2011-07-05 Eric Seidel <eric@webkit.org>
[WebKit-https.git] / Tools / ChangeLog
1 2011-07-05  Eric Seidel  <eric@webkit.org>
2
3         new-run-webkit-tests fails to start http server if one is already running
4         https://bugs.webkit.org/show_bug.cgi?id=63956
5
6         Reviewed by Adam Barth.
7
8         * BuildSlaveSupport/kill-old-processes:
9          - Removed ^M line endings and sorted the process names.
10          - There are no functional changes to this file.
11            (except that now that its using unix line endings it can be executed directly!)
12         * Scripts/webkitpy/layout_tests/servers/apache_http_server.py:
13          - Change the name to 'httpd' to match old-run-webkit-test paths.
14
15 2011-07-05  Adam Barth  <abarth@webkit.org>
16
17         Add basic ajax support to garden-o-matic
18         https://bugs.webkit.org/show_bug.cgi?id=63874
19
20         Reviewed by Eric Seidel.
21
22         This patch adds some basic infrastructure to garden-o-matic.
23         Currently, the infrastructure is just used to make the "quit" command
24         use Ajax, but in the future, this infrastructure will be used to do
25         more sophistocated remote proceedure calls.
26
27         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html:
28         * Scripts/webkitpy/tool/servers/data/gardeningserver/main.js: Added.
29         * Scripts/webkitpy/tool/servers/gardeningserver.py:
30         * Scripts/webkitpy/tool/servers/reflectionhandler.py:
31
32 2011-07-05  Adam Barth  <abarth@webkit.org>
33
34         Add trivial garden-o-matic command (with server)
35         https://bugs.webkit.org/show_bug.cgi?id=63872
36
37         Reviewed by Eric Seidel.
38
39         This patch adds a trivial "Hello, world" garden-o-matic command,
40         complete with an HTTP server.  This command re-uses a bunch of code
41         from the existing rebaseline-server command.  Over time, this command
42         will grow to be a tool useful for gardening the WebKit tree.
43
44         * Scripts/webkitpy/tool/commands/__init__.py:
45         * Scripts/webkitpy/tool/commands/abstractservercommand.py: Added.
46         * Scripts/webkitpy/tool/commands/gardenomatic.py: Added.
47         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
48         * Scripts/webkitpy/tool/servers/data/gardeningserver/index.html: Added.
49         * Scripts/webkitpy/tool/servers/gardeningserver.py: Added.
50         * Scripts/webkitpy/tool/servers/rebaselineserver.py:
51
52 2011-07-05  Adam Roben  <aroben@apple.com>
53
54         Make prepare-ChangeLog include modified Perl functions in its ChangeLog template
55
56         This is a very simple first cut. Functions must start with a line that starts with "sub "
57         and end with a line that starts with a closing brace. No leading whitespace is allowed.
58         Package names aren't parsed at all.
59
60         Fixes <http://webkit.org/b/21591> prepare-ChangeLog should know how to find functions in
61         Perl files
62
63         Reviewed by David Kilzer.
64
65         * Scripts/prepare-ChangeLog: Removed redundant code that ignored certain files when
66         searching for function line ranges. This is already done inside the get_function_line_ranges
67         function.
68         (get_function_line_ranges): Cleaned up coding style a little bit. Call
69         get_function_line_ranges_for_perl for files with .pl and .pm extensions. For files with an
70         unknown extension or no extension, read the shebang line to try to determine the script
71         interpreter. Call get_function_line_ranges_for_perl if the interpreter seems to be Perl.
72         (get_function_line_ranges_for_perl): Added. Does extremely basic parsing of the file to find
73         lines starting with "sub " or "}".
74
75 2011-07-05  Adam Roben  <aroben@apple.com>
76
77         Clean up run-api-tests output on Windows
78
79         We were mixing run-api-tests output with gtest output, and the result was a mess.
80
81         Fixes <http://webkit.org/b/63954> run-api-tests output is very confusing on Windows
82
83         Reviewed by Dan Bates.
84
85         * Scripts/run-api-tests:
86         (runTest):
87         (populateTests):
88         Made Windows use the formerly-Mac-only codepaths which correctly suppress gtest output
89         except in verbose mode. Most changes are due to indentation. Use of the arch utility is now
90         the only Mac-specific thing in these functions. Some other platform-specific code was moved
91         from here...
92
93         (prepareEnvironmentForRunningTestTool):
94         (testToolPath):
95         ...to here.
96
97 2011-07-04  Gavin Barraclough  <barraclough@apple.com>
98
99         Reviewed by Sam Weinig.
100
101         https://bugs.webkit.org/show_bug.cgi?id=16652
102         Firefox and JavaScriptCore differ in Number.toString(integer)
103
104         Added forwarding header.
105
106         * DumpRenderTree/ForwardingHeaders/wtf/StdLibExtras.h: Added.
107
108 2011-07-04  Anders Carlsson  <andersca@apple.com>
109
110         Try to fix Qt build.
111
112         * DumpRenderTree/TestNetscapePlugIn/Tests/NPRuntimeRemoveProperty.cpp:
113
114 2011-07-04  Anders Carlsson  <andersca@apple.com>
115
116         NP_RemoveProperty is not called back by Safari when delete npObject.prop is encountered in JavaScript
117         https://bugs.webkit.org/show_bug.cgi?id=63915
118         <rdar://problem/7124300>
119
120         Reviewed by Sam Weinig.
121
122         Extend the NPRuntimeRemoveProperty to handle delete object.property from JavaScript.
123
124         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
125         (PluginTest::NPN_IdentifierIsString):
126         (PluginTest::NPN_UTF8FromIdentifier):
127         (PluginTest::NPN_IntFromIdentifier):
128         (PluginTest::NPN_RetainObject):
129         (PluginTest::NPN_ReleaseObject):
130         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
131         (PluginTest::Object::removeProperty):
132         (PluginTest::Object::identifierIs):
133         (PluginTest::Object::NP_RemoveProperty):
134         (PluginTest::Object::npClass):
135         * DumpRenderTree/TestNetscapePlugIn/Tests/NPRuntimeRemoveProperty.cpp:
136         (NPRuntimeRemoveProperty::TestObject::TestObject):
137         (NPRuntimeRemoveProperty::TestObject::hasProperty):
138         (NPRuntimeRemoveProperty::TestObject::getProperty):
139         (NPRuntimeRemoveProperty::TestObject::removeProperty):
140         (NPRuntimeRemoveProperty::PluginObject::PluginObject):
141         (NPRuntimeRemoveProperty::PluginObject::~PluginObject):
142         (NPRuntimeRemoveProperty::PluginObject::hasMethod):
143         (NPRuntimeRemoveProperty::PluginObject::invoke):
144         (NPRuntimeRemoveProperty::PluginObject::hasProperty):
145         (NPRuntimeRemoveProperty::PluginObject::getProperty):
146         (NPRuntimeRemoveProperty::NPP_GetValue):
147
148 2011-07-04  Leandro Pereira  <leandro@profusion.mobi>
149
150         Reviewed by Kent Tamura.
151
152         [EFL] Add LayoutTestControllerEfl.
153         https://bugs.webkit.org/show_bug.cgi?id=61974
154
155         Only the core features are implemented. Most functions are actually
156         stubs and will be implemented as soon as its features are needed.
157         Build system changes will be made as soon as DRT is completely
158         upstreamed.
159
160         * DumpRenderTree/efl/LayoutTestControllerEfl.cpp: Added.
161         (LayoutTestController::~LayoutTestController):
162         (LayoutTestController::addDisallowedURL):
163         (LayoutTestController::clearBackForwardList):
164         (LayoutTestController::copyDecodedHostName):
165         (LayoutTestController::copyEncodedHostName):
166         (LayoutTestController::dispatchPendingLoadRequests):
167         (LayoutTestController::display):
168         (LayoutTestController::counterValueForElementById):
169         (LayoutTestController::keepWebHistory):
170         (LayoutTestController::computedStyleIncludingVisitedInfo):
171         (LayoutTestController::nodesFromRect):
172         (LayoutTestController::layerTreeAsText):
173         (LayoutTestController::pageNumberForElementById):
174         (LayoutTestController::numberOfPages):
175         (LayoutTestController::pageProperty):
176         (LayoutTestController::isPageBoxVisible):
177         (LayoutTestController::pageSizeAndMarginsInPixels):
178         (LayoutTestController::webHistoryItemCount):
179         (LayoutTestController::workerThreadCount):
180         (LayoutTestController::notifyDone):
181         (LayoutTestController::pathToLocalResource):
182         (LayoutTestController::queueLoad):
183         (LayoutTestController::setAcceptsEditing):
184         (LayoutTestController::setAlwaysAcceptCookies):
185         (LayoutTestController::setCustomPolicyDelegate):
186         (LayoutTestController::waitForPolicyDelegate):
187         (LayoutTestController::setScrollbarPolicy):
188         (LayoutTestController::addOriginAccessWhitelistEntry):
189         (LayoutTestController::removeOriginAccessWhitelistEntry):
190         (LayoutTestController::setMainFrameIsFirstResponder):
191         (LayoutTestController::setTabKeyCyclesThroughElements):
192         (LayoutTestController::setUseDashboardCompatibilityMode):
193         (LayoutTestController::setUserStyleSheetEnabled):
194         (LayoutTestController::setUserStyleSheetLocation):
195         (LayoutTestController::setValueForUser):
196         (LayoutTestController::setViewModeMediaFeature):
197         (LayoutTestController::setWindowIsKey):
198         (LayoutTestController::setSmartInsertDeleteEnabled):
199         (waitToDumpWatchdogFired):
200         (LayoutTestController::setWaitToDump):
201         (LayoutTestController::windowCount):
202         (LayoutTestController::setPrivateBrowsingEnabled):
203         (LayoutTestController::setJavaScriptCanAccessClipboard):
204         (LayoutTestController::setXSSAuditorEnabled):
205         (LayoutTestController::setFrameFlatteningEnabled):
206         (LayoutTestController::setSpatialNavigationEnabled):
207         (LayoutTestController::setAllowUniversalAccessFromFileURLs):
208         (LayoutTestController::setAllowFileAccessFromFileURLs):
209         (LayoutTestController::setAuthorAndUserStylesEnabled):
210         (LayoutTestController::setAutofilled):
211         (LayoutTestController::disableImageLoading):
212         (LayoutTestController::setMockDeviceOrientation):
213         (LayoutTestController::setMockGeolocationPosition):
214         (LayoutTestController::setMockGeolocationError):
215         (LayoutTestController::setGeolocationPermission):
216         (LayoutTestController::numberOfPendingGeolocationPermissionRequests):
217         (LayoutTestController::addMockSpeechInputResult):
218         (LayoutTestController::setIconDatabaseEnabled):
219         (LayoutTestController::setJavaScriptProfilingEnabled):
220         (LayoutTestController::setSelectTrailingWhitespaceEnabled):
221         (LayoutTestController::setPopupBlockingEnabled):
222         (LayoutTestController::setPluginsEnabled):
223         (LayoutTestController::elementDoesAutoCompleteForElementWithId):
224         (LayoutTestController::execCommand):
225         (LayoutTestController::findString):
226         (LayoutTestController::isCommandEnabled):
227         (LayoutTestController::setCacheModel):
228         (LayoutTestController::setPersistentUserStyleSheetLocation):
229         (LayoutTestController::clearPersistentUserStyleSheet):
230         (LayoutTestController::clearAllApplicationCaches):
231         (LayoutTestController::setApplicationCacheOriginQuota):
232         (LayoutTestController::clearApplicationCacheForOrigin):
233         (LayoutTestController::localStorageDiskUsageForOrigin):
234         (LayoutTestController::originsWithApplicationCache):
235         (LayoutTestController::applicationCacheDiskUsageForOrigin):
236         (LayoutTestController::clearAllDatabases):
237         (LayoutTestController::setDatabaseQuota):
238         (LayoutTestController::originsWithLocalStorage):
239         (LayoutTestController::deleteAllLocalStorage):
240         (LayoutTestController::deleteLocalStorageForOrigin):
241         (LayoutTestController::observeStorageTrackerNotifications):
242         (LayoutTestController::syncLocalStorage):
243         (LayoutTestController::setDomainRelaxationForbiddenForURLScheme):
244         (LayoutTestController::goBack):
245         (LayoutTestController::setDefersLoading):
246         (LayoutTestController::setAppCacheMaximumSize):
247         (LayoutTestController::pauseAnimationAtTimeOnElementWithId):
248         (LayoutTestController::pauseTransitionAtTimeOnElementWithId):
249         (LayoutTestController::sampleSVGAnimationForElementAtTime):
250         (LayoutTestController::numberOfActiveAnimations):
251         (LayoutTestController::suspendAnimations):
252         (LayoutTestController::resumeAnimations):
253         (LayoutTestController::overridePreference):
254         (LayoutTestController::addUserScript):
255         (LayoutTestController::addUserStyleSheet):
256         (LayoutTestController::setDeveloperExtrasEnabled):
257         (LayoutTestController::setAsynchronousSpellCheckingEnabled):
258         (LayoutTestController::showWebInspector):
259         (LayoutTestController::closeWebInspector):
260         (LayoutTestController::evaluateInWebInspector):
261         (LayoutTestController::evaluateScriptInIsolatedWorld):
262         (LayoutTestController::removeAllVisitedLinks):
263         (LayoutTestController::callShouldCloseOnWebView):
264         (LayoutTestController::apiTestNewWindowDataLoadBaseURL):
265         (LayoutTestController::apiTestGoToCurrentBackForwardItem):
266         (LayoutTestController::setWebViewEditable):
267         (LayoutTestController::markerTextForListItem):
268         (LayoutTestController::authenticateSession):
269         (LayoutTestController::setEditingBehavior):
270         (LayoutTestController::abortModal):
271         (LayoutTestController::hasSpellingMarker):
272         (LayoutTestController::hasGrammarMarker):
273         (LayoutTestController::dumpConfigurationForViewport):
274         (LayoutTestController::setSerializeHTTPLoads):
275         (LayoutTestController::setMinimumTimerInterval):
276         (LayoutTestController::setTextDirection):
277         (LayoutTestController::allowRoundingHacks):
278
279 2011-06-24  Vsevolod Vlasov  <vsevik@chromium.org>
280
281         Unreviewed, add myself as committer.
282
283         * Scripts/webkitpy/common/config/committers.py:
284
285 2011-07-03  Sheriff Bot  <webkit.review.bot@gmail.com>
286
287         Unreviewed, rolling out r90347.
288         http://trac.webkit.org/changeset/90347
289         https://bugs.webkit.org/show_bug.cgi?id=63886
290
291         Build breaks on Leopard, Chromium-win, WinCairo, and WinCE.
292         (Requested by tkent on #webkit).
293
294         * DumpRenderTree/ForwardingHeaders/wtf/StdLibExtras.h: Removed.
295
296 2011-07-03  Daniel Bates  <dbates@webkit.org>
297
298         Import Checkout in scm_unittests.py and logging in svn.py to fix
299         Python missing global variable errors
300
301         Rubber-stamped by Adam Barth.
302
303         * Scripts/webkitpy/common/checkout/scm/scm_unittest.py: Import Checkout.
304         * Scripts/webkitpy/common/checkout/scm/svn.py: Import logging.
305
306 2011-07-02  Gavin Barraclough  <barraclough@apple.com>
307
308         Reviewed by Sam Weinig.
309
310         https://bugs.webkit.org/show_bug.cgi?id=16652
311         Firefox and JavaScriptCore differ in Number.toString(integer)
312
313         Added forwarding header.
314
315         * DumpRenderTree/ForwardingHeaders/wtf/StdLibExtras.h: Added.
316
317 2011-04-02  Robert Hogan  <robert@webkit.org>
318
319         Reviewed by Benjamin Poulain.
320
321         [Qt] Support third-party cookie policy for Qt clients
322         https://bugs.webkit.org/show_bug.cgi?id=45455
323
324         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
325         (LayoutTestController::reset):
326         (LayoutTestController::setAlwaysAcceptCookies):
327         (LayoutTestController::setAlwaysBlockCookies):
328         * DumpRenderTree/qt/LayoutTestControllerQt.h:
329
330 2011-07-03  Adam Barth  <abarth@webkit.org>
331
332         Factor ReflectionHandler out of the rebaseline server.  Just code motion.
333
334         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
335         * Scripts/webkitpy/tool/servers/rebaselineserver.py:
336         * Scripts/webkitpy/tool/servers/reflectionhandler.py: Added.
337
338 2011-07-03  Adam Barth  <abarth@webkit.org>
339
340         Move the rebaseline server out of the commands package and into the
341         (new) servers package.  This patch prepares for refactoring this code
342         to share server infrastructure with the (forthcoming) garden-o-matic
343         command.
344
345         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
346         * Scripts/webkitpy/tool/servers/__init__.py: Added.
347         * Scripts/webkitpy/tool/servers/data/rebaselineserver/loupe.js:
348         * Scripts/webkitpy/tool/servers/data/rebaselineserver/main.css:
349         * Scripts/webkitpy/tool/servers/data/rebaselineserver/queue.js:
350         * Scripts/webkitpy/tool/servers/rebaselineserver.py: Added.
351
352 2011-07-01  Adam Barth  <abarth@webkit.org>
353
354         Fix typo in master.cfg, hoping to trigger a master restart.
355
356         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
357
358 2011-07-01  Adam Barth  <abarth@webkit.org>
359
360         Turn off new-run-webkit-tests and remove some debug logging.
361
362         * Scripts/run-webkit-tests:
363         * Scripts/webkitpy/layout_tests/port/http_lock.py:
364
365 2011-07-01  Adam Barth  <abarth@webkit.org>
366
367         Add some more debug logging to try to understand why we're getting
368         stuck acquring the HTTP lock.
369
370         * Scripts/webkitpy/layout_tests/port/http_lock.py:
371
372 2011-07-01  Adam Barth  <abarth@webkit.org>
373
374         Refactor NRWT parsing logic in master.cfg to understand that run-webkit-tests might run NRWT
375         https://bugs.webkit.org/show_bug.cgi?id=63854
376
377         Reviewed by Dirk Pranke.
378
379         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
380
381 2011-07-01  Eric Seidel  <eric@webkit.org>
382
383         new-run-webkit-tests results does not understand that mac uses test_expectations files
384         https://bugs.webkit.org/show_bug.cgi?id=63838
385
386         Reviewed by Adam Barth.
387
388         We've decided to "enable" test_expectations for the Mac port for now.
389         It makes the results.html results much more confusing to read, but at least they're
390         no longer lying to us.
391
392         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
393         * Scripts/webkitpy/layout_tests/port/base.py:
394         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
395
396 2011-07-01  Adam Barth  <abarth@webkit.org>
397
398         Temporarily disable waiting for ReportCrash to see how this code
399         affects performance on the bots.
400
401         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
402         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
403
404 2011-07-01  Dmitry Lomov  <dslomov@google.com>
405
406         Reviewed by Adam Roben.
407
408         https://bugs.webkit.org/show_bug.cgi?id=63818
409         Test bots shouldn't attempt to build TestWebKitAPI on Windows.
410         TestWebKitAPI builds as a part of general webkit build and is a
411         part of build products archive.
