593         Reviewed by Eric Seidel.
3594
3595         new-run-webkit-tests can't find ImageDiff on Windows
3596         https://bugs.webkit.org/show_bug.cgi?id=37403
3597
3598         It turns out the build directory on Windows is structured differently
3599         than it is on other platforms.  Instead of assuming the normal
3600         structure, we should just ask perl to figure it out for us.
3601
3602         * Scripts/webkitpy/layout_tests/port/webkit.py:
3603
3604 2010-04-10  Adam Barth  <abarth@webkit.org>
3605
3606         Rubber-stamped by Eric Seidel.
3607
3608         Change "Gathering files" status line to "Collecting tests".  Gathering
3609         the files sounds silly to me.
3610
3611         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3612
3613 2010-04-10  Adam Barth  <abarth@webkit.org>
3614
3615         Reviewed by Eric Seidel.
3616
3617         Fix bugs to make new-run-webkit-tests almost run on windows
3618         https://bugs.webkit.org/show_bug.cgi?id=37400
3619
3620         Fix some minor bugs that prevent new-run-webkit-tests from being run on
3621         Windows.  I still haven't run it to completion, but I'm getting
3622         further.
3623
3624         * Scripts/webkitpy/layout_tests/port/factory.py:
3625         * Scripts/webkitpy/layout_tests/port/webkit.py:
3626
3627 2010-04-10  Robert Hogan  <robert@webkit.org>
3628
3629         Reviewed by Kenneth Rohde Christiansen.
3630
3631         Refactor Qt DRT support in QtWebKit
3632
3633         Update Qt DRT to use new DumpRenderTreeSupportQt static class.
3634
3635         https://bugs.webkit.org/show_bug.cgi?id=35844
3636
3637         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
3638         (WebCore::DumpRenderTree::DumpRenderTree):
3639         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
3640         (WebCore::DumpRenderTree::dumpFramesAsText):
3641         * DumpRenderTree/qt/DumpRenderTreeQt.h:
3642         * DumpRenderTree/qt/GCControllerQt.cpp:
3643         (GCController::collect):
3644         (GCController::collectOnAlternateThread):
3645         (GCController::getJSObjectCount):
3646         * DumpRenderTree/qt/GCControllerQt.h:
3647         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3648         (LayoutTestController::counterValueForElementById):
3649         (LayoutTestController::setViewModeMediaFeature):
3650         (LayoutTestController::setMediaType):
3651         (LayoutTestController::closeWebInspector):
3652         (LayoutTestController::showWebInspector):
3653         (LayoutTestController::evaluateInWebInspector):
3654         (LayoutTestController::setFrameFlatteningEnabled):
3655         (LayoutTestController::setJavaScriptProfilingEnabled):
3656         (LayoutTestController::setTimelineProfilingEnabled):
3657         (LayoutTestController::pauseAnimationAtTimeOnElementWithId):
3658         (LayoutTestController::pauseTransitionAtTimeOnElementWithId):
3659         (LayoutTestController::sampleSVGAnimationForElementAtTime):
3660         (LayoutTestController::numberOfActiveAnimations):
3661         (LayoutTestController::whiteListAccessFromOrigin):
3662         (LayoutTestController::setCaretBrowsingEnabled):
3663         (LayoutTestController::setDomainRelaxationForbiddenForURLScheme):
3664         (LayoutTestController::workerThreadCount):
3665         (LayoutTestController::pageNumberForElementById):
3666         (LayoutTestController::numberOfPages):
3667         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3668
3669 2010-04-10  Adam Barth  <abarth@webkit.org>
3670
3671         Reviewed by Eric Seidel.
3672
3673         Move global queue log to the logs directory so it survives git clean -f
3674         https://bugs.webkit.org/show_bug.cgi?id=37395
3675
3676         * Scripts/webkitpy/tool/commands/queues.py:
3677
3678 2010-04-10  Adam Barth  <abarth@webkit.org>
3679
3680         Reviewed by Eric Seidel.
3681
3682         SheriffBot should spam when it encounters errors
3683         https://bugs.webkit.org/show_bug.cgi?id=37329
3684
3685         We need to always update the status server so we don't get stuck in a
3686         spam loop.  I tried writing a test for this change, but it kind of
3687         got out of control.  We need a better way to do failure injection.
