WebKitTools/ChangeLog

   1 2010-04-28  Sam Weinig  <sam@webkit.org>
   2
   3         Reviewed by Mark Rowe.
   4
   5         Make running MiniBrowser.app without explicitly setting DYLD_FRAMEWORK_PATH.
   6
   7         * MiniBrowser/MiniBrowser.xcodeproj/project.pbxproj:
   8         * MiniBrowser/mac/make-launchable.sh: Added.
   9
  10 2010-04-28  Eric Seidel  <eric@webkit.org>
  11
  12         Reviewed by David Levin.
  13
  14         Audit all uses of subprocess in webkitpy
  15         https://bugs.webkit.org/show_bug.cgi?id=38284
  16
  17         After further discussions with Jeffrey Yasskin
  18         about http://bugs.python.org/issue2320
  19         and related issues of using subprocess from
  20         multiple threads, I have learned that subprocess
  21         is known to be non-threadsafe through recent
  22         Python 2.7 builds.
  23
  24         I'm attempting to lessen our exposure to these
  25         subprocess bugs by auditing each use of subprocess
  26         in webkitpy.  I did not find any unsafe calls
  27         in my audit, but I did remove numerous unneeded
  28         import subprocess lines.
  29
  30         * Scripts/webkitpy/common/checkout/api.py:
  31         * Scripts/webkitpy/common/net/bugzilla.py:
  32         * Scripts/webkitpy/common/system/deprecated_logging_unittest.py:
  33         * Scripts/webkitpy/common/system/user.py:
  34         * Scripts/webkitpy/layout_tests/port/base.py:
  35         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
  36         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
  37         * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py: Added.
  38         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
  39         * Scripts/webkitpy/layout_tests/port/gtk.py:
  40         * Scripts/webkitpy/layout_tests/port/mac.py:
  41         * Scripts/webkitpy/layout_tests/port/qt.py:
  42         * Scripts/webkitpy/layout_tests/port/webkit.py:
  43         * Scripts/webkitpy/layout_tests/port/win.py:
  44
  45 2010-04-28  Darin Adler  <darin@apple.com>
  46
  47         Ignore a directory the Python tools creates.
  48
  49         * Scripts/webkitpy/style: Modified property svn:ignore.
  50
  51 2010-04-28  Darin Adler  <darin@apple.com>
  52
  53         * Scripts/extract-localizable-strings: Fix minor mistake in
  54         argument checking.
  55
  56 2010-04-28  Luiz Agostini  <luiz.agostini@openbossa.org>
  57
  58         Reviewed by Kenneth Rohde Christiansen.
  59
  60         [Qt] QWebPage viewMode property
  61         https://bugs.webkit.org/show_bug.cgi?id=38119
  62
  63         Replacing method qt_wrt_setViewMode by wrt_viewMode property.
  64
  65         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
  66         (LayoutTestController::setViewModeMediaFeature):
  67
  68 2010-04-28  Chris Jerdonek  <cjerdonek@webkit.org>
  69
  70         Reviewed by Shinichiro Hamaji.
  71
  72         Adjusted check-webkit-style so that files with file type NONE
  73         are automatically skipped without warning.
  74
  75         https://bugs.webkit.org/show_bug.cgi?id=38197
  76
  77         This change simplifies configuring which files to skip.  It also
  78         addresses an issue whereby check-webkit-style was unintentionally
  79         checking .vcproj files for carriage returns.
  80
  81         * Scripts/webkitpy/style/checker.py:
  82           - Moved the C++, Python, and text file extensions to new
  83             module-level configuration variables.
  84           - Removed .pyc from the _SKIPPED_FILES_WITHOUT_WARNING configuration
  85             variable.
  86           - Changed the numeric values of the FileType enum so that
  87             FileType.NONE evaluates to False.
  88           - For ProcessorDispatcher.should_skip_without_warning():
  89             - Changed the method to return True for FileType.NONE files.
  90             - Made ChangeLog files an exception to getting skipped.
  91           - Changed the StyleProcessor.process() method to raise an
  92             exception if given a file path that should not be processed.
  93
  94         * Scripts/webkitpy/style/checker_unittest.py:
  95           - Updated the unit tests and added more test cases as necessary.
  96
  97 2010-04-28  Eric Seidel  <eric@webkit.org>
  98
  99         Reviewed by Jeremy Orlow.
 100
 101         webkitpy: ScriptError('Failed to run "[u\'taskkill.exe\', u\'/f\', u\'/im\', u\'httpd.exe\']" exit_code: 128',)
 102         https://bugs.webkit.org/show_bug.cgi?id=38248
 103
 104         The previous code did not check the return code of taskkill.
 105         When I moved that callsite from using subprocess.call to
 106         Executive.run_command having a non-zero return code became an error.
 107
 108         In this change I've centralized our killall handling in executive,
 109         and added tests for it to make sure it works.
 110
 111         Currently kill_process and kill_all swallow exceptions in the cases
 112         where the process(es) to be killed do(es) not exist.
 113
 114         * Scripts/webkitpy/common/system/executive.py:
 115         * Scripts/webkitpy/common/system/executive_unittest.py:
 116         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
 117         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
 118         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
 119         * Scripts/webkitpy/layout_tests/port/gtk.py:
 120         * Scripts/webkitpy/layout_tests/port/mac.py:
 121         * Scripts/webkitpy/layout_tests/port/qt.py:
 122         * Scripts/webkitpy/layout_tests/port/win.py:
 123
 124 2010-04-28  Eric Seidel  <eric@webkit.org>
 125
 126         Reviewed by Shinichiro Hamaji.
 127
 128         wdiff_text throws ScriptError because wdiff returns non-zero when files differ
 129         https://bugs.webkit.org/show_bug.cgi?id=38246
 130
 131         wdiff returns 0 when files are the same, 1 when they differ.
 132         run_command by default raises ScriptError if the return code is non-zero.
 133         Fixed this by adding a custom error handler which only raises if the
 134         return code is not 1.
 135
 136         I broke up the huge wdiff_text() method into little pieces
 137         for easier unit testing.  There is only one functional change here
 138         and that is the addition of the custom error handler.
 139
 140         * Scripts/webkitpy/layout_tests/port/base.py:
 141         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
 142
 143 2010-04-28  Fumitoshi Ukai  <ukai@chromium.org>
 144
 145         Unreviewed build fix.
 146
 147         * DumpRenderTree/chromium/TestShell.cpp:
 148         (TestShell::dumpImage): format '%u' expects type 'unsigned int', but argument 2 has type 'size_t'.
 149         * DumpRenderTree/chromium/WebViewHost.cpp:
 150         (WebViewHost::willSendRequest): too few arguments for format.
 151
 152 2010-04-27  Shinichiro Hamaji  <hamaji@chromium.org>
 153
 154         Reviewed by Darin Adler and Eric Seidel.
 155
 156         Add layoutTestController.setPrinting()
 157         https://bugs.webkit.org/show_bug.cgi?id=37203
 158
 159         * DumpRenderTree/LayoutTestController.cpp:
 160         (LayoutTestController::LayoutTestController):
 161         (setPrintingCallback):
 162         (LayoutTestController::staticFunctions):
 163         * DumpRenderTree/LayoutTestController.h:
 164         (LayoutTestController::isPrinting):
 165         (LayoutTestController::setIsPrinting):
 166         * DumpRenderTree/mac/DumpRenderTree.mm:
 167         (dump):
 168
 169 2010-04-27  Michael Nordman  <michaeln@google.com>
 170
 171         Reviewed by Dmitry Titov.
 172
 173         [Chromium] Add two things to the webkit API to support appcaches in workers.
 174         1) WebURLRequest TargetTypes for worker and shared worker main resources.
 175         2) Factory method on class WebCommonWorkerClient to
 176            createApplicationCacheHost() for the associated worker.
 177
 178         https://bugs.webkit.org/show_bug.cgi?id=38147
 179
 180         * DumpRenderTree/chromium/TestWebWorker.h add a stub impl of the factory method
 181         (TestWebWorker::createApplicationCacheHost):
 182
 183 2010-04-27  Adam Barth  <abarth@webkit.org>
 184
 185         Reviewed by Eric Seidel.
 186
 187         run-bindings-tests doesn't work in an SVN checkout
 188         https://bugs.webkit.org/show_bug.cgi?id=38225
 189
 190         Previously detect_scm_system needed an absolute path for SVN.  Now we
 191         accept a relative path.
 192
 193         * Scripts/webkitpy/common/checkout/scm.py:
 194         * Scripts/webkitpy/common/checkout/scm_unittest.py:
 195
 196 2010-04-27  Sam Weinig  <sam@webkit.org>
 197
 198         Reviewed by Stephanie Lewis.
 199
 200         Always build WebKit2 when building on SnowLeopard and later.
 201
 202         * Scripts/build-webkit:
 203
 204 2010-04-27  Sam Weinig  <sam@webkit.org>
 205
 206         Reviewed by Maciej Stachowiak.
 207
 208         Fix for https://bugs.webkit.org/show_bug.cgi?id=38238
 209         Allow both WebKit and WebKit2 to link to the same WebCore.framework
 210
 211         * Scripts/build-webkit: Remove the UMBRELLA_LDFLAGS override when building
 212         WebCore for WebKit2, it is no longer necessary.
 213
 214 2010-04-27  James Robinson  <jamesr@chromium.org>
 215
 216         Reviewed by David Levin.
 217
 218         Fix a typo in chromium.py that causes NRWT to fail in --verbose
 219         mode in a Chromium checkout.
 220         https://bugs.webkit.org/show_bug.cgi?id=38234
 221
 222         * Scripts/webkitpy/layout_tests/port/chromium.py:
 223
 224 2010-04-27  Darin Fisher  <darin@chromium.org>
 225
 226         Reviewed by Dimitri Glazkov.
 227
 228         [Chromium] Remove deprecated form of didChangeLocationWithinPage
 229         https://bugs.webkit.org/show_bug.cgi?id=38178
 230
 231         Switch over to implementing didNavigateWithinPage.
 232
 233         * DumpRenderTree/chromium/WebViewHost.cpp:
 234         (TestWebViewDelegate::didNavigateWithinPage):
 235         (WebViewHost::didChangeLocationWithinPage):
 236         * DumpRenderTree/chromium/WebViewHost.h:
 237
 238 2010-04-27  Evan Martin  <evan@chromium.org>
 239
 240         Unreviewed.
 241
 242         Adding myself to commmitters list.
 243
 244         * Scripts/webkitpy/common/config/committers.py:
 245
 246 2010-04-27  Adam Barth  <abarth@webkit.org>
 247
 248         Reviewed by Darin Adler.
 249
 250         run-bindings-tests should use --reset-results instead of --overwrite
 251         https://bugs.webkit.org/show_bug.cgi?id=38200
 252
 253         As requested by Ojan.
 254
 255         * Scripts/run-bindings-tests:
 256
 257 2010-04-27  Adam Barth  <abarth@webkit.org>
 258
 259         Reviewed by Eric Seidel.
 260
 261         REGRESSION(r58261): webkit-patch edit-changelogs is broken
 262         https://bugs.webkit.org/show_bug.cgi?id=38204
 263
 264         In 58261, we added code to abstract step that interrogates the squash
 265         and git_commit options, but it doesn't declare that it uses those
 266         options.  That means any command that doesn't happen to declare those
 267         options might be broken if it uses the cached_lookup mechanism.
 268
 269         * Scripts/webkitpy/tool/steps/abstractstep.py:
 270         * Scripts/webkitpy/tool/steps/applypatch.py:
 271         * Scripts/webkitpy/tool/steps/applypatchwithlocalcommit.py:
 272         * Scripts/webkitpy/tool/steps/build.py:
 273         * Scripts/webkitpy/tool/steps/checkstyle.py:
 274         * Scripts/webkitpy/tool/steps/cleanworkingdirectory.py:
 275         * Scripts/webkitpy/tool/steps/closebug.py:
 276         * Scripts/webkitpy/tool/steps/closebugforlanddiff.py:
 277         * Scripts/webkitpy/tool/steps/commit.py:
 278         * Scripts/webkitpy/tool/steps/confirmdiff.py:
 279         * Scripts/webkitpy/tool/steps/createbug.py:
 280         * Scripts/webkitpy/tool/steps/ensurebuildersaregreen.py:
 281         * Scripts/webkitpy/tool/steps/ensurelocalcommitifneeded.py:
 282         * Scripts/webkitpy/tool/steps/obsoletepatches.py:
 283         * Scripts/webkitpy/tool/steps/postcodereview.py:
 284         * Scripts/webkitpy/tool/steps/postdiff.py:
 285         * Scripts/webkitpy/tool/steps/preparechangelog.py:
 286         * Scripts/webkitpy/tool/steps/runtests.py:
 287         * Scripts/webkitpy/tool/steps/update.py:
 288         * Scripts/webkitpy/tool/steps/updatechangelogswithreviewer.py:
 289         * Scripts/webkitpy/tool/steps/validatereviewer.py:
 290
 291 2010-04-27  Eric Seidel  <eric@webkit.org>
 292
 293         Reviewed by David Levin.
 294
 295         prepare-ChangeLog --bug fails on some CYGWIN installs due to missing certs
 296         https://bugs.webkit.org/show_bug.cgi?id=38212
 297
 298         * Scripts/prepare-ChangeLog:
 299          - Pass --insecure to curl to work around CYGWIN missing certs.
 300
 301 2010-04-27  Sam Weinig  <sam@webkit.org>
 302
 303         Rubber-stamped by Beth Dakin.
 304
 305         It is no longer necessary to set ENABLE_EXPERIMENTAL_SINGLE_VIEW_MODE
 306         or WTF_USE_WEB_THREAD. Remove them.
 307
 308         * Scripts/build-webkit:
 309
 310 2010-04-27  Chris Jerdonek  <cjerdonek@webkit.org>
 311
 312         Reviewed by David Levin.
 313
 314         Changed Chris Jerdonek from committer to reviewer.
 315
 316         * Scripts/webkitpy/common/config/committers.py:
 317
 318 2010-04-27  Benjamin Poulain  <benjamin.poulain@nokia.com>
 319
 320         Reviewed by Kenneth Rohde Christiansen.
 321
 322         [Qt] QGraphicsWebView: Arrow keys scroll the graphics-view instead of the web-page
 323         https://bugs.webkit.org/show_bug.cgi?id=35834
 324
 325         The scene should always have the size of the web view otherwhise it is
 326         possible to scroll the graphics view.
 327
 328         * QtLauncher/webview.cpp:
 329         (WebViewGraphicsBased::resizeEvent):
 330
 331 2010-04-27  Diego Gonzalez  <diegohcg@webkit.org>
 332
 333         Unreviewed.
 334
 335         Adding myself to committers.py
 336
 337         * Scripts/webkitpy/common/config/committers.py:
 338
 339 2010-04-27  Eric Seidel  <eric@webkit.org>
 340
 341         Reviewed by Adam Barth.
 342
 343         [chromium] new-run-webkit-tests hangs on Chromium Bots (OS X and Linux)
 344         https://bugs.webkit.org/show_bug.cgi?id=37987
 345
 346         After further research, I believe the hang is caused by:
 347         http://bugs.python.org/issue2320
 348         Basically Popen() is not reentrant.
 349         The workaround is to pass close_fds=True to Popen() on Mac/Linux.
 350
 351         I fixed our main Popen wrapper "Executive.run_command" to use close_fds=True
 352         when appropriate.
 353
 354         I audited all places we call Popen() and either moved them to run_command
 355         or left a FIXME that they are not thread safe.  A few places I added the
 356         close_fds workaround there and left an explanitory note.
 357
 358         * Scripts/webkitpy/common/checkout/scm_unittest.py:
 359          - Added note that this Popen use is not threadsafe.
 360         * Scripts/webkitpy/common/system/executive.py:
 361          - Fixed our Executive.run_* to workaround python bug 2320.
 362         * Scripts/webkitpy/common/system/user.py:
 363          _ Added note that this Popen use is not threadsafe.
 364         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py: ditto.
 365         * Scripts/webkitpy/layout_tests/port/apache_http_server.py: ditto.
 366         * Scripts/webkitpy/layout_tests/port/base.py:
 367          - Change wdiff back to using run_command now that we believe it
 368            to be threadsafe.
 369         * Scripts/webkitpy/layout_tests/port/chromium.py:
 370          - Fix to use Executive in places.
 371          - Pass self._executive down to the Driver for easier unit testing.
 372         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
 373          - Re-factor to use a _kill_all method.
 374          - Made the _kill_all method use run_command to be threadsafe.
 375         * Scripts/webkitpy/layout_tests/port/http_server.py:
 376          - Add FIXME about using Executive.
 377         * Scripts/webkitpy/layout_tests/port/server_process.py:
 378          - Use Executive to be threadsafe.
 379         * Scripts/webkitpy/layout_tests/port/webkit.py:
 380          - Pass self._executive down to the Driver.
 381         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
 382          - Add note about Popen not being threadsafe.
 383         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
 384          - Move one caller to run_command add notes about moving others.
 385
 386 2010-04-27  Adam Barth  <abarth@webkit.org>
 387
 388         Reviewed by Maciej Stachowiak.
 389
 390         REGRESSION(r58261): webkit-patch upload does not work in an SVN checkout.
 391         https://bugs.webkit.org/show_bug.cgi?id=38186
 392
 393         Unfortunately, we don't have a good way of testing this change because
 394         our test coverage of the scm.py API is poor...
 395
 396         * Scripts/webkitpy/common/checkout/scm.py:
 397         * Scripts/webkitpy/tool/steps/preparechangelog.py:
 398
 399 2010-04-26  Eric Seidel  <eric@webkit.org>
 400
 401         Reviewed by Adam Barth.
 402
 403         Remove unused code from text_diff.py
 404         https://bugs.webkit.org/show_bug.cgi?id=38170
 405
 406         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
 407          - Remove is_render_tree_dump which appears unused.
 408
 409 2010-04-26  Adam Barth  <abarth@webkit.org>
 410
 411         Reviewed by Eric Seidel.
 412
 413         webkit-patch pretty-diff is broken
 414         https://bugs.webkit.org/show_bug.cgi?id=38172
 415
 416         We need to register for these options because they're used when we look
 417         up the diff.
 418
 419         * Scripts/webkitpy/tool/steps/confirmdiff.py:
 420
 421 2010-04-26  Adam Barth  <abarth@webkit.org>
 422
 423         Reviewed by Eric Seidel.
 424
 425         Add ObjC and GObject to run-bindings-test
 426         https://bugs.webkit.org/show_bug.cgi?id=38168
 427
 428         * Scripts/run-bindings-tests:
 429
 430 2010-04-26  Adam Barth  <abarth@webkit.org>
 431
 432         Reviewed by Eric Seidel.
 433
 434         Add testing infrastructure for JSC bindings generator
 435         https://bugs.webkit.org/show_bug.cgi?id=38167
 436
 437         Add support for testing more than one bindings.  Also, converted the
 438         script to PEP8 style.
 439
 440         * Scripts/run-bindings-tests:
 441
 442 2010-04-26  Eric Seidel  <eric@webkit.org>
 443
 444         Reviewed by Adam Barth.
 445
 446         Refactor results.html generation out into a new method and test it
 447         https://bugs.webkit.org/show_bug.cgi?id=38164
 448
 449         Hopefully this results in no change in functionality.
 450
 451         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
 452         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
 453
 454 2010-04-26  Adam Barth  <abarth@webkit.org>
 455
 456         Unreviewed, rolling out r58265.
 457         http://trac.webkit.org/changeset/58265
 458         https://bugs.webkit.org/show_bug.cgi?id=38021
 459
 460         This change prevents me from uploading patches.  It also breaks sheriff-bot.
 461
 462         * Scripts/webkitpy/tool/steps/options.py:
 463
 464 2010-04-26  Eric Seidel  <eric@webkit.org>
 465
 466         Reviewed by Adam Barth.
 467
 468         [chromium] new-run-webkit-tests hangs on Chromium Bots (OS X and Linux)
 469         https://bugs.webkit.org/show_bug.cgi?id=37987
 470
 471         Rolled out:
 472         http://trac.webkit.org/changeset/58062
 473         http://trac.webkit.org/changeset/58060
 474         http://trac.webkit.org/changeset/58059
 475         http://trac.webkit.org/changeset/58055
 476         http://trac.webkit.org/changeset/58054
 477         and parts of:
 478         http://trac.webkit.org/changeset/58050
 479
 480         I also wrote some new comments and a tiny amount of new
 481         code to help make ChromiumDriver.run_test easier to read.
 482
 483         In order to unit-test my new code, I had to change ChromiumDriver
 484         to not automatically start itself when created.  That ended up
 485         being a lot of plumbing, but is hopefully easier to understand now.
 486
 487         There are no tests for the (restored) wdiff code.  wdiff does not
 488         exist on all systems, so for now we will assume it worked since
 489         it is just old code being reverted.
 490
 491          * Scripts/webkitpy/layout_tests/driver_test.py:
 492           - Use create_driver instead of start_driver, and be sure to call .stop()
 493          * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
 494           - Use create_driver instead of start_driver
 495          * Scripts/webkitpy/layout_tests/port/base.py:
 496           - Added a comment to explain that diffs are binary files.
 497           - Various patch reverts relating to wdiff
 498           - Add Driver._command_wrapper to share code between WebKitDriver and ChromiumDriver.
 499           - Made _command_wrapper use shlex.split to get rid of the FIXME.
 500          * Scripts/webkitpy/layout_tests/port/base_unittest.py: Added.
 501           - test the new _command_wrapper
 502          * Scripts/webkitpy/layout_tests/port/chromium.py:
 503          - Use _command_wrapper to get rid of a bunch of ugly code.
 504          - Make __init__ stop auto-starting.
 505          - Rename create_driver to start_driver.
 506          - Added _write_command_and_read_line to make it possible to
 507            put a FIXME next to read_line() w/o having to put it in two places.
 508          - Moved test_shell command building into _test_shell_command and tested it.
 509          - Fix comments to say test_shell since ChromiumDriver is test_shell only.
 510          * Scripts/webkitpy/layout_tests/port/chromium_unittest.py: Added.
 511           - Test the new test_shell_command method.
 512          * Scripts/webkitpy/layout_tests/port/dryrun.py:
 513           - Rename create_driver to start_driver.
 514          * Scripts/webkitpy/layout_tests/port/test.py:
 515           - Rename create_driver to start_driver.