412
413         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
414         * Scripts/run-api-tests:
415
416 2011-07-01  Adam Barth  <abarth@webkit.org>
417
418         We shouldn't wait for ReportCrash while the timeout is running
419         https://bugs.webkit.org/show_bug.cgi?id=63837
420
421         Reviewed by Eric Seidel.
422
423         Instead, we should move this check to worker.py.
424
425         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
426         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
427
428 2011-07-01  Adam Roben  <aroben@apple.com>
429
430         Stop setting unsafe headers on POSTs via XHR from TestFailures page
431
432         Fixes <http://webkit.org/b/63833> TestFailures page causes a bunch of spew in the
433         Inspector's console about unsafe headers
434
435         Reviewed by Dan Bates.
436
437         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
438         (fetchResource): Removed unnecessary code to set Content-Length and Connection headers.
439
440 2011-07-01  Eric Seidel  <eric@webkit.org>
441
442         Reviewed by Adam Barth.
443
444         Switch Snow Leopard over to new-run-webkit-tests (except for wk2)
445         https://bugs.webkit.org/show_bug.cgi?id=63786
446
447         We're down to one bug blocking the "move to NRWT" bug:
448         https://bugs.webkit.org/show_bug.cgi?id=34984
449         so we're trying transitioning Snow Leopard to NRWT as an experiment.
450
451         If we run into any blocking issues we'll of course roll this back out
452         (or anyone reading this should feel encouraged to).
453
454         * Scripts/run-webkit-tests:
455
456 2011-07-01  Adam Barth  <abarth@webkit.org>
457
458         Add a compat shim for --leaks.  Support for --leaks will be added in
459         https://bugs.webkit.org/show_bug.cgi?id=63832.
460
461         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
462
463 2011-07-01  Adam Barth  <abarth@webkit.org>
464
465         Reviewed by Darin Adler.
466
467         ReportCrash destabilizes new-run-webkit-tests
468         https://bugs.webkit.org/show_bug.cgi?id=63795
469
470         ReportCrash chews up too many system resources that it destablizes
471         tests run concurrently.  This patch causes us to spin down all the
472         worker processes when ReportCrash is running.
473
474         Also, this patch causes the master process to tell the user we're
475         waiting for the crash reporter rather than having that be the worker's
476         job.  This stops the user from getting spammed with too many "waiting
477         for crash reporter" messages and also cleans up some minor UI issues
478         with the meter.
479
480         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
481         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
482         * Scripts/webkitpy/layout_tests/port/base.py:
483         * Scripts/webkitpy/layout_tests/port/server_process.py:
484
485 2011-07-01  Adam Roben  <aroben@apple.com>
486
487         Teach webkitpy about the new format of our ChangeLog template
488
489         r90229 moved the "Need a short description and bug URL" line to the top of the ChangeLog
490         template. But webkitpy didn't know this, so the various rollout-related commands were
491         leaving that line in the ChangeLog.
492
493         Fixes <http://webkit.org/b/63815> REGRESSION (r90229): webkit-patch rollout and sheriffbot
494         rollout are broken
495
496         Reviewed by Dan Bates.
497
498         * Scripts/webkitpy/common/checkout/changelog.py:
499         (ChangeLog.update_with_unreviewed_message): Changed to look for the "Need a short
500         description and bug URL" line instead of the "Reviewed by" line as the first line of the
501         ChangeLog template.
502
503         * Scripts/webkitpy/common/checkout/changelog_unittest.py:
504         (ChangeLogTest): Updated _new_entry_boilerplate to match what prepare-ChangeLog now
505         generates.
506
507 2011-07-01  Adam Roben  <aroben@apple.com>
508
509         Make TestFailures stop loading old builds once we've explained all current failures
510
511         We used to do this, but some refactoring in r90165 broke this logic.
512
513         Fixes <http://webkit.org/b/63808> REGRESSION (r90165): TestFailures page always loads all
514         builds back to the beginning of time
515
516         Reviewed by Dan Bates.
517
518         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
519         (LayoutTestHistoryAnalyzer.prototype.start): Stop fetching old builds once we're told that
520         all current failures have been explained.
521
522 2011-07-01  Adam Roben  <aroben@apple.com>
523
524         commit-log-editor reorders ChangeLog entries in unexpected ways
525         https://bugs.webkit.org/show_bug.cgi?id=63804
526
527         Reviewed by Darin Adler.
528
529         commit-log-editor was changed in r46899 to move the bug title and URL to the top of the
530         commit message in order to make git-based tools, which expect the first line of the commit
531         message to be a summary of the change, to work better. But this reordering of the ChangeLog
532         entry was unpredictable and unexpected.
533
534         I changed prepare-ChangeLog to put the bug title and URL at the top of the ChangeLog entry
535         so that commit-log-editor doesn't have to move them. Then I changed commit-log-editor not to
536         try to move them.
537
538         * Scripts/commit-log-editor: Partially reverted r46899. We no longer save the "Reviewed by"
539         line and try to move it down below the bug title and URL. The order of the text in the
540         ChangeLog is now preserved, except that we insert a "Patch by" line just above the "Reviewed
541         by" line in cases where the committer and author are not the same person.
542
543         * Scripts/prepare-ChangeLog: Moved the "Reviewed by" line down below the bug title and URL.
544
545 2011-07-01  Mihnea Ovidenie  <mihnea@adobe.com>
546
547         Reviewed by Adam Roben.
548
549         [CSSRegions] Add build slave
550         https://bugs.webkit.org/show_bug.cgi?id=62372
551
552         * BuildSlaveSupport/build.webkit.org-config/config.json:
553         Add an entry for a release build of Mac.
554
555 2011-07-01  Sheriff Bot  <webkit.review.bot@gmail.com>
556
557         Unreviewed, rolling out r90213.
558         http://trac.webkit.org/changeset/90213
559         https://bugs.webkit.org/show_bug.cgi?id=63792
560
561         The SL bot has an errant HTTP server holding port 8080
562         (Requested by abarth on #webkit).
563
564         * Scripts/run-webkit-tests:
565
566 2011-07-01  Carlos Garcia Campos  <cgarcia@igalia.com>
567
568         Unreviewed. Fix WebKit2 GTK build after r90163.
569
570         * MiniBrowser/gtk/WebBundle/WebBundleMain.c:
571         (didCreatePage):
572
573 2011-07-01  Adam Barth  <abarth@webkit.org>
574
575         Exit faster when we get a keyboard interrupt.  Previously, we'd open
576         Safari if you ^C NRWT after it found a failure.
577
578         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
579
580 2011-07-01  Eric Seidel  <eric@webkit.org>
581
582         Reviewed by Adam Barth.
583
584         Switch Snow Leopard over to new-run-webkit-tests (except for wk2)
585         https://bugs.webkit.org/show_bug.cgi?id=63786
586
587         We're down to one bug blocking the "move to NRWT" bug:
588         https://bugs.webkit.org/show_bug.cgi?id=34984
589         so we're trying transitioning Snow Leopard to NRWT as an experiment.
590
591         If we run into any blocking issues we'll of course roll this back out
592         (or anyone reading this should feel encouraged to).
593
594         * Scripts/run-webkit-tests:
595
596 2011-07-01  Sheriff Bot  <webkit.review.bot@gmail.com>
597
598         Unreviewed, rolling out r90192.
599         http://trac.webkit.org/changeset/90192
600         https://bugs.webkit.org/show_bug.cgi?id=63788
601
602         Appears to have caused NRWT on Chromium WebKit Vista to hang
603         (Requested by abarth on #webkit).
604
605         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
606         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
607
608 2011-06-30  Adam Barth  <abarth@webkit.org>
609
610         Reviewed by Eric Seidel.
611
612         Remove the concept of "being wedged" from new-run-webkit-tests
613         https://bugs.webkit.org/show_bug.cgi?id=63767
614
615         Worker processes shouldn't ever become wedged.  My understanding is
616         that this code was originally motivated by the old threading-based
617         design but no longer servers any purpose.
618
619         Note: If we actually have a problem with the test harness getting
620         stuck, buildbot will kill us.
621
622         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
623         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py:
624         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
625
626 2011-06-30  Adam Barth  <abarth@webkit.org>
627
628         The win32 implementation of fork is sad about forking test-webkitpy.
629         Hopefully this will heal the bot.
630
631         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
632
633 2011-06-30  Eric Seidel  <eric@webkit.org>
634
635         Reviewed by Adam Barth.
636
637         Update new-run-webkit-tests now that the wk2 shared skipped list has moved
638         https://bugs.webkit.org/show_bug.cgi?id=63780
639
640         The shared list was moved from mac-wk2 to wk2 as part of
641         http://trac.webkit.org/changeset/90125.
642
643         * Scripts/webkitpy/layout_tests/port/webkit.py:
644         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
645
646 2011-06-30  Kenichi Ishibashi  <bashi@chromium.org>
647
648         Reviewed by Darin Adler.
649
650         Style Checker should flag removal of "developmentRegion = English;" from project.pbxproj.
651         https://bugs.webkit.org/show_bug.cgi?id=62022
652
653         Adds XcodeProjectFileChecker to check removal of "developmentRegion".
654
655         * Scripts/webkitpy/style/checker.py: Added XcodeProjectFileChecker as a checker for *.pbxproj.
656         * Scripts/webkitpy/style/checkers/xcodeproj.py: Added.
657         * Scripts/webkitpy/style/checkers/xcodeproj_unittest.py: Added.
658
659 2011-06-30  Dirk Pranke  <dpranke@chromium.org>
660
661         Reviewed by Ojan Vafai.
662
663         nrwt: make sharding tests needing locks less hard-coded
664         https://bugs.webkit.org/show_bug.cgi?id=63112
665
666         This change also changes the manager logic so that it will
667         drop the server lock as soon as all of the shards requiring
668         the lock have completed.
669
670         This change includes some minor namespace/import changes in the
671         unit tests, and also makes the Manager a new-style object, which
672         it should've been all along.
673
674         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
675         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
676
677 2011-06-30  Adam Barth  <abarth@webkit.org>
678
679         Reviewed by Eric Seidel.
680
681         Remove threaded mode from new-run-webkit-tests
682         https://bugs.webkit.org/show_bug.cgi?id=63771
683
684         This mode is not used and is buggy.  Rather than carry around a bunch
685         of unused buggy code, we should rip it out and focus on the
686         multiprocess implementation.
687
688         * Scripts/webkitpy/common/system/executive.py:
689         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker.py:
690         * Scripts/webkitpy/layout_tests/layout_package/manager_worker_broker_unittest.py:
691         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
692         * Scripts/webkitpy/layout_tests/port/base.py:
693         * Scripts/webkitpy/layout_tests/port/chromium.py:
694         * Scripts/webkitpy/layout_tests/port/mac.py:
695         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
696         * Scripts/webkitpy/layout_tests/port/server_process.py:
697         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
698         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
699
700 2011-06-30  Adam Barth  <abarth@webkit.org>
701
702         Reviewed by Darin Adler.
703
704         Clean up output from new-run-webkit-tests
705         https://bugs.webkit.org/show_bug.cgi?id=63759
706
707         Printing messages from the child process looks super ugly because of
708         the way the pretty-printer works.  Printing a blank line first is a
709         hack, but it makes things at least partially sane.
710
711         Also, handle the case where calling sample throws an exception.
712
713         * Scripts/webkitpy/layout_tests/port/server_process.py:
714         * Scripts/webkitpy/layout_tests/port/server_process_unittest.py:
715
716 2011-06-27  Diego Gonzalez  <diegohcg@webkit.org>
717
718         Reviewed by Antonio Gomes.
719
720         [Qt] DRT: Provide LayoutTestController::setDefersLoading and ::goBack support
721         https://bugs.webkit.org/show_bug.cgi?id=63489
722
723         Unskipped:
724             loader/navigation-while-deferring-loads.html
725
726         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
727         (LayoutTestController::goBack):
728         (LayoutTestController::setDefersLoading):
729         * DumpRenderTree/qt/LayoutTestControllerQt.h:
730
731 2011-06-30  Adam Barth  <abarth@webkit.org>
732
733         Reviewed by Dirk Pranke.
734
735         new-run-webkit-tests should stop when I ^C it
736         https://bugs.webkit.org/show_bug.cgi?id=63763
737
738         It still doesn't stop fast enough, but at least with this patch it
739         doesn't run more tests.
740
741         Also, I tweaked the UI that happens when there's a keyboard interrupt
742         to not be amazingly ugly.  There's still a garbage "Te" left on the
743         console, but that's a problem for another day.
744
745         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
746
747 2011-06-30  Adam Roben  <aroben@apple.com>
748
749         Make TestFailures never treat a too-many-failures build as a last-passing build for a test
750
751         Fixes <http://webkit.org/b/63758> TestFailures sometimes incorrectly says that a build was
752         the last passing build for a test, when in fact the test wasn't run due to too many failures
753         causing old-run-webkit-tests to exit early
754
755         Reviewed by Anders Carlsson.
756
757         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
758         (LayoutTestHistoryAnalyzer.prototype._incorporateBuildHistory): If there were too many
759         failures in this build, assume that all the tests failed in the last build would have failed
760         in this build, too, had they been run. Introduced some new local variables to reduce typing,
761         and used them throughout.
762
763         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
764         (LayoutTestResultsLoader.prototype.start): Fix a regression from r89841 where we wouldn't
765         correctly remember whether old-run-webkit-tests exited early due to too many failures for
766         builds older than r89610. Also removed an unused parameter while I was at it.
767
768 2011-06-30  Adam Roben  <aroben@apple.com>
769
770         Make TestFaiulres only load old-results directories as needed
771
772         Fixes <http://webkit.org/b/63752> Tester pages on TestFailures page load very slowly
773
774         Reviewed by Anders Carlsson.
775
776         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
777         (Builder.prototype.getBuildNames): Moved up into the API section of the class. Now just
778         calls through to _getBuildNamesFromResultsDirectory.
779         (Builder.prototype.getOldBuildNames): Added. Just calls through to
780         _getBuildNamesFromResultsDirectory.
781         (Builder.prototype._getBuildNamesFromResultsDirectory): Renamed from getBuildNames. Now
782         takes the directory URL as an argument and only fetches that single URL.
783
784         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
785         (LayoutTestHistoryAnalyzer.prototype.start): Moved most logic to _analyzeBuilds. First
786         analyzes builds from Builder.getBuildNames, then from Builder.getOldBuildNames if needed.
787         (LayoutTestHistoryAnalyzer.prototype._analyzeBuilds): Moved logic here from start. (Most
788         changes are just indentation.) Now takes a callback to call when we've finished analyzing
789         all builds in buildNames so that we can try to fetch more build names if needed.
790
791 2011-06-30  Mark Rowe  <mrowe@apple.com>
792
793         Reviewed by Anders Carlsson.
794
795         Update MiniBrowser and WebKitTestRunner to use the client version constants.
796
797         * MiniBrowser/gtk/BrowserWindow.c:
798         (browserWindowLoaderClientInit):
799         (browserWindowUIClientInit):
800         * MiniBrowser/gtk/WebBundle/WebBundleMain.c:
801         (didCreatePage):
802         (WKBundleInitialize):
803         * MiniBrowser/gtk/main.c:
804         (createWKContextWithInjectedBundle):
805         * MiniBrowser/mac/AppDelegate.m:
806         (-[BrowserAppDelegate init]):
807         * MiniBrowser/mac/BrowserWindowController.m:
808         (-[BrowserWindowController awakeFromNib]):
809         * MiniBrowser/mac/WebBundle/WebBundleMain.m:
810         (WKBundleInitialize):
811         * MiniBrowser/win/BrowserView.cpp:
812         (BrowserView::create):
813         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
814         (WTR::InjectedBundle::initialize):
815         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
816         (WTR::InjectedBundlePage::InjectedBundlePage):
817         * WebKitTestRunner/TestController.cpp:
818         (WTR::TestController::createOtherPage):
819         (WTR::TestController::initialize):
820
821 2011-06-30  Adam Roben  <aroben@apple.com>
822
823         Identify revisions that modified failing tests on TestFailures page
824
825         Fixes <http://webkit.org/b/63716> TestFailures page should try to infer which commit caused
826         a failure by looking at commit logs
827
828         Reviewed by David Kilzer.
829
830         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
831         (.existing-bugs-list, .suspect-revisions-list): Made this rule apply to the suspect
832         revisions list, too.
833
834         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Trac.js:
835         (Trac): Initialize our memory cache.
836         (Trac.prototype.getCommitDataForRevisionRange): Added. Fetches the commit log in RSS form,
837         then parses out the revision number, Trac's idea of the commit title, and the files modified
838         by the commit. (To get the modified files, we rely on the commit message including a
839         prepare-ChangeLog-style file list.)
840         (Trac.prototype.logURL): Added a new formatAsRSS parameter.
841
842         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
843         (removePathExtension): Added. Returns a new string with the last dot and everything after it
844         removed.
845         (sorted): Added sortFunction parameter.
846         (Node.prototype.removeAllChildren): Added. Does what it says.
847
848         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
849         (ViewController.prototype._displayBuilder): Updated for change to _domForRegressionRange.
850         (ViewController.prototype._domForRegressionRange): Changed argument order to match
851         _domForNewAndExistingBugs. Now takes the list of failing tests and uses it to figure out
852         which revisions modified the failing tests and displays the suspect revisions in a list.
853
854 2011-06-30  Eric Seidel  <eric@webkit.org>
855
856         Reviewed by Adam Barth.
857
858         Rename should_kill_dump_render_tree to driver_needs_restart in preparation for supporting WebKitTestRunner
859         https://bugs.webkit.org/show_bug.cgi?id=63743
860
861         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
862         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
863
864 2011-06-30  Martin Robinson  <mrobinson@igalia.com>
865
866         Reviewed by Anders Carlsson.
867
868         [GTK] Crash observed with nspluginwrapper and flash
869         https://bugs.webkit.org/show_bug.cgi?id=62249
870
871         Added a TestNetscapePlugin test which verifies that WebKit properly
872         handles situations where InvalidateRect is called with a null instance.
873
874         * DumpRenderTree/TestNetscapePlugIn/Tests/x11/CallInvalidateRectWithNullNPPArgument.cpp: Added.
875         (CallInvalidateRectWithNullNPPArgument::CallInvalidateRectWithNullNPPArgument):
876         (CallInvalidateRectWithNullNPPArgument::NPP_New):
877         * GNUmakefile.am: Add the new file to sources list.
878
879 2011-06-30  Eric Seidel  <eric@webkit.org>
880
881         Reviewed by Adam Barth.
882
883         Remove redundant tests in TestFailuresTest
884         https://bugs.webkit.org/show_bug.cgi?id=63742
885
886         These tests are all already covered by TestFailuresTest.test_loads.
887         I also added the missing "T" from TestFailuresest. :)
888
889         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
890
891 2011-06-30  Adam Barth  <abarth@webkit.org>
892
893         Reviewed by Eric Seidel.
894
895         REGRESSION(r84294): new-run-webkit-tests results.html generate links to diffs.html or diff.png that don't exist
896         https://bugs.webkit.org/show_bug.cgi?id=63494
897
898         This patch does two things:
899
900         1) Refactor diff_image to return the image diff instead of writing the
901         diff to a file.  This fixes a bunch of hacks and disentangles a bunch
902         of code.
903
904         2) When there's a checksum mismatch but not image diff, we no longer
905         report an IMAGE failure to results.html.  That fixes the bug in
906         question because results.html won't try to link to a non-existent image
907         diff.