3688
3689         * Scripts/webkitpy/tool/commands/sheriffbot.py:
3690
3691 2010-04-10  Adam Barth  <abarth@webkit.org>
3692
3693         Unreviewed attempt to fix the Chromium Mac canary.
3694
3695         * Scripts/webkitpy/common/config/ports.py:
3696         * Scripts/webkitpy/common/config/ports_unittest.py:
3697         * Scripts/webkitpy/common/system/executive.py:
3698
3699 2010-04-10  Adam Barth  <abarth@webkit.org>
3700
3701         Unreviewed.
3702
3703         Add the Apache bits to win.py for new-run-webkit-tests
3704         https://bugs.webkit.org/show_bug.cgi?id=37397
3705
3706         I still have run this yet, but I looked around to figure out what the
3707         various paths appear to be.  I'll figure out a way to remove the
3708         copy/paste code in a future patch.
3709
3710         * Scripts/webkitpy/layout_tests/port/mac.py:
3711         * Scripts/webkitpy/layout_tests/port/webkit.py:
3712         * Scripts/webkitpy/layout_tests/port/win.py:
3713
3714 2010-04-10  Adam Barth  <abarth@webkit.org>
3715
3716         Reviewed by Eric Seidel.
3717
3718         Sketch out the win port for new-run-webkit-tests
3719         https://bugs.webkit.org/show_bug.cgi?id=37393
3720
3721         I haven't tried running this yet, but we've got to start somewhere.
3722
3723         * Scripts/webkitpy/layout_tests/port/factory.py:
3724         * Scripts/webkitpy/layout_tests/port/mac.py:
3725         * Scripts/webkitpy/layout_tests/port/webkit.py:
3726         * Scripts/webkitpy/layout_tests/port/win.py: Added.
3727
3728 2010-04-10  Ojan Vafai  <ojan@chromium.org>
3729
3730         Reviewed by Adam Barth.
3731
3732         webkit-patch land should not build and test by default
3733         https://bugs.webkit.org/show_bug.cgi?id=33631
3734
3735         Reverse the sense of --no-build and --no-test to be --build and --test.
3736         Also, decoupled the build and test options so you can test without
3737         building.
3738
3739         (Patch manngled by Adam Barth.  All bugs are his fault.)
3740
3741         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
3742         * Scripts/webkitpy/tool/commands/queues.py:
3743         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3744         * Scripts/webkitpy/tool/steps/options.py:
3745         * Scripts/webkitpy/tool/steps/runtests.py:
3746
3747 2010-04-10  Adam Barth  <abarth@webkit.org>
3748
3749         Reviewed by Eric Seidel.
3750
3751         Factor WebKitPort out of MacPort to allow for WinPort
3752         https://bugs.webkit.org/show_bug.cgi?id=37388
3753
3754         The split is a bit of a guess.  We might have to adjust things once we
3755         actually have a second port to work with.
3756
3757         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
3758         * Scripts/webkitpy/layout_tests/port/mac.py:
3759         * Scripts/webkitpy/layout_tests/port/webkit.py: Added.
3760         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
3761
3762 2010-04-10  Adam Barth  <abarth@webkit.org>
3763
3764         Reviewed by Eric Seidel.
3765
3766         run_webkit_tests.py shouldn't have platform-specific logic
3767         https://bugs.webkit.org/show_bug.cgi?id=37387
3768
3769         Dirk Pranke pointed out that my last patch was wrong because I
3770         introduced platform-specific logic into run_webkit_tests.py, limiting
3771         the parallelism in Chromium to work around a bug in the main Mac port.
3772
3773         * Scripts/webkitpy/common/system/executive.py:
3774             - Fix a typo pointed out by Chris Jerdonek.
3775         * Scripts/webkitpy/layout_tests/port/base.py:
3776         * Scripts/webkitpy/layout_tests/port/mac.py:
3777         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3778
3779 2010-04-10  Robert Hogan  <robert@webkit.org>
3780
3781         Unreviewed fix to regressions in r57416.
3782
3783         [Qt] Fix regressions in http/tests/navigation from r57416
3784
3785         Reset willSendRequestReturnsNullOnRedirect after each test to
3786         prevent it leaking to subsequent tests.