 516          * Scripts/webkitpy/layout_tests/port/webkit.py:
 517           - Rename create_driver to start_driver.
 518           - Treat output as binary arrays.
 519          * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
 520           - Treat diff files as binary.
 521          * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
 522           - Treat diff files as binary.
 523
 524 2010-04-26  Adam Barth  <abarth@webkit.org>
 525
 526         Reviewed by Eric Seidel.
 527
 528         Move bindings test directory into the scripts directory
 529         https://bugs.webkit.org/show_bug.cgi?id=38161
 530
 531         Change script to point to the new location of these data files.
 532
 533         * Scripts/run-bindings-tests:
 534
 535 2010-04-26  Adam Barth  <abarth@webkit.org>
 536
 537         Unreviewed.  Mark run-bindings-tests executable so we can execute it.
 538
 539         * Scripts/run-bindings-tests:
 540
 541 2010-04-26  Tony Chang  <tony@chromium.org>
 542
 543         Reviewed by David Levin.
 544
 545         [chromium] build-webkit --chromium should build release by default
 546         https://bugs.webkit.org/show_bug.cgi?id=38028
 547
 548         * Scripts/build-dumprendertree: Also should build release by default
 549         * Scripts/build-webkit: Make sure to pass command line args through
 550         * Scripts/webkitdirs.pm: Build the right configuration
 551
 552 2010-04-26  Ojan Vafai  <ojan@chromium.org>
 553
 554         Reviewed by Eric Seidel.
 555
 556         complete rietveld bugzilla integration
 557         https://bugs.webkit.org/show_bug.cgi?id=38021
 558
 559         Makes --fancy-review the default. All this means is that the patch will
 560         be uploaded to rietveld in addition to bugs.webkit.org.
 561
 562         * Scripts/webkitpy/tool/steps/options.py:
 563
 564 2010-04-26  Chris Jerdonek  <cjerdonek@webkit.org>
 565
 566         Reviewed by Shinichiro Hamaji.
 567
 568         For check-webkit-style, renamed the style/processors/ directory
 569         to style/checkers/.
 570
 571         https://bugs.webkit.org/show_bug.cgi?id=38122
 572
 573         * Scripts/webkitpy/style/checker.py:
 574           - Updated import statements.
 575
 576         * Scripts/webkitpy/style/checker_unittest.py:
 577           - Updated import statements.
 578
 579         * Scripts/webkitpy/style/checkers: Copied from WebKitTools/Scripts/webkitpy/style/processors.
 580         * Scripts/webkitpy/style/processors: Removed.
 581         * Scripts/webkitpy/style/processors/__init__.py: Removed.
 582         * Scripts/webkitpy/style/processors/common.py: Removed.
 583         * Scripts/webkitpy/style/processors/common_unittest.py: Removed.
 584         * Scripts/webkitpy/style/processors/cpp.py: Removed.
 585         * Scripts/webkitpy/style/processors/cpp_unittest.py: Removed.
 586         * Scripts/webkitpy/style/processors/python.py: Removed.
 587         * Scripts/webkitpy/style/processors/python_unittest.py: Removed.
 588         * Scripts/webkitpy/style/processors/python_unittest_input.py: Removed.
 589         * Scripts/webkitpy/style/processors/text.py: Removed.
 590         * Scripts/webkitpy/style/processors/text_unittest.py: Removed.
 591
 592 2010-04-06  Ojan Vafai  <ojan@chromium.org>
 593
 594         Reviewed by Adam Barth.
 595
 596         Include git commits in the diff for webkit-patch upload/land.
 597         https://bugs.webkit.org/show_bug.cgi?id=36394
 598
 599         Adds --squash, --no-squash and --git-commit.
 600
 601         --git-commit will use a specific local commit for land/upload.
 602         If a commit-range is specified, then that range is treated as
 603         a single squashed commit.
 604
 605         --squash will squash all local changes including working copy changes
 606         into a single patch.
 607
 608         --no-squash is the legacy behavior (upload only considers the working copy,
 609         land commits the working copy and then each local commit separately to SVN)
 610
 611         If neither is specified, then an informative error is raised if there is
 612         more than one local commit or when there are local commit(s) and working
 613         copy changes.
 614
 615         If the webkit-patch.squash git config parameter is set, then
 616         that will be respected instead of raising an error.
 617
 618         * Scripts/check-webkit-style:
 619         * Scripts/webkitpy/common/checkout/api.py:
 620         * Scripts/webkitpy/common/checkout/api_unittest.py:
 621         * Scripts/webkitpy/common/checkout/scm.py:
 622         * Scripts/webkitpy/common/checkout/scm_unittest.py:
 623         * Scripts/webkitpy/style/optparser.py:
 624         --git-since is removed and --git-commit no longer implies commit_id..
 625         Instead, it treats that individual commit, but also supports commit ranges
 626         (e.g. commit_id..) as arguments.
 627         * Scripts/webkitpy/style/optparser_unittest.py:
 628         * Scripts/webkitpy/style_references.py:
 629         * Scripts/webkitpy/tool/commands/download.py:
 630         * Scripts/webkitpy/tool/commands/upload.py:
 631         * Scripts/webkitpy/tool/main.py:
 632         * Scripts/webkitpy/tool/mocktool.py:
 633         * Scripts/webkitpy/tool/steps/abstractstep.py:
 634         * Scripts/webkitpy/tool/steps/applypatchwithlocalcommit.py:
 635         * Scripts/webkitpy/tool/steps/checkstyle.py:
 636         * Scripts/webkitpy/tool/steps/commit.py:
 637         * Scripts/webkitpy/tool/steps/options.py:
 638         * Scripts/webkitpy/tool/steps/postdiff.py:
 639         * Scripts/webkitpy/tool/steps/preparechangelog.py:
 640         * Scripts/webkitpy/tool/steps/preparechangelogforrevert.py:
 641         * Scripts/webkitpy/tool/steps/updatechangelogswithreviewer.py:
 642         * Scripts/webkitpy/tool/steps/validatereviewer.py:
 643
 644 2010-04-26  Chris Jerdonek  <cjerdonek@webkit.org>
 645
 646         Reviewed by Shinichiro Hamaji.
 647
 648         Deleted the StyleChecker-related classes that are no longer
 649         being used.
 650
 651         https://bugs.webkit.org/show_bug.cgi?id=38118
 652
 653         * Scripts/webkitpy/style/checker.py:
 654           - Deleted the DeprecatedStyleChecker class.
 655
 656         * Scripts/webkitpy/style/checker_unittest.py:
 657           - Deleted the StyleCheckerTest, StyleCheckerCheckFileBase,
 658             StyleCheckerCheckFileTest, and StyleCheckerCheckPathsTest classes.
 659
 660 2010-04-26  Chris Jerdonek  <cjerdonek@webkit.org>
 661
 662         Reviewed by Shinichiro Hamaji.
 663
 664         Changed the StyleChecker class to use the new TextFileReader class.
 665
 666         https://bugs.webkit.org/show_bug.cgi?id=37066
 667
 668         This revision separates the code responsible for reading and iterating
 669         over text files from the rest of check-webkit-style.
 670
 671         * Scripts/check-webkit-style:
 672           - Changed the script to use the new StyleProcessor and
 673             TextFileReader classes instead of the StyleChecker class.
 674
 675         * Scripts/webkitpy/style/checker.py:
 676           - Added a FIXME to rename many of uses of the word "processor" to
 677             "checker".  We did this to clarify the difference between
 678             ProcessorBase instances passed to the TextFileReader and
 679             classes that process and check lines for style.
 680           - Added a FIXME to remove FileType.NONE as a possible return value
 681             of ProcessorDispatcher._file_type().  This will better consolidate
 682             the logic of which files should be skipped.
 683           - Added a FIXME to delete the StyleChecker class.
 684           - Added the StyleProcessor class which implements ProcessorBase.
 685             This class is responsible for processing lines to check style
 686             (but not for reading files).  For each file, this class creates
 687             creates both a carriage-return checker and a format-specific
 688             style checker (e.g. one of C++, Python, etc).
 689           - Updated the PatchChecker class to use a TextFileReader instance
 690             instead of a StyleChecker.
 691
 692         * Scripts/webkitpy/style/checker_unittest.py:
 693           - Added the StyleProcessor_EndToEndTest class to do "end-to-end"
 694             tests of the StyleProcessor class.
 695           - Added the StyleProcessor_CodeCoverageTest to test the
 696             StyleProcessor class with more complete code coverage.
 697             Unlike the StyleProcessor_EndToEndTest class, this class makes
 698             heavy use of mock parameters.
 699           - Added FIXME's to delete the unit test classes that are no
 700             longer needed.
 701           - Updated the PatchCheckerTest class to use a MockTextFileReader
 702             instead of a MockStyleChecker.
 703
 704         * Scripts/webkitpy/style/filereader.py:
 705           - Updated the TextFileReader class to use the preferred logic
 706             of checking file existence at the beginning of the process_file()
 707             method instead of in the except block, per
 708             https://bugs.webkit.org/show_bug.cgi?id=37122
 709
 710         * Scripts/webkitpy/style/filereader_unittest.py:
 711           - In the TextFileReaderTest class:
 712             - Moved the test_process_file__should_not_process() method.
 713             - Added a test_process_file__file_stdin() method to test
 714               the file path "-".
 715
 716 2010-04-20  Robert Hogan  <robert@webkit.org>
 717
 718         Reviewed by Simon Hausmann.
 719
 720         [Qt] Add more support for textInputController
 721
 722         Add support for selectedRange(), setMarkedText(), insertText(),
 723         and firstRectForCharacterRange().
 724
 725         Unskip tests:
 726
 727         fast/forms/input-maxlength-ime-preedit.html
 728         fast/forms/input-maxlength-ime-completed.html
 729         fast/text/international/thai-cursor-position.html
 730         fast/events/ime-composition-events-001.html
 731         editing/selection/5825350-1.html
 732         editing/selection/5825350-2.html
 733         editing/selection/mixed-editability-10.html
 734
 735         https://bugs.webkit.org/show_bug.cgi?id=35702
 736
 737         * DumpRenderTree/qt/TextInputControllerQt.cpp:
 738         (TextInputController::setMarkedText):
 739         (TextInputController::insertText):
 740         (TextInputController::selectedRange):
 741         (TextInputController::firstRectForCharacterRange):
 742         * DumpRenderTree/qt/TextInputControllerQt.h:
 743
 744 2010-04-23  Eric Seidel  <eric@webkit.org>
 745
 746         Reviewed by Adam Barth.
 747
 748         check-webkit-style complains about non-utf8 data in layout test result
 749         https://bugs.webkit.org/show_bug.cgi?id=38027
 750
 751         The problem was we were assuming patch files/diff output as utf-8.
 752         Turns out they're not.  We have to treat them as binary data because
 753         a single patch may have multiple text files in it with conflicting encodings!
 754
 755         * Scripts/webkitpy/common/checkout/api.py:
 756          - contents_at_revision returns a byte array, so decode it to unicode
 757            before passing it to parse_latest_entry_from_file
 758         * Scripts/webkitpy/common/checkout/api_unittest.py:
 759          - Update our mock mock_contents_at_revision to match the encoding
 760            semantics of the real one.
 761         * Scripts/webkitpy/common/checkout/scm.py:
 762          - Be careful not to decode output which may contain file contents
 763            (like diff, cat or show) as the encoding for that content is unknown.
 764         * Scripts/webkitpy/common/checkout/scm_unittest.py:
 765          - Update our tests to use both latin1 and utf-8 encoded data.
 766         * Scripts/webkitpy/common/net/bugzilla.py:
 767          - _fill_attachment_form should not assume unicode data.  Callers
 768            may wish to attach other types of files to bugs.
 769         * Scripts/webkitpy/common/prettypatch.py:
 770          - Diffs are byte arrays, deal with them as such.
 771         * Scripts/webkitpy/common/prettypatch_unittest.py:
 772          - Test to make sure we handle diffs with multiple conflicting encodings.
 773         * Scripts/webkitpy/common/system/executive_unittest.py:
 774          - Make sure that our unicode support does not break our
 775            byte array input support for run_command.
 776
 777 2010-04-23  Sam Weinig  <sam@webkit.org>
 778
 779         Reviewed by David Levin.
 780
 781         Fix for https://bugs.webkit.org/show_bug.cgi?id=38060
 782         Split up Threading.h
 783
 784         Add necessary forwarding headers.
 785
 786         * DumpRenderTree/ForwardingHeaders/wtf/Atomics.h: Added.
 787         * DumpRenderTree/ForwardingHeaders/wtf/ThreadSafeShared.h: Added.
 788         * DumpRenderTree/ForwardingHeaders/wtf/ThreadingPrimitives.h: Added.
 789
 790 2010-04-23  Xiaomei Ji  <xji@chromium.org>
 791
 792         No need to review.
 793
 794         Add xji as committer.
 795
 796         * Scripts/webkitpy/common/config/committers.py:
 797
 798 2010-04-23  Kevin Ollivier  <kevino@theolliviers.com>
 799
 800         [wx] Build fix, add platform stub for new LayoutTestController method.
 801
 802         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
 803         (LayoutTestController::markerTextForListItem):
 804
 805 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
 806
 807         Reviewed by Adam Barth.
 808
 809         new-run-webkit-tests: add a "--retry-failures" flag and a
 810         "--no-retry-failures" flag (the former is the default). Also, rename
 811         "--print-unexpected-results" and "--retry-unexpected-results" to
 812         "--print-last-failures" and "--retry-last-failures" because the
 813         retry flag was confusing. The new flag names aren't great, but
 814         hopefully they're less confusing.
 815
 816         https://bugs.webkit.org/show_bug.cgi?id=37838
 817
 818         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
 819
 820 2010-04-22  Sam Weinig  <sam@webkit.org>
 821
 822         Reviewed by Anders Carlsson.
 823
 824         Fix for https://bugs.webkit.org/show_bug.cgi?id=38022
 825         Move isMainThread predicate function to MainThread.h
 826
 827         Added forwarding header for MainThread.h
 828
 829         * DumpRenderTree/ForwardingHeaders/wtf/MainThread.h: Added.
 830
 831 2010-04-23  Laszlo Gombos  <laszlo.1.gombos@nokia.com>
 832
 833         Reviewed by Kenneth Rohde Christiansen.
 834
 835         [Qt] Disable Netscape plugin support for minimal configuration
 836         https://bugs.webkit.org/show_bug.cgi?id=38026
 837
 838         Pass the minimal configuration option to Qt build system
 839         as part of the CONFIG variable.
 840
 841         * Scripts/build-webkit:
 842
 843 2010-04-22  Fumitoshi Ukai  <ukai@chromium.org>
 844
 845         Unreviewed fix.
 846
 847         * Scripts/webkitpy/layout_tests/port/websocket_server.py: fix NameError: global name 'f' is not defined.
 848
 849 2010-04-22  James Robinson  <jamesr@chromium.org>
 850
 851         Reviewed by Adam Barth.
 852
 853         Revert 58077 and follow-ups.  It's broken.
 854         https://bugs.webkit.org/show_bug.cgi?id=37664
 855
 856         * Scripts/new-run-webkit-httpd:
 857         * Scripts/new-run-webkit-websocketserver:
 858         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
 859         * Scripts/webkitpy/layout_tests/port/http_server.py:
 860         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
 861
 862 2010-04-22  Yaar Schnitman  <yaar@chromium.org>
 863
 864         Reviewed by Adam Barth.
 865
 866         Integrate v8 testing utility with webkit tests
 867         https://bugs.webkit.org/show_bug.cgi?id=37731
 868
 869         * Scripts/run-bindings-tests: Added.
 870
 871 2010-04-22  Eric Seidel  <eric@webkit.org>
 872
 873         Reviewed by Dimitri Glazkov.
 874
 875         new-run-webkit-tests --verbose shows ever-increasing #EOF lines
 876         https://bugs.webkit.org/show_bug.cgi?id=37794
 877
 878         * Scripts/webkitpy/layout_tests/port/webkit.py:
 879          - Remove the assert() since our ServerProcess code does not always
 880            seem to be reading the full stderr output (or we're not waiting for it to).
 881
 882 2010-04-22  Eric Seidel  <eric@webkit.org>
 883
 884         Reviewed by Adam Barth.
 885
 886         Add code to help debug new-run-webkit-test hangs on the Chromium bots
 887         https://bugs.webkit.org/show_bug.cgi?id=38011
 888
 889         I can see no reasonable way to test this change.
 890         Stubbing out sys._current_frames() and traceback.extract_stack
 891         seems folly.  Dumping real data would have line number
 892         (and possibly other call stack) variance between runs.
 893
 894         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
 895          - Add _dump_thread_states and _dump_thread_states_if_necessary
 896            to have our main thread dump the states of all threads every
 897            60 seconds when running in verbose mode.
 898          - Better document what is going on in our main loop.
 899
 900 2010-04-22  Sam Weinig  <sam@webkit.org>
 901
 902         Reviewed by Anders Carlsson.
 903
 904         Fix for https://bugs.webkit.org/show_bug.cgi?id=38002
 905         Add rudimentary statistics gathering for WebKit2
 906
 907         * MiniBrowser/MiniBrowser.xcodeproj/project.pbxproj:
 908         * MiniBrowser/mac/AppDelegate.h:
 909         * MiniBrowser/mac/AppDelegate.m:
 910         (-[BrowserAppDelegate init]):
 911         (-[BrowserAppDelegate newWindow:]):
 912         (-[BrowserAppDelegate getCurrentPageNamespace]):
 913         (-[BrowserAppDelegate _setProcessModel:]):
 914         (-[BrowserAppDelegate showStatisticsWindow:]):
 915         (-[BrowserAppDelegate applicationWillTerminate:]):
 916         * MiniBrowser/mac/BrowserStatisticsWindow.xib: Added.
 917         * MiniBrowser/mac/BrowserStatisticsWindowController.h: Added.
 918         * MiniBrowser/mac/BrowserStatisticsWindowController.m: Added.
 919         (-[BrowserStatisticsWindowController initWithThreadedWKContextRef:processWKContextRef:]):
 920         (-[BrowserStatisticsWindowController dealloc]):
 921         (-[BrowserStatisticsWindowController windowDidLoad]):
 922         (-[BrowserStatisticsWindowController refreshStatistics:]):
 923         * MiniBrowser/mac/BrowserWindowController.h:
 924         * MiniBrowser/mac/English.lproj/MainMenu.xib:
 925         * MiniBrowser/mac/MiniBrowser_Prefix.pch:
 926
 927 2010-04-22  Dave Moore  <davemoore@chromium.org>
 928
 929         Reviewed by Dimitri Glazkov.
 930
 931         Add test support for icon changes.
 932
 933         https://bugs.webkit.org/show_bug.cgi?id=33812
 934
 935         * DumpRenderTree/LayoutTestController.cpp:
 936         (LayoutTestController::LayoutTestController):
 937         (dumpIconChangesCallback):
 938         (LayoutTestController::staticFunctions):
 939         * DumpRenderTree/LayoutTestController.h:
 940         (LayoutTestController::dumpIconChanges):
 941         (LayoutTestController::setDumpIconChanges):
 942         * DumpRenderTree/win/FrameLoadDelegate.cpp:
 943         (FrameLoadDelegate::didChangeIcons):
 944         * DumpRenderTree/win/FrameLoadDelegate.h:
 945         * WinLauncher/WinLauncher.h:
 946         (WinLauncherWebHost::didChangeIcons):
 947
 948 2010-04-22  Eric Seidel  <eric@webkit.org>
 949
 950         Unreviewed.  Script fix.  Will ask dpranke to look tomorrow.
 951
 952         new-run-webkit-tests --verbose shows ever-increasing #EOF lines
 953         https://bugs.webkit.org/show_bug.cgi?id=37794
 954
 955         The bots are seeing cases where .error is sometimes empty.
 956         Lets make the code not crash in that case for now.
 957
 958         * Scripts/webkitpy/layout_tests/port/webkit.py:
 959
 960 2010-04-22  Fumitoshi Ukai  <ukai@chromium.org>
 961
 962         Unreviewed fix.
 963
 964         * Scripts/webkitpy/layout_tests/port/websocket_server.py: fix NameError: global name 'f' is not defined.
 965
 966 2010-04-22  Fumitoshi Ukai  <ukai@chromium.org>
 967
 968         Unreviewed fix.
 969
 970         * Scripts/new-run-webkit-websocketserver: PyWebSocket no longer takes keyword argument 'register_cygwin'.
 971
 972 2010-04-22  Eric Seidel  <eric@webkit.org>
 973
 974         Unreviewed.  Restore the Leopard bots to using
 975         old-run-webkit-tests for now.
 976
 977         * Scripts/run-webkit-tests:
 978
 979 2010-04-22  Fumitoshi Ukai  <ukai@chromium.org>
 980
 981         Unreviewed fix.
 982
 983         * Scripts/webkitpy/layout_tests/port/http_server.py: fix NameError: global name 'env' is not defined.
 984
 985 2010-04-22  Fumitoshi Ukai  <ukai@chromium.org>
 986
 987         Unreviewed fix.
 988
 989         * Scripts/webkitpy/layout_tests/port/websocket_server.py: fix NameError: global name 'env' is not defined.
 990
 991 2010-04-22  Fumitoshi Ukai  <ukai@chromium.org>
 992
 993         Reviewed by Adam Barth.
 994
 995         Chromium: Add --chromium option to new-run-webkit-websocketserver
 996         https://bugs.webkit.org/show_bug.cgi?id=37664
 997
 998         Move cygwin setup logic in chromium_win.py.
 999
1000         * Scripts/new-run-webkit-httpd: remove register_cygwin parameter to pass http_server.
1001         * Scripts/new-run-webkit-websocketserver: add --chromium flag
1002         * Scripts/webkitpy/layout_tests/port/chromium_win.py: setup for cygwin
1003         * Scripts/webkitpy/layout_tests/port/http_server.py: remove cygwin setup logic
1004         * Scripts/webkitpy/layout_tests/port/websocket_server.py: remove cygwin setup logic
1005
1006 2010-04-22  Eric Seidel  <eric@webkit.org>
1007
1008         Reviewed by Adam Barth.
1009
1010         new-run-webkit-tests --verbose shows ever-increasing #EOF lines
1011         https://bugs.webkit.org/show_bug.cgi?id=37794
1012
1013         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1014          - Fix the log message to explain that this is stderr output, not test output.