908
909         * Scripts/webkitpy/common/system/executive_mock.py:
910         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
911         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py:
912         * Scripts/webkitpy/layout_tests/port/base.py:
913         * Scripts/webkitpy/layout_tests/port/chromium.py:
914         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
915         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
916         * Scripts/webkitpy/layout_tests/port/test.py:
917         * Scripts/webkitpy/layout_tests/port/webkit.py:
918         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
919         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests.py:
920
921 2011-06-28  Hans Wennborg  <hans@chromium.org>
922
923         Reviewed by Tony Chang.
924
925         IndexedDB: Prepare for running layout tests with LevelDB
926         https://bugs.webkit.org/show_bug.cgi?id=63593
927
928         The TestShell must always provide a temporary folder for LevelDB,
929         not just when the backing store type is overridden.
930
931         * DumpRenderTree/chromium/LayoutTestController.cpp:
932         (LayoutTestController::setOverrideIndexedDBBackingStore):
933         (LayoutTestController::clearAllDatabases):
934         * DumpRenderTree/chromium/LayoutTestController.h:
935         * DumpRenderTree/chromium/TestShell.cpp:
936         (TestShell::TestShell):
937         * DumpRenderTree/chromium/TestShell.h:
938
939 2011-06-30  Zsolt Fehér  <h490267@stud.u-szeged.hu>
940
941         Reviewed by Csaba Osztrogonác.
942
943         [WK2] Create a general wk2 specific LayoutTests/platform/wk2 platform
944         qt-wk2, win-wk2, mac-wk2 use this common Skipped list too
945         https://bugs.webkit.org/show_bug.cgi?id=61785 
946
947         * Scripts/old-run-webkit-tests:
948
949 2011-06-30  Adam Roben  <aroben@apple.com>
950
951         Use a colon to separate test names from failure info on TestFailures
952
953         This matches how the "Possibly Flaky Tests" list works, and is much prettier when the
954         failure info itself includes parentheses.
955
956         Fixes <http://webkit.org/b/63722> Crashing symbols on TestFailures page look ugly due to
957         double parentheses
958
959         Reviewed by Darin Adler.
960
961         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
962         (ViewController.prototype._domForFailedTest): Use a colon instead of surrounding the failure
963         info in parentheses.
964
965 2011-06-30  Adam Roben  <aroben@apple.com>
966
967         Robustify PersistentCache against cached data that contains the string ': '
968
969         Fixes <http://webkit.org/b/63721> Exceptions thrown while loading TestFailures page due to
970         passing a bad string to JSON.parse
971
972         Reviewed by Darin Adler.
973
974         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/PersistentCache.js:
975         (PersistentCache._parseDateAndJSONFromString): Manually split the string on the separator
976         using indexOf and substring. (String.prototype.split always splits on all instances of the
977         separator.)
978
979 2011-06-30  Adam Roben  <aroben@apple.com>
980
981         Include the crashing symbol in crash logs links on TestFailures
982
983         Fixes <http://webkit.org/b/63465> Links to crash logs on TestFailures page should include
984         the crashing symbol (like links in results.html do)
985
986         Reviewed by David Kilzer.
987
988         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
989         (LayoutTestResultsLoader.prototype.start): Bumped the cache version because we now store
990         crashing symbols for crashing tests. Renamed testsForResultTable to parseResultTable because
991         it now returns more than just the test names. Specifically, it now looks for crash log links
992         and extracts the crashing symbol name from them. Updated callers of parseResultTable to
993         match its new behavior. Changed to store the crashing symbol along with the failure type in
994         the data we pass to the callback for tests which crashed.
995
996         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
997         (code): Make <code> elements a little smaller because their contents can be quite long.
998
999         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1000         (ViewController.prototype._domForFailureDiagnosis): Include the crashing symbol inside a
1001         <code> element in the link, if there is a crashing symbol.
1002
1003 2011-06-30  Adam Roben  <aroben@apple.com>
1004
1005         Use objects instead of strings to represent a test result in TestFailures code
1006
1007         This will eventually allow us to store more than just the type of failure for each test.
1008         (E.g., we can store the name of the crashing symbol for tests which crashed.)
1009
1010         Prep work for <http://webkit.org/b/63465> Links to crash logs on TestFailures page should
1011         include the crashing symbol (like links in results.html do)
1012
1013         Reviewed by David Kilzer.
1014
1015         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
1016         (Builder.prototype.failureDiagnosisTextAndURL): Changed to expect a testResult object
1017         instead of just a failureType string.
1018
1019         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/FlakyLayoutTestDetector.js:
1020         (FlakyLayoutTestDetector.prototype.incorporateTestResults): Changed to store a
1021         testResult-like object for passing tests.
1022         (FlakyLayoutTestDetector.prototype.flakinessExamples): Changed to expect testResult-like
1023         objects.
1024
1025         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
1026         (LayoutTestResultsLoader.prototype.start): Store a version number along with the cached data
1027         so we can throw away cached data that's in an old format. Store a testResult object for each
1028         test instead of just its failure type.
1029
1030         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1031         (ViewController.prototype._domForFailedTest):
1032         (ViewController.prototype._domForFailureDiagnosis):
1033         Changed to expect testResult objects instead of failureType strings.
1034
1035 2011-06-30  Adam Roben  <aroben@apple.com>
1036
1037         Show full commit logs when visiting Trac from TestFailures page
1038
1039         Fixes <http://webkit.org/b/63715> Links to Trac on TestFailures page
1040         should show full commit logs
1041
1042         Reviewed by David Kilzer.
1043
1044         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Trac.js:
1045         (Trac.prototype.logURL): Added new 'showFullCommitLogs' parameter. When specified, we add
1046         verbose=on to the log URL to tell Trac to show full commit logs.
1047
1048         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1049         (ViewController.prototype._domForRegressionRange): Specify true for showFullCommitLogs.
1050
1051 2011-06-30  Adam Roben  <aroben@apple.com>
1052
1053         Add a link to file a bug about TestFailures to the TestFailures page
1054
1055         Fixes <http://webkit.org/b/63466> TestFailures page should include a link to file a bug
1056         about TestFailures itself
1057
1058         Reviewed by Anders Carlsson.
1059
1060         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
1061         (aside): Added some styles for the new link. It is positioned in the upper-right and
1062         hopefully not to glaringly obvious.
1063
1064         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1065         (ViewController.prototype.loaded): Set up our main content element and auxiliary UI.
1066
1067         (ViewController.prototype._displayBuilder):
1068         (ViewController.prototype._displayTesters):
1069         Modify the main content element instead of the entire body, as there are other things in the
1070         body that we want to preserve.
1071
1072         (ViewController.prototype._domForAuxiliaryUIElements): Create and return an <aside> element
1073         that contains a link to file a bug about the TestFailures page.
1074
1075 2011-06-30  Adam Roben  <aroben@apple.com>
1076
1077         Ignore callbacks that correspond to old navigations of the TestFailures page
1078
1079         Fixes <http://webkit.org/b/63711> TestFailures page sometimes ignores back/forward
1080         navigations
1081
1082         Reviewed by Anders Carlsson.
1083
1084         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
1085         (LayoutTestHistoryAnalyzer.prototype.start): Check the return value of the callback. If the
1086         callback returns false, stop fetching more data.
1087
1088         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1089         (ViewController): Initialize new _navigationID member. This member is used to assign each
1090         navigation by the user a unique ID.
1091         (ViewController.prototype.parseHash): Increment the navigation ID.
1092
1093         (ViewController.prototype._displayBuilder):
1094         (ViewController.prototype._displayTesters):
1095         At the start of these functions, record the current navigation ID. Bail out from callbacks
1096         if the navigation ID has changed, indicating that the user has navigated elsewhere.
1097
1098 2011-06-30  Carlos Garcia Campos  <cgarcia@igalia.com>
1099
1100         Reviewed by Martin Robinson.
1101
1102         [GTK] Add back/forward menu to MiniBrowser toolbar
1103         https://bugs.webkit.org/show_bug.cgi?id=63445
1104
1105         * MiniBrowser/gtk/BrowserWindow.c:
1106         (browser_window_init): Use a GtkMenuToolButton for navigation
1107         widgets.
1108         (browserWindowConstructed): Get the page BackForwardList.
1109         (browserWindowHistoryItemActivated): Go to activated history item.
1110         (browserWindowHistoryItemSelected): Show the url of currently
1111         selected history item in the statusbar.
1112         (createGtkActionFromBackForwardItem): Create a GtkAction from a
1113         WKBackForwardListItemRef.
1114         (browserWindowCreateMenuItemFromBackForwardItem): Create a
1115         GtkMenuItem from a WKBackForwardListItemRef.
1116         (browserWindowCreateBackForwardMenu): Create a GtkMenu with the
1117         given history items.
1118         (browserWindowUpdateNavigationActions): Update navigation widgets
1119         sensitivity and history menus.
1120         (didChangeBackForwardList): Call
1121         browserWindowUpdateNavigationActions().
1122         (browserWindowLoaderClientInit): Add didChangeBackForwardList
1123         callback.
1124
1125 2011-06-30  Dmitry Lomov  <dslomov@google.com>
1126
1127         Reviewed by Adam Roben.
1128
1129         https://bugs.webkit.org/show_bug.cgi?id=63660
1130         This enables unit-tests step on Windows and MacOS SnowLeopard.
1131
1132         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1133
1134 2011-06-30  Balazs Kelemen  <kbalazs@webkit.org>
1135
1136         Reviewed by Kenneth Rohde Christiansen.
1137
1138         [Qt][WK2] Implement the rest of WTR::PlatformWebView
1139         https://bugs.webkit.org/show_bug.cgi?id=63630
1140
1141         * WebKitTestRunner/qt/PlatformWebViewQt.cpp:
1142         (WTR::PlatformWebView::windowFrame):
1143         (WTR::PlatformWebView::setWindowFrame):
1144
1145 2011-06-29  Dirk Pranke  <dpranke@chromium.org>
1146
1147         Reviewed by Adam Barth.
1148
1149         test-webkitpy is spamming lots of messages about updating test expectations
1150         https://bugs.webkit.org/show_bug.cgi?id=63680
1151
1152         This patch changes the passing_run() function to capture the
1153         output to stdout and stderr and assert that it is empty, and
1154         also specifies --no-new-test-results by default to avoid getting
1155         "generating baseline" messages for the expected-missing results.
1156
1157         Also check that we are logging messages when we generate new
1158         baselines.
1159
1160         Also fix a minor bug in the way test_tolerance() was
1161         checking for whether the tests were passing or not.
1162
1163         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1164
1165 2011-06-29  Yuta Kitamura  <yutak@chromium.org>
1166
1167         Reviewed by Kent Tamura.
1168
1169         DRT: Make Hixie76WebSocketProtocolEnabled preference flag configurable from LayoutTestController
1170         https://bugs.webkit.org/show_bug.cgi?id=63532
1171
1172         LayoutTestController::overridePreference() will be used to change the value of
1173         the Hixie76WebSocketProtocolEnabled preference flag from layout tests. In this way,
1174         we can avoid writing a lot of boilerplate code in each port's DumpRenderTree, but
1175         at least we need to reset preference values in DumpRenderTree every time we start
1176         a new test because preference values changed in some test should not be leaked
1177         to the next test.
1178
1179         * DumpRenderTree/chromium/LayoutTestController.cpp:
1180         (LayoutTestController::overridePreference):
1181         The name of preference key is defined in WebKit/{mac/WebView,win}/WebPreferenceKeysPrivate.h.
1182         * DumpRenderTree/chromium/WebPreferences.cpp:
1183         (WebPreferences::reset):
1184         Set the default value. This will be called from TestShell::resetTestController().
1185         (WebPreferences::applyTo):
1186         * DumpRenderTree/chromium/WebPreferences.h:
1187         * DumpRenderTree/mac/DumpRenderTree.mm:
1188         (resetDefaultsToConsistentValues): Set the default value.
1189         * DumpRenderTree/win/DumpRenderTree.cpp:
1190         (resetDefaultsToConsistentValues): Ditto.
1191
1192 2011-06-29  Kenichi Ishibashi  <bashi@chromium.org>
1193
1194         Reviewed by Adam Barth.
1195
1196         check-webkit-style fails to detect https://webkit.org/b/XXXXX as a bug number
1197         https://webkit.org/b/57360
1198
1199         Modify the regular expression to match https scheme.
1200
1201         * Scripts/webkitpy/common/checkout/changelog_unittest.py: Added a test case for https://webkit.org/b/XXXXX.
1202         * Scripts/webkitpy/common/config/urls.py: Modified bug_url_short to match https scheme.
1203
1204 2011-06-29  Adam Barth  <abarth@webkit.org>
1205
1206         Reviewed by Eric Seidel.
1207
1208         new-run-webkit-tests: images page switch actual and expected images before they're loaded
1209         https://bugs.webkit.org/show_bug.cgi?id=63199
1210
1211         We now preload the images and keep the DOM nodes in memory instead of
1212         going back to the MemoryCache all the time.  (Also, remove some nutty
1213         </img> HTML.)
1214
1215         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py:
1216
1217 2011-06-29  Adam Barth  <abarth@webkit.org>
1218
1219         Reviewed by Eric Seidel.
1220
1221         new-run-webkit-tests should tell you when it creates new expectations
1222         https://bugs.webkit.org/show_bug.cgi?id=63675
1223
1224         Unforunately, this patch creates a bunch of test-webkitpy spam because
1225         we don't understand how Python logging works well enough to stop the
1226         spam.  :(
1227
1228         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
1229
1230 2011-06-29  Eric Seidel  <eric@webkit.org>
1231
1232         Reviewed by Adam Barth.
1233
1234         Refactor WebKitDriver.run_test into smaller pieces
1235         https://bugs.webkit.org/show_bug.cgi?id=63673
1236
1237         There is no functional change here.  Just moving code around.
1238         This is in preparation for adding support for #CRASHED and #CRASHED - WebProcess
1239
1240         * Scripts/webkitpy/layout_tests/port/webkit.py:
1241
1242 2011-06-29  Eric Seidel  <eric@webkit.org>
1243
1244         Reviewed by Adam Barth.
1245
1246         Simplify skipped file finding in preparation for adding wk2 skipped list fallback
1247         https://bugs.webkit.org/show_bug.cgi?id=63501
1248
1249         Re-landing after fixing a bug in WebKitPort.baseline_search_path which broke
1250         the rebaselineserver unit tests.  Also fixed nits brought up by Dirk after previous landing.
1251
1252         The goal was to add support for wk2 skipped lists.
1253         However, I found that our skipped list computation was a
1254         manual hack (which only worked for the mac port).
1255
1256         So I fixed a FIXME to move the skipped list finding
1257         into WebKitPort instead of MacPort.
1258         Doing so required the concept of a "port_name", since previously
1259         the only name accessible from a port object was name()
1260         which includes many things beyond "mac" or "chromium", etc.
1261
1262         Eventually I believe we'll want to re-think the way that we pass
1263         in a port_name argument to Port subclasses and expect them to parse
1264         it.  But for now I just added a cls.port_name variable which contains
1265         the static information needed to compute wk2 names as well as
1266         compute Skipped list fallback which works for Mac/Win/Qt and Gtk.
1267
1268         In order to test my new _skipped_file_search_paths method, I
1269         fixed another FIXME by making it return relative paths.
1270
1271         I also fixed the test_expectations_path code in WebKitPort to use port_name.
1272         It was using name() which would return PORT-VERSION so MacPort was overriding
1273         it to use just PORT.  After fixing test_expectations_path to use port_name
1274         (and making it aware of webkit2) I was able to remove the MacPort implementation.
1275
1276         * Scripts/webkitpy/layout_tests/port/base.py:
1277          - Add port_name() to access "mac", since name() returns "mac-leopard" etc.
1278          - Document that real_name() seems to have no purpose.
1279         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1280          - Add _parse_port_name(), eventually we might call this from WebKitPort directly.
1281         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1282          - Add _parse_port_name.
1283         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1284          - Add _parse_port_name.
1285         * Scripts/webkitpy/layout_tests/port/gtk.py:
1286         * Scripts/webkitpy/layout_tests/port/mac.py:
1287          - Move Skipped-file finding code down to WebKitPort
1288         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1289         * Scripts/webkitpy/layout_tests/port/qt.py:
1290         * Scripts/webkitpy/layout_tests/port/webkit.py:
1291         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
1292
1293 2011-06-29  Sheriff Bot  <webkit.review.bot@gmail.com>
1294
1295         Unreviewed, rolling out r90070.
1296         http://trac.webkit.org/changeset/90070
1297         https://bugs.webkit.org/show_bug.cgi?id=63676
1298
1299         Broke test-webkitpy rebaseline tests (Requested by eseidel2 on
1300         #webkit).
1301
1302         * Scripts/webkitpy/layout_tests/port/base.py:
1303         * Scripts/webkitpy/layout_tests/port/chromium.py:
1304         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1305         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1306         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1307         * Scripts/webkitpy/layout_tests/port/gtk.py:
1308         * Scripts/webkitpy/layout_tests/port/mac.py:
1309         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1310         * Scripts/webkitpy/layout_tests/port/qt.py:
1311         * Scripts/webkitpy/layout_tests/port/webkit.py:
1312         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
1313         * Scripts/webkitpy/layout_tests/port/win.py:
1314         * Scripts/webkitpy/tool/commands/rebaselineserver_unittest.py:
1315
1316 2011-06-29  Adam Barth  <abarth@webkit.org>
1317
1318         Reviewed by Eric Seidel.
1319
1320         new-run-webkit-tests doesn't support sample-on-timeout
1321         https://bugs.webkit.org/show_bug.cgi?id=56731
1322
1323         * Scripts/webkitpy/layout_tests/port/server_process.py:
1324         * Scripts/webkitpy/layout_tests/port/server_process_unittest.py:
1325         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1326
1327 2011-06-29  Eric Seidel  <eric@webkit.org>
1328
1329         Reviewed by Adam Barth.
1330
1331         Simplify skipped file finding in preparation for adding wk2 skipped list fallback
1332         https://bugs.webkit.org/show_bug.cgi?id=63501
1333
1334         The goal was to add support for wk2 skipped lists.
1335         However, I found that our skipped list computation was a
1336         manual hack (which only worked for the mac port).
1337
1338         So I fixed a FIXME to move the skipped list finding
1339         into WebKitPort instead of MacPort.
1340         Doing so required the concept of a "port_name", since previously
1341         the only name accessible from a port object was name()
1342         which includes many things beyond "mac" or "chromium", etc.
1343
1344         Eventually I believe we'll want to re-think the way that we pass
1345         in a port_name argument to Port subclasses and expect them to parse
1346         it.  But for now I just added a cls.port_name variable which contains
1347         the static information needed to compute wk2 names as well as
1348         compute Skipped list fallback which works for Mac/Win/Qt and Gtk.
1349
1350         In order to test my new _skipped_file_search_paths method, I
1351         fixed another FIXME by making it return relative paths.
1352
1353         I also fixed the test_expectations_path code in WebKitPort to use port_name.
1354         It was using name() which would return PORT-VERSION so MacPort was overriding
1355         it to use just PORT.  After fixing test_expectations_path to use port_name
1356         (and making it aware of webkit2) I was able to remove the MacPort implementation.
1357
1358         * Scripts/webkitpy/layout_tests/port/base.py:
1359          - Add port_name() to access "mac", since name() returns "mac-leopard" etc.
1360          - Document that real_name() seems to have no purpose.
1361         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1362          - Add _parse_port_name(), eventually we might call this from WebKitPort directly.
1363         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1364          - Add _parse_port_name.
1365         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
1366          - Add _parse_port_name.
