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