1015         * Scripts/webkitpy/layout_tests/port/webkit.py:
1016          - Be sure to reset the server_process.error after reading (seems like the wrong
1017            place to do this, but at least this fixes the bug and dpranke and I can talk
1018            about better designs later).
1019          - Also remove the #EOF from the stderr output before returning it.
1020
1021 2010-04-22  Eric Seidel  <eric@webkit.org>
1022
1023         Unreviewed.  Fixing new-run-webkit-tests on the bots.
1024
1025         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1026          - Fixing _compat_shim_option to take variable keyword args.
1027          - Turns out that nargs + callback is ignored unless type is specified.
1028            So I added the type so that the option was properly consumed.
1029            This was why new-run-webkit-tests couldn't find any tests, it
1030            was looking for them under "20" because that was the argument
1031            which should have been consumed by --exit-after-n-failures.
1032
1033 2010-04-22  Tony Chang  <tony@chromium.org>
1034
1035         Not reviewed, build fix for chromium Windows.
1036
1037         [chromium] fix ImageDiff compile on windows
1038         https://bugs.webkit.org/show_bug.cgi?id=37979
1039
1040         * DumpRenderTree/chromium/ImageDiff.cpp:
1041
1042 2010-04-22  Adam Barth  <abarth@webkit.org>
1043
1044         Unreviewed, rolling out r58069.
1045         http://trac.webkit.org/changeset/58069
1046         https://bugs.webkit.org/show_bug.cgi?id=27751
1047
1048         Broke compile on Windows.
1049
1050         * DumpRenderTree/LayoutTestController.cpp:
1051         (LayoutTestController::staticFunctions):
1052         * DumpRenderTree/LayoutTestController.h:
1053         * DumpRenderTree/chromium/LayoutTestController.cpp:
1054         (LayoutTestController::LayoutTestController):
1055         (LayoutTestController::overridePreference):
1056         * DumpRenderTree/chromium/LayoutTestController.h:
1057         * DumpRenderTree/chromium/TestShell.cpp:
1058         (TestShell::resetWebSettings):
1059         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1060         * DumpRenderTree/mac/DumpRenderTree.mm:
1061         (resetDefaultsToConsistentValues):
1062         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1063         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1064         (WebCore::WebPage::WebPage):
1065         (WebCore::WebPage::resetSettings):
1066         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1067         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1068         * DumpRenderTree/win/DumpRenderTree.cpp:
1069         (resetDefaultsToConsistentValues):
1070         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1071         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
1072
1073 2010-04-22  Eric Seidel  <eric@webkit.org>
1074
1075         Unreviewed.  Just adding logging.
1076
1077         Adding logging to help debug why the Leopard Bot
1078         can't find any tests to run.
1079
1080         * Scripts/webkitpy/layout_tests/layout_package/test_files.py:
1081         * Scripts/webkitpy/layout_tests/port/base.py:
1082         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1083          - Make it a fatal error to have no tests to run.
1084
1085 2010-04-22  Abhishek Arya  <inferno@chromium.org>
1086
1087         Reviewed by Adam Barth.
1088
1089         Add support for controlling clipboard access from javascript.
1090         Clipboard access from javascript is enabled in test framework.
1091         https://bugs.webkit.org/show_bug.cgi?id=27751
1092
1093         * DumpRenderTree/LayoutTestController.cpp:
1094         (setJavaScriptCanAccessClipboardCallback):
1095         (LayoutTestController::staticFunctions):
1096         * DumpRenderTree/LayoutTestController.h:
1097         * DumpRenderTree/chromium/LayoutTestController.cpp:
1098         (LayoutTestController::LayoutTestController):
1099         (LayoutTestController::setJavaScriptCanAccessClipboard):
1100         (LayoutTestController::overridePreference):
1101         * DumpRenderTree/chromium/LayoutTestController.h:
1102         * DumpRenderTree/chromium/TestShell.cpp:
1103         (TestShell::resetWebSettings):
1104         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1105         (LayoutTestController::setJavaScriptCanAccessClipboard):
1106         * DumpRenderTree/mac/DumpRenderTree.mm:
1107         (resetDefaultsToConsistentValues):
1108         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1109         (LayoutTestController::setJavaScriptCanAccessClipboard):
1110         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1111         (WebCore::WebPage::WebPage):
1112         (WebCore::WebPage::resetSettings):
1113         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1114         (LayoutTestController::setJavaScriptCanAccessClipboard):
1115         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1116         * DumpRenderTree/win/DumpRenderTree.cpp:
1117         (resetDefaultsToConsistentValues):
1118         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1119         (LayoutTestController::setJavaScriptCanAccessClipboard):
1120         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
1121         (LayoutTestController::setJavaScriptCanAccessClipboard):
1122
1123 2010-04-21  Adam Barth  <abarth@webkit.org>
1124
1125         Reviewed by Eric Seidel.
1126
1127         webkit-patch can't fetch attachments on security bugs
1128         https://bugs.webkit.org/show_bug.cgi?id=37975
1129
1130         Instead of calling CURL, we just need to use our Mechanize object,
1131         which understand bugs.webkit.org authentication.
1132
1133         * Scripts/webkitpy/common/checkout/api.py:
1134         * Scripts/webkitpy/common/net/bugzilla.py:
1135
1136 2010-04-21  Eric Seidel  <eric@webkit.org>
1137
1138         Unreviewed.  Temporary commit, will roll-out before morning.
1139
1140         Turning on new-run-webkit-tests for the Leopard build bot
1141         for testing of the harness.
1142         Users should not noctice (except for the results.html difference).
1143
1144         * Scripts/run-webkit-tests:
1145
1146 2010-04-21  Eric Seidel  <eric@webkit.org>
1147
1148         Unreviewed.  Fix typo in my previous fix attempt.
1149
1150         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1151         https://bugs.webkit.org/show_bug.cgi?id=37765
1152
1153         More on-the-bot debugging, sigh.  I wish I had a local build.
1154
1155         * Scripts/webkitpy/layout_tests/port/base.py:
1156
1157 2010-04-21  Eric Seidel  <eric@webkit.org>
1158
1159         Unreviewed.  Fix typo in my previous fix attempt.
1160
1161         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1162         https://bugs.webkit.org/show_bug.cgi?id=37765
1163
1164         * Scripts/webkitpy/layout_tests/port/base.py:
1165          - Add self. to class variable access.
1166
1167 2010-04-21  Eric Seidel  <eric@webkit.org>
1168
1169         Unreviewed.  Another attempt to fix NRWT for chromium.
1170
1171         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1172         https://bugs.webkit.org/show_bug.cgi?id=37765
1173
1174         * Scripts/webkitpy/layout_tests/port/base.py:
1175          - wdiff_text was returning a byte array instead of a
1176            unicode string.  The simple fix was to just decode
1177            the result.  However, seeing so much duplicated code
1178            with Executive made me cry, so I re-wrote the function
1179            to be more like pretty_patch_text and use run_command
1180            (which already knows how to handle unicode).
1181
1182 2010-04-21  Adam Barth  <abarth@webkit.org>
1183
1184         Reviewed by Eric Seidel.
1185
1186         Windows tests buildbots are too slow to be core builders
1187         https://bugs.webkit.org/show_bug.cgi?id=37970
1188
1189         It's 10:45.  The Windows test bots are still hours behind.  They're too
1190         slow to be core builders.  When they get fast, we can add them back.
1191
1192         * Scripts/webkitpy/common/net/buildbot.py:
1193         * Scripts/webkitpy/common/net/buildbot_unittest.py:
1194
1195 2010-04-21  Eric Seidel  <eric@webkit.org>
1196
1197         Unreviewed.  Attempt one more time to fix NRWT for chromium.
1198
1199         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1200         https://bugs.webkit.org/show_bug.cgi?id=37765
1201
1202         I'm debugging in the blind because I don't have a chromium
1203         build on this laptop.
1204
1205         * Scripts/webkitpy/layout_tests/port/chromium.py:
1206          - Apply the previous fix to a second caller.
1207
1208 2010-04-21  Eric Seidel  <eric@webkit.org>
1209
1210         Unreviewed.  Attempt one more time to fix NRWT for chromium.
1211
1212         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1213         https://bugs.webkit.org/show_bug.cgi?id=37765
1214
1215         * Scripts/webkitpy/layout_tests/port/chromium.py:
1216          - Fix handling of test_shell output so that we always
1217            decode it as utf-8.
1218            Unlike DumpRenderTree test_shell does not ever return
1219            pixel data.  It spits out the pixel dumps in a separate
1220            file, thus all output over stdout is utf-8 text.
1221
1222 2010-04-21  Eric Seidel  <eric@webkit.org>
1223
1224         Unreviewed.  Attempt one more time to fix NRWT for chromium.
1225
1226         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1227         https://bugs.webkit.org/show_bug.cgi?id=37765
1228
1229         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
1230          - Pass encoding to _save_baseline_data and write_output_files
1231         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
1232          - Make _save_baseline_data and write_output_files take an encoding.
1233         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
1234         - Pass encoding to _save_baseline_data and write_output_files
1235
1236 2010-04-21  No'am Rosenthal  <noam.rosenthal@nokia.com>
1237
1238         Reviewed by Kenneth Rohde Christiansen.
1239
1240         [Qt] QtLauncher: make FPS measurement accurate
1241         https://bugs.webkit.org/show_bug.cgi?id=37934
1242
1243         Instead of counting paints, which are not interchangeable with frames that
1244         the user sees, we now set an arbitrary timer for FPS measurements. The idea is
1245         that if the main thread is delayed for any reason, that timer would be delayed
1246         as well.
1247
1248         * QtLauncher/QtLauncher.pro:
1249         * QtLauncher/fpstimer.cpp: Added.
1250         (FpsTimer::FpsTimer):
1251         (FpsTimer::numFrames):
1252         (FpsTimer::start):
1253         (FpsTimer::stop):
1254         (FpsTimer::timerEvent):
1255         * QtLauncher/fpstimer.h: Added.
1256         * QtLauncher/webview.cpp:
1257         (WebViewGraphicsBased::setFrameRateMeasurementEnabled):
1258         (WebViewGraphicsBased::updateFrameRate):
1259         (WebViewGraphicsBased::paintEvent):
1260         * QtLauncher/webview.h:
1261
1262 2010-04-21  Eric Seidel  <eric@webkit.org>
1263
1264         Unreviewed, just adding missing ":" in python file.
1265
1266         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1267         https://bugs.webkit.org/show_bug.cgi?id=37765
1268
1269         new-run-webkit-tests --chromium was borked.
1270
1271         * Scripts/webkitpy/layout_tests/port/chromium.py:
1272
1273 2010-04-21  Eric Seidel  <eric@webkit.org>
1274
1275         Reviewed by Adam Barth.
1276
1277         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1278         https://bugs.webkit.org/show_bug.cgi?id=37765
1279
1280         Third time is the charm.  I've fixed all the
1281         new-run-webkit-tests regressions from previous attempts.
1282
1283         I fixed the queue to not ignore Tor as a reviwer in r57531,
1284         but instead it throws an exception every time his name is in a patch.
1285
1286         This fixes our Executive.run_command code to work around a Popen
1287         bug http://bugs.python.org/issue5290 whereby python versions before 2.6
1288         do not correctly handle unicode objects as input or output to
1289         Popen.communicate.
1290
1291         Following the advice of:
1292         http://farmdev.com/talks/unicode/
1293         I have changed all of webkitpy to use unicode() objects as strings
1294         instead of str objects (which in Python 3 are renamed "bytes").
1295
1296         String literals were left as "foo" instead of converting to u"foo"
1297         as u"foo" is only required if the string has a non-ascii code point.
1298         Python is smart about comparing str() and unicode() values and will
1299         log an error to the console if the comparison is ever invalid.
1300
1301         Executive.run* now correctly accept and return unicode() objects.
1302         I attempted to fix all the places that we call .write() to make sure we
1303         encode any unicode() objects into utf-8.
1304
1305         I removed several uses of StringIO.  StringIO should generally always be
1306         passed a unicode() value.
1307
1308         Likewise I replaced most calls to open() with codecs.open().
1309         codecs.open() matches Python 3 open semantics in requiring an encoding
1310         parameter.  Files opened with codecs.open() with a unicode-compatible
1311         encoding will vend unicode() objects from their read() calls, like how
1312         StringIO created with a unicode() object will do.
1313
1314         I also deployed "with" statements wider (now that the project has
1315         settled on Python 2.5) to close a bunch of file descriptor leaks.
1316
1317         * Scripts/webkitpy/common/checkout/api_unittest.py:
1318          - Read/write utf-8 files instead of ascii.
1319          - Update the tests to use test for proper unicode() handling.
1320         * Scripts/webkitpy/common/checkout/changelog.py:
1321          - Document that parse_latest_entry_from_file expects
1322            file-like objects which return unicode strings.
1323         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
1324          - Use unicode() strings instead of str() byte arrays.
1325          - Deploy "with" to close file descriptor leaks.
1326         * Scripts/webkitpy/common/checkout/commitinfo.py:
1327          - Remove unneeded import.
1328         * Scripts/webkitpy/common/checkout/scm.py:
1329          - Remove use of str().
1330         * Scripts/webkitpy/common/checkout/scm_unittest.py:
1331          - Read/write utf-8 files and use unicode() strings in testing.
1332         * Scripts/webkitpy/common/config/committers.py:
1333          - Use \u instead of \x to make slightly clearer what we're doing.
1334         * Scripts/webkitpy/common/net/bugzilla.py:
1335          - Add a new _string_contents() method and explain why
1336             we have to call unicode() on the result of soup.string
1337             and why it's safe to do so w/o needing to pass a codec name.
1338           - Remove the (unused) support for passing a file object to add_patch_to_bug().
1339         * Scripts/webkitpy/common/net/buildbot.py:
1340          - Use unicode() instead of str() when needing to coax a
1341             NavigableString object into a unicode() object.
1342         * Scripts/webkitpy/common/net/buildbot_unittest.py:
1343          - Add a test which contains a unicode builder name.
1344         * Scripts/webkitpy/common/net/statusserver.py:
1345          - Remove use of str()
1346         * Scripts/webkitpy/common/prettypatch.py:
1347          - Write out the patch file as utf-8.
1348         * Scripts/webkitpy/common/system/autoinstall.py:
1349          - Write out files with a explicit encodings.
1350          - Deploy "with" to close file descriptor leaks.
1351         * Scripts/webkitpy/common/system/deprecated_logging.py:
1352          - Write out log files as utf-8.
1353         * Scripts/webkitpy/common/system/executive.py:
1354          - Make run* properly take and return unicode() objects.
1355          - Cleaned up input handling in run_command a little by adding
1356            a _compute_input() method which can return early instead of having
1357            such a long/cluttered if-block.
1358         * Scripts/webkitpy/common/system/executive_unittest.py:
1359          - Added a unit test to make sure we don't break Tor again!
1360         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1361          - Write out the test list as utf-8.
1362         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1363          - Write out json files as utf-8.
1364         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1365          - Deploy "with" to close file descriptor leaks.
1366         * Scripts/webkitpy/layout_tests/port/chromium.py: Add Executive.py FIXME.
1367         * Scripts/webkitpy/layout_tests/port/chromium_linux.py: ditto.
1368         * Scripts/webkitpy/layout_tests/port/chromium_mac.py: ditto.
1369         * Scripts/webkitpy/layout_tests/port/gtk.py: ditto.
1370         * Scripts/webkitpy/layout_tests/port/mac.py: ditto.
1371         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1372          - Make the skipped file parsing test unicode.
1373         * Scripts/webkitpy/layout_tests/port/qt.py: Add Executive.py FIXME.
1374         * Scripts/webkitpy/layout_tests/port/server_process.py: ditto.
1375         * Scripts/webkitpy/layout_tests/port/webkit.py:
1376          - Deploy "with" to close file descriptor leaks.
1377         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
1378          - Make explicit the encodings of log files and pid files.
1379         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
1380          - Make encodings explicit and deploy "with".
1381         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: ditto.
1382         * Scripts/webkitpy/layout_tests/test_types/image_diff.py: ditto.
1383         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py: ditto.
1384         * Scripts/webkitpy/layout_tests/test_types/text_diff.py: ditto.
1385         * Scripts/webkitpy/style/filereader_unittest.py: ditto.
1386         * Scripts/webkitpy/thirdparty/__init__.py: ditto.
1387         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
1388          - Removed extra import.
1389         * Scripts/webkitpy/tool/commands/queues.py:
1390          - No need to map args to strings now that run_command does.
1391          - Update test results to match args changes.
1392          - Document our global argument hacks.
1393         * Scripts/webkitpy/tool/commands/upload.py:
1394          - Pass the diff directly to add_patch_to_bug instead of creating a StringIO file wrapper.
1395         * Scripts/webkitpy/tool/mocktool.py:
1396          - Rename add_patch_to_bug argument to match bugzilla.py
1397         * Scripts/webkitpy/tool/steps/abstractstep.py:
1398          - Executive.run_* now require lists instead of strings.
1399            The lack of this change was what broke webkit-patch
1400            for svn users the first time this was landed.
1401         * Scripts/webkitpy/tool/steps/postdiff.py:
1402          - Pass the diff directly to add_patch_to_bug instead of creating a StringIO file wrapper.
1403         * Scripts/webkitpy/tool/steps/postdiffforcommit.py: ditto
1404         * Scripts/webkitpy/tool/steps/postdiffforrevert.py: ditto
1405         * Scripts/webkitpy/tool/steps/steps_unittest.py:
1406          - Fixed spurious logging seen when running test-webkitpy
1407
1408 2010-04-21  Chris Fleizach  <cfleizach@apple.com>
1409
1410         Reviewed by Alexey Proskuryakov.
1411
1412         aria-liveregion-notifications.html fails on leopard release bot
1413         https://bugs.webkit.org/show_bug.cgi?id=37112
1414
1415         Change the way that notifications are listened for by forcing clients
1416         to call a remove listener as well to match the add listener. DRT will
1417         assert if those are not done in the correct order.
1418
1419         * DumpRenderTree/AccessibilityUIElement.cpp:
1420         (removeNotificationListenerCallback):
1421         (AccessibilityUIElement::getJSClass):
1422         * DumpRenderTree/AccessibilityUIElement.h:
1423         * DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
1424         (AccessibilityUIElement::removeNotificationListener):
1425         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
1426         (-[AccessibilityNotificationHandler initWithPlatformElement:]):
1427         (-[AccessibilityNotificationHandler dealloc]):
1428         (-[AccessibilityNotificationHandler _notificationReceived:]):
1429         (-[AccessibilityNotificationHandler setCallback:]):
1430         (AccessibilityUIElement::AccessibilityUIElement):
1431         (AccessibilityUIElement::~AccessibilityUIElement):
1432         (AccessibilityUIElement::addNotificationListener):
1433         (AccessibilityUIElement::removeNotificationListener):
1434         * DumpRenderTree/win/AccessibilityUIElementWin.cpp:
1435         (AccessibilityUIElement::removeNotificationListener):
1436
1437 2010-04-21  Anders Carlsson  <andersca@apple.com>
1438
1439         Reviewed by Sam Weinig.
1440
1441         Add debug menu items to show/hide the Web View.
1442         https://bugs.webkit.org/show_bug.cgi?id=37958
1443
1444         * MiniBrowser/mac/BrowserWindowController.h:
1445         * MiniBrowser/mac/BrowserWindowController.m:
1446         (-[BrowserWindowController showHideWebView:]):
1447         (-[BrowserWindowController removeReinsertWebView:]):
1448         (-[BrowserWindowController validateMenuItem:]):
1449         * MiniBrowser/mac/English.lproj/MainMenu.xib:
1450
1451 2010-04-21  Eric Seidel  <eric@webkit.org>
1452
1453         Unreviewed.  Rolling out unicode() changes as they broke NRWT for chromium.
1454         Rolling out:
1455         http://trac.webkit.org/changeset/58014
1456         http://trac.webkit.org/changeset/58016
1457         http://trac.webkit.org/changeset/58020
1458
1459         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbo
1460         https://bugs.webkit.org/show_bug.cgi?id=37765
1461
1462         * Scripts/webkitpy/common/checkout/api_unittest.py:
1463         * Scripts/webkitpy/common/checkout/changelog.py:
1464         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
1465         * Scripts/webkitpy/common/checkout/commitinfo.py:
1466         * Scripts/webkitpy/common/checkout/scm.py:
1467         * Scripts/webkitpy/common/checkout/scm_unittest.py:
1468         * Scripts/webkitpy/common/config/committers.py:
1469         * Scripts/webkitpy/common/net/bugzilla.py:
1470         * Scripts/webkitpy/common/net/buildbot.py:
1471         * Scripts/webkitpy/common/net/buildbot_unittest.py:
1472         * Scripts/webkitpy/common/net/statusserver.py:
1473         * Scripts/webkitpy/common/prettypatch.py:
1474         * Scripts/webkitpy/common/system/autoinstall.py:
1475         * Scripts/webkitpy/common/system/deprecated_logging.py:
1476         * Scripts/webkitpy/common/system/executive.py:
1477         * Scripts/webkitpy/common/system/executive_unittest.py:
1478         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1479         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1480         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1481         * Scripts/webkitpy/layout_tests/port/chromium.py:
1482         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1483         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1484         * Scripts/webkitpy/layout_tests/port/gtk.py:
1485         * Scripts/webkitpy/layout_tests/port/mac.py:
1486         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1487         * Scripts/webkitpy/layout_tests/port/qt.py:
1488         * Scripts/webkitpy/layout_tests/port/server_process.py:
1489         * Scripts/webkitpy/layout_tests/port/webkit.py:
1490         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
1491         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
1492         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1493         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
1494         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
1495         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
1496         * Scripts/webkitpy/style/filereader_unittest.py:
1497         * Scripts/webkitpy/thirdparty/__init__.py:
1498         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
1499         * Scripts/webkitpy/tool/commands/queues.py:
1500         * Scripts/webkitpy/tool/commands/queues_unittest.py:
1501         * Scripts/webkitpy/tool/commands/upload.py:
1502         * Scripts/webkitpy/tool/mocktool.py:
1503         * Scripts/webkitpy/tool/steps/abstractstep.py:
1504         * Scripts/webkitpy/tool/steps/postdiff.py:
1505         * Scripts/webkitpy/tool/steps/postdiffforcommit.py:
1506         * Scripts/webkitpy/tool/steps/postdiffforrevert.py:
1507         * Scripts/webkitpy/tool/steps/steps_unittest.py:
1508
1509 2010-04-21  Eric Seidel  <eric@webkit.org>
1510
1511         Unreviewed, fixing NRWT for real this time.
1512
1513         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1514         https://bugs.webkit.org/show_bug.cgi?id=37765
1515
1516         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
1517          - Add a hack to fix new-run-webkit-tests
1518            my understanding of codecs.open(encoding=None)
1519            must have been wrong.