1367         * Scripts/webkitpy/layout_tests/port/gtk.py:
1368         * Scripts/webkitpy/layout_tests/port/mac.py:
1369          - Move Skipped-file finding code down to WebKitPort
1370         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1371         * Scripts/webkitpy/layout_tests/port/qt.py:
1372         * Scripts/webkitpy/layout_tests/port/webkit.py:
1373         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
1374
1375 2011-06-29  Adam Barth  <abarth@webkit.org>
1376
1377         Reviewed by Dirk Pranke.
1378
1379         new-run-webkit-tests complains about missing pixel results instead of plopping down new expectations
1380         https://bugs.webkit.org/show_bug.cgi?id=38063
1381
1382         This patch changes new-run-webkit-tests to match old-run-webkit-tests
1383         in generating new expected results when the expected results are
1384         missing.
1385
1386         There are still a couple details that are different:
1387
1388         1) Image baselines aren't generated unless you run with -p.
1389         2) Render tree dumps are places in the cross-platform directory instead
1390         of the platform-specific directory.
1391
1392         I'm inclined to deal with both of these issues in follow-up patches.
1393
1394         * Scripts/webkitpy/common/net/layouttestresults.py:
1395         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
1396         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
1397         * Scripts/webkitpy/layout_tests/layout_package/test_results.py:
1398         * Scripts/webkitpy/layout_tests/port/test.py:
1399         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1400         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1401
1402 2011-06-29  Adam Roben  <aroben@apple.com>
1403
1404         Teach TestFailures to detect possibly flaky tests and list them separately
1405
1406         Fixes <http://webkit.org/b/61061> <rdar://problem/9452796> TestFailures page blames
1407         arbitrary revisions for breaking flaky tests
1408
1409         Reviewed by Dan Bates.
1410
1411         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/FlakyLayoutTestDetector.js: Added.
1412         (FlakyLayoutTestDetector): This class identifies flaky tests when given the test results
1413         from various builds (in reverse-chronological order).
1414         (FlakyLayoutTestDetector.prototype.incorporateTestResults): Detects flaky tests. Tests move
1415         monotonically through three states: LastSeenFailing, LastSeenPassing, and PossiblyFlaky.
1416         (FlakyLayoutTestDetector.prototype.flakinessExamples): Finds examples of flakiness for the
1417         given test. Essentially, finds all the transitions from passing to failing (or vice-versa)
1418         and puts them in an array in reverse-chronological order.
1419         (FlakyLayoutTestDetector.prototype.get possiblyFlakyTests): Returns all tests we've detected
1420         to be possibly flaky.
1421
1422         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
1423         (LayoutTestHistoryAnalyzer): Initialize new members.
1424         (LayoutTestHistoryAnalyzer.prototype.start): Now passes the callback an object with two
1425         properties: history and possiblyFlaky. history holds the data this function used to pass to
1426         the callback, while possiblyFlaky lists all tests that might be flaky and examples of their
1427         flakiness. Updated documentation comment to match.
1428         (LayoutTestHistoryAnalyzer.prototype._incorporateBuildHistory): Now uses a
1429         FlakyLayoutTestDetector to identify possibly flaky tests. Any possibly flaky tests are
1430         removed from the failure history, since when they started failing is no longer meaningful.
1431         We tell our caller to keep calling until all current failures have been explained and we've
1432         gone through 5 builds without any new flaky tests being identified.
1433
1434         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
1435         (sorted): New helper function to return a sorted copy of an array.
1436         (Array.prototype.findLast): New helper function. Like findFirst, but finds the last item
1437         that matches the predicate.
1438
1439         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1440         (ViewController.prototype._displayBuilder): Updated for change in the object passed to us by
1441         the analyzer. Now puts the list of possibly flaky tests after the failure history.
1442         (ViewController.prototype._domForFailedTest): Moved some code from here...
1443         (ViewController.prototype._domForFailureDiagnosis): ...to here.
1444         (ViewController.prototype._domForPossiblyFlakyTests): New function, builds up a list of
1445         possibly flaky tests and examples of their flakiness and returns it.
1446
1447         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/index.html: Pull in
1448         FlakyLayoutTestDetector.js.
1449
1450 2011-06-29  Eric Seidel  <eric@webkit.org>
1451
1452         Adam says cowboys don't review (or unit test).
1453
1454         * Scripts/webkitpy/tool/commands/download.py:
1455
1456 2011-06-29  Eric Seidel  <eric@webkit.org>
1457
1458         Reviewed by Adam Barth.
1459
1460         Make port/config.py mockable for easier unit testing
1461         https://bugs.webkit.org/show_bug.cgi?id=63661
1462
1463         Fix an exception from the unit tests.
1464
1465         * Scripts/webkitpy/layout_tests/port/config.py:
1466
1467 2011-06-29  Brent Fulgham  <bfulgham@webkit.org>
1468
1469         Unreviewed build correction.
1470
1471         * DumpRenderTree/win/ImageDiffWinCairoCommon.vsprops:
1472         Link against zlib1.dll, not the static zlib.lib.
1473
1474 2011-06-29  Eric Seidel  <eric@webkit.org>
1475
1476         Reviewed by Adam Barth.
1477
1478         Make port/config.py mockable for easier unit testing
1479         https://bugs.webkit.org/show_bug.cgi?id=63661
1480
1481         Right now config.py leaks real system paths for
1482         webkit_base_dir due to using __file__.  This patch
1483         adds a filsystem.py wrapper around __file__ so that
1484         we don't have to jump through hoops in webkit_unittest.py
1485         to avoid leaking local paths to the unit test results.
1486
1487         There are probably many more places where we should now
1488         use filesystem.path_for_module, but starting with just this one.
1489
1490         * Scripts/webkitpy/common/system/filesystem.py:
1491         * Scripts/webkitpy/common/system/filesystem_mock.py:
1492         * Scripts/webkitpy/layout_tests/port/config.py:
1493         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
1494         * Scripts/webkitpy/tool/commands/rebaselineserver_unittest.py:
1495
1496 2011-06-29  Eric Seidel  <eric@webkit.org>
1497
1498         Reviewed by Adam Barth.
1499
1500         Remove duplicate methods in filesystem.py
1501         https://bugs.webkit.org/show_bug.cgi?id=63658
1502
1503         Looks like there was a bad merge at some point.
1504
1505         I also removed a bunch of redundant docstrings.
1506
1507         * Scripts/webkitpy/common/system/filesystem.py:
1508
1509 2011-06-29  Adam Barth  <abarth@webkit.org>
1510
1511         Reviewed by Eric Seidel.
1512
1513         Pass --no-new-test-results to NRWT on build.webkit.org
1514         https://bugs.webkit.org/show_bug.cgi?id=63653
1515
1516         This is in preparation for https://bugs.webkit.org/show_bug.cgi?id=38063.
1517
1518         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1519
1520 2011-06-29  Xan Lopez  <xlopez@igalia.com>
1521
1522         Reviewed by Adam Barth.
1523
1524         [GTK] overlapping drag&drop tests fail on NRWT
1525         https://bugs.webkit.org/show_bug.cgi?id=57640
1526
1527         Make a GTK test driver that spawns one Xvfb instance per
1528         thread. This avoids bad interactions in DnD tests between threads.
1529
1530         * Scripts/webkitpy/layout_tests/port/gtk.py:
1531
1532 2011-06-29  Sheriff Bot  <webkit.review.bot@gmail.com>
1533
1534         Unreviewed, rolling out r89888.
1535         http://trac.webkit.org/changeset/89888
1536         https://bugs.webkit.org/show_bug.cgi?id=63609
1537
1538          "Seems to be causing merging issues on the dashboard, e.g.
1539         http://test-
1540         results.appspot.com/dashboards/flakiness_dashboard.html#group=%40ToT%20-%20chromium.org&tests=fast%2Fcanvas"
1541         (Requested by aboxhall on #webkit).
1542
1543         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
1544         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1545         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
1546
1547 2011-06-29  Philippe Normand  <pnormand@igalia.com>
1548
1549         Reviewed by Martin Robinson.
1550
1551         [webkitpy] Executive.running_pids() sometimes doesn't correctly process "ps" output
1552         https://bugs.webkit.org/show_bug.cgi?id=63624
1553
1554         * Scripts/webkitpy/common/system/executive.py: In some cases the
1555         ps output lines can contain one or more leading white-spaces, so
1556         strip it before splitting the pid and command name.
1557
1558 2011-06-29  Noel Gordon  <noel.gordon@gmail.com>
1559
1560         Reviewed by Kent Tamura.
1561
1562         [Chromium] [DRT] ImageDiff should compile in VS2010
1563         https://bugs.webkit.org/show_bug.cgi?id=63614
1564
1565         * DumpRenderTree/chromium/ImageDiff.cpp: Include <iterator> for back_inserter.
1566
1567 2011-06-27  Diego Gonzalez  <diegohcg@webkit.org>
1568
1569         Reviewed by Kenneth Rohde Christiansen.
1570
1571         [Qt] DRT support for setInteractiveFormValidationEnabled
1572         https://bugs.webkit.org/show_bug.cgi?id=63496
1573
1574         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
1575         (WebCore::DumpRenderTree::DumpRenderTree):
1576
1577 2011-06-28  Eric Seidel  <eric@webkit.org>
1578
1579         Reviewed by Adam Barth.
1580
1581         Make new-run-webkit-tests -2 actually use WebKitTestRunner
1582         https://bugs.webkit.org/show_bug.cgi?id=63572
1583
1584         new-run-webkit-tests -2 works with this change.  However a zillion
1585         tests fail since it can't find the right skipped lists for wk2.
1586         The Skipped-list fixing is covered in bug 63501.
1587
1588         * Scripts/webkitpy/layout_tests/port/gtk.py:
1589         * Scripts/webkitpy/layout_tests/port/qt.py:
1590         * Scripts/webkitpy/layout_tests/port/webkit.py:
1591
1592 2011-06-28  Ryosuke Niwa  <rniwa@webkit.org>
1593
1594         Reviewed by Adam Roben.
1595
1596         Add more machines to Chromium Mac Release and Chromium Mac Release (Tests)
1597         https://bugs.webkit.org/show_bug.cgi?id=63500
1598
1599         Added google-mac-3 and google-mac-4.
1600
1601         * BuildSlaveSupport/build.webkit.org-config/config.json:
1602
1603 2011-06-28  Eric Seidel  <eric@webkit.org>
1604
1605         Reviewed by Dirk Pranke.
1606
1607         Remove webkitpy's tiger support now that WebKit no longer supports tiger
1608         https://bugs.webkit.org/show_bug.cgi?id=63568
1609
1610         I don't think our support actually fully worked anyway.  Regardless
1611         tiger is dead now (thankfully)!
1612
1613         I also unwrapped some short lines and added a couple FIXMEs while
1614         touring the code.
1615
1616         No new tests, only removing old ones which tested our Tiger support.
1617
1618         * Scripts/webkitpy/layout_tests/port/base.py:
1619         * Scripts/webkitpy/layout_tests/port/builders.py:
1620         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
1621         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
1622         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
1623         * Scripts/webkitpy/layout_tests/port/mac.py:
1624         * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
1625         * Scripts/webkitpy/tool/commands/rebaseline.py:
1626         * Scripts/webkitpy/tool/commands/rebaselineserver_unittest.py:
1627
1628 2011-06-28  Eric Seidel  <eric@webkit.org>
1629
1630         Unreviewed, just fixing a typo.
1631
1632         Fix misspellings of "thead" to "thread".
1633
1634         * Scripts/webkitpy/common/checkout/scm/git.py:
1635         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1636
1637 2011-06-28  Eric Seidel  <eric@webkit.org>
1638
1639         Reviewed by Tony Chang.
1640
1641         new-run-webkit-tests needs a --webkit-test-runner option
1642         https://bugs.webkit.org/show_bug.cgi?id=63439
1643
1644         This was accidentally fixed in bug 63501 as well.
1645         The root problem here was that the webkit unit tests were not using
1646         enough mocks.  I've fixed them here to use more mocks
1647         which should unbreak windows.
1648
1649         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
1650
1651 2011-06-28  Eric Seidel  <eric@webkit.org>
1652
1653         Reviewed by Tony Chang.
1654
1655         new-run-webkit-tests should not run mthml tests when MHTML support is disabled
1656         https://bugs.webkit.org/show_bug.cgi?id=63549
1657
1658         This way mhtml tests won't show up as "missing" results on all non-chromium platforms.
1659         I happen to go looking for what code was deciding that .mht was a supported
1660         extension, found test_files.py and then removed some redundant doc-strings and
1661         if branching.
1662
1663         * Scripts/webkitpy/layout_tests/port/test_files.py:
1664         * Scripts/webkitpy/layout_tests/port/webkit.py:
1665         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
1666
1667 2011-06-28  Roland Steiner  <rolandsteiner@chromium.org>
1668
1669         Reviewed by Tony Chang.
1670
1671         Fix filesystem_unittest.FileSystemTest test_read_and_write_file()
1672         https://bugs.webkit.org/show_bug.cgi?id=63514
1673
1674         - change write_text_file to write_binary_file
1675         - check that text_path and binary_path are actually existant files
1676         - change cleanup block from 'except' to 'finally'
1677
1678         * Scripts/webkitpy/common/system/filesystem_unittest.py:
1679
1680 2011-06-28  Adam Roben  <aroben@apple.com>
1681
1682         Roll out r89924 and r89925
1683
1684         I didn't mean to commit them!
1685
1686         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
1687         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
1688         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1689
1690 2011-06-27  Adam Barth  <abarth@webkit.org>
1691
1692         Reviewed by Eric Seidel.
1693
1694         NRWT should wait for ReportCrash
1695         https://bugs.webkit.org/show_bug.cgi?id=63495
1696
1697         When a test crashes on Mac, ReportCrash confuses us into thinking the
1698         test times out.
1699
1700         I'm not sure how to test this in a meaningful way in our current test
1701         harness...
1702
1703         * Scripts/webkitpy/common/system/crashlogs.py:
1704         * Scripts/webkitpy/common/system/crashlogs_unittest.py:
1705         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py:
1706         * Scripts/webkitpy/layout_tests/port/base.py:
1707         * Scripts/webkitpy/layout_tests/port/mac.py:
1708         * Scripts/webkitpy/layout_tests/port/server_process.py:
1709         * Scripts/webkitpy/tool/commands/queries.py:
1710
1711 2011-06-27  Sheriff Bot  <webkit.review.bot@gmail.com>
1712
1713         Unreviewed, rolling out r89881.
1714         http://trac.webkit.org/changeset/89881
1715         https://bugs.webkit.org/show_bug.cgi?id=63513
1716
1717         Makes Qt massively slow (Requested by abarth on #webkit).
1718
1719         * Scripts/webkitpy/common/system/crashlogs.py:
1720         * Scripts/webkitpy/common/system/crashlogs_unittest.py:
1721         * Scripts/webkitpy/common/system/executive.py:
1722         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py:
1723         * Scripts/webkitpy/layout_tests/port/base.py:
1724         * Scripts/webkitpy/layout_tests/port/mac.py:
1725         * Scripts/webkitpy/layout_tests/port/server_process.py:
1726         * Scripts/webkitpy/tool/commands/queries.py:
1727
1728 2011-06-27  Alice Boxhall  <aboxhall@chromium.org>
1729
1730         Reviewed by Ojan Vafai.
1731
1732         Convert json_results_generator.py to output version 4 JSON.
1733         https://bugs.webkit.org/show_bug.cgi?id=60869
1734
1735         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
1736         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
1737         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
1738
1739 2011-06-27  Adam Barth  <abarth@webkit.org>
1740
1741         Reviewed by Eric Seidel.
1742
1743         NRWT should wait for ReportCrash
1744         https://bugs.webkit.org/show_bug.cgi?id=63495
1745
1746         When a test crashes on Mac, ReportCrash confuses us into thinking the
1747         test times out.
1748
1749         I'm not sure how to test this in a meaningful way in our current test
1750         harness...
1751
1752         * Scripts/webkitpy/common/system/crashlogs.py:
1753         * Scripts/webkitpy/common/system/crashlogs_unittest.py:
1754         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py:
1755         * Scripts/webkitpy/layout_tests/port/base.py:
1756         * Scripts/webkitpy/layout_tests/port/mac.py:
1757         * Scripts/webkitpy/layout_tests/port/server_process.py:
1758         * Scripts/webkitpy/tool/commands/queries.py:
1759
1760 2011-06-27  Adam Barth  <abarth@webkit.org>
1761
1762         Reviewed by Eric Seidel.
1763
1764         new-run-webkit-tests should upload crash logs
1765         https://bugs.webkit.org/show_bug.cgi?id=55907
1766
1767         NRWT used to have a different name for crash logs.  I've changed the
1768         name to patch ORWT.
1769
1770         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py:
1771         * Scripts/webkitpy/layout_tests/run_webkit_tests_integrationtest.py:
1772
1773 2011-06-27  Eric Seidel  <eric@webkit.org>
1774
1775         Reviewed by Adam Barth.
1776
1777         new-run-webkit-tests needs a --webkit-test-runner option
1778         https://bugs.webkit.org/show_bug.cgi?id=63439
1779
1780         NRWT doesn't actually know how to run with the WebKitTestRunner yet
1781         but it does have a --webkit-test-runner option and will build WebKitTestRunner correctly.
1782
1783         There is a bunch of other little cleanup in this patch which I added as I
1784         took a tour through all of our hard-coded DumpRenderTree strings.
1785
1786         * Scripts/webkitpy/common/net/layouttestresults.py: Added FIXME.
1787         * Scripts/webkitpy/common/system/outputcapture.py: Fixed spacing to pass PEP8.
1788         * Scripts/webkitpy/layout_tests/layout_package/manager.py: Made the FIXME slightly stronger (bad bug).
1789         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py: test_shell is dead.
1790         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py: Unwrapped silly wrapping.
1791         * Scripts/webkitpy/layout_tests/port/base.py: Removed two dead methods.
1792         * Scripts/webkitpy/layout_tests/port/config.py:
1793          - build_dumprendertree had no business in this class, removed it.
1794          - Exposed _FLAGS_FROM_CONFIGURATIONS through flag_for_configuration()
1795          - Exposed _script_path as script_path() (this probably belongs elsewhere).
1796         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1797          - Moved these tests to webkit_unittests.
1798         * Scripts/webkitpy/layout_tests/port/webkit.py:
1799          - Now respects the --webkit-test-runner option.
1800          - setup_test_run is empty in base.py too, no need to override it.
1801         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
1802          - Test the new hotness.
1803          - Bad, bad, bad!  The old code was using a real Executive during unit-testing!
1804            I think this code is still hitting disk during the unit tests. :(
1805         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1806          - Add the option.
1807         * Scripts/webkitpy/tool/bot/queueengine_unittest.py:
1808          - Remove optional args option.
1809         * Scripts/webkitpy/tool/commands/commandtest.py:
1810          - Remove optional args option.
1811         * Scripts/webkitpy/tool/commands/queues_unittest.py:
1812          - Remove optional args option.
1813
1814 2011-06-27  Adam Barth  <abarth@webkit.org>
1815
1816         Reviewed by Dirk Pranke.
1817
1818         new-run-webkit-tests does not report stderr output
1819         https://bugs.webkit.org/show_bug.cgi?id=37739
1820
1821         Almost all the infrastructure appears to be in place to report stderr
1822         in results.html.  We just need to wire up a bool to tell results.html
1823         to create the hyperlink.