3787
3788         Error pointed out by Jakub Wieczorek.
3789
3790         https://bugs.webkit.org/show_bug.cgi?id=37237
3791
3792         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3793         (LayoutTestController::reset):
3794
3795 2010-04-11  Robert Hogan  <robert@webkit.org>
3796
3797         Reviewed by Simon Hausmann.
3798
3799         [Qt] Update layoutTestController.DumpResourceLoadCallbacks to match other ports.
3800
3801         Add support for layoutTestController.setWillSendRequestReturnsNullOnRedirect to Qt DRT.
3802         Prevent dumping resource load callbacks once layout test has dumped.
3803
3804         https://bugs.webkit.org/show_bug.cgi?id=37237
3805
3806         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
3807         (WebCore::DumpRenderTree::dump):
3808         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3809         (LayoutTestController::setWillSendRequestReturnsNullOnRedirect):
3810         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3811
3812 2010-04-10  Laszlo Gombos  <laszlo.1.gombos@nokia.com>
3813
3814         Reviewed by Kenneth Rohde Christiansen.
3815
3816         [Qt] Implement Desktop Notifications API for QtWebKit
3817         https://bugs.webkit.org/show_bug.cgi?id=35503
3818
3819         DRT stubs for notification. Enables to run and pass
3820         3 (currently disabled) tests.
3821
3822         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3823         (LayoutTestController::LayoutTestController):
3824         (LayoutTestController::grantDesktopNotificationPermission):
3825         (LayoutTestController::checkDesktopNotificationPermission):
3826         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3827
3828 2010-04-09  Dumitru Daniliuc  <dumi@chromium.org>
3829
3830         Unreviewed, another change to executive.py to make it run with
3831         python 2.4.
3832
3833         * Scripts/webkitpy/common/system/executive.py:
3834
3835 2010-04-09  Dumitru Daniliuc  <dumi@chromium.org>
3836
3837         Unreviewed, attempting to make executive.py run with python 2.4
3838         (which is still used on Chromium's Windows canary bot).
3839
3840         * Scripts/webkitpy/common/system/executive.py:
3841
3842 2010-04-09  Kevin Watters  <kevinwatters@gmail.com>
3843
3844         Reviewed by Eric Seidel.
3845
3846         [wx] Basic implementation of SVG support for wx port.
3847
3848         * wx/build/settings.py:
3849
3850 2010-04-09  Adam Barth  <abarth@webkit.org>
3851
3852         Reviewed by Eric Seidel.
3853
3854         Don't reinvent Executive.cpu_count for every port
3855         https://bugs.webkit.org/show_bug.cgi?id=37377
3856
3857         mac.py and chromium_mac.py had some copy/paste code.  This code doesn't
3858         actually have anything to do with WebKit ports.  It's really just
3859         something in the multiprocessing package.  The lame bit is that package
3860         isn't available in older versions of Python, so we need to implement a
3861         fallback.  However, we already have the fallback in common.  We don't
3862         need to reinvent it specificly for layout_tests.
3863
3864         * Scripts/webkitpy/common/system/executive.py:
3865         * Scripts/webkitpy/layout_tests/port/base.py:
3866         * Scripts/webkitpy/layout_tests/port/chromium.py:
3867         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
3868         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
3869         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
3870         * Scripts/webkitpy/layout_tests/port/mac.py:
3871         * Scripts/webkitpy/layout_tests/port/test.py:
3872         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3873
3874 2010-04-09  Adam Barth  <abarth@webkit.org>
3875
3876         Reviewed by Eric Seidel.
3877
3878         new-run-webkit-tests --release fails to build release DRT when global configuration is Debug
3879         https://bugs.webkit.org/show_bug.cgi?id=37376
3880
3881         We need to explicitly pass the --release flag.  I bet there are more
3882         instances of this bug.