1520
1521 2010-04-21  Eric Seidel  <eric@webkit.org>
1522
1523         Unreviewed, just fixing exception seen on builders.
1524
1525         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1526         https://bugs.webkit.org/show_bug.cgi?id=37765
1527
1528         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
1529          - Pass and encoding to _write_into_file_at_path
1530
1531 2010-04-21  Eric Seidel  <eric@webkit.org>
1532
1533         Reviewed by Adam Barth.
1534
1535         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
1536         https://bugs.webkit.org/show_bug.cgi?id=37765
1537
1538         I fixed the queue to not ignore Tor as a reviwer in r57531,
1539         but instead it throws an exception every time his name is in a patch.
1540
1541         This fixes our Executive.run_command code to work around a Popen
1542         bug http://bugs.python.org/issue5290 whereby python versions before 2.6
1543         do not correctly handle unicode objects as input or output to
1544         Popen.communicate.
1545
1546         Following the advice of:
1547         http://farmdev.com/talks/unicode/
1548         I have changed all of webkitpy to use unicode() objects as strings
1549         instead of str objects (which in Python 3 are renamed "bytes").
1550
1551         String literals were left as "foo" instead of converting to u"foo"
1552         as u"foo" is only required if the string has a non-ascii code point.
1553         Python is smart about comparing str() and unicode() values and will
1554         log an error to the console if the comparison is ever invalid.
1555
1556         Executive.run* now correctly accept and return unicode() objects.
1557         I attempted to fix all the places that we call .write() to make sure we
1558         encode any unicode() objects into utf-8.
1559
1560         I removed several uses of StringIO.  StringIO should generally always be
1561         passed a unicode() value.
1562
1563         Likewise I replaced most calls to open() with codecs.open().
1564         codecs.open() matches Python 3 open semantics in requiring an encoding
1565         parameter.  Files opened with codecs.open() with a unicode-compatible
1566         encoding will vend unicode() objects from their read() calls, like how
1567         StringIO created with a unicode() object will do.
1568
1569         I also deployed "with" statements wider (now that the project has
1570         settled on Python 2.5) to close a bunch of file descriptor leaks.
1571
1572         * Scripts/webkitpy/common/checkout/api_unittest.py:
1573          - Read/write utf-8 files instead of ascii.
1574          - Update the tests to use test for proper unicode() handling.
1575         * Scripts/webkitpy/common/checkout/changelog.py:
1576          - Document that parse_latest_entry_from_file expects
1577            file-like objects which return unicode strings.
1578         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
1579          - Use unicode() strings instead of str() byte arrays.
1580          - Deploy "with" to close file descriptor leaks.
1581         * Scripts/webkitpy/common/checkout/commitinfo.py:
1582          - Remove unneeded import.
1583         * Scripts/webkitpy/common/checkout/scm.py:
1584          - Remove use of str().
1585         * Scripts/webkitpy/common/checkout/scm_unittest.py:
1586          - Read/write utf-8 files and use unicode() strings in testing.
1587         * Scripts/webkitpy/common/config/committers.py:
1588          - Use \u instead of \x to make slightly clearer what we're doing.
1589         * Scripts/webkitpy/common/net/bugzilla.py:
1590          - Add a new _string_contents() method and explain why
1591             we have to call unicode() on the result of soup.string
1592             and why it's safe to do so w/o needing to pass a codec name.
1593           - Remove the (unused) support for passing a file object to add_patch_to_bug().
1594         * Scripts/webkitpy/common/net/buildbot.py:
1595          - Use unicode() instead of str() when needing to coax a
1596             NavigableString object into a unicode() object.
1597         * Scripts/webkitpy/common/net/buildbot_unittest.py:
1598          - Add a test which contains a unicode builder name.
1599         * Scripts/webkitpy/common/net/statusserver.py:
1600          - Remove use of str()
1601         * Scripts/webkitpy/common/prettypatch.py:
1602          - Write out the patch file as utf-8.
1603         * Scripts/webkitpy/common/system/autoinstall.py:
1604          - Write out files with a explicit encodings.
1605          - Deploy "with" to close file descriptor leaks.
1606         * Scripts/webkitpy/common/system/deprecated_logging.py:
1607          - Write out log files as utf-8.
1608         * Scripts/webkitpy/common/system/executive.py:
1609          - Make run* properly take and return unicode() objects.
1610          - Cleaned up input handling in run_command a little by adding
1611            a _compute_input() method which can return early instead of having
1612            such a long/cluttered if-block.
1613         * Scripts/webkitpy/common/system/executive_unittest.py:
1614          - Added a unit test to make sure we don't break Tor again!
1615         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1616          - Write out the test list as utf-8.
1617         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1618          - Write out json files as utf-8.
1619         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
1620          - Deploy "with" to close file descriptor leaks.
1621         * Scripts/webkitpy/layout_tests/port/chromium.py: Add Executive.py FIXME.
1622         * Scripts/webkitpy/layout_tests/port/chromium_linux.py: ditto.
1623         * Scripts/webkitpy/layout_tests/port/chromium_mac.py: ditto.
1624         * Scripts/webkitpy/layout_tests/port/gtk.py: ditto.
1625         * Scripts/webkitpy/layout_tests/port/mac.py: ditto.
1626         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1627          - Make the skipped file parsing test unicode.
1628         * Scripts/webkitpy/layout_tests/port/qt.py: Add Executive.py FIXME.
1629         * Scripts/webkitpy/layout_tests/port/server_process.py: ditto.
1630         * Scripts/webkitpy/layout_tests/port/webkit.py:
1631          - Deploy "with" to close file descriptor leaks.
1632         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
1633          - Make explicit the encodings of log files and pid files.
1634         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
1635          - Make encodings explicit and deploy "with".
1636         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: ditto.
1637         * Scripts/webkitpy/layout_tests/test_types/image_diff.py: ditto.
1638         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py: ditto.
1639         * Scripts/webkitpy/layout_tests/test_types/text_diff.py: ditto.
1640         * Scripts/webkitpy/style/filereader_unittest.py: ditto.
1641         * Scripts/webkitpy/thirdparty/__init__.py: ditto.
1642         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
1643          - Removed extra import.
1644         * Scripts/webkitpy/tool/commands/queues.py:
1645          - No need to map args to strings now that run_command does.
1646          - Update test results to match args changes.
1647          - Document our global argument hacks.
1648         * Scripts/webkitpy/tool/commands/upload.py:
1649          - Pass the diff directly to add_patch_to_bug instead of creating a StringIO file wrapper.
1650         * Scripts/webkitpy/tool/mocktool.py:
1651          - Rename add_patch_to_bug argument to match bugzilla.py
1652         * Scripts/webkitpy/tool/steps/abstractstep.py:
1653          - Executive.run_* now require lists instead of strings.
1654            The lack of this change was what broke webkit-patch
1655            for svn users the first time this was landed.
1656         * Scripts/webkitpy/tool/steps/postdiff.py:
1657          - Pass the diff directly to add_patch_to_bug instead of creating a StringIO file wrapper.
1658         * Scripts/webkitpy/tool/steps/postdiffforcommit.py: ditto
1659         * Scripts/webkitpy/tool/steps/postdiffforrevert.py: ditto
1660         * Scripts/webkitpy/tool/steps/steps_unittest.py:
1661          - Fixed spurious logging seen when running test-webkitpy
1662
1663 2010-04-21  Kinuko Yasuda  <kinuko@chromium.org>
1664
1665         Unreviewed.
1666
1667         Add myself in committers.py.
1668
1669         * Scripts/webkitpy/common/config/committers.py:
1670
1671 2010-04-21  Dirk Pranke  <dpranke@chromium.org>
1672
1673         Reviewed by Dimitri Glazkov.
1674
1675         new-run-webkit-tests: fix a bug in the Chromium port where we would
1676         try to talk to a crashed test_shell and raise exceptions that weren't
1677         being caught.
1678
1679         https://bugs.webkit.org/show_bug.cgi?id=37941
1680
1681         * Scripts/webkitpy/layout_tests/port/chromium.py:
1682
1683 2010-04-21  Laszlo Gombos  <laszlo.1.gombos@nokia.com>
1684
1685         Unreviewed.
1686
1687         [Qt] [Symbian] Build fix.
1688
1689         Work around a Qt quirk. Some versions of Symbian port Qt
1690         QFontDatabase::removeAllApplicationFonts symbol is not available.
1691
1692         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1693         (WebCore::DumpRenderTree::open):
1694
1695 2010-04-21  Alexey Proskuryakov  <ap@apple.com>
1696
1697         Unreviewed.
1698
1699         https://bugs.webkit.org/show_bug.cgi?id=37933
1700         <rdar://problem/7719540> XMLHttpRequest.withCredentials should be better enforced.
1701
1702         Adding stub implementation of authenticateSession(). Depending on platform loader behavior,
1703         a real implementation may or may not be necessary for the one test that currently uses it
1704         to pass.
1705
1706         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1707         (LayoutTestController::authenticateSession):
1708         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1709
1710 2010-04-21  Alexey Proskuryakov  <ap@apple.com>
1711
1712         Reviewed by Brady Eidson.
1713
1714         https://bugs.webkit.org/show_bug.cgi?id=37933
1715         <rdar://problem/7719540> XMLHttpRequest.withCredentials should be better enforced.
1716
1717         Adding authenticateSession() method that adds credentials to per-process credential storage
1718         (for platforms that even have such). No Windows implementation, because writing another
1719         loader for DRT is painful.
1720
1721         * DumpRenderTree/LayoutTestController.cpp:
1722         (authenticateSessionCallback):
1723         (LayoutTestController::staticFunctions):
1724         * DumpRenderTree/LayoutTestController.h:
1725         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1726         (-[SynchronousLoader dealloc]):
1727         (-[SynchronousLoader connectionShouldUseCredentialStorage:]):
1728         (-[SynchronousLoader connection:didReceiveAuthenticationChallenge:]):
1729         (-[SynchronousLoader connection:didFailWithError:]):
1730         (-[SynchronousLoader connectionDidFinishLoading:]):
1731         (+[SynchronousLoader makeRequest:withUsername:password:]):
1732         (LayoutTestController::authenticateSession):
1733
1734         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1735         (LayoutTestController::authenticateSession):
1736         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
1737         (LayoutTestController::authenticateSession):
1738         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1739         (LayoutTestController::authenticateSession):
1740         Stub implementations.
1741
1742 2010-04-21  Kent Tamura  <tkent@chromium.org>
1743
1744         Reviewed by Dimitri Glazkov.
1745
1746         [DRT/Chromium] Make DRT compilable in Chromium tree
1747         https://bugs.webkit.org/show_bug.cgi?id=37923
1748
1749         We need to use different GYPs in a case of WebKit-only checkout
1750         and a case of whole Chromium checkout because the relative paths
1751         from webkit/ to WebKit/chromium/features.gypi are different in
1752         these cases and we can't use 'conditions' for 'includes' in GYPs.
1753
1754         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
1755
1756 2010-04-21  Jakub Wieczorek  <jwieczorek@webkit.org>
1757
1758         Reviewed by Darin Adler.
1759
1760         List item markers are not always updated after changes in the DOM.
1761         https://bugs.webkit.org/show_bug.cgi?id=37060
1762
1763         * DumpRenderTree/LayoutTestController.cpp:
1764         (markerTextForListItemCallback): A function that returns the marker text for a given list item.
1765         (LayoutTestController::staticFunctions):
1766         * DumpRenderTree/LayoutTestController.h:
1767         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
1768         (LayoutTestController::markerTextForListItem): Implement it in the GTK port.
1769         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
1770         (LayoutTestController::markerTextForListItem): Add a stub.
1771         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1772         (LayoutTestController::LayoutTestController):
1773         (LayoutTestController::markerTextForListItem): Implement it in the Qt port.
1774         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1775         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
1776         (LayoutTestController::markerTextForListItem): Add a stub.
1777
1778 2010-04-21  Adam Roben  <aroben@apple.com>
1779
1780         Exclude leaks in CGGradientCreateWithColorComponents from
1781         run-webkit-tests leaks output
1782
1783         The leak in CG is covered by <rdar://problem/7888492>.
1784
1785         Fixes <http://webkit.org/b/37927>.
1786
1787         Reviewed by Eric Carlson.
1788
1789         * Scripts/old-run-webkit-tests:
1790         (sub countAndPrintLeaks): Exclude leaks in
1791         CGGradientCreateWithColorComponents on certain OSs.
1792
1793 2010-04-21  Kent Tamura  <tkent@chromium.org>
1794
1795         Reviewed by Shinichiro Hamaji.
1796
1797         [DRT/Chromium] Import MockSpellCheck from Chromium
1798         https://bugs.webkit.org/show_bug.cgi?id=37910
1799
1800         Import webkit/tools/test_shell/mock_spellcheck.{cc,h} rev.37241 of Chromium.
1801
1802         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
1803           Add ICU explicitly because WTFString.h includes ICU headers.
1804           Add MockSpellCheck.{cpp,h}.
1805         * DumpRenderTree/chromium/MockSpellCheck.cpp: Added.
1806         * DumpRenderTree/chromium/MockSpellCheck.h: Added.
1807         * DumpRenderTree/chromium/WebViewHost.cpp:
1808         (WebViewHost::spellCheck):
1809         * DumpRenderTree/chromium/WebViewHost.h:
1810         * DumpRenderTree/chromium/config.h: Define JS_EXPORTDATA, which is used
1811           by wtf/text/AtomicString.h included by wtf/text/WTFString.h.
1812
1813 2010-04-21  Adam Barth  <abarth@webkit.org>
1814
1815         Reviewed by Eric Seidel.
1816
1817         Add webkit-patch pretty-diff
1818         https://bugs.webkit.org/show_bug.cgi?id=37892
1819
1820         This is slightly lame because it asks you whether the diff is correct,
1821         but it's a starting point.
1822
1823         * Scripts/webkitpy/tool/commands/__init__.py:
1824         * Scripts/webkitpy/tool/commands/prettydiff.py: Added.
1825         * Scripts/webkitpy/tool/main.py:
1826
1827 2010-04-21  Adam Barth  <abarth@webkit.org>
1828
1829         Reviewed by Darin Adler.
1830
1831         Remove mention of non-existant --no-build option
1832         https://bugs.webkit.org/show_bug.cgi?id=37893
1833
1834         The option doesn't exist!
1835
1836         * Scripts/webkitpy/tool/commands/stepsequence.py:
1837
1838 2010-04-21  Balazs Kelemen  <kb@inf.u-szeged.hu>
1839
1840         Reviewed by Kenneth Rohde Christiansen.
1841
1842         [Qt] Interrupting JavaScript is cumbersome when you use QtLaucher for testing or profiling.
1843         https://bugs.webkit.org/show_bug.cgi?id=37198
1844
1845         * QtLauncher/main.cpp:
1846         (LauncherWindow::toggleInterruptingJavaScriptEnabled):
1847         (LauncherWindow::newWindow):
1848         (LauncherWindow::cloneWindow):
1849         (LauncherWindow::createChrome):
1850         (main):
1851         * QtLauncher/webpage.cpp:
1852         (WebPage::WebPage):
1853         (WebPage::shouldInterruptJavaScript):
1854         * QtLauncher/webpage.h:
1855         (WebPage::setInterruptingJavaScriptEnabled):
1856
1857 2010-04-21  Eric Seidel  <eric@webkit.org>
1858
1859         Unreviewed, rolling out r57963.
1860         http://trac.webkit.org/changeset/57963
1861         https://bugs.webkit.org/show_bug.cgi?id=37759
1862
1863         Three tests started crashing on the Qt bot.
1864
1865         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1866         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1867
1868 2010-04-21  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
1869
1870         Reviewed by Eric Seidel.
1871
1872         Make new-run-webkit-tests work for the Qt port
1873         https://bugs.webkit.org/show_bug.cgi?id=37588
1874
1875         * Scripts/webkitpy/layout_tests/port/qt.py:
1876
1877 2010-04-21  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
1878
1879         Reviewed by Eric Seidel.
1880
1881         new-run-webkit-tests: try to detect alternate apache path
1882         https://bugs.webkit.org/show_bug.cgi?id=37587
1883
1884         _check_port_build() also needs to return true in the
1885         base implementation to not fail the check_build step.
1886
1887         * Scripts/webkitpy/layout_tests/port/webkit.py:
1888
1889 2010-04-21  Yi Shen  <yi.4.shen@nokia.com>
1890
1891         Reviewed by Simon Hausmann.
1892
1893         [Qt] Add LayoutTestController interface: computedStyleIncludingVisitedInfo
1894         https://bugs.webkit.org/show_bug.cgi?id=37759
1895
1896         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
1897         (LayoutTestController::computedStyleIncludingVisitedInfo):
1898         * DumpRenderTree/qt/LayoutTestControllerQt.h:
1899
1900 2010-04-21  Eric Seidel  <eric@webkit.org>
1901
1902         Unreviewed, test fix only.
1903
1904         new-run-webkit-tests: implement a --log trace message to be able to display detailed output of an individual test run
1905         https://bugs.webkit.org/show_bug.cgi?id=37726
1906
1907         This change seems to have broken a test.
1908         Attempting to handle the case where we don't have any
1909         timing information.  Dirk may have to correct this change.
1910
1911         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1912
1913 2010-04-19  Eric Seidel  <eric@webkit.org>
1914
1915         Reviewed by Adam Barth.
1916
1917         new-run-webkit-tests has much higher startup latency than run-webkit-tests
1918         https://bugs.webkit.org/show_bug.cgi?id=37643
1919
1920         I got rid of the -expected.checksum reads during startup.
1921         This makes startup noticably better on my laptop.
1922
1923         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1924          - Use image_hash() instead of .image_hash now that expected.checksum
1925            file reads are done lazily.
1926         * Scripts/webkitpy/layout_tests/port/http_server_base.py:
1927          - Add debug logging for this sleep call.
1928            In my testing I never saw this sleep() hit.
1929         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
1930          - Sleep a shorter interval to make websocket server
1931            startup more responsive.  On my machine startup was
1932            taking around 1 second.
1933          - Remove the unconditional .5s delay on startup.
1934         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1935          - Make image_hash file reads done lazily in a new image_hash() function.
1936          - Add a "Starting testing ..." meter update after DRT threads have
1937            been started, but before we get updates from the first one.
1938          - Rename variable "t" to a full english name to match WebKit style.
1939
1940 2010-04-20  Daniel Bates  <dbates@rim.com>
1941
1942         Reviewed by Eric Seidel.
1943
1944         https://bugs.webkit.org/show_bug.cgi?id=37748
1945
1946         Make Sheriffbot more inspirational.
1947
1948         * Scripts/webkitpy/common/net/bugzilla.py:
1949         * Scripts/webkitpy/common/net/bugzilla_unittest.py:
1950         * Scripts/webkitpy/tool/bot/irc_command.py:
1951         * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:
1952         * Scripts/webkitpy/tool/mocktool.py:
1953
1954 2010-04-20  Kent Tamura  <tkent@chromium.org>
1955
1956         Reviewed by Dimitri Glazkov.
1957
1958         [DRT/Chromium] Import Chromium image_diff as ImageDiff
1959         https://bugs.webkit.org/show_bug.cgi?id=37790
1960
1961         ImageDiff.cpp is based on tools/imagediff/image_diff.cc r41911 of Chromium.
1962
1963         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
1964         * DumpRenderTree/chromium/ImageDiff.cpp: Added.
1965         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1966
1967 2010-04-20  Dirk Pranke <dpranke@chromium.org>
1968
1969         This patch to new-run-webkit-tests adds a --log 'trace' option
1970         that prints out detailed info about a given test as it executes
1971         (where the baselines are, what the expectation is, what we got,
1972         how long it took).
1973
1974         https://bugs.webkit.org/show_bug.cgi?id=37726
1975
1976         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1977           - use the newly exposed TestResult class and implement
1978             --log trace
1979         * Scripts/webkitpy/layout_tests/layout_package/dump_render_thread.py:
1980           - rename TestStats to TestResult and make it more public, resulting
1981             in cleaner code
1982         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
1983           - add expectation_to_string() as a separate callable function
1984
1985 2010-04-20  Eric Seidel  <eric@webkit.org>
1986
1987         Unreviewed, rolling out r57907.
1988         http://trac.webkit.org/changeset/57907
1989         https://bugs.webkit.org/show_bug.cgi?id=37765
1990
1991         Appears to have broken MacEWS and possibly webkit-patch upload
1992         for svn users.  Needs further investigation.
1993
1994         * Scripts/webkitpy/common/checkout/api.py:
1995         * Scripts/webkitpy/common/checkout/api_unittest.py:
1996         * Scripts/webkitpy/common/checkout/changelog.py:
1997         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
1998         * Scripts/webkitpy/common/checkout/commitinfo.py:
1999         * Scripts/webkitpy/common/checkout/scm.py:
2000         * Scripts/webkitpy/common/checkout/scm_unittest.py:
2001         * Scripts/webkitpy/common/config/committers.py:
2002         * Scripts/webkitpy/common/net/bugzilla.py:
2003         * Scripts/webkitpy/common/net/buildbot.py:
2004         * Scripts/webkitpy/common/net/buildbot_unittest.py:
2005         * Scripts/webkitpy/common/net/statusserver.py:
2006         * Scripts/webkitpy/common/prettypatch.py:
2007         * Scripts/webkitpy/common/system/autoinstall.py:
2008         * Scripts/webkitpy/common/system/deprecated_logging.py:
2009         * Scripts/webkitpy/common/system/executive.py:
2010         * Scripts/webkitpy/common/system/executive_unittest.py:
2011         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2012         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
2013         * Scripts/webkitpy/layout_tests/layout_package/metered_stream.py:
2014         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
2015         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
2016         * Scripts/webkitpy/tool/commands/queues.py:
2017         * Scripts/webkitpy/tool/commands/queues_unittest.py:
2018         * Scripts/webkitpy/tool/commands/upload.py:
2019         * Scripts/webkitpy/tool/mocktool.py:
2020         * Scripts/webkitpy/tool/steps/abstractstep.py:
2021         * Scripts/webkitpy/tool/steps/postdiff.py:
2022         * Scripts/webkitpy/tool/steps/postdiffforcommit.py:
2023         * Scripts/webkitpy/tool/steps/postdiffforrevert.py:
2024         * Scripts/webkitpy/tool/steps/steps_unittest.py:
2025
2026 2010-04-20  Nate Chapin  <japhet@chromium.org>
2027
2028         Unreviewed.
2029
2030         Update my irc handle in committers.py (natechapin -> japhet).
2031
2032         * Scripts/webkitpy/common/config/committers.py:
2033
2034 2010-04-20  Eric Seidel  <eric@webkit.org>
2035
2036         Reviewed by Adam Barth.
2037
2038         REGRESSION(57531): the commit-queue still hates Tor Arne Vestbø
2039         https://bugs.webkit.org/show_bug.cgi?id=37765
2040
2041         I fixed the queue to not ignore Tor as a reviwer in r57531,
2042         but instead it throws an exception every time his name is in a patch.