1824
1825         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
1826         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
1827         * Scripts/webkitpy/layout_tests/layout_package/test_result_writer.py:
1828         * Scripts/webkitpy/layout_tests/layout_package/test_results.py:
1829         * Scripts/webkitpy/layout_tests/port/base.py:
1830
1831 2011-06-27  Adam Barth  <abarth@webkit.org>
1832
1833         Reviewed by Dirk Pranke.
1834
1835         webkitpy unit tests should have more descriptive names than just "Test"
1836         https://bugs.webkit.org/show_bug.cgi?id=63479
1837
1838         Luckily we kept these classes in separate namespaces in the harness so
1839         we were actually running them all even though they had the same name.
1840
1841         * Scripts/webkitpy/common/system/stack_utils_unittest.py:
1842         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
1843         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:
1844
1845 2011-06-27  Greg Simon  <gregsimon@chromium.org>
1846
1847         No review necessary.
1848
1849         Add email addr to contributor list
1850         https://bugs.webkit.org/show_bug.cgi?id=63478
1851
1852         * Scripts/webkitpy/common/config/committers.py:
1853
1854 2011-06-27  Sheriff Bot  <webkit.review.bot@gmail.com>
1855
1856         Unreviewed, rolling out r89834.
1857         http://trac.webkit.org/changeset/89834
1858         https://bugs.webkit.org/show_bug.cgi?id=63476
1859
1860         Broke test-webkitpy (Requested by abarth on #webkit).
1861
1862         * Scripts/webkitpy/common/net/layouttestresults.py:
1863         * Scripts/webkitpy/common/system/outputcapture.py:
1864         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
1865         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py:
1866         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
1867         * Scripts/webkitpy/layout_tests/port/base.py:
1868         * Scripts/webkitpy/layout_tests/port/config.py:
1869         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1870         * Scripts/webkitpy/layout_tests/port/webkit.py:
1871         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
1872         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1873         * Scripts/webkitpy/tool/bot/queueengine_unittest.py:
1874         * Scripts/webkitpy/tool/commands/commandtest.py:
1875         * Scripts/webkitpy/tool/commands/queues_unittest.py:
1876
1877 2011-06-27  Jeff Miller  <jeffm@apple.com>
1878
1879         Reviewed by Adam Roben.
1880
1881         Stop copying obsolete WebKit libraries on Windows
1882         https://bugs.webkit.org/show_bug.cgi?id=63474
1883         
1884         Don't copy obsolete WebKit library directories and files on Windows, they no longer exist and just generate extraneous warnings. These include the CharacterSets directory, dnssd.dll, and icudt40.dll.
1885
1886         * DumpRenderTree/win/DumpRenderTreePostBuild.cmd:
1887         * MiniBrowser/MiniBrowserPostBuild.cmd:
1888         * WebKitTestRunner/win/WebKitTestRunnerPostBuild.cmd:
1889
1890 2011-06-27  Ryosuke Niwa  <rniwa@webkit.org>
1891
1892         Add Wyatt Carss to the list of non-committer contributors.
1893
1894         * Scripts/webkitpy/common/config/committers.py:
1895
1896 2011-06-27  Adam Barth  <abarth@webkit.org>
1897
1898         Reviewed by Eric Seidel.
1899
1900         webkitpy should understand crash logs
1901         https://bugs.webkit.org/show_bug.cgi?id=63468
1902
1903         We're planning to use this functionality to upload crash logs along
1904         with test results for new-run-webkit-tests.
1905
1906         * Scripts/webkitpy/common/system/crashlog.py: Added.
1907         * Scripts/webkitpy/common/system/crashlog_unittest.py: Added.
1908         * Scripts/webkitpy/common/system/executive.py:
1909         * Scripts/webkitpy/common/system/executive_unittest.py:
1910         * Scripts/webkitpy/common/system/filesystem.py:
1911         * Scripts/webkitpy/common/system/filesystem_mock.py:
1912         * Scripts/webkitpy/tool/commands/queries.py:
1913
1914 2011-06-27  Adam Roben  <aroben@apple.com>
1915
1916         Make LayoutTestResultsLoader cache whether old-run-webkit-tests exited early due to too many
1917         failures
1918
1919         Fixes <http://webkit.org/b/63470> TestFailures page for a particular builder forgets
1920         old-run-webkit-tests exited early after reload
1921
1922         Reviewed by Anders Carlsson.
1923
1924         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
1925         (LayoutTestResultsLoader.prototype.start): Store both the set of failing tests and whether
1926         old-run-webkit-tests exited early due to too many failures in PersistentCache.
1927
1928 2011-06-27  Adam Roben  <aroben@apple.com>
1929
1930         Make TestFailures load build names from build.webkit.org/old-results too
1931
1932         Build results are periodically moved from results to old-results. This change makes those
1933         builds still visible to TestFailures.
1934
1935         Fixes <http://webkit.org/b/63453> TestFailures page doesn't show information for builds that
1936         have been moved to build.webkit.org/old-results
1937
1938         Reviewed by Anders Carlsson.
1939
1940         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
1941         (Builder.prototype.getBuildNames): Extracted code to parse build names from a
1942         build.webkit.org directory listing into a separate function. Instead of omitting .zip files,
1943         we now only include directory entries whose names are parseable as build names (since
1944         old-results sometimes contains other random files/directories from who knows what). We now
1945         fetch both results and old-results (with a FIXME about loading old-results on demand),
1946         extract build names from each, and concatenate the two sets of names.
1947
1948         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/WebKitBuildbot.js:
1949         (WebKitBuildbot.prototype.parseBuildName): Changed to return null when the build name isn't
1950         parseable, rather than throwing an exception.
1951
1952 2011-06-26  Adam Roben  <aroben@apple.com>
1953
1954         Store analyzed history directly in LayoutTestHistoryAnalyzer instead of in a generic cache
1955
1956         It seems silly to have a whole generic cache just to store a single object.
1957
1958         Fixes <http://webkit.org/b/63407> LayoutTestHistoryAnalyzer has a generic cache but doesn't
1959         need one
1960
1961         Reviewed by Anders Carlsson.
1962
1963         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
1964         (LayoutTestHistoryAnalyzer): Replaced this._cache with this._history.
1965         (LayoutTestHistoryAnalyzer.prototype.start): Ditto.
1966         (LayoutTestHistoryAnalyzer.prototype._incorporateBuildHistory): Ditto, and removed the
1967         no-longer-needed history argument.
1968
1969 2011-06-26  Adam Roben  <aroben@apple.com>
1970
1971         Move some inappropriate code out of Builder
1972
1973         Fixes <http://webkit.org/b/63406> TestFailure page's Builder class has a bunch of code that
1974         isn't really about a builder
1975
1976         Reviewed by Anders Carlsson.
1977
1978         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
1979         (Builder.prototype.getBuildNames): Renamed from _getBuildNames, since it now needs to be
1980         called by other objects.
1981
1982         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestHistoryAnalyzer.js:
1983         Added. Code came from Builder.
1984         (LayoutTestHistoryAnalyzer):
1985         (LayoutTestHistoryAnalyzer.prototype.start):
1986         (LayoutTestHistoryAnalyzer.prototype._incorporateBuildHistory):
1987
1988         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/LayoutTestResultsLoader.js:
1989         Added. Code came from Builder.
1990         (LayoutTestResultsLoader):
1991         (LayoutTestResultsLoader.prototype.start):
1992
1993         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
1994         (ViewController.prototype._displayBuilder): Changed to use LayoutTestHistoryAnalyzer.
1995
1996         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/index.html: Pull in new
1997         scripts.
1998
1999 2011-05-17  Nat Duca  <nduca@chromium.org>
2000
2001         Reviewed by James Robinson.
2002
2003         [chromium] Implement CCLayerTreeHost and CCLayerTreeHostImpl portions of threaded compositor
2004         https://bugs.webkit.org/show_bug.cgi?id=58408
2005
2006         * DumpRenderTree/chromium/WebViewHost.cpp:
2007         (WebViewHost::paintInvalidatedRegion):
2008
2009 2011-06-27  Eric Seidel  <eric@webkit.org>
2010
2011         Reviewed by Adam Barth.
2012
2013         new-run-webkit-tests needs a --webkit-test-runner option
2014         https://bugs.webkit.org/show_bug.cgi?id=63439
2015
2016         NRWT doesn't actually know how to run with the WebKitTestRunner yet
2017         but it does have a --webkit-test-runner option and will build WebKitTestRunner correctly.
2018
2019         There is a bunch of other little cleanup in this patch which I added as I
2020         took a tour through all of our hard-coded DumpRenderTree strings.
2021
2022         * Scripts/webkitpy/common/net/layouttestresults.py: Added FIXME.
2023         * Scripts/webkitpy/common/system/outputcapture.py: Fixed spacing to pass PEP8.
2024         * Scripts/webkitpy/layout_tests/layout_package/manager.py: Made the FIXME slightly stronger (bad bug).
2025         * Scripts/webkitpy/layout_tests/layout_package/single_test_runner.py: test_shell is dead.
2026         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py: Unwrapped silly wrapping.
2027         * Scripts/webkitpy/layout_tests/port/base.py: Removed two dead methods.
2028         * Scripts/webkitpy/layout_tests/port/config.py:
2029          - build_dumprendertree had no business in this class, removed it.
2030          - Exposed _FLAGS_FROM_CONFIGURATIONS through flag_for_configuration()
2031          - Exposed _script_path as script_path() (this probably belongs elsewhere).
2032         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
2033          - Moved these tests to webkit_unittests.
2034         * Scripts/webkitpy/layout_tests/port/webkit.py:
2035          - Now respects the --webkit-test-runner option.
2036          - setup_test_run is empty in base.py too, no need to override it.
2037         * Scripts/webkitpy/layout_tests/port/webkit_unittest.py:
2038          - Test the new hotness.
2039          - Bad, bad, bad!  The old code was using a real Executive during unit-testing!
2040            I think this code is still hitting disk during the unit tests. :(
2041         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2042          - Add the option.
2043         * Scripts/webkitpy/tool/bot/queueengine_unittest.py:
2044          - Remove optional args option.
2045         * Scripts/webkitpy/tool/commands/commandtest.py:
2046          - Remove optional args option.
2047         * Scripts/webkitpy/tool/commands/queues_unittest.py:
2048          - Remove optional args option.
2049
2050 2011-06-27  Chang Shu  <cshu@webkit.org>
2051
2052         Reviewed by Darin Adler.
2053
2054         [WK2] WebKitTestRunner should reset FrameFlattening flag to disable before each test
2055         https://bugs.webkit.org/show_bug.cgi?id=63457
2056
2057         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
2058         (WTR::InjectedBundle::beginTesting):
2059
2060 2011-06-27  Raphael Kubo da Costa  <kubo@profusion.mobi>
2061
2062         Reviewed by Kent Tamura.
2063
2064         [EFL] DRT: Add an ewk_view specialization for EFL's DumpRenderTree.
2065         It will be plugged into the build system as soon as the rest of the
2066         DRT code is fully upstreamed.
2067         https://bugs.webkit.org/show_bug.cgi?id=63086
2068
2069         * DumpRenderTree/efl/DumpRenderTreeView.cpp: Added.
2070         (onConsoleMessage):
2071         (onJavaScriptAlert):
2072         (onJavaScriptConfirm):
2073         (onJavaScriptPrompt):
2074         (drtViewTiledAdd):
2075         * DumpRenderTree/efl/DumpRenderTreeView.h: Added.
2076
2077 2011-06-27  Leandro Pereira  <leandro@profusion.mobi>
2078
2079         Reviewed by Kent Tamura.
2080
2081         [EFL] Add an EventSender implementation to EFL's DumpRenderTree along
2082         with support files.
2083         Build system changes will be made as soon as the whole DumpRenderTree
2084         is upstreamed.
2085         https://bugs.webkit.org/show_bug.cgi?id=61962
2086
2087         * DumpRenderTree/efl/EventSender.cpp: Added.
2088         (setEvasModifiers):
2089         (translateMouseButtonNumber):
2090         (sendMouseEvent):
2091         (sendClick):
2092         (scheduleAsynchronousClickCallback):
2093         (updateClickCount):
2094         (modifierFromJSValue):
2095         (modifiersFromJSValue):
2096         (mouseDownCallback):
2097         (mouseUpCallback):
2098         (mouseMoveToCallback):
2099         (evasMouseEventFromHorizontalAndVerticalOffsets):
2100         (mouseScrollByCallback):
2101         (continuousMouseScrollByCallback):
2102         (keyPadNameFromJSValue):
2103         (keyNameFromJSValue):
2104         (keyDownCallback):
2105         (scalePageByCallback):
2106         (getClass):
2107         (makeEventSender):
2108         * DumpRenderTree/efl/EventSender.h: Added.
2109
2110 2011-06-27  Eric Seidel  <eric@webkit.org>
2111
2112         Reviewed by Adam Barth.
2113
2114         Remove evil uses of hasattr
2115         https://bugs.webkit.org/show_bug.cgi?id=63430
2116
2117         For some reason these classes believe that they may be called with
2118         various flavors of "option" elements and so carefully check to make
2119         sure that the options element has their option before checking it.
2120
2121         We had a set_option_default method which was never called, so I made it
2122         do what callsites seemed to want it to do and replaced 3 callers
2123         who previously used hasattr manually to use set_option_default instead.
2124
2125         * Scripts/webkitpy/layout_tests/port/base.py:
2126         * Scripts/webkitpy/layout_tests/port/webkit.py:
2127
2128 2011-06-27  Kent Tamura  <tkent@chromium.org>
2129
2130         Reviewed by Hajime Morita.
2131
2132         [Chromium] WebViewHost should be destructed before LayoutTestController, etc.
2133         https://bugs.webkit.org/show_bug.cgi?id=63427
2134
2135         ~WebViewHost() navigates to about:blank, so it calls some
2136         callbacks and LayoutTestController functions.
2137
2138         * DumpRenderTree/chromium/TestShell.h:
2139          Move m_printer upper and move m_webViewHost lower.
2140
2141 2011-06-27  Adam Barth  <abarth@webkit.org>
2142
2143         Reviewed by Eric Seidel.
2144
2145         Move server config files to webkitpy.layout_tests.servers
2146         https://bugs.webkit.org/show_bug.cgi?id=63436
2147
2148         The config files belong with the servers.
2149
2150         * Scripts/webkitpy/layout_tests/port/httpd2.pem: Removed.
2151         * Scripts/webkitpy/layout_tests/port/lighttpd.conf: Removed.
2152         * Scripts/webkitpy/layout_tests/servers/http_server.py:
2153         * Scripts/webkitpy/layout_tests/servers/httpd2.pem: Copied from Tools/Scripts/webkitpy/layout_tests/port/httpd2.pem.
2154         * Scripts/webkitpy/layout_tests/servers/lighttpd.conf: Copied from Tools/Scripts/webkitpy/layout_tests/port/lighttpd.conf.
2155
2156 2011-06-27  Adam Barth  <abarth@webkit.org>
2157
2158         Reviewed by Eric Seidel.
2159
2160         Move websocket_server and http_server to webkitpy.layout_package.servers
2161         https://bugs.webkit.org/show_bug.cgi?id=63434
2162
2163         These classes do not represent ports.  They represent servers.
2164
2165         * Scripts/new-run-webkit-httpd:
2166         * Scripts/new-run-webkit-websocketserver:
2167         * Scripts/webkitpy/layout_tests/port/base.py:
2168         * Scripts/webkitpy/layout_tests/port/chromium.py:
2169         * Scripts/webkitpy/layout_tests/port/http_server.py: Removed.
2170         * Scripts/webkitpy/layout_tests/port/http_server_base.py: Removed.
2171         * Scripts/webkitpy/layout_tests/port/http_server_integrationtest.py: Removed.
2172         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
2173         * Scripts/webkitpy/layout_tests/port/websocket_server.py: Removed.
2174         * Scripts/webkitpy/layout_tests/servers/apache_http_server.py:
2175         * Scripts/webkitpy/layout_tests/servers/http_server.py: Copied from Tools/Scripts/webkitpy/layout_tests/port/http_server.py.
2176         * Scripts/webkitpy/layout_tests/servers/http_server_base.py: Copied from Tools/Scripts/webkitpy/layout_tests/port/http_server_base.py.
2177         * Scripts/webkitpy/layout_tests/servers/http_server_integrationtest.py: Copied from Tools/Scripts/webkitpy/layout_tests/port/http_server_integrationtest.py.
2178         * Scripts/webkitpy/layout_tests/servers/websocket_server.py: Copied from Tools/Scripts/webkitpy/layout_tests/port/websocket_server.py.
2179
2180 2011-06-27  Eric Seidel  <eric@webkit.org>
2181
2182         Reviewed by Adam Barth.
2183
2184         Remove Chromium and ChromiumMac implementations of driver_name now that everything uses DumpRenderTree
2185         https://bugs.webkit.org/show_bug.cgi?id=63428
2186
2187         * Scripts/webkitpy/layout_tests/port/chromium.py:
2188         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
2189
2190 2011-06-27  Adam Barth  <abarth@webkit.org>
2191
2192         Reviewed by Eric Seidel.
2193
2194         Move apache_http_server.py to webkitpy.layout_package.servers
2195         https://bugs.webkit.org/show_bug.cgi?id=63433
2196
2197         The apache_http_server.py is a server, not a port.
2198
2199         * Scripts/webkitpy/layout_tests/port/apache_http_server.py: Removed.
2200         * Scripts/webkitpy/layout_tests/port/base.py:
2201         * Scripts/webkitpy/layout_tests/servers: Added.
2202         * Scripts/webkitpy/layout_tests/servers/__init__.py: Added.
2203         * Scripts/webkitpy/layout_tests/servers/apache_http_server.py: Copied from Tools/Scripts/webkitpy/layout_tests/port/apache_http_server.py.
2204
2205 2011-06-27  Adam Barth  <abarth@webkit.org>
2206
2207         Reviewed by Eric Seidel.
2208
2209         Move rebaseline-chromium-webkit-tests.py to webkitpy.to_be_moved
2210         https://bugs.webkit.org/show_bug.cgi?id=63432
2211
2212         This code isn't in the dependency cone of new-run-webkit-tests. It's
2213         unclear what it's final home should be, but this patch moves it out of
2214         layout_tests to simply the package.
2215
2216         * Scripts/rebaseline-chromium-webkit-tests:
2217         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py: Removed.
2218         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py: Removed.
2219         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests.py: Copied from Tools/Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py.
2220         * Scripts/webkitpy/to_be_moved/rebaseline_chromium_webkit_tests_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py.
2221
2222 2011-06-27  Adam Barth  <abarth@webkit.org>
2223
2224         Reviewed by Eric Seidel.
2225
2226         Move update_webgl_conformance_tests.py to webkitpy.to_be_moved
2227         https://bugs.webkit.org/show_bug.cgi?id=63429
2228
2229         This package as nothing to do with the layout_tests package.
2230
2231         * Scripts/update-webgl-conformance-tests:
2232         * Scripts/webkitpy/layout_tests/update_webgl_conformance_tests.py: Removed.
2233         * Scripts/webkitpy/layout_tests/update_webgl_conformance_tests_unittest.py: Removed.
2234         * Scripts/webkitpy/to_be_moved/update_webgl_conformance_tests.py: Copied from Tools/Scripts/webkitpy/layout_tests/update_webgl_conformance_tests.py.
2235         * Scripts/webkitpy/to_be_moved/update_webgl_conformance_tests_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/update_webgl_conformance_tests_unittest.py.