3883
3884         * Scripts/webkitpy/layout_tests/port/base.py:
3885         * Scripts/webkitpy/layout_tests/port/mac.py:
3886
3887 2010-04-09  Tony Chang  <tony@chromium.org>
3888
3889         Reviewed by Eric Seidel.
3890
3891         [chromium] DRT compile fix on win/linux
3892         https://bugs.webkit.org/show_bug.cgi?id=37314
3893
3894         Looks like this was missed when upstreaming.
3895
3896         * DumpRenderTree/chromium/EventSender.cpp:
3897         (applyKeyModifier):
3898
3899 2010-04-09  Adam Barth  <abarth@webkit.org>
3900
3901         Unreviewed, but approved by Dirk Pranke.
3902
3903         rename test_expectations_test.py to test_expectations_unittest.py so it actually gets run
3904         https://bugs.webkit.org/show_bug.cgi?id=37372
3905
3906         We need to end unit tests with _unittest.py for them to be autodetected
3907         by the test harness.  +6 tests.
3908
3909         * Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py: Renamed from WebKitTools/Scripts/webkitpy/layout_tests/layout_package/test_expectations_test.py.
3910
3911 2010-04-09  Adam Barth  <abarth@webkit.org>
3912
3913         Reviewed by Eric Seidel.
3914
3915         new-run-webkit-tests should talk about DumpRenderTree not test_shell
3916         https://bugs.webkit.org/show_bug.cgi?id=37371
3917
3918         test_shell is some strange Chromium thing.
3919         DumpRenderTree (tm) is the real deal.
3920
3921         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py: Added.
3922         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
3923         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
3924         * Scripts/webkitpy/layout_tests/layout_package/test_shell_thread.py: Removed.
3925         * Scripts/webkitpy/layout_tests/port/chromium.py:
3926         * Scripts/webkitpy/layout_tests/port/server_process.py:
3927         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3928
3929 2010-04-09  Zoltan Horvath  <zoltan@webkit.org>
3930
3931         Reviewed by Alexey Proskuryakov.
3932
3933         Fix 2 issues (what were introduced in r56524) in svn-create-patch's generateDiff()
3934         https://bugs.webkit.org/show_bug.cgi?id=32582
3935
3936         Add missing return variable. Initialize $patch variable and remove unnecessary condition.
3937
3938         * Scripts/svn-create-patch:
3939
3940 2010-04-09  Kevin Ollivier  <kevino@theolliviers.com>
3941
3942         [wx] Build fix after addition of LayoutTestController method.
3943
3944         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3945         (LayoutTestController::computedStyleIncludingVisitedInfo):
3946
3947 2010-04-09  Sam Weinig  <sam@webkit.org>
3948
3949         Reviewed by Anders Carlsson.
3950
3951         Fix for https://bugs.webkit.org/show_bug.cgi?id=37368
3952         Add MiniBrowser.
3953
3954         * MiniBrowser: Added.
3955         * MiniBrowser/MiniBrowser.vcproj: Added.
3956         * MiniBrowser/MiniBrowser.xcodeproj: Added.
3957         * MiniBrowser/MiniBrowser.xcodeproj/project.pbxproj: Added.
3958         * MiniBrowser/mac: Added.
3959         * MiniBrowser/mac/AppDelegate.h: Added.
3960         * MiniBrowser/mac/AppDelegate.m: Added.
3961         * MiniBrowser/mac/BrowserWindowController.h: Added.
3962         * MiniBrowser/mac/BrowserWindowController.m: Added.
3963         * MiniBrowser/mac/English.lproj: Added.
3964         * MiniBrowser/mac/English.lproj/BrowserWindow.xib: Added.
3965         * MiniBrowser/mac/English.lproj/InfoPlist.strings: Added.
3966         * MiniBrowser/mac/English.lproj/MainMenu.xib: Added.
3967         * MiniBrowser/mac/MiniBrowser-Info.plist: Added.
3968         * MiniBrowser/mac/MiniBrowser_Prefix.pch: Added.
3969         * MiniBrowser/mac/main.m: Added.
3970         * MiniBrowser/win: Added.
3971         * MiniBrowser/win/BrowserView.cpp: Added.
3972         * MiniBrowser/win/BrowserView.h: Added.
3973         * MiniBrowser/win/BrowserWindow.cpp: Added.
3974         * MiniBrowser/win/BrowserWindow.h: Added.
3975         * MiniBrowser/win/MiniBrowser.cpp: Added.
3976         * MiniBrowser/win/MiniBrowser.h: Added.
3977         * MiniBrowser/win/MiniBrowser.rc: Added.
3978         * MiniBrowser/win/Resources: Added.
3979         * MiniBrowser/win/main.cpp: Added.