2043
2044         This fixes our Executive.run_command code to work around a Popen
2045         bug http://bugs.python.org/issue5290 whereby python versions before 2.6
2046         do not correctly handle unicode objects as input or output to
2047         Popen.communicate.
2048
2049         Following the advice of:
2050         http://farmdev.com/talks/unicode/
2051         I'm attempting to take the python unicode plunge and use unicode()
2052         objects as strings instead of str() objects everywhere in webkitpy.
2053
2054         We do not have to use u"" instead of "" because u"a" == "a" as expected
2055         in Python.  Python will generate a warning to the console in cases where
2056         a unicode() == str() operation cannot be performed.
2057
2058         I also cleaned up the input handling in run_command a little by adding
2059         a new _compute_input() method which can return early instead of having
2060         such a long/cluttered if-block.
2061
2062         Executive.run* now correctly accept and return unicode() objects.
2063         I attempted to fix all the places that we call .write() to make sure we
2064         encode any unicode() objects into utf-8.
2065
2066         All places which use StringIO need to be sure to pass StringIO a
2067         pre-encoded byte-array (str object) instead of unicode so that
2068         clients which read from the StringIO don't have encoding exceptions.
2069         To make this easier, I removed the patch_file_object support from
2070         add_patch_to_bug, and changed the 4 places which previously used
2071         StringIO to create a fake patch file.
2072
2073         I attempted to document any places where we are not correctly converting
2074         to/from bytes (str() objects) to strings (unicode() objects).
2075
2076         * Scripts/webkitpy/common/checkout/api_unittest.py:
2077          - Read/write utf-8 files instead of ascii.
2078          - Update the tests to use test for proper unicode() handling.
2079         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
2080          - Use unicode() strings instead of str() byte arrays.
2081         * Scripts/webkitpy/common/checkout/scm.py:
2082          - Remove use of str().
2083         * Scripts/webkitpy/common/checkout/scm_unittest.py:
2084          - Read/write utf-8 files and use unicode() strings in testing.
2085         * Scripts/webkitpy/common/config/committers.py:
2086          - Use \u instead of \x to make slightly clearer what we're doing.
2087         * Scripts/webkitpy/common/net/bugzilla.py:
2088          - Add a new _string_contents() method and explain why
2089            we have to call unicode() on the result of soup.string
2090            and why it's safe to do so w/o needing to pass a codec name.
2091          - Remove the (unused) support for passing a file object to add_patch_to_bug().
2092         * Scripts/webkitpy/common/net/buildbot.py:
2093          - Use unicode() instead of str() when needing to coax a
2094            NavigableString object into a unicode() object.
2095         * Scripts/webkitpy/common/net/statusserver.py:
2096          - Remove use of str()
2097         * Scripts/webkitpy/common/prettypatch.py:
2098          - Write out the patch file as utf-8.
2099         * Scripts/webkitpy/common/system/autoinstall.py:
2100          - Add a FIXME about encoding.
2101         * Scripts/webkitpy/common/system/deprecated_logging.py:
2102          - Document that tee() works on bytes, not strings.
2103         * Scripts/webkitpy/common/system/executive.py:
2104          - Make run* properly take and return unicode() objects.
2105         * Scripts/webkitpy/common/system/executive_unittest.py:
2106          - Added a unit test to make sure we don't break Tor again!
2107         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2108          - Write out the test list as utf-8.
2109         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
2110          - Write out json files as utf-8.
2111         * Scripts/webkitpy/layout_tests/layout_package/metered_stream.py:
2112          - Add FIXME about encoding handling.
2113         * Scripts/webkitpy/tool/commands/upload.py:
2114          - Pass the diff directly to add_patch_to_bug instead of creating a StringIO file wrapper.
2115         * Scripts/webkitpy/tool/mocktool.py:
2116          - Rename add_patch_to_bug argument to match bugzilla.py
2117         * Scripts/webkitpy/tool/steps/postdiff.py:
2118          - Pass the diff directly to add_patch_to_bug instead of creating a StringIO file wrapper.
2119         * Scripts/webkitpy/tool/steps/postdiffforcommit.py: ditto.
2120         * Scripts/webkitpy/tool/steps/postdiffforrevert.py: ditto.
2121         * Scripts/webkitpy/tool/steps/steps_unittest.py:
2122          - Fixed spurious logging seen when running test-webkitpy
2123
2124 2010-04-20  Chris Jerdonek  <cjerdonek@webkit.org>
2125
2126         Reviewed by Shinichiro Hamaji.
2127
2128         For check-webkit-style, implemented __eq__() and __ne__() (the
2129         built-in equality and inequality methods) for the
2130         DefaultStyleErrorHandler class.
2131
2132         https://bugs.webkit.org/show_bug.cgi?id=37850
2133
2134         This will facilitate unit-testing for a subsequent patch,
2135         namely for https://bugs.webkit.org/show_bug.cgi?id=37850
2136
2137         * Scripts/webkitpy/style/error_handlers.py:
2138           - Added __eq__() and __ne__() to the DefaultStyleErrorHandler
2139             class.
2140
2141         * Scripts/webkitpy/style/error_handlers_unittest.py:
2142           - Added unit tests for __eq__() and __ne__().
2143           - Also included a minor clean-up refactoring of combining the
2144             StyleErrorHandlerTestBase class (which has not needed to
2145             be separate due to previous changes) into the
2146             DefaultStyleErrorHandlerTest class.
2147
2148 2010-04-20  Jakub Wieczorek  <jwieczorek@webkit.org>
2149
2150         Unreviewed.
2151
2152         Add my IRC nick to the committers.py list.
2153
2154         * Scripts/webkitpy/common/config/committers.py:
2155
2156 2010-04-20  Kim Grönholm  <kim.gronholm@nomovok.com>
2157
2158         Reviewed by Simon Hausmann.
2159
2160         [Qt] Multitouch mocking in QtLauncher doesn't work with QGraphicsView
2161         https://bugs.webkit.org/show_bug.cgi?id=37816
2162
2163         Fix multi-touch mocking in QtLauncher when using QGraphicsView.
2164         Test: https://bug-32434-attachments.webkit.org/attachment.cgi?id=44955
2165
2166         * QtLauncher/main.cpp:
2167         (LauncherWindow::eventFilter):
2168         (LauncherWindow::initializeView):
2169
2170 2010-04-20  MORITA Hajime <morrita@google.com>
2171
2172         Unreviewed, add myself to the committers list.
2173
2174         * Scripts/webkitpy/common/config/committers.py:
2175
2176 2010-04-20  Kent Tamura  <tkent@chromium.org>
2177
2178         Reviewed by Shinichiro Hamaji.
2179
2180         [DRT/Chromium] Fix some unexpected results of editing
2181         https://bugs.webkit.org/show_bug.cgi?id=37843
2182
2183         This change fixes about 70 unexpected results.
2184         The original test_webview_delegate.cc doesn't have this bug.
2185         The bug was introduced when I ported it to WebKit tree.
2186
2187         * DumpRenderTree/chromium/WebViewHost.cpp:
2188         (printRangeDescription): Replace the latter startContainer() with endContainer().
2189
2190 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2191
2192         Unreviewed, build fix.
2193
2194         Turn off some unit tests for now - the new-run-webkit-tests dryrun
2195         tests for chromium won't work if you don't have a chromium checkout.
2196
2197         https://bugs.webkit.org/show_bug.cgi?id=37841
2198
2199         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2200
2201 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2202
2203         Unreviewed, build fix.
2204
2205         (Re-)add dryrun.py; this was renamed from passing.py in the previous
2206         CL but apparently somehow didn't get checked in.
2207
2208         https://bugs.webkit.org/show_bug.cgi?id=37841
2209
2210         * Scripts/webkitpy/layout_tests/port/dryrun.py: Added.
2211
2212 2010-04-19  Shinichiro Hamaji  <hamaji@chromium.org>
2213
2214         Reviewed by David Levin.
2215
2216         check-webkit-style: exits when encountering a deleted file
2217         https://bugs.webkit.org/show_bug.cgi?id=37122
2218
2219         This reverts the quick fix done by r57119 and makes check_patch
2220         not call check_file for deleted files.
2221
2222         Also this change fixes the behavior for "-", which should mean
2223         stdin.  Before this change, the style checker just ignored "-"
2224         with a warning message.
2225
2226         * Scripts/webkitpy/style/checker.py:
2227         * Scripts/webkitpy/style/checker_unittest.py:
2228
2229 2010-04-19  Daniel Bates  <dbates@rim.com>
2230
2231         No review, rolling out 57868.
2232         http://trac.webkit.org/changeset/57868
2233         https://bugs.webkit.org/show_bug.cgi?id=37748
2234
2235         Sheriffbot wasn't very inspirational after this patch.
2236         Instead, he was silent when you said hi :-(. Rolling
2237         out this patch so that I can debug/test this some more.
2238
2239         * Scripts/webkitpy/common/net/bugzilla.py:
2240         * Scripts/webkitpy/common/net/bugzilla_unittest.py:
2241         * Scripts/webkitpy/tool/bot/irc_command.py:
2242         * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:
2243         * Scripts/webkitpy/tool/mocktool.py:
2244
2245 2010-04-19  Daniel Bates  <dbates@rim.com>
2246
2247         Reviewed by Adam Barth.
2248
2249         https://bugs.webkit.org/show_bug.cgi?id=37748
2250
2251         Make Sheriffbot more inspirational.
2252
2253         * Scripts/webkitpy/common/net/bugzilla.py:
2254         * Scripts/webkitpy/common/net/bugzilla_unittest.py:
2255         * Scripts/webkitpy/tool/bot/irc_command.py:
2256         * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:
2257         * Scripts/webkitpy/tool/mocktool.py:
2258
2259 2010-04-19  Kevin Ollivier  <kevino@theolliviers.com>
2260
2261         [wx] Build fix, add missing header.
2262
2263         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2264
2265 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2266
2267          Reviewed by Adam Barth.
2268
2269          new-run-webkit-tests - repurpose the "Passing" port as "Dryrun" port
2270          that can be used to test platforms other than the one you are running
2271          on. This can be useful for checking baselines and testing code
2272          coverage.
2273
2274          Note that running the code on the "wrong" port requires each
2275          port-specific implementation to actually not require any
2276          platform-specific python code (e.g., the chromium-win port must
2277          test for the existence of windows functions before calling them).
2278
2279          https://bugs.webkit.org/show_bug.cgi?id=37782
2280
2281          * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2282          * Scripts/webkitpy/layout_tests/port/dryrun.py: Renamed from WebKitTools/Scripts/webkitpy/layout_tests/port/passing.py.
2283          * Scripts/webkitpy/layout_tests/port/factory.py:
2284          * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
2285
2286 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2287
2288         Reviewed by Eric Seidel.
2289
2290         new-run-webkit-tests: add --build (default) and --no-build
2291         options to make that step optional. This flag modifies what happens
2292         in port.check_build().
2293
2294         https://bugs.webkit.org/show_bug.cgi?id=37786
2295
2296         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2297         * Scripts/webkitpy/layout_tests/port/chromium.py:
2298         * Scripts/webkitpy/layout_tests/port/webkit.py:
2299
2300 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2301
2302         Unreviewed, build fix.
2303
2304         new-run-webkit-tests - fix a typo in r57480 that caused us to stop
2305         logging the actual list of unexpected results.
2306
2307         https://bugs.webkit.org/show_bug.cgi?id=37831
2308
2309         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2310
2311 2010-04-19  Dirk Pranke <dpranke@chromium.org>
2312
2313         Reviewed by Adam Barth.
2314
2315         From a patch by Tor Arne Vestbo  <tor.arne.vestbo@nokia.com>
2316
2317         new-run-webkit-tests: make the retry step more explicit
2318         https://bugs.webkit.org/show_bug.cgi?id=37606
2319
2320         It might be confusing to see the test and percent counters
2321         reset without any notice of what's going on, so we make the
2322         message that a retry-run is started explicit.
2323
2324         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2325
2326 2010-04-19  Sam Weinig  <weinig@apple.com>
2327
2328         Reviewed by Anders Carlsson.
2329
2330         Add support for opening a new window (File->New Window) to
2331         Windows MiniBrowser. Accelerator doesn't work.
2332
2333         * MiniBrowser/win/BrowserView.cpp:
2334         (createNewPage): Use BrowserView::create.
2335         * MiniBrowser/win/BrowserWindow.cpp:
2336         (BrowserWindow::onCommand): Respond to ID_FILE_NEW_WINDOW
2337         by creating a new window.
2338         * MiniBrowser/win/BrowserWindow.h:
2339         (BrowserWindow::create): Added. Don't allow creating
2340         BrowserWindows on the stack by making constructor
2341         private and exposing the create function.
2342         * MiniBrowser/win/MiniBrowser.cpp:
2343         (MiniBrowser::createNewWindow): Move new window creation
2344         logic here.
2345         * MiniBrowser/win/MiniBrowser.h:
2346         * MiniBrowser/win/MiniBrowser.rc:
2347         * MiniBrowser/win/main.cpp:
2348         (_tWinMain): Use the new MiniBrowser::createNewWindow().
2349
2350 2010-04-19  Chris Fleizach  <cfleizach@apple.com>
2351
2352         Reviewed by Beth Dakin.
2353
2354         AX: aria-haspopup needs to be exposed
2355         https://bugs.webkit.org/show_bug.cgi?id=37808
2356
2357         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
2358         (AccessibilityUIElement::hasPopup):
2359
2360 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2361
2362         Reviewed by Eric Seidel.
2363
2364         Submit a better workaround for r57806 than the one in r57831 - log
2365         an error and exit if you try to run new-run-webkit-tests with --use-drt
2366         on Windows.
2367
2368         https://bugs.webkit.org/show_bug.cgi?id=37822
2369
2370         * Scripts/webkitpy/layout_tests/port/chromium.py:
2371
2372 2010-04-19  Jesus Sanchez-Palencia  <jesus@webkit.org>
2373
2374         Unreviewed.
2375
2376         Just adding myself as a committer.
2377
2378         * Scripts/webkitpy/common/config/committers.py:
2379
2380 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2381
2382         Reviewed by Eric Seidel.
2383
2384         Add slightly better logging to the websocket python wrapper script,
2385         including a --verbose flag for debug output.
2386
2387         https://bugs.webkit.org/show_bug.cgi?id=37233
2388
2389         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
2390         * Scripts/new-run-webkit-websocketserver:
2391
2392 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2393
2394         Reviewed by Eric Seidel.
2395
2396         new-run-webkit-tests: add a way (--print-unexpected-results) to
2397         (re-)print out the tests that produced unexpected results in the
2398         last run. Also add a way (--retry-unexpected-results) to
2399         automatically re-test them.
2400
2401         https://bugs.webkit.org/show_bug.cgi?id=37783
2402
2403         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2404
2405 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2406
2407         Reviewed by Eric Seidel.
2408
2409         Restructure the logging in new-run-webkit-tests so that many of log
2410         messages that were logged to the MeteredStream also get logged in
2411         --verbose mode.
2412
2413         https://bugs.webkit.org/show_bug.cgi?id=37780
2414
2415         * Scripts/webkitpy/layout_tests/layout_package/metered_stream.py:
2416         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2417
2418 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2419
2420         Reviewed by Eric Seidel.
2421
2422         new-run-webkit-tests: add 'summary' and 'unexpected-results' options
2423         to the --log flag. Also add a 'progress' flag to enable the regular
2424         progress bar (as opposed to 'detailed-progress', which enables the
2425         dots). Also add a 'nothing' flag to allow you to be explicit that
2426         you don't want any logging on the command line.
2427
2428         The default is
2429         '--log detailed-progress,summary,unexpected,unexpected-results'
2430
2431         (The default logging is unchanged by this patch, this just makes things
2432         properly configurable).
2433
2434         Note that 'nothing' doesn't work properly yet; I need a couple other
2435         patches to land to avoid rewriting things several different ways.
2436
2437         https://bugs.webkit.org/show_bug.cgi?id=37785
2438
2439         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2440
2441 2010-04-19  Kenneth Rohde Christiansen  <kenneth@webkit.org>
2442
2443         Reviewed by Simon Hausmann.
2444
2445         Make the URL change on committed load.
2446
2447         * QtLauncher/mainwindow.cpp:
2448         (MainWindow::buildUI):
2449         (MainWindow::setAddressUrl):
2450         * QtLauncher/mainwindow.h:
2451
2452 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2453
2454         Reviewed by Eric Seidel.
2455
2456         Fix a typo in the rebaselining tool that causes us to use "debug"
2457         instead of "Debug" in a directory path, which fails on platforms with
2458         case-sensitive filesystems.
2459
2460         https://bugs.webkit.org/show_bug.cgi?id=37819
2461
2462         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2463
2464 2010-04-19  Dirk Pranke  <dpranke@chromium.org>
2465
2466         Unreviewed, build fix.
2467
2468         Fix breakage of chromium-win canary bots caused by r57806. That patch
2469         introduced the option of using Chrome's new port of DumpRenderTree,
2470         but unfortunately that port relies on the webkit.py class
2471         implementation which uses non-blocking I/O that isn't available on
2472         Windows. This patch turns off that option and doesn't import the
2473         class if we're running on Windows.
2474
2475         https://bugs.webkit.org/show_bug.cgi?id=37817
2476
2477         * Scripts/webkitpy/layout_tests/port/chromium.py:
2478
2479 2010-04-19  James Robinson  <jamesr@chromium.org>
2480
2481         Reviewed by abarth.
2482
2483         Fix a typo
2484
2485         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
2486
2487 2010-04-19  Adam Roben  <aroben@apple.com>
2488
2489         Fix run-webkit-tests when there are spaces in the path
2490
2491         Fixes <http://webkit.org/b/37809>
2492
2493         Reviewed by Adam Barth.
2494
2495         * Scripts/run-webkit-tests: Use an "indirect object" to specify the
2496         path to the harness to exec(). According to perldoc, this usage will
2497         prohibit perl from parsing the arguments to exec() via the shell,
2498         which would incorrectly split paths with spaces in them, etc.
2499
2500 2010-04-19  Dimitri Glazkov  <dglazkov@chromium.org>
2501
2502         Reviewed by Adam Roben.
2503
2504         WinLauncher.h should use LF line-endings and use native line-endings style.
2505         https://bugs.webkit.org/show_bug.cgi?id=37807
2506
2507         * WinLauncher/WinLauncher.h: Added property svn:eol-style, converted to LF line-endings.
2508
2509 2010-04-19  Kent Tamura  <tkent@chromium.org>
2510
2511         Reviewed by Shinichiro Hamaji.
2512
2513         [DRT/Chromium] Fix a test initialization problem
2514         https://bugs.webkit.org/show_bug.cgi?id=37791
2515
2516         * DumpRenderTree/chromium/DumpRenderTree.cpp:
2517         (runTest): Call resetTestController() before runFileTest(). Some
2518           controllers initialize their fields in reset() and not in their
2519           constructors.
2520
2521 2010-04-19  Kent Tamura  <tkent@chromium.org>
2522
2523         Reviewed by Eric Seidel.
2524
2525         [Chromium] new-run-webkit-tests should use WebKitDriver for --use-drt
2526         https://bugs.webkit.org/show_bug.cgi?id=37793
2527
2528         We need to use WebKitDriver instead of ChromiumDriver for Chromium
2529         DRT because its interface is different from test_shell.
2530
2531         Chromium DRT has no UI.  So we can't use it to show test results.
2532
2533         * Scripts/webkitpy/layout_tests/port/chromium.py:
2534
2535 2010-04-18  Eric Seidel  <eric@webkit.org>
2536
2537         Reviewed by Adam Roben.
2538
2539         Make failure-reason more forgiving
2540         https://bugs.webkit.org/show_bug.cgi?id=37525
2541
2542         Removed search_limit, which wasn't very useful anyway.
2543         Added a log about the long load time loading from the builders.
2544         Prompt the user for what revision to start walking from (makes it easy to restart upon failure).
2545
2546         * Scripts/webkitpy/common/net/buildbot.py:
2547         * Scripts/webkitpy/tool/commands/queries.py:
2548
2549 2010-04-18  Chris Jerdonek  <cjerdonek@webkit.org>
2550
2551         Reviewed by Shinichiro Hamaji.
2552
2553         Created a class for check-webkit-style that encapsulates iterating
2554         over text files and reading them.
2555
2556         https://bugs.webkit.org/show_bug.cgi?id=37754
2557
2558         This revision is an intermediate step towards separating our
2559         style-checking code from the logic of iterating over files and
2560         reading them.
2561
2562         * Scripts/webkitpy/common/system/logtesting.py:
2563           - Added a logMessages() method to the LoggingTestCase class.
2564             This method provides unit tests with access to the raw list
2565             of log messages in case the tester needs to do something more
2566             than simply assert the list of existing messages.
2567
2568         * Scripts/webkitpy/style/checker.py:
2569           - Added a ProcessorBase class that processors of lists of lines
2570             should eventually inherit from.
2571           - Also added a FIXME to use the ProcessorBase class and the
2572             TextFileReader class added below.
2573
2574         * Scripts/webkitpy/style/filereader.py: Added.
2575           - Created a TextFileReader class that encapsulates reading
2576             and iterating over text files.