2236
2237 2011-06-27  Adam Barth  <abarth@webkit.org>
2238
2239         Reviewed by Eric Seidel.
2240
2241         Move deduplicate_tests.py into webkitpy.to_be_moved
2242         https://bugs.webkit.org/show_bug.cgi?id=63426
2243
2244         deduplicate_tests.py doesn't haven anything to do with layout tests.
2245         It's really a separate script.  This patch moves it to the (new)
2246         to_be_moved package so that it doesn't distract us when working on NRWT.
2247
2248         * Scripts/deduplicate-tests:
2249         * Scripts/webkitpy/layout_tests/deduplicate_tests.py: Removed.
2250         * Scripts/webkitpy/layout_tests/deduplicate_tests_unittest.py: Removed.
2251         * Scripts/webkitpy/test/main.py:
2252         * Scripts/webkitpy/to_be_moved: Added.
2253         * Scripts/webkitpy/to_be_moved/__init__.py: Added.
2254         * Scripts/webkitpy/to_be_moved/deduplicate_tests.py: Copied from Tools/Scripts/webkitpy/layout_tests/deduplicate_tests.py.
2255         * Scripts/webkitpy/to_be_moved/deduplicate_tests_unittest.py: Copied from Tools/Scripts/webkitpy/layout_tests/deduplicate_tests_unittest.py.
2256
2257 2011-06-26  Adam Barth  <abarth@webkit.org>
2258
2259         Rubber-stamped by Eric Seidel.
2260
2261         Remove empty directory.
2262
2263         * Scripts/webkitpy/layout_tests/test_types: Removed.
2264
2265 2011-06-25  Naoki Takano  <honten@chromium.org>
2266
2267         Reviewed by Eric Seidel.
2268
2269         Change my entry from contributors to committers.
2270         https://bugs.webkit.org/show_bug.cgi?id=63043
2271
2272         * Scripts/webkitpy/common/config/committers.py: Change my e-mail from takano.naoki@gmail.com to honten@chromium.org. And move the entry to commiters from contributors.
2273
2274 2011-06-25  Adam Barth  <abarth@webkit.org>
2275
2276         Reviewed by Eric Seidel.
2277
2278         [Chromium] Update WebKit for WEBKIT_FRAME_TO_DOCUMENT_API_MOVE
2279         https://bugs.webkit.org/show_bug.cgi?id=63394
2280
2281         * DumpRenderTree/chromium/LayoutTestController.cpp:
2282         (LayoutTestController::queueLoad):
2283
2284 2011-06-25  Brent Fulgham  <bfulgham@webkit.org>
2285
2286         Unreviewed build correction after r89426.
2287
2288         * WebKitTestRunner/Configurations/WebKitTestRunnerCFLite.vsprops:
2289           Link with Cairo DLL for new functionality.
2290         * WebKitTestRunner/cairo/TestInvocationCairo.cpp: Include compatibility
2291           code for snprintf.
2292
2293 2011-06-25  Dimitri Glazkov  <dglazkov@chromium.org>
2294
2295         Reviewed by Adam Barth.
2296
2297         run-bindings-tests does not work on Chromium Windows.
2298         https://bugs.webkit.org/show_bug.cgi?id=63382
2299
2300         * BuildSlaveSupport/build.webkit.org-config/master.cfg: Removed run-bindings-tests as a build step for cr-win.
2301
2302 2011-06-25  Dan Bernstein  <mitz@apple.com>
2303
2304         Try to fix the Qt and Chromium builds.
2305
2306         * DumpRenderTree/chromium/LayoutTestController.cpp:
2307         (LayoutTestController::setEncodedAudioData):
2308         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2309
2310 2011-06-24  Dan Bernstein  <mitz@apple.com>
2311
2312         Reviewed by Anders Carlsson.
2313
2314         Add an option to enable legacy rounding hacks
2315         https://bugs.webkit.org/show_bug.cgi?id=63363
2316
2317         Added layoutTestController.allowRoundingHacks() and implemented it
2318         on OS X.
2319
2320         * DumpRenderTree/LayoutTestController.cpp:
2321         (allowRoundingHacksCallback):
2322         (LayoutTestController::staticFunctions):
2323         * DumpRenderTree/LayoutTestController.h:
2324         * DumpRenderTree/chromium/LayoutTestController.cpp:
2325         (LayoutTestController::allowRoundingHacks):
2326         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2327         (LayoutTestController::allowRoundingHacks):
2328         * DumpRenderTree/mac/DumpRenderTree.mm:
2329         (resetWebViewToConsistentStateBeforeTesting):
2330         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2331         (LayoutTestController::allowRoundingHacks):
2332         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2333         (LayoutTestController::allowRoundingHacks):
2334         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2335         (LayoutTestController::allowRoundingHacks):
2336         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2337         (LayoutTestController::allowRoundingHacks):
2338
2339 2011-06-24  Adam Roben  <aroben@apple.com>
2340
2341         Include the directory containing the failing tests in titles of bugs filed from TestFailures
2342         when all the tests won't fit
2343
2344         Fixes <http://webkit.org/b/63350> Short bug titles from TestFailures page give no indication
2345         which tests are failing
2346
2347         Reviewed by David Kilzer.
2348
2349         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
2350         (longestCommonPathPrefix): Added. Given a set of paths, returns the longest common prefix
2351         that ends in a path separator.
2352
2353         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
2354         (ViewController.prototype._domForNewAndExistingBugs): When we can't fit all the test names
2355         in the title, first try to include the longest common prefix of the test names, then, if
2356         that's still too long or there's no common prefix, fall back to not including any
2357         information about which tests are failing.
2358
2359 2011-06-24  Adam Roben  <aroben@apple.com>
2360
2361         Include links to Trac in bugs filed from TestFailures
2362
2363         Fixes <http://webkit.org/b/63348> Bugs filed from TestFailures page should include links to
2364         Trac for suspect revisions
2365
2366         Reviewed by David Kilzer.
2367
2368         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Trac.js:
2369         (Trac.prototype.changesetURL): New function, returns the URL for a particular revision in
2370         Trac.
2371
2372         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
2373         (ViewController.prototype._domForNewAndExistingBugs): If there's only one suspect revision,
2374         include a link to Trac for it in the first sentence of the description. Otherwise include a
2375         link to the suspect range after the list of failing tests.
2376
2377 2011-06-24  Nat Duca  <nduca@chromium.org>
2378
2379         Unreviewed, rolling out r89694.
2380         http://trac.webkit.org/changeset/89694
2381         https://bugs.webkit.org/show_bug.cgi?id=58408
2382
2383         Test shell still not ready for animate changes.
2384
2385         * DumpRenderTree/chromium/WebViewHost.cpp:
2386         (WebViewHost::paintInvalidatedRegion):
2387
2388 2011-06-24  Adam Roben  <aroben@apple.com>
2389
2390         Fix typos introduced in r89696
2391
2392         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
2393         (Builder.prototype.getNumberOfFailingTests): Pass the result data's constituent parts to the
2394         callback, since that's what it expects.
2395
2396 2011-06-24  Adam Roben  <aroben@apple.com>
2397
2398         Make TestFailures correctly remember whether old-run-webkit-tests exited early
2399
2400         Fixes <http://webkit.org/b/63342> TestFailures page incorrectly claims test run ran to
2401         completion after reload
2402
2403         Reviewed by David Kilzer.
2404
2405         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
2406         (Builder.prototype.getNumberOfFailingTests): Instead of just storing the number of failures
2407         in the PeristentCache, store an object that contains both the number of failures and whether
2408         old-run-webkit-tests exited early.
2409
2410 2011-06-24  Adam Roben  <aroben@apple.com>
2411
2412         Add links to regression ranges in Trac to the TestFailures page
2413
2414         Fixes <http://webkit.org/b/61060> <rdar://problem/9452153> TestFailures page should provide
2415         links to regression ranges in Trac
2416
2417         Reviewed by David Kilzer.
2418
2419         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Trac.js: Added.
2420         (Trac): This new class represents a particular instance of Trac for a single project.
2421         (Trac.prototype.logURL): Returns the URL that can be used to see the log of the given
2422         repository path between the two specified revisions
2423         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
2424         (ViewController): Added new trac argument.
2425         (ViewController.prototype._displayBuilder): Moved code to create the DOM for the
2426         passed/failed builds from here...
2427         (ViewController.prototype._domForRegressionRange): ...to here. Now also includes a link to
2428         Trac if there are multiple suspect revisions.
2429
2430         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/index.html: Pull in
2431         Trac.js and pass a Trac instance for trac.webkit.org to the ViewController.
2432
2433 2011-05-17  Nat Duca  <nduca@chromium.org>
2434
2435         Reviewed by James Robinson.
2436
2437         [chromium] Implement CCLayerTreeHost and CCLayerTreeHostImpl portions of threaded compositor
2438         https://bugs.webkit.org/show_bug.cgi?id=58408
2439
2440         * DumpRenderTree/chromium/WebViewHost.cpp:
2441         (WebViewHost::paintInvalidatedRegion):
2442
2443 2011-06-24  Dominic Cooney   <dominicc@chromium.org>
2444
2445         Reviewed by Dimitri Glazkov.
2446
2447         Convert shadow DOM-related tests to use window.internals
2448         https://bugs.webkit.org/show_bug.cgi?id=61671
2449
2450         Remove ensureShadowRoot, shadowRoot, shadowPseudoId and
2451         removeShadowRoot from layoutTestController; these have been
2452         replaced by equivalents in window.internals instead.
2453
2454         * DumpRenderTree/LayoutTestController.cpp:
2455         (LayoutTestController::staticFunctions):
2456         * DumpRenderTree/LayoutTestController.h:
2457         * DumpRenderTree/chromium/LayoutTestController.cpp:
2458         (LayoutTestController::LayoutTestController):
2459         * DumpRenderTree/chromium/LayoutTestController.h:
2460         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2461         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2462         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2463         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2464         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2465         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2466         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
2467         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
2468         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
2469
2470 2011-06-24  Csaba Osztrogonác  <ossy@webkit.org>
2471
2472         Reviewed by Dirk Schulze.
2473
2474         Make run-bindings-tests remove its temporary directories.
2475         https://bugs.webkit.org/show_bug.cgi?id=63320
2476
2477         * Scripts/run-bindings-tests:
2478
2479 2011-06-23  MORITA Hajime  <morrita@google.com>
2480
2481         Reviewed by Kent Tamura.
2482
2483         [Chromium] Invalid write inside WebKit::FrameLoaderClientImpl::dispatchDidClearWindowObjectInWorld
2484         https://bugs.webkit.org/show_bug.cgi?id=63216
2485
2486         Deleted m_webViewHost which looks leaked.
2487         
2488         * DumpRenderTree/chromium/TestShell.h:
2489         * DumpRenderTree/chromium/TestShell.cpp:
2490         (TestShell::~TestShell):
2491
2492 2011-06-24  Carlos Garcia Campos  <cgarcia@igalia.com>
2493
2494         Unreviewed. Fix WebKit2 GTK build.
2495
2496         * WebKitTestRunner/GNUmakefile.am:
2497
2498 2011-06-23  Ryuan Choi  <ryuan.choi@samsung.com>
2499
2500         Unreviewed, add myself as committer.
2501
2502         * Scripts/webkitpy/common/config/committers.py:
2503
2504 2011-06-23  Nat Duca  <nduca@chromium.org>
2505
2506         Unreviewed, add myself as committer.
2507
2508         * Scripts/webkitpy/common/config/committers.py:
2509
2510 2011-06-23  Dirk Pranke  <dpranke@chromium.org>
2511
2512         Reviewed by Tony Chang.
2513
2514         nrwt: move http locking code into manager
2515         https://bugs.webkit.org/show_bug.cgi?id=63103
2516
2517         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
2518         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
2519
2520 2011-06-23  Alok Priyadarshi  <alokp@chromium.org>
2521
2522         Reviewed by James Robinson.
2523
2524         [chromium] Add support for running layout tests with accelerated rendering
2525         https://bugs.webkit.org/show_bug.cgi?id=63274
2526
2527         * Scripts/webkitpy/layout_tests/port/chromium.py:
2528         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2529
2530 2011-06-23  Adam Roben  <aroben@apple.com>
2531
2532         Avoid fetching JSON data when possible on TestFailures page to determine if
2533         old-run-webkit-tests exited early
2534
2535         As of r89610, results.html includes information about whether old-run-webkit-tests exited
2536         early, so we don't need to look at JSON data to get this information for builds newer than
2537         that.
2538
2539         Fixes <http://webkit.org/b/63281> TestFailures page fetches build JSON for every build,
2540         which is slow and limits history to the last build master restart
2541
2542         Reviewed by David Kilzer.
2543
2544         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
2545         (Builder.prototype._getFailingTests): Moved code dealing with results.html into a new
2546         function. If we're dealing with a build that has early-exit information in results.html,
2547         don't bother fetching JSON data and go straight to parsing results.html. Otherwise first use
2548         the JSON data to determine if old-run-webkit-tests exited early, and then parse results.html
2549         (as before).
2550
2551 2011-06-23  Eric Seidel  <eric@webkit.org>
2552
2553         Reviewed by Adam Barth.
2554
2555         sheriffbot can no longer restart
2556         https://bugs.webkit.org/show_bug.cgi?id=63221
2557
2558         Fixed and unit tested.
2559
2560         * Scripts/webkitpy/tool/bot/queueengine.py:
2561         * Scripts/webkitpy/tool/bot/sheriffircbot.py:
2562         * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:
2563
2564 2011-06-23  Martin Robinson  <mrobinson@igalia.com>
2565
2566         Reviewed by Xan Lopez.
2567
2568         [GTK] [WebKit2] WebKitTestRunner does not properly load TestNetscapePlugin
2569         https://bugs.webkit.org/show_bug.cgi?id=63287
2570
2571         Send the path to the TestNetscapePlugin to the WKTR via an environment variable
2572         so that the plugin loads properly.
2573
2574         * Scripts/old-run-webkit-tests: Pass the appropriate environment variable to WKTR.
2575         * WebKitTestRunner/gtk/TestControllerGtk.cpp:
2576         (WTR::getEnvironmentVariableAsUTF8String): Added this helper which abstracts away some
2577         of the work for getting an environment variable and converting it to UTF-8.
2578         (WTR::TestController::initializeInjectedBundlePath): Use the helper.
2579         (WTR::TestController::initializeTestPluginDirectory): Get the path from an environment variable now.
2580
2581 2011-06-23  Adam Roben  <aroben@apple.com>
2582
2583         Record whether old-run-webkit-tests aborted early in results.html
2584
2585         Fixes <http://webkit.org/b/63280> results.html gives no indication of whether
2586         old-run-webkit-tests stopped running tests early
2587
2588         Reviewed by David Kilzer.
2589
2590         * Scripts/old-run-webkit-tests:
2591         (top level): Switched to heredoc syntax for printing the start of results.html for clarity.
2592         Added a <style> element with some styles for the new message. If we stopped running early,
2593         put the reason why at the top of results.html.
2594         (stopRunningTestsEarlyIfNeeded): Store the message we print in the new
2595         $stoppedRunningEarlyMessage global.
2596
2597 2011-06-23  Adam Roben  <aroben@apple.com>
2598
2599         Cache some of the TestFailures page's data in localStorage
2600
2601         This makes reloading TestFailures much faster. Right now we only store the number of failing
2602         tests and the list of failing tests for each build. We may choose to store more later, but
2603         it's easy to run up against quota limits.
2604
2605         Fixes <http://webkit.org/b/61520> TestFailures page should take advantage of LocalStorage
2606         APIs (or similar) to improve loading performance
2607
2608         Reviewed by David Kilzer.
2609
2610         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
2611         (Builder.prototype.getNumberOfFailingTests): Changed to use the new PersistentCache object.
2612         (Builder.prototype.startFetchingBuildHistory): Changed to pass whether or not we're still
2613         fetching data to the callback.
2614         (Builder.prototype._getFailingTests): Changed to use the new PersistentCache object. We now
2615         store the tests in the cache just before calling the callback. (The previous code in this
2616         function relied on being able to modify the tests object after storing it in the cache and
2617         having the cached version be updated. This worked while it was a non-serialized cache, but
2618         PersistentCache uses serialization.)
2619
2620         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/PersistentCache.js: Added.
2621         (PersistentCache): This object wraps localStorage. It uses JSON to serialize/deserialize
2622         values, and stores the date that each value was initially stored along with it. This is
2623         later used for pruning the cache.
2624         (PersistentCache.contains): Checks whether the key exists in localStorage.
2625         (PersistentCache.get): Fetch the string we stored in localStorage and extract the original
2626         value out of it.
2627         (PersistentCache.set): Serialize the value, add the date to it, and store it in
2628         localStorage. If this fails due to quota limits, empty the whole cache and try again.
2629         (PersistentCache.prune): Delete any cached data that is deemed old enough.
2630         (PersistentCache._addDateToJSONString): Prepend the current date to the string.
2631         (PersistentCache._emptyCache): Delete everything from localStorage.
2632         (PersistentCache._parseDateAndJSONFromString): Split apart the date and the JSON string and
2633         return them.
2634
2635         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
2636         (ViewController.prototype._displayBuilder): Updated for change to callback signature. When
2637         we finish fetching data, prune the PersistentCache. While I was here I also fixed a bug
2638         where we'd never show the new bug link for tests for which we couldn't determine a passing
2639         revision.
2640
2641         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/index.html: Pull in
2642         PersistentCache.js.
2643
2644 2011-06-23  Adam Roben  <aroben@apple.com>
2645
2646         Make finding existing bugs and filing new bugs work on TestFailures even when lots of tests are failing
2647
2648         One bug this fixes is <http://webkit.org/b/61660> New bug links on TestFailures page often
2649         contain titles that are so long they are rejected by Bugzilla
2650
2651         Reviewed by David Kilzer.
2652
2653         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Bugzilla.js:
2654         (Bugzilla.prototype.quickSearch): Use fetchResource to POST the search query rather than
2655         using a query string on the URL. If the search query is very long, using a query string can
2656         cause the request to be rejected due to the URL being too long. POSTing avoids this issue.
2657         (Bugzilla.maximumBugTitleLength): Added this constant based on Bugzilla's implementation.
2658
2659         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
2660         (.new-bug-form): Hide the form that we secretly use to file a new bug.
2661
2662         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
2663         (fetchResource): Added. Code came from getResource. If we're using a GET request, add the
2664         query parameters to the URL. Otherwise, send them as the body of the request along with the
2665         appropriate headers.
2666         (getResource): Now just calls through to fetchResource.
2667         (urlEncodedQueryParameters): Added. Moved some code here...
2668         (addQueryParametersToURL): ...from here.
2669
2670         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
2671         (ViewController.prototype._domForNewAndExistingBugs): Shorten the title to just mention the
2672         number of failing tests if mentioning all the tests would make the title too long. Improved
2673         the description for large numbers of failures by listing the tests one-per-line instead of
2674         just having them be comma-separated, which was hard to read. Use a form to file the new bug
2675         instead of a link so that we can POST the form data. (Using a URL with a query string can
2676         result in the URL being too long.) The new bug link now just submits the form.
2677
2678 2011-06-23  Adam Roben  <aroben@apple.com>
2679
2680         Show closed bugs on the TestFailures page in addition to open ones
2681
2682         Fixes <http://webkit.org/b/63194> TestFailures page should show closed bugs, too
2683
2684         Reviewed by David Kilzer.
2685
2686         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Bugzilla.js:
2687         (Bugzilla.prototype.quickSearch): Added code to extract the bug's status and store it in the
2688         returned data.