3980         * MiniBrowser/win/resource.h: Added.
3981         * MiniBrowser/win/stdafx.cpp: Added.
3982         * MiniBrowser/win/stdafx.h: Added.
3983
3984 2010-04-09  Eric Seidel  <eric@webkit.org>
3985
3986         Reviewed by Adam Barth.
3987
3988         webkit-patch attached my patch to the wrong bug
3989         https://bugs.webkit.org/show_bug.cgi?id=37015
3990
3991         The problem here is that SVN was violating SCM's implicit
3992         contract of always returning paths relative to the repository root.
3993         That can easily be fixed by telling SVN that the CWD is the repository root.
3994
3995         When fixing this I realized there are a large number of places in SCM.py where
3996         we want to consider explicitly passing self.checkout_root as the CWD.
3997         That would allow scm methods to be executed even when the CWD is not inside
3998         the scm tree at all, and would also make sure (in the case of SVN) that paths
3999         returned are relative to the root.  Git (almost always) returns paths relative
4000         to the repository root.
4001
4002         * Scripts/webkitpy/common/checkout/scm.py:
4003          - Explicitly pass self.checkout_root as cwd in run_status_and_extract_filenames
4004          - Add a ton of FIXMEs about the need to go back and decide which methods require cwd=self.checkout_root
4005            and which do not.  We'll probably add a helper function to scm (likely SCM._run) which
4006            always passes cwd=self.checkout_root to Executive.run_command
4007         * Scripts/webkitpy/common/checkout/scm_unittest.py:
4008          - Add a test for this change.
4009         * Scripts/webkitpy/tool/commands/upload.py:
4010          - Removed the explicit os.chdir to the repository root, since scm.py methods
4011            should be robust against the cwd not being equal to the root.
4012
4013 2010-04-09  Adam Roben  <aroben@apple.com>
4014
4015         Don't return 0 as a JSValueRef
4016
4017         That is an illegal use of the JSC API.
4018
4019         Fixes <http://webkit.org/b/37333> REGRESSION (r57292): :visited tests
4020         are asserting on debug Windows and GTK builds
4021
4022         Reviewed by Anders Carlsson.
4023
4024         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
4025         (LayoutTestController::computedStyleIncludingVisitedInfo):
4026         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
4027         (LayoutTestController::computedStyleIncludingVisitedInfo):
4028         Return an "undefined" JSValueRef instead of 0.
4029
4030 2010-04-09  Csaba Osztrogonác  <ossy@webkit.org>
4031
4032         Reviewed by Eric Seidel.
4033
4034         Make DumpRenderTree parallelizable
4035         https://bugs.webkit.org/show_bug.cgi?id=36899
4036
4037         * DumpRenderTree/mac/DumpRenderTree.mm:
4038         (libraryPathForDumpRenderTree): Use DUMPRENDERTREE_TEMP if exist.
4039         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
4040         (WebCore::DumpRenderTree::DumpRenderTree): Use DUMPRENDERTREE_TEMP if exist.
4041         * Scripts/run-webkit-tests:
4042           - Create a unique temporary directory and pass its path to
4043           DumpRenderTree with DUMPRENDERTREE_TEMP environment variable.
4044
4045 2010-04-09  Mark Rowe  <mrowe@apple.com>
4046
4047         Reviewed by Oliver Hunt.
4048
4049         <http://webkit.org/b/37326> IDL files are being copied in to the WebCore framework again
4050
4051         Add a script to detect the presence of inappropriate files in the frameworks.  At present
4052         it only looks for .css, .idl and .in files in the top level of WebCore.framework/Resources,
4053         as these are the only cases we've encountered recently.  It can be extended to check the
4054         other frameworks or for other inappropriate files in the future.
4055
4056         * Scripts/check-for-inappropriate-files-in-framework: Added.
4057
4058 2010-04-08  Sam Weinig  <sam@webkit.org>
4059
4060         Reviewed by Darin Adler.
4061
4062         Add option to build WebKit2 to build-webkit.
4063
4064         * Scripts/build-webkit:
4065
4066 2010-04-08  Darin Adler  <darin@apple.com>
4067
4068         Reviewed by Alexey Proskuryakov.
4069
4070         run-webkit-tests should respect argument order
4071         https://bugs.webkit.org/show_bug.cgi?id=37257
4072
4073         * Scripts/run-webkit-tests: Changed so that sorting is done only
4074         on the results of iterating directories. Test order is based on
4075         what's passed on the command line. Removed code that aimed to
4076         eliminate duplicates since it can be useful to run the same test
4077         more than once.