2577
2578         * Scripts/webkitpy/style/filereader_unittest.py: Added.
2579           - Added a TextFileReaderTest class to unit-test the
2580             new TextFileReader class.
2581
2582 2010-04-15  Tony Chang  <tony@chromium.org>
2583
2584         Reviewed by Adam Barth.
2585
2586         [chromium] new-run-webkit-tests should be able to use chromium DRT
2587         https://bugs.webkit.org/show_bug.cgi?id=37645
2588
2589         Make sure that the lack of a chromium checkout doesn't cause the script to
2590         fail.
2591
2592         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
2593         * Scripts/webkitpy/layout_tests/port/chromium.py:
2594         * Scripts/webkitpy/layout_tests/port/chromium_mac.py: Fix up a few paths
2595             to be relative to an upstream output dir.
2596         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: Add --use-drt flag.
2597
2598 2010-04-18  Eric Seidel  <eric@webkit.org>
2599
2600         Unreviewed, fixing the Qt bot.
2601
2602         Add a layer of indirection when calling run-webkit-tests to allow testing new-run-webkit-tests on various platforms
2603         https://bugs.webkit.org/show_bug.cgi?id=37632
2604
2605         * Scripts/run-webkit-tests:
2606          - Exit non-zero of launching the harness fails.
2607
2608 2010-04-18  Eric Seidel  <eric@webkit.org>
2609
2610         Unreviewed, fixing the Qt bot.
2611
2612         Add a layer of indirection when calling run-webkit-tests to allow testing new-run-webkit-tests on various platforms
2613         https://bugs.webkit.org/show_bug.cgi?id=37632
2614
2615         * Scripts/run-webkit-tests:
2616          - Fix the wrapper to work for users who do not
2617            have WebKitTools/Scripts in their path.
2618
2619 2010-04-14  Eric Seidel  <eric@webkit.org>
2620
2621         Reviewed by Adam Barth.
2622
2623         Add a layer of indirection when calling run-webkit-tests to
2624         allow testing new-run-webkit-tests on various platforms.
2625         https://bugs.webkit.org/show_bug.cgi?id=37632
2626
2627         This will let us test and fix bugs in new-run-webkit-tests
2628         without needing to restart the buildbot master between tests.
2629
2630         Currently this change leaves run-webkit-tests as-is, but once
2631         its landed we will easily be able to turn on/off
2632         new-run-webkit-tests for various ports/configurations.
2633
2634         I will send a note out to webkit-dev about how we will
2635         be using this launcher script to test on the bots.
2636
2637         * Scripts/old-run-webkit-tests: Copied from WebKitTools/Scripts/run-webkit-tests.
2638         * Scripts/run-webkit-tests:
2639          - A new script which decides whether to run new- or old-
2640            run-webkit-tests based on the platform.
2641         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2642          - Add some dummy argument handling for arguments which
2643            old-run-webkit-tests supports but new-run-webkit-tests
2644            does not yet.
2645
2646 2010-04-18  Eric Seidel  <eric@webkit.org>
2647
2648         Reviewed by Adam Barth.
2649
2650         Add Gtk bots to the list of "core builders" (builders which stop the commit-queue when they turn red)
2651         https://bugs.webkit.org/show_bug.cgi?id=33295
2652
2653         The Gtk builders have been green every time I've looked
2654         at them in the last 5 days or so.  It would appear webkit
2655         is now keeping them green and we should update the core
2656         builder list to match reality.
2657
2658         * Scripts/webkitpy/common/net/buildbot.py:
2659         * Scripts/webkitpy/common/net/buildbot_unittest.py:
2660
2661 2010-04-18  Robert Hogan  <robert@webkit.org>
2662
2663         Reviewed by Simon Hausmann.
2664
2665         [Qt] Add support for LayoutTestController commands:
2666                    setSmartInsertDeleteEnabled
2667                    setSelectTrailingWhitespaceEnabled
2668                    execCommand
2669                    isCommandEnabled
2670
2671         https://bugs.webkit.org/show_bug.cgi?id=35844
2672
2673         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
2674         (WebCore::WebPage::resetSettings):
2675         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2676         (LayoutTestController::setSmartInsertDeleteEnabled):
2677         (LayoutTestController::setSelectTrailingWhitespaceEnabled):
2678         (LayoutTestController::execCommand):
2679         (LayoutTestController::isCommandEnabled):
2680         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2681
2682 2010-04-17  Adam Barth  <abarth@webkit.org>
2683
2684         Reviewed by Eric Seidel.
2685
2686         Remove steps_references and commands_references
2687         https://bugs.webkit.org/show_bug.cgi?id=37758
2688
2689         We tried using the mumble_references convention to manage our
2690         dependencies, but it doesn't seem to be providing much value for the
2691         steps and commands module because these modules are small pieces of the
2692         larger tool module.  In this patch, I've removed the references file
2693         for these modules.
2694
2695         I've left the style_references file for the style module because that
2696         module seems better isolated from the rest of webkitpy and the
2697         style_references file appears to be providing some value.
2698
2699         * Scripts/webkitpy/tool/commands/commandtest.py:
2700         * Scripts/webkitpy/tool/commands/download_unittest.py:
2701         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
2702         * Scripts/webkitpy/tool/commands/queries_unittest.py:
2703         * Scripts/webkitpy/tool/commands/queues_unittest.py:
2704         * Scripts/webkitpy/tool/commands/queuestest.py:
2705         * Scripts/webkitpy/tool/commands/upload.py:
2706         * Scripts/webkitpy/tool/commands/upload_unittest.py:
2707         * Scripts/webkitpy/tool/commands_references.py: Removed.
2708         * Scripts/webkitpy/tool/steps/closebugforlanddiff_unittest.py:
2709         * Scripts/webkitpy/tool/steps/steps_unittest.py:
2710         * Scripts/webkitpy/tool/steps/updatechangelogswithreview_unittest.py:
2711         * Scripts/webkitpy/tool/steps/validatereviewer_unittest.py:
2712         * Scripts/webkitpy/tool/steps_references.py: Removed.
2713
2714 2010-04-17  Eric Seidel  <eric@webkit.org>
2715
2716         Reviewed by Adam Barth.
2717
2718         WebKit needs a Chromium Mac EWS Builder
2719         https://bugs.webkit.org/show_bug.cgi?id=37742
2720
2721         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
2722
2723 2010-04-17  Adam Barth  <abarth@webkit.org>
2724
2725         Fix expected results for unit test broken by
2726         http://trac.webkit.org/changeset/57772
2727
2728         * Scripts/webkitpy/common/net/buildbot_unittest.py:
2729
2730 2010-04-17  Sam Weinig  <weinig@apple.com>
2731
2732         Reviewed by Adam Roben.
2733
2734         Teach windows MiniBrowser how to work with window.open()
2735         and targeted links.
2736
2737         * MiniBrowser/win/BrowserView.cpp:
2738         (createNewPage): Create a new BrowserWindow and return its page.
2739         (showPage): Show the page.
2740         (closePage): Empty implementation.
2741         (runJavaScriptAlert): Empty implementation.
2742         (BrowserView::create): Register a UIClient.
2743         * MiniBrowser/win/BrowserView.h:
2744         (BrowserView::webView): Added.
2745         Change create to take a BrowserWindow instead of an HWND.
2746
2747         * MiniBrowser/win/BrowserWindow.cpp:
2748         (BrowserWindow::wndProc): Respond to WM_NCDESTROY.
2749         (BrowserWindow::goToURL): Added. Forwards to BrowserView.
2750         (BrowserWindow::onCreate): Don't always go to the default
2751         url. Let the caller do this.
2752         (BrowserWindow::onNCDestroy): Delete the window.
2753         * MiniBrowser/win/BrowserWindow.h:
2754         (BrowserWindow::view): Added.
2755         (BrowserWindow::window): Added.
2756
2757         * MiniBrowser/win/main.cpp:
2758         (_tWinMain):
2759         Go to the default URL for the initial page. Allocate the initial
2760         window on the heap for correctness.
2761
2762 2010-04-16  Adam Roben  <aroben@apple.com>
2763
2764         Add the Windows Debug (Test) builder to the list of core builders
2765
2766         It's been green for a few days now, and all the known Windows
2767         flakiness is Release-only.
2768
2769         Rubber-stamped by Mark Rowe.
2770
2771         * Scripts/webkitpy/common/net/buildbot.py:
2772         (BuildBot.core_builder_name_regexps): Added a regular expression to
2773         match the "Windows Debug (Test)" builder.
2774
2775 2010-04-16  Sam Weinig  <sam@webkit.org>
2776
2777         Reviewed by Anders Carlsson.
2778
2779         Fix window.open() and targeted links.
2780
2781         * MiniBrowser/mac/BrowserWindowController.m:
2782         (_createNewPage): Use the correct initializer to and load the window.
2783
2784 2010-04-16  Adam Roben  <aroben@apple.com>
2785
2786         Fix links to layout test results from build status pages
2787
2788         Reviewed by Mark Rowe.
2789
2790         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
2791         (ExtractTestResults.finished): Prepend "/" on the URL of the test
2792         results page so that it is treated as an absolute URL.
2793
2794 2010-04-16  Tony Chang  <tony@chromium.org>
2795
2796         Reviewed by Dimitri Glazkov.
2797
2798         [chromium] build DRT on Linux
2799         https://bugs.webkit.org/show_bug.cgi?id=37690
2800
2801         * Scripts/build-dumprendertree: Add support for win and linux
2802
2803 2010-04-16  Sam Weinig  <weinig@apple.com>
2804
2805         Reviewed by Adam Roben.
2806
2807         Use the threaded process model for MiniBrowser if holding down
2808         the shift key on startup.
2809
2810         * MiniBrowser/win/BrowserView.cpp:
2811         (BrowserView::create):
2812
2813 2010-04-15  Tony Chang  <tony@chromium.org>
2814
2815         Reviewed by Dimitri Glazkov.
2816
2817         [chromium] Add TestShellGtk.cpp so we can link on Linux
2818         https://bugs.webkit.org/show_bug.cgi?id=37561
2819
2820         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp: Add new file and
2821             add platform file exceptions.
2822         * DumpRenderTree/chromium/TestShellGtk.cpp: Added.
2823         (AlarmHandler):
2824         (TestShell::waitTestFinished):
2825
2826 2010-04-15  Tony Chang  <tony@chromium.org>
2827
2828         Reviewed by Dimitri Glazkov.
2829
2830         build DRT on chromium mac
2831         https://bugs.webkit.org/show_bug.cgi?id=37639
2832
2833         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
2834         * Scripts/build-dumprendertree: enable build-dumprendertree --chromium
2835
2836 2010-04-15  Kent Tamura  <tkent@chromium.org>
2837
2838         Reviewed by Dimitri Glazkov.
2839
2840         [DRT/Chromium] Add LayoutTestHelper for Mac
2841         https://bugs.webkit.org/show_bug.cgi?id=37668
2842
2843         LayouTestHelper.mm is based on webkit/tools/test_shell/mac/layout_test_helper.mm
2844         of Chromium.
2845
2846         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
2847         * DumpRenderTree/chromium/LayoutTestHelper.mm: Added.
2848
2849 2010-04-15  Alexey Proskuryakov  <ap@apple.com>
2850
2851         Reviewed by Geoff Garen.
2852
2853         https://bugs.webkit.org/show_bug.cgi?id=37494
2854         <rdar://problem/7857060> REGRESSION (r57340): fast/events/mouse-drag-from-frame-to-other-frame.html
2855         fails on Windows
2856
2857         * DumpRenderTree/win/EventSender.cpp: (makeEventSender):
2858         * DumpRenderTree/win/EventSender.h:
2859         Tell EventSender if it's being created for a top frame.
2860
2861         * DumpRenderTree/win/FrameLoadDelegate.cpp:
2862         (FrameLoadDelegate::didClearWindowObjectForFrameInStandardWorld): We only want to reset
2863         EventSender machinery when a new test is loaded, not when an iframe (or just its global
2864         object) is created.
2865
2866 2010-04-15  Adam Roben  <aroben@apple.com>
2867
2868         Fix Windows WebKit2 build.
2869
2870         * MiniBrowser/win/MiniBrowser.cpp:
2871
2872 2010-04-15  Ojan Vafai  <ojan@chromium.org>
2873
2874         Reviewed by Adam Barth.
2875
2876         Include codereview issue number in patch description
2877         https://bugs.webkit.org/show_bug.cgi?id=37677
2878
2879         This lets us know which rietveld issue this patch is tied to.
2880
2881         Also, make it so that --fancy-review overrides --no-review.
2882
2883         * Scripts/webkitpy/tool/steps/postcodereview.py:
2884         * Scripts/webkitpy/tool/steps/postdiff.py:
2885
2886 2010-04-15  Adam Roben  <aroben@apple.com>
2887
2888         Make --exit-after-n-failures work when all tests are timing out or crashing
2889
2890         Fixes <http://webkit.org/b/37679>.
2891
2892         Reviewed by Jon Honeycutt.
2893
2894         * Scripts/run-webkit-tests:
2895         (top level): When a test crashes or times out, break out of the main loop if
2896         stopRunningTestsEarlyIfNeeded returns true. Moved some code from the bottom of the main loop
2897         from here...
2898         (stopRunningTestsEarlyIfNeeded): ...to here.
2899
2900 2010-04-15  Sam Weinig  <sam@webkit.org>
2901
2902         Reviewed by Anders Carlsson.
2903
2904         Add WebHistoryClient support.
2905         https://bugs.webkit.org/show_bug.cgi?id=37671
2906
2907         Add HistoryClient logging.
2908
2909         * MiniBrowser/mac/BrowserWindowController.m:
2910         (_didNavigateWithNavigationData):
2911         (_didPerformClientRedirect):
2912         (_didPerformServerRedirect):
2913         (_didUpdateHistoryTitle):
2914         (-[BrowserWindowController awakeFromNib]):
2915
2916 2010-04-15  Anders Carlsson  <andersca@apple.com>
2917
2918         Reviewed by Sam Weinig.
2919
2920         Start the mini browser in threaded mode if shift is pressed during startup.
2921         https://bugs.webkit.org/show_bug.cgi?id=37670
2922
2923         * MiniBrowser/mac/AppDelegate.m:
2924         (-[BrowserAppDelegate init]):
2925
2926 2010-04-15  Chris Fleizach  <cfleizach@apple.com>
2927
2928         Reviewed by Beth Dakin.
2929
2930         AXHelp is being appended from ancestors incorrectly
2931         https://bugs.webkit.org/show_bug.cgi?id=37659
2932
2933         * DumpRenderTree/AccessibilityUIElement.cpp:
2934         (getHelpTextCallback):
2935         (AccessibilityUIElement::getJSClass):
2936         * DumpRenderTree/AccessibilityUIElement.h:
2937         * DumpRenderTree/gtk/AccessibilityUIElementGtk.cpp:
2938         (AccessibilityUIElement::helpText):
2939         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
2940         (AccessibilityUIElement::helpText):
2941         * DumpRenderTree/win/AccessibilityUIElementWin.cpp:
2942         (AccessibilityUIElement::helpText):
2943
2944 2010-04-15  Dirk Pranke  <dpranke@chromium.org>
2945
2946         Reviewed by Adam Barth.
2947
2948         Re-format run-webkit-tests to fit in 80-columns for PEP-8 compliance.
2949         (broken by r57463 and r57381, at least). I've also filed bug 37477
2950         to fix check-webkit-style to catch these things.
2951
2952         https://bugs.webkit.org/show_bug.cgi?id=38586
2953
2954         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2955
2956 2010-04-15  Dirk Pranke  <dpranke@chromium.org>
2957
2958         Reviewed by Dimitri Glazkov.
2959
2960         Add a way to override the user-visible name for the test binary since
2961         some ports don't call it DumpRenderTree (e.g., Chromium Win uses
2962         test_shell, Chromium Mac uses TestShell) by adding a driver_name()
2963         method to the Port interface.
2964
2965         https://bugs.webkit.org/show_bug.cgi?id=37631
2966
2967         * Scripts/webkitpy/layout_tests/port/base.py:
2968         * Scripts/webkitpy/layout_tests/port/chromium.py:
2969         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
2970         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2971
2972 2010-04-14  Anders Carlsson  <andersca@apple.com>
2973
2974         Reviewed by Adam Roben.
2975
2976         Add "Force Repaint" to debug menu.
2977         https://bugs.webkit.org/show_bug.cgi?id=37627
2978
2979         * MiniBrowser/mac/BrowserWindowController.h:
2980         * MiniBrowser/mac/BrowserWindowController.m:
2981         (-[BrowserWindowController forceRepaint:]):
2982         * MiniBrowser/mac/English.lproj/MainMenu.xib:
2983
2984 2010-04-15  Sam Weinig  <sam@webkit.org>
2985
2986         Reviewed by Adam Roben.
2987
2988         Add debug-minibrowser script.
2989
2990         * Scripts/debug-minibrowser: Copied from Scripts/run-minibrowser.
2991         * Scripts/webkitdirs.pm:
2992
2993 2010-04-15  Roland Steiner  <rolandsteiner@chromium.org>
2994
2995         Reviewed by Dimitri Glazkov.
2996
2997         Bug 37636 - [DRT/Chromium] Implement DRT/Chromium for Windows
2998         https://bugs.webkit.org/show_bug.cgi?id=37636
2999
3000         Second patch: add Windows-specific implementation parts
3001         of TestShell.
3002
3003         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
3004         * DumpRenderTree/chromium/TestShell.h:
3005         (TestShell::finishedEvent):
3006         * DumpRenderTree/chromium/TestShellWin.cpp:
3007         (watchDogThread):
3008         (TestShell::waitTestFinished):
3009
3010 2010-04-15  Roland Steiner  <rolandsteiner@chromium.org>
3011
3012         Reviewed by Dimitri Glazkov.
3013
3014         Bug 37636 - [DRT/Chromium] Implement DRT/Chromium for Windows
3015         https://bugs.webkit.org/show_bug.cgi?id=37636
3016
3017         First patch: fix compiler errors.
3018
3019         * DumpRenderTree/chromium/EventSender.cpp:
3020         (EventSender::reset):
3021         (EventSender::dispatchMessage):
3022         * DumpRenderTree/chromium/LayoutTestController.cpp:
3023         (LayoutTestController::pathToLocalResource):
3024         * DumpRenderTree/chromium/TestWebWorker.h:
3025         * DumpRenderTree/chromium/TextInputController.cpp:
3026         * DumpRenderTree/chromium/WebViewHost.h:
3027         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
3028
3029 2010-04-14  Luiz Agostini  <luiz.agostini@openbossa.org>
3030
3031         Reviewed by Simon Hausmann.
3032
3033         [Qt] Moving setViewMode from DumpRenderTreeSupportQt to qwebpage.cpp
3034         https://bugs.webkit.org/show_bug.cgi?id=37622
3035
3036         Method qt_wrt_setViewMode was removed from qwebpage.cpp by mistake in r57433
3037         (bug 35844). Moving it back.
3038
3039         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3040         (LayoutTestController::setViewModeMediaFeature):
3041
3042 2010-04-15  Dirk Pranke  <dpranke@chromium.org>
3043
3044         Reviewed by Adam Barth.
3045
3046         Add some very minimal unit tests for new-run-webkit-tests. This should
3047         be enough to catch egregious brokenness like syntax errors and import
3048         declaration issues.
3049
3050         https://bugs.webkit.org/show_bug.cgi?id=37432
3051
3052         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3053         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py: Added.
3054
3055 2010-04-14  Brian Weinstein  <bweinstein@apple.com>
3056
3057         Reviewed by Adam Roben.
3058
3059         Add a way for the buildbot to kill any old processes that are running. This
3060         is useful because the Windows bots can get in states where a process remains
3061         running (httpd.exe, DumpRenderTree.exe), which causes the bots to get in a red
3062         state, and the processes must be killed manually.
3063
3064         * BuildSlaveSupport/build.webkit.org-config/master.cfg: If we are on Windows, kill
3065             the old processes that might be running.
3066         * BuildSlaveSupport/win/kill-old-processes: Added.
3067
3068 2010-04-14  Sam Weinig  <sam@webkit.org>
3069
3070         Reviewed by Anders Carlsson.
3071
3072         Tear down WebKit more completely on window closing and application
3073         termination. We still don't block application termination for pending
3074         close, but this is a step in the right direction.
3075
3076         * MiniBrowser/mac/AppDelegate.m:
3077         (-[BrowserAppDelegate applicationWillTerminate:]):
3078         * MiniBrowser/mac/BrowserWindowController.h:
3079         * MiniBrowser/mac/BrowserWindowController.m:
3080         (-[BrowserWindowController dealloc]):
3081         (-[BrowserWindowController windowWillClose:]):
3082         (-[BrowserWindowController applicationTerminating]):
3083
3084 2010-04-14  Adam Barth  <abarth@webkit.org>
3085
3086         Reviewed by Eric Seidel.
3087
3088         Use pretty patch for confirming webkit-patch diffs
3089         https://bugs.webkit.org/show_bug.cgi?id=37489
3090
3091         * Scripts/webkitpy/common/prettypatch.py: Added.
3092         * Scripts/webkitpy/tool/commands/download_unittest.py:
3093         * Scripts/webkitpy/tool/commands/upload_unittest.py:
3094         * Scripts/webkitpy/tool/mocktool.py:
3095         * Scripts/webkitpy/tool/steps/confirmdiff.py:
3096
3097 2010-04-14  Eric Seidel  <eric@webkit.org>
3098
3099         Reviewed by Adam Barth.
3100
3101         Teach webkit-patch how to handle revisions missing ChangeLogs
3102         https://bugs.webkit.org/show_bug.cgi?id=37519
3103
3104         Make commit_info_for_revision return None when revision
3105         is missing a ChangeLog.  Previously we would throw an array index
3106         exception.
3107         Teach callers how to handle None.
3108
3109         * Scripts/webkitpy/common/checkout/api.py:
3110         * Scripts/webkitpy/common/checkout/api_unittest.py:
3111         * Scripts/webkitpy/tool/commands/download.py:
3112         * Scripts/webkitpy/tool/commands/queries.py:
3113         * Scripts/webkitpy/tool/commands/sheriffbot.py:
3114
3115 2010-04-13  Eric Seidel  <eric@webkit.org>
3116
3117         Reviewed by Adam Roben.
3118
3119         Add Qt Bot to the list of "core builders" (builders which block the commit-queue when red)
3120         https://bugs.webkit.org/show_bug.cgi?id=33297
3121
3122         This is an experiment.  The bots have been green for
3123         a while.  We'll see if adding them under sheriff-bot protection
3124         will keep them green.