2689         (Bugzilla.isOpenStatus): New function, returns true if the passed-in status indicates that
2690         the associated bug is still open.
2691
2692         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
2693         (addQueryParametersToURL): Add a missing semicolon.
2694         (Node.prototype.appendChildren): New function, like appendChild but takes an array-like
2695         object and appends each of the values stored within.
2696
2697         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
2698         (ViewController.prototype._domForNewAndExistingBugs): Prepend 'ALL' to the query so closed
2699         bugs will be included in the results. Split the returned bugs into two sets: those which are
2700         open, and those which are closed. Put the open bugs at the top level of the list, and the
2701         closed bugs in a second level.
2702
2703 2011-06-23  Adam Roben  <aroben@apple.com>
2704
2705         Don't count new tests as failures on the TestFailures page
2706
2707         Fixes <http://webkit.org/b/63254> TestFailures page calls new tests "failures", even though
2708         they aren't
2709
2710         Reviewed by Anders Carlsson.
2711
2712         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Builder.js:
2713         (Builder.prototype.getNumberOfFailingTests): Don't add the number of new tests to the number
2714         of failures.
2715
2716 2011-06-23  Dirk Pranke  <dpranke@chromium.org>
2717
2718         Reviewed by Eric Seidel.
2719
2720         nrwt: make TestInput objects printable
2721         https://bugs.webkit.org/show_bug.cgi?id=63225
2722
2723         This makes debugging slighlty easier.
2724
2725         * Scripts/webkitpy/layout_tests/layout_package/test_input.py:
2726
2727 2011-06-22  Daniel Bates  <dbates@webkit.org>
2728
2729         Reviewed by Adam Barth.
2730
2731         test-webkitpy --all errors out because scm_unittest.py can't find module checkout
2732         https://bugs.webkit.org/show_bug.cgi?id=62943
2733
2734         Remove "from .checkout import Checkout" from scm_unittest.py since it's included
2735         by Scripts/webkitpy/common/checkout/__init__.py.
2736
2737         * Scripts/webkitpy/common/checkout/scm/scm_unittest.py:
2738
2739 2011-06-22  Dominic Cooney  <dominicc@chromium.org>
2740
2741         Reviewed by Mark Rowe.
2742
2743         Add window.internals to WebKit2's WebKitTestRunner.
2744         https://bugs.webkit.org/show_bug.cgi?id=61073
2745
2746         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pxbproj:
2747         * WebKitTestRunner/Configurations/Base.xcconfig:
2748         * WebKitTestRunner/Configurations/InjectedBundleCFLite.vsprops:
2749         * WebKitTestRunner/Configurations/InjectedBundleCommon.vsprops:
2750         * WebKitTestRunner/Configurations/InjectedBundleCoreFoundation.vsprops:
2751         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2752         (WTR::InjectedBundlePage::didClearWindowForFrame):
2753         * WebKitTestRunner/InjectedBundle/qt/InjectedBundle.pro:
2754         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
2755
2756 2011-06-22  Eric Seidel  <eric@webkit.org>
2757
2758         Reviewed by Adam Barth.
2759
2760         Make sheriff-bot rollout messages a little nicer
2761         https://bugs.webkit.org/show_bug.cgi?id=63107
2762
2763         itertools.chain.from_iterable is new in 2.6,
2764         use itertools.chain(*list) for 2.5 compat.
2765
2766         * Scripts/webkitpy/tool/bot/irc_command.py:
2767
2768 2011-06-22  Eric Seidel  <eric@webkit.org>
2769
2770         Reviewed by Adam Barth.
2771
2772         Make sheriff-bot robust against exceptions from commands
2773         https://bugs.webkit.org/show_bug.cgi?id=63211
2774
2775         sheriff-bot was acting strange this afternoon.
2776         We don't know if this will fix the cause, but
2777         at least it adds some unit tests and catches
2778         one possible cause.
2779
2780         * Scripts/webkitpy/tool/bot/sheriffircbot.py:
2781         * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:
2782
2783 2011-06-22  Mark Rowe  <mrowe@apple.com>
2784
2785         Reviewed by David Levin.
2786
2787         <http://webkit.org/b/63212> TestWebKitAPI Xcode project has bogus settings since r86287
2788
2789         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Fix the reference to libgtest.a to
2790         be relative to the built products directory rather than using a hard-coded path. Remove the
2791         explicit settings of FRAMEWORK_SEARCH_PATHS, HEADER_SEARCH_PATHS and LIBRARY_SEARCH_PATHS
2792         as they're all unnecessary. If they were necessary they should be set in the .xcconfig file
2793         rather than in the .xcodeproj directly.
2794
2795 2011-06-22  Dirk Pranke  <dpranke@chromium.org>
2796
2797         Reviewed by Ojan Vafai.
2798
2799         new-run-webkit-tests: remove obsolete port.shut_down_http_server method
2800         https://bugs.webkit.org/show_bug.cgi?id=59993
2801
2802         * Scripts/webkitpy/layout_tests/port/base.py:
2803         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
2804         * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
2805         * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
2806         * Scripts/webkitpy/layout_tests/port/chromium_win.py:
2807         * Scripts/webkitpy/layout_tests/port/gtk.py:
2808         * Scripts/webkitpy/layout_tests/port/mac.py:
2809         * Scripts/webkitpy/layout_tests/port/mock_drt_unittest.py:
2810         * Scripts/webkitpy/layout_tests/port/qt.py:
2811         * Scripts/webkitpy/layout_tests/port/win.py:
2812
2813 2011-06-22  Luke Macpherson   <macpherson@chromium.org>
2814
2815         Reviewed by Andreas Kling.
2816
2817         Move macpherson from contributor list to committer list.
2818         https://bugs.webkit.org/show_bug.cgi?id=63179
2819
2820         * Scripts/webkitpy/common/config/committers.py:
2821         Move macpherson from contributor list to committer list.
2822
2823 2011-06-22  David Levin  <levin@chromium.org>
2824
2825         Reviewed by Adam Barth.
2826
2827         check-webkit-style should detect returning (Own|Ref)Ptr instead of the Pass*Ptr version.
2828         https://bugs.webkit.org/show_bug.cgi?id=63204
2829
2830         * Scripts/webkitpy/style/checkers/cpp.py: Added a check for the return value and combined
2831           with similar code for the parameter checking.
2832         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Removed pass_ptr checks from
2833           those done for single lines since they don't make sense in that case (variable decls look like function decls).
2834           Removed some redundant comments (one of which was slightly wrong).
2835           Added checks for the new functionality and minor other test changes.
2836
2837 2011-06-22  Nate Chapin  <japhet@chromium.org>
2838
2839         Reviewed by Adam Barth.
2840
2841         Add libsoup to libWebCoreInternals build.
2842         https://bugs.webkit.org/show_bug.cgi?id=62066
2843
2844         * GNUmakefile.am:
2845
2846 2011-06-22  Dirk Pranke  <dpranke@chromium.org>
2847
2848         Reviewed by Tony Chang.
2849
2850         nrwt: handle missing httpd cleanly
2851         https://bugs.webkit.org/show_bug.cgi?id=62027
2852
2853         We had reverted the change in r89414, so this adds it back in
2854         (modifying check_sys_deps() in port/base.py to check for an
2855         installed web server).
2856
2857         This change then fixes the change in r89414 to stub out the
2858         check_sys_deps() in the test port so that the unit tests run correctly.
2859
2860         * Scripts/webkitpy/layout_tests/port/base.py:
2861         * Scripts/webkitpy/layout_tests/port/test.py:
2862
2863 2011-06-22  David Levin  <levin@chromium.org>
2864
2865         Reviewed by Adam Barth.
2866
2867         check-webkit-style should check for invalid uses of RefPtr/OwnPtr as parameters.
2868         https://bugs.webkit.org/show_bug.cgi?id=63188
2869
2870         * Scripts/webkitpy/style/checkers/cpp.py: Added the check.
2871         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added tests.
2872
2873 2011-06-22  Tony Chang  <tony@chromium.org>
2874
2875         Reviewed by Ojan Vafai.
2876
2877         Add a buildbot for testing ENABLE(CSS3_FLEXBOX)
2878         https://bugs.webkit.org/show_bug.cgi?id=62891
2879
2880         * BuildSlaveSupport/build.webkit.org-config/config.json:
2881
2882 2011-06-22  Dirk Pranke  <dpranke@chromium.org>
2883
2884         Reviewed by Ojan Vafai.
2885
2886         nrwt: don't look for http lock when running the test port
2887         https://bugs.webkit.org/show_bug.cgi?id=63158
2888
2889         * Scripts/webkitpy/layout_tests/port/mock_drt.py:
2890         * Scripts/webkitpy/layout_tests/port/test.py:
2891
2892 2011-06-22  Balazs Kelemen  <kbalazs@webkit.org>
2893
2894         Reviewed by Andreas Kling.
2895
2896         [Qt][WK2] Set up plugin tests
2897         https://bugs.webkit.org/show_bug.cgi?id=63066
2898
2899         * WebKitTestRunner/qt/TestControllerQt.cpp:
2900         (WTR::TestController::initializeTestPluginDirectory):
2901         Use the QTWEBKIT_PLUGIN_PATH environment variable as the path
2902         of the test plugin like DRT does.
2903
2904 2011-06-22  Eric Seidel  <eric@webkit.org>
2905
2906         Reviewed by Ojan Vafai.
2907
2908         Make sheriff-bot rollout messages a little nicer
2909         https://bugs.webkit.org/show_bug.cgi?id=63107
2910
2911         It annoyed me this afternoon that I had to convert sheriff-bots "r12345" revisions
2912         into urls myself.  So I have now fixed its "preparing" message to include a url.
2913
2914         I also figured that I should make the messages mention all of the responsible parties
2915         so that rollouts are never surprises.  If you're in the channel and were involved
2916         in a patch, you will see if someone is using sheriff-bot to rollout a patch.
2917
2918         As part of doing this I also changed (and tested) _parse_args to fail-fast
2919         when given invalid args.
2920
2921         * Scripts/webkitpy/tool/bot/irc_command.py:
2922         * Scripts/webkitpy/tool/bot/irc_command_unittest.py:
2923         * Scripts/webkitpy/tool/bot/sheriff.py:
2924         * Scripts/webkitpy/tool/bot/sheriffircbot_unittest.py:
2925
2926 2011-06-22  Dirk Pranke  <dpranke@chromium.org>
2927
2928         Reviewed by Tony Chang.
2929
2930         run-webkit-websocketserver fails to stop websocket server
2931         https://bugs.webkit.org/show_bug.cgi?id=63123
2932
2933         There's a ten-character change in new-run-webkit-websocketserver
2934         that fixes the actual bug, and then we update the code in
2935         new-run-webkit-httpd to do the actual stop call, and add more tests
2936         (refactoring the existing test code so that it is shared).
2937
2938         * Scripts/new-run-webkit-httpd:
2939         * Scripts/new-run-webkit-websocketserver:
2940         * Scripts/webkitpy/layout_tests/port/http_server_integrationtest.py:
2941
2942 2011-06-22  Adam Roben  <aroben@apple.com>
2943
2944         Add links to existing bugs related to failing tests on TestFailures page
2945
2946         I changed the layout of the page a little to make it easier to read with all the new
2947         information. Passing/failing revisions have been moved down below the list of tests to be
2948         closer to the existing bugs and the new bug link. And each set of tests and its relevant
2949         information is in a light gray box.
2950
2951         Fixes <http://webkit.org/b/61665> TestFailures page should link to existing bugs when
2952         possible
2953
2954         Reviewed by Darin Adler.
2955
2956         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Bugzilla.js: Added.
2957         (Bugzilla): This new class represents a single Bugzilla instance.
2958         (Bugzilla.prototype.quickSearch): Searches Bugzilla using its Quick Search functionality,
2959         passing the resulting bug titles and URLs to the callback when complete. If called multiple
2960         times with the same query before the query returns, caches the callbacks so that only one
2961         query is sent over the wire. When the query completes, all pending callbacks are called.
2962
2963         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/TestFailures.css:
2964         (#failure-history): Reduce the margin/padding on the top-level list a bit.
2965         (#failure-history > li): Put each set of tests in a gray box, and indent most information
2966         inside the box.
2967         (.test-list): Unindent the list of failing tests so it is visually at the top level.
2968         (.new-and-existing-bugs): Reduce the space at the bottom of this area so that the bottom of
2969         each box isn't a big empty space.
2970         (.existing-bugs-list): Use a smaller text size for existing bugs, since their titles can be
2971         quite long.
2972
2973         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/Utilities.js:
2974         (addQueryParametersToURL): New function extracted from
2975         ViewController.prototype._domForNewAndExistingBugs.
2976
2977         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/ViewController.js:
2978         (ViewController): Take and store an optional Bugzilla instance.
2979         (ViewController.prototype._displayBuilder): Give the top-level list an id for styling
2980         purposes and move the list of failing tests above all other information. Only show bug
2981         information once we've determined the most-recent passing revision for a set of tests. It's
2982         not that useful to file a new bug before this information has been determined, and searching
2983         for existing bugs before we've figured out which tests started failing at the same time
2984         would end up giving you information about a bunch of unrelated tests.
2985         (ViewController.prototype._domForNewAndExistingBugs): Renamed from _domForNewBugLink. Now
2986         returns a DocumentFragment instead of an HTMLParagraphElement. If we don't have a Bugzilla
2987         instance, just returns an empty DocumentFragment. Starts a search for bugs related to the
2988         failing tests, and adds links to the bugs when the search completes.
2989
2990         * BuildSlaveSupport/build.webkit.org-config/public_html/TestFailures/index.html: Pass a
2991         Bugzilla instance for bugs.webkit.org to the ViewController.
2992
2993 2011-06-22  Carlos Garcia Campos  <cgarcia@igalia.com>
2994
2995         Reviewed by Martin Robinson.
2996
2997         [GTK] Apply current view settings to newly created views in GtkLauncher
2998         https://bugs.webkit.org/show_bug.cgi?id=63142
2999
3000         * GtkLauncher/main.c:
3001         (createWebViewCb):
3002
3003 2011-06-22  Carlos Garcia Campos  <cgarcia@igalia.com>
3004
3005         Reviewed by Martin Robinson.
3006
3007         [GTK] Use GOption to parse main arguments in GtkLauncher
3008         https://bugs.webkit.org/show_bug.cgi?id=63060
3009
3010         Instead of getting the settings from the view, create a new
3011         settings object not associated with any view so that we can parse
3012         the command line arguments before creating the view. It will also
3013         allows to use the same settings for more than one view if we
3014         eventually support passing more than one uri from the command
3015         line.
3016
3017         * GtkLauncher/main.c:
3018         (parseOptionEntryCallback): Make this function static. Also use
3019         g_ascii_strtoll/g_ascii_strtod to parse int/float values instead
3020         of atoi/atof to provide better error messages. And always fill the
3021         error struct when the function fails.
3022         (getOptionEntriesFromWebKitWebSettings): Use an array of
3023         GOptionEntry structs instead of a GArray, since that is what
3024         g_option_group_add_entries() expects and the array size doesn't
3025         change.
3026         (addWebSettingsGroupToContext): Helper function to add the
3027         websettings group to the option context.
3028         (main): Create the option context to parse command line arguments.
3029
3030 2011-06-22  Martin Robinson  <mrobinson@igalia.com>
3031
3032         Reviewed by Adam Roben.
3033
3034         [GTK] Implement pixel dump support for WebKitTestRunner
3035         https://bugs.webkit.org/show_bug.cgi?id=58242
3036
3037         * DumpRenderTree/gtk/GNUmakefile.ImageDiff.am: Added.
3038         * GNUmakefile.am: Separate out the ImageDiff bits of the GNUmakefile.
3039         * Scripts/old-run-webkit-tests: Update to add support for the GTK+ port.
3040         * WebKitTestRunner/GNUmakefile.am: Added Cairo specific files to the build.
3041         * WebKitTestRunner/TestInvocation.cpp: Moved this code to the platform-independent
3042         file so that we can share it across platforms.
3043         (WTR::TestInvocation::compareActualHashToExpectedAndDumpResults):
3044         * WebKitTestRunner/TestInvocation.h: Updated method list.
3045         * WebKitTestRunner/cairo/TestInvocationCairo.cpp: Added Cairo implementation.
3046         * WebKitTestRunner/cg/TestInvocationCG.cpp:
3047         (WTR::TestInvocation::dumpPixelsAndCompareWithExpected): Updated to use new helper.
3048
3049 2011-06-22  Sheriff Bot  <webkit.review.bot@gmail.com>
3050
3051         Unreviewed, rolling out r89414.
3052         http://trac.webkit.org/changeset/89414
3053         https://bugs.webkit.org/show_bug.cgi?id=63130
3054
3055         It broke webkitpy tests (Requested by Ossy on #webkit).
3056
3057         * Scripts/webkitpy/layout_tests/port/base.py:
3058
3059 2011-06-22  Kristóf Kosztyó  <Kosztyo.Kristof@stud.u-szeged.hu>
3060
3061         Reviewed by Dirk Pranke.
3062
3063         [NRWT] Print error message when there is no httpd present on the system
3064         https://bugs.webkit.org/show_bug.cgi?id=62027
3065
3066         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
3067
3068 2011-06-21  Dirk Pranke  <dpranke@chromium.org>
3069
3070         Unreviewed, build fix.
3071         Fix crashes in new-run-webkit-tests resulting from the
3072         change to the http_server logic in r89400. Python 2.5
3073         on Mac 10.5 has some weird error paths.
3074
3075         * Scripts/webkitpy/layout_tests/port/http_server_base.py:
3076         * Scripts/webkitpy/common/system/executive.py:
3077
3078 2011-06-21  MORITA Hajime  <morrita@google.com>
3079
3080         Unreviewed, rolling out r89401 and r89403.
3081         http://trac.webkit.org/changeset/89401
3082         http://trac.webkit.org/changeset/89403
3083         https://bugs.webkit.org/show_bug.cgi?id=62970
3084
3085         Breaks mac build and mistakenly enables the spellcheck API
3086
3087         * Scripts/build-webkit:
3088         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
3089
3090 2011-06-21  Kent Tamura  <tkent@chromium.org>
3091
3092         [Mac] Sort Xcode project files.
3093
3094         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
3095
3096 2011-06-20  MORITA Hajime  <morrita@google.com>
3097
3098         Reviewed by Kent Tamura.
3099
3100         Spellcheck API should be build-able.
3101         https://bugs.webkit.org/show_bug.cgi?id=62970
3102
3103         * Scripts/build-webkit:
3104
3105 2011-06-21  Dirk Pranke  <dpranke@chromium.org>
3106
3107         Reviewed by Tony Chang.
3108
3109         nrwt: attempt #3 at fixing server startup/shutdown
3110         https://bugs.webkit.org/show_bug.cgi?id=62829
3111
3112         Attempt yet again to make starting and stopping servers
3113         reliable. It turns out that apache has races between when the
3114         ctl process exists and when it writes/removes the pid file.
3115         This change accounts for the races.
3116
3117         * Scripts/new-run-webkit-httpd:
3118         * Scripts/webkitpy/common/system/executive.py:
3119         * Scripts/webkitpy/layout_tests/port/apache_http_server.py:
3120         * Scripts/webkitpy/layout_tests/port/base.py:
3121         * Scripts/webkitpy/layout_tests/port/http_server.py:
3122         * Scripts/webkitpy/layout_tests/port/http_server_base.py:
3123         * Scripts/webkitpy/layout_tests/port/http_server_integrationtest.py: Added.