4078
4079 2010-04-07  David Hyatt  <hyatt@apple.com>
4080
4081         Reviewed by Oliver Hunt.
4082
4083         https://bugs.webkit.org/show_bug.cgi?id=24300, don't expose history info via CSS.  Add a new method for
4084         obtaining computed style with :visited info included.  This allows layout tests to actually tell that
4085         :visited is in effect.
4086
4087         * DumpRenderTree/LayoutTestController.cpp:
4088         (computedStyleIncludingVisitedInfoCallback):
4089         (LayoutTestController::staticFunctions):
4090         * DumpRenderTree/LayoutTestController.h:
4091         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
4092         (LayoutTestController::computedStyleIncludingVisitedInfo):
4093
4094 2010-04-07  Dirk Pranke  <dpranke@chromium.org>
4095
4096         Reviewed by Adam Barth.
4097
4098         Enable rebaseline-chromium-webkit-tests to run from a webkit-only
4099         checkout (i.e., you don't need anything from the Chromium tree checked
4100         out). This requires us to introduce the concept of a "target"
4101         port/platform that we use to get configuration information from as well
4102         as the "running" port that we use to make directories and diff images
4103         and the "rebaselining" port we use to actually manage baselines.
4104
4105         https://bugs.webkit.org/show_bug.cgi?id=37238
4106
4107         * Scripts/webkitpy/layout_tests/port/mac.py:
4108         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
4109
4110 2010-04-05  Antonio Gomes  <tonikitoo@webkit.org>
4111
4112         Reviewed by Simon Hausmann.
4113
4114         [Qt] REGRESSION:(r50665) QWebFrame::setScrollBarPolicy(Qt::Vertical,Qt::ScrollBarAlwaysOff) has no effect.
4115         https://bugs.webkit.org/show_bug.cgi?id=29431
4116
4117         Added stubs for Mac, win, gtk and wx DRTs to implement setScrollbarPolicy method.
4118
4119         * DumpRenderTree/LayoutTestController.cpp:
4120         (setScrollbarPolicyCallback):
4121         (LayoutTestController::staticFunctions):
4122         * DumpRenderTree/LayoutTestController.h:
4123         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
4124         (LayoutTestController::setScrollbarPolicy):
4125         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
4126         (LayoutTestController::setScrollbarPolicy):
4127         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
4128         (LayoutTestController::setScrollbarPolicy):
4129         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
4130         (LayoutTestController::setScrollbarPolicy):
4131
4132 2010-04-01  Antonio Gomes  <tonikitoo@webkit.org>
4133
4134         Reviewed by David Hyatt.
4135
4136         [Qt] REGRESSION:(r50665) QWebFrame::setScrollBarPolicy(Qt::Vertical,Qt::ScrollBarAlwaysOff) has no effect.
4137         https://bugs.webkit.org/show_bug.cgi?id=29431
4138
4139         Make possible to DRT to set scrollbar policies (on, off or auto).
4140
4141         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
4142         (LayoutTestController::setScrollbarPolicy):
4143         * DumpRenderTree/qt/LayoutTestControllerQt.h:
4144         * QtLauncher/main.cpp:
4145         (LauncherWindow::toggleScrollbars):
4146         (LauncherWindow::createChrome):
4147
4148 2010-04-08  Adam Barth  <abarth@webkit.org>
4149
4150         Reviewed by Eric Seidel.
4151
4152         new-run-webkit-tests should give a percent complete indication
4153         https://bugs.webkit.org/show_bug.cgi?id=37258
4154
4155         Because it's awesome.
4156
4157         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
4158
4159 2010-04-08  Adam Barth  <abarth@webkit.org>
4160
4161         Unreviewed.
4162
4163         Add back the --target option because it's needed by the downstream
4164         Chromium bots.
4165
4166         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
4167
4168 2010-04-08  Adam Barth  <abarth@webkit.org>
4169
4170         Reviewed by Eric Seidel.
4171