3125
3126         * Scripts/webkitpy/common/net/buildbot.py:
3127         * Scripts/webkitpy/common/net/buildbot_unittest.py:
3128
3129 2010-04-13  Eric Seidel  <eric@webkit.org>
3130
3131         Unreviewed, just fixing a constant in the Rietveld unit test.
3132
3133         * Scripts/webkitpy/common/net/rietveld_unittest.py
3134
3135 2010-04-13  Eric Seidel  <eric@webkit.org>
3136
3137         Unreviewed, just adding missing Mock to fix python tests.
3138
3139         * Scripts/webkitpy/tool/mocktool.py:
3140
3141 2010-04-13  Ojan Vafai  <ojan@chromium.org>
3142
3143         Reviewed by David Levin.
3144
3145         Add experimental prototype Rietveld integration to webkit-patch upload
3146         https://bugs.webkit.org/show_bug.cgi?id=37418
3147
3148         This patch adds bare-bones integration with Rietveld for code reviews.
3149         The behavior is hidden behind the --fancy-review command line flag.
3150         Currently, there's no support for uploading more than one patch per
3151         issue (which is a nice feature of Rietveld).  The plan is to play with
3152         this for a bit and see if it's useful.
3153
3154         Modified from Adam's original patch to autoinstall the rietveld upload script.
3155
3156         * Scripts/webkitpy/common/config/__init__.py:
3157         * Scripts/webkitpy/common/net/rietveld.py: Added.
3158         * Scripts/webkitpy/common/net/rietveld_unitttest.py: Added.
3159         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3160         * Scripts/webkitpy/tool/commands/upload.py:
3161         * Scripts/webkitpy/tool/commands/upload_unittest.py:
3162         * Scripts/webkitpy/tool/main.py:
3163         * Scripts/webkitpy/tool/mocktool.py:
3164         * Scripts/webkitpy/tool/steps/__init__.py:
3165         * Scripts/webkitpy/tool/steps/options.py:
3166         * Scripts/webkitpy/tool/steps/postcodereview.py: Added.
3167         * Scripts/webkitpy/tool/steps/postdiff.py:
3168
3169 2010-04-13  Sam Weinig  <sam@webkit.org>
3170
3171         Rubber-stamped by Mark Rowe.
3172
3173         Add Makefile to MiniBrowser.
3174
3175         * MiniBrowser/Makefile: Added.
3176
3177 2010-04-13  Eric Seidel  <eric@webkit.org>
3178
3179         Unreviewed, just adding a sanity check.
3180
3181         Add check to make sure commit-queue can never commit too short a message
3182         https://bugs.webkit.org/show_bug.cgi?id=37528
3183
3184         The commit-queue made bogus messages here:
3185         http://trac.webkit.org/changeset/57532
3186         http://trac.webkit.org/changeset/57534
3187
3188         This was a regression caused by adding unicode parsing for
3189         our ChangeLog files.  Popen does not seem to play nice with
3190         unicode strings.
3191
3192         I'm also adding an "assert" to make sure short ChangeLogs never happen again.
3193
3194         * Scripts/webkitpy/common/system/executive.py:
3195          - Cast input to strings before passing to POpen
3196         * Scripts/webkitpy/tool/steps/commit.py:
3197          - Validate that commit messages are not to short.
3198
3199 2010-04-13  Adam Roben  <aroben@apple.com>
3200
3201         Robustify new-run-webkit-tests against paths with spaces in them
3202
3203         Reviewed by Eric Seidel.
3204
3205         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
3206         (LayoutTestApacheHttpd.__init__): Quote all paths that we pass to
3207         Apache to ensure that paths with spaces in them are interpreted
3208         correctly.
3209
3210 2010-04-13  Csaba Osztrogonác  <ossy@webkit.org>
3211
3212         Unreviewed buildfix after r57537.
3213
3214         * DumpRenderTree/qt/LayoutTestControllerQt.h: Declaration of removeOriginAccessWhitelistEntry() added.
3215
3216 2010-04-13  Sam Weinig  <sam@webkit.org>
3217
3218         Reviewed by Maciej Stachowiak.
3219
3220         Make building new webkit API and MiniBrowser a little easier.
3221
3222         * Scripts/build-webkit: Make building with --webkit2 build the
3223         MiniBrowser as well and tell you how to use it.
3224         * Scripts/run-minibrowser: Copied from Scripts/run-safari.
3225         * Scripts/webkitdirs.pm: Add runMiniBrowser function.
3226
3227 2010-04-12  Timothy Hatcher  <timothy@apple.com>
3228
3229         SecurityOrigin needs a way to remove individual OriginAccessEntries
3230         https://bugs.webkit.org/show_bug.cgi?id=37449
3231
3232         Reviewed by Dave Hyatt.
3233
3234         * DumpRenderTree/LayoutTestController.cpp:
3235         (removeOriginAccessWhitelistEntryCallback): Added. Call LayoutTestController::removeOriginAccessWhitelistEntry.
3236         (LayoutTestController::staticFunctions): Added removeOriginAccessWhitelistEntry.
3237         * DumpRenderTree/LayoutTestController.h:
3238         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3239         (LayoutTestController::removeOriginAccessWhitelistEntry): Added. FIXME to implement.
3240         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3241         (LayoutTestController::removeOriginAccessWhitelistEntry): Added.
3242         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3243         (LayoutTestController::removeOriginAccessWhitelistEntry): Added. FIXME to implement.
3244         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3245         (LayoutTestController::removeOriginAccessWhitelistEntry): Added.
3246         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3247         (LayoutTestController::whiteListAccessFromOrigin): FIXME to implement.
3248         (LayoutTestController::removeOriginAccessWhitelistEntry): Added. FIXME to implement.
3249
3250 2010-04-13  Timothy Hatcher  <timothy@apple.com>
3251
3252         Rename SecurityOrigin::whiteListAccessFromOrigin to addOriginAccessWhitelistEntry.
3253         And LayoutTestController.whiteListAccessFromOrigin to addOriginAccessWhitelistEntry.
3254         And SecurityOrigin::resetOriginAccessWhiteLists to resetOriginAccessWhitelists.
3255
3256         SecurityOrigin needs a way to remove individual OriginAccessEntries
3257         https://bugs.webkit.org/show_bug.cgi?id=37449
3258
3259         Reviewed by Dave Hyatt.
3260
3261         * DumpRenderTree/LayoutTestController.cpp:
3262         (addOriginAccessWhitelistEntryCallback):
3263         (LayoutTestController::staticFunctions):
3264         * DumpRenderTree/LayoutTestController.h:
3265         * DumpRenderTree/chromium/LayoutTestController.cpp:
3266         (LayoutTestController::LayoutTestController):
3267         (LayoutTestController::addOriginAccessWhitelistEntry):
3268         * DumpRenderTree/chromium/LayoutTestController.h:
3269         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3270         (LayoutTestController::addOriginAccessWhitelistEntry):
3271         * DumpRenderTree/mac/DumpRenderTree.mm:
3272         (resetWebViewToConsistentStateBeforeTesting):
3273         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3274         (LayoutTestController::addOriginAccessWhitelistEntry):
3275         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3276         (LayoutTestController::addOriginAccessWhitelistEntry):
3277         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3278         * DumpRenderTree/win/DumpRenderTree.cpp:
3279         (resetWebViewToConsistentStateBeforeTesting):
3280         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3281         (LayoutTestController::addOriginAccessWhitelistEntry):
3282         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3283         (LayoutTestController::addOriginAccessWhitelistEntry):
3284
3285 2010-04-13  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
3286
3287         Reviewed by Eric Seidel.
3288
3289         Remove duplicate function for new-run-webkit-tests
3290         https://bugs.webkit.org/show_bug.cgi?id=37517
3291
3292         The version() function was already implemented.
3293
3294         * Scripts/webkitpy/layout_tests/port/webkit.py:
3295
3296 2010-04-13  Eric Seidel  <eric@webkit.org>
3297
3298         Reviewed by David Levin.
3299
3300         the commit-queue hates Tor Arne Vestbø
3301         https://bugs.webkit.org/show_bug.cgi?id=37511
3302
3303         We were failing to read reviewers out of ChangeLogs
3304         when the reviewer has unicode characters in his/her name.
3305         I fixed this by explicitly decoding from utf8 every time we
3306         read in a ChangeLog file (they are always UTF8).
3307
3308         * Scripts/webkitpy/common/checkout/changelog.py:
3309         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
3310
3311 2010-04-13  Adam Roben  <aroben@apple.com>
3312
3313         Fix run-webkit-tests on Windows with spaces in the path
3314
3315         Fixes <http://webkit.org/b/37509>.
3316
3317         Reviewed by Steve Falkenburg.
3318
3319         * Scripts/run-webkit-tests:
3320         (convertPathUsingCygpath): Remove spaces from the path before passing
3321         them to cygpath, then add them back in after conversion, as some
3322         versions of cygpath seem to convert spaces into newlines.
3323
3324 2010-04-13  Adam Barth  <abarth@webkit.org>
3325
3326         Unreviewed, but approved by Dumitru Daniliuc.  (This patch is intended
3327         to fix the downstream Chromium build bots.  Hopefully it will work!)
3328
3329         Add a driver script for the new websocket server
3330         https://bugs.webkit.org/show_bug.cgi?id=37495
3331
3332         websocket_server.py can't be run directly because its a module and not
3333         a standalone script.  This used to work by accident because it didn't
3334         depend on any other modules.
3335
3336         * Scripts/new-run-webkit-websocketserver: Added.
3337         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
3338
3339 2010-04-12  Eric Seidel  <eric@webkit.org>
3340
3341         Reviewed by Adam Barth.
3342
3343         Make new-run-webkit-test PrettyPatch failure reporting more awesome
3344         https://bugs.webkit.org/show_bug.cgi?id=37487
3345
3346         I also fixed an Executive/executive typo.
3347
3348         * Scripts/webkitpy/layout_tests/port/base.py:
3349
3350 2010-04-12  Eric Seidel  <eric@webkit.org>
3351
3352         Reviewed by Adam Barth.
3353
3354         new-run-webkit-tests should only build java support files on Mac
3355         https://bugs.webkit.org/show_bug.cgi?id=37482
3356
3357         Only the mac needs java support files, so I pushed _build_java
3358         down into the Mac port using a new hook "_check_port_build".
3359         In the process I noticed a bunch of code which could be shared
3360         between all ports and thus got rid of _tests_for_disabled_features
3361         and version() copy/paste between all webkit ports.
3362         I also made check_build only bother to check for ImageDiff if we're
3363         using pixel tests.
3364
3365         * Scripts/webkitpy/layout_tests/port/gtk.py:
3366         * Scripts/webkitpy/layout_tests/port/mac.py:
3367         * Scripts/webkitpy/layout_tests/port/qt.py:
3368         * Scripts/webkitpy/layout_tests/port/webkit.py:
3369         * Scripts/webkitpy/layout_tests/port/win.py:
3370
3371 2010-04-12  Dumitru Daniliuc  <dumi@chromium.org>
3372
3373         Unreviewed, trying to make scripts work on machines without
3374         Ruby...
3375
3376         * Scripts/webkitpy/layout_tests/port/base.py:
3377
3378 2010-04-12  Tor Arne Vestbø <tor.arne.vestbo@nokia.com>
3379
3380         Reviewed by Eric Seidel.
3381
3382         Add stub files for running new-run-webkit-tests for the Qt port
3383
3384         * Scripts/webkitpy/layout_tests/port/factory.py:
3385         * Scripts/webkitpy/layout_tests/port/qt.py: Added.
3386
3387 2010-04-12  Dirk Pranke  <dpranke@chromium.org>
3388
3389         Reviewed by Eric Seidel.
3390
3391         Modify run_webkit_tests.py to not call sys.exit() at the end of test
3392         run; doing so makes it more difficult to embed the routine for,
3393         among other things, unit tests. We push the exit calling up into
3394         new-run-webkit-tests.
3395
3396         https://bugs.webkit.org/show_bug.cgi?id=37464
3397
3398         * Scripts/new-run-webkit-tests:
3399         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3400
3401 2010-04-12  Eric Seidel  <eric@webkit.org>
3402
3403         Unreviewed.
3404
3405         Add stub Gtk implementation for new-run-webkit-tests.
3406
3407         * Scripts/webkitpy/layout_tests/port/factory.py:
3408         * Scripts/webkitpy/layout_tests/port/gtk.py: Added.
3409
3410 2010-04-12  Eric Seidel  <eric@webkit.org>
3411
3412         Reviewed by Adam Barth.
3413
3414         new-run-webkit-tests fails with exception on systems missing ruby
3415         https://bugs.webkit.org/show_bug.cgi?id=37441
3416
3417         * Scripts/webkitpy/layout_tests/port/base.py:
3418          - Catch failures similar to how wdiff code path does.
3419          - After one failure, stop trying.
3420
3421 2010-04-12  Tony Chang  <tony@chromium.org>
3422
3423         Reviewed by Eric Seidel.
3424
3425         Fix how we import simplejson based on how it's used in this file.
3426         This fixes exceptions raised when trying to write the simplejson output.
3427
3428         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3429
3430 2010-04-11  Chris Jerdonek  <cjerdonek@webkit.org>
3431
3432         Reviewed by Shinichiro Hamaji.
3433
3434         Refactored check-webkit-style so that the StyleChecker class
3435         has no dependencies on patch-related concepts.
3436
3437         https://bugs.webkit.org/show_bug.cgi?id=37065
3438
3439         This patch is an intermediate step towards making the StyleChecker
3440         class a generalized file processor that can do arbitary operations
3441         on the files corresponding to a list of paths.  This patch
3442         also simplifies the unit-testing of patch-checking code.
3443
3444         * Scripts/check-webkit-style:
3445           - Updated to use the new PatchChecker class.
3446
3447         * Scripts/webkitpy/style/checker.py:
3448           - Refactored the StyleChecker.check_patch() method into the
3449             check() method of a new PatchChecker class.
3450
3451         * Scripts/webkitpy/style/checker_unittest.py:
3452           - Refactored the unit tests as necessary, changing the
3453             StyleCheckerCheckPatchTest class to a PatchCheckerTest class.
3454
3455 2010-04-11  Adam Barth  <abarth@webkit.org>
3456
3457         Unreviewed.
3458
3459         Fix new-run-webkit-tests regressions cased by Eric's option parsing patch
3460         https://bugs.webkit.org/show_bug.cgi?id=37430
3461
3462         We need some basic unit testing of this script, or we're going to keep
3463         breaking it like this.  Added missing namespace qualifiers and
3464         propagated renaming of an option.
3465
3466         * Scripts/webkitpy/common/config/ports.py:
3467         * Scripts/webkitpy/layout_tests/driver_test.py:
3468         * Scripts/webkitpy/layout_tests/port/base.py:
3469         * Scripts/webkitpy/layout_tests/port/mac.py:
3470         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3471
3472 2010-04-11  Eric Seidel  <eric@webkit.org>
3473
3474         Reviewed by Adam Barth.
3475
3476         Break new-run-webkit-tests options into groups for easier re-use and possible relocation
3477         https://bugs.webkit.org/show_bug.cgi?id=37408
3478
3479         new-run-webkit-tests currently has one huge function for
3480         dealing with all options-parsing.
3481         This patch is a first attempt at trying to split that large
3482         function down into smaller (hopefully more readable?) chunks
3483         dealing with the different areas of options.
3484         For example, it would make sense to move configuration
3485         options off into some module which deals with the vagries of
3486         WebKit's configuration system.  It would also make sense to move
3487         Chromium options off onto the Chromium port object (where they are used).
3488         It may make sense to move results.json options over to the results.json code.
3489         This change is a first iteration, and we will certainly need more
3490         refinement to this code over time.  Hopefully I didn't make things
3491         harder to read here.
3492
3493         * Scripts/webkitpy/layout_tests/driver_test.py:
3494         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3495
3496 2010-04-11  Adam Barth  <abarth@webkit.org>
3497
3498         Reviewed by Eric Seidel.
3499
3500         create-rollout copy needs to be updated to reflect removal of --no-build
3501         https://bugs.webkit.org/show_bug.cgi?id=37425
3502
3503         Removed --no-build and --no-test from the instructions because these
3504         don't exist anymore.
3505
3506         * Scripts/webkitpy/tool/commands/download_unittest.py:
3507             - Updated the expected results to reflect the new copy.
3508         * Scripts/webkitpy/tool/steps/postdiffforrevert.py:
3509
3510 2010-04-11  Sheriff Bot  <webkit.review.bot@gmail.com>
3511
3512         Unreviewed, rolling out r57460.
3513         http://trac.webkit.org/changeset/57460
3514         https://bugs.webkit.org/show_bug.cgi?id=37424
3515
3516         broke chromium builders (Requested by tony^work on #webkit).
3517
3518         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:
3519
3520 2010-04-11  Tony Chang  <tony@chromium.org>
3521
3522         Reviewed by Darin Fisher.
3523
3524         [chromium] update chromium DEPS for upstream compile
3525         https://bugs.webkit.org/show_bug.cgi?id=36578
3526
3527         * DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp: Small fix to avoid a circular dependency between
3528           WebKit.gyp and webkit.gyp.
3529
3530 2010-04-11  Joseph Pecoraro  <joepeck@webkit.org>
3531
3532         Reviewed by Eric Seidel.
3533
3534         svn-apply errors out when removing directories in git
3535         https://bugs.webkit.org/show_bug.cgi?id=34871
3536
3537         * Scripts/svn-apply:
3538         (isDirectoryEmptyForRemoval): early break if the directory doesn't exist
3539         (scmRemove): have git ignore unmatched files
3540
3541 2010-04-11  Daniel Bates  <dbates@rim.com>
3542
3543         No review, rolling out 57440.
3544         http://trac.webkit.org/changeset/57440
3545         https://bugs.webkit.org/show_bug.cgi?id=27204
3546
3547         Did not handle Git patches that included both file and property
3548         changes to the same file. Rolling this change out while I look
3549         into this.
3550
3551         * Scripts/VCSUtils.pm:
3552         * Scripts/svn-apply:
3553         * Scripts/svn-unapply:
3554         * Scripts/webkitperl/VCSUtils_unittest/appendSVNExecutableBitChangeToPatch.pl: Removed.
3555         * Scripts/webkitperl/VCSUtils_unittest/parseGitFileMode.pl: Removed.
3556         * Scripts/webkitperl/VCSUtils_unittest/parseStartOfPatchOrPropertyChangeAndEndOfPropertyChange.pl: Removed.
3557
3558 2010-04-11  Adam Barth  <abarth@webkit.org>
3559
3560         Reviewed by Eric Seidel.
3561
3562         Fix typo in log path for AbstractQueue
3563         https://bugs.webkit.org/show_bug.cgi?id=37414
3564
3565         * Scripts/webkitpy/tool/commands/queues.py:
3566         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3567
3568 2010-04-11  Adam Barth  <abarth@webkit.org>
3569
3570         Unreviewed.  Add missing import statement.
3571
3572         * Scripts/webkitpy/common/system/executive.py:
3573
3574 2010-04-11  Robert Hogan  <robert@webkit.org>
3575
3576         Reviewed by Simon Hausmann.
3577
3578         [Qt] Add setWillSendRequestReturnsNull and setWillSendRequestClearHeader
3579
3580         https://bugs.webkit.org/show_bug.cgi?id=37410
3581
3582         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3583         (LayoutTestController::reset):
3584         (LayoutTestController::setWillSendRequestReturnsNull):
3585         (LayoutTestController::setWillSendRequestClearHeader):
3586         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3587
3588 2010-04-11  Csaba Osztrogonác  <ossy@webkit.org>
3589
3590         [Qt] Unreviewed buildfix for --debug build after r57433.
3591
3592         Refactor Qt DRT support in QtWebKit
3593         https://bugs.webkit.org/show_bug.cgi?id=35844
3594
3595         * QtLauncher/main.cpp: qt_drt_garbageCollector_collect(); renamed to DumpRenderTreeSupportQt::garbageCollectorCollect();
3596         (launcherMain):
3597
3598 2010-04-11  Joseph Pecoraro  <joepeck@webkit.org>
3599
3600         Reviewed by Mark Rowe.
3601
3602         Make commit-log-editor Rubber-stamp aware. And other minor cleanups.
3603         https://bugs.webkit.org/show_bug.cgi?id=37407
3604
3605         * Scripts/commit-log-editor:
3606
3607 2010-04-11  Eric Seidel  <eric@webkit.org>
3608
3609         Reviewed by Adam Barth.
3610
3611         Add PrettyPatch links to new-run-webkit-tests output
3612         https://bugs.webkit.org/show_bug.cgi?id=37406
3613
3614         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3615          - We're leaking a file handle here, add a FIXME.
3616         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
3617          - Add pretty diff links.
3618         * Scripts/webkitpy/layout_tests/port/base.py:
3619          - Add support for generating pretty diffs using PrettyPatch.
3620         * Scripts/webkitpy/layout_tests/port/webkit.py:
3621          - We're leaking another file handle here, another FIXME.
3622         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
3623          - Update write_output_files signature.
3624         * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
3625          - Remove unused arguments from write_output_files.
3626          - Add support for dumping pretty diffs to write_output_files.
3627          - Fix a bunch of file descriptor leaks in this file.
3628         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
3629          - Update write_output_files signature.
3630
3631 2010-04-10  Adam Barth  <abarth@webkit.org>
3632
3633         Reviewed by Eric Seidel.
3634
3635         kill_process is copy/pasted in five places
3636         https://bugs.webkit.org/show_bug.cgi?id=37405
3637
3638         We shouldn't replicate the kill_process logic in every port.  Instead,
3639         we should move the process interaction to Executive.
3640
3641         Dirk mentioned that he wanted this abstraction to make it easier to
3642         mock things out for testing.  It turns out this function is only used
3643         in one place where it can't be used as a mock point for testing because
3644         the corresponding create process actually creates a real process.  In
3645         the long term, we should indirect both these calls through a non-static
3646         Executive as a mock point.  However, we should wait on that until we
3647         actually want to write the test.