3124         * Scripts/webkitpy/layout_tests/port/port_testcase.py:
3125         * Scripts/webkitpy/layout_tests/port/websocket_server.py:
3126
3127 2011-06-21  Dmitry Lomov  <dslomov@google.com>
3128
3129         Reviewed by Adam Roben.
3130
3131         https://bugs.webkit.org/show_bug.cgi?id=63080
3132         Unit-tests step on test-only bot is broken.
3133         Removing the step until run-api-tests is fixed.
3134
3135         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3136
3137 2011-06-20  Dmitry Lomov  <dslomov@google.com>
3138
3139         Reviewed by David Levin.
3140
3141         https://bugs.webkit.org/show_bug.cgi?id=63017
3142         "Unit tests" step on bots should provide more information in case of failure.
3143
3144         * BuildSlaveSupport/build.webkit.org-config/master.cfg: added --verbose flag to UnitTests step.
3145
3146 2011-06-20  Dirk Pranke  <dpranke@chromium.org>
3147
3148         Reviewed by Ojan Vafai.
3149
3150         new-run-webkit-tests: order shards by name, not number of tests
3151         https://bugs.webkit.org/show_bug.cgi?id=62753
3152
3153         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3154
3155 2011-06-20  Dirk Pranke  <dpranke@chromium.org>
3156
3157         Reviewed by Ojan Vafai.
3158
3159         we should only log the test names to the tests_run.txt files during layout_tests
3160         https://bugs.webkit.org/show_bug.cgi?id=62751
3161
3162         * Scripts/old-run-webkit-tests:
3163         * Scripts/webkitpy/layout_tests/layout_package/worker.py:
3164
3165 2011-06-20  Eric Seidel  <eric@webkit.org>
3166
3167         Reviewed by Adam Barth.
3168
3169         commit-queue/ews-bots should run run-bindings-tests for every patch
3170         https://bugs.webkit.org/show_bug.cgi?id=63010
3171
3172         * Scripts/webkitpy/common/config/ports.py:
3173         * Scripts/webkitpy/tool/commands/download_unittest.py:
3174         * Scripts/webkitpy/tool/steps/runtests.py:
3175         * Scripts/webkitpy/tool/steps/runtests_unittest.py:
3176         * Scripts/webkitpy/tool/steps/steps_unittest.py:
3177
3178 2011-06-20  Chang Shu  <cshu@webkit.org>
3179
3180         Reviewed by Adam Barth.
3181
3182         fast/js/sputnik/Conformance/10_Execution_Contexts/10.2_Entering_An_Execution_Context/10.2.2_Eval_Code tests fail when showModalDialog isn't supported
3183         https://bugs.webkit.org/show_bug.cgi?id=53676
3184
3185         * WebKitTestRunner/TestController.cpp:
3186         (WTR::TestController::initialize):
3187
3188 2011-06-20  Eric Seidel  <eric@webkit.org>
3189
3190         Reviewed by Adam Barth.
3191
3192         buildbots should run run-bindings-tests
3193         https://bugs.webkit.org/show_bug.cgi?id=63001
3194
3195         run-bindings-tests takes 2.4 seconds on my machine, so
3196         I expect this will not be noticeable on the bots.
3197
3198         I also removed tiger-specific checks since Tiger is no
3199         longer a supported platform for webkit.
3200
3201         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3202
3203 2011-06-20  Carlos Garcia Campos  <cgarcia@igalia.com>
3204
3205         Unreviewed. Fix WebKit2 GTK build after r89249.
3206
3207         * GtkLauncher/main.c:
3208         (main):
3209
3210 2011-06-20  Pavel Feldman  <pfeldman@chromium.org>
3211
3212         Reviewed by Yury Semikhatsky.
3213
3214         Web Inspector: remove LayoutTestController::setTimelineProfilingEnabled.
3215         https://bugs.webkit.org/show_bug.cgi?id=62994
3216
3217         * DumpRenderTree/LayoutTestController.cpp:
3218         (closeWebInspectorCallback):
3219         (LayoutTestController::staticFunctions):
3220         * DumpRenderTree/LayoutTestController.h:
3221         * DumpRenderTree/chromium/DRTDevToolsAgent.cpp:
3222         * DumpRenderTree/chromium/DRTDevToolsAgent.h:
3223         * DumpRenderTree/chromium/LayoutTestController.cpp:
3224         (LayoutTestController::LayoutTestController):
3225         * DumpRenderTree/chromium/LayoutTestController.h:
3226         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3227         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
3228         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3229         * DumpRenderTree/qt/LayoutTestControllerQt.h:
3230         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
3231         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
3232         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
3233         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
3234         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
3235
3236 2011-06-20  Lukasz Slachciak  <l.slachciak@samsung.com>
3237
3238         Reviewed by Martin Robinson.
3239
3240         [GTK] General mechanism for adjusting WebKitWebSettings in GtkLauncher.
3241         https://bugs.webkit.org/show_bug.cgi?id=55308
3242
3243         * GtkLauncher/main.c: Added general mechanism for adjusting WebKitWebSettings.
3244         (parseOptionEntryCallback): Callback for parsing option entry.
3245         (getOptionEntriesFromWebKitWebSettings): Basing on the WebKitWebSettings properties prepare
3246         list of option entries.
3247         (transformStringToBoolean): Transform function for boolean convertion used by parseOptionEntryCallback.
3248         (transformStringToInt): Transform function for int convertion used by parseOptionEntryCallback.
3249         (transformStringToFloat): Transform function for float convertion used by parseOptionEntryCallback.
3250         (parseAdditionalOptions): Help function parsing additional commandline options.
3251         (main): Added call to parseAdditionalOptions.
3252
3253 2011-06-20  Viatcheslav Ostapenko  <ostapenko.viatcheslav@nokia.com>
3254
3255         Reviewed by Andreas Kling.
3256
3257         [Qt] [WK2] Compile error in ActivateFontsQt.cpp on Mac
3258         https://bugs.webkit.org/show_bug.cgi?id=62869
3259
3260         Add QFont include to fix "incomlete type QFont" error.
3261
3262         * WebKitTestRunner/InjectedBundle/qt/ActivateFontsQt.cpp:
3263
3264 2011-06-19  Adam Bergkvist  <adam.bergkvist@ericsson.com>
3265
3266         Reviewed by Martin Robinson.
3267
3268         [GTK] Enable Media Stream feature and make it default on
3269         https://bugs.webkit.org/show_bug.cgi?id=60394
3270
3271         * Scripts/build-webkit:
3272
3273 2011-06-19  Sheriff Bot  <webkit.review.bot@gmail.com>
3274
3275         Unreviewed, rolling out r89198.
3276         http://trac.webkit.org/changeset/89198
3277         https://bugs.webkit.org/show_bug.cgi?id=62953
3278
3279         Chromium breakage on "Linux Tests" (Requested by senorblanco
3280         on #webkit).
3281
3282         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
3283         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
3284         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
3285
3286 2011-06-18  Daniel Bates  <dbates@webkit.org>
3287
3288         Reviewed by Eric Seidel.
3289
3290         Git.push_local_commits_to_server() should use passed username and
3291         password when we don't have cached credentials
3292         https://bugs.webkit.org/show_bug.cgi?id=62941
3293
3294         Fixes a issue where Git.push_local_commits_to_server() would always look
3295         for cached credentials regardless of whether a username and password
3296         were passed as arguments. Instead, if given both a username and password
3297         then we shouldn't look for cached credentials (since we were explicitly given
3298         credentials).
3299
3300         * Scripts/webkitpy/common/checkout/scm/git.py:
3301         * Scripts/webkitpy/common/checkout/scm/scm_unittest.py: Added unit tests.
3302
3303 2011-06-18  Daniel Bates  <dbates@rim.com>
3304
3305         Rubber-stamped by Eric Seidel.
3306
3307         Remove run-javascriptcore-tests --skip-build option; Instead add --build/no-build options
3308         https://bugs.webkit.org/show_bug.cgi?id=62767
3309
3310         Part 3 of 3
3311
3312         Remove the command-line option --skip-build now that the master BuildBot has been restarted
3313         following <http://trac.webkit.org/changeset/89138>.
3314
3315         Instead, specify --no-build to get analogous functionality.
3316
3317         * Scripts/run-javascriptcore-tests:
3318
3319 2011-06-18  Dirk Pranke  <dpranke@chromium.org>
3320
3321         Reviewed by Tony Chang.
3322
3323         new-run-webkit-tests: fix test sorting
3324         https://bugs.webkit.org/show_bug.cgi?id=58691
3325
3326         This changes the sorting to use a "natural" sort order such that
3327         foo2.html sorts before foo13.html, and sorts the tests so
3328         that all the tests in a directory are run before any test in a
3329         subdirectory.
3330
3331         This change also adds a split() method to the filesystem since
3332         that is needed to cluster files into a directory otherwise.
3333
3334         * Scripts/webkitpy/common/system/filesystem.py:
3335         * Scripts/webkitpy/common/system/filesystem_mock.py:
3336         * Scripts/webkitpy/layout_tests/layout_package/manager.py:
3337         * Scripts/webkitpy/layout_tests/layout_package/manager_unittest.py:
3338
3339 2011-06-18  Evan Martin  <evan@chromium.org>
3340
3341         Reviewed by Darin Adler.
3342
3343         Compiler warning in double->int conversion in TestNetscapePlugIn
3344         https://bugs.webkit.org/show_bug.cgi?id=51554
3345
3346         This code is providing a JS API for removing a property by numeric ID.
3347         We get it from JS as a double, and must convert it to an int before using it.
3348
3349         Test: LayoutTests/plugins/npruntime/remove-property.html
3350
3351         * DumpRenderTree/TestNetscapePlugIn/Tests/NPRuntimeRemoveProperty.cpp:
3352         (NPRuntimeRemoveProperty::TestObject::invoke):
3353
3354 2011-06-18  Patrick Gansterer  <paroga@webkit.org>
3355
3356         Reviewed by Eric Seidel.
3357
3358         Don't exit webkit-patch when running on Win32 python
3359         https://bugs.webkit.org/show_bug.cgi?id=55811
3360
3361         Most of the basic actions work in the meantime,
3362         so let people play with in on Win32 now.
3363
3364         * Scripts/webkit-patch:
3365
3366 2011-06-18  Alice Boxhall  <aboxhall@chromium.org>
3367
3368         Reviewed by Ojan Vafai.
3369
3370         Convert json_results_generator.py to output version 4 JSON.
3371         https://bugs.webkit.org/show_bug.cgi?id=60869
3372
3373         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
3374         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
3375         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
3376
3377 2011-06-18  Berend-Jan Wever  <skylined@chromium.org>
3378
3379         Reviewed by Eric Seidel.
3380
3381         Some names with spaces are incorrectly reported as not containing spaces
3382         while running prepare-ChangeLog
3383         https://bugs.webkit.org/show_bug.cgi?id=55572
3384
3385         * Scripts/VCSUtils.pm: Updated CHANGE_LOG_NAME check for spaces to allow
3386           non-alphanumeric characters in names.
3387
3388 2011-06-18  Michael Saboff  <msaboff@apple.com>
3389
3390         Reviewed by Eric Seidel.
3391
3392         Enhancement: New Script to Format malloc_history output as tree
3393         https://bugs.webkit.org/show_bug.cgi?id=56378
3394
3395         New script to process malloc_history output to tree format.
3396
3397         * Scripts/malloc-tree: Added.
3398
3399 2011-06-17  Dmitry Lomov  <dslomov@google.com>
3400
3401         Reviewed by Daniel Bates.
3402
3403         https://bugs.webkit.org/show_bug.cgi?id=62918 
3404         Pass configuration to UnitTests step in master.cfg
3405
3406         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3407
3408 2011-06-17  Leandro Pereira  <leandro@profusion.mobi>
3409
3410         Reviewed by Eric Seidel.
3411
3412         [EFL] Add Evas-based ImageDiff implementation.
3413         https://bugs.webkit.org/show_bug.cgi?id=61942
3414
3415         Build system changes will be performed as soon as the complete
3416         DumpRenderTree implementation is upstreamed.
3417
3418         * DumpRenderTree/efl/ImageDiff.cpp: Added.
3419         (pixelFromImageData):
3420         (differenceImageFromDifferenceBuffer):
3421         (computeDistanceBetweenPixelComponents):
3422         (calculatePixelDifference):
3423         (calculateDifference):
3424         (getTemporaryFile):
3425         (printImage):
3426         (printImageDifferences):
3427         (resizeEcoreEvasIfNeeded):
3428         (readImageFromStdin):
3429         (parseCommandLineOptions):
3430         (shutdownEfl):
3431         (abortWithErrorMessage):
3432         (initEfl):
3433         (main):
3434
3435 2011-06-17  Tony Chang  <tony@chromium.org>
3436
3437         Reviewed by Dirk Pranke.
3438
3439         [chromium] force rebaseline-chromium-webkit-tests to use the chromium port
3440         https://bugs.webkit.org/show_bug.cgi?id=62907
3441
3442         This fixes a bug where the ImageDiff on the user's machine doesn't
3443         match the logic used on the Chromium buildbots.  This has the downside
3444         of requiring that the Chromium ImageDiff is built, but it turns out
3445         that even with --tolerance 0, the CG ImageDiff will allow small pixel
3446         differences through.
3447
3448         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
3449
3450 2011-06-17  Brent Fulgham  <bfulgham@webkit.org>
3451
3452         [WinCairo] Unreviewed build correction.
3453
3454         * DumpRenderTree/win/DumpRenderTreeCairo.vsprops:
3455         Update WinCairo property sheet to link against the Cairo DLL,
3456         now that we are using a more recent Cairo build. 
3457
3458 2011-06-17  Brent Fulgham  <bfulgham@webkit.org>
3459
3460         Reviewed by Brian Weinstein.
3461
3462         Create a simple example of using the COM-based C++ listeners.
3463         A new onclick event is attached (via C++) to the WebKit logo
3464         in the default test pattern displayed on startup.  Clicking on
3465         the logo causes a message box to be displayed.
3466         https://bugs.webkit.org/show_bug.cgi?id=61885
3467
3468         * WinLauncher/DOMDefaultImpl.h: Added.  Stub implementation of
3469           the WebScriptObject and DOMEventListener.
3470         * WinLauncher/WinLauncher.cpp:
3471         (SimpleEventListener::SimpleEventListener): Example implementation
3472           of a simple DOM event listener.
3473         (SimpleEventListener::handleEvent): 
3474         (WinLauncherWebHost::didFinishLoadForFrame): Added implementation
3475          to bind a C++ method to the 'onclick' event for the WebKit logo.
3476         (_tWinMain):
3477         * WinLauncher/WinLauncher.h:
3478         * WinLauncher/WinLauncher.vcproj: Add new DOMDefaultImpl.h file.
3479
3480 2011-06-17  Chang Shu  <cshu@webkit.org>
3481
3482         Reviewed by Andreas Kling.
3483
3484         [Qt] [WK2] Support undo/redo in MiniBrowser
3485         https://bugs.webkit.org/show_bug.cgi?id=62809
3486
3487         * MiniBrowser/qt/BrowserWindow.cpp:
3488         (BrowserWindow::BrowserWindow):
3489
3490 2011-06-17  Daniel Bates  <dbates@rim.com>
3491
3492         Rubber-stamped by Eric Seidel.
3493
3494         Remove run-javascriptcore-tests --skip-build option; Instead add --build/no-build options
3495         https://bugs.webkit.org/show_bug.cgi?id=62767
3496
3497         Part 2 of 3
3498
3499         Update the BuildBot master.cfg to use the new --no-build run-javascriptcore-tests
3500         command line option instead of --skip-build.
3501
3502         We'll remove --skip-build from run-javascriptcore-test in a follow up patch once the
3503         master buildbot picks up the change.
3504
3505         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3506
3507 2011-06-17  Daniel Bates  <dbates@rim.com>
3508
3509         Rubber-stamped by Eric Seidel.
3510
3511         Remove run-javascriptcore-tests --skip-build option; Instead add --build/no-build options
3512         https://bugs.webkit.org/show_bug.cgi?id=62767
3513
3514         Part 1 of 3
3515
3516         Towards renaming the option --skip-build to --no-build. Add --build/--no-build options to
3517         run-javascriptcore-tests. Also, add a remark about these options to the usage message
3518         printed by run-javascriptcore-tests.
3519
3520         We'll remove --skip-build in a follow up patch once the master buildbot picks up the changes.
3521
3522         * Scripts/run-javascriptcore-tests: Additionally, remove unnecessary initialization
3523           of $showHelp.
3524
3525 2011-06-17  Joone Hur  <joone.hur@collabora.co.uk>
3526
3527         Reviewed by Martin Robinson.
3528
3529         [GTK] Replace GdkRectangle by cairo_rectangle_int_t
3530         https://bugs.webkit.org/show_bug.cgi?id=60687
3531
3532         Replace GdkRectangle by cairo_rectangle_int_t.
3533
3534         * DumpRenderTree/gtk/PixelDumpSupportGtk.cpp: Replace GdkRectangle by cairo_rectangle_int_t.
3535         (createBitmapContextFromWebView):
3536         * DumpRenderTree/gtk/TextInputController.cpp: Ditto.
3537         (firstRectForCharacterRangeCallback):
3538
3539 2011-06-16  Dmitry Lomov  <dslomov@google.com>
3540
3541         Reviewed by David Levin.
3542
3543         https://bugs.webkit.org/show_bug.cgi?id=62603
3544         Run run-api-tests on bots in debug mode.
3545         Release mode is disabled due to https://bugs.webkit.org/show_bug.cgi?id=61812.
3546
3547         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
3548
3549 2011-06-16  Leandro Pereira  <leandro@profusion.mobi>
3550
3551         Reviewed by Eric Seidel.
3552
3553         [EFL] DumpRenderTree: Add GCController, PixelDumpSupport and
3554         WorkQueueItem.
3555         They will be plugged into the build system once the EFL's DRT is
3556         fully upstreamed.
3557         https://bugs.webkit.org/show_bug.cgi?id=61958
3558
3559         * DumpRenderTree/efl: Added.
3560         * DumpRenderTree/efl/GCControllerEfl.cpp: Added.
3561         (GCController::collect):
3562         (GCController::collectOnAlternateThread):
3563         (GCController::getJSObjectCount):
3564         * DumpRenderTree/efl/PixelDumpSupportEfl.cpp: Added.
3565         (createBitmapContextFromWebView):
3566         * DumpRenderTree/efl/WorkQueueItemEfl.cpp: Added.
3567         (LoadItem::invoke):
3568         (LoadHTMLStringItem::invoke):
3569         (ReloadItem::invoke):
3570         (ScriptItem::invoke):
3571         (BackForwardItem::invoke):
3572         * DumpRenderTree/efl/WorkQueueItemEfl.h: Added.
3573         (WorkQueueItem::~WorkQueueItem):
3574         (LoadItem::LoadItem):
3575         (LoadHTMLStringItem::LoadHTMLStringItem):
3576         (ScriptItem::ScriptItem):
3577         (LoadingScriptItem::LoadingScriptItem):
3578         (LoadingScriptItem::invoke):
3579         (NonLoadingScriptItem::NonLoadingScriptItem):
3580         (NonLoadingScriptItem::invoke):
3581         (BackForwardItem::BackForwardItem):
3582         (BackItem::BackItem):
3583         (ForwardItem::ForwardItem):
3584
3585 2011-06-16  Kevin Ollivier  <kevino@theolliviers.com>
3586
3587         [wx]&nbs