3648
3649         * Scripts/webkitpy/layout_tests/port/base.py:
3650         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
3651         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
3652         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
3653         * Scripts/webkitpy/layout_tests/port/mac.py:
3654         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
3655         * Scripts/webkitpy/layout_tests/port/win.py:
3656
3657 2010-04-10  Adam Barth  <abarth@webkit.org>
3658
3659         Reviewed by Eric Seidel.
3660
3661         new-run-webkit-tests shouldn't alter its path to import packages
3662         https://bugs.webkit.org/show_bug.cgi?id=37404
3663
3664         * Scripts/new-run-webkit-tests:
3665         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
3666         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
3667         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
3668         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3669         * Scripts/webkitpy/layout_tests/test_types/fuzzy_image_diff.py:
3670         * Scripts/webkitpy/layout_tests/test_types/image_diff.py:
3671         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
3672         * Scripts/webkitpy/thirdparty/simplejson/decoder.py:
3673
3674 2010-04-10  Eric Seidel  <eric@webkit.org>
3675
3676         Reviewed by Adam Barth.
3677
3678         new-run-webkit-tests should store results to a directory under the build tree
3679         https://bugs.webkit.org/show_bug.cgi?id=37380
3680
3681         * Scripts/webkitpy/layout_tests/port/webkit.py:
3682
3683 2010-04-10  Daniel Bates  <dbates@rim.com>
3684
3685         Reviewed by Eric Seidel.
3686
3687         https://bugs.webkit.org/show_bug.cgi?id=27204
3688
3689         Implement support for changing the executable bit of a file.
3690         The executable bit is among the most changed file properties.
3691         Future support can include other property changes.
3692
3693         Currently, if a patch changes the executable bit of a file
3694         it is not respected by svn-apply or svn-unapply. Since the
3695         commit-queue bot uses these tools as part of its workflow,
3696         such patches cannot be committed by it. That is, such patches
3697         need to be committed by hand. Instead, we should add support
3698         for the executable bit so that such patches can be committed
3699         by the commit-queue bot.
3700
3701         * Scripts/VCSUtils.pm: Also change reference to Apple Computer, Inc.
3702         in copyright to Apple, Inc.
3703         * Scripts/svn-apply:
3704         * Scripts/svn-unapply:
3705         * Scripts/webkitperl/VCSUtils_unittest/appendSVNExecutableBitChangeToPatch.pl: Added.
3706         * Scripts/webkitperl/VCSUtils_unittest/parseGitFileMode.pl: Added.
3707         * Scripts/webkitperl/VCSUtils_unittest/parseStartOfPatchOrPropertyChangeAndEndOfPropertyChange.pl: Added.
3708
3709 2010-04-10  Eric Seidel  <eric@webkit.org>
3710
3711         Reviewed by Adam Barth.
3712
3713         WinEWS bot fails to svn update because scm.clean_working_directory leaves files around
3714         https://bugs.webkit.org/show_bug.cgi?id=37401
3715
3716         The Git-based bots don't have this trouble because
3717         Git.clean_working_directory fully removes files that were
3718         marked as "add".  SVN.clean_working_directory previously just
3719         called "svn revert" which would leave added files in the
3720         working directory untracked.  This patch makes
3721         SVN.clean_working_directory function more like
3722         Git.clean_working_directory by removing added files after revert.
3723
3724         * Scripts/webkitpy/common/checkout/scm.py:
3725          - Add SCM.absolute_path for easy conversion between
3726            repository-relative paths and absolute paths.
3727          - Add SCM.add and SCM.added_files
3728          - Make SVN.clean_working_directory remove any added_files after svn revert.
3729          - The new unit tests found a bug in Git.status_command, change to use git diff --name-status instead.
3730         * Scripts/webkitpy/common/checkout/scm_unittest.py:
3731          - Add tests for added code.
3732
3733 2010-04-10  Adam Barth  <abarth@webkit.org>
3734
3735         Reviewed by Eric Seidel.
3736
3737         new-run-webkit-tests can't find ImageDiff on Windows
3738         https://bugs.webkit.org/show_bug.cgi?id=37403
3739
3740         It turns out the build directory on Windows is structured differently
3741         than it is on other platforms.  Instead of assuming the normal
3742         structure, we should just ask perl to figure it out for us.
3743
3744         * Scripts/webkitpy/layout_tests/port/webkit.py:
3745
3746 2010-04-10  Adam Barth  <abarth@webkit.org>
3747
3748         Rubber-stamped by Eric Seidel.
3749
3750         Change "Gathering files" status line to "Collecting tests".  Gathering
3751         the files sounds silly to me.
3752
3753         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3754
3755 2010-04-10  Adam Barth  <abarth@webkit.org>
3756
3757         Reviewed by Eric Seidel.
3758
3759         Fix bugs to make new-run-webkit-tests almost run on windows
3760         https://bugs.webkit.org/show_bug.cgi?id=37400
3761
3762         Fix some minor bugs that prevent new-run-webkit-tests from being run on
3763         Windows.  I still haven't run it to completion, but I'm getting
3764         further.
3765
3766         * Scripts/webkitpy/layout_tests/port/factory.py:
3767         * Scripts/webkitpy/layout_tests/port/webkit.py:
3768
3769 2010-04-10  Robert Hogan  <robert@webkit.org>
3770
3771         Reviewed by Kenneth Rohde Christiansen.
3772
3773         Refactor Qt DRT support in QtWebKit
3774
3775         Update Qt DRT to use new DumpRenderTreeSupportQt static class.
3776
3777         https://bugs.webkit.org/show_bug.cgi?id=35844
3778
3779         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
3780         (WebCore::DumpRenderTree::DumpRenderTree):
3781         (WebCore::DumpRenderTree::resetToConsistentStateBeforeTesting):
3782         (WebCore::DumpRenderTree::dumpFramesAsText):
3783         * DumpRenderTree/qt/DumpRenderTreeQt.h:
3784         * DumpRenderTree/qt/GCControllerQt.cpp:
3785         (GCController::collect):
3786         (GCController::collectOnAlternateThread):
3787         (GCController::getJSObjectCount):
3788         * DumpRenderTree/qt/GCControllerQt.h:
3789         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3790         (LayoutTestController::counterValueForElementById):
3791         (LayoutTestController::setViewModeMediaFeature):
3792         (LayoutTestController::setMediaType):
3793         (LayoutTestController::closeWebInspector):
3794         (LayoutTestController::showWebInspector):
3795         (LayoutTestController::evaluateInWebInspector):
3796         (LayoutTestController::setFrameFlatteningEnabled):
3797         (LayoutTestController::setJavaScriptProfilingEnabled):
3798         (LayoutTestController::setTimelineProfilingEnabled):
3799         (LayoutTestController::pauseAnimationAtTimeOnElementWithId):
3800         (LayoutTestController::pauseTransitionAtTimeOnElementWithId):
3801         (LayoutTestController::sampleSVGAnimationForElementAtTime):
3802         (LayoutTestController::numberOfActiveAnimations):
3803         (LayoutTestController::whiteListAccessFromOrigin):
3804         (LayoutTestController::setCaretBrowsingEnabled):
3805         (LayoutTestController::setDomainRelaxationForbiddenForURLScheme):
3806         (LayoutTestController::workerThreadCount):
3807         (LayoutTestController::pageNumberForElementById):
3808         (LayoutTestController::numberOfPages):
3809         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3810
3811 2010-04-10  Adam Barth  <abarth@webkit.org>
3812
3813         Reviewed by Eric Seidel.
3814
3815         Move global queue log to the logs directory so it survives git clean -f
3816         https://bugs.webkit.org/show_bug.cgi?id=37395
3817
3818         * Scripts/webkitpy/tool/commands/queues.py:
3819
3820 2010-04-10  Adam Barth  <abarth@webkit.org>
3821
3822         Reviewed by Eric Seidel.
3823
3824         SheriffBot should spam when it encounters errors
3825         https://bugs.webkit.org/show_bug.cgi?id=37329
3826
3827         We need to always update the status server so we don't get stuck in a
3828         spam loop.  I tried writing a test for this change, but it kind of
3829         got out of control.  We need a better way to do failure injection.
3830
3831         * Scripts/webkitpy/tool/commands/sheriffbot.py:
3832
3833 2010-04-10  Adam Barth  <abarth@webkit.org>
3834
3835         Unreviewed attempt to fix the Chromium Mac canary.
3836
3837         * Scripts/webkitpy/common/config/ports.py:
3838         * Scripts/webkitpy/common/config/ports_unittest.py:
3839         * Scripts/webkitpy/common/system/executive.py:
3840
3841 2010-04-10  Adam Barth  <abarth@webkit.org>
3842
3843         Unreviewed.
3844
3845         Add the Apache bits to win.py for new-run-webkit-tests
3846         https://bugs.webkit.org/show_bug.cgi?id=37397
3847
3848         I still have run this yet, but I looked around to figure out what the
3849         various paths appear to be.  I'll figure out a way to remove the
3850         copy/paste code in a future patch.
3851
3852         * Scripts/webkitpy/layout_tests/port/mac.py:
3853         * Scripts/webkitpy/layout_tests/port/webkit.py:
3854         * Scripts/webkitpy/layout_tests/port/win.py:
3855
3856 2010-04-10  Adam Barth  <abarth@webkit.org>
3857
3858         Reviewed by Eric Seidel.
3859
3860         Sketch out the win port for new-run-webkit-tests
3861         https://bugs.webkit.org/show_bug.cgi?id=37393
3862
3863         I haven't tried running this yet, but we've got to start somewhere.
3864
3865         * Scripts/webkitpy/layout_tests/port/factory.py:
3866         * Scripts/webkitpy/layout_tests/port/mac.py:
3867         * Scripts/webkitpy/layout_tests/port/webkit.py:
3868         * Scripts/webkitpy/layout_tests/port/win.py: Added.
3869
3870 2010-04-10  Ojan Vafai  <ojan@chromium.org>
3871
3872         Reviewed by Adam Barth.
3873
3874         webkit-patch land should not build and test by default
3875         https://bugs.webkit.org/show_bug.cgi?id=33631
3876
3877         Reverse the sense of --no-build and --no-test to be --build and --test.
3878         Also, decoupled the build and test options so you can test without
3879         building.
3880
3881         (Patch manngled by Adam Barth.  All bugs are his fault.)
3882
3883         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
3884         * Scripts/webkitpy/tool/commands/queues.py:
3885         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3886         * Scripts/webkitpy/tool/steps/options.py:
3887         * Scripts/webkitpy/tool/steps/runtests.py:
3888
3889 2010-04-10  Adam Barth  <abarth@webkit.org>
3890
3891         Reviewed by Eric Seidel.
3892
3893         Factor WebKitPort out of MacPort to allow for WinPort
3894         https://bugs.webkit.org/show_bug.cgi?id=37388
3895
3896         The split is a bit of a guess.  We might have to adjust things once we
3897         actually have a second port to work with.
3898
3899         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
3900         * Scripts/webkitpy/layout_tests/port/mac.py:
3901         * Scripts/webkitpy/layout_tests/port/webkit.py: Added.
3902         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
3903
3904 2010-04-10  Adam Barth  <abarth@webkit.org>
3905
3906         Reviewed by Eric Seidel.
3907
3908         run_webkit_tests.py shouldn't have platform-specific logic
3909         https://bugs.webkit.org/show_bug.cgi?id=37387
3910
3911         Dirk Pranke pointed out that my last patch was wrong because I
3912         introduced platform-specific logic into run_webkit_tests.py, limiting
3913         the parallelism in Chromium to work around a bug in the main Mac port.
3914
3915         * Scripts/webkitpy/common/system/executive.py:
3916             - Fix a typo pointed out by Chris Jerdonek.
3917         * Scripts/webkitpy/layout_tests/port/base.py:
3918         * Scripts/webkitpy/layout_tests/port/mac.py:
3919         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3920
3921 2010-04-10  Robert Hogan  <robert@webkit.org>
3922
3923         Unreviewed fix to regressions in r57416.
3924
3925         [Qt] Fix regressions in http/tests/navigation from r57416
3926
3927         Reset willSendRequestReturnsNullOnRedirect after each test to
3928         prevent it leaking to subsequent tests.
3929
3930         Error pointed out by Jakub Wieczorek.
3931
3932         https://bugs.webkit.org/show_bug.cgi?id=37237
3933
3934         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3935         (LayoutTestController::reset):
3936
3937 2010-04-11  Robert Hogan  <robert@webkit.org>
3938
3939         Reviewed by Simon Hausmann.
3940
3941         [Qt] Update layoutTestController.DumpResourceLoadCallbacks to match other ports.
3942
3943         Add support for layoutTestController.setWillSendRequestReturnsNullOnRedirect to Qt DRT.
3944         Prevent dumping resource load callbacks once layout test has dumped.
3945
3946         https://bugs.webkit.org/show_bug.cgi?id=37237
3947
3948         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
3949         (WebCore::DumpRenderTree::dump):
3950         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3951         (LayoutTestController::setWillSendRequestReturnsNullOnRedirect):
3952         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3953
3954 2010-04-10  Laszlo Gombos  <laszlo.1.gombos@nokia.com>
3955
3956         Reviewed by Kenneth Rohde Christiansen.
3957
3958         [Qt] Implement Desktop Notifications API for QtWebKit
3959         https://bugs.webkit.org/show_bug.cgi?id=35503
3960
3961         DRT stubs for notification. Enables to run and pass
3962         3 (currently disabled) tests.
3963
3964         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3965         (LayoutTestController::LayoutTestController):
3966         (LayoutTestController::grantDesktopNotificationPermission):
3967         (LayoutTestController::checkDesktopNotificationPermission):
3968         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3969
3970 2010-04-09  Dumitru Daniliuc  <dumi@chromium.org>
3971
3972         Unreviewed, another change to executive.py to make it run with
3973         python 2.4.
3974
3975         * Scripts/webkitpy/common/system/executive.py:
3976
3977 2010-04-09  Dumitru Daniliuc  <dumi@chromium.org>
3978
3979         Unreviewed, attempting to make executive.py run with python 2.4
3980         (which is still used on Chromium's Windows canary bot).
3981
3982         * Scripts/webkitpy/common/system/executive.py:
3983
3984 2010-04-09  Kevin Watters  <kevinwatters@gmail.com>
3985
3986         Reviewed by Eric Seidel.
3987
3988         [wx] Basic implementation of SVG support for wx port.
3989
3990         * wx/build/settings.py:
3991
3992 2010-04-09  Adam Barth  <abarth@webkit.org>
3993
3994         Reviewed by Eric Seidel.
3995
3996         Don't reinvent Executive.cpu_count for every port
3997         https://bugs.webkit.org/show_bug.cgi?id=37377
3998
3999         mac.py and chromium_mac.py had some copy/paste code.  This code doesn't
4000         actually have anything to do with WebKit ports.  It's really just
4001         something in the multiprocessing package.  The lame bit is that package
4002         isn't available in older versions of Python, so we need to implement a
4003         fallback.  However, we already have the fallback in common.  We don't
4004         need to reinvent it specificly for layout_tests.
4005
4006         * Scripts/webkitpy/common/system/executive.py:
4007         * Scripts/webkitpy/layout_tests/port/base.py:
4008         * Scripts/webkitpy/layout_tests/port/chromium.py:
4009         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
4010         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
4011         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
4012         * Scripts/webkitpy/layout_tests/port/mac.py:
4013         * Scripts/webkitpy/layout_tests/port/test.py:
4014         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
4015
4016 2010-04-09  Adam Barth  <abarth@webkit.org>
4017
4018         Reviewed by Eric Seidel.
4019
4020         new-run-webkit-tests --release fails to build release DRT when global configuration is Debug
4021         https://bugs.webkit.org/show_bug.cgi?id=37376
4022
4023         We need to explicitly pass the --release flag.  I bet there are more
4024         instances of this bug.
4025
4026         * Scripts/webkitpy/layout_tests/port/base.py:
4027         * Scripts/webkitpy/layout_tests/port/mac.py:
4028
4029 2010-04-09  Tony Chang  <tony@chromium.org>
4030
4031         Reviewed by Eric Seidel.
4032
4033         [chromium] DRT compile fix on win/linux
4034         https://bugs.webkit.org/show_bug.cgi?id=37314
4035
4036         Looks like this was missed when upstreaming.
4037
4038         * DumpRenderTree/chromium/EventSender.cpp:
4039         (applyKeyModifier):
4040
4041 2010-04-09  Adam Barth  <abarth@webkit.org>
4042
4043         Unreviewed, but approved by Dirk Pranke.
4044
4045         rename test_expectations_test.py to test_expectations_unittest.py so it actually gets run
4046         https://bugs.webkit.org/show_bug.cgi?id=37372
4047
4048         We need to end unit tests with _unittest.py for them to be autodetected
4049         by the test harness.  +6 tests.
4050
4051         * Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py: Renamed from WebKitTools/Scripts/webkitpy/layout_tests/layout_package/test_expectations_test.py.
4052
4053 2010-04-09  Adam Barth  <abarth@webkit.org>
4054
4055         Reviewed by Eric Seidel.
4056
4057         new-run-webkit-tests should talk about DumpRenderTree not test_shell
4058         https://bugs.webkit.org/show_bug.cgi?id=37371
4059
4060         test_shell is some strange Chromium thing.
4061         DumpRenderTree (tm) is the real deal.
4062
4063         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py: Added.
4064         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
4065         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
4066         * Scripts/webkitpy/layout_tests/layout_package/test_shell_thread.py: Removed.
4067         * Scripts/webkitpy/layout_tests/port/chromium.py:
4068         * Scripts/webkitpy/layout_tests/port/server_process.py:
4069         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
4070
4071 2010-04-09  Zoltan Horvath  <zoltan@webkit.org>
4072
4073         Reviewed by Alexey Proskuryakov.
4074
4075         Fix 2 issues (what were introduced in r56524) in svn-create-patch's generateDiff()
4076         https://bugs.webkit.org/show_bug.cgi?id=32582
4077
4078         Add missing return variable. Initialize $patch variable and remove unnecessary condition.
4079
4080         * Scripts/svn-create-patch:
4081
4082 2010-04-09  Kevin Ollivier  <kevino@theolliviers.com>
4083
4084         [wx] Build fix after addition of LayoutTestController method.
4085
4086         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
4087         (LayoutTestController::computedStyleIncludingVisitedInfo):
4088
4089 2010-04-09  Sam Weinig  <sam@webkit.org>
4090
4091         Reviewed by Anders Carlsson.
4092
4093         Fix for https://bugs.webkit.org/show_bug.cgi?id=37368
4094         Add MiniBrowser.
4095
4096         * MiniBrowser: Added.
4097         * MiniBrowser/MiniBrowser.vcproj: Added.
4098         * MiniBrowser/MiniBrowser.xcodeproj: Added.
4099         * MiniBrowser/MiniBrowser.xcodeproj/project.pbxproj: Added.
4100         * MiniBrowser/mac: Added.
4101         * MiniBrowser/mac/AppDelegate.h: Added.
4102         * MiniBrowser/mac/AppDelegate.m: Added.
4103         * MiniBrowser/mac/BrowserWindowController.h: Added.
4104         * MiniBrowser/mac/BrowserWindowController.m: Added.
4105         * MiniBrowser/mac/English.lproj: Added.
4106         * MiniBrowser/mac/English.lproj/BrowserWindow.xib: Added.
4107         * MiniBrowser/mac/English.lproj/InfoPlist.strings: Added.
4108         * MiniBrowser/mac/English.lproj/MainMenu.xib: Added.
4109         * MiniBrowser/mac/MiniBrowser-Info.plist: Added.
4110         * MiniBrowser/mac/MiniBrowser_Prefix.pch: Added.
4111         * MiniBrowser/mac/main.m: Added.
4112         * MiniBrowser/win: Added.
4113         * MiniBrowser/win/BrowserView.cpp: Added.
4114         * MiniBrowser/win/BrowserView.h: Added.
4115         * MiniBrowser/win/BrowserWindow.cpp: Added.
4116         * MiniBrowser/win/BrowserWindow.h: Added.
4117         * MiniBrowser/win/MiniBrowser.cpp: Added.
4118         * MiniBrowser/win/MiniBrowser.h: Added.
4119         * MiniBrowser/win/MiniBrowser.rc: Added.
4120         * MiniBrowser/win/Resources: Added.
4121         * MiniBrowser/win/main.cpp: Added.
4122         * MiniBrowser/win/resource.h: Added.
4123         * MiniBrowser/win/stdafx.cpp: Added.
4124         * MiniBrowser/win/stdafx.h: Added.
4125
4126 2010-04-09  Eric Seidel  <eric@webkit.org>
4127
4128         Reviewed by Adam Barth.
4129
4130         webkit-patch attached my patch to the wrong bug
4131         https://bugs.webkit.org/show_bug.cgi?id=37015
4132
4133         The problem here is that SVN was violating SCM's implicit
4134         contract of always returning paths relative to the repository root.
4135         That can easily be fixed by telling SVN that the CWD is the repository root.
4136
4137         When fixing this I realized there are a large number of places in SCM.py where
4138         we want to consider explicitly passing self.checkout_root as the CWD.
4139         That would allow scm methods to be executed even when the CWD is not inside
4140         the scm tree at all, and would also make sure (in the case of SVN) that paths
4141         returned are relative to the root.  Git (almost always) returns paths relative
4142         to the repository root.
4143
4144         * Scripts/webkitpy/common/checkout/scm.py:
4145          - Explicitly pass self.checkout_root as cwd in run_status_and_extract_filenames
4146          - Add a ton of FIXMEs about the need to go back and decide which methods require cwd=self.checkout_root
4147            and which do not.  We'll probably add a helper function to scm (likely SCM._run) which
4148            always passes cwd=self.checkout_root to Executive.run_command
4149         * Scripts/webkitpy/common/checkout/scm_unittest.py:
4150          - Add a test for this change.
4151         * Scripts/webkitpy/tool/commands/upload.py:
4152          - Removed the explicit os.chdir to the repository root, since scm.py methods
4153            should be robust against the cwd not being equal to the root.
4154
4155 2010-04-09  Adam Roben  <aroben@apple.com>
4156
4157         Don't return 0 as a JSValueRef
4158
4159         That is an illegal use of the JSC API.
4160
4161         Fixes <http://webkit.org/b/37333> REGRESSION (r57292): :visited tests
4162         are asserting on debug Windows and GTK builds
4163
4164         Reviewed by Anders Carlsson.
4165
4166         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
4167